2025-01126 SUPPLEMENTAL MATERIAL - Airport South Industrial Annexation (P21-017) (Noticed on 11072025; Published 11072025; Passed for Publication 10212025; Published 10242025) [Updated 11/18/2025 @4:41 PM] - Part 5 — Sacramento City Council, December 2, 2025 (part 2 of 3)

Second, what causes size variations in California red scale? Scale are largest on fruit, of intermediate size on leaves, and smallest on wood. Scale on the twigs (above the fourth internode) are smaller than those on the leaves. High laboratory temperatures (above 27 °C) produce smaller scale (D. Sicki-Yu and R. F. Luck, unpublished data), and our field data on scale size suggest that smaller scale are present during the hottest months. Scale size influences vulnerability to Aphytis melinus and whether
BIOLOGICAL CONTROL OF CALIFORNIA RED SCALE 177
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. the scale receives a female or a male egg. Biological control is less effective on lemon and grapefruit cultivars; we suspect that the difference is associated with scale size and the relative abundance of the substrate supporting the smaller scale. We suspect that these substrates offer red scale a refuge from A. melinus:
the scale are too small to be hosts for female wasp progeny, so the wasps do not spend much time searching these substrates.

Third, how does a parasitoid recognize a scale as a potential host? The parasitoids usually hesitate on encountering a scale, and they drum and turn if the scale is initially accepted (Luck et al., 1982). However, not all apparently suitable individual scale are parasitized. Parasitoids walk over some without hesitating. Host recognition is based, in part, on chemical cues called kairomones— transspecific chemical messengers that favor the recipient, rather than the emitter (Brown et al., 1970)— according to Luck and Uygun (in press) and Quednau and Hubsch (1964). Kairomones and their recognition might also depend partly on the chemistry of the scale's host plant and on the parasitoid's experience, associative learning.

SOURCES OF KNOWLEDGE
Generally Accepted Ecological Facts
Much is known about the natural history of California red scale, its predators, and its parasitoids, especially the parasitoid genus Aphytis (Bodenheimer, 1951; Ebeling, 1959; Rosen and DeBach, 1978, 1979). Less is known about processes that govern the scale's biological control. Information is also available on the red scale's pheromone and its use as a predictive tool (Moreno and Kennett, 1985). All these sources were consulted when this research project was designed and were initially summarized in a population model (Luck et al., 1980). The only body of literature we ignored was that associated with chemical control of the scale.

General Theory and General Principles of Ecology
Four ecological theories provided the structure for our California red scale research— those dealing with predator-prey relations (e.g., Hassell, 1978, 1980), competition (e.g., Pianka, 1981), foraging (e.g., Pyke et al., 1977), and sex allocation (e.g., Charnov, 1982). All these theories are relevant and helped to direct the development of the experimental designs. Predator-prey relations are at the heart of biological control. As stated earlier, biological control is predicated on the notion that host or
BIOLOGICAL CONTROL OF CALIFORNIA RED SCALE 178
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. prey densities are limited by natural enemies, because these agents compete intraspecifically and interspecifically for hosts or prey. Thus, the agents are themselves limited by the densities of their prey or hosts in a reciprocal interaction. However, DeBach and colleagues (DeBach, 1966; DeBach and Sundby, 1963; Rosen and DeBach, 1979) concluded, on the basis of laboratory and field experiments, that the two principal parasitoids responsible for biological control of California red scale were not limited by host density. But if hosts were not limiting for the parasitoids, how did Aphytis melinus control California red scale biologically? We used the principles of foraging theory to identify what scale types the parasitoids chose as hosts, to rank the scale for quality as hosts in the laboratory, and to test this ranking in the field. Because Aphytis allocates sons to small hosts and daughters to large hosts, sex-allocation theory became an extension of foraging theory. Thus, the hypothesis we tested was an amalgam of these principles — that resources are indeed limiting and that the limitation is evinced as a male-biased sex ratio in the field population of A. melinus. As scale grow, they reach the size used by Aphytis to produce sons before they can attain the size used for daughters. Thus, a male-biased sex ratio results from a scarcity of hosts large enough to produce daughters. This pattern might be episodic, depending on the dynamics of the scale's age structure.

A second body of theory used in the design of the experiments concerns the effect of temperature on the developmental rate of insects (Wagner et al., 1984). An insect's rate of development increases with increasing temperature (to a maximum), but the relative duration spent in a developmental stage at a given temperature and the developmental rate of that stage often differ with developmental stage. These effects, coupled with the influence of temperature on reproduction, lead to variations in an insect's age structure as a consequence of temperature variations in the field. We hypothesized that climatically induced variation in the scale's age structure, coupled with the requirement of A. melinus and A. lingnanensis for a specific size or large scale on which to produce daughters, leads to bottlenecks in the availability of suitable hosts and that displacement of A. lingnanensis by A. melinus occurred during such bottlenecks.

Specific Models
First, we used a simulation model of the life history of California red scale and its parasitoids that organized information about these organisms (Luck et al., 1980) and included a modified host-parasitoid model similar to that of Hassell (1978). This model, which embodied the verbal model BIOLOGICAL CONTROL OF CALIFORNIA RED SCALE 179
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. specific to biological control, was inadequate, because biological knowledge was insufficient. Thus, we used the verbal model to predict the field host- parasitoid interaction. This model specifies that natural enemies reduce their host or prey populations to densities at which intraspecific competition and interspecific competition among the predators and parasitoids for prey or hosts reciprocally limit the natural enemies' densities (DeBach, 1974; Huffaker and Laing, 1972). Because it became clear in the course of the research that, from the parasitoid's perspective, the scale was not a homogeneous resource, we used foraging and sex-allocation theory to define, in the laboratory, both host resource requirements for A. melinus and A. lingnanensis progeny production (especially daughters) and the value of a host to a parental female in terms of the quality of the daughters (size = fecundity) that the scale produced. Finally, we used a day-degree model — developmental rate of the scale and parasitoid as a function of temperature (Wagner et al., 1984) — to evaluate the effects of diurnal and seasonal temperature patterns on age structure of California red scale.

Analog Studies
No other evaluations of the effectiveness of a natural enemy appear to have used a similar conceptual approach. Most of the available studies (CIBC, 1971; Clausen, 1978; Greathead, 1976; Greathead et al., 1971; McGuggan and Coppell, 1962; McLeod, 1962; Rao et al., 1971; Turnbull and Chant, 1961; Wilson, 1960) at best simply documented success or failure. Important exceptions were the studies of Wellington (1960, 1964, 1965) and Whitham (1978, 1979, 1980).

Project as Experiment
The California red scale research project was designed to improve the biological control of California red scale and to test conventional wisdom as to the attributes that characterize an effective biological control agent, in this case a parasitoid. Huffaker et al. (1977), Hassell (1978), and Waage and Hassell (1982) proposed the following attributes as characterizing an effective biological control agent: high search rates, ability to aggregate in patches of high host density, close synchrony with the host population, high degree of host specificity, and sufficient reproductive rate to overtake and suppress the host population. Because we knew that A. melinus and A. lingnanensis were successful biological control agents in part of their ranges, we began testing them to see whether they had these attributes and whether A. melinus was better endowed with them.

BIOLOGICAL CONTROL OF CALIFORNIA RED SCALE 180
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

We quickly faced the problem of how to measure a high rate of search or determine whether a parasitoid has a sufficiently high reproductive rate to overtake its host in the field. Moreover, on the basis of a reproductiverate criterion, A. lingnanensis would be judged the better of the two parasitoids, because it is the more fecund (DeBach and Sundby, 1963). Unfortunately, A. lingnanensis was unable to realize its reproductive advantage in the form of daughters, because it was more restrictive in its host (red scale) requirements for daughters than was A. melinus: it required larger scale. Similar problems arose with several other attributes. Thus, after our initial effort to measure some of the attributes for A. lingnanensis and A. melinus in the laboratory (Kfir and Luck, 1979, 1984; Luck et al., 1980; Podoler, 1981), we changed our approach and asked what the specific host requirements were for these two parasitoids with respect to host stage (instar) and size and the pattern of host availability in the field. We used the parasitoids' behavior and the size of the daughters to find the value of a host to the female wasp. On the basis of the results of these laboratory experiments, we are continuing to test the pattern of resource use in the field. We are nearing the end of this phase of our research in late 1985.

Thus, we have used a flexible research strategy of hypothesis formation and testing that incorporates laboratory and field experiments mixed with computer simulation models of aspects of California red scale and parasitoid biology.

Expert Judgment
Clearly, a research project of this sort is not undertaken by a single researcher. First, several colleagues have participated in the evolution and conduct of the research. Rami Kfir, Haggai Podoler, Devin Carrol, Dani Blumberg, Dicky Sicki-Yu, Daniel Moreno, W. W. Murdoch, Charles Kennett, and John Reeves — all associated with academic or research institutions — have been involved in various aspects of hypothesis formation, experimental design, experimental conduct, analysis, and interpretation. Several persons — Harry Griffith, Jim Stewart, Jim Gorden, and Stanley Warner — with many years of experience in citrus pest management contributed as participants in the research and as critics of our interpretations of the results. Finally, in response to seminars, in conversations with colleagues — especially F. Galis, J. M. M. van Alphen, J. C. van Lentern, P. DeBach, L. E. M. Vet, L. Nunny, and D. R. Strong — or in peer reviews of submitted papers, questions or comments often called attention to an aspect of the problem of which we were unaware or stimulated us to see the problem from a different perspective. Thus, the project represents a collage of contributions from a variety of sources.

BIOLOGICAL CONTROL OF CALIFORNIA RED SCALE 181
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

CONTRIBUTION OF ECOLOGICAL KNOWLEDGE TO

PROJECT RESULTS
Classical biological control has been very much a trial-and-error enterprise. It has been based on the notion that natural enemies regulate phytophagous insect populations and that an effective natural enemy is well adapted to and synchronized with its host or prey, is a good searcher, has a high reproductive rate, and aggregates at denser host or prey patches. The large number of successful biological control projects shows that natural enemies indeed limit the densities of their host or prey populations. Yet, if a measure of a successful theory is its ability to predict the outcome of a particular experiment, then predator-prey theory and biological control theory are inadequate. We cannot predict whether a given introduction will succeed or fail.

More important, we cannot describe a natural enemy's attributes that characterize a successful biological control agent. Thus, predator-prey theory is to a great extent untested and, in its current formulation, not easily falsified.

Therefore, we studied a biological control project in which both successful and unsuccessful parasitoids have been introduced, so that we could test several hypotheses arising from predator-prey theory and biological control theory and compare successful and unsuccessful parasitoids. Theory played a role both in formulating the hypotheses and in specifying the type of project to look at.

The California red scale biological control project was chosen for several reasons: the natural histories of the organisms involved were well known, the history of the project was well documented, the scale is controlled biologically in only a part of its California range, the scale and its natural enemies are easily reared in the laboratory, and the parasitoids responsible for the biological control of the scale are known. Initially, our experimental designs were guided by the Nicholson-Bailey (1935) model as modified by Hassell and Varley (Hassell, 1978). Although we obtained a modeled host-parasitoid interaction that persisted, it provided little insight into why A. melinus was so successful.

The characteristics that defined the host-parasite interaction could not be measured in the field. Optimal-foraging theory became an alternative when evidence emerged that A. melinus and A. lingnanensis did not use the same host stages for offspring, as DeBach and co-workers had asserted (DeBach, 1966; DeBach and Sundby, 1963; Rosen and DeBach, 1979). Optimal-foraging theory provides ways of ranking hosts based on their value to the parasitoid, and we were able to use the size of the female progeny arising from a host as a surrogate measure of host quality. Thus, foraging theory played a role in reformulating the questions. More important, it led us to stop thinking of the BIOLOGICAL CONTROL OF CALIFORNIA RED SCALE 182
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. host-parasitoid interaction as one in which hosts and parasitoids are homogeneous units (i.e., the old predator-prey theory) and to start thinking of hosts as resources of different value.

ACKNOWLEDGMENTS
The work described in this study was supported by grants from the National Science Foundation (BSR 84-03394) and the Binational Agricultural Research and Development Fund (1-138-79).

REFERENCES
Abdelrahman, I. 1974. Studies in the ovipositional behavior and control of sex in Aphytis melinus DeBach, a parasite of California red scale, Aonidiella aurantii (Mask.). Aust. J. Zool. 22:231-247 .

Bedford, E. C. G., and M. B. Georgala. 1978. Citrus pests in the Republic of South Africa. Sci. Bull.

Dept. Agric. Tech. Ser., Rep. S. Afr. 391:109-118 .

Bodenheimer, F. S. 1951. Citrus Entomology in the Middle East. W. Junk, The Hague.

Brown, W. L., Jr., T. Eisner, and R. H. Whittaker. 1970. Allomones and kairomones: Transspecific chemical messengers. BioScience 20:21-22 .

Charnov, E. L. 1982. The Theory of Sex Allocation. Princeton University Press, Princeton, N.J.

Charnov, E. L., R. L. Los-den Hartogh, W. T. Jones, and J. van den Assem. 1981. Sex ratio evolution in a variable environment. Nature 289:27-33.

Chewyreuv, I. 1913. Le role des femelles dans la determination du sexe et leur descendance dans le groupe des Ichneumonides. C. R. Soc. Biol. Paris 74:695-699 .

CIBC (Commonwealth Institute of Biological Control). 1971. Biological Control Programmes Against Insects and Weeds in Canada, 1959-1968. Technical Communication 4.

Commonwealth Agricultural Bureaux, Farnham Royal, Slough, Eng.

Clausen, C. P. 1939. The effects of host size upon the sex ratio of hymenopterous parasites and its relation to methods of rearing and colonization. J. N.Y. Entomol. Soc. 47:1-9.

Clausen, C. P. 1978. Introduced Parasites and Predators of Arthropod Pests and Weeds: A World Review. USDA/ARS Agric. Handbk. 480. Agricultural Research Service, U.S. Department of Agriculture, Washington, D.C.

DeBach, P. 1965. Some biological and ecological phenomena associated with colonizing entomophagous insects. Pp. 287-303 in H. G. Baker and G. L. Stebbins, eds. The Genetics of Colonizing Species. Academic Press, New York.

DeBach, P. 1966. The competitive displacement and coexistence principles. Annu. Rev. Entomol. 11:184-212 .

DeBach, P. 1974. Biological Control by Natural Enemies. Cambridge University Press, New York.

DeBach, P., and L. C. Argyriou. 1967. The colonization and success in Greece of some imported species (Hym., Aphelinidae) parasitic on citrus scale insects (Hom., Diaspididae).

Entomophaga 12:325-342 .

DeBach, P., and P. Sisojevic. 1960. Some effects of temperature and competition on the BIOLOGICAL CONTROL OF CALIFORNIA RED SCALE 183
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. distribution and relative abundance of Aphytis lingnanensis and A. chrysomphali (Hymenoptera: Aphelinidae). Ecology 41:153-160.

DeBach, P., and R. A. Sundby. 1963. Competitive displacement between ecological homologues.

Hilgardia 34:105-166.

Dittrich, V. P., P. Streibert, and P. A. Bathe. 1974. An old case reopened: Mite stimulation by insecticide residues. Environ. Entomol. 3:534-539 .

Ebeling, W. 1959. Subtropical Fruit Pests. University of California Division of Agricultural Science, Berkeley.

Flaherty, D. L., J. E. Pehrson, and C. E. Kennett. 1973. Citrus pest management studies in Tulare County. Calif. Agric. Nov., pp. 3-7.

Furness, G. O., G. A. Buchanan, R. S. Georgie, and N. L. Richardson. 1983. A history of the biological and integrated control of red scale, Aonidiella aurantii, on citrus in the lower Murry Valley of Australia. Entomophaga 28:199-212.

Georghiou, G. P., and C. E. Taylor. 1977. Pesticide resistance as an evolutionary phenomenon.

Proc. Int. Congr. Entomol. 15:759-792 .

Greathead, D. J., ed. 1976. A Review of Biological Control in Western and Southern Europe.

Technical Communication 7. Commonwealth Agricultural Bureaux, Farnham Royal, Slough, Eng.

Greathead, D. J., J. F. G. Lionnet, N. Lodos, and J. A. Whellan. 1971. A Review of Biological Control in the Ethiopian Region. Technical Communication 5. Commonwealth Agricultural Bureaux, Farnham Royal, Slough, Eng.

Hassell, M. P. 1978. The Dynamics of Arthropod Predator-Prey Systems. Princeton University Press, Princeton, N.J.

Hassell, M. P. 1980. Foraging strategies, population models and biological control: A case study. J.

Anim. Ecol. 49:603-628.

Huffaker, C. B., and J. E. Laing. 1972. ''Competitive displacement" without a shortage of resources? Res. Pop. Ecol. 14:1-17.

Huffaker, C. B., R. F. Luck, and P. S. Messenger. 1977. The ecological basis of biological control.

Proc. Int. Congr. Entomol. 15:560-586.

Jones, V. P., and M. P. Parrella. 1984. The sublethal effects of selected insecticides on life table parameters of Panonychus citri (Acari: Tetranychidae). Can. Entomol. 76:1178-1180.

Kfir, R., and R. F. Luck. 1979. Effects of constant and variable temperature extremes on sex ratio and progeny production by Aphytis melinus and A. lingnanensis (Hymenoptera:

Aphelinidae). Ecol. Entomol. 4:335-344.

Kfir, R., and R. F. Luck. 1984. Effects of temperature and relative humidity on the developmental rate and adult life span of three Aphytis species (Hym., Aphelinidae) parasitizing California red scale. Z. Angew. Entomol. 97:314-320.

Luck, R. F. 1984. Integrated pest management in California citrus. Proc. Int. Soc. Citricul. 2:630-635.

Luck, R. F., and H. Podoler. 1985. The potential role of host size in the competitive exclusion of Aphytis lingnanensis by A. melinus. Ecology 66:904-913.

Luck, R. F., and N. Uygun. In press. Host recognition and selection by Aphytis species: Response to California red, Aonidiella aurantii (Mask.), oleander, Aspidiotus nerii Bouche, and cactus, Diaspis echinocacti (Bouche) scale cover extracts. Entomol. Exp. Appl.

Luck, R. F., R. van den Bosch, and R. Garcia. 1977. Chemical insect control — A troubled pest management strategy. BioScience 27:606-611.

Luck, R. F., J. C. Allen, and D. Baasch. 1980. A systems approach to research and decision making in the citrus ecosystem. Pp. 366-395 in C. B. Huffaker, ed. New Technology of Pest Control. John Wiley & Sons, New York.

BIOLOGICAL CONTROL OF CALIFORNIA RED SCALE 184
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Luck, R. F., H. Podoler, and R. Kfir. 1982. Host selection and egg allocation behaviour by Aphytis melinus and A. lingnanensis: A comparison of two facultatively gregarious parasitoids.

Ecol. Entomol. 7:397-408.

Maggi, V. L., and T. F. Leigh. 1983. Fecundity response of the two-spotted spider mite to cotton treated with methyl parathion or phosphoric acid. J. Econ. Entomol. 75:616-619 .

McGuggan, B. M., and H. C. Coppell. 1962. Biological control of forest insects, 19101958. Pp. 35-127 in A Review of the Biological Control Attempts Against Insects and Weeds in Canada. Technical Communication 2. Commonwealth Agricultural Bureaux, Farnham Royal, Slough, Eng.

McLeod, J. H. 1962. Biological control of pests of crops, fruit trees, ornamentals and weeds in Canada up to 1959. Pp. 1-33 in A Review of Biological Control Attempts Against Insects and Weeds in Canada. Technical Communication 2. Commonwealth Agricultural Bureaux, Farnham Royal, Slough, Eng.

Moreno, D., and C. E. Kennett. 1985. Predictive year-end California red scale (Homoptera:

Diaspididae) orange fruit infestation based on male trap catches in the San Joaquin Valley, California. J. Econ. Entomol. 78:1-9.

Nicholson, A. J., and V. A. Bailey. 1935. The balance of animal populations. Proc. Zool. Soc. Lond. 3:551-598 .

Orphanides, G. M. 1984. Competition displacement between Aphytis spp. [Hym.: Aphelinidae] parasites of the California red scale in Cyprus. Entomophaga 29:275-281 .

Pianka, E. R. 1981. Competition and niche theory. Pp. 167-196 in R. M. May, ed. Theoretical Ecology: Principles and Applications. Sinauer Associates, Sunderland, Mass.

Podoler, H. 1981. Effects of variable temperatures on responses of Aphytis melinus and A. lingnanensis to host density. Phytoparasitica 9:179-190.

Pyke, G. H., H. R. Pulliam, and E. L. Charnov. 1977. Optimal foraging: A selective review of theory and tests. Q. Rev. Biol. 52:137-154.

Quayle, H. J. 1938. Insects of Citrus and Other Subtropical Fruits. Comstock Publ. Co., Ithaca, N.Y.

Quednau, F. W., and H. M. Hubsch. 1964. Factors influencing the host-finding and host acceptance pattern in some Aphytis species (Hymenoptera: Aphelinidae). S. Afr. J. Agric. Sci. 7:543-554.

Rao, V. P., M. A. Ghani, T. Sankaran, and K. C. Mathur. 1971. A Review of the Biological Control of Insects and Other Pests in South-east Asia and the Pacific Region. Commonw. Inst.

Biol. Control Tech. Comm. 6. Commonwealth Agricultural Bureaux, Farnham Royal, Slough, Eng.

Rosen, D., and P. DeBach. 1978. Diaspididae. Pp. 78-128 in C. P. Clausen, ed. Introduced Parasites and Predators of Arthropod Pests and Weeds: A World Review. USDA/ARS Agriculture Handbook 480. Agricultural Research Service, U.S. Department of Agriculture, Washington, D.C.

Rosen, D., and P. DeBach. 1979. Species of Aphytis of the World (Hymenoptera: Aphelinidae). W.

Junk, The Hague.

Turnbull, A. L., and D. A. Chant. 1961. The practice and theory of biological control of insects in Canada. Can. J. Zool. 39:697-753 .

Waage, J. K., and M. P. Hassell. 1982. Parasitoids as biological control agents — A fundamental approach. Parasitology 84:241-268.

Wagner, T. L., H. Wu, P. J. H. Sharpe, and R. N. Coulson. 1984. Modeling distributions of insect development time: A literature review and application of the Weibull function. Ann.

Entomol. Soc. Am. 77:475-483 .

BIOLOGICAL CONTROL OF CALIFORNIA RED SCALE 185
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Wellington, W. G. 1960. Qualitative changes in natural populations during changes in abundance.

Can. J. Zool. 38:289-314 .

Wellington, W. G. 1964. Qualitative changes in populations in unstable environments. Can.

Entomol. 96:436-451 .

Wellington, W. G. 1965. Some maternal influences on progeny quality in the western tent caterpillar, Malacosoma pluviale (Dyar). Can Entomol. 97:1-14 .

Whitham, T. G. 1978. Habitat selection by Pemphigus aphids in response to resource limitation and competition. Ecology 59:1164-1176 .

Whitham, T. G. 1979. Territorial behaviour of Pemphigus gall aphids. Nature 279:324-325 .

Whitham, T. G. 1980. The theory of habitat selection: Examined and extended using Pemphigus aphids. Am. Nat. 115:449-466 .

Wilson, F. 1960. A Review of the Biological Control of Insects and Weeds in Australia and Australian New Guinea. Commonw. Inst. Biol. Control Tech. Comm. 1. Commonwealth Agricultural Bureaux, Farnham Royal, Slough, Eng.

Committee Comment
The first point that emerges from the California red scale case is not treated in detail by Luck — that, even before the study that he describes was initiated, important partial control of red scale was achieved at little cost and with little attention to ecological theory (Clausen, 1978; DeBach, 1974). The key new knowledge deployed was taxonomic — until 1947, virtually all Aphytis reared overseas from California red scale were thought to be A. chrysomphali.

Because this species had already been introduced and had usually not exerted very good control (partly because of pesticides and ants), specimens attributed to A. lingnanensis were not sent to the United States. Thus, in 1905-1909, a species that was probably A. lingnanensis was discovered in southern China, but not shipped. With the recognition of distinct species, A. lingnanensis was introduced in 1947 and very quickly exerted important control. The ecological knowledge used later to increase this control was rudimentary autecology. It appeared that A. lingnanensis was hindered by the harsh climate of interior southern California, so a search was initiated for other parasitoids in areas with similar climate. Success was achieved in northwestern India and Pakistan with the discovery of A. melinus. This approach is typical of biological control efforts; the first consideration in the search for natural enemies is usually that the climate of their native region approximate the climate where control is sought.

Luck's description is of the effort to extend control to the desert and the San Joaquin Valley, and his thrust is an interesting mixture of classical biological control (much like the introduction of the Aphytis species that are already present) and academic evolutionary ecology. One could simply BIOLOGICAL CONTROL OF CALIFORNIA RED SCALE 186
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. have proceeded by trying more and more introductions from many parts of the world, perhaps focusing the search by looking for areas with citrus and with climates very similar to those of the desert and the San Joaquin Valley and hoping that some species would prove effective where the 52 already tried had not. Instead, Luck tried to understand why control operates differently in different regions and why A. melinus is replacing A. lingnanensis, despite the greater fecundity of the latter. The goal of this understanding is either to predict better what sort of parasitoid would succeed where A. melinus and A. lingnanensis have failed or to modify cultivation or other conditions so as to increase the effects of the Aphytis that are already present.

A number of apparently relevant ecological theories were examined and incorporated into this effort. For the most part, these proved inspirational or didactic, rather than directly applicable. The key to the entire biological control rationale is that the host (red scale) will ultimately be controlled by the parasitoids and will thus come to be scarce, so that parasitoid populations will ultimately be limited by intraspecific or interspecific competition; this is an entomological version of the "green earth" hypothesis of Hairston, Smith, and Slobodkin (1960). Luck observes that the degree of resolution afforded by this hypothesis did not allow answers to the most pertinent questions. He thus proceeded to the more specific traditional predation models of Nicholson and Bailey (1935) and Hassell and Varley (Hassell, 1978), but found that the model parameters were ambiguous and unmeasurable and so could not help to explain the geographic variation in parasitism rate or the apparent replacement of one parasitoid by another. The key to a more useful approach seems to have been the recognition that red scale individuals are not a uniform resource for the wasps, which led rather naturally to casting the problem in terms of optimal foraging. Although no particular existing optimal-foraging model was applicable, simply the formulation of the problem in these terms led to major insights on both geographic variation in parasitism and the superiority of A. melinus over A. lingnanensis. It is also noteworthy that Luck's familiarity with sex-allocation theory led him to examine sex ratio as an indicator of degree of competition for large hosts.

In sum, idiosyncrasies of key species in this study rendered even the most detailed theories inapplicable, but the theories, particularly optimal foraging, suggested new and fruitful ways of looking at the problem. Without the massive information on the biology of the organisms, one could not even have begun the study, but without familiarity with the theory, the beginnings of a solution might have been much longer in coming.

The entire project is conceived not only as an attempt to solve a particular BIOLOGICAL CONTROL OF CALIFORNIA RED SCALE 187
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. environmental problem, but as an experiment to test hypotheses generated by various ecological theories and perhaps to aid in modifying the theories or hypotheses. For example, the intersection of optimal-foraging theory, competition theory, and sex-allocation theory in the explanation of why A. melinus outcompetes A. lingnanensis even when the latter would be successful by itself is more than just interesting entomology — it is an exciting theoretical advance that should spur further theoretical research. A quantitative generalization of this intersection does not exist, and the attention of theorists should be drawn to the problem. Whether the theoretical advance can ultimately help to solve the practical environmental problem — noninsecticidal control of red scale in the desert and the San Joaquin Valley — remains to be seen.

However, the entire study is certainly an advance over the trial-and-error approach of much classical biological control, and the focus on precise mechanisms of species interaction seems to lend itself to more theoretical treatment.

It is interesting that the scoping of this project, as for all biological control (and, in fact, for most of applied ecology), is somewhat myopic. The myriad effects of introduced species are well documented in the ecological literature (e.g., Elton, 1958; Simberloff, 1981), and biological control rests fundamentally on introducing species, such as the Aphytis wasps used to control red scale.

Although attention is paid to more than just the effects on red scale, it is restricted to effects that impinge on agricultural economy. For example, one wishes to avoid introducing a parasitoid or predator that attacks a beneficial species, such as a pollinator, and one screens potential imports to ensure that they are not carrying hyperparasites that would interfere with the parasites already active in controlling agricultural pests. However, little attention is paid to the possible effects of an introduced species on native species that are not agriculturally important.

Howarth (1983), for example, laments the reduction of native Hawaiian lepidopterans, partially by species introduced for biological control. He calls for a more narrowly focused release effort, rather than the hit-and-miss release of many potentially beneficial species that typifies some biological control efforts.

The direction that the red scale study has taken certainly qualifies as a desirable narrowing of focus. Howarth also calls for more studious consideration of potential effects of control agents on nontarget species, just as one seeks such effects for pesticides. The three Aphytis species of most interest in red scale control— A. chrysomphali, A. lingnanensis, and A. melinus— all attack other insects; but, as far as is known, all the other hosts are themselves agricultural pests, such as yellow scale and dictyospermum scale (Clausen, 1978).

Consequently, unexpected detrimental effects of their introduction are unlikely.

However,
BIOLOGICAL CONTROL OF CALIFORNIA RED SCALE 188
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. the fact that a successful introduction is virtually impossible to eradicate (Sailer, 1983) and can easily spread from the site of introduction dictates a particularly broad scoping process.

References
Clausen, C. P. 1978. Introduced Parasites and Predators of Arthropod Pests and Weeds: A World Review. USDA/ARS Agriculture Handbook 480. Agricultural Research Service, U.S.

Department of Agriculture, Washington, D.C.

DeBach, P. 1974. Biological Control by Natural Enemies. Cambridge University Press, New York.

Elton, C. S. 1958. The Ecology of Invasions by Animals and Plants. Methuen, London.

Hairston, N. G., F. E. Smith, and L. B. Slobodkin. 1960. Community structure, population control, and competition. Am. Nat. 94:421-425 .

Hassell, M. P. 1978. The Dynamics of Arthropod Predator-Prey Systems. Princeton University Press, Princeton, N.J.

Howarth, F. G. 1983. Classical biocontrol: Panacea or Pandora's box? Proc. Hawaii. Entomol. Soc. 24:239-244.

Nicholson, A. J., and V. A. Bailey. 1935. The balance of animal populations. Proc. Zool. Soc. Lond. 3:551-598.

Sailer, R. I. 1983. History of insect introductions. Pp. 15-38 in C. Graham and C. Wilson, eds.

Exotic Plant Pests and North American Agriculture. Academic Press, New York.

Simberloff, D. 1981. Community effects of introduced species. Pp. 53-81 in M. H. Nitecki, ed.

Biotic Crises in Ecological and Evolutionary Time. Academic Press, New York.

BIOLOGICAL CONTROL OF CALIFORNIA RED SCALE 189
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. 15 Experimental Control of Malaria in West
Africa
Epidemiology encompasses a sophisticated field of ecological knowledge, and complex mathematical models are available to describe and predict patterns of disease incidence in space and time. The Garki project in Nigeria began in 1970 to provide detailed information on all factors in the transmission of malaria and to test hypotheses about the dynamics of malaria. The project was based on epidemiological models of proven value, it was carefully planned, and it had appropriate controls. It was viewed as a prototype for more extensive malaria control projects elsewhere in Africa.

Although the project was carefully designed and executed in an integrated way over several years and data were gathered on the basis of a realistic mathematical model of malaria transmission, it failed to produce the predicted results, because incorrect assumptions were made about the biology of the mosquito vectors. In particular, it was assumed that all mosquitoes behave alike, whereas resting behavior actually varied in such a way that a large fraction of the mosquitoes did not land inside buildings, where they would have been exposed to the contact insecticide used. This case study highlights the importance of accurate and detailed natural history in the design of effective control strategies and the effect of initial assumptions on the results of using even sophisticated models.

EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 190
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Case Study
ROBERT M. MAY, Department of Biology, Princeton University

INTRODUCTION
Malaria might well be the most important public health problem in tropical Africa. The intensity of malaria transmission is so high that control is very difficult, and many malariologists have suggested that humans should not interfere with the established natural immunity in the population, in as much as intervention could increase the severity of the clinical manifestations of malaria and the mortality caused by it in older children and adults. Largely for that reason— and because the public health infrastructure was inadequate — theory predicted that available programs to control malaria would not work, and Africa was not included in the global malaria eradication program launched by the World Health Organization (WHO) in the 1950s.

However, spraying houses with DDT was locally effective in interrupting malaria transmission, and the results of later pilot projects in various African countries suggested that malaria transmission could be interrupted, particularly in forested areas, if insecticide coverage were total and were followed by full surveillance. But in none of these pilot projects were quantitative epidemiological data collected. In view of this background, WHO decided to begin a field research project, carefully designed to provide information on all the factors involved in malaria maintenance and transmission. The northern part of Nigeria, specifically the Garki district, was selected for this project.

The designers of the project decided to put more resources than usual into collection of baseline data and into evaluation of the impact of spraying houses with an effective residual insecticide (either alone or in combination with mass administration of drugs). As part of this effort to understand the dynamics of malaria transmission, a mathematical model was developed and tested against data. The model was developed in an attempt to simulate malaria transmission realistically, in the hopes of understanding the various factors involved and improving the planning of control programs. The Garki project was also designed to study an array of seroimmunological tests administered before, during, and after the application of control measures.

The project is described in detail in The Garki Project, edited by Molineaux and Gramiccia (1980).

EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 191
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

THE ENVIRONMENTAL PROBLEMS
Malaria is caused by the multiplication of parasitic protozoa of the family Plasmodiidae in the blood cells or other tissues of the vertebrate host; the clinical symptoms in humans arise from multiplication of the blood stages of the parasite. The several genera of malarial parasites are associated with different hosts. Four main species of Plasmodium are found in humans: P. falciparum, P. vivax, P. malariae, and P. ovale (comparatively rare); P. falciparum is the most common in tropical regions and is of greatest concern in the Garki region, and P. vivax is most common in temperate zones.

Infection of a human host begins with the bite of a female anopheline mosquito and the injection of sporozoite stages of the parasite into the bloodstream. These stages are carried to the liver, where they develop in the parenchymal cells. After an incubation period of several days (or months, as some researchers have argued for P. vivax ), these exoerythrocytic stages (i.e., from outside the red cells) grow, divide, and release merozoites into the bloodstream. The merozoites penetrate red cells, where they grow and subdivide to produce more merozoites, which rupture the host cells and invade other red cells. At some ill-understood point in this process, some of the merozoites develop into sexual stages, the gametocytes. Only the gametocytes are infective to the mosquito. When a vector mosquito bites a human and ingests male and female gametocytes, these are freed from the ingested blood cell, and the female gamete is fertilized and develops into an oocyst on the wall of the mosquito's gut. After about 10 days (the development time depends on temperature), immature sporozoites migrate from the ruptured oocyst to the mosquito's salivary glands and mature to infectivity, and the cycle is ready to repeat itself.

Thus, the overall cycle of the malaria parasite involves transmission both from mosquito to human and from human to mosquito. One method of control or eradication of malaria, therefore, is to reduce the mosquito population sufficiently to keep the parasite cycle from maintaining itself. After World War II, DDT was effective in reducing mosquito populations and the incidence of malaria in many parts of the world (more in Asia than in Africa, where transmission rates have always been very high). The pesticide chosen in the Garki project was propoxur, which is more specifically targeted against anopheline mosquitoes than are broad-spectrum insecticides, such as DDT.

Malaria is commonly believed to be responsible for more deaths than any other infectious disease, both today and throughout history. It is listed as the principal cause of morbidity and mortality by most African countries (notably by Nigeria, the most populous). In Latin America, respiratory EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 192
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. tract infection tends to be listed as the most common cause of morbidity and mortality, followed by malaria and then measles; malaria ranks first in Brazil and Colombia. In Asia, gastrointestinal infection is the most common, followed by respiratory infection, with malaria ranking third as a source of morbidity; malaria ranks first in Malaysia. In short, malaria is a very serious public health problem throughout the tropical world (which is largely coincident with the developing world), and it is of especial importance in tropical Africa.

GENERAL APPROACH
The general approach of the Garki project was formulated on the basis of earlier pilot studies in Africa and elsewhere, all of which were qualitative, rather than quantitative. In contrast with the earlier studies, the Garki project had clearly formulated quantitative objectives. The main objectives were as follows (Molineaux and Gramiccia, 1980, p. 21):
"To study the epidemiology of malaria in the lowland rural Sudan savanna. This means in particular a concentration of the study on the measurement of entomological, parasitological and seroimmunological variables and on their relationships. Also included were some meteorological, demographic and clinical variables, and the study of the prevalence of abnormal haemoglobins in the population." To measure the effect of specified activities directed toward interrupting transmission. Such intervention included spraying with a residual insecticide, propoxur, alone and in combination with mass administration of drugs.
To undertake control measures in conjunction with the construction and testing of a mathematical model of malaria transmission. The general aim was to develop such a model into a planning tool that permitted comparison of the expected effects of different control strategies. Specifically, the model was developed to link entomological variables with parasitological variables (particularly the prevalence and transmissibility of P. falciparum ) and to calculate the expected parasitological effect of defined changes in mosquito population density. The model also sought to permit evaluation of the effects of specific schemes for the mass administration of drugs, including the estimated role of immunity.

DEVELOPMENT OF THE PROJECT
Boundaries
In the preparatory phase, from September 1969 through September 1970, preliminary entomological and parasitological surveys were made, EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 193
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. and specific areas and clusters of villages were chosen for detailed study.

Protocols, forms, and operation manuals were drafted, and field and laboratory methods were tested. The boundaries of the study were, in effect, defined as specific villages in the Garki region.

Study Strategy and Monitoring
Drawing on experience gained in earlier, qualitative pilot projects in Africa and elsewhere, the designers of the Garki project produced a study strategy that divided the operation into three main phases of data collection and monitoring.

In the baseline phase, from October 1970 through March 1972 (which included a dry season, a wet season, and a second dry season), baseline epidemiological and parasitological data were collected in the clusters of villages that had been selected for followup. The preliminary insecticide and drug trials were also run.

In the intervention phase, from April 1972 through October 1973 (which included a wet season, a dry season, and a second wet season), the intervention strategies and the monitoring of epidemiological variables in the treated and untreated clusters of villages were implemented. The length of this phase of the project was not fixed in advance, but rather was left open to be guided by accumulated knowledge. In February 1972, a review of the results obtained up to then suggested that the additional information likely to be gained from a third year of intervention and monitoring would not justify the expense and the expected additional loss of population immunity (as discussed in the introduction of this case study).

In the postintervention phase, from November 1973 to the termination of the project in February 1976, selective administration of drugs in the villages covered during the intervention phase was continued, as was epidemiological monitoring in the villages treated with the most intensive strategies and in one cluster of untreated (control) villages.

The antimosquito insecticide propoxur was sprayed, primarily inside buildings, in 164 villages, distributed over roughly 1,000 km 2— a total of
around 30,000 huts that housed a total population of some 50,000 persons.

Propoxur is a contact chemical, so a mosquito must land on a sprayed surface if it is to pick up the insecticide. It remains active against mosquitoes on sprayed surfaces long after application. The coverage with insecticide, expressed as a proportion of huts sprayed, averaged about 99%. This coverage was estimated immediately after each spraying; the true coverage must have been somewhat lower, owing to building and repair activities between rounds of insecticide application.

A combination of the antimalarial drugs sulfalene and pyrimethamine EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 194
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. was given to about 16,000 people in 60 of the sprayed villages. Infants were excluded, but visitors were included. The proportion of the total population given the drugs averaged about 85% (higher in the wet season, lower in the dry season).

USES OF KNOWLEDGE
The relevant ecological facts pertaining to the dynamics of malaria transmission are conveniently discussed under two headings: the mosquito and the malaria parasite. Human immunology, although extensively studied (Molineaux and Gramiccia, 1980), did not yield any findings particularly relevant to the overall design of control programs and is not treated here.

The Mosquito
The extent of contact between mosquitoes and humans in the Sudan savanna is very high indeed: the biting rates of Anopheles gambiae sensu lato and A. funestus were found to attain seasonal peaks of 174 and 94 bites per person per night, respectively (in studies that averaged more than 8 nights). The vectorial capacity, defined as the rate of contact between human hosts via the mosquito vectors, reached a seasonal peak of about 40, which is some 2,000 times the critical value for maintaining endemic malaria. The cumulative inoculation rate by mosquitoes reached a maximum of 145 sporozoite-positive bites per person in 1 year (of which 132 were in the wet season).

The mosquito density, and hence the extent of malaria transmission, in the Sudan savanna varies widely with season, year, and locale, according to Molineaux and Gramiccia (1980), who write:

In most villages, the vectorial capacity drops below its critical level for about half of the year (this does not necessarily prevent transmission, given the large reservoir of parasites), while in some it remains well above the critical level throughout the year. The variations from year to year are relatively important; over a period of 3 years, these variations followed the variations in total rainfall in the case of A. funestus but not in the case of A. gambiae s.l. Villages, even when not obviously different on inspection and located only a few kilometres apart, differ in vector density, anopheline fauna ( A. funestus has a very uneven distribution) and probably in vector behaviour and karyotypes.

Among 8 villages, the cumulative inoculation rate ranged from 18 to 145 sporozoite-positive bites in one year. The local and yearly variations stress the need for adequate baseline and comparison (control) data for a correct evaluation of the impact of control measures.

EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 195
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

The Malaria Parasite
In general, the baseline prevalence of malaria parasitemia was high, and the age-specific curves of both prevalence and density of infection were typical of a high rate of transmission with a high level of acquired immunity. The findings were the same on each of the three parasites ( P. falciparum, P. malariae, and P. ovale). There was relatively little variation among villages or between years.

The rate at which infants acquired malaria and the rate of onset of episodes of patent parasitemia (with fever, etc.) in the general population both testified to the high rate of transmission. The marked increase in the rate at which patent parasitemia disappeared with increasing age and the high ratio (nearly 1:1) of the entomological inoculation rate to the infant conversion rate (the ineffectiveness of sporozoite-positive bites) confirmed the high level of population immunity.

The combination of propoxur spraying and mass administration of sulfalene and pyrimethamine every 10 weeks reduced the prevalence of P. falciparum to a very low point in the dry season, but it did not significantly interrupt malaria transmission. Nor did it prevent an increase in prevalence in the wet season, when natural conditions favored mosquito breeding. A new equilibrium, with incidence oscillating between wet and dry seasons, was reached rapidly, and continuation of the intervention program would probably not have affected this result significantly. It was concluded that mobility of the human population, which was relatively pronounced, was unlikely to be the main cause of the maintenance of transmission. As discussed at length below, the main limiting factor was probably the exophily of some of the mosquitoes — i.e., their preference for resting outdoors.

When propoxur was sprayed and drugs administered more often (every 2 weeks in the wet season, every 10 weeks in the dry season), the prevalence of P. falciparum decreased to around 1% in the dry season and to 5% or less in the wet season. But transmission was still not interrupted greatly. As with the lower- frequency intervention, a new oscillating equilibrium was fairly rapidly attained, and it seems unlikely that continuation of the intervention would have modified the result. Again, it was concluded that exophily of the mosquitoes was the main cause, rather than population mobility.

As expected, mass application of the drugs at high frequencies for 1.5 years caused a temporary loss of immunity against P. falciparum, which was reflected in the resurgence of malaria in the postintervention phase of the project.

EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 196
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

GENERAL THEORY
The basic model for malaria was first set out by Ross (1911, 1916) and later refined by MacDonald (1952, 1957, 1973), who included the latent, infected-but-not-yet-infectious period for the mosquito. If we divide the human host population into those who are susceptible and those who are infected and divide the female mosquito population into the susceptible and the infected, we can construct a pair of differential equations to describe the essentials of the dynamics of infection.
dx/dt = (abM/N ) y (1 – x) – rx, (1)
dy/dt = ax(1 – y) – µy, (2)
where
x is the proportion of the human population infected,
y is the proportion of the female mosquito population infected, N is the size of the human population,
M is the size of the female mosquito population,
a is the rate of biting on a human by a single mosquito (number of bites per unit time),
b is the proportion of infected bites on humans that produce an infection, r is the per capita rate of recovery for humans ( 1/r is the average duration of infection in a human host), and
µ is the per capita mortality rate for female mosquitoes (1/µ is the average lifetime of a mosquito).

In this simplest model, the total populations of both humans and female mosquitoes are assumed to be unchanging ( N and M constant), so the dynamic variables are the proportions infected in the two populations ( x and y). Equation 1 describes changes in the proportion of humans infected. New infections are acquired at a rate that depends on the number of mosquito bites per person per unit time ( aM/N), on the probabilities that the biting mosquito is infected ( y) and that a bitten human is not already infected (1 - x), and on the chance that an uninfected person bitten by an infected mosquito will become infected ( b); and infections are lost by the return of infected people to the uninfected class, at a characteristic recovery rate ( rx). Equation 2 describes changes in the proportion of mosquitoes infected. The gain term is proportional to the number of bites per mosquito per unit time ( a) and to the probabilities that the biting mosquito is uninfected (1 - y) and that the bitten human is infected already ( x); the loss term arises from the death of infected mosquitoes (y ).

EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 197
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

More formally, the loss terms for infected humans and for infected mosquitoes both involve death and recovery. But for human hosts the recovery rate is typically greater than the death rate (by 1-2 orders of magnitude), whereas for mosquitoes the opposite is typically the case; the above formulation is therefore a sensible approximation.

This model is, of course, highly simplified. One of its glaring flaws is the failure to disarticulate the ''infected" categories of human and mosquito hosts to take account of the various developmental stages of the parasite. For instance, the model does not incorporate the incubation period of the parasite in the mosquito (during which no sporozoites are present in the salivary glands of the "infected" mosquitoes), even though this incubation period is comparable with the mean life span of the mosquito. We pursue this complication below.

Likewise, the model does not distinguish between the pathological asexual merozoite blood stages and the infectious gametocyte sexual stages in the human. Encounters between biting mosquitoes and the humans they bite are assumed to be random.

Notwithstanding those shortcomings, the simple model defined by Equations 1 and 2 is useful in laying bare the essentials of the transmission process and in elucidating patterns in the diverse array of epidemiological data on different geographical regions. In particular, the model makes plain the significance of the "basic reproductive rate" of the parasite, in this context conventionally called z
0(MacDonald, 1952, 1957). The basic reproductive rate
is essentially the number of secondary infections generated by one infected individual in a population of susceptibles. If this number is, on the average, less than unity, the disease will be unable to maintain itself; if it equals or exceeds unity, the disease will be able to maintain itself. In general, the larger the basic reproductive rate, the greater the resistance of the disease to eradication. In this simple model,
z0 = ma2b/µr, (3)
where m is the number of female mosquitoes per human host (MIN).

This result is intuitively understandable: the mosquito biting rate, a, enters twice in the cycle (hence, a2); transmission is helped by large numbers of mosquitoes per human host (large m) and by large b; and transmission is hindered by a high mosquito death rate or by fast recovery (large µ and r, respectively).

To be more accurate, we should recognize that a latent period of duration T must elapse before an infected mosquito becomes infectious. This introduces an additional factor, exp(- µ T ), into Equation 3 to represent the probability that an infected mosquito will survive the latent period to
EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 198
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. become a transmitter of infection. The resulting expression for z 0 (as first derived by MacDonald, 1952, 1957) is then
zo = ma2b exp(– µT )/µr. (4)

A SPECIFIC MODEL
Dietz et al. (1974) developed a more realistic modification of the above model, which is presented in detail in Molineaux and Gramiccia (1980, Ch. 10).

The essentials of the model are indicated schematically in Figure 1. There are two classes of people: one class has a low rate of recovery from malaria, and all infections can be detected; the other class has a high recovery rate, and infections have only a 70% chance of being detected (owing to the low densities of parasites). Members of both classes repeatedly are exposed, become infected, and recover, remaining within their own class except for a fixed rate of transition from the relatively susceptible class to the relatively immune class (the transition rate is determined by fitting the model to the data). The model thus takes into account the basic characteristics of immunity to malaria.

Within the framework of this model, it is possible to calculate the prevalence of P. falciparum as a function of the vectorial capacity and of its spontaneous and man-induced changes. It is also possible to calculate the effect of mass administration of drugs on the prevalence of malaria.

The model gave a good account of the basic dynamics of the interactions among humans, mosquitoes, and malaria. But the effects of insecticide FIGURE 1 Schematic illustration of malaria model of Dietz et al. (1974). As discussed more fully in text, diagram shows various categories of individuals considered in model and possible transitions between categories.

EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 199
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. on mosquito longevity were not accurately estimated by treating the mosquito population as homogeneous and as all resting indoors. Accurate prediction of the effects of insecticide on mosquito mortality, and on overall transmission dynamics, required an appreciation of the heterogeneity of the environment, which was not built into the original model.

The study concluded that the specific model described above does indeed simulate the epidemiology of P. falciparum infections with acceptable realism and that it can be used both for planning malaria control and for teaching the epidemiology and control of malaria. But important complications enter into the estimation of the parameters in the model, particularly the estimation of the effects of insecticides on mosquito mortality rates. These complications in many ways constitute the most interesting of the lessons learned in this case.

CONCLUSION: THE CASE STUDY AND ECOLOGICAL

KNOWLEDGE
It will be seen that the Dietz et al. model does not discriminate among different classes of mosquito vectors. A substantial complication, however, arises from inhomogeneities in the effects of insecticides. In the Garki project, an insecticide was applied to interior surfaces of houses and was therefore more effective against mosquitoes that rest indoors after a blood meal. If insecticides do not affect all mosquitoes equally, measurements of the average biting rate and of average longevity might be severely distorted (Molineaux et al., 1979).

Suppose that after spraying there are in effect two mosquito populations: a minority (initially p
1 = 0.2) that rest outdoors (exophilous) and are consequently relatively unaffected by the spraying, and a majority (initially p2 = 0.8) that rest indoors (endophilous) and suffer greater mortality. Let a1 = 1 bite per day and µ1 = 0.05 per day for the exophilic mosquitoes, and a2 = 2 and µ 2 = 0.5 for the endophilic ones. Let the latent period, T, equal 10 days for all mosquitoes. The correct way to calculate the effective index, a exp(– µ T )/µ, which determines the transmission rate from infected mosquitoes to human hosts and is therefore relevant to evaluating the effects of intervention, is to take the appropriate arithmetic average of the separate indexes:
p1a1 exp( – µ1T )/µ1 + p2a2exp( – µ2T)/µ2 = 2.5.

But if the endophilic and exophilic categories are not properly distinguished, then the above index is likely to be estimated with average values of a and of µ: a exp( – µT )/µ = 0.073 (where a = p1a1 + p2a2= 1.8 EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 200
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. and µ = p1µ1 + p2µ2 = 0.41). Molineaux et al. (1979) analyzed the consequences of this aggregation phenomenon in considerable detail and emphasized that aggregating the groups will always underestimate the biting capacity of the mosquito population and hence will always underestimate the basic reproductive rate, z
0, and overestimate the impact of insecticides. As the above example makes clear, these incorrect estimates can lead to seriously wrong conclusions.

The above discussion captures the essentials of the most important surprise found in the Garki project. It was known from the outset that the intensity of transmission of malaria in these parts of Africa was very high. Indeed, the study showed the basic reproductive rate of the infection to be about 1,000; in other words, the transmission rate, or "vectorial capacity of the mosquito," was about 1,000 times the critical value required for the maintenance of endemic malaria.

This very high transmission rate put malaria in tropical Africa in a category of its own and constituted one reason that a control program like that tried in the Garki project was not attempted earlier. Against this background, however, it was not understood that heterogeneities in the environmental setting of the mosquitoes might invalidate simple estimates of the efficacy of insecticide application, which were based on assumptions that the mosquito population was effectively homogeneous and that most mosquitoes rested indoors much of the time.

Thus, it was found that spraying with the residual insecticide propoxur did not have the expected effect on the prevalence of malaria, although coverage was as nearly complete as possible and the insecticide was very effective against the mosquito vectors (it produced a high mortality rate even at the beginning of the third wet season after the last application). Immigration of vectors or humans from unsprayed villages did not appear to be a significant factor. The decisive factor appears to have been the exophily of a substantial fraction of the mosquito vectors. The combination of exophily with a high biting rate maintained a high transmission rate. Thus, although the insecticide program was supplemented by administration of drugs at high frequency and with high coverage and it reduced malaria to a very low incidence, it failed to interrupt transmission.

In more detail, it was found that the effect of residual spraying in reducing A. gambiae s.l. populations, and hence malaria transmission, varied significantly among villages. The variation was probably related not to variations in spraying coverage or to distance from unsprayed villages, but to variation in the amount of exophily. The amount of exophily appeared to be a relatively stable characteristic of particular villages, and it also appeared to be associated with genetic differences within species of the A. gambiae complex.

The vector population attached to a village
EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 201
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. appeared to be relatively isolated genetically most of the time. This was also in accord with the observation that the effect of spraying was influenced little by the size of the sprayed area.

As to the possible genetic mechanisms underlying the behavioral differences between endophilic and exophilic mosquitoes, Molineaux and Gramiccia (1980) stated:

A. gambiae s.l. in the study area is composed of A. gambiae s.s. and A. arabiensis. The dominant species is usually A. arabiensis but the relative abundance of the 2 species varies between times and places in ways which are not explained. A. gambiae s.s. is the more anthropophilic and has higher sporozoite rates; no clear-cut difference was demonstrated regarding exophily or effectiveness of propoxur. The vectorial capacity was estimated as if A. gambiae s.l. were a single species; appropriate simulations show that this is unlikely to have introduced a large error in the estimate. The cytogenetic investigations of Coluzzi suggest that resting behaviour and exposure to propoxur are related less to the relative abundance of the 2 species than to the intraspecific frequency of certain chromosomal inversions, some of which may be associated with a relatively stable behaviour pattern of the individual.

In short, the essential conclusion of the Garki study, which was not anticipated in the early planning, was that,
if the resting behavior of a mosquito species is genetically determined, exophily will be a stable characteristic of individual vectors, and the usual method of interpreting the impact of residual insecticides on longevity, which tacitly assumes uniform behaviour, is overoptimistic.

From the parasitological point of view, the study was interesting, in that its longitudinal nature permitted it to demonstrate that almost everyone was infected early in life, not only by P. falciparum, but very probably also by P. malariae and even by P. ovale, which is commonly regarded as a "rare" parasite. In addition, early demonstration of the effect of parasitism on immunity confirmed that a degree of immunity is evoked by malarial infection and that continual reinfection is necessary to maintain such immunity over the long term. Other clinical studies showed interesting relationships between body temperature and parasitemia and showed a significant effect of malaria control on frequency of fever and on anthropometric indicators of the nutritional status of children.

In another conclusion, Molineaux and Gramiccia (1980) stated that "the new mathematical model, painstakingly tested against hard facts, allows much more realistic simulations of the epidemiology of malaria, both before and after the application of control measures, than was previously possible." Cohen and Singer (1979) developed the model further, incorporating additional elements of realism (such as infection with more than
EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 202
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. one species of malaria and a more accurate description of immune responses in human hosts) while retaining an essential understanding of the process of malaria transmission.

REFERENCES
Aron, J. L., and R. M. May. 1982. The population dynamics of malaria. Pp. 139-179 in R. M.

Anderson, ed. Population Dynamics of Infectious Diseases. Chapman and Hall, London.

Cohen, J. E., and B. Singer. 1979. Malaria in Nigeria: Constrained continuous-time Markov models for discrete-time longitudinal data on human mixed-species infections. Pp. 69-133 in S. A.

Levin, ed. Lectures on Mathematics in the Life Sciences. Vol. 12. Some Mathematical Questions in Biology. American Mathematical Society, Providence, R.I.

Dietz, K., L. Molineaux, and A. Thomas. 1974. A malaria model tested in the African savannah.

Bull. WHO 50:347-357 .

MacDonald, G. 1952. The analysis of equilibrium in malaria. Trop. Dis. Bull. 49:813-828 .

MacDonald, G. 1957. The Epidemiology and Control of Malaria. Oxford University Press, London.

Macdonald, G. 1973. Dynamics of Tropical Diseases. Oxford University Press, London.

Molineaux, L., and G. Gramiccia, eds. 1980. The Garki Project. World Health Organization, Geneva.

Molineaux, L., G. R. Shidrawi, J. L. Clarke, J. R. Boulzaguet, and T. S. Ashkar. 1979. Assessment of insecticidal impact on the malaria mosquito's vectorial capacity, from data on the man- biting rate and age-composition. Bull.,WHO 57:265-274 .

Ross, R. 1911. The Prevention of Malaria. 2nd ed. Murray, London.

Ross, R. 1916. An application of the theory of probabilities to the study of a priori pathometry.

Proc. R. Soc. A 92:204-230.

Committee Comment
In general, the interactions between hosts and parasites are examples of prey-predator associations that have some simplifying features — especially for human host-parasite systems, in which the population size of the human host is usually determined by other factors and much information on transmission and maintenance of the parasite is often available. Many public health studies can be viewed as examples of the interaction of environmental problems (broadly defined) with ecological principles.

The Garki project is notable in that it was very thoughtfully designed, was maintained in an integrated way over several years, and carefully interdigitated the data gathered with a relatively realistic mathematical model. Moreover, the mathematical model, based on hard data and tested against data gathered in the course of the study, could be thought of as a
EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 203
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. paradigm for a prey-predator model. It was successful in explaining the patterns of prevalence before, during, and after the intervention program.

Interestingly, during construction of the basic model, no thought had been given to the possible effects of genetic and behavioral heterogeneity in the vector mosquito population. Such effects turned out to be the most important factor working against control. That a large fraction of the mosquitoes were exophilic meant that simple preliminary estimates of the overall transmission rate and of the likely effect of insecticides were incorrect.

This story illustrates the need for careful design in the gathering of data (and the extent to which a thoughtful mathematical model can guide the process) and shows that, no matter how carefully these things are thought through, unexpected complications are likely. Beyond these platitudes, the story also illustrates a specific theme found in much contemporary ecological theory.

It seems increasingly likely that spatial heterogeneity, with different dynamic processes going on in separate patches, will be the most important factor in the overall persistence of many natural prey-predator or host-parasite associations (Hassell and May, 1985). The Garki project is certainly one example:
behavioral heterogeneity in the mosquito population was, on the one hand, the most important factor in maintaining infection in the presence of an intervention program and, on the other hand, a factor not initially reckoned with (in a preliminary analysis drawn from the conventional traditions of ecological theory, which too often treat the world as homogeneous).

Reference
Hassell, M. P., and R. M. May. 1985. From individual behaviour to population dynamics. Pp. 3-32 in R. Sibly and R. Smith, eds. Behavioural Ecology. Blackwell, Oxford, Eng.

EXPERIMENTAL CONTROL OF MALARIA IN WEST AFRICA 204
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. 16 Protecting Caribou During Hydroelectric
Development in Newfoundland
Single species are often judged to be the valued ecosystem components likely to be adversely affected by construction and operation of a development project. When that happens, ecological information and monitoring studies are oriented specifically toward the biology of the target species. Even if such a target species is well known, however, predicting the impact of the project on it can be difficult, and the project might have to proceed on a contingency basis, taking cautionary measures and altering them if effects on the target species are different from those anticipated. The effects of the Upper Salmon Hydroelectric Development on woodland caribou constitute an example of this type of problem.

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
205
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Case Study
DAVID J. KIELL and EDWARD L. HILL, Environmental Services Department, Newfoundland & Labrador Hydro, St. John's, Newfoundland SHANE P. MAHONEY, Newfoundland Wildlife Division, Department of Tourism, Recreation and Youth, St. John's, Newfoundland

INTRODUCTION
An analysis completed in 1977 of projected power requirements for insular Newfoundland, compared with installed generating capacity, suggested that a shortfall would occur by about the end of 1982. A number of alternatives were available to meet this shortfall, but preliminary information suggested that the Upper Salmon Hydroelectric Development (USD) was the most economical option.

The USD is in south-central Newfoundland, approximately 170 km southwest of Gander and 50 km northwest of St. Albans, on Bay d'Espoir (Figure 1 ). It exploits a portion of the available head between Meelpaeg Reservoir and Round Pond, which are elements of the existing Bay d'Espoir hydroelectric project.

Before the USD, water flowed from Meelpaeg Reservoir through Crooked Lake and Great Burnt Lake down the North Salmon River to Round Pond (Figure 1). Water from Cold Spring Pond flowed down the West Salmon River to Godaleich Pond and then into Round Pond.

A dam on the North Salmon River now diverts water from Great Burnt Lake through two diversion canals into Cold Spring Pond. A dam on the West Salmon River creates a reservoir at Cold Spring Pond. Water from this reservoir is transmitted to the intake structure via a power canal 3.6 km long. A penstock, approximately 455 m long, delivers water to the generating station on the shores of Godaleich Brook, a tributary of Godaleich Pond. The lower 1,000 m of Godaleich Brook has been excavated and serves as a tailrace to transport water to Godaleich Pond. A channel improvement was constructed to control water levels in Godaleich Pond.

The powerhouse contains one Francis turbine, which generates 84 MW of power, using a hydraulic head of about 51 m.

Access to the USD area was achieved by construction of a 50-km road from St. Albans. The 230-kV transmission line from the USD parallels the access road within a corridor 1 km wide for most of its length and terminates at the powerhouse at Bay d'Espoir.

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
206
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

FIGURE 1 The Upper Salmon Hydroelectric Development.

A preliminary environmental analysis was undertaken in 1975 (Airphoto/ Beak, 1976). Detailed engineering and environmental analyses of the USD were not begun until the spring of 1978. A study designed to evaluate the possible impacts of the USD on wildlife was commissioned with other studies on fish, forest and land resources, and socioeconomics.

In November 1978, discussion between Newfoundland and Labrador

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
207
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Hydro (Hydro), the proponent of the project, and the Newfoundland Wildlife Division (NWD) revealed that the potential impact of the USD on woodland caribou (Rangifer tarandus), particularly the large Grey River herd, was of great concern. It was agreed that the wildlife study in the area would not be sufficiently rigorous to permit adequate decisions related to caribou.

The USD received approval in principle from the provincial government in 1979, before the preparation of the environmental impact statement (EIS). At that time, it was generally recognized that the construction and operation of the USD would affect caribou, but the type and magnitude of impact were open to debate. Therefore, Hydro and NWD were faced with the problem of how to measure the impact of the USD on caribou.

THE ENVIRONMENTAL PROBLEMS
Caribou and the Upper Salmon Development
The USD is within the range of the Grey River caribou herd, one of the largest herds in Newfoundland (approximately 5,000 animals). The calving ranges of two smaller herds, the Sandy Lake (approximately 250 caribou) and the Pot Hill (approximately 500), are about 30 and 50 km, respectively, northeast of Godaleich Pond. Although the two smaller herds were included in the impact study, this case study deals exclusively with the Grey River herd.

The Grey River herd winters between the USD and the southern coast of the island. Its migration route to and from its summer range north of Great Burnt Lake is through and around the USD area ( Figure 2). However, caribou can be found within or near the USD area throughout the year.

The Grey River caribou usually calve in the Wolf Lake-Dolland Pond area in late May to early June, but some have been known to calve near the southern end of Cold Spring Pond. Traditionally, these caribou have used the Cold Spring Pond area as part of their postcalving range during the last half of June and early July. This is where the animals strengthen the cow-calf bond and take advantage of emerging vegetation to regain the energy they have lost during the winter, pregnancy, and parturition. The cows require a highly nutritious diet at this season, because of the demands of lactation. After calving, the caribou gradually move north along both sides of Cold Spring Pond. By the middle of July, most have moved away from the upper Salmon area and onto their summer range.

The USD area is not merely a migration corridor through which caribou move as quickly as possible from calving grounds to summer range. It also contains essential grazing range and terrain used to minimize

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
208
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

FIGURE 2 General spring and fall migration routes of the Grey River caribou herd through project area.

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
209

Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. harassment by predators and insects. The apparently erratic movements with reversals of direction and stationary periods (other than those caused by construction activity) are normal in early summer. It is also normal for some animals to spend the summer in the USD area.

In this brief discussion of the movement patterns of the Grey River caribou, specific locations — such as point ''x" where they winter and point "y" through which they migrate — have been avoided. The purpose of the avoidance was to indicate the inherent variability in caribou movements. A general range and general migration routes can be described for a herd, but its precise location at any time is very difficult to predict. An exception to this broad generalization is that caribou are usually faithful to a calving area. A number of factors can account for these relatively unpredictable movements: snow; other weather conditions; vegetation; harassment by predators, insects, or hunters; disturbance by developments, such as roads, pipelines, or hydroelectric projects; and caribou wandering.

During the last decade, much information has been accumulated about the effects of developments on wildlife. Some species, such as moose ( Alces alces), benefit when clearing operations result in the growth of their preferred food items. Others, such as caribou, might be adversely affected. It has been shown that linear developments— such as roads, pipelines, and transmission lines — can block, delay, or deflect caribou movements (Banfield, 1974; Child, 1974; Klein, 1971). The amounts and types of traffic and human activity associated with such developments can affect the reactions of caribou to them (Horejsi, 1981; McCourt et al., 1974). Their responses also vary with group size, sex, and age.

Females accompanied by young exhibit the strongest reactions (Cameron et al., 1979). The effect of blocking, delaying, or deflecting caribou movements might be to reduce or eliminate the use of part of their range. These effects might not become apparent for years.

In addition to the physical barrier effect, harassment can result in decreased rates of growth, development, and reproduction (Geist, 1975).

Injuries and death from accidents could occur, as could avoidance of areas, which would lead to loss of important resources and ultimately a population decline. Increases in human access itself can also have an impact (Bergerud, 1979), for example, in the form of increased hunting pressure.

The sensitivity of caribou to development is debatable. Bergerud (1974) stated that caribou "seem to be both highly adapted and highly adaptable. " He observed that caribou in Newfoundland do not have an aversion to roads or railroads themselves, but that traffic is a stimulus for flight. Banfield (1974) noted that caribou "are quick" to use cleared seismic lines for travel routes.

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
210
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Ecological Questions and Issues
Ecological questions needed to be answered before the impact of the USD on caribou could be determined, including the following:
Will the migration and distribution patterns of the herd change? Will calf production and mortality change?
Will the age and sex ratios of the herd change?
Will there be a measurable behavioral reaction to the USD? What will be the effects of increased human access to the previously isolated caribou range?
It was also recognized that measurement of changes was not sufficient.

The reasons for changes and the long-term implications for the herd also needed to be known, particularly if corrective or preventive measures were to be undertaken. Both Hydro and NWD agreed that the key issue was whether statistically significant changes in population characteristics, such as herd size and age and sex ratios, could be attributed to the USD.

There were points of agreement and disagreement about the makeup of an adequate study. It was agreed that the measurement of changes in population characteristics (productivity and age and sex ratios) over time would be the ultimate indicator of the impact of the USD on caribou. Recognized wildlife techniques were available to obtain these measurements.

The most likely magnitude of impacts was debated early in the conceptualization of the study. The design of the study depended on such predictions. If many small impacts were expected, the study would have to be designed to measure them. However, if one or two major events were expected, a study of a different kind would be in order. The methods to be used, the area and duration of the study, and the analytical techniques to be used depended on the decisions reached.

NWD was concerned that construction of the USD would have a detrimental impact on the size of the herds. It argued that the main elements of the USD (the West Salmon dam, the power canal, the intake, and the penstock complex) were perpendicular to the general direction of movement of the Grey River herd from the postcalving area to the summer range. As a result, the USD could act as a barrier to normal migration and distribution. In the worst scenario, the postcalving migration would be blocked, and the herd would not migrate past the USD. If the blockage occurred for several consecutive years, it was feared, the caribou would lose their migratory tradition, and their range would in effect be halved.

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
211
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Presumably, the population would eventually decrease to the carrying capacity of the new range. It was considered more likely, however, that the animals would move through the area and that the stress of confronting the barriers to migration would cause an energy deficit sufficient to reduce survivorship, reproductive success, and ultimately herd size. It was also hypothesized that caribou might migrate around the USD on the west side of Cold Spring Pond and so not use habitat within the USD area. Loss of this habitat might have a major impact on the annual energy budget of at least part of the herd.

Hydro's environmental staff shared Bergerud's (1974) opinion that caribou are relatively adaptable animals. Although there was reason for concern over the possible impacts of the USD on caribou, it was thought that any alterations in migration and distribution would be localized and brief. This opinion was supported by Bergerud (personal communication) when he was given the details of the proposed development. Hydro's position was that any project-related impacts would not have a significant effect on herd productivity. Hydro's biologists continued to argue that, to evaluate the impacts of a hydroelectric project on caribou, a study should emphasize individuals and groups of animals near the development. This approach would permit the identification of interactions that could be used to clarify the causes of impacts, if any, at the population level. It would also facilitate the development of effective mitigation measures for this and future projects.

OBTAINING RELEVANT ECOLOGICAL KNOWLEDGE
Population Characteristics, Migration, and Distribution A study was developed during the winter of 1978-1979 to document the productivity and other population characteristics of the herds. It was not clear whether impacts would be severe enough to manifest themselves at this level, but the ultimate indicator of impact would be a significant decrease in productivity. Therefore, these important characteristics had to be measured.

NWD anticipated a large impact as a result of the physical and behavioral barriers presented by the USD. It suggested that this impact would be reflected in changes in herd migration and distribution patterns. Therefore, NWD proposed a study to document caribou movement and distribution on their range (8,000 km
2) with radiotelemetry. It also proposed an evaluation of the range aimed at understanding and predicting the implications of major changes in migration and distribution patterns.

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
212
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Hydro was critical of some aspects of the proposed studies. There was agreement that productivity and other population characteristics had to be monitored. However, Hydro's biologists argued that impacts would probably be more subtle than major shifts in movement patterns and habitat uses. They suggested that remote sensing techniques, such as radiotelemetry, were too coarse to detect and document impacts and that field work near the USD area was more appropriate.

With some reluctance, Hydro agreed to cofund part of the studies proposed by NWD. The study was undertaken by NWD personnel and continued for 6 years (1979-1985)— through planning, construction, and 2 years of operation of the USD.

Behavior
In 1981, Hydro initiated a study to observe systematically the activities of caribou near the USD (the experimental area) and in control areas. The time devoted by caribou to distinct activities — such as feeding, lying, standing, and running— would be used to prepare activity budgets for animals in the experimental and control areas. Disturbance or harassment would alter activity budgets by increasing the incidence and duration of unproductive activities. It was hypothesized that disturbance of caribou by construction activities would be reflected in their activity budgets, even if major shifts in migration and distribution could not be detected. A similar study design was used to examine the reaction of caribou to the trans-Alaska pipeline and haul road (Roby, 1978).

Observations were made through the summer of 1984, but most intensively in 1981 and 1982, when construction was going on.

Sensitivity Criteria
The environmental impact statement for the USD identified a number of mitigating measures that would facilitate caribou movement nearby. Among these measures was a restriction of blasting to times when fewer than 100 caribou were within 1.6 km of the blast site. The 1.6-km distance was based on a comment by Bergerud (1974) that one Newfoundland herd overwintered about 1 mile (1.6 km) from the trans-Canada highway. The sensitivity criterion of 100 animals was chosen arbitrarily.

As the USD proceeded, it was decided that similar protection was required for other construction activities, such as vehicular traffic on access roads and construction at various civil works areas and at borrow pits and quarries. The criteria varied with the sensitivity of caribou at the time of

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
213
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. the activity— calving (May-June 15), postcalving aggregation (June 15 July 1), and migration (July). More protection was afforded caribou during calving than after calving and during migration, because they are more sensitive to disturbance at that time. The criteria also provided for an early warning system, so that project managers could be alerted if a large number of caribou were moving into an area and construction activities might have to be temporarily curtailed. When caribou were moving through the area, daily counts were taken from a helicopter (whenever weather permitted), to locate animals in reference to USD activities and to determine whether some elements should be shut down to minimize adverse impacts on caribou. NWD approved the criteria before implementation. Brazil (1983) and Northcott (1984) have described how the criteria were developed.

The main purpose of these mitigating measures was to facilitate decision- making with regard to caribou. They were not envisioned as integral parts of the caribou studies. However, as the caribou locations were mapped and data began to accumulate, a valuable picture of caribou movements in and around the USD emerged.

Conclusions
Although the data from the component studies have not all been completely analyzed, some conclusions have been reached. The productivity and the size and age structure of the herds studied have not changed significantly.
The migration and distribution patterns of the Grey River herd appear to have been somewhat affected by construction of the USD. Calving activity was not observed near Cold Spring Pond during construction and by 1984 had not returned to the level reported in 1979. Rutting activity has not been observed in the USD area since 1979. Large numbers of migrating caribou have not been observed during the fall near Godaleich Pond since 1980. The major fall movement appears to occur west of Cold Spring Pond. The postcalving aggregation was concentrated west of the West Salmon dam during construction. Use of the area between the West Salmon dam and Godaleich Pond increased in 1983 and 1984 to form a pattern similar to that observed before construction. Dispersal from the postcalving aggregation appeared to change during construction as animals avoided crossing the North Salmon road.
The maximal number of caribou observed in the USD area during the postcalving aggregation decreased from 2,100 in 1979 and 1980 (before construction) to 1,307 in 1981 and 821 in 1982 (during construction).

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
214
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Numbers increased after completion of construction to 1,777 in 1983 and 2,039 in 1984 (Northcott, in press).
Caribou appear to have become sensitized to construction activities in 1982. The proportion of behavior associated with energy expenditure or a state of disturbance increased significantly from 1981 to 1982, but then decreased in 1983 and 1984. This suggests that the major cause of disturbance was construction activity, rather than the structures (Hill, in press).
The effectiveness of the sensitivity criteria as a mitigating measure has been questioned (Brazil, 1983; Northcott, 1984). Much of the criticism is related to problems associated with determining when the criteria were met and with stopping work. Although it cannot be proved, it can be argued that these measures helped some caribou to cross the USD area. These mitigation techniques require further refinement, if they are to have generic applications.

USES OF ECOLOGICAL KNOWLEDGE AND

UNDERSTANDING
Valued Ecosystem Components
A preliminary environmental analysis of the USD was undertaken in 1975-1976 (Airphoto/Beak, 1976). Only two of the 16 recommendations pertaining to the predesign, design, and construction phases of the USD were related to wildlife:
"Undertake field surveys of wildlife population, habitat, movement, and breeding characteristics. Place special emphasis on the ecology of the Godaleich delta and North Cold Spring Pond region" (Figure 1). "Recognize caribou migration requirements in design and construction of the access road, transmission lines, diversion canals and during reservoir filling, especially between Cold Spring Pond and Great Burnt Lake."
Caribou were not recognized as an ecosystem component of greater importance than many others, even though the study was managed by a government committee that included a representative of NWD. However, in November 1978, NWD voiced its concern that a general wildlife inventory was insufficient. It suggested that studies at the USD concentrate on caribou, for the following ecological and socioeconomic reasons:
Caribou were a species of recreational and commercial importance. Caribou represented the wilderness experience to Newfoundlanders and therefore had aesthetic value.

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
215
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. Caribou had historically been hunted by local Indians and were important to the Indians' heritage.
Known range preferences and patterns of migration and distribution conflicted directly with proposed construction activities. Caribou had been perceived to be very sensitive to human disturbance. Caribou populations had been steadily recovering from serious overhunting in the early 1900s, and it was feared that this trend could be halted or reversed if the USD substantially reduced the Grey River herd.

Therefore, during the fall of 1978, woodland caribou populations came to be seen as a valued ecosystem component. Monitoring studies and mitigating measures were adopted in response to this new concern.

Significance of Impacts
The Newfoundland government approved construction of the USD in the spring of 1979, in spite of possible environmental impacts. Approval was based on the need to prevent the energy shortfall that was projected to occur in 1982.

The government, however, required that all reasonable mitigation to reduce possible impacts on caribou be undertaken and that the cooperative study be undertaken. These requirements were stipulated in the Order-in-Council that authorized Hydro to proceed with the USD. Thus, the political decision regarding the relative significance of the two resources (hydroelectric and caribou) was made very early in the development of the study program.

On a technical level, Hydro and NWD biologists generally agreed that a statistically significant change in herd productivity would demonstrate that an important impact had occurred as a result of the project. This was a "primary" level of significance.

Impacts that might not have a direct effect on herd productivity (e.g., changes in distribution or behavior) were assigned "secondary" significance.

They were important for two main reasons. First, such impacts are often easy to mitigate; for example, vehicular traffic could be curtailed to allow animals to cross access roads. Second, many small impacts could accumulate; therefore, if a statistical decrease in productivity were observed, an analysis of secondary impacts might help to explain it.

Establishment of Boundaries
Establishing spatial boundaries for the studies was not difficult once the scope of the studies was determined. The decision that herd productivity was the ultimate measure of impact required an investigation of

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
216
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. animals in their whole range. Calving and postcalving areas were emphasized because of the vulnerability of animals around the project area. However, some investigations— particularly those regarding predation, distribution, and range quality— were undertaken on the summer and winter ranges.

The temporal boundary for the NWD-Hydro cooperative study was established shortly after the start of the first field season. It was thought that any substantial USD-related impacts would occur during construction, but that it was necessary to continue the study for at least 2 years into the operation phase, to determine whether the animals were becoming accustomed to the presence of the USD within their range.

The main thrust of the behavior study was to monitor changes in activity budgets of individuals and groups of caribou as they confronted the USD. This required that the study concentrate on areas near the USD, although control data were obtained away from the experimental area.

The sensitivity criteria established for the USD specifically defined spatial boundaries. The boundaries were directly related to the location of construction activities and varied with the caribou's life-history stage. In as much as the sensitivity criteria were implemented to make decisions on construction activities, the temporal boundary was defined by the construction schedule of the USD.

Development and Implementation of Study Strategy
Detailed planning of the study program began after two political decisions were taken: that the development would be built and that caribou constituted a valued ecosystem component. Therefore, the program was oriented toward the monitoring of impacts, rather than assisting in making a "go, no-go" decision, which is usually the underlying goal of environmental impact assessment.

Impact prediction was one objective during the first year of the cooperative NWD-Hydro study. However, in following years, the focus was on monitoring impacts.

In this context, Hydro and NWD biologists began developing the required studies. The program's first study was to monitor the population characteristics, migration, and distribution pattern of the herd. The study had to be undertaken in a technically competent manner, within a schedule and within a reasonable budget. Various options were examined, including the use of consulting firms.

However, it was decided that such a study would afford the third author (S. P.

M.) an opportunity to undertake doctoral studies at the University of Calgary.

This provided a valuable

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
217
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. academic orientation to the study and reduced the cost and administrative burden to Hydro, the proponent.

It was also agreed that NWD would manage the study, provide technical assistance from its existing staff, and provide some of the aircraft time. Hydro would pay for the remainder of the flying time and the radiotelemetry equipment and drugs that were needed. Thus, the study developed as a cooperative effort involving the responsible resource agency and the proponent.

Further comments on this relationship have been provided by Mahoney (1983) and Kiell (1984).

As previously discussed, Hydro initiated behavior studies near the USD.

But this was not a unilateral action; NWD supported the initiative, helped to develop the terms of reference for the study, and reviewed proposals received from consultants. During the study, data were shared with NWD, and in the later stages NWD assisted in obtaining some of the necessary control data.

The main expense and responsibility of implementing the sensitivity criteria were borne by the proponent. Helicopter surveys of the USD area, in cooperation with the Newfoundland Department of Environment, provided valuable information on caribou numbers and distribution to both NWD and the consultants conducting the behavior study. Telemetry surveys by NWD assisted in the scheduling of these monitoring flights. Trained observers for the surveys were provided periodically by NWD and the consultants.

The overall strategy was one of cooperation. There were periods of disagreement, but generally this approach was satisfactory for the proponent, NWD, and the graduate student.

Development of Predictions and Hypotheses
The program involved three distinct groups, or interests: Hydro, NWD, and the academic interest, each with a staff of biologists. Therefore, hypothesis- testing, although not always specifically mentioned, was fundamental to the design of the studies. The testable null hypotheses included the following:
The reproductive rates of the herd before and after the USD would not differ.
The age and sex ratios of the herd would not change during the study. The numbers of caribou using the construction area as a postcalving site would not change during the study.
The behavior of caribou encountering the USD area would not change

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
218
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. from one year to another and would not differ between control (undisturbed) situations and disturbed circumstances.
Caribou mortality patterns would not change during the study.

These null hypotheses were tested with accepted wildlife techniques, and the data were analyzed in a statistically rigorous manner.

Monitoring
The primary objective of the program was to monitor the impact of the USD on caribou. The program treated the USD as an experiment. In 1979-1980, control (predevelopment) data were obtained on caribou population biology for comparison with similar data obtained during and immediately after construction. Control data for the behavior study were obtained in the same year as the experimental data, but were collected away from the influence of the USD. These studies were designed so that any changes due to the USD would be detected and statistically verified.

Information from these studies was often used to make USD-related decisions. For example, knowledge of the migration routes of caribou from telemetry data obtained in 1979 was used to determine areas of the reservoirs that should be cleared to facilitate future caribou migration (Kiell, 1981).

Underlying the reservoir-clearing program was the assumption or prediction that caribou would use these areas in future years.

Information from the sensitivity-criteria surveys was provided directly to the decision-makers who planned the daily activities during construction. Data from this study served such a specific use that they were not expected to contribute to the general knowledge of caribou movements or behavior. Hydro in particular saw this study in a very restricted way. In retrospect, slight changes in the design of the program would have yielded more rigorous data and would have added another element to the knowledge of caribou migration, distribution, and behavior.

Cumulative Effects
The study program was designed to examine the impact of the USD on caribou at the individual and population levels. It was necessary to collect data on caribou ecology before, during, and after construction, to evaluate cumulative effects. A short-term study would have been unable to detect or assess cumulative effects.

One of the deficiencies in the program has been the lack of thought regarding the effect of increased human access caused by the USD. It now appears that this may be the most important impact of the develop

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
219
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. ment, because of general disturbance and increased human legal or illegal hunting. The problem has been addressed in policy, but no data from this program were collected to assist in policy development.

SOURCES OF ECOLOGICAL KNOWLEDGE
Ecological Facts
The main product of the program was ecological facts about caribou that could be compared before, during, and after construction of the USD. Well- documented wildlife research techniques were used to obtain information on herd size, sex and age ratios, calf production and survival (productivity), and habitats. Radiotelemetry allowed researchers to determine general migration routes and spatial and temporal distribution patterns. A specifically designed study documented the effect of the USD on caribou behavior.

Before the study program was implemented, information about the biology of the Grey River herd was obtained from published papers, mostly by Bergerud, in the late 1960s and early 1970s (e.g., Bergerud, 1971) and from NWD data files. Studies of barren-ground caribou (northern Canada and Alaska) and of reindeer (Scandinavia) provided some guidance for interpretation of Newfoundland data, although natural history information was not directly applicable. Experiences of local people and wildlife technicians provided a valuable historical perspective during the program.

The study program supplied most of the data needed for decision-making during construction and operation of the USD. These data were used in making decisions, such as on where to clear the reservoir and whether vehicular traffic should be allowed to operate as usual. Specific, high-quality, timely information was needed for these decisions; management data obtained by NWD in the 1960s were not adequate.

Time and budget constraints did not preclude studying some ecological elements, but were important in determining the depth of studies and precision of the data. For example, studies on the winter range were minimized by logistics.

Theory and General Principles
The theoretical framework of basic animal ecology and ethology formed the foundation of the studies. Population dynamics and bioenergetics underlay the approach to the program. It was felt that, if there were significant impacts from the USD, they would affect the energy budget of the caribou and ultimately be reflected in changed reproductive and mortality rates.

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
220
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Specific Models
Airphoto/Beak (1976) reviewed available data on caribou and developed an overview of caribou migration and distribution patterns around the USD. On the basis of data from the first year of the study, Mahoney (1980) improved this overview to depict the general boundaries of calving and postcalving areas, migration corridors, and summer and winter ranges. A similar picture was produced for each later year to help organize data and to see whether there were marked differences between years. Figure 2 is an example of the types of information summarized in these descriptions.

Analog Studies
The development of oil resources in the Canadian and American North in the last decade required construction of pipelines to markets in the South or to tanker ports. Many studies have been undertaken to investigate the impact of pipelines, vehicular traffic, compressor noise, helicopter harassment, etc., on caribou migration, e.g., those of Banfield (1974), Roby (1978), Cameron et al. (1979), Miller and Gunn (1979), Hanson (1981), Horejsi (1981), and Smith and Cameron (1983). These studies were useful to the program at the USD, because they highlighted the issues that needed attention, provided suggestions on study techniques, assisted in formulating study design, and generated a group of experienced resource people with whom to discuss ideas.

Pilot Studies
The speed with which decisions were made and studies initiated precluded the use of pilot studies. If time had allowed, pilot studies to check the effectiveness of radiotelemetry equipment might have been useful.

Development of receiving equipment, the use of radio collars, and calibration of equipment took place as the program proceeded.

The Project as an Experiment
The considerable flexibility in this program was exercised when there was a compelling reason to do so. For example, as the program developed, collars were installed almost exclusively on calves. Calves were considered to be most vulnerable to potential disturbance, and their migration and distribution patterns were considered to reflect those of cows. It was thought that more and better data would be obtained from this approach.

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
221
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

The evolution of the caribou sensitivity criteria also illustrates how activities were modified according to need. Originally, the criteria applied only to blasting (Newfoundland and Labrador Hydro, 1980). As construction activities multiplied, the criteria were modified to include all project activities and to be cognizant of the differing degrees of sensitivity to disturbance expected during the calving, postcalving, and migratory periods. This flexibility worked to protect the caribou around the USD.

Expert Judgment
Table 1 summarizes the expertise used and the role of the various ecologists in the program. The program was being funded and administered mainly by NWD and Hydro (both of which agencies had environmental staffs), so the need for purely academic input was identified at an early stage. This group of ecologists helped to design the studies, provided expert advice on request, and reviewed study components periodically to ensure a high degree of scientific integrity. Two scientists, D. R. Klein and E. Reimers, moved to the USD from Alaska and Norway, respectively, to gain firsthand experience with the situation. The technical input from these experts was valuable, and the confidence that they instilled in study administrators and less experienced biologists was very important.

CONTRIBUTION OF ECOLOGICAL KNOWLEDGE
This case had three attributes that contribute to its usefulness: scientific methods were applied in designing the program, sound environmental management strategies were adopted in scoping the studies and organizing their implementation, and ecological knowledge was used to address the problems at hand.

The main features of the scientific method are the framing of testable hypotheses and the design of studies in such a way that statistical confidence can be attached to the results through spatial and temporal control and bounding of observations. These elements were aspects of the study described here.

All environmental studies that contribute to decision-making are constrained. Prudent management of time, funds, and staff resources is essential for success. In this case, the decision by NWD and Hydro to undertake the studies in-house with the assistance of selected outside experts (e.g., academic advisors and consultants) was a major step toward maintaining scientific flexibility and obtaining the most from the resources available.

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
222
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Finally, it was recognized early in the program that an ecological approach was needed to answer the complex questions about the interaction of caribou with a hydroelectric development. Information of different types was obtained from various sources. Personal communication with scientists knowledgeable about the biology of caribou in Newfoundland and about the interaction between caribou and industrial development elsewhere provided an ecological perspective. Data in the primary and ''gray" literature provided a relatively good model of caribou distribution and migration patterns in and around the USD and identified the boundaries of important habitats, but did not provide much insight into the effects of building a hydroelectric project in their range. Original research with

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
223
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. standard wildlife management techniques (population census, aging, and classification techniques), modern methods (radiotelemetry), and behavior analysis (time and energy budgets) generated specific ecological data on the animals. These data were interpreted in the context of experiences elsewhere (analog studies) and ecological principles, such as population dynamics and energetics.

REFERENCES
Airphoto/Beak (Airphoto Analysis Associates Consultants Ltd. and Beak Consultants Ltd.). 1976.

Upper Salmon/Cat Arm: Environmental Impact Assessment (Preliminary). Airphoto Analysis Associates Consultants Ltd., Toronto, Ont.

Banfield, A. W. F. 1974. The relationship of caribou migration behaviour to pipeline construction.

Pp. 797-804 in V. Geist and F. Walther, eds. The Behaviour of Ungulates and Its Relation to Management. International Union for Conservation of Native and Natural Resources, Gland, Switz.

Bergerud, A. T. 1971. The Population Dynamics of Newfoundland Caribou. Wildl. Monogr. 25:1-55.

Bergerud, A. T. 1974. The role of the environment in the aggregation, movement and disturbance behaviour of caribou. Pp. 552-584 in V. Geist and F. Walther, eds. The Behaviour of Ungulates and Its Relation to Management. International Union for Conservation of Native and Natural Resources, Gland, Switz.

Bergerud, A. T. 1979. Access: Greatest threat to caribou. West. Guidelines 2:5-9.

Brazil, J. 1983. Reactions of Caribou Observed During Construction at the Upper Salmon Hydroelectric Project, October, 1979-August, 1981. Unpublished manuscript.

Newfoundland Wildlife Division, St. John's, Nfld.

Cameron, R. D., K. R. Whitten, W. T. Smith, and D. D. Roby. 1979. Caribou distribution and group composition associated with construction of the trans-Alaska pipeline. Can. Field-Nat. 93:155-162.

Child, K. N. 1974. Reaction of caribou to various types of simulated pipelines at Prudhoe Bay, Alaska . Pp. 805-812 in V. Geist and F. Walther, eds. The Behaviour of Ungulates and Its Relation to Management. International Union for Conservation of Native and Natural Resources, Gland, Switz.

Geist, V. 1975. Harassment of large mammals and birds. Unpublished report to the Berger Comm., University of Calgary, Calgary, Alta.

Hanson, W. C. 1981. Caribou (Rangifer tarandus) encounters with pipelines in northern Alaska.

Can. Field-Nat. 95:57-62.

Hill, E. L. In press. A preliminary examination of the behavioural reaction of caribou to the Upper Salmon Hydroelectric Development in Newfoundland. In McGill Subarctic Research Papers. Vol. 40. McGill University Centre for Northern Studies, Montreal.

Horejsi, B. L. 1981. Behavioural responses of barren ground caribou to a moving vehicle. Arctic 34:180-185.

Kiell, D. J. 1981. Development of a reservoir preparation strategy. Can. Water Res. J. 7:112-131.

Kiell, D. J. 1984. Environmental decision-making during planning, construction, and early operation of the Upper Salmon Hydroelectric Development in Newfoundland, Canada: A case study.

Pp. 352-374 in Proceedings of the Facility Siting and Routing '84 Symposium, Banff, Alberta. Environmental Protection Service, Environment Canada, Ottawa, Ont.

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
224
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Klein, D. R. 1971. Reaction of reindeer to obstructions and disturbances. Science 173:393398.

Mahoney, S. P. 1980. The Grey River Caribou Study. Newfoundland Wildlife Division and Newfoundland and Labrador Hydro, St. John's, Nfld.

Mahoney, S. P. 1983. The trend toward bio-politics. Caribou News 3(3):12-13.

McCourt, K. H., J. D. Feist, D. Doll, and J. J. Russell. 1974. Disturbance studies of caribou and other mammals in the Yukon and Alaska, 1972. Arctic Gas Biological Report Series. Vol.

5. Canadian Arctic Gas Study Ltd. and Alaskan Arctic Gas Study Co., Calgary, Alta.

Miller, F. L., and A. Gunn. 1979. Responses of Peary caribou and musk oxen to helicopter harassment. Occasional Paper 40. Canadian Wildlife Service, Minister of Supply and Services Canada, Ottawa, Ont.

Newfoundland and Labrador Hydro. 1980. Upper Salmon Hydroelectric Development:

Environmental Impact Statement. Newfoundland and Labrador Hydro, St. John's, Nfld.

Northcott, P. L. Impact of the Upper Salmon Hydroelectric Development on the Distribution and Movement of the Grey River Caribou Herd in 1982. Unpublished manuscript.

Newfoundland Department of Environment.

Northcott, P. L. In press. Impact of the Upper Salmon Hydroelectric Development on the Grey River caribou herd. In McGill Subarctic Research Papers. Vol. 40. McGill University Centre for Northern Studies, Montreal.

Roby, D. D. 1978. Behavioral Patterns of Barren-Ground Caribou of the Central Arctic Herd Adjacent to the Trans-Alaska Oil Pipeline. Unpublished master's thesis, University of Alaska, Fairbanks.

Smith, W. T., and R. D. Cameron. 1983. Responses of caribou to industrial development of Alaska's Arctic slope. Acta. Zool. Fennica 175:43-45.

Committee Comment
Development projects, such as hydroelectric dams, have impacts on ecosystems both during construction and during their operation. Measures designed to reduce undesirable impacts of these two phases of a project are often different, as they were in the case of the Upper Salmon Hydroelectric Development. Construction of the dam was accompanied by much vehicular traffic and blasting. The latter ceased on project completion, but traffic, made possible in part by access roads built for construction, was clearly a continuing and underestimated problem whose nature changed as people began to use the area for purposes other than those occurring during construction.

During the planning phase, caribou were perceived as the ecosystem component of greatest concern and value. There were reasons to believe that the large Grey River herd might be affected during migration and calving. Planners, however, were confronted with conflicting information that suggested that caribou either might be very sensitive to the project or might be adaptable and become accustomed readily to the kinds of disturbance caused by construction and operation of the dam. As a result,

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
225
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. although a great deal was known about caribou biology and this information was used extensively during the planning and construction of the project, it was necessary to develop a plan for monitoring the behavior of animals and taking cautionary measures when animals were close to the construction sites, even if their responses to construction activity could not be predicted. In addition, roads were designed to be readily crossed by caribou and revegetation was encouraged, in the hope that the roads would pose minimal barriers to caribou movements once the project was completed.

The monitoring during construction revealed that use of the general project area by caribou was greatly reduced, but that use increased after construction.

The long-term effects of the project remain to be determined. Indeed, one of the disappointing aspects of the project is that monitoring was carried out for only 2 years after project completion. This is insufficient to determine whether caribou are moving through the project area as they did before construction, whether the efforts to make roads readily crossable have been successful, whether caribou are crossing the new lakes where old trees were cleared to facilitate their crossing, and whether there are long-term changes in the site and behavior of the herd.

A major problem, and one that is likely to occur regularly in similar projects, is that public access to previously roadless areas is greatly facilitated by roads built for project construction. The general disturbance and illegal hunting in the area might be much more serious for the long-term welfare of the Grey River caribou herd than was the disturbance during construction itself.

The problem of increased use of areas after construction of projects needs to be addressed formally and incorporated into policies governing such projects, perhaps including methods of limiting access of people to the areas at especially critical times.

PROTECTING CARIBOU DURING HYDROELECTRIC DEVELOPMENT IN

NEWFOUNDLAND
226
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. 17 Conserving a Regional Spotted Owl
Population
Human influence on the environment has reduced the populations of many plants and animals. Small populations are vulnerable to extinction, because of the difficulty of finding mates and because of random fluctuations, disease, unfavorable weather, and other catastrophic events. When populations of a species are small and fragmented (a common situation when habitat destruction is the cause of the reduction in population size), long-term survival can also be threatened by genetic deterioration — the deleterious effects of inbreeding and loss of genetic variability.

The survival of spotted owls is of increasing concern, because their required old-growth forest habitat is being reduced and fragmented by logging.

The planning for spotted owl management described here is based on theoretical population genetics and ecology, as well as on quantitative natural history— observations on reproductive ecology, dispersal, and foraging behavior. Although the underlying theory is not yet fully developed, it is applicable to many attempts to conserve populations of vertebrates with low reproductive potential.

CONSERVING A REGIONAL SPOTTED OWL POPULATION 227
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Case Study
HAL SALWASSER, USDA Forest Service, Wildlife and Fisheries Staff, Washington, D.C.

INTRODUCTION
Maintaining the full diversity of native vertebrates is a legal mandate of the federal government (Endangered Species Act of 1973, as amended; National Forest Management Act of 1976), as well as a policy of many state resource management agencies. In the late 1970s, the United States Department of Agriculture (USDA) Forest Service adopted regulations (36 CFR 219) requiring habitats to be managed to maintain viable populations of all native vertebrates in the national forest system— over 190 million acres of land and water.

Since the establishment of the national forests in the early 1900s, state and federal policies and lack of attention have resulted in the loss of species, such as wolves (Canis lupus) and grizzly bears ( Ursus arctos), from some forests. Land and resource management planning under the National Forest Management Act of 1976 is intended to prevent further loss of species.

The spotted owl (Strix occidentalis) constitutes a major test of the policy of maintaining species in national forests. The northern subspecies ( S. o. caurina Merriam 1898) inhabits mature and old-growth coniferous forests in the Cascade, Klamath, and Coast Range mountains of Washington, Oregon, and California. It appears to need stands of trees more than 24 in. in diameter, a multilayered canopy more than 70% closed, and large standing and fallen dead trees (Forsman et al., 1984). Such forests have extremely high commercial timber value (Heinrichs, 1983), and therein lies the spotted owl dilemma: the kind and amount of habitat required for survival of each pair of owls is a resource highly valued by an industry that is the economic backbone of the Pacific Northwest.

Further fragmentation of old-growth forest might impede dispersal of owls and isolate populations that are too small to survive for long. Thirty to forty years ago, half the original 15 million acres of old-growth forest in the Pacific Northwest remained (Franklin, 1984), and the spotted owl was not an issue in forest management. Timber harvests in the last few decades have removed nearly all the easily accessible lowland old-growth forest, and the much- reduced spotted owl population now exists primarily in rugged, mountainous terrain.

A major purpose of the national forests is to sustain yields of different CONSERVING A REGIONAL SPOTTED OWL POPULATION 228
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. kinds of resources. Timber, wildlife, and wilderness are specifically mentioned in the law (Multiple-Use Sustained Yield Act of 1960; Wilderness Act of 1964).

What is the most efficient way to manage forests to maintain both a continuous flow of timber products and viable populations of spotted owls and their coinhabitants of old-growth forests?
By the mid-1970s, the biology of the spotted owl was under study, and inventories had been initiated. In 1974, biologists began to develop management guidelines based on the new information, to prevent the owl from declining so much that formal listing as a threatened species would be necessary. Research and inventories have continued, and the management guidelines have evolved to incorporate new findings. In 1981, the guidelines were revised, largely under the influence of the work of Soul é (1980), to include new theories and observations concerning the genetic basis of population viability.

This chapter presents an overview of an evolving population management plan for the fragmented spotted owl population over a large area of the Pacific Northwest. The plan attempts to integrate management of individual national forests over the whole Pacific Northwest region. Rather than attempting to review the considerable research on spotted owl biology in detail, this discussion incorporates demographic estimates based on the research.

THE BASIC PROBLEM AND APPROACH
One objective of the Forest Service has been to develop plans for the national forests in the Pacific Northwest that protect resident populations of the spotted owl while allowing the multiple uses of the forest required by law. The primary threat to the owl is further reduction and fragmentation of its old- growth forest habitat through logging. Thus, the plan must deal with the effects of logging on the suitability of habitat for long-term maintenance of individual pairs, populations in specific national forests, and the whole regional population. Hence, the problem is to manage logging and related activities so that remaining old growth will support long-term survival of spotted owls in all national forests in the Pacific Northwest. Four general ecological issues are most important:
It is necessary to determine the habitat characteristics required for the survival and successful nesting of individual pairs of owls. This ultimately involves detailed studies of owls to determine patterns of habitat use and factors that influence habitat quality, the relationship between habitat quality and home range size, and specific requirements for nesting, foraging, roosting, and dispersal. Several studies are being conducted.

CONSERVING A REGIONAL SPOTTED OWL POPULATION 229
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Owls have been censused to establish which habitats they occur in and with what frequency and to determine the spacing between pairs.

Radiotelemetry has helped to determine patterns of habitat use and the size of home ranges of individual birds, and cast pellets have been examined to determine what prey are taken. These studies have provided the input for habitat-suitability index models that can evaluate habitat quality on the basis of specific features (Laymon et al., in press).
The distribution of habitats is critical to the successful dispersal of juveniles and movement of adults into suitable vacant habitat.

Radiotelemetry has been used to determine patterns of dispersal of juveniles and to determine what kinds of habitats they will not cross.

Adults with radio transmitters have also been followed to determine their movements between areas and throughout the year. The information can be used to develop criteria for minimal distances between patches of habitat during dispersal and for resident birds and to determine the best configuration of habitat that connects populations in adjacent forests.
The population in each forest must be large enough to withstand normal environmental fluctuations and random demographic changes without becoming locally extinct. In addition, the regional population must be sufficiently large and well distributed to withstand severe environmental fluctuations and reductions and even extinction of local populations. These requirements are being studied with a general model of regional population dynamics and with demographic values determined for the spotted owl from field studies and from the literature on better-studied species. The model is evolving to incorporate estimates of adult and juvenile mortality, range of dispersal, reproductive success, and other variables.
An effective population large enough to minimize the deleterious genetic effects associated with small and isolated populations must be maintained. The northwest spotted owl population is a metapopulation (Levins, 1970), i.e., a large regional population made up of many smaller populations of varied sizes, densities, and degrees of isolation from one another. Metapopulations are subject to two genetic problems: a loss of average individual fitness through an increase in the frequency of breeding of close relatives (inbreeding depression) and a reduction in potential for evolutionary adaptation because of a loss of genetic variability. Both these problems can be reduced by maintaining a number of relatively large local populations with substantial gene flow among them. A model similar to that used for population dynamics is being used with values for the spotted owl from field studies and from the literature on more intensively studied species.

Planning for protection of the spotted owl began in the early 1970s and CONSERVING A REGIONAL SPOTTED OWL POPULATION 230
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. continues today. An interagency task force of state and federal scientists and managers developed interim spotted owl management guidelines in the middle and late 1970s; later research led to a revision in 1981. The focus was initially on the kinds and amounts of habitat needed by individual pairs. The late 1970s saw increased academic interest in long-term population viability as the threat of mass extinction of species became widely recognized. The result has been more research on factors underlying extinction (e.g., Frankel and Soul é, 1981; Schonewald-Cox, 1983; Soul é and Wilcox, 1980). A 1982 workshop that included management biologists and academic ecologists studying population viability used the spotted owl as a basis for developing a general risk- assessment planning process for long-term population management (Salwasser et al., 1984). The approach has been revised to incorporate new information on other species and the results of a second workshop that was held in the fall of 1984.

Debate concerning a plan for managing spotted owls has focused on several issues. First, what are the specific habitat requirements of individual pairs of owls? The management unit for individual pairs or small groups of pairs with contiguous home ranges is called a spotted owl management area (SOMA). The type, amount, and distribution of habitat that should constitute an adequately managed SOMA have been controversial. Underestimation of minimal home range size or quality of sites needed or too much fragmentation of habitat within a SOMA could reduce the probability of survival and reproduction of a pair. Second, are the demographic estimates used in the management models sufficiently accurate? At issue are estimates of adult and juvenile mortality, reproductive rate, dispersal distance, habitat occupancy, and related population characteristics. Third, are the models themselves adequate and appropriate? Are demographic or genetic problems more important to long- term viability? The current guidelines are based on the assumptions that genetic problems are more critical at the scale of the regional population and that preventing genetic deterioration will prevent demographic collapse.

Demographics and biogeography are assumed to be most critical at the scale of a forest population.

The overriding issue concerns the minimal regional population size necessary for long-term survival and consequently the amount and distribution of old-growth forest to be provided in the future. Points of view range from the position of some environmental interest groups, that too much old growth has already been cut and that habitat management criteria (particularly minimal SOMA characteristics) are inadequate, to the view of some representatives of the timber industry, that current management guidelines are too stringent and that the current estimate of minimal acceptable population size is too large. The current guidelines were under
CONSERVING A REGIONAL SPOTTED OWL POPULATION 231
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. administrative appeal in March 1985, and legal action might be taken by parties at both extremes.

ECOLOGICAL KNOWLEDGE USED IN DEVELOPING THE

MANAGEMENT GUIDELINES
The management guidelines are based on data, principles, and theory. In addition to field research on the biology of spotted owls themselves, the guidelines rely heavily on the findings of recent theoretical and empirical research on the long-term viability of populations of other species, both in captivity and in the wild. Studies on spotted owls are used to estimate values for general models of population viability based on this recent literature.

Studies of the Ecology of Spotted Owls
Research on spotted owls has focused on breeding biology, foraging ecology, habitat use, and general distribution. Because spotted owls are nocturnal and now inhabit rugged, mountainous terrain almost exclusively, field research is very difficult, and much of their biology is still poorly known.

Individual pairs of owls do not reproduce every year, and clutch size averages only about two eggs (Forsman, 1980). In addition, the survival of juveniles appears to be extremely low; no juveniles with radio transmitter harnesses have survived to breed. Hence, spotted owls appear to have a very low reproductive potential and thus poor ability to recover from reductions in population size.

Spotted owls in the Northwest appear to be nonmigratory, and radiotelemetry has shown that adults move over extremely large home ranges in the course of a year. A home range can be as large as 8,300 acres for an individual owl and 10,400 acres for a pair (Forsman et al., 1984). The old- growth coniferous forests occupied by spotted owls in the Pacific Northwest are over 200 years old and have several layers, with many standing dead trees and new trees coming up in gaps where old ones have fallen. The multilayered structure of these forests is believed to affect foraging success, partly through an effect on the abundance of the owl's prey — primarily arboreal rodents, such as flying squirrels, voles, and wood rats. Multilayered forests also provide cool microsites that allow owls to avoid heat stress, to which they are apparently sensitive. The broken-off tops of mature trees characteristic of old-growth forest also provide nest sites for the owls (Forsman, 1980). Spotted owls are not known to breed in young second-growth forests and are rarely found there.

CONSERVING A REGIONAL SPOTTED OWL POPULATION 232
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Although it is generally assumed that home ranges must be larger in areas with lower habitat quality, and home ranges are indeed larger in heavily logged areas (Forsman, 1980), there is no good evidence of a general correlation between habitat quality and home range size within a study area. Territories might be abandoned as a result of timber harvest (Forsman et al., 1984). In any year, not all territories are occupied; the average occupancy rate can be as low as 50% and does not exceed 75% of potentially suitable sites. The minimal home range size that would adequately support a breeding pair is still being debated and probably varies with the individual pair, terrain, forest type, and forest distribution. Sizes estimated from censuses and radiotelemetry studies range from 740 acres to 8,300 acres in heavily logged areas (Forsman et al., 1984; Marcot, 1978). Logging in a pair's home range reduces habitat quality, not only by reducing the amount of high-quality habitat, but also by fragmenting the forest. Fragmentation leads to greater travel time for foraging and greater exposure to predators, such as great horned owls. The effects of habitat removal can be minimized by retaining travel corridors of old-growth forest to connect the larger stands.

Some juvenile spotted owls have been reported to disperse as far as 100 miles from their areas of birth, although most do not travel that far (Gutierrez, personal communication). It is not known why no radiotagged juveniles have survived to breed. They might be too inefficient in finding suitable habitat or suffer high mortality when occupying low-quality habitats during dispersal. It is also possible that the radio transmitters themselves increase the chance of death in young and inexperienced owls. Adults apparently do not like to cross large open areas (Forsman et al., 1984), and they use corridors of old-growth forest, when available, to travel between old-growth stands, even if much longer distances must be covered.

Long-Term Population Viability
Studies of population viability have focused on how and why the risk of extinction increases as populations become smaller and more isolated from other populations. Interest has been not only in the fate of local populations, but also in how long-term viability of a regional population or an entire species can be affected by changes in the size, makeup, and distribution of its constituent local populations.

Reviews of the factors that have led to recent extinctions indicate that natural agents — such as predation, competition, parasitism, and disease — have rarely been the cause of extinction and that the reduction, alteration, CONSERVING A REGIONAL SPOTTED OWL POPULATION 233
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. and isolation of habitats through human activity are ultimately more important (Frankel and Soul é, 1981; Hester, 1967; Soul é, 1983; Terborgh and Winter, 1980; Ziswiler, 1967).

Two categories of factors increase the risk of extinction of small and isolated populations (Shaffer, 1981; Soul é, 1983; Terborgh, 1974): internal changes and external stresses. Population productivity can decrease as a result of random changes in fertility rate, litter size, sex ratio, death rate, immigration rate, and so on. Behavioral dysfunction occurs in some species below a threshold population size, often disrupting breeding. Genetic drift can lead to loss of genetic variation, with a loss of potential for adapting to environmental change. Studies of captive populations have also shown that inbreeding depression occurs when fewer individuals — all closely related— are available to mate; this situation often decreases reproduction and survival (Soul é, 1980).

External factors also have more severe effects as populations become smaller and more isolated. Habitat change, intense predation or competition, fire, drought, and floods can leave too few individuals for recovery to occur. In short, when populations become very small, the probability increases dramatically that random or even regular periodic events will reduce population numbers so much that recovery can occur only if individuals immigrate from other populations. As populations become more fragmented and isolated, the supply of potential immigrants diminishes.

For a population to survive, it must have demographic resilience, i.e., resistance to extinction due to random demographic changes or environmental fluctuations. Theoretical studies suggest that as few as 10 adults can sustain a population for decades; but, for any given species, this minimum depends on social system, reproductive potential, generation length, the nature of random events, and other factors (May, 1973). Some empirical studies suggest that 50 adults might be necessary for a reasonable chance of surviving for several decades (Shaffer, 1981). For a population to survive, it must also maintain adequate average individual fitness in the face of inbreeding depression (Chambers, 1983) and must have sufficient genetic variability to allow adaptation to environmental change (Soul é, 1980). Populations lose genetic variation through random changes in gene frequency (Kimura, 1983). In small populations, inbreeding is more likely and genetic variation is more likely to be lost through genetic drift. Genetic variation arises through mutation and immigration, and rough estimates indicate that, to avoid extinction through genetic deterioration, a population must be about 10 times larger than that necessary to maintain demographic resilience (Soulé, 1980).

The initial guidelines for management described below were based on the assumption that adequate demographic resilience will be maintained CONSERVING A REGIONAL SPOTTED OWL POPULATION 234
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. if a population is managed to minimize genetic deterioration. Current planning is assessing the roles of demographic variation and population fragmentation as well.

In a genetically ''ideal" population, individuals have an equal probability of mating with each other, fecundity is constant, population size is constant, and generations are discrete. Few real populations meet these criteria (Hartl, 1980; Kimura, 1983; Kimura and Crow, 1963; Wright, 1938). To predict random genetic change in a population of average size N, when the population characteristics depart from the ideal, an effective population size, N e, is often
used in lieu of the actual census size. The calculation of Ne typically incorporates adjustments for deviation from the ideal in sex ratio, variation in litter size or survival of offspring to reproductive age, overlapping of generations, random mating, and fluctuation of population number (Kimura, 1983). N
e is used to calculate inbreeding coefficients, and it is usually less than the census N— sometimes less than 20% of it. Formulas for estimating the effects of population structure on Ne have been given by Franklin (1980), Hartl (1980), Frankel and Soul é (1981), Kimura (1983), and Thomas and Ballou (1983).

Corrections for three factors are often incorporated into the calculation of Ne: sex ratio different from 1:1, variation in offspring survival, and population fluctuations.

If not all adults breed, some genes are less likely than others to be passed on. A biased sex ratio can create such an effect, and Ne should be adjusted as follows:

Ne = 1/[1/(4Nm) + 1/(4Nf )], (1)
where Nm and Nf are the numbers of adult males and females.

The reproduction of genes is biased when the production of offspring varies between parents. The effect of this variation on Ne is given by:

Ne = 4N /(2 + V), (2)
where V is the variance in survival of offspring per parent and N is the actual population size.

When a fluctuating population is well below the average size, genes can be reproduced in a biased fashion as a result of genetic drift. The longer a population remains small (i.e., fails to recover from decrease), the greater the effect. It will be more pronounced, therefore, in species with low powers of increase. The effect on N
e is as follows:

Ne = t/(1/Ne1 + 1/Ne2 + . . . + 1/Net). (3)
where t is the number of generations stipulated and Net is the effective number (N adjusted for sex ratio and offspring variance) in generation t.

CONSERVING A REGIONAL SPOTTED OWL POPULATION 235
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

The total effect of these three factors can be approximated by:

Ne = NSOP, (4)
where S, 0, and P are the ratios of Ne to N calculated for sex ratio, offspring variance, and population fluctuation, respectively. This approximation should be sufficiently precise for general wildlife habitat management, in which case other information is likely to be much less precise than the estimate of N e.

The inbreeding coefficient, F, is calculated with the overall value of Ne as derived above. Empirical data on the actual effects of inbreeding in small populations of normally free-ranging and outbred species of mammals have shown that even a small amount of inbreeding is correlated with reduced fecundity and reduced survival of offspring (Ralls and Ballou, 1982, 1983; Ralls et al., 1979). Free-ranging populations normally have behavioral and ecological patterns that keep natural inbreeding low. The formula for F is as follows:

Ft = 1 – (1 – [1/(2Ne) + 0.5]t, (5)
where Ft is the inbreeding coefficient in generation t, t is the number of generations from time zero, and Ne is the effective population number during the period of interest.

Some of the variables in these formulas cannot be measured in wild populations, and Soul é (1980) cautioned that basing management on estimates of Ne can yield only rules of thumb, rather than reliable quantitative results. As a rule of thumb, Franklin (1980) and Soul é (1980) suggested that an Ne of 500 or more might approach the balance point between random loss and addition of genetic variation for many species, and Lande (1980) suggested that populations of several hundred are at little more risk than very large ones. The life history and population structure of a species must be considered before such a rule of thumb is applied, but there is now little basis for determining a more specific threshold value for any species.

Soulé (1980) has proposed that wild populations are at risk of extinction because of genetic factors when F
t reaches 0.5 ( Figure 1). Because Formula 5
does not consider the mitigating effects of migration between populations or mutations, it overestimates the inbreeding coefficient, sometimes substantially (Hartl, 1980; Kimura, 1983; Figure 2 ). Figure 1 shows that, assuming a generation time of 2 years for the spotted owl, a regional population of 500 or more will be sufficient to provide protection against genetic deterioration for many centuries. It is assumed that the inbreeding coefficient calculated for the regional population also applies
CONSERVING A REGIONAL SPOTTED OWL POPULATION 236
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

FIGURE 1 Inbreeding coefficient, F, increases as function of effective population number, Ne, and number of generations. At low Ne, F approaches dangerous extent of inbreeding in fewer generations.

FIGURE 2 Higher migration rates of reproductively successful individuals into small population from larger population offset effects of inbreeding.

CONSERVING A REGIONAL SPOTTED OWL POPULATION 237
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. to all local forest populations, because they are not totally isolated from each other.

THE PLANNING PROCESS FOR SPOTTED OWLS
The planning process for spotted owls is derived from a general protocol (Salwasser et al., 1984) and now consists of eight steps (Figure 3). Step 1: The northern spotted owl in the Pacific Northwest was identified as a species of concern, for several reasons. The owl is obligately dependent on a habitat that is now rare (old-growth coniferous forest) and that is being reduced rapidly, and individuals require large amounts of the habitat. The population is becoming increasingly fragmented, as a result of continued timber operations, and some individual forest populations might be close to the lower limit of adequate short-term demographic resilience — about 50-100 individuals. Planning is required now, so that the regional population does not decrease below the size necessary for long-term survival.

Because the spotted owl has the most stringent requirements for old- growth habitat, managing old growth to protect spotted
FIGURE 3 Eight-step process for planning viable
population and analyzing risk.

CONSERVING A REGIONAL SPOTTED OWL POPULATION 238
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. owls will probably also protect most, if not all, other species that depend on old-growth forest (Raphael and Barrett, 1984). Step 2: Planning responsibility is assigned on the basis of range and distribution of the species. Censuses have been performed on all national forest land and evaluated in conjunction with information from adjacent lands to identify areas that might become effective biological reserves for the species, regardless of ownership and prevailing land use.
Step 3: Habitat requirements for spotted owls and the best distribution of habitat within and between individual forest populations are determined, and the results lead to development of habitat capability models that describe the full range of habitats over which the species occurs (Nelson and Salwasser, 1982). The models can be used to guide habitat planning and to determine the effects of different land-use patterns on the species (Laymon et al., in press). On the basis primarily of radiotelemetry studies of spotted owl habitat use in Oregon, a minimally suitable year-round home range for a pair is presumed to include 300 contiguous acres of mature to old-growth forest and 700 acres within 1.5 miles of the nest site (Forsman et al., 1984).

To facilitate interchange among local populations and to make the occupancy of suitable habitat more likely, each managed habitat area must be within the normal dispersal distance of the species — 6-12 miles, according to the results of radiotelemetry studies. The key concern is that loss of a piece of habitat not lead to permanent isolation of a local population. The best pattern of habitat distribution would entail several connections among suitable pieces of habitat, so that the loss of one connection would not isolate any piece. Linear patterns should be avoided.
Step 4: Population and habitat requirements are translated into land-use planning variables. This is accomplished by assigning each individual forest (or other planning unit) a quota for the number of pairs to be maintained (as determined from the overall risk analysis, discussed below) and by specifying how habitat in each forest is to be managed.

In the national forests, this plan must be flexible to accommodate multiple land uses. There will inevitably be alternative plans for meeting these multiple objectives, each of which will involve trade- offs among individual goals, such as protecting spotted owls. Step 5: The alternative management strategies are projected to estimate their effects on key population dimensions, such as N, N e, and
the structure of local demes. The projections use habitat and population simulation models with explicit assumptions about rates of systematic changes and the importance of random variation.

For each planning alternative, two substeps are performed: estimation of the population size in each forest that the planned habitat distribution CONSERVING A REGIONAL SPOTTED OWL POPULATION 239
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. can support (habitat capability) and estimation of the effective population size (Ne).

Habitat capability for each forest can be estimated with the habitat management criteria for owl pairs described above and with censuses of owls and maps of projected habitat distribution. The formulas presented earlier can be used with population data on spotted owls to estimate N e both for individual
forests and for the regional population (Formulas 4 and 5). Spotted owls breed monogamously, and, because there is no indication of a biased sex ratio, the coefficient S is assumed to be 1.0. Data on variance in offspring production are few, but studies have shown that complete reproductive failure occurs in some years (Gutierrez, personal communication). Assuming that variance in offspring production would be at the high end of the range for vertebrates, the coefficient O is assigned a value of 0.66 (inferred from Crow and Morton, 1955).

Accurate counting of the nocturnal spotted owl is difficult, because the failure of individuals to respond to imitated or recorded calls might indicate only unresponsiveness, rather than absence. Population fluctuation in spotted owls is probably small, because populations of other species with low reproductive rate and high adult survivorship generally fluctuate little. It is assumed that populations do not decrease to less than 50% of the average more frequently than 1 year in 5 and that it takes about 3 years to recover from a decrease to 50% of average N. Use of these assumptions in Wright's (1938) harmonic-mean formula gives a value of 0.76 for the coefficient P. Combining these coefficients (Formula 4) yields an estimated N
e of half the census N (1.0 ´

0.66 ´ 0.76). Thus, Ne for the northwest regional spotted owl population, currently estimated at more than 2,000 adults, would be about 1,000, or 500 pairs. The N
e for any isolated population can be calculated similarly by multiplying the census number by 0.5.
Step 6: An effective degree of protection ( Table 1) is determined for a species through a risk analysis of the estimated demographic, generic, and geographic results of the management alternatives. The eventual management goals will depend explicitly on the degree of protection desired for the species and will be in essence a value judgment involving a cost-benefit analysis of possible protective measures.

Obviously, the greatest protection would be obtained if no more habitat were altered or removed, but that is not always practical.

Assessing the degree of protection with each alternative involves evaluating the expected size of the whole regional population and the size and degree of isolation of each forest unit. If isolation is not expected, the potential demographic resilience of a forest population can be evaluated CONSERVING A REGIONAL SPOTTED OWL POPULATION 240
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. with the habitat capability model, because some immigration can be expected to offset the risk of temporary population declines. Local populations that are not isolated from others are assumed to experience the same inbreeding coefficient as the regional population as a whole (Figure 1).

CONSERVING A REGIONAL SPOTTED OWL POPULATION 241
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

When a proposed alternative is expected to result in isolation of a local population, Steps 5 and 6 should be performed for that population. Land-use planning should evaluate the likelihood that planned or unplanned events will eliminate adjacent habitats at the weakest points in the habitat distribution. Step 7: This step, the decision process, involves consideration of other resource concerns and values for the areas of land in question.

Biological assessments are blended with social, political, and economic issues. The decisions often entail a social preference among competing uses of the land.
Step 8: A monitoring and research program is developed. It should reflect the degree of protection and the potential environmental costs of management and should allow for evaluation of critical assumptions used in the risk assessment protocol (Salwasser et al., 1983). It should also stipulate the variables to be measured and the frequency of measurement and should address the issue of measurement reliability.

A monitoring program has three major goals. First, compliance with the plan must be monitored. For example, if 100 SOMAs are allocated to one national forest, it must be determined whether all SOMAs are being maintained properly and whether the actual spatial relationships of the various SOMAs are acceptable. Second, monitoring must show whether management is achieving its resource goals. For example, are the SOMAs supporting as many pairs of owls as expected, is reproductive rate adequate for long-term population stability, and is genetic variation being maintained? Third, information must continually be gathered and used to update and revise the plan. Some of this information must come from experimental research.

Key assumptions in current planning involve dispersal behavior of juveniles, the nature of the owls' dependence on stands of old-growth forest for reproduction and survival, and the adequacy of population sizes for maintaining demographic resilience. These assumptions are being studied by a number of agencies (Ruggiero and Carey, 1984). There are no plans to conduct intensive research on genetic factors, as long as at least 1,000 adults are maintained in the regional population. A demographic model for the spotted owl is being modified to incorporate genetic considerations and the effects that demographic variation might have on viability.

CONTRIBUTION OF ECOLOGICAL KNOWLEDGE TO THE

CASE STUDY RESULTS
Until recently, efforts to protect populations were based primarily on site- specific habitat management, to ensure the survival and reproductive CONSERVING A REGIONAL SPOTTED OWL POPULATION 242
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. success of individuals and breeding pairs. Although such efforts are still critical to species management, it is now recognized that local populations cannot be protected for long without studies of interactions among populations that might exchange individuals — studies that have been central to developing a regional management plan for the spotted owl.

But metapopulation and other genetic models are only as good as the values that go into them. In spite of recent research on spotted owls, there is little assurance that the values of variables used in the models are accurate.

Many must be derived from studies on unrelated species. Information on distribution and abundance is generally accurate, but knowledge of survivorship, reproductive rate, dispersal, and other demographic characteristics is sparse. Thus, it is essential to view current management plans and their assumptions as hypotheses to be tested. Both monitoring and research must be involved in the tests. The current plan is flexible and adaptive, and it takes the uncertainty of the effects of our proposed management actions into account.

ACKNOWLEDGMENTS
Contributions to this chapter have come from many people. Michael Soulé's lucid writing on conservation biology prompted the planning process described. Eric Forsman, Rod Canutt, and Dean Carrier worked out the basic method for specifying suitable habitats and distribution (Step 3). Jack Ward Thomas showed how to use the species habitat niche to represent resource needs in a practical way (Steps 3 and 4). Karl Siderits and Bob Radtke pioneered the use of diversity standards to provide for the habitat needs of all wildlife (Step 4). John Lehkuhl, Ed Harshman, and Daniell Jerry tested the application of population theories to determining the number of individuals needed (Step 5). Michael Soul é, Daniel Goodman, Michael Gilpin, James Brown, Linda Joyce, Tom Hoekstra, Mark Shaffer, Curt Flather, Dick Holthausen, Bill Burbridge, Brad Gilbert, Charlie Phillips, Maureen Beckstead, Tom Burke, and Paul Brouha assisted in workshops, brainstorming, and review of manuscripts to develop the analytical strategy and process.

REFERENCES
Chambers, S. M. 1983. Genetic principles for managers. Pp. 15-46 in C. M. Schonewald-Cox, S. M.

Chambers, B. MacBryde, and W. L. Thomas, eds. Genetics and Conservation: A Reference for Managing Wild Animal and Plant Populations. Benjamin/Cummings, Menlo Park, Calif.

Crow, J. F., and N. E. Morton. 1955. Measurement of gene frequency drift in small populations.

Evolution 9:202-214.

CONSERVING A REGIONAL SPOTTED OWL POPULATION 243
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Forsman, E. D. 1980. Habitat Utilization by Spotted Owls in the West-Central Cascades of Oregon.

Ph.D. thesis, The Oregon State University, Corvallis.

Forsman, E. D., E. C. Meslow, and H. M. Wight. 1984. Distribution and biology of the spotted owl in Oregon. Wildl. Monogr. 87:5-68.

Frankel, D. H., and M. E. Soul é. 1981. Conservation and Evolution. Cambridge University Press, New York.

Franklin, I. R. 1980. Evolutionary change in small populations. Pp. 135-149 in M. E. Soul é and B.

A. Wilcox, eds. Conservation Biology: An Evolutionary-Ecological Perspective. Sinauer, Sunderland, Mass.

Franklin, J. F. 1984. Characteristics of old-growth Douglas-fir forest. Pp. 328-334 in Society of American Foresters. New Forests for a Changing World. Society of American Foresters, Bethesda, Md.

Hartl, D. L. 1980. Principles of Population Genetics. Sinauer, Sunderland, Mass.

Heinrichs, J. 1983. The winged snail darter. J. For. 81:212-262.

Hester, J. 1967. The agency of man in animal extinctions. Pp. 169-192 in P. S. Martin and H. E.

Wright, eds. Pleistocene Extinctions: The Search for a Cause. Yale University Press, New Haven, Conn.

Kimura, M. 1983. The Molecular Theory of Evolution. Cambridge University Press, New York.

Kimura, M., and J. F. Crow. 1963. The measurement of effective population number. Evolution 17:279-288.

Lande, R. 1980. Genetic variation and phenotypic evolution during allopatric speciation. Am. Nat. 116:463-479.

Laymon, S. A., H. Salwasser, and R. H. Barrett. In press. Habitat suitability index models: Spotted owl. U.S. Fish Wildl. Serv. Biol. Rep.

Levins, R. 1970. Extinction. Pp. 77-107 in M. Gerstenhaber, ed. Some Mathematical Questions in Biology. Vol. II. American Mathematical Society, Providence, R.I.

Marcot, B. G. 1978. Prolegomena of the Spotted Owl (Strix occidentalis) in Six Rivers National Forest. Tech. Rept. USDA For. Serv. Six Rivers National Forest, Eureka, Calif.

May, R. M. 1973. Stability and Complexity in Model Ecosystems. Princeton University Press, Princeton, N.J.

Nelson, R. D., and H. Salwasser. 1982. The Forest Service wildlife and fish habitat relationship program. Trans. N. Am. Wildl. Nat. Resour. Conf. 47:174-183.

Rails, K., and J. Ballou. 1982. Effect of inbreeding on juvenile mortality in some small mammal species. Lab. Anim. 16:159-166.

Ralls, K., and J. Ballou. 1983. Extinction: Lessons from zoos. Pp. 164-184 in C. M. Schonewald- Cox, S. M. Chambers, B. MacBryde, and W. L. Thomas, eds. Genetics and Conservation:

A Reference for Managing Wild Animal and Plant Populations. Benjamin/Cummings, Menlo Park, Calif.

Ralls, K., K. Brugger, and J. Ballou. 1979. Inbreeding and juvenile mortality in small populations of ungulates. Science 206:1101-1103.

Raphael, M. G., and R. H. Barrett. 1984. Diversity and abundance of wildlife in late successional Douglas-fir forest. Pp. 352-360 in Society of American Foresters. New Forests for a Changing World. Society of American Foresters, Bethesda, Md.

Ruggiero, L. F., and A. B. Carey. 1984. A programmatic approach to the study of oldgrowth forest- wildlife habitat relationships. Pp. 328-334 in Society of American Foresters. New Forests for a Changing World. Society of American Foresters, Bethesda, Md.

Salwasser, H., C. K. Hamilton, W. B. Krohn, J. F. Lipscomb, and C. H. Thomas. 1983.

CONSERVING A REGIONAL SPOTTED OWL POPULATION 244
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Monitoring wildlife and fish: Mandates and their implications. Trans. N. Am. Wildl. Nat.

Resour. Conf. 48:297-307
Salwasser, H., S. P. Mealey, and K. Johnson. 1984. Wildlife population viability — A question of risk. Trans. N. Am. Wildl. Nat. Resour. Conf. 49:421-439.

Schonewald-Cox, C. M. 1983. Conclusions: Guidelines for management: A beginning attempt. Pp. 414-446 in C. M. Schonewald-Cox, S. M. Chambers, B. MacBryde, and W. L. Thomas, eds. Genetics and Conservation: A Reference for Managing Wild Animal and Plant Populations. Benjamin/Cummings, Menlo Park, Calif.

Shaffer, M. 1981. Minimum population sizes for species conservation. BioScience 31:131-134.

Soulé, M. E. 1980. Thresholds for survival: Maintaining fitness and evolutionary potential . Pp. 151-169 in M. E. Soul é and B. A. Wilcox, eds. Conservation Biology: An Evolutionary- Ecological Perspective. Sinauer, Sunderland, Mass.

Soulé, M. E. 1983. What do we really know about extinction? Pp. 414-446 in C. M. Schonewald- Cox, S. M. Chambers, B. MacBryde, and W. L. Thomas, eds. Genetics and Conservation:

A Reference for Managing Wild Animal and Plant Populations. Benjamin/Cummings, Menlo Park, Calif.

Soulé, M. E., and B. A. Wilcox. 1980. Conservation Biology: An Evolutionary-Ecological Perspective. Sinauer, Sunderland, Mass.

Terborgh, J. 1974. Preservation of natural diversity: The problem of extinction-prone species.

BioScience 24:715-722.

Terborgh, J., and B. Winter. 1980. Some causes of extinction. Pp. 119-134 in M. E. Soul é and B. A.

Wilcox, eds. Conservation Biology: An Evolutionary-Ecological Perspective. Sinauer, Sunderland, Mass.

Thomas, W. L., and J. Ballou. 1983. Equations and population management. Pp. 414-446 in C. M.

Schonewald-Cox, S. M. Chambers, B. MacBryde, and W. L. Thomas, eds. Genetics and Conservation: A Reference for Managing Wild Animal and Plant Populations. Benjamin/ Cummings, Menlo Park, Calif.

Wright, S. 1938. Size of population and breeding structure in relation to evolution. Science 87:430-431.

Ziswiler, Z. 1967. Extinct and Vanishing Animals. Springer, New York.

Committee Comment
As we continue to alter and fragment the habitat of more and more species of plants and animals, we add to the list of species that are composed of metapopulations at an alarming rate. Only recently has it been widely recognized that such species cannot be protected against the threat of extinction by managing habitat for their local constituent populations. This recognition is the crucial first step toward maintaining the long-term viability of threatened species, but the scientific issues raised will not be easy to deal with, and the social and political steps required will not be easy to accomplish. Habitats of many species are distributed without regard for national boundaries and are often in countries where pressure for modifying those habitats is intense.

Even if we had control over habitat destruction, we would be far from CONSERVING A REGIONAL SPOTTED OWL POPULATION 245
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. understanding how to protect these species. Theoretical research in metapopulation management is in its infancy, and the data needed for emerging models are difficult to acquire and are lacking for most species. The spotted owl management plan constitutes one of the first attempts to incorporate metapopulation modeling into usable management guidelines. But, as pointed out by Salwasser, even the basic issue of whether demographic or genetic constraints are more critical is being hotly debated, and only very general rules of thumb for determining minimal population sizes are available. In Salwasser's spotted owl case study, genetic constraints are assumed to be more stringent, and the model used is based primarily on limiting genetic deterioration. Until more research is done on demographic models, we cannot be confident that this assumption is valid.

In a recent study, Lande (1985) suggested that demographic factors might be more critical than genetic factors to the survival of spotted owls. Lande used two independent analytical methods — a basic life-table analysis and an analysis of habitat occupancy — and concluded that current management plans for the spotted owl will eventually result in demographic collapse. Lande's study demonstrates the importance of exploring every available analytical approach before making irreversible habitat management decisions.

In addition to the deterministic approaches used by Lande, Shaffer (1981) and others have been developing a Monte Carlo simulation approach for determining minimal viable population size. Various demographic values can be used in an iterative stochastic simulation model that projects population size and makeup for many generations. By simulating both environmental and demographic variability, the models clearly demonstrate that population viability is a matter of probability. How long do we wish the population to avoid extinction? What risk (probability of extinction) are we willing to accept? Management decisions ultimately depend on what are essentially value judgments.

On a metapopulation level, we must also address the issue of whether near- term (centuries) protection against demographic collapse is more important than providing the potential for long-term evolutionary change. Overall management plans aimed at these two different goals will often be very different. Is one very large population— less likely to become extinct — more desirable than a system of smaller, connected populations that are more likely to facilitate evolutionary change in response to locally changing environments? Clearly, protecting a species against demographic collapse is a first priority, for, unless we maintain a viable population, we will not even have the chance to tackle the longer-term problems of loss of genetic variation.

The management plan presented in this chapter will undergo many CONSERVING A REGIONAL SPOTTED OWL POPULATION 246
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. changes with a better understanding of metapopulation dynamics and genetics and with more demographic and ecological data on spotted owls. The clear and accessible presentation of the plan, however, is of great value, because it focuses the debate and research needed for the achievement of its goals.

References
Lande, R. 1985. Report on the Demography and Survival of the Spotted Owl. Paper prepared for the National Wildlife Federation, Portland. Oreg.

Shaffer, M. 1981. Minimum population sizes for species conservation. BioScience 31:131-134.

CONSERVING A REGIONAL SPOTTED OWL POPULATION 247
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. 18 Restoring Derelict Lands in Great Britain
Reclamation of derelict lands is of increasing importance for industrialized societies, both because quantities of lands so affected are increasing and because the value attached to restoring natural vegetation on these lands is increasing. Early reclamation attempts tried to produce vegetation that looked pleasing and that reduced soil erosion. More recently, society has demanded the creation or restoration of more complex natural communities of plants and animals, even though this goal is more difficult and expensive. The task requires knowledge of the tolerances of plants for the environmental conditions; of the ways in which plants compete with one another for access to water, light, and mineral resources; and of the course of succession of plant species after establishment of pioneering vegetation. Less attention has been paid to restoration of animal communities, in part because animals often invade these sites naturally once suitable vegetation has been established, and in part because less is known about members of animal communities and how to manipulate them.

The Longstone Edge reclamation project in Derbyshire, England, represents the current state of the art of land reclamation. The project drew on existing ecological knowledge about plant competition and succession, but also included experiments to help develop new theories more applicable to the specific conditions of reclamation with which the investigators had to work. A particular part of the challenge — one that is generally a component of such projects— was to produce vegetation that would be self-regenerating once established, so that costly maintenance would not need to be continued for long periods. In Britain, most derelict sites are on
RESTORING DERELICT LANDS IN GREAT BRITAIN 248
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. very poor soils; the specifics of the Longstone Edge results are, accordingly, applicable to many other sites in that country. In North America, however, many sites being restored are on richer soils, and techniques adapted to those conditions, as well as to the more continental climates prevailing in most of North America, will need to be developed to meet challenges there.

Case Study
PETER WATHERN, Department of Botany and Microbiology, University College of Wales, Aberystwyth

INTRODUCTION
Bradshaw and Chadwick (1980) estimated the amount of derelict land in the United Kingdom in April 1971 at over 55,000 hectares (ha), and the consensus is that the amount has increased steadily in recent years. Much of this dereliction is the result of mining activities, and pressures on resources in the United Kingdom are such that mineral deposits are likely to be worked with little or no concern for their location. Thus, mining operations often occur in some of the most scenically attractive and environmentally sensitive areas, such as national parks. Examples include Dartmoor (kaolin), Yorkshire Dales (heavy metals and fluorspar), and the Peak District (heavy metals, fluorspar, and limestone). Many areas, including western and northeastern Wales and Cornwall, carry a legacy of dereliction from mines long defunct.

In the United Kingdom, reclamation of despoiled land has been a fertile area for research for more than 2 decades, since the pioneering work in the Lower Swansea Valley of Wales reported by Hilton (1967). Reclamation research has been geared to developing techniques for achieving three main objectives: to restore land so that it can be used productively, to remove local sources of environmental pollution associated with toxic materials in discarded wastes, and to make a site visually attractive again. Reclamation has generally been achieved by establishing a vegetation cover that prevents further weathering and allows such productive uses as agriculture, forestry, housing, manufacturing, and recreation. Thus, reinstatement can be regarded as an essential component of the recycling of land through a succession of uses.

In some instances, derelict land detracts markedly from an otherwise pleasing landscape and must be reclaimed in the interests of visual amenity, irrespective of a readily apparent productive after-use. The major concern RESTORING DERELICT LANDS IN GREAT BRITAIN 249
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. in the reinstatement of such sites is the creation of an area that, at worst, blends with the surrounding landscape or, preferably, is pleasing in its own right. This case study describes experimental work undertaken to develop techniques for dealing with such situations.

Longstone Edge, Derbyshire, is a limestone scarp situated in a prominent position in the heart of the Peak District National Park. It visually dominates an area intensively used for recreation. Fluorspar deposits are found in a vertical vein (locally known as a rake) running for about 1.5 km along its crest. The deposits are worked by a combination of open-cast and deep mining techniques.

The area has a long history of lead mining, and the steep scarp face of Longstone Edge is covered with waste heaps tipped to a depth of some 5 m by medieval miners. These waste heaps of gangue minerals, such as fluorspar and barytes, are being reworked by the mining company. Much of the site, including the medieval waste heaps, is covered with species-rich calcareous grassland, and heath dominated by heather ( Calluna vulgaris) has developed on superficial loessic drift (Balme, 1953; Pigott, 1962). The limestone grassland contains several rare species, including Epipactis atrorubens, Minuartia verna, Orchis apifera, and the terricolous lichen Cetraria islandica. As a result of mining activities, seminatural vegetation is being destroyed over much of the site.

The general objective of the experimental work described here was to learn whether it would be possible to recreate the seminatural grasslands being destroyed by mining operations. The experimental areas were flat plots beside the worked-out rake. In addition, plots were established on the steep south- facing slopes of the Edge itself. The mining company had attempted reinstatement of some areas that provided a valuable contrast with the experimental plots. The experiments were set up in February-September 1972 and covered about 2 ha. The experiments are continuing, in an attempt to establish the long-term fate of the sown vegetation — the only real criterion of success in reinstatement.

ENVIRONMENTAL PROBLEMS IN REINSTATEMENT
Derelict land reclamation has been the subject of intensive research in the United Kingdom so long that few waste materials or mine sites present insurmountable problems for reinstatement. The existence of large amounts of derelict land reflects the limited resources that have been made available for such work, rather than a lack of appropriate techniques. As a result of the work mainly of A. D. Bradshaw and his co-workers at Liverpool University since the early 1960s, even seemingly intractable wastes contaminated by toxic substances can be revegetated with ecotypes able to
RESTORING DERELICT LANDS IN GREAT BRITAIN 250
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. tolerate the presence of these materials. Toxic metals present such a special problem that there is probably no alternative to the use of this approach.

However, the treatment of nontoxic wastes has been much debated. The primary consideration has been to develop a recipe that can be used to produce a quick green cover, irrespective of its appropriateness as a long-term solution.

This search for universal recipes that eliminate, rather than accommodate, constraints on vegetation development has colored much of the research on derelict land reclamation. The major constraints on plant growth on derelict land, apart from possible toxic effects, are low nutrient status, low soil organic content, drought stress, and, if the material has been mechanically sorted, uniform particle size.

Abandoned mineral workings and waste materials — such as alkali wastes (Lee and Greenwood, 1976), quarries (Davies, 1976; Hodgson, 1982), and wastes contaminated with heavy metals (Holliday and Johnson, 1979; Johnson, 1978; Johnson et al., 1978) — often develop into important preserves for seminatural vegetation. The evidence suggests that there is no need to reclaim derelict or abandoned mineral workings, because an interesting vegetation will develop naturally, given sufficient time. But that stance neglects important changes that have rendered old sites unreliable as models. The changes affect all three main phases of the diversification of sown vegetation (Wathern, 1977):
Transfer. Propagules of the species must be dispersed from an established stand and arrive at the reinstated site.
Establishment. Propagules must become established and develop into mature plants.
Periodic recruitment. From time to time, new individuals derived from the original source or from the new population must become established, and the rate of recruitment should at least equal the death rate, if the population is to be sustained.

The abandoned mineral workings that are now biologically so interesting began their development in the early decades of this century. The landscape over much of the United Kingdom was different from that seen today.

Seminatural vegetation was the basis of much of the land use. Wood products were derived largely from broad-leaved woodlands, rather than softwood plantations. Pastoral agriculture relied on the use of seminatural grassland for grazing and hay. Thus, the countryside had a basic matrix of seminatural vegetation that could serve as a source of propagules for the diversification of abandoned mine workings. Transfer distances were short and within the capabilities of most species. The preponderance of wind-dispersed species on the alkali wastes highlighted by Lee and Greenwood (1976) indicates the importance of isolation, in that these waste
RESTORING DERELICT LANDS IN GREAT BRITAIN 251
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. areas are many miles from potential seed sources. Species with other types of dispersal mechanism usually move only a short distance with each generation and can move long distances only incrementally through a chain of suitable habitats. At present, however, species are becoming concentrated in fewer and fewer refuges. It is likely, therefore, that a quarry site abandoned today will be many miles from potential sources of propagules and that transfer distances will be too great for many species.

In addition, because mining and quarry operations have changed over the years, it is unlikely that modern sites will acquire the same ecological interest as the old sites. Large-scale workings are preferred, and mineral extraction produces uniform conditions over extensive areas. A modern limestone quarry lacks the physical diversity of an old quarry with its uneven face, rubble-strewn floor, and scattered waste heaps, which account for much of the biological interest. Such diversity can be created only by a program of treatment after quarrying operations have ceased. Many of the old skills have been lost, but Humphries (1977) describes how, for example, a quarry face could be treated.

The final objection to a laissez-faire attitude to site rehabilitation is not biological. Society will no longer tolerate the visual intrusion of a derelict mineral working or quarry over a long period in the hope that interesting habitats will eventually develop. Industry must be seen to be rehabilitating environments degraded by its activities. The challenge in derelict land reclamation is to reinstate a site so that its visual impact is reduced, while maintaining its potential for development into a biologically interesting area.

The traditional approach to reclamation of derelict land for visual amenity has been based on agricultural techniques. In the period 1967-1972, the mining company made considerable efforts to revegetate areas on Longstone Edge where extraction was complete or where there was a temporary halt in operations. The approach adopted was to sow agricultural seed mixtures (Lolium perenne, Dactylis glomerata, and Phleum bertolonii ) on raw subsoils with heavy fertilizer applications. It soon became apparent that this agricultural approach would provide no long-term solution. Whenever such grass-seed mixtures were sown, a green cover developed briefly. After a short period of vigorous growth, swards became moribund, the grasses showed signs of nutrient deficiency, and standing crops declined. On these free-draining soils lacking organic matter, nutrients, particularly nitrogen, were leached out.

Periodic applications of fertilizers became the main form of management used by the mining company. Clearly, there would be considerable advantages in developing some form of single-application treatment that would lead to stable vegetation. It would be more visually acceptable to the public and to the RESTORING DERELICT LANDS IN GREAT BRITAIN 252
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. controlling authority, the Peak Park Planning Board, and it would have obvious economic benefit.

Either of two approaches could be adopted: constraints on the productivity of high-yield species could be eliminated, or constraints on the system could be accommodated by creating some form of low-productivity system whose demands for nutrients would be better attuned to the ability of the system to supply them. The first approach has been used successfully in the revegetation of kaolin wastes in Cornwall. Kaolin wastes are nutrient-deficient, drought- stressed environments (Bradshaw et al., 1978), and fertilizers must be added regularly to maintain vegetation cover (Bradshaw et al., 1975). To overcome this problem, a biological source of nitrogen was sought. Rhizosphere fixation was inadequate to achieve the target of about 700 kg/ha estimated by Bradshaw et al. (1975) to be necessary to produce a self-sustaining system. The potential of nitrogen fixation by nodule bacteria in legumes was investigated in a series of trials. The results showed the use of legumes to be a means of overcoming nutrient limitations (Jefferies et al., 1981); Lupinus arboreus, the tree lupine, was the most effective species (Palaniappan et al., 1979), but is not native to the United Kingdom. Although sowing legumes is useful for reinstating some nutrient-deficient substrates, it was rejected for Longstone Edge, because the seminatural vegetation of Longstone Edge comprises calcareous grassland and Calluna heath, both of which are naturally nutrient-deficient systems. A productive agricultural sward created on the mineral working would be out of context with the surrounding vegetation.

Consequently, it was decided to attempt to produce a low-productivity vegetation in harmony with the existing nutrient-deficient system. The potential of these swards appeared great. Grime (1973a) has argued that environments moderately stressed by drought, nutrient deficiency, or grazing support diverse communities. In contrast, high-productivity systems are often species-poor.

These low-productivity grasslands could be expected to develop into diverse swards akin to the existing seminatural vegetation.

Local populations of plants could not be used as the sole source of propagules in this research, because of the difficulty of interpreting the experimental results. Because the natural arrival of propagules, mainly seeds, in the experimental plots would be a chance occurrence, it would be impossible to determine whether the failure of a particular species to invade an area resulted from a lack of propagules or from an inability to establish under the prevailing conditions. The ability of a sward to accommodate a new species is termed its receptivity (Wathern, 1976). Natural seed rain provides a flow of only a small irregular number of propagules into experimental areas; that makes it difficult to measure the receptivity
RESTORING DERELICT LANDS IN GREAT BRITAIN 253
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. of contrasted swards. Therefore, we diversified the swards artificially by introducing known quantities of seeds of a range of native species, hoping to create a visually pleasing community rapidly.

The decision to introduce native species deliberately raised an important issue, which has since achieved greater prominence. Deliberate introductions are discouraged by scientists, mainly because they disrupt natural distribution patterns, particularly of local ecotypes. However, because plant distribution in the United Kingdom has been modified drastically by human activity over a long period, that objection has little validity. In these experiments, only commonly occurring species were used, and they were collected from local seed sources whenever possible. This is an ethical consideration that requires more thorough discussion, because failure to transfer species artificially might doom them to local extinction. The ethics of introductions in the construction of natural ecosystems are discussed by van der Hoek (1982).

APPROACHES TO RECREATING SEMINATURAL

GRASSLAND
Experiments were established to investigate whether the yield of species on the Longstone Edge subsoils was related to relative growth rate (RGR).

Monocultures of Lolium perenne (RGR
max, 1.30), Festuca rubra (1.18), and F.
ovina (1.00) showed an inverse relationship between yield and RGR max after 3 years. Thus, the species with the greatest potential for growth, L. perenne, performed least well. Other trial areas were sown with two contrasting mixtures — one consisting of F. ovina and F. rubra and the other consisting of L. perenne, Dactylis glomerata (1.31), and Poa trivialis (1.40)— and monitored for 5 years.

After 1 year, the Lolium sward had almost twice the standing crop of the Festuca sward. After 2 years, both grasslands had increased in yield, but they did not differ significantly from one another. After 5 years, the Lolium sward had the same standing crop as it had after 2 years; in the intervening 3 years, it had become moribund. In contrast, the Festuca sward continued to increase its standing crop throughout the experiment. The results of this experiment, described more fully by Wathern and Gilbert (1979), confirmed that nutrient- deficient subsoils could be reinstated successfully with low-RGR grasses.

Grasses, however, provide only the general matrix of the community in seminatural grasslands. In the United Kingdom, many of the distinctive elements, and certainly the visually dominant species, are dicotyledons. Indeed, the seminatural grasslands of the carboniferous limestone are renowned for their floral richness. Consequently, in parallel with the trials RESTORING DERELICT LANDS IN GREAT BRITAIN 254
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. on low-productivity swards, the feasibility of recreating seminatural grasslands was investigated. Experiments were established as long-term trials in September 1972 and still cannot be considered complete. For the first 5 years, the plots were monitored fairly intensively, but they have since been assessed only infrequently.

One area of Longstone Edge provided an opportunity to investigate the effect of a grass seed mixture on the rate of invasion by native species under natural conditions. Three years before experiments on Longstone Edge began, the company had regraded a large area. Half this area had been sown with a grass seed mixture, and the remainder had been inexplicably left fallow. The vegetation of the two sections was surveyed, but no important differences in the number of adventive species could be found. Initial seeding with a mixture producing an open sward did not appear to affect the rate of diversification.

This suggested that the two-part objective of reinstatement — reducing visual impact without significantly affecting sward receptivity — could be achieved.

The second major conclusion from this work was that, although diversification occurred naturally, most of the adventives were species that produced large numbers of highly dispersable seeds. It appeared that, without deliberate introduction, the more characteristic species of climax limestone grasslands with larger, less readily dispersed seeds might not become established in the short term.

In the light of these observations, experiments were set up to establish the most appropriate means of diversifying swards and to determine the receptivity of contrasted swards to native species. The Festuca- and Lolium-based mixtures used in the low-productivity sward experiment described above were used.

Various sources of propagules of native species were investigated. Hand- collected seed was broadcast at known rates in the swards and fallow areas.

After 2 years, the Festuca sward contained 22 sown species; six species showed an establishment rate of over 10%. The Lolium and fallow areas contained 15 and 17 species, respectively, at low frequency. After 5 years, although the frequency of native species had increased in all areas as a result of additional recruitment, self-set seed, and vegetative spread, the differences among the Festuca, Lolium, and fallow plots were even more marked than they had been after 2 years. Thus, not only was the Festuca sward initially more receptive to native dicotyledons, but this increased receptivity was sustained over a long period. Adventive species also showed higher frequencies in Festuca areas and after 12 years included important elements of the limestone grassland flora.

In a separate experiment, sowings along a soil depth gradient showed that sward performance determined establishment rates for native species.

RESTORING DERELICT LANDS IN GREAT BRITAIN 255
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Establishment was greatest in swards of intermediate density. In such swards, competition is not severe, but there are enough plants to ameliorate the microclimate of developing seedlings. Dense swards eliminate native species by competitive exclusion, whereas failure rates of seedlings in open situations are high because of harsh environmental conditions (Wathern and Gilbert, 1978).

These experiments suggest that low-productivity swards increase the rate of diversification by native species and that sowing seed of native species by hand can create visually pleasing grasslands rapidly.

Other sources of native species were investigated. From casual observations on the site, it was noted that fragments of old turf had become established, particularly in compacted bulldozer tracks, providing small islands of seminatural grassland. Therefore, turf stripped from an area about to be reworked by the mining company was used in the diversification trials with an oversowing of the Festuca mixture (Gilbert and Wathern, 1980). The use of seed-rich topsoil on this site was not successful, but the approach should not be dismissed, in as much as it has been successful in other experimental areas.

Seed-rich Calluna litter (Gimingham, 1960) has been used to produce a Calluna heath community on an acidic subsoil road verge on highway A57 west of Sheffield (Gilbert and Wathern, 1976). Similarly, Farmer et al. (1982) and Tacey and Glossop (1980) have since used seed-rich topsoils for the reinstatement of mine areas.

To obtain additional information on the fate of man-made swards, a chronosequence of grasslands up to 150 years old was analyzed. A sample of 69 quadrats covering the complete range of grassland age was recorded. Soils were taken for laboratory analysis to determine whether edaphic factors were controlling the composition of the vegetation. For grasslands less than 25 years old, specifications for the original seed mixtures were known. The detailed procedures, particularly the seed mixtures, used to create the older grasslands could not be discovered. The major finding of this study was that soil nutrient status was the most important factor in controlling succession in urban grasslands (Wathern, 1976, 1980).

USES OF KNOWLEDGE
Of the ecosystem modifications listed by Wathern (1984), the most important changes resulting from development that influence reinstatement appear to be simplification, the opening of previously closed nutrient cycles, and the loss of vegetation types. The objective of revegetation RESTORING DERELICT LANDS IN GREAT BRITAIN 256
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. should be to redress that progression. Many of the objections to development are based on amenity, so it seems appropriate for reinstatement to aim at creating vegetation that has or can rapidly acquire amenity value. The application of agricultural technology to this problem, however, has often failed to achieve this objective. In the case reported here, ecological knowledge was used to develop a more subtle approach to reinstatement.

The revegetation of mineral workings on Longstone Edge was designed to satisfy five criteria: reinstated swards should be inexpensive to produce, result from a single simple treatment, require no intensive management, be visually pleasing, and resemble surrounding seminatural vegetation. The research involved field observations and theoretical considerations aimed at understanding the relationships between reinstated grasslands and the environment and then application of this information in devising a reclamation scheme. The application of ecological considerations pervaded the study.

However, ecological knowledge figured most prominently in four ways:
production of low-productivity swards, sward diversification, assessment of the potential of soil seed banks, and urban grassland succession.

Low-Productivity Swards
The most important concept used in this phase of the work involved the ecological importance of relative growth rate. Hunt (1970) and Grime and Hunt (1975) measured the RGR of 115 species in a uniform high-nutrient environment. The RGR
max of each species was determined. This study showed
that low-productivity environments are characterized by species with low RGR
max, whereas productive systems contain species with high RGRmax.

The agricultural species used previously in landscape reinstatement have high RGRmax. When sown on Longstone Edge, they initially produced a green cover, because artificial fertilizers provided the high concentrations of nutrients required for their growth. These skeletal soils, however, have a low organic content. Under these conditions, nutrients are not held within the system, but rapidly leach out. For the highly productive species to flourish, fertilizers must be added regularly. From Hunt's observations, it can be predicted that such soils would naturally carry species with low RGR. If those species were sown during reinstatement, the results should be more satisfactory than those obtained with agricultural species. Thus, the work on RGR provides the theoretical basis of the reinstatement of low-productivity swards. In turn, the experiments provide a field trial of the importance of growth rate in explaining plant distribution.

RESTORING DERELICT LANDS IN GREAT BRITAIN 257
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Sward Diversification
High RGR, which reflects the ability of a species to use resources that are freely available, is only one of the characteristics by which one plant can gain a competitive advantage over another. Structure and height affect competition for light, and the ability to produce litter that persists from one growing season to the next is an aid to occupying space when a plant is not actively growing.

Grime (1973a) argued that these four attributes can be amalgamated into an index of competitive ability. Species with high competitive index (''competitors" in Grime's terminology) grow in productive environments where resources, such as nutrients and water, are freely available. Such environments are species-poor, because competitors eliminate other species that are less able to exploit the resources. Environments where resources or physical and chemical constraints severely limit plant production, described as "stressed," are also species-poor. Only a few species are adapted to tolerate extreme conditions of, for example, drought or heavy-metal contamination. Such species are characterized by very low RGR
max.

Composition of a particular community, however, is dictated not only by the growth rates of the constituent species and by the physical and chemical characteristics that can limit the full realization of this potential. The degree of disturbance of the environment, such as grazing, is a third factor. Poverty of species is associated with a high incidence of disturbance.

Grime advocated a triangular ordination as the means of analyzing community composition, the axes being the relative importance of competition, stress, and disturbance. At the extremes, species-poor communities are present.

Diverse communities are found where competition, stress, and disturbance are all moderate, and they are characterized by species that exhibit the "C-S-R" strategy in Grime's terminology. C, S, and R stand for competitive, stress- resistant, and ruderal (weedy). Such plants are "confined to habitats in which competition is restricted to moderate intensities by the combined effects of stress and disturbance . . ." (Grime, 1977). The Longstone Edge reinstatement areas have this combination of features, so they should ultimately develop into rich communities of C-S-R forbs. The major representatives of C-S-R strategies in the U.K. flora are small tussock grasses, small deep-rooted forbs, small stoloniferous species, forbs with short rhizomes, legumes, and small sedges and rushes (Grime, 1979).

The major grasses used successfully to form the low-productivity sward on Longstone Edge were two small tussock-formers, Festuca ovina and RESTORING DERELICT LANDS IN GREAT BRITAIN 258
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. a nonstoloniferous variety of F. rubra (var. fallax). The grasslands were diversified by using the seeds of wildflower species that could be collected in the area. Many achieved high rates of seedling establishment and conversion to mature plants. Several of the species that Grime (1979) lists among a short illustrative group of U.K. C-S-R strategists established successfully from seed in the Longstone Edge trials. In contrast, several tall herb species included for comparative purposes showed initial high rates of seedling establishment, but then failed to survive.

Thus, Grime's concepts of community ecology are directly relevant to derelict land reclamation. Although the concepts were known in general terms at the start of this research, the detailed design of the experiments was dictated by pragmatic considerations, such as seed availability. Before the trials, it was known that diverse natural grasslands were characterized not only by species with low growth rates, but also by environmental constraints on plant productivity. Such considerations influenced the basic suppositions underlying the diversification trials. The substantial theoretical developments that took place during the project permitted a more thorough and systematic analysis of the results and set the conclusions in a broader context.

Soil Seed Banks
Established ecological theory was used to appraise the potential of soil as a source of propagules for diversification. Initially, only the general characteristics of soil seed banks had been determined. Seed banks from beneath particular types of seminatural vegetation have now been described (Brown and Oosterhuis, 1981; Donelan and Thompson, 1980; Thompson and Grime, 1979), but when our experiments were being formulated, most of the literature was related to the weed populations of arable systems (see Harper, 1977).

Spray (1970) and Grime (1978) characterized the soil seed bank of tall herb communities. On the basis of high numbers of some native species, Spray suggested that seed-rich topsoil could be used to diversify reinstated vegetation.

Experiments were undertaken to establish whether his techniques could be used on a large scale. Soil from beneath limestone turf was taken from a part of the site about to be worked for minerals. In laboratory germination tests, this soil was found to contain seeds of various native species. The field trials were not successful, but the reasons for the failure have never been adequately explained.

RESTORING DERELICT LANDS IN GREAT BRITAIN 259
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Succession in Urban Grassland
Sown vegetation is not static. Over the course of time, it changes as a result of a complex array of interacting factors. Succession is a phenomenon that must be accommodated by any vegetational scientist involved in derelict land reinstatement. In the United Kingdom, few data on grassland successions are available. Wells (1965) monitored the changing composition of a sown agricultural grassland. Of particular interest, however, is the description of a complex of grassland types that have developed on formerly arable land on Salisbury Plain after various periods of abandonment (Wells et al., 1976).

Changing composition was associated with nutrient accumulation within the ecosystem.

Analysis of a chronosequence of sown urban grasslands provided some insight into the fate of vegetation created for visual amenity. In Sheffield, extensive areas of grassland have been produced for public open space around municipal housing in recent years. The age of the grasslands, up to 25 years, and the seed mixtures originally sown were known from the records of the local authority. Older grasslands were discovered around large country houses in the vicinity. In the case of the latter grasslands, the composition of the sowing mixtures could not be determined, but the chronosequence could be extended over a 150-year period with estate records. The results of this survey are discussed by Wathern (1976, 1980).

Diversification is the most important successional phenomenon for revegetation studies. The process in natural successions is accompanied by the accumulation of materials, particularly nutrients, within the system (Wells et al., 1976). In the urban grassland chronosequence, no increased diversity with age could be detected; older swards did not contain more species. However, the soils beneath these grasslands are subject to leaching. Grime (1973a) has shown that the number of species in seminatural grassland in the Sheffield region is inversely related to the acidity of the soil. In as much as few species tolerate extremely acidic soils, a chronosequence in which there is a decrease in pH is likely to be one in which the number of potential recruits also decreases.

Grime's data make it possible to separate the effects of diversification from pH changes. For each quadrat in the urban grassland survey, the number of species present was divided by the number of species anticipated in a soil with the pH of that quadrat. Values of a relative diversity index (Wathern, 1976), when plotted against age, showed that as a grassland gets older it acquires more of the species that are capable of growing at the appropriate pH. The results confirm the conclusion that succession is important in molding the composition of sown grasslands and indicate that many species invade
RESTORING DERELICT LANDS IN GREAT BRITAIN 260
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. only after long periods. Invasion of Calluna into these grasslands under conditions of natural seedfall, for example, takes about a century.

The second major conclusion with important implications for the diversification trials concerned the significance of soil nutrient status in determining composition. This was assessed with the Rumex acetosa seedling bioassay developed by Rorison (1967), a standard technique used in Sheffield.

Increasing relative diversity was associated with decreasing nutrient status.

These results are at variance with the observations of Wells et al. (1976), but it should be noted that soils in the urban grassland chronosequence initially had high nutrient concentrations as a result of fertilizer applications to topsoils. The implications with respect to the triangular ordination of community characteristics are discussed below. The observations suggested that nutrient deficiency in the limestone subsoil of the mineral workings on Longstone Edge should not be regarded as a major constraint on vegetation development, but potentially as an asset that should lead ultimately to high diversity if an appropriate method can be developed.

SOURCES OF ECOLOGICAL INFORMATION
The overall objective of the Longstone Edge trials was to create seminatural vegetation resembling the surrounding diverse grasslands on the nutrient-deficient subsoils of the reinstated mineral workings. The research had two basic assumptions: sward production depends on the relation of soil nutrient status to the growth rate of constituent species, and diversity depends on sward productivity. Each experiment was established to address these factors, with the intention of combining the results in the development of an approach for the reinstatement of this and similar sites. The important ecological sources in this case, therefore, were those which either aided identification of soil nutrient status and relative growth rate as key issues in the research or influenced experimental design.

As an illustration of the application of ecological theory to the resolution of ecological problems, the Longstone Edge revegetation trials are unusual. The development of the experimental work preceded the publication of much of the ecological theory to which it is most closely allied. This apparent anomaly can be explained only with a thorough understanding of the situation in which the work was conducted.

The evolution of an experimental research project is a complex process and one in which it is exceedingly difficult to unravel the detailed history of an idea. In a research environment like a university, many important innovations begin in a few chance remarks, in discussions with colleagues RESTORING DERELICT LANDS IN GREAT BRITAIN 261
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. working in related fields, or in questions raised in research seminars. It is often difficult later to apportion progress among these causal factors.

Chance occurrences often change the course of research or provide valuable information that permits a more comprehensive analysis of results. For example, the soil-gradient experiment described earlier provided much information on the effect of sward performance on the establishment of native species. The experiment was not part of the original scheme; in some ways, its establishment was accidental. The company requested the use of the low- productivity sward mixture on this part of the site and, because surplus seeds of native species were available, this was added almost as an afterthought.

Therefore, although planning is an essential component of good research, no program should be so intensely planned that there is no room for exploiting chance occurrences.

The factors that influenced the evolution of approaches to the reinstatement of disturbed land can be considered in four categories: ecological facts, ecological theory, analog studies, and expert judgment.

Ecological Facts
The major impetus in experimental design came from the field observations of research workers on site, who analyzed unsuccessful attempts at reclamation undertaken by the company, and from a detailed appraisal of the likely causes of failure in the light of ecological knowledge. A number of published and unpublished sources aided the study of low-productivity swards and grassland diversification.

The experimental grass sowings on Longstone Edge were designed to assess the performance of different species on limestone subsoil and to investigate the feasibility of creating low-productivity swards. Initial species selection therefore had to take account of the commercial availability of seed, as well as of potential productivity. Hunt (1970) compiled data on the relative growth rate of a range of species found in different habitats in the Sheffield region. These data confirmed the differences in growth rates of the species selected for the grass monoculture and the low-productivity sward experiments and allowed these differences to be quantified. The urban grassland survey described above was also important in species selection. The results showed that the younger grasslands on soils with high nutrient status were dominated by Lolium perenne swards, whereas older nutrient-poor soils carried grasslands dominated by fine-leaved Festuca species.

Although the older urban grasslands were more interesting, there was no evidence from an initial analysis of the data that they were more diverse than younger grasslands. Not until after the publication of the Grime RESTORING DERELICT LANDS IN GREAT BRITAIN 262
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. (1973a) paper, after the Longstone Edge trials were established, were the data reinterpreted to take account of the pH effects discussed above. The effect of modifications of soil nutrients on species composition in the United Kingdom, however, has been known for many years, with respect to upland grasslands (Milton, 1940), lowland meadows (Brenchley and Warington, 1958), dunes (Willis, 1963), and limestone grasslands (Jeffrey, 1971; Jeffrey and Pigott, 1973).

The major conclusion from these experiments is that grassland composition can be manipulated by changing nutrient status with fertilizer applications. The park grass experiment described by Brenchley and Warington (1958) resulted in the production of grassland types totally different from a common sward over the course of a century of contrasted fertilizer applications to trial plots. Most important, these investigations showed the high diversity of swards on soils of low nutrient status. This information motivated our expectation that rich grassland would develop on the nutrient-poor subsoils of the reinstated areas.

Data on the relative growth rate and competitive ability of different species proved to be the most valuable unpublished materials available to the study.

The interrelationships among environmental characteristics, competitive ability, and plant strategies were already being investigated by members of the Unit of Comparative Ecology at Sheffield University. Grime proposed that nutrient-rich communities were species-poor because of the competitive dominance of the constituent species. A competitive index, quantifying this dominance, had been developed and calculated for 441 species present in the Sheffield region. The index was based on maximal height, structure, and maximal accumulation of persistent litter. The data collected by Hunt (1970) on relative growth rates were also included. Only incomplete data related to 115 species were available, so growth rate was not included formally in the index. The index confirmed the high competitive ability of agricultural grasses compared with that of the low- productivity sward mixture. In more recent discussions of competition (Grime, 1973a), growth rate has been incorporated into the index as a result of the availability of additional data.

The Unit of Comparative Ecology at Sheffield has been collecting systematic data on the characteristics of grasslands in the Sheffield region for many years. These data were used to produce a computerized data base on grasslands and their constituent species, but the first analysis of grasslands in the area (Lloyd et al., 1971) was too general to be valuable during this study.

The data were also used to produce an ecological atlas of common grassland plants (Grime and Lloyd, 1973), including detailed information on the environmental characteristics of the habitats in which each of 94 RESTORING DERELICT LANDS IN GREAT BRITAIN 263
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. species occurs. This atlas has been described as a practical aid for landscape managers involved in, for example, revegetation, but that appears to be a post hoc rationalization of its function. Selection of species for landscape reinstatement is determined by pragmatic considerations, such as seed availability, ease of establishment, and likelihood of success. These considerations are not addressed in the atlas. For example, productivity of the systems in which a species occurs, acknowledged elsewhere by Grime as a major factor in determining distribution, is not considered. The atlas therefore has the wrong emphasis to be valuable in revegetation.

Consequently, the information on distribution was not used in the revegetation trials. Species from a wide range of habitats were selected for the diversification experiments. It was felt that this approach would be better, because it would provide information on the conditions necessary not only for success, but also for failure.

Ecological Theory
The results of the seral study on urban grasslands and the subsequent diversification trials illustrate well the implications of the theory of plant strategies and community ecology for reinstatement studies. There were important changes in the status of this theoretical background over the course of the work described here. At the outset, the theory was still being developed by Grime in Sheffield. It was not published until most of the experiments had been set up, although the general tenor of the ideas was known. Thus, during formulation of the reinstatement work, the urban-grassland study and papers on the relationship between diversity and nutrient status were more important.

Preliminary analysis of the experimental work began in 1974, after publication of the basic theoretical concepts (Grime, 1973a,b, 1974). The theory has been further developed by Grime (1977, 1979), but not all aspects have been accepted without dissent (see, for example, Newman, 1973). With this more comprehensive theoretical base, the results of the experimental work became more important and could be set within a broader context. The results demonstrate that diverse vegetation can be created artificially under conditions of moderate stress, as the theory predicts, and that the techniques used constitute a worthy strategy for reinstatement of disturbed land for visual amenity.

Analog Studies
No reports of previous attempts to reinstate mineral workings to seminatural vegetation were discovered during a systematic search. Work on RESTORING DERELICT LANDS IN GREAT BRITAIN 264
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. diversification of road verges, however, is relevant to the study. Additional diversification trials were established on road verges in the Sheffield area, because they have the potential to make a major contribution to the reserves of seminatural habitats in the United Kingdom (Mellanby, 1974). The results were reported by Wathern and Gilbert (1978). A number of previous attempts at establishing native species in verges have been reported in the conservation press, and these were investigated in detail.

Most cases were in the form of "rescue" operations, in which populations of rare species were threatened by highway construction. Generally, mature plants were returned to the verge after a period in a nursery to bridge the construction phase. The most spectacular success has been the mass planting of native Narcissus pseudonarcissus (daffodil) on the M40. There were a few attempts at more general diversification. Calluna turf was used successfully for reinstatement at the Devil's Punch Bowl, Surrey, after roadwork and at Cumbernaud New Town. Commercially available herbs have been used on road verges in southeastern England. The results of this work have never been published. Significantly, some members of voluntary organizations had suggested more general diversification of highway verges in their areas, but such proposals were invariably abandoned after objections on the grounds that distribution patterns would be disrupted. The main value of the analog studies was in confirming that artificial diversification could be achieved. In general, the labor-intensive approaches were rejected as being inappropriate for large- scale projects.

Experimental work in the Netherlands on landscape reinstatement was of particular interest, but it was not seen until the Longstone Edge trials had been established. The Dutch experience was useful in interpretation of the experimental results, because a similar approach had been adopted independently in many instances. For example, the problem of low-productivity swards on sandy soils was investigated early (Hoogerkamp, 1971), and trials on the establishment of Calluna on highway verges had shown the importance of a nurse crop in seedling establishment, as was found in the United Kingdom (Gilbert and Wathern, 1976).

Expert Judgment
It should be clear from this discussion that a university provides a unique atmosphere for research. The interplay of personnel leads to a rapid evolution of ideas. The Longstone Edge study was undertaken in the Department of Landscape Architecture of Sheffield University as a doctoral research project.

The supervisor was an experienced ecologist, and the research student a botany graduate. There were also important links with the Botany Department, in particular its Unit of Comparative Ecology.

RESTORING DERELICT LANDS IN GREAT BRITAIN 265
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

These contacts occurred during a formative and productive period for the Unit, when ideas concerning plant strategies and community ecology were being developed. The work on the comparative ecology of individual species and the composition of plant communities might seem esoteric, but the revegetation work on Longstone Edge has shown its relevance to reinstatement.

Although not published, information was freely available to the study, and informal discussions around this work were useful in the development of ideas concerning reinstatement.

Consultations with other organizations were not generally important in the evolution of the work. However, discussions with experts outside the university system, particularly government scientists in the Institute of Terrestrial Ecology (ITE) and the Nature Conservancy Council (NCC), took place. In addition, a British Council travel fellowship enabled the work undertaken in the Netherlands to be discussed with government scientists at IBS Wageningen in 1974.

The responses of research scientists differed. Most research workers were concerned primarily with investigating the effects of management on grassland composition. A few research scientists at ITE, however, were also interested in creating seminatural grassland; indeed, they set up experimental work in 1973 and 1978 (Wells, 1983). A guidance manual on diversification has also been prepared (Wells et al., 1981). The discussions in 1971 generally concerned the need to develop techniques, rather than how this might be done.

Sowings of wildflower species became quite common in the Netherlands at one time (Cole and Keen, 1976). During the discussions with Dutch scientists, however, the ethics of the approach were called into question. This hinged not on theoretical considerations of disrupted distributions of different ecotypes, but on a more fundamental point: whether people should interfere in the development of an ecosystem.

The attitude of conservation officers in the United Kingdom was defensive and antagonistic. Many saw within the techniques a potential diminution in their ability to object to development proposals on conservation grounds, because they believed that any suggestion that plant communities could be recreated after development would undermine their position. This attitude reflected a failure to understand the concepts underlying the techniques. There is no suggestion of constructing communities that contain all manner of rare species.

The approach seeks merely to redress somewhat the losses of seminatural vegetation that are occurring. The alternative is that yet more highly productive grazing land will be created in reinstatement, irrespective of location. This misconception concerning the overall approach, unfortunately, continues.

RESTORING DERELICT LANDS IN GREAT BRITAIN 266
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

CONTRIBUTION OF ECOLOGICAL KNOWLEDGE
A Peak Park Planning Board official has described the Longstone Edge trials discussed here as "the most exciting piece of reinstatement work which has been undertaken within the area." The excitement had two sources. First, the resulting sward is diverse and contains many of the elements of seminatural calcareous grassland. The presence of these characteristic species is the result of both deliberate introductions of propagules and chance invasions. Even 12 years after establishment, the grasslands are still acquiring additional species. Second, development of these techniques marks an important departure from previous approaches to reinstatement.

Formerly, reclamation was based on agricultural technology with remedial measures, so the environment was modified primarily by fertilization to support the species sown. The scheme adopted on Longstone Edge was based on a more objective ecological approach. In the simplest terms, it can be described as an attempt to sow species capable of surviving under the environmental regime of the limestone subsoils. The change in emphasis was possible because the objectives of reinstatement were reconsidered and ecological theories concerning the control of community composition were reappraised.

Development of theory was particularly useful in two respects: it could be used to explain the composition of seminatural communities in the area and the grassland chronosequence studied early in the research (it was on the basis of these field observations and the literature on the effects of fertilizers on grassland diversity that the more intuitive experiments were established), and, during interpretation of the experimental results, it enabled success and failure to be explained, rather than merely noted.

There is one major criticism of the work. Covering no more than 2 ha, the plots are at best large trials. With one exception, they have yet to be scaled up.

Although increased scale should not affect the outcome of reinstatement, it would establish whether the approach is usable on an industrial scale. The possibility of developing a "wildflower" mixture for reinstatement schemes based on commercial sources of native species has been considered and rejected. Commercial seed is prohibitively expensive, but there are more serious ecological objections to such a mixture. The research was developed in an attempt to avoid the previous approach — namely, producing a recipe to be applied in all situations — even if it could guarantee a visually attractive grassland. The species that are available in commercial quantities constitute inappropriate ecotypes. For example,
RESTORING DERELICT LANDS IN GREAT BRITAIN 267
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. native populations of Lotus corniculatus figured prominently and successfully in the experiments, and this species seems to be useful in reinstatement, whereas commercial material is a robust, erect ecotype from central Europe that is selected for its high yield and ease of harvesting and is quite unlike the calcareous grassland ecotype found on Longstone Edge. Hand collection of local ecotypes is far more satisfactory, but its feasibility and, more important, its impact on local populations are not known. Some form of bulking up might be required, because large quantities of seed could be involved. Wells (1983), for example, recommends up to 2 kg/ha for some species. These practical considerations must be resolved.

Work on Calluna has been scaled up by other workers, and the creation of Calluna heathland has generated considerable interest. Large areas of moorland and heath over much of the United Kingdom are degraded and in need of reinstatement. In addition, serious visual intrusion has resulted from the use of traditional methods of reinstatement in the revegetation of such developments as pipeline rights of way. Reinstatement of Calluna heathland is now feasible on a large scale as a result of further field trials (Meaden, 1984; Moorland Restoration Project, 1983; Putwain et al., 1982).

The Longstone Edge research has stressed the necessity for a long-term perspective in analyzing the results of reinstatement schemes. Results over short periods can have misleading implications for small-scale pilot schemes.

Differences that were encountered in these trials over time could be explained by applying theories of community ecology.

The main value of the case described here concerns the interrelationship between theoretical ecology and applied ecology. The study has provided experimental support for theories of community ecology, and theoretical considerations, particularly of plant strategy and community ecology, help to weld the experimental work into a coherent whole. This unity converts what might have been regarded merely as an interesting set of empirical observations on a set of reinstatement trials on a small site in Derbyshire into a radical alternative strategy for the revegetation of derelict and despoiled land.

ACKNOWLEDGMENTS
The work described here was undertaken in collaboration with my research supervisor, Oliver Gilbert, and formed the basis of a doctoral thesis submitted to the University of Sheffield in 1976. The work was funded by a research studentship from the Natural Environment Research Council. The assistance of the mining company, Laporte Industries Ltd.,
RESTORING DERELICT LANDS IN GREAT BRITAIN 268
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. and Jack Harwood in providing facilities for this work is gratefully acknowledged. The research should be regarded as the joint work of Oliver Gilbert and me, but the observations and interpretation of events included in this report are my responsibility.

REFERENCES
Balme, O. E. 1953. Edaphic and vegetational zoning on the carboniferous limestone of the Derbyshire Dales. J. Ecol. 41:331-344.

Bradshaw, A. D., and M. J. Chadwick. 1980. The Restoration of Land. Blackwell, Oxford, Eng.

Bradshaw, A. D., W. S. Dancer, J. F. Handley, and J. C. Sheldon. 1975. The biology of land revegetation and the reclamation of the china clay wastes in Cornwall. Pp. 363-384 in M.

J. Chadwick and G. T. Goodman, eds. The Ecology of Resource Degradation and Renewal. Br. Ecol. Soc. Symp. 15. Blackwell, Oxford, Eng.

Bradshaw, A. D., R. N. Humphries, M. S. Johnson, and R. D. Roberts. 1978. The restoration of vegetation on derelict land produced by industrial activity. Pp. 249-274 in M. W. Holdgate and M. J. Woodmand, eds. The Breakdown and Restoration of Ecosystems. Plenum, New York.

Brenchley, W. E., and K. Warington. 1958. The Park Grass Plots at Rothamsted, 1856-1949.

Rothamsted Experimental Station, Harpenden, Eng.

Brown, A., and L. Oosterhuis. 1981. The role of buried seeds in coppiced woods. Biol. Conserv. 21:19-38.

Cole, L., and C. Keen. 1976. Dutch techniques for the establishment of natural plant communities in urban areas. Landscape Design 116:31-34.

Davies, B. N. K. 1976. Wildlife, urbanisation and industry. Biol. Conserv. 10:249-291.

Donelan, M., and K. Thompson. 1980. The distribution of buried and viable seeds along a successional series. Biol. Conserv. 17:297-312.

Farmer, R. E., M. Cunningham, and M. A. Barnhill. 1982. First-year development of plant communities originating from forest topsoils placed on southern Appalachian minesoils. J.

Appl. Ecol. 19:283-294.

Gilbert, O. L., and P. Wathern. 1976. Towards the production of extensive Calluna swards.

Landscape Design 114:35.

Gilbert, O. L., and P. Wathern. 1980. The creation of flower-rich swards on mineral workings.

Reclam. Rev. 3:217-221.

Gimingham, C. H. 1960. Biological flora of the British Isles: Calluna vulgaris (L.) Hull. J. Ecol. 48:455-483.

Grime, J. P. 1973a. Control of species density in herbaceous vegetation. J. Environ. Manage. 1:151-167.

Grime, J. P. 1973b. Competitive exclusion in herbaceous vegetation. Nature 242:344-347.

Grime, J. P. 1974. Vegetation classification by reference to strategies. Nature 250:26-31.

Grime, J. P. 1977. Evidence for the existence of three primary strategies in plants and its relevance to ecological and evolutionary theory. Am. Na. 111:1169-1194.

Grime, J. P. 1978. Interpretation of small-scale patterns in the distribution of plant species in space and time. Pp. 101-124 in A. H. J. Freyson and J. W. Woldendorp, eds. Structure and Functioning of Plant Populations. Elsevier-North Holland, Amsterdam.

Grime, J. P. 1979. Plant Strategies and Vegetation Processes. John Wiley & Sons, Chichester, Eng.

RESTORING DERELICT LANDS IN GREAT BRITAIN 269
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Grime, J. P., and R. Hunt. 1975. Relative growth rate, its range and adaptive significance in a local flora. J. Ecol. 63:521-534.

Grime, J. P., and P. S. Lloyd. 1973. An Ecological Atlas of Grassland Plants. Edward Arnold, London.

Harper, J. L. 1977. Population Biology of Plants. Academic Press, London.

Hilton, K. J. 1967. The Lower Swansea Valley Project. Longmans Green, London.

Hodgson, J. G. 1982. The botanical interest and value of quarries. Pp. 3-11 in B. N. K. Davies, ed.

Ecology of Quarries. ITE Symposium II. Institute of Terrestrial Ecology, Cambridge, Eng.

Holliday, R. J., and M. S. Johnson. 1979. The contribution of derelict mineral and industrial sites to the conservation of rare plants in the United Kingdom. Min. Environ. 1:1-7.

Hoogerkamp, M. 1971. Probleme bei Ansaaten an Strassenrändern. Rasen 2:85-86.

Humphries, R. N. 1977. An ecological approach to the revegetation of limestone quarries in the United Kingdom. Pp. 2-39 in E. J. Perry and N. A. Richards, eds. Limestone Quarries.

Allied Chemical Corporation, Jamesville, N.Y.

Hunt, R. 1970. Relative Growth Rate: Its Range and Adaptive Significance in a Local Flora. Ph.D. thesis, University of Sheffield.

Jefferies, R. A., A. D. Bradshaw, and P. D. Putwain. 1981. Growth, nitrogen accumulation and nitrogen transfer by legume species established on mine spoils. J. Appl. Ecol. 18:945-956.

Jeffrey, D. W. 1971. The experimental alteration of a Kobresia-rich sward in Upper Teesdale. Pp. 79-89 in E. Duffey and A. S. Watt, eds. The Scientific Management of Animal and Plant Communities for Conservation. Br. Ecol. Soc. Symp. 11. Blackwell, Oxford, Eng.

Jeffrey, D. W., and C. D. Pigott. 1973. The response of grasslands on sugar-limestone in Teesdale to application of phosphorus and nitrogen. J. Ecol. 61:85-92.

Johnson, M. S. 1978. The botanical significance of derelict industrial sites in Britain. Environ.

Conserv. 5:223-238.

Johnson, M. S., P. D. Putwain, and R. J. Holliday. 1978. Wildlife conservation value of derelict metalliferous mine workings in Wales. Biol. Conserv. 14:131-148.

Lee, J. A., and B. Greenwood. 1976. The colonisation by plants of calcareous wastes from the salt and alkali industry in Cheshire, England. Biol. Conserv. 10:131-149.

Lloyd, P. S., J. P. Grime, and I. H. Rorison. 1971. The grassland vegetation of the Sheffield region.

J. Ecol. 59:863-886.

Meaden, D. P. 1984. The Restoration and Creation of Heather Moorland Vegetation. Ph.D. thesis, University of Liverpool.

Mellanby, K. 1974. The changing environment. Pp. 1-6 in D. L. Hawksworth, ed. The Changing Flora and Fauna of Britain. Academic Press, London.

Milton, W. E. J. 1940. The effects of manuring, grazing and cutting on the yield, botanical and chemical composition of natural hill pastures. I. Yield and botanical composition section.

J. Ecol. 28:326-356.

Moorland Restoration Project. 1983. Phase 2 Report: Re-vegetation Trials. Moorland Restoration Project, Bakewell, Eng.

Newman, E. I. 1973. Competition and diversity in herbaceous vegetation. Nature 244:310.

Palaniappan, V. M., R. H. Marrs, and A. D. Bradshaw. 1979. The effect of Lupinus arboreus on the nitrogen status of china clay wastes. J. Appl. Ecol. 16:825-831.

Pigott, C. D. 1962. Soil formation and development on the carboniferous limestone of Derbyshire. I.

Parent materials. J. Ecol. 50:145-156.

Putwain, P. D., D. A. Gillham, and R. J. Holliday. 1982. Restoration of heather moorland RESTORING DERELICT LANDS IN GREAT BRITAIN 270
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. and lowland heathland with special reference to heathlands. Environ. Conserv. 9:225-235.

Rorison, I. H. 1967. A seedling bioassay on some soils in the Sheffield area. J. Ecol. 55:725-741.

Spray, M. 1970. Management of roadsides, hedgerows and grasslands for nature conservation and amenity. Pp. 28-33 in P. J. Grime, ed. People and Plants. Derbyshire Naturalists' Trust, Matlock, Eng.

Tacey, W. H., and B. L. Glossop. 1980. Assessment of topsoil handling techniques for rehabilitation of sites mined for bauxite within the Jarrah Forest of Western Australia. J. Appl. Ecol. 17:195-201.

Thompson, K., and J. P. Grime. 1979. Seasonal variation in herbaceous seed banks. J. Ecol. 67:893-922.

Van der Hoek, D. 1982. New man-made nature. Pp. 287-288 in S. P. Tjallingii and A. A. van der Veer, eds. Perspectives in Landscape Ecology. Centre for Agricultural Publishing and Documentation, Wageningen, Neth.

Wathern, P. 1976. The Ecology of Development Sites. Ph.D. thesis, University of Sheffield.

Wathern, P. 1977. The ecology of herb establishment and survival in swards and its relevance to grassland reinstatement. Rasen 8:102-108.

Wathern, P. 1980. The creation of natural vegetation after development. Pp. 5-20in Proceedings of the Second International Conference on Environment Protection. Vol. II. Scientific Society for Building, Budapest.

Wathern, P. 1984. Ecological impact assessment. Pp. 273-292 in B. D. Clark, A. Gilad, R. Bisset, and P. Tomlinson, eds. Perspectives on Environmental Impact Assessment. Reidel, Dordrecht, Neth.

Wathern, P., and O. L. Gilbert. 1978. Artificial diversification of grassland with native herbs. J.

Environ. Manage. 7:29-42.

Wathern, P., and O. L. Gilbert. 1979. The production of grassland on subsoil. J. Environ. Manage. 8:269-275.

Wells, T. C. E. 1965. Changes in the botanical composition of a sown pasture on the chalk in Kent, 1956-65. J. Br. Grassl. Soc. 22:277-281.

Wells, T. C. E. 1983. The Creation of Species Rich Grasslands. Pp. 215-232 in A. Warren and F. B.

Goldsmith, eds. Conservation in Perspective. John Wiley & Sons, Chichester, Eng.

Wells, T. C. E., J. Sheail, D. F. Ball, and L. K. Ward. 1976. Ecological studies on the Porton Ranges. Relationships between vegetation, soils and land-use history. J. Ecol. 64:589-626.

Wells, T. C. E., S. A. Bell, and A. Frost. 1981. Creating Attractive Grasslands Using Native Plant Species. Nature Conservancy Council, Shrewsbury, Eng.

Willis, A. J. 1963. Braunton Burrows: The effects on the vegetation of addition of mineral nutrients to the dune soils. J. Ecol. 51:353-374.

RESTORING DERELICT LANDS IN GREAT BRITAIN 271
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Committee Comment
An increasingly important activity of environmental scientists is the reinstatement and reclamation of derelict land that has been modified for various reasons with various outcomes. Reclamation practices are most highly developed in the densely populated countries of western Europe, where human modifications affect virtually every piece of terrain. Much of this experience is summarized in a valuable book by A. D. Bradshaw and M. J. Chadwick, The Restoration of Land, published in 1980. Increasing interest in this subject in the United States is reflected in the launching of a new journal, Restoration and Management Notes, published by the University of Wisconsin Arboretum. As in most fields of applied ecology, early attempts at restoration relied almost entirely on an empirical approach — try a variety of approaches and see which ones work best. They were also highly constrained by cost, availability of materials for revegetation purposes, and the objectives expressed by society for reinstatement projects.

Happily, the combination of several factors has resulted in important changes in the practice of restoration. Society is placing increasing value on land restoration, plant ecology has seen substantial recent development of relevant theory and data, and the goals of restoration have become much more sophisticated than merely getting something pretty and green. The project at Longstone Edge is an admirable example of applied ecology. It showed the importance of drawing on a variety of sources — such as ecological facts, ecological theories, analog studies, and the reasoned judgment of persons familiar with the local areas— if sound solutions are to be achieved.

The Longstone Edge project took its direction from the recognition that most derelict land in Britain occurs on nutrient-poor soils and that attempts to make such areas fertile in the long run require repeated fertilization, which, because of its cost, is seldom feasible. But it was observed that diverse natural grasslands often occurred on nutrient-poor sites; apparently, therefore, the site could be ecologically interesting without being productive. These important observations played a role in the development of the powerful experimental design that was used in this project.

Species-rich swards can be established on nutrient-poor soils when constituent species are poor competitors characterized by low maximal growth rates and when little competitive exclusion of species occurs. Under rich soil conditions, a few of the strongest competitors often come to dominate — that is, interestingly, an empirical observation for which there is no satisfactory theoretical explanation. We do not understand competition among plants in enough detail to be able to identify the mechanisms
RESTORING DERELICT LANDS IN GREAT BRITAIN 272
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. of interactions. Restoration projects might turn out to be especially valuable for gaining further knowledge about these fundamental processes. They are carried out on a scale larger than is usually possible in purely academic research and, given current interest in them, are likely to be followed long enough to reveal aspects of the processes whose outcomes require decades and longer to be seen.

Better understanding of competition among plants might offer new insights into the practical problems of reinstatement. For example, the Longstone Edge project used primarily information on relative growth rates to assess long-term success of species on the site. Other factors that influence interactions among plants, such as growth form and the relative allocation of resources to different tissues and organs, might also be used in the selection of species for seeding of experimental plots.

Similarly, since the initiation of the Longstone Edge project, considerable progress has been made in understanding the role of particular combinations of nutrients in determining the outcome of competition among plants, particularly algae. The experimental results and a theoretical interpretation of them have been provided by Tilman (1982). Because derelict lands are likely to be diverse in their nutrient status, even if they are mostly on poor soils, better understanding of the importance of particular combinations of soil nutrients for the success of competing plant species should enable both better selection of species for colonization and the cultivation of specific genotypes for use in particular circumstances.

Restoration ecology might also contribute substantially to our knowledge and understanding of seed banks in the soils and their roles in the development of vegetation after disturbance. In the Longstone Edge project, important use was made of topsoil with seeds; this technology appears to have great potential.

New advances in the understanding of the causes of seed longevity and the specific conditions that result in termination of dormancy in various species might reveal which species have seeds that persist well in different circumstances.

An important lesson of the Longstone Edge project is that in restoration ecology there are unlikely to be formulas of broad applicability that can be used in restoring most derelict lands. Solving problems for a particular site will inevitably require knowledge of its environmental conditions and of the genotypes of local plants that are potential colonizers of the site. Nonetheless, general concepts— such as the relationships among soil nutrients, growth rates, and competition — are likely to be broadly applicable. In addition, society is likely to have multiple goals for reclaimed lands. Many sites in North America, for example, unlike those in Britain, are on potentially valuable agricultural land or are in highly productive sites that are to be restored to some seminatural state. Thus, apart from highly
RESTORING DERELICT LANDS IN GREAT BRITAIN 273
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. specific knowledge of local conditions, continued developments in the theory of competition, plant succession, seed ecology, and plant nutrition are likely to remain vital to restoration ecology far into the future.

References
Bradshaw, A. D., and M. J. Chadwick. 1980. The Restoration of Land. Blackwell, Oxford, Eng.

Tilman, D. 1982. Resource Competition and Community Structure. Princeton University Press, Princeton, N.J.

RESTORING DERELICT LANDS IN GREAT BRITAIN 274
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. 19 Optimizing Timber Yields in New
Brunswick Forests
Management of commercially important species to increase yields is usually a local matter, independent of conditions and management actions undertaken elsewhere. That is especially true in forestry, in which individual trees are stationary, long-lived, and easily countable. Nevertheless, plot-by-plot management can create long-term supply problems on a larger scale. These larger-scale effects are the most important ones that influence the vitality of economic activity based on exploitation of the resource. This case study illustrates an attempt to link local decisions with regional ones where the valued ecosystem component (wood) is readily identified and the processes producing wood are well known. The major uncertainties are associated with changes in scale, and they require a balancing of scientific and social issues.

OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 275
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Case Study
THOM A. ERDLE, Forest Management Branch, New Brunswick
Department of Natural Resources, Fredericton, Canada
GORDON L. BASKERVILLE. Faculty of Forestry, University of New Brunswick, Fredericton, Canada

INTRODUCTION
Silviculture and wood technology are highly developed, and the factors that influence wood production rates are well known, despite the very long cycle of the crop. Like many other areas, New Brunswick depends heavily on industrial forestry. However, the forest base probably cannot continue indefinitely to supply the quantity and quality of wood required by industry without a large and expensive effort to develop the forest resources.

Recognition of this need by government and industry forestry decision-makers has resulted in a strong commitment to carry out the necessary development, which entails legislative changes in forest administration, changes in the allocation of raw material to industrial wood consumers, and intensified action aimed at making the forest biologically more productive (Bird, 1980).

Increasing the productivity of the forest requires special consideration, because it includes direct intervention in the development of a natural system.

Tools for direct biological intervention in forest development include harvesting, protection, and silviculture; and the use of those tools requires explicit decisions on how much, where, and when to implement them. Their implementation initiates a biological response that alters forest development and production of wood, and the response must be forecast to permit the design of useful intervention strategies. This is all made difficult by disparities of scale, in both time and space, between the stand level at which the actions are taken and the forest level at which the response must be assessed.

Stands, considered here as small homogeneous communities of trees (10-50 hectares), collectively constitute a forest. A forest can be enormously complex, encompassing many thousands of stands and many hundreds of thousands of hectares. Forest management attempts to deal with complexity of this scale by orchestrating the implementation of harvesting, silviculture, and protection at the stand level.

The direct results of local interventions are immediate and visible, as stands are harvested, established, and protected. But the interventions OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 276
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. alter local dynamics that govern the development of stands and that contribute to the unfolding of forest dynamics. Because of the complexity of the forest and the long periods associated with stand growth, forest-level responses triggered by local interventions are not readily apparent and can accumulate into unforeseen, undesirable, and essentially unalterable proportions.

How can the time and scale differences between the stand level at which actions are taken and the forest level at which management is planned and evaluated be bridged to permit the formulation of local intervention tactics that address the forest management problem in the most appropriate, biologically realistic manner? The case described here, involving a 300,000-ha forest management license in New Brunswick, is an attempt to construct a part of that local-global bridge. The results of similar analyses have led to management decisions that affect a large fraction of the forests of New Brunswick.

THE ENVIRONMENTAL PROBLEMS
The issue in this case is impact assessment, i.e., designing local interventions that change the development of a biological system and forecasting the nature and extent of the change (impact) to permit evaluation of the desirability of proposed interventions.

Three entities must be considered with respect to wood supply: the quantity of wood, the quality of available wood, and the timing of availability.

Analyses have revealed that development of New Brunswick's forest under management restricted solely to harvesting and protection, as has been the practice, yields wood whose volume is insufficient to meet industrial demand, whose quality is below minimal standards, and whose availability is discontinuous and erratic (Baskerville, 1982). Studies have shown that incorporation of silviculture, particularly tree planting and tending in the present case, can mitigate these supply problems (Baskerville, 1983). This case study addresses the specific questions: What are the probable forest-level responses to an array of local, stand-level decisions regarding the amount of planting and density at which plantations are established, and what are the immediate and long-term impacts of the decisions on wood supply? The restriction of the present analysis to planting is solely for illustrative purposes. Stand spacing, thinning, and fertilization are other powerful silvicultural tools. The procedure described here could well be extended to those other tactics.

In planting, two major points are at issue. First, at what rate (hectares per year) should plantations be established? Tree planting, an expensive OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 277
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. undertaking initially, forces a continuing commitment to tend planted stands as they develop. Second, at what tree density (trees per hectare) should plantations be established? High volume per unit area and large trees are both desirable in a stand. However, the density dependence of individual tree growth in a stand is such that volume per stand and volume per tree cannot be maximized simultaneously. The former tends to be directly related to stand density, the latter inversely related to it. Thus, any chosen planting density represents a trade-off between volume per unit area and volume per tree, and the impact of this trade-off must be forecast for the proposed strategy and expressed in a form suitable for review by a decision-maker.

Sophisticated routines have been developed that help to optimize stand performance (Brodie and Kao, 1979; Hann et al., 1983). But they concentrate on stand-level responses and ignore the interplay of stand dynamics, governed by the interactions among trees in a stand, and forest dynamics, governed by the collective development of stands in the forest. Failure to place stand development in the context of forest development has separated the elegant silvicultural solutions from the forest management problem and has impeded evaluation and selection of efficient means to address overall wood-supply concerns.

THE APPROACH
The approach presented here is an analytical framework that presents probable outcomes of a wide array of decisions and yet retains the linkages between trees, stands, and the forest. The framework is built around six steps, as follows:

Step 1. Select a specific forest holding and determine the future wood supply and the precise nature of the biological limitations on wood supply produced by restricting management actions to harvesting and protection. This is the biological definition of the economic problem of wood supply.

Step 2. Identify an array of remedial measures that can be applied to the strategic problem of wood supply. These are tactical approaches and, in this case study, are limited to the rate and density at which plantations are established.

Step 3. Forecast the stand-level outcomes of each alternative in Step 2.

Step 4. Link the stand-level responses into the forest mosaic and forecast the new forest dynamics that would result from implementing each of the planting interventions of Step 2.

Step 5. Translate forest-level outcomes into performance indicators relevant to the wood-supply problem (identified in Step 1).

OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 278
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Step 6. Assemble the performance indicators of all the alternatives in Step 2 to create a response surface (impact statement) that displays relationships between stand-level decision variables and forest- level performance indicators.

USES OF KNOWLEDGE AND UNDERSTANDING
Valued Ecosystem Component
Within the six-step framework described above, the valued ecosystem component is the wood supply, as described by the performance indicators at the stand and forest levels. These indicators can be divided into two sets. The first set, at the forest level, includes short-term indicators of maximal sustainable harvest and associated unit cost and long-term indicators of potential harvest expansion and associated unit cost. These are the most important measures, because they are related directly to the wood supply. The second set, at the stand level, includes accumulated salable volume and average tree size at any point in the life of the stand. Although these are not the direct basis of decision-making, they collectively contribute to the forest-level indicators.

Significance of Impacts
The significance attached to changes in forest-level indicators brought about by the exercising of a set of stand-level control options is largely a function of the goals of the decision-making agency. Interpretations of significance can vary considerably between and within public and private agencies. Therefore, interpretations of significance are deliberately omitted here; ''significance" has meaning peculiar to each agency, given its current perspective on the problem. The magnitude of reforestation effects on wood supply must be shown in a form that enables decision-makers to impose their own values and to draw their own conclusions regarding the significance of the impacts of a wide variety of strategies.

Bounding the Problem
The scale differences between forest-management control of stands and silvicultural control of trees necessitate different spatial bounding. At the forest level, spatial bounds can be logically established around the wood-supply base for a mill or group of mills controlled by one management agency. In this case, that supply base comprises 300,000 ha in northwestern New Brunswick. At the stand level, the spatial bounds must be
OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 279
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. expressed with a resolution consistent with silvicultural intervention (plantations), usually about 20 ha. The density dependence of tree growth operates at a very local level, so there is little interaction among non-neighbors.

For a given stand structure, this leads to a perfectly linear relationship between stand production and size of stand (or group of stands of a given size).

Two considerations were involved in imposing temporal bounds at the stand and forest management levels. Because of the long period required for stand development and the societal importance of the forest as a long-term renewable resource, the appropriate period for examining wood flow is necessarily long. Furthermore, the forest age structure has a biological memory:
effects of disturbances, such as harvesting and planting, will surface and persist as the forest develops. A time horizon of 80 years was therefore deemed appropriate. In wood-supply problems, it is not sufficient merely to know the wood yield of a plantation or the volume of wood available from a forest. The timing of wood availability, and therefore the timing of supply problems, is of utmost importance. Detailed time resolution within the long-term horizon is necessary, to address timing in design of strategies. Consequently, the 80-year horizon was divided into 2-year steps for forecasting forest development and for tracking performance indicators.

For purposes of this chapter, it was decided not to consider the impact of spruce budworm on the forest and not to consider the numerous other issues that pervade real forest management. The budworm has played a major role in the development of the New Brunswick forest, as reflected in the present age- class structure of the forest. Yields from forest lands also are influenced by the sizes of budworm populations. Nonetheless, attempting to capture the complexities of the budworm-forest interaction would so dominate the analysis as to obscure the central issue considered here — assessing the cumulative effects of harvesting practices. (For an overview of the budworm problem, see Baskerville, 1976.)
Study Strategy Development
A clear understanding of population dynamics at two levels is required for strategy formulation. At the stand level, dynamics are governed by the density- dependent forces that influence growth of individual trees. At the forest level, dynamics are governed by the structure and development patterns of the various types of stands that make up a forest. Density-dependent stand development and age-dependent forest development must be functionally integrated to provide a path for local intervention to be projected to the forest level.

OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 280
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

FIGURE 1 Stand development over time characterized by merchantable volume (above) and stems/m3 (below) for various initial stand densities.

The results of density-dependent tree competition in plantation development can be captured in two stand variables and their change over time:
salable volume per hectare (m 3/ha) and average tree size (trees/m 3). Typical patterns for stands of three different initial densities are shown in Figure 1 .

Within limits, the density dependence of tree growth shifts the patterns up in higher-density stands and down in lower-density stands. For each stand type or plantation in the forest, these relationships must be quantitative.

For a stand to be considered economically available for harvest, minimal thresholds in salable volume per hectare and average tree size must be satisfied.

These thresholds are the criteria that stands must meet to be OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 281
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. recruited into the operable (or available) volume inventory. The age range over which these thresholds are satisfied defines the timing of a stand's availability for harvest (Figure 2). For a given set of volume and tree-size thresholds, initial stand density can be silviculturally controlled to influence the timing of the stand's availability for harvest (Figure 3).

FIGURE 2 Determination of operable window for stand that defines its time of availability for harvest. Broken horizontal lines represent operability constraints of 100 m3/ha (above) and 9 stems/m 3 (below). Broken vertical lines indicate earliest age at which both constraints are satisfied.

The potential power of stand-level control of availability becomes readily apparent in the context of forest-level dynamics. Because of forces of origin, species mixture, site fertility, and stocking, each of the many stands in a forest has its own pattern of development. A description of a forest shows the stage (age class) of each stand in its development pattern. Forest OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 282
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. dynamics are generated by the collective development of the constituent stands as they progress (age) along their own patterns of development. Stands do not influence one another's growth, but they are linked at other levels. First, there is an analytical relationship between stands, in that the abundance of each developmental stage collectively constitutes an age-class structure for the forest.

Second, interventions, such as harvesting and silviculture, performed in one place can force the withholding of such interventions in some other place.

FIGURE 3 Effect of initial stand density on timing of stand availability for harvest. Operability thresholds, indicated by broken horizontal lines, are 100 m3/ha (above) and 9 stems/m 3 (below). Broken vertical lines represent earliest time of availability for stands at each initial density.

Managing the forest for wood supply is primarily a matter of regulating OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 283
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. the availability of stands. That is achieved through a balancing of the liquidation of mature stands with the recruitment of immature ones across the lower threshold of operability. It requires a harvest schedule that defines the rate and sequence at which the harvest will proceed through the age structure of the forest. The rate of harvest is constrained by the timing of the availability of replacement stands as mature, operable ones are harvested. Availability of wood over time is a function of the initial forest age-class structure, particularly the relative abundance of stands in each age class, and the rate of development of these stands to operability. Consequently, any actions that hasten or retard availability of young stands for harvest have obvious implications for the rate at which mature ones can be harvested. That is the path by which density control at the stand level influences wood supply at the forest level.

The forest age-class structure both determines availability and constrains the manner in which implemented stand-level actions translate into forest-level responses. As a result of the natural and man-made forces behind their development histories, New Brunswick forests have irregular and different age structures. A fixed set of stand actions, carried out on similar sites, will yield identical stand-level results that are wholly independent of the forest in which they are applied. However, the same set of stand actions might generate dramatically different responses, depending on the age structure of the forest.

This case study deals with one such initial structure.

Hypotheses
The use of these conceptual tools to manage forest-level responses requires testing and evaluation of three hypotheses.
A quantitatively specific hypothesis must be made about how stand density affects tree growth and, ultimately, stand growth. It must be comprehensive enough to address such a question as: What will be the pattern of volume per hectare and average tree size over the life of a black spruce (Picea mariana (Mill.) B.S.P.) plantation established in northern New Brunswick at an initial density of 500, 1,000, 2,500, or 4,000 stems/ ha? The variables expressed in the forecast must be, at least, average tree size and volume per hectare, because of their important role in defining stand availability for harvest. A hypothesis is necessary with respect to change over time in the minimal thresholds of tree size and stand volume required for stand op OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 284
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. erability (utilization standards). Rapid changes in use indicate that operability constraints change as a result of better technology in harvesting and product development. Instead of attempts to forecast how utilization standards would change, a range of reasonable future values was established, and the impact of these values on the forest- level indicators were analyzed. All the outcomes were incorporated into the impact statement in the form of a response surface, so that decision-makers could locate their own expectations with respect to changing operability limits and assess their significance. A third hypothesis must relate development of the whole forest quantitatively to that of the component stands, particularly those for which the plantation effort is contemplated. This hypothesis is also the basis for a forecast. The following are relevant and representative groups of questions: (1) How will the forest structure and resulting total growing stock change over time if no harvest is performed? If harvesting is carried out in the oldest stands first at a rate of, say, 400,000 m
3/year? 500,000 m 3/year? (2) How will forest structure and resulting growing stock change, if the same harvest schedules are attempted, but 4,000 stems are planted per hectare at a rate of 1,500 ha/ year? 3,000 stems/ha at 3,000 ha/year? This second type of question is essential in designing a strategy of plantation use, and it highlights the necessary linkage between the local decision variables (planting rate and density) and the forest performance indicators (wood supply from the forest).

Cumulative Effects
To evaluate the cumulative effects of stand-level tactics on forest-level performance, the hypotheses with respect to stand development, operability limits, and forest development were systematically knitted together. Plantation performance was forecast for each density alternative between 500 and 4,000 stems/ha. Various sets of operability constraints were imposed on each of these, to establish the pattern of stand availability for harvest. Forest dynamics were then forecast iteratively as plantation tactics were systematically varied over all combinations of 0-4,000 hectares planted per year and 500-4,000 trees planted per hectare. For this purpose, the planting rate and density variables were changed at intervals of 500 ha/ year and 500 trees/ha, respectively, to generate 72 unique planting strategies, each with a pattern of future wood availability.

Performance indicators were tracked, through a model of forest development, to describe the cumulative wood-supply effects resulting from each strategy.

OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 285
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

SOURCES OF KNOWLEDGE AND UNDERSTANDING
Generally Accepted Ecological Facts
Many empirical studies have described the effects of density on tree and stand growth for several species and locations in North America (e.g., Baskerville, 1965; Ker, 1981; Lundgren, 1981; Stiell and Berry, 1973). The findings, qualitatively displayed in Figure 1, are generally accepted by foresters, although quantification has proved more difficult for particular species and site conditions.

Major planting programs were initiated only in the 1960s, so few comprehensive data sources are available to help in quantifying the developmental relationships in plantations in New Brunswick. For efficient use of the limited data, a plantation growth model, responsive to density control, was constructed for the species and sites relevant to this case study.

The forest structure and natural stand growth are more adequately described by data sets available in the New Brunswick forest inventory and the records of the industrial land owner in this case. The forest-level characterization is presented in simplified form in Figure 4 , which shows the present forest to comprise three types of stands: softwood stands dominated by fir (Abies balsamea (L.) Mill.), softwood stands dominated FIGURE 4 Age-class structures and yields for three stand types — fir (top), spruce (middle), and mixed wood (bottom) — that constitute forest used in case study.

OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 286
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. by spruce ( Picea sp.), and mixed-wood stands with fir and spruce components.

The development pattern of each type, in terms of wood yield, reflects the biological characteristics of the species that make it up.

The present forest structure is a product of history and is obviously nonuniform, with a preponderance of mature stands, a moderate presence of newly regenerating stands, and a dearth of stands at stages of development between the extremes. This age structure bears the ecological imprint of the two main historical forces behind the forest'sdevelopment — harvesting by man and harvesting by spruce budworm ( Choristoneura fumiferana (Clem.)). Intensive harvesting by man over the last 30 years has created many stands in the regenerating category. The forest suffered severe "harvesting" as a result of defoliation by the spruce budworm between 1910 and 1920; large-scale destruction of the mature forest at that time resulted in a large number of regenerating stands that appear today as the sizable area in the 60- to 80-year age classes.

Most of the current stands are at or very near their peak volume. The futures of these stands, particularly fir, will be characterized by substantial decrease in volume, owing to natural decadence and breakup. Thus, even in the absence of harvesting, there will be a decrease in growing stock for the whole forest as the individual stands age. The paucity of stands in the 30- to 60-year age classes means that little in the way of replacement stands will become available for harvest as the mature stands are eliminated through harvest or by natural decline. The wood-availability problem hinges on the latter point. To maintain a constant wood flow, harvest of mature stands must be paced so that they last until sufficient young stands are available for harvest.

Together, these factors reveal that measures that hasten the entry of regenerating stands into the available inventory might have a powerful effect in overcoming the problem of continuity of available stands posed by the broken forest age structure. If cut-over sites are immediately restocked with desirable species, spatially arranged to use as much of each site as possible and to control intraspecies competition, individual stem growth will be more rapid than in untreated stands. The more rapid development of plantations increases the availability of replacement stands, which in turn permits a greater rate of liquidation of mature stands and therefore an increase in the annual wood supply. That is the principle that underlies the use of planting tactics to address the wood-supply problem for this particular forest.

The model used to forecast plantation development is based on two general principles: in the absence of any competition, individual tree height and diameter growth are bounded by inherent upper limits that are functions of species and site conditions, and realized growth (less than these OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 287
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. limits) is regulated by the amount of growing space afforded the individual tree and its competitive status within the canopy of the stand.

Specific Models
Two models were used to make the required forecasts of stand and forest dynamics. At the stand level, a simulation model was constructed by giving quantitative form to the principles stated above and using available field data to define the maximal potential growth rates for black spruce in the region.

The model views the stand as a population of trees described by a diameter and height distribution at any time. Growing space and competitive status for each size class are determined from this distribution and applied to potential dimension increments to obtain forecast increments for the current competitive situation. The growth-estimation process is iterative, each iteration taking into account the accumulating changes in stand structure in the computation of growing space and competitive status, the two growth-regulation factors. The model is therefore dynamic enough to grow stands in a structurally realistic manner and to allow incorporation of planting-density alternatives.

Lack of long-term plantation growth data on New Brunswick made creation of the model and assessment of its predictive behavior particularly difficult. Comparisons were made against relevant data sets; in all cases, the model forecasts of volume per hectare and tree growth were qualitatively in accord with the empirical studies. Model forecasts for four initial stand densities are shown in Figure 5. The model provides a specific focus for carrying out a plantation-growth monitoring program, and the explicit assumptions behind the model are in a form amenable to specific field-testing and revision.

Forest development was forecast for various harvest and planting interventions with slight modification of an existing model. The original model was built by Erik Wang (Fraser Inc.) and parallels the concept behind the Wood Supply and Forest Productivity (WOSFOP) model of Hall (1978). This type of simulation model has seen extensive use in New Brunswick and other provinces (Cuff and Baskerville, 1982) and was used here because it was appropriate for the forest-level questions asked and because forest managers in the province are generally comfortable with it.

The forest-development model accepts as input an array of stand growth types that constitute the initial description of the selected forest, by giving stand areas, age-class structures, and yield functions (as in Figure 4 ); operability limits (as in Figure 3); and a rate of harvest specifying the OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 288
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. amount to be removed annually and the priority rules for allocating that harvest to the various stand types and to stages of development within a stand type. At 2-year intervals, the model advances the age-class structures for each stand type along their yield curves, performs the harvest and planting activities, and calculates forest-level indicators, such as total available growing stock, volume loss in mortality, and the evolving forest age structure.

FIGURE 5 Model forecasts of merchantable volume (above) and tree size development (below) for black spruce plantations at five initial densities.

Broken vertical lines represent minimal time required to satisfy current operability thresholds of 115 m
3/ha and 7 stems/m3 simultaneously.

The two models are linked, in that the stand model is an input source OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 289
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. for the forest model. Through this linkage, stand-level tactics are carried through to the forest-level performance indicators.

CONTRIBUTION OF RESULTS TO ECOLOGICAL

KNOWLEDGE
The study shows that limited preliminary data, combined with basic ecological principles, can be systematically used to display the impacts of a range of management strategies. Several hundred simulations were performed by linking the plantation-development model with the forest-development model to evaluate the forest-level outcomes of various combinations of planting rate and density. The results are presented as a set of response surfaces, or nomograms, in Figure 6 . This format is particularly useful when two decision variables are to be assessed via their control over the valued ecosystem components through several indicator variables (Peterman, 1975). Decision- makers can easily review outcomes of a wide range of strategies before setting a decision process (or optimization) in motion. In this case, the decision variables are local (stand level) implementation options of plantation density and planting rate, shown on the Y and X axes, respectively. The surface evaluation, or Z variable, represents a forest-level performance indicator associated with each stand-level decision combination. This surface is a type of prediction containing not only absolute outcomes of a host of alternative combinations of stand tactics, but also the sensitivity of the forest indicators to those tactics.

Figure 6 contains a number of noteworthy relationships between local stand actions and overall forest outcomes that have important implications for selection of a strategy. For example, Figure 6A shows that increasing the area planted annually increases the sustainable harvest immediately available from the forest, regardless of planting density. The increase in forest response surface is nonlinear, however, as evidenced by the widening intercontour gaps associated with increasing the planting rate. At the onset of horizontal contours (e.g., at 2,000 ha/year and 2,500 stems/ ha), the positive immediate impact of planting on wood supply disappears altogether. Obviously, at a given density, each additional hectare planted is growing identically with the rest, but the effect of the marginal hectare on total forest performance changes dramatically with the degree of activity. Each additional hectare is progressively less ''productive" in promoting immediate increased harvest. Thus, increases in planting are not accompanied by proportional harvest increases, and, beyond particular amounts of planting, there is no immediate harvest increase at all. At this point, harvesting has so drastically altered the forest structure that the resulting low abundance of mature stands constrains further immediate OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 290
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. harvest increases. Essentially, the biological limitation on wood supply switches from scarcity of young stands at low planting rates to scarcity of mature stands at high planting rates. The time to stand operability associated with each plantation density regulates the planting rate at which this switch occurs. This has two important implications. First, the powerful influence of forest structure on the effectiveness of planting is not evident
FIGURE 6 Impact of planting rate ( X axis) and plantation density ( Y axis) on four indicators of wood supply. Indicators A and B are related to immediate impacts; indicators C and D are related to impacts 60 years later.

OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 291
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. if planting is evaluated solely on the basis of local, stand-level performance, as is common in investment approaches. Stand-level assessment would indicate that, if some planting is good, then more is better, and it is better in direct proportion to the extent of effort. Clearly, that is not the case: the linearity breaks down when plantations are viewed in the context of the whole forest.

Second, the forest-structure constraint on planting effectiveness necessitates plantation tactic design specific to the forest in question. The surface in Figure 6A would be different if the same stand response were applied to a forest of different initial age structure from the one used here. Thus, management strategies embodying design of stand tactics are not freely transportable between forests.

The maximal sustainable harvest for this forest is 480,000 m 3/year
(Figure 6A) and can be attained by planting 1,500 stems/ha at a rate of 3,000 ha/ year. This combination, of course, pertains to a specific set of operability constraints and an 80-year horizon. The maximum resides in a very sensitive area in Figure 6A. The tightly packed contours in the lower third of the figure form a "cliff" that represents a high-risk zone for the decision-maker. If strategies at the edge of the cliff are pursued to maximize the harvest, then substandard plantation performance, poor competition control, ineffective protection, or bad model forecasts could have disastrous consequences. Such occurrences would mean that the sustainable harvest was over the brink and either a severely reduced sustainable harvest or an unexpected disruption in wood flow would result from implementing a harvest rate set in accordance with plantation yields that never materialize. Strategies in this high-sensitivity zone might be chosen by the risk-taking decision-maker, whereas one with a more conservative approach to risk might opt for a higher plantation density (2,000-3,000 stems/ha) and accept a slightly lower sustainable harvest to ensure a less sensitive response.

At the stand level, the highest-yield-per-hectare option is planting at a density of 4,000/ha ( Figure 5 ). This does indeed keep stands in the most productive state, which is seen as a desirable goal by some foresters, but, as shown in Figure 6 A, devastates the productivity of the forest as a whole, because it severely delays availability of plantations by reducing growth rates of individual trees. Again, this exposes the important difference between local stand response and overall forest response to selected sets of tactics.

Figure 6 C contains information relevant to possible future increases in harvest, showing the growing stock available 60 years from now. Appreciable gains in growing stock occur at planting rates higher than those needed to achieve the maximal immediate harvest. Furthermore, with respect to future increases, the greatest gains are realized at higher densities OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 292
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. than those which generate the maximal immediate harvest. Thus, high present harvests and high future harvests cannot be ensured simultaneously through the implementation of one plantation tactic. The decision-maker faces a trade-off between immediate gains and future gains. It can be resolved by striking a compromise between the two or by using a strategy with two simultaneous tactics: a low-density one aimed at immediate harvest gains and a high-density one aimed at future harvest gains. Regardless of the choice, failure to link stand and forest performance would not even reveal the problem to the decision- maker, let alone suggest potential solutions.

To illustrate the impact of operability limits on the decision-maker, several stand- and forest-level simulations were performed in which the minimal acceptable operability standards were systematically altered. The annual planting rates were fixed at the number of hectares required to maximize the current harvest for each planting density (e.g., 1,500 ha/ year at 4,000 stems/ha and 3,000 ha/year at 1,500 stems/ha). The results are presented in nomogram form in Figure 7 . The X , Y , and Z axes show minimal average tree size, plantation density, and maximal sustainable harvest, respectively.

Each of the four surfaces in Figure 7 represents a different minimal- volume-per-hectare requirement. It is clear that the desirability of alternative densities, with respect to wood supply, varies considerably with the tree-size constraints that will be in effect. In Figure 7B, for example, under the stringent requirement of 6 stems/m
3, plantation densities of 1,000-1,500 stems/ha provide the maximal harvest. However, as the tree-size constraint is relaxed, both the maximal sustainable harvest and the plantation density at which the maximal harvest is realized increase. This is evidenced by the ridge that runs through the surface in the figure. That the surface elevation increases with relaxation of tree- size constraint is a reflection of the decreased time to stand operability and the consequent increased availability of stands when harvest of smaller trees is acceptable. That the ridge has a positive slope in the X-Y plane is indicative of the earlier achievement of these lower operability thresholds by higher-density (and consequently higher-volume) plantations when the tree-size constraint is reduced.

Examination of all four surfaces reveals that the increase in harvest associated with relaxed tree-size constraints holds for the minimal-volume-per- hectare constraint as well. The maximal possible harvest increases as the minimal volume per hectare is decreased from 205 m
3/ha (Figure 7D) to 70 m 3/
ha ( Figure 7 A). Similarly, the planting density that yields the greatest harvest increases as the volume-per-hectare constraint increases. That is consistent with Figure 5, which shows the advantages of low and
OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 293
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. high densities to be rapid tree-size development and rapid volume-per-hectare development, respectively. Of importance to decision-making in Figure 7 is the powerful influence that utilization constraints can have over forest-level outcomes. The actual biological response underlying each surface in Figure 7 is identical. What generates the marked differences in forest outcome is a logistical harvesting constraint that the decision-maker imposes on the biological system, which regulates the availability of stands FIGURE 7 Impact on wood supply of plantation density ( Y axis) and minimal acceptable tree size (X axis) at four constraints on minimal volume per hectare.

OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 294
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. for harvest — this is no mere detail, in as much as utilization constraints are largely under the decision-maker's control. Figure 7 reveals likely outcomes that the decision-maker might achieve by exercising that control.

APPLICATION OF THE TOOL
The modeling process captured in Figures 6 and 7 constitutes a useful tool in policy design. Nomograms like those shown are powerful aids in structuring policy questions so that detailed modeling and analysis can be directed incisively. The diagrams are not used to choose a "best" planting policy. Rather, they are displayed to senior decision-makers, first to discover how they weight various indicators and second to discover where the decision-makers would like to be within the possible policy domain. Detailed analyses can then proceed with the indicators of choice within limits in terms of the policy variables. Thus, the nomograms are not used to answer operational questions, but rather to structure management questions. In this context, they make the scientific advisor more efficient both in the use of his own time and in establishing an understanding of the problems between advisor and decision-maker.

CONCLUSION
The case study presented here shows how a wide range of stand-level actions would influence forest-level outcomes. The purpose is to make the decision-maker aware of the importance of linking the two levels of consideration. No recommendations of optimal planting densities or planting rates are presented, because such decisions are strongly influenced by specific industrial strategies and objectives and by the degree of risk aversion of decision-makers. Furthermore, the decision-making picture is incomplete.

Analysis becomes more complex as wood value, harvesting cost, and additional silvicultural tactics (such as spacing and thinning) are considered. Once the range of impacts is understood, more sophisticated analytical tools, like mathematical programing, might be effectively brought to bear on strategy design.

Different planting alternatives, applied at the stand level, generate different forest-level outcomes, because of the interaction of stand and forest dynamics.

These differences highlight risks, sensitivities, and trade-offs that, although of prime importance in decision-making, would not be readily evident from stand- level or forest-level analyses alone. As a package, the analytical framework and the results it provides form a rich body of information relevant to the forest management problem of evaluating the impact of interventions on wood supply.

OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 295
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

The techniques described in this study have been applied in the design of New Brunswick's silvicultural program. Effective control over planting is possible, because about 50% of forest land in the province is Crown land and the large timber companies need access to Crown land to maintain economical operations. Of the remaining land, half is owned by large companies and half has varied ownership. Thus, about three-fourths of the forest land in the province has been readily incorporated into a harvesting program suggested by the analysis as necessary to even the flow of timber and counteract the unfavorable age distribution of the stands.

There is not yet a firm link between stand dynamics and budworm population dynamics. Therefore, it is not possible to predict the effects of the harvesting and planting program now being implemented on population dynamics of the insect. Whether the program will have to be modified to accommodate problems generated by the budworm remains to be seen.

REFERENCES
Baskerville, G. L. 1965. Dry matter production in immature balsam fir stands. For. Sci. Monogr. 9:1-42.

Baskerville, G. L. 1976. Report of the Task Force for Evaluation of Spruce Budworm Control Alternatives. New Brunswick Cabinet Committee on Economic Development, Fredericton, N.B.

Baskerville, G. L. 1982. The Spruce/Fir Wood Supply in New Brunswick. New Brunswick Department of Natural Resources, Fredericton, N.B.

Baskerville, G. L. 1983. Good Forest Management — A Commitment to Action. New Brunswick Department of Natural Resources, Fredericton, N.B.

Bird, J. W. 1980. Forest management — A provincial perspective. Pp. 33-37 in The Forest Imperative. Proc. Can. For. Congr., Toronto, Ont., September 22-23, 1979. Canadian Pulp and Paper Association, Montreal.

Brodie, J. D., and C. Kao. 1979. Optimizing thinning in Douglas-fir with three descriptor dynamic programming to account for accelerated diameter growth. For. Soc. 25:665-674.

Cuff, W., and G. L. Baskerville. 1982. Ecological Modeling and Management of Spruce Budworm Infested Fir-Spruce Forest of New Brunswick, Canada. Paper presented at 3rd Int. Conf. on State-of-the Art in Ecological Modeling, Colo. State Univ., May 24-28, 1982.

Hall, T. H. 1978. Toward a Framework for Forest Management Decision-Making in New Brunswick. Report TRI-78. New Brunswick Department of Natural Resources, Fredericton, N.B.

Hann, D. W., J. D. Brodie, and K. H. Riitters. 1983. Optimum stand prescriptions for ponderosa pine. J. For. 81:595-598.

Ker, M. F. 1981. Early Response of Balsam Fir to Spacing in Northwestern New Brunswick.

Maritime Forest Research Center Information Report M-X-129. Canadian Forest Service, Fredericton, N.B.

Lundgren, A. L. 1981. The Effects of Initial Number of Trees Per Acre and Thinning OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 296
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Densities on Timber Yields from Red Pine Plantations in the Lake States. Forest Service Research Paper NS-193. U.S. Department of Agriculture, Washington, D.C.

Peterman, R. M. 1975. New techniques for policy evaluation in ecological systems: Methodology for a case study of Pacific salmon fisheries. J. Fish. Res. Bd. Can. 32:2179-2188.

Stiell, W., and A. B. Berry. 1973. Development of Unthinned White Spruce Plantations to Age 50 at Petawawa Forest Experiment Station. Publication 1317. Canadian Forest Service, Ottawa, Ont.

Committee Comment
The work reported in this case study provides a concrete example of how to assess the long-term, regional consequences of immediate, local forest- management actions in a manner useful for the design of sustainable strategies for resource development. The story is interesting as a particularly simple and clear case of effective assessment of cumulative effects, and it has proved useful in actual practice — by the Department of Natural Resources (DNR) of New Brunswick. Indeed, Baskerville initiated the analysis described here to deal with the practical difficulties he encountered as assistant deputy minister in DNR; Erdle is now responsible for applying the results of the analysis in DNR's Forest Management Branch.

The problem addressed here is widespread in forest management. A forest is a large-scale mosaic of individual stands of trees. Different stands can generally be characterized by age distributions, species mixtures, and habitat.

The forest therefore is also a mosaic of stand-development trajectories, different stands reaching maturity at different times and rates. Management actions can alter those trajectories.

Although wood is considered the only valued ecosystem component in this model, additional components, such as control of spruce budworm and maintenance of an aesthetically pleasing mosaic of forest patches, have been considered in other analyses of the system (Baskerville, 1976; Clark et al., 1979). The focus developed in this study was the result of a consensus that emerged among government, industry, and academic participants in the forest- management debate that, if the wood-supply problem were solved, most other concerns would be met automatically. Interestingly, the supply of wood is from the outset defined in cumulative terms over the regional scale and long periods relevant to the economics of the provincial forest industry. What matters ultimately is not the production from an individual stand of trees, but rather how the production from all the stands of the forest, taken together, can best be managed to meet society's needs.

OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 297
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Erdle and Baskerville studied the cumulative effects of alternative management actions by integrating local models of stand growth, regional models of forest age-structure dynamics, and a detailed inventory of the forest's existing age structure. They used available models of the relations among site quality, tree density, and growth rates of individual trees as the major sources of ecological knowledge. In essence, a commonly used wood-supply and forest- productivity model was taken "off the shelf" and provided with data relevant to the New Brunswick situation. This procedure made their task simpler and made their results more acceptable to managers already familiar with the models.

The forest model used by Erdle and Baskerville is essentially a bookkeeping one that tracks the aging of forest stands as they respond to additions to wood (through planting and natural regeneration) and removals of wood (through harvest, insect damage, and natural death). The underlying ecological theories are simple. Growth rates are assumed to be density- dependent, and the yield curve has a maximum. These results have been empirically determined, but they can also be derived theoretically from basic principles of plant competition. The simplicity of these demographic accounting models is also their main strength. They are robust, and, if solid biological data are available, they can be readily used for a wide variety of situations (see also Chapter 12 ). More complex, general, and rigorously tested stand models (Shugart, 1984) could have been used in the analysis and probably would have afforded improved credibility in academic circles. But there is no reason to believe that these more elegant models would have appreciably improved the results obtained by Erdle and Baskerville.

An additional aspect of ecological knowledge central to the success of the Erdle and Baskerville analysis was the existence of an accurate description of the present age distribution of provincial forests. Such baseline data on the heterogeneity of tree stands with respect to age distribution, species mixture, and growth potential are necessary to coordinate local, short-term management actions in a way that achieves desired regional cumulative consequences.

Unfortunately, such data are extremely rare, especially in the case of age distribution, in inventories of forests and other renewable resources.

New Brunswick has a forest inventory data base that is one of the most accurate and useful for forecasting purposes in North America. This is in part because of the data-base shortcomings uncovered in Baskerville's earlier modelings and analyses and his later tenure in DNR. Generally speaking, however, useful background data are difficult and unglamorous to obtain, monitor, and update, although a characteristic feature of the OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 298
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. few success stories in resource management is the existence and intelligent use of such data bases (see, for example, Chapter 12).

One final aspect of the Erdle and Baskerville analysis, the method of presenting results, is relevant not only to the assessment of cumulative effects, but also to a wide variety of efforts to provide usable ecological knowledge.

They use nomograms to illustrate the trade-offs in valued ecosystem components that are likely to result from alternative management actions or other development interventions. In this forest-management case, as in many other environmental problems, interpretations of what is important vary considerably within and among government agencies, private interest groups, and the ecological profession. Moreover, as Erdle and Baskerville point out, these perceptions of significance change with time. Rather than adopting a single definition that would render their analysis usable from only a single perspective, Erdle and Baskerville have devised a framework in which users can specify (and indeed often discover) their own definitions of significance and explore the implications of their definitions for a wide range of possible decisions.

An added benefit of the nomogram approach, used to good effect in this case study, is that the spacing of the nomograms contours of valued ecosystem components indicates the sensitivity of a predicted outcome to errors or incompleteness in implementing the decisions. Peterman (1981) has shown how minor modifications of the nomogram technique can be used to show the significance of uncertainties in the ecological models for the expected effects on valued ecosystem components. An analysis of the uncertainties inherent in the regional demographic and local competition models underlying this case study would have increased its usefulness even more.

Nomograms of this sort have been applied to environmental problem solving in cases of renewable-resource management (e.g., Holling, 1978; Peterman, 1975, 1977; Regier, 1976), river-basin planning (e.g., Rabinovich, 1978), and regional development (e.g., Miller, 1982). They have a long history of application in business and industrial management. Nomograms are not a cure-all, especially given their restriction to two or at most three simultaneous trade-offs. But they are useful in appropriate circumstances, and they deserve wider application.

Not all the approaches and techniques for cumulative-effects assessment described by Erdle and Baskerville are relevant to the more complicated situations of, say, river-basin planning, or even to the more closely analogous situations of regional fisheries management. As difficult as forest age structure is to deal with, it is nonetheless measurable, and it provides a firm handle on the assessment of future forest development— a handle
OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 299
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. that most cumulative-effects studies will not have. In addition, the mere existence of a forest-level ''planning authority" (the province's Department of Natural Resources) makes the case reviewed here much easier than the more typical one in which no agency is either solely responsible for or has sole power over the long-term, regional impacts of a class of related development decisions. More generally, it is worth emphasizing with the authors that a great proportion of their analysis involved the careful "tuning" to local conditions of a few relatively simple ecological concepts and models. No general prescriptions for cumulative-effects assessment or management emerged from this study, nor are they likely to emerge from others.

References
Baskerville, G. L. 1976. Report of the Task Force for Evaluation of Spruce Budworm Control Alternatives. New Brunswick Cabinet Committee on Economic Development, Fredericton, N.B.

Clark, W. C., D. D. Jones, and C. S. Holling. 1979. Lessons for ecological policy design: A case study of ecosystem management. Ecol. Model. 7:1-53.

Holling, C. S., ed. 1978. Adaptive Environmental Assessment and Management. John Wiley & Sons, Chichester, Eng.

Miller, P. C. 1982. Simulation of socio-ecological impacts. Environ. Manage. 6:123-144.

Peterman, R. M. 1975. New techniques for policy evaluation in ecological systems: Methodology for a case study of Pacific salmon fisheries. J. Fish. Res. Bd. Can. 32:2179-2188.

Peterman, R. M. 1977. Graphical evaluation of environmental management options: Examples from a forest-insect pest system. Ecol. Model. 3:133-148.

Peterman, R. M. 1981. Form of random variation in salmon smolt-to-adult relations and its influence on production estimates. Can. J. Fish. Aquat. Sci. 38:1113-1119.

Rabinovich, J. E. 1978. An analysis of regional development in Venezuela. Pp. 243-278 in C. S.

Holling, ed. Adaptive Environmental Assessment and Management. John Wiley & Sons, Chichester, Eng.

Regier, H. A. 1976. Science for scattered fisheries of the Canadian interior. J. Fish. Res. Bd. Can. 33:1213-1232.

Shugart, H. H. 1984. A Theory of Forest Dynamics. Springer, New York.

OPTIMIZING TIMBER YIELDS IN NEW BRUNSWICK FORESTS 300
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. 20 Control of Eutrophication in Lake
Washington
Most large cities in the world are situated on coastlines or the shores of rivers or lakes. Freshwaters and estuaries are the initial or eventual recipients of much of the waste products of technological societies. Consequently, water pollution is one of the first environmental problems to arise, and it continues to be pervasive even when discharges of wastes are reduced and waste materials are treated in more sophisticated ways. Because limnology is one of the more advanced fields of ecology, the key factors influencing the responses of lakes and rivers are rather well known, and the reasons for those response patterns are often understood. The Lake Washington case study is an example of creative interaction between the scientific community and the political arena in the development and execution of a plan that resulted in striking and rapid improvement of the quality of the waters of this lake, which was being increasingly influenced by growth of the metropolitan Seattle area.

CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 301
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Case Study
JOHN T. LEHMAN, Division of Biological Sciences, University of Michigan, Ann Arbor
Lake Washington at Seattle (47 °37 ' N, 122 °14 ' W) is a moderately deep (65 m), warm, monomictic basin in the drainage of the Cedar River and Sammamish River. The lake discharges into Puget Sound via a system of locks and canals built in 1916. Situated in an expanding metropolitan area, Lake Washington has for years experienced varied and intense demands for transportation, recreation, and waste disposal. The condition of the lake has changed greatly over the years in response to changes in nutrient income brought about by the sewerage arrangements. That its water quality today is as good as or better than at any other time in its history is due to a unique blend of scientific judgment and public action. During the expansion of suburban Seattle in the years after World War II, the lake deteriorated in proportion to the pressures applied by a growing populace. Reversing that trend by design, the citizens of the region responded voluntarily to the environmental problem. The solution was costly, but public decision was guided by a firm statement of the problem and a plain alternative. Scientific knowledge helped to define possible future conditions of Lake Washington, and voting citizens selected the course.

The story of deterioration and recovery of water quality in Lake Washington on the one hand reflects changes in demographics and politics of a city and its suburbs and on the other hand shows a development and application of scientific thought on a problem that required special qualities of scientific leadership and communication for public education. Action and expenditure of public funds were linked to scientific arguments and to quantitative predictions about conditions of the lake. From a scientific perspective, the actions constituted an experiment and an opportunity to refine hypotheses about how lakes function. From a civil perspective, the case exemplifies transition from parochial, local concerns to a regional outlook in environmental matters.

Seattle began discharging raw sewage into Lake Washington at the start of the twentieth century. The large (86.5 km
2) and deep basin became the
repository of street and septic discharges as the city expanded eastward from Puget Sound. In 1926, however, Seattle created a bond issue for a series of intercepting and trunk sewers to divert sewage from Lake Washington to a treatment plant on the Duwamish River that discharged directly into Puget Sound. By 1941, the last sewer outfall into Lake Washington CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 302
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. from Seattle had been removed. Thereafter, water quality in the lake reflected development not of Seattle, but of its suburbs.

Between 1941 and 1953, 10 sewage treatment plants began operating at points around the lake, with a combined daily effluent of 80 million liters.

Alternative discharge options were not as readily available to the small municipalities as they had been to Seattle. By 1953, James Ellis, a lawyer whose clients included some of the sewer districts around Lake Washington, sought diversion of sewage from Bellevue, but could not get the cooperation of neighboring districts for the necessary routing. He therefore spoke to the Seattle- King County Municipal League, proposing a system of metropolitan organization that could oversee such regional issues.

While these first steps were being discussed in the political arena, scientific investigations of Lake Washington attracted public notice with release of Technical Bulletin 18 of the Washington Pollution Control Commission, An Investigation of Pollution Effects in Lake Washington (1952-1953) (Peterson, 1955). This was the first substantial report of nutrient enrichment of the lake. It cited the work and data of Anderson (1954) and Comita (1953), who had conducted their doctoral studies under the guidance of W. T. Edmondson of the University of Washington. The Seattle Times trumpeted the report with its July 11, 1955, article, "Lake's Play Use Periled by Pollution." The Times returned to the issue a month later, when it reported the complaints of lakeshore residents and spectators at the Gold Cup yacht races (August 11, "Algae Increase Noted in Lake Washington"). Sanitation authorities doubted that the problems stemmed from the increased entry of effluent into the lake. They blamed sunshine, weather, and water conditions for the changes. It turned out, however, that the conditions of the lake that day triggered a new dimension of scientific curiosity in Edmondson's laboratory. Anderson had returned from the lake with a water sample containing an alga never encountered during his doctoral investigations of the lake: the blue-green alga (cyanobacterium) Oscillatoria rubescens.

To Edmondson, the appearance of Oscillatoria signaled that the lake was deteriorating in classical fashion. The large, deep lakes of Western Europe, particularly Lake Zurich, had been similarly enriched with high nutrient effluents decades earlier, and water quality had declined. A series of lakes near Madison, Wisconsin, had received treated discharges from that municipality and had deteriorated. Accounts of these cases were in the scientific literature (Hasler, 1947), and Edmondson was struck by the fact that the name Oscillatoria appeared in each account in connection with the earliest stages of decline. For Lake Washington, previous data were available for comparison from 1933 (Scheffer and Robinson, 1939)
CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 303
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. and from 1949-1950 (Comita and Anderson, 1959). The doctoral studies of Anderson and Comita and the earlier investigation from 1933 provided a baseline by which to judge changes (Edmondson et al., 1956). The main change in the watershed had been the increased load of nutrients from secondary- treatment waste discharge. Edmondson shared his observations in the October 13, 1955, University of Washington Daily ("Edmondson Announces Pollution May Ruin Lake"), recounting the appearance of Oscillatoria and its likely meaning. He defined his own interest as observing and analyzing the transitional nature of the lake and adding to the research done in Germany and Switzerland.

By 1956, the stage was set for developments that would bring civic leaders and scientists together. The scientists took the first step. Edmondson and two University of Washington engineering faculty members, R. O. Sylvester and R.

H. Bogan, published a popular-science article in the university's journal The Trend in Engineering, "A New Critical Phase of the Lake Washington Pollution Problem" (Sylvester et al., 1956). The article told the history of sewage treatment for the area, described the problems posed by nutrient enrichment, and proposed three procedures for solution: comprehensive regional administration and planning, complete elimination of sewage discharge into Lake Washington, and research on the relationships among temperature, nutrients, and algal growth. It provided a concise layman's explanation of nutrient enrichment and its effects, including the appearance of O . rubescens, and it focused particularly on enrichment with the mineral nutrient phosphorus and the difficulty of removing it from sewage. The Seattle Times publicized the article with the headline (April 18, 1956), "Lesson of Switzerland Lakes Brought Home to Seattle Area." By October, Edmondson and a new postdoctoral associate, J. Shapiro, had received funds from the National Institutes of Health to study water chemistry and photosynthesis by algae in Lake Washington. In December, Edmondson wrote a letter to James Ellis that marked his first involvement in the public action. James Ellis had been appointed chairman of Seattle's newly established Metropolitan Problems Advisory Committee by Mayor Gordon Clinton, and Edmondson wanted to ensure that Ellis and the committee understood that even well-treated sewage contained enough nutrients to stimulate the growth of plants in the lake. Lake Washington was already showing signs of the same series of changes toward deterioration as had been observed elsewhere.

After his initial letter had produced a cordial and positive response from Ellis, Edmondson sent him, on February 13, 1957, a nine-page summary of the effect of drainage and effluent entry into Lake Washington. The letter was phrased as a question-and-answer document and included references to the professional literature. Edmondson listed answers to 15 CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 304
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. questions that he thought Ellis might be asked in connection with his work on the advisory committee — such questions as: How has Lake Washington changed? What will happen if fertilization continues? Why not poison the algae? Edmondson included a rudimentary nutrient budget for the lake constructed from available data, and he developed his case that the mass of algae present varied in strict proportion to the amounts of fertilizing nutrients added to the water. The letter included a mass of limnological cause-and-effect reasoning phrased in jargonless, objective tones. Ellis responded enthusiastically within the week, requesting copies of the letter for distribution to interested groups.

The initiative passed back to the political arena for the remainder of 1957.

The immediate obstacle was the absence of provisions whereby municipalities could combine some of their government functions in comprehensive regional matters. Moreover, the notion faced opposition from some, on the grounds that it smacked of "big government." The next step required action at the state level.

That aim was met when the Washington state legislature passed a bill permitting the establishment of a metropolitan government ("Metro") with specified functions (Ch. 217, Laws of 1957). The floor manager in the House had been Daniel J. Evans, a first-term representative, former King County engineer, and future governor. The act permitted the formation of a metropolitan government with any or all of six functions: water supply, sewage and garbage disposal, transportation, comprehensive planning, and park administration. Establishment of a Metro would require passage of a public referendum.

The first effort to win public acceptance for spending money to clean up the lake occurred in March 1958, when a proposal to establish a Metro charged with sewage disposal, transportation, and comprehensive planning was placed on the ballot. The proposal won 54.4% of the vote, but was defeated through a complicated system of weighting votes separately in Seattle and the rest of King County. Many people outside Seattle believed that the plan was an effort to tax them for the expenses of the city. Ellis and his committee revised the scope of their plan, targeting only water pollution control and reflecting the urgency posed by the deteriorating state of Lake Washington. A revised proposal, with the single function of sewage disposal, was approved on September 9, 1958, winning 58% of the vote in Seattle and 67% in the rest of the county.

Lake Washington obviously had become a focus of regional concern among the numerous communities that populated its shores. The water and beaches served for recreation, and the lake itself was deemed of aesthetic value.

A genuine sense of pride and responsibility is evident in the political arguments that surrounded the issue. Citizens were asked to undertake, at the expense of about $2 per month for each household served,
CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 305
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. a public project of sewage diversion that was at the time the most costly pollution control effort in the nation. The plan called for construction of a massive trunk sewer to divert all effluent from around the lake, to treat it, and to discharge it at great depth in Puget Sound. Tidal flushing guaranteed that objectionable quantities of nutrients would not accumulate in the estuary.

Edmondson played no part in the partisan politics, but his scientific knowledge and judgment were a deciding asset in the Metro campaign. By supplying facts and generally making himself available to answer questions from the mass media or private citizens, he provided the authority that backed the movement with facts and logic. Ellis praised Edmondson's stand years afterward for providing the facts needed to quiet the critics. When he spoke, Ellis reported, "he made us feel that the Lord God was standing right behind us on this one" (Chasan, 1971, p. 11). Privately and professionally, Edmondson reported the Lake Washington case as an experiment in lake fertilization. His scientific publications during this period traced the departure of chemical and biological conditions from the historical conditions and attempted to discern the general quantitative relationships between nutrient additions and primary productivity in lakes.

Edmondson had been able to predict in his letter to Ellis of February 1957 a serious and rapid decline in water quality. He wrote: "Within a few years we can expect to have serious scum and odor nuisances. . . . Judging by the speed with which the process has gone in other lakes, I would expect distinct trouble here within five years, although isolated occurrences might come earlier." The important elements of his predictions that gave heart to the proponents of the diversion plan were that Edmondson thought that the lake had not yet been irreversibly damaged and that diversion would lead to a decrease in the abundance of blue-green algae. These were the plants that clouded the fertilized water, rafted to shore, and decomposed or otherwise fouled the lake. His predictions were based on fundamental principles of mass balance, stoichiometry, and an opinion that, to a large degree, changes in lake conditions are reversible when factors forcing the changes are reversed.

The specific basis for quantitative predictions was a conceptual and graphic model that related changes in lake properties from known initial conditions to changes in nutrients (Edmondson, 1979). The model assumed a limited return of nutrients from sediment deposits on the basis of chemical conditions in the lake and work done decades earlier in Germany and England (Mortimer, 1941, 1942; Ohle, 1934). Finally, it required knowledge of the lake's water budget, to permit calculation of a rate of dilution. From these basic facts and hypotheses, it was possible to project not only
CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 306
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. the speed of deterioration, but also the rate of recovery with different diversion schemes. Years later, the same ideas were used by other limnologists to construct mathematical models of lake conditions in response to nutrient income and hydrology (Piontelli and Tonolli, 1964; Vollenweider, 1969, 1975, 1976); and the principles remain guiding tenets of modern lake management (Chapra and Reckhow, 1983; Reckhow, 1979). From a strictly scientific viewpoint, the exercise encouraged thought about nutrient budgets and helped to integrate studies of lakes with their watersheds. Equally important, the scientific studies helped to elaborate the quantitative links between nutrients and productivity. Many limnologists of the early twentieth century had been trained in the shadow of Forbes's (1925) philosophical essay "The Lake as a Microcosm" and had confined their investigations within lakeshore boundaries.

Forbes himself took a much broader view of lakes and their watersheds than the title suggests, but most limnologists at the time began and ended their studies at the shoreline. Edmondson, however, had visited Wisconsin as a graduate student while C. N. Sawyer was laboring to construct the first budgets of fertilization for the lakes at Madison (Sawyer, 1947). The perspective he gained was holistic and comprehensive. The new view might better be termed "the lake in an ecosystem."
Groundbreaking ceremonies for the new project were held in July 1961.

Meanwhile, the lake deteriorated according to predictions. On July 3, 1962, the Seattle Post-Intelligencer reported "Lake Washington Brown — That's Algae, Not Mud and It'll Be There For the Next 10 Years." Visibility in lake water had declined from 4 m in 1950 to less than 1 m by 1962. The first diversions were slated for the next year, and, on the basis of the timetable for later diversions, Edmondson estimated that the lake would revert to its condition of 1949 by about 5 years after completion of the project. By October 5, 1963, the Post- Intelligencer had dubbed Lake Washington "Lake Stinko," with nuisance conditions at their peak just before effluent diversion.

The rest of the public record is a series of congratulatory editorials and progress reports in city and suburban newspapers. One by one, waste treatment plants around the lake had their effluent diverted. The first diversion was in 1963, and the last was in 1968. The trend of deterioration stopped in 1964; conditions that summer were no worse than in 1963. By 1965, it was apparent that water transparency, algal abundance, and phosphate concentrations were improving. On November 19, 1965, Edmondson predicted in his address to Sigma Xi, the scientific research society, that the lake would return to its pre-1950s condition within 6 years and that it would be possible to see the bottom as deep as 6 m. In the previous summer, he had made the same prediction to K. Wuhrmann, a skeptical
CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 307
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. colleague from Zurich. At meetings of the International Association for Theoretical and Applied Limnology, Wuhrmann had argued that sediment release of phosphorus would extend the recovery for decades. The scientists disagreed about one of the principal hypotheses that Edmondson had used for his quantitative predictions.

Phosphorus arrives at the sediments in the form of detrital material that decomposes more slowly than it becomes buried. Thus, the water only a few millimeters below the surface would contain great reservoirs of phosphate ions that owed their presence to the rich conditions of eutrophication. If the ions diffused through porous and unconsolidated sediment back into the water, they could renew productivity from that "internal" source. Edmondson reasoned that the oxidation-reduction potential of surficial sediments guaranteed that iron would exist in its ferric (+3) oxidized form. In that state, it could form insoluble ferric-hydroxyl-phosphates of indeterminate stoichiometry, and the sediments would become an "iron trap" for the phosphorus. This was what Mortimer had shown in microcosm experiments with mud from the English Lake District.

Rates of decomposition in Lake Washington were not high enough to exhaust the oxygen content of the deep water during stratification each summer. As long as the large hypolimnion remained toxic, redox potentials would favor ferric iron, and the "iron trap" could halt the upward diffusive flux of phosphate.

Wuhrmann doubted that principles governing ion speciation and fluxes inside model tanks could be freely extrapolated to whole lakes. Their friendly wager that summer and Wuhrmann's delivery of one bottle of Scotch during the 1971 congress in Leningrad highlight the intellectual excitement and new understanding that the Lake Washington experiment afforded to professional limnologists.

Edmondson's professional publications during the period reported the progress of physical, chemical, and biological changes in the lake (Edmondson, 1961, 1966, 1968, 1969a,b, 1970, 1972a,b). Transparency and algal abundance responded very quickly to the nutrient diversions. Species composition proved somewhat more intransigent. Even though biomass was reduced, Oscillatoria persisted into the early 1970s, making occasional appearances each summer.

Finally, it too was gone. Trophic equilibrium in response to altered nutrient loading was complete in 1975 (Edmondson, 1977a,b; Edmondson and Lehman, 1981). Concentrations of phosphorus were reduced nearly to equilibrium and were similar from year to year. Chlorophyll concentrations and algal biomass were dramatically reduced, in parallel with the nutrient changes. Transparency had increased, and all filamentous blue-green algae, including Oscillatoria, had been eliminated.

The experiment was complete, and the scientific community had learned CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 308
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. new lessons about dynamic processes in lakes. The general public enjoyed its own measure of international praise, as witness articles in Harpers (Clark, 1967), Smithsonian (Chasan, 1971), and Audubon (Kenworthy, 1971). Metro became one of the few noncities to win an "All-American City" award.

The public record ends here, but new sources of scientific curiosity grew from continuing investigations. After the few years of constancy, water transparency suddenly increased to a point never before recorded in the lake.

Visibility was 12 m at times. Accompanying the change was a further, drastic reduction in algal abundance. This time, there had been little or no change in watershed relations; the limnological conditions were changing despite relatively constant nutrient loading.

What had changed were the herbivores (Edmondson and Litt, 1982).

Throughout the doctoral studies of Comita and throughout the episode of enrichment and diversion, Lake Washington had been dominated by copepods, a group of microscopic crustacean plankton. The changing nutrient supply to the lake exerted controls on the biota from the base of the food chain, because nutrients are essential for plant growth. By the late 1970s, however, Lake Washington was at times dominated by cladocerans, particularly by members of the genus Daphnia. Controls on algal abundance had shifted to much higher in the food chain. The cladocerans are able to reproduce faster than the copepods, and their success reduced the algae by sheer numbers and grazing pressure.

Why had the zooplankton community changed? Daphnia had not dominated the lake even in 1933. Did it have anything to do with changes set in motion by the enrichment episode? The answers to this new puzzle are being debated now, because similar shifts have been recorded in Lake Tahoe, in Lake Michigan, and in ponds of central Europe. It is known that the answer lies in deciphering the balance of forces that affect birth and death rates among the potentially dominant populations. The success of predatory invertebrates and planktivorous fish is involved in present hypotheses. Edmondson and Litt (1982) proposed that the changes might be traced to the decline of an important predator on Daphnia. Selective predation is known to be a major force in determining species composition in zooplankton communities (Hrbacek, 1962; Hrbacek et al., 1961). Neomysis mercedis, which was very abundant during the 1950s and early 1960s, suddenly declined in the mid-1960s with the rise of the longfin smelt Spirinchus thaleichthys. Neomysis strongly selects Daphnia over other planktonic crustaceans (Murtaugh, 1981), and Spirinchus feeds largely on Neomysis (Eggers et al., 1978). Released from this predation, Daphnia nonetheless took 10 years to dominate the lake plankton. Delays inherent in life histories or colonization times were insufficient to explain the gap.

CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 309
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

The reason for the delay seems to have been the continued presence of Oscillatoria. Long individual trichomes of O . rubescens and a few other filamentous species clog the feeding mechanism of Daphnia and force the animal to eject entire boli of food and to engage in elaborate grooming behavior (Infante and Abella, 1985). Copepods like Diaptomus do not seem to exhibit similar evidence of interference and can thrive in the presence of the trichomes, possibly because of hydromechanical differences in food capture. Thus, it was not until manipulations of the nutrient base excluded Oscillatoria from the lake that Daphnia could assert its dominance among the zooplankton.

Changes at many trophic levels are thus relevant to the scientific side of the case study of Lake Washington. They illustrate the ease with which the solving of environmental ''problems" can blend with ecological investigation.

Retrospective analyses of the public record make a good lesson in civics, but thoughtful progress in science is cheated if investigations do not uncover new challenges and point to new paths of inquiry, as does the study of Lake Washington and its biological community.

REFERENCES
Algae increase noted in Lake Washington. Seattle Times. August 11, 1955.

Anderson, G. C. 1954. A Limnological Study of the Seasonal Variation of Phytoplankton Populations. Ph.D. thesis, University of Washington, Seattle.

Chapra, S. C., and K. H. Reckhow. 1983. Engineering Approaches for Lake Management. Vol. 2.

Mechanistic Modeling. Ann Arbor Sciences, Ann Arbor, Mich.

Chasan, D. J. 1971. The Seattle area wouldn't allow the death of its lake. Smithsonian 2(4):6-13.

Clark, E. 1967. How Seattle is beating water pollution: Metro's project. Harpers 234:91-95 (June).

Comita, G. W. 1953. A Limnological Study of Planktonic Copepod Populations. Ph.D. thesis, University of Washington, Seattle.

Comita, G. W., and G. C. Anderson. 1959. The seasonal development of a population of Diaptomus ashlandi Marsh, and related phytoplankton cycles in Lake Washington. Limnol. Oceanogr. 4:37-52.

Edmondson announces pollution may ruin lake. University of Washington Daily. October 13, 1955.

Edmondson, W. T. 1961. Changes in Lake Washington following an increase in the nutrient income. Verh. Int. Verein. Limnol. 14:167-175.

Edmondson, W. T. 1966. Changes in the oxygen deficit of Lake Washington. Verh. Int. Verein.

Limnol. 16:153-158.

Edmondson, W. T. 1968. Water-quality management and lake eutrophication: The Lake Washington case. Pp. 139-178 in T. H. Campbell and R. O. Sylvester, eds. Water Resources Management and Public Policy. University of Washington Press, Seattle.

Edmondson, W. T. 1969a. Cultural eutrophication with special reference to Lake Washington. Mitt.

Int. Verein. Limnol. 17:19-32.

CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 310
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Edmondson, W. T. 1969b. Eutrophication in North America. Pp. 124-149 in Eutrophication:

Causes, Consequences, Correctives. Proceedings of a Symposium. National Academy of Sciences, Washington, D.C.

Edmondson, W. T. 1970. Phosphorus, nitrogen and algae in Lake Washington after diversion of sewage. Science 169:690-691.

Edmondson, W. T. 1972a. The present condition of Lake Washington. Verh. Int. Verein. Limnol. 18:284-291.

Edmondson, W. T. 1972b. Nutrients and phytoplankton in Lake Washington. Am. Soc. Limnol.

Oceanogr. Spec. Symp. 1:172-193.

Edmondson, W. T. 1977a. Recovery of Lake Washington from eutrophication. Pp. 102-109 in J.

Cairns, Jr., K. L. Dickson, and E. E. Herricks, eds. Recovery and Restoration of Damaged Ecosystems. University Press of Virginia, Charlottesville.

Edmondson, W. T. 1977b. Trophic Equilibrium of Lake Washington. EPA-600/3-77-087.

Environmental Research Laboratory, U.S. Environmental Protection Agency, Corvallis, Oreg.

Edmondson, W. T. 1979. Lake Washington and the predictability of limnological events. Arch.

Hydrobiol. Beih. 13:234-241.

Edmondson, W. T., and J. T. Lehman. 1981. The effect of changes in the nutrient income on the condition of Lake Washington. Limnol. Oceanogr. 26:1-29.

Edmondson, W. T., and A. H. Litt. 1982. Daphnia in Lake Washington. Limnol. Oceanogr. 27:272-293.

Edmondson, W. T., G. C. Anderson, and D. R. Peterson. 1956. Artificial eutrophication of Lake Washington. Limnol. Oceanogr. 1:47-53.

Eggers, D. M., et al. 1978. The Lake Washington ecosystem: The perspective from the fish community production and forage base. J. Fish. Res. Bd. Can. 35:1553-1571.

Forbes, S. A. 1925. The lake as a microcosm. [Reprinted.] Bull. III. Nat. Hist. Surv. 15:537-550.

Hasler, A. D. 1947. Eutrophication of lakes by domestic drainage. Ecology 28:383-395.

Hrbacek, J. 1962. Species composition and the amount of the zooplankton in relation to the fish stock. Rozpr. Cesk. Akad. Ved. Rada Mat. Prir. Ved. 10:1-116.

Hrbacek, J., M. Dvorakova, M. Korinek, and L. Prochazkova. 1961. Demonstration of the effect of fish stock on the species composition of zooplankton and the intensity of metabolism of the whole plankton association. Verh. Int. Verein. Limnol. 14:192-195.

Infante, A., and S. E. B. Abella. 1985. Inhibition of Daphnia by Oscillatoria in Lake Washington.

Limnol. Oceanogr. 30:1046-1052.

Kenworthy, E. W. 1971. How Seattle cleaned up. Audubon 73:105-106.

Lake's play use periled by pollution. Seattle Times. July 11, 1955.

Lake Stinko. Seattle Post-Intelligencer. October 5, 1963.

Lake Washington brown— That's algae, not mud. . . . Seattle Post-Intelligencer. July 3, 1962.

Lesson of Switzerland lakes brought home to Seattle area. Seattle Times. April 18, 1956.

Metro area citizens should be proud of achievement. Seattle Times. September 19, 1965.

Mortimer, C. H. 1941. The exchange of dissolved substances between mud and water in lakes. Parts 1 and 2. J. Ecol. 29:280-329.

Mortimer, C. H. 1942. The exchange of dissolved substances between mud and water in lakes. Parts 3 and 4. J. Ecol. 30:147-201.

Murtaugh, P. A. 1981. Selective predation by Neomysis mercedis in Lake Washington. Limnol.

Oceanogr. 26:445-453.

Ohle, W. 1934. Chemische und physikalische Untersuchungen norddeutscher Seen. Arch.

Hydrobiol. 26:386-464, 584-658.

CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 311
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Peterson, D. R. 1955. An Investigation of Pollution Effects in Lake Washington (1952-1953).

Washington Pollution Control Commission Tech. Bull. 18, Seattle, Wash.

Piontelli, R., and V. Tonolli. 1964. Residence time of lake water in relation to enrichment, with special reference to Lago Maggiore. Mem. Ist. Ital. Idrobiol. 17:247-266. [in Italian] Reckhow, K. H. 1979. Empirical lake models for phosphorus: Development, applications, limitations and uncertainty. Pp. 193-221 in D. Scavia and A. Robertson, eds. Perspectives on Lake Ecosystem Modeling. Ann Arbor Sciences, Ann Arbor, Mich.

Sawyer, C. N. 1947. Fertilization of lakes by agricultural and urban drainage. J. N. Engl. Water Works Assoc. 61:109-127.

Scheffer, V. B., and R. J. Robinson. 1939. A limnological study of Lake Washington. Ecol. Monogr. 9:95-143.

Sylvester, R. O., W. T. Edmondson, and R. H. Bogan. 1956. A new critical phase of the Lake Washington pollution problem. Trend in Engineering 8(2):8-14.

Vollenweider, R. A. 1969. M öglichkeiten und Grenzen element ärer Modelle der Stoffbilanz von Seen. Arch. Hydrobiol. 66:1-36.

Vollenweider, R. A. 1975. Input-output models with special reference to the phosphorus loading concept in limnology. Schweiz. Z. Hydrol. 37:53-84.

Vollenweider, R. A. 1976. Advances for defining critical loading levels for phosphorus in lake eutrophication. Mem. Ist. Ital. Idrobiol. 33:53-83.

Committee Comment
Several features of environmental problem-solving are illustrated by the Lake Washington example. The project itself was regarded as an experiment, and scientists were able to test their hypotheses during the study. Analogs existed in the scientific literature, so investigators were able to reason partly from first principles and partly by reference to other examples. At one point, Edmondson read through an article by A. D. Hasler (1947) that reviewed the history of cultural eutrophication in Europe and North America. He underlined "Oscillatoria" each time the word appeared in the text and discovered that the organism was a nearly ubiquitous indicator of eutrophication. Similarly, the relationship between Neomysis and Daphnia was suggested in part by experiences in Lake Tahoe when Mysis relicta was introduced as a forage food for fish (Richards et al., 1975).

To understand the course of events in Lake Washington, one must draw on most of the sources of knowledge identified here. The initial events were treated as an experiment in lake fertilization that could improve our understanding of the ways that nutrient inputs control the biological character of lakes. Changes in the plankton community after the enrichment experiment required analyses of biological events at the population and community levels. Plankton community structure was seen to be governed by a variety of species interactions, including predation and interference. Superficially, the lake appeared to exhibit alternative stable states with
CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 312
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. regard to plankton composition. In fact, the transition from one state to another was a logical consequence of changing fields of predators and algae.

Most important, the scientists were able to identify a few factors and processes that were acting with special force among the myriad present.

Edmondson and Litt (1982) wrote: "We assume that the processes controlling the population are simultaneously affected by many factors and that changing any one can affect the population. At some times, one factor may dominate the others quantitatively." Scientific judgment was needed to identify the stoichiometries and relations among nutrient income, hydrology, and algal production. Similar reasoning helped to establish the likely causes for species alterations. No amount of descriptive field study alone would establish causality firmly enough to permit quantitative management decisions. Judgments had to be made about how algal production would respond to alterations in nutrient supplies and about the importance of nutrient returns from the sediments. That required knowledge of more than the biota alone. It was necessary to regard the organisms as being integrated with their physical and chemical surroundings.

Spatial relations and the vertical differentiation that arises from thermal stratification were important, too. Lake Washington is a warm, monomictic lake; therefore, it circulates all winter long. Winter is also the time of greatest fluvial discharge, and on the average one-third of the lake's volume is renewed each year. Most of the water drains from the Cascade Range and is very low in dissolved salts or nutrients of any kind. Each winter, the lake is thus diluted by water of low nutrient content. Edmondson could argue securely that, if fertilizing discharges from municipalities ceased, the accumulated nutrients and algal biomass could be flushed from the basin within a few years. Because Lake Washington possessed a large hypolimnion with more than adequate reserves of oxygen to last through summer stratification, most of the phosphorus locked in the sediments would stay there.

Furthermore, in the case of Lake Washington, scientific judgment backed by logic and data was separated from emotional statements. The decision to raise public funds and divert the effluent was political, not scientific; indeed, sound and important discoveries probably would have accompanied a study of continued deterioration of the lake.

The case of Lake Washington is exemplary, not because the forecasts were so accurate, but because events were documented and reported in comprehensible fashion. The documents reveal a remarkable synergism of scientific and public awareness about an environmental issue. The retrospective account makes the scientific issues sound perhaps more cut and dried than they were at the time. It might seem that the only suspense
CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 313
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. was related to the public's willingness to spend money to improve water quality.

In the 1960s, however, debates about the causes of lake eutrophication were common, and the debates eventually spawned an inquiry by the National Research Council (1969). The Research Council's report directed attention to phosphorus, but the evidence came in part from results seen in Lake Washington. Vocal scientific lobbies had argued that carbon and nitrogen could be limiting elements in many aquatic habitats and that phosphorus control would therefore be insufficient to halt eutrophication. Many of the conflicting opinions were published in the proceedings of a special symposium of the American Society of Limnology and Oceanography (Likens, 1972). It was in the early stages of this debate that the Lake Washington experiment was conceived and executed.

Despite the obvious success in curbing pollution of the lake, the experiment could not by itself prove that phosphorus was the culprit, even though predictions had been based on that assumption. The action of removing waste treatment effluent from a lake lacks the rigor of a conventional laboratory experiment, in that many factors are manipulated simultaneously. Opponents could argue that improvement arose because some unmeasured trace metal or unknown growth factor was removed with the effluent and that phosphorus control elsewhere might be costly and irrelevant. Indeed, many investigators had discovered that, when lake phytoplankton was enclosed in bottles and subjected to single-nutrient additions, carbon, nitrogen, or trace metals could often stimulate their metabolism and growth (Likens, 1972). This very type of observation is the basis of present views about nitrogen limitation in the oceans (Ryther and Dunstan, 1971).

The principal difference between bottle bioassays and Lake Washington, however, lies not in methodological detail, but in the scale of the manipulation.

In the case of Lake Washington, an entire ecosystem was manipulated. At lake- wide scales, exchange processes at air-water and sediment-water interfaces become important, and responses can be followed over long periods. Lake Washington became one of the pioneer "whole-lake" experiments. Within a few years, Canadian limnologists had established an Experimental Lakes Area in northwestern Ontario (Johnson and Vallentyne, 1971) and were setting out to test nutrient controls of eutrophication more rigorously than could ever be possible in Lake Washington. Experimental lakes were purposely fertilized with nitrogen, phosphorus, and carbon, singly and in combinations. The results showed beyond doubt that phosphorus was the master controlling nutrient, as far as eutrophication was concerned (Schindler, 1977). When lakes were fertilized with
CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 314
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. phosphate alone, algal growth reached bloom proportions, because inorganic carbon entered the lake water from the atmosphere and continuously replaced the carbon used by the algae during photosynthesis. Similarly, species composition in these lakes became more strongly represented by nitrogen-fixing blue-green algae (cyanobacteria), which formed the ultimate reservoir for the nutrient. Phosphorus, however, has no gaseous atmospheric phase, so its rate of supply to a lake basin sets an absolute limit on standing crops. With only bottle bioassays or small-scale experiments of short duration, the ultimate consequences of a manipulation could not be forecast. Bottles are closed to gas exchange with the atmosphere, and the experiments are too brief to permit species assemblages to change. In short, no study short of a whole-lake manipulation could have provided an adequate analogy to this experiment. That is why the lessons from Lake Zurich and the lakes in Madison, Wisconsin, were so valuable in the early, predictive stages of the project.

As a historical footnote, the success story of Lake Washington might have heartened those in Switzerland who were trying to clean up Lake Zurich.

Cultural eutrophication in Lake Zurich had been accelerating since 1896 (Thomas, 1969). Three-stage waste treatment plants around the lake with chemical precipitation processes for the removal of the phosphate were introduced. The first started operating in 1967; since then, all Zurich treatment plants have had precipitation installations to eliminate phosphate (Dietlicher, 1974). The improvements coincided with reductions in phosphate concentrations in the lake and improvements in water quality.

References
Dietlicher, K. 1974. The Water Quality of the Lakes of Zurich and "Walensee." Zurich Waterworks, Zurich, Switz.

Edmondson, W. T., and A. H. Litt. 1982. Daphnia in Lake Washington. Limnol. Oceanogr. 27:272-293.

Hasler, A. D. 1947. Eutrophication of lakes by domestic drainage. Ecology 28:383-395.

Johnson, W. E., and J. R. Vallentyne. 1971. Rationale, background, and development of experimental lake studies in northwestern Ontario. J. Fish Res. Bd. Can. 28:123-128.

Likens, G. E., ed. 1972. Nutrients and eutrophication: The limiting nutrients controversy. Am. Soc.

Limnol. Oceanogr. Spec. Symp. 1:1-328.

National Research Council. 1969. Eutrophication: Causes, Consequences, Correctives. Proceedings of a Symposium. National Academy of Sciences, Washington, D.C.

Richards, R. C., C. R. Goldman, T. C. Frantz, and R. Wickwire. 1975. Where have all the Daphnia gone? The decline of a major cladoceran in Lake Tahoe, California-Nevada. Verh. Int.

Verein. Limnol. 19:835-842.

CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 315
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Ryther, J. H., and W. M. Dunstan. 1971. Nitrogen, phosphorus, and eutrophication in the coastal marine environment. Science 171:1008-1013.

Schindler, D. W. 1977. Evolution of phosphorus limitation in lakes. Science 195:260-262.

Thomas, E. A. 1969. Kulturbeeinflusste chemische und biologische Ver änderungen des Z ürichsees im Verlaufe von 70 Jahren. Mitt. Int. Verein. Limnol. 17:226-239.

CONTROL OF EUTROPHICATION IN LAKE WASHINGTON 316
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. 21 Raising the Level of a Subarctic Lake
A common form of environmental intervention is damming a river to produce an artificial lake or raising the level of an existing lake to create a larger one. The environmental problems arising from these manipulations are more complex than those resulting from pollution of lakes, because the manipulations establish new flow patterns, inundate existing shorelines, initiate erosional processes that create new shorelines, and change the morphologic characteristics of the lakes. For these reasons, prediction of the consequences of altering the dimensions of lakes is more difficult and more uncertain than prediction of the responses of lakes to pollutants and their elimination.

Southern Indian Lake, in northern Manitoba, was raised 3 m in 1976, and rivers were diverted so that the flow of water through the lake was reduced by about 75%. The case study reveals that, despite careful planning of the project, many of the predictions were seriously wrong. The limnologists who carried out the preproject studies later analyzed the disparities between their predictions and the outcomes. As a result, future studies for similar projects will be based on better models.

RAISING THE LEVEL OF A SUBARCTIC LAKE 317
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Case Study
JOHN T. LEHMAN, Division of Biological Sciences, University of Michigan, Ann Arbor
Southern Indian Lake (57 ° N, 99 ° W), in northern Manitoba, is a preproject basin along the Churchill River that drains northward to Hudson Bay.

In 1976, the natural lake outlet was dammed and the lake level raised 3 m, so that the Churchill could be diverted southward across a drainage divide. Of the natural river flow of almost 1,000 m
3/second, 75% was diverted into the Nelson
River and no longer flows through the main lake. The diversion scheme permitted the combined flow of drainage from over 1.4 million square kilometers of the northern Great Plains to flow through a single series of dams and hydroelectric generating stations. The project had been planned in principle since the 1950s, and the lake had been the object of careful impact assessments.

Extensive and detailed knowledge of the later ecological events has come through the efforts of scientists at the Freshwater Institute in Winnipeg. Institute scientists were charged in 1976 with a long-term study to evaluate the predictive capabilities of impact assessments and to generate new capabilities with increased quantitative precision.

The preimpoundment assessment work was done by professional limnologists who used existing modeling techniques and a literature-based paradigm regarding reservoir dynamics. The project was treated as a large-scale experiment to test the hypotheses of the impact assessments and to modify the paradigm. Many predictions were realized, and others differed from expectation only in quantitative detail. Most biological responses above the primary trophic level, however, were predicted either incorrectly or not at all. Assessments based on analogy and scaling were often misleading. The most striking unpredicted change involved the commercial whitefish fishery and the local economy that depended on it.

Southern Indian Lake supported the largest commercial fishery in northern Manitoba. About 85% of the catch was Coregonus clupeaformis, the lake whitefish. During the 3 decades before impoundment, the annual catch averaged 334 metric tons, almost exclusively high-quality fish of "export" grade (Bodaly et al., 1984b). The dendritic basin of Southern Indian Lake consists of several subbasins that are separated by islands and channel constrictions. The high quality of the fishery was maintained by selective fishing in the most profitable basins. Elsewhere, stocks were dominated by darker fish with high incidences of muscle cysts of the parasitic cestode Triaenophorus crassus.

RAISING THE LEVEL OF A SUBARCTIC LAKE 318
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Southern Indian Lake is in a region underlain by Precambrian shield bedrock. The overlying sedimentary deposits are glaciolacustrine or glaciofluvial, tracing either to the glacial Lake Agassiz or to outwash during the retreat of the last continental glaciation. Before the 3-m impoundment, 76% of the shoreline was exposed bedrock. Mean annual temperature in the region is - 5°C, so permafrost is widespread in the glacial deposits. Vegetation is characteristic of boreal forest or taiga (Newbury et al., 1984).

On the basis of federal and provincial feasibility studies, Manitoba Hydro in the 1960s proposed raising Southern Indian Lake by 10 m. This high-level impoundment was selected to optimize the generation of electricity, without regard for environmental issues or a thorough study of the cost-effectiveness of hydrologic storage in the lake. In response to public concern, the utility contracted with Underwood-McLellan and Associates Ltd., a consulting engineering firm, to produce a predevelopment impact assessment of several diversion scheme options. Published in 1970, that study caused Manitoba Hydro to opt instead for a less costly low-level (3-m) impoundment, on the grounds that it would have proportionally smaller effects on watershed and shoreline. A federal-provincial study board was later commissioned to investigate the effects of the new plan. The Lake Winnipeg, Churchill and Nelson Rivers Study Board undertook its work concurrently with construction activities and issued its final report in April 1975, shortly before the diversion. The two studies differed considerably in their objectives and data bases. The first involved no substantial field work and was aimed at determining costs and benefits of different diversion schemes. The second was initiated after the configuration and construction schedule of the project were fixed; scientists were charged with providing reliable baseline data, predicting future conditions, and then measuring them.

The federal-provincial study was comprehensive. It included studies of hydrology, soils, fisheries, limnology, wildlife, geology, recreation, forestry, archaeology, navigation, and socioeconomic factors. University scientists, private consultants, and government laboratories were involved in different aspects of the investigations. The fisheries-limnology component was summarized by Hecky and Ayles (1974), who described the existing conditions of the basin in concise, quantitative terms and then presented a series of predictions.

The limnologists recognized that the diversion scheme would alter flow patterns in the lake and that some regions would be flushed more quickly than others. In their opinion, the altered flow would have a greater impact on water chemistry than would inundation itself, because Churchill River water greatly influenced the main basins of the lake under natural conditions. If normal flowthrough were reduced by 80%, which was the
RAISING THE LEVEL OF A SUBARCTIC LAKE 319
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. proposed long-term average, local drainage could dominate conditions in a large area of the lake.

Projections were based in part on conceptual and numerical models that related nutrient loading and water exchange with algal abundance and productivity. These were new, sophisticated versions of the approach that Edmondson had used to evaluate nutrient loading in Lake Washington (see Chapter 20 and, e.g., Dillon and Rigler, 1974, and Vollenweider, 1975). The models had sound empirical support and widespread acceptance as a means to predict algal abundance and trophic state. Predictions were for steady-state conditions that would follow transient inundation effects. Reduced inflow of Churchill River water would lead to diminished turbidity and nutrients from fluvial sediment loading in the long run, but water residence time would increase. Overall annual algal production was expected to decline by 33-50% in the main basins. The diversion was expected to have a positive effect in one basin along the diversion route, which previously had received little Churchill water directly.

Inundation was expected to affect water quality, but not as drastically in the long term as diversion. The conventional reservoir paradigm predicted that nutrient release from soils and vegetation would increase postimpoundment production at all levels in the food chain. The overburden of frozen glaciolacustrine clays around Southern Indian Lake, however, had great potential for erosion. Increased turbidity would negate benefits of increases in nutrients, according to surveys of light penetration and primary productivity.

The inundation was thus expected to have no net favorable effects and to be detrimental in littoral regions.

The limnologists expected higher trophic levels generally to track changes in the algae, although not necessarily immediately or in strict proportion.

Declines in planktonic and zoobenthic production in main basins were expected to diminish fish production gradually. Suitable spawning sites for whitefish in the inundated areas were of concern, as was the chance that diversion could disorient migrating fish. Hecky and Ayles (1974) foresaw difficulties for the commercial fishery that would cause it to decline more than overall biological production would decline. The traditional fishery was concentrated in the largest basin, which was certain to experience the greatest deprivation from diversion. As production potential fell, fishermen would see their yields per unit of fishing effort decline. No evidence of overfishing yet existed, but many marginal operators would probably leave the fishery if faced with a need to increase their effort substantially. The scientists expected an overall decrease of 13% in commercial yield. Lakewide average biological production was expected to decline by 10%. The declines were forecast on the basis of expected redistributions of nutrient income from the Churchill River. They RAISING THE LEVEL OF A SUBARCTIC LAKE 320
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. applied to the long-term, steady-state outcome of the manipulation. Greater uncertainty was attached to short-term responses, and statements on that matter were largely qualitative.

Early in the study, the federal-provincial study board had identified the need for long-term ecological monitoring of the impact areas. Hecky and his colleagues at the Freshwater Institute initiated their own case study of the Southern Indian Lake reservoir in 1974 with the preimpoundment work. The operational regimen has been roughly that considered by the impact study, so predictions can legitimately be compared with results. The lake provides a good opportunity to evaluate scientific judgments retrospectively. The technical results of the decade-long effort were published in 17 papers in Volume 41 of the Canadian Journal of Fisheries and Aquatic Sciences (1984).

Quantitative predictions had been generally lacking in the impact assessments of 1970 and 1975. The 10% reduction in biological production offered by Hecky and Ayles was cited repeatedly in the 1975 summary report as one hard prediction (Lake Winnipeg, Churchill and Nelson Rivers Study Board, 1975). The lack of postimpoundment case studies in northern Canada and the absence of verified models for important processes precluded other precise estimates. The limnologists recommended research into process models as one way to improve predictions. They had been forced to rely, as had those who conducted the 1970 study, on results from reservoirs in temperate and tropical regions. Excellent studies had been done in Siberia, but those reservoirs were deep, steep-sided impoundments, and their physical characteristics differed from those of Southern Indian Lake. The Siberian reservoirs, for instance, are deep enough to develop thermal stratification with reduced oxygen in the bottom water during the summer, but Hecky and Ayles expected no such thermal stratification or deoxygenation in Southern Indian Lake. The potential for extensive shoreline erosion caused their predictions to differ from conventional wisdom regarding productivity and fish yield in new reservoirs.

In retrospect, the predevelopment predictions were extremely good with regard to nutrients and algal production. No thermal stratification developed, and shoreline erosion was extensive (Hecky, 1984; Hecky and McCullough, 1984; Newbury and McCullough, 1984. The increased nutrients from erosion and decay of vegetation increased phosphorus concentrations as expected, but turbidity increased also. Physiological studies revealed that phosphorus deficiency among the algae had been replaced by light limitation in turbid regions of the lake, and no significant increase in productivity occurred (Hecky and Guildford, 1984; Planas and Hecky, 1984). Regions with high transparency showed increased productivity. Few changes were predicted incorrectly. The notable exception was in
RAISING THE LEVEL OF A SUBARCTIC LAKE 321
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. fish recruitment. Both impact assessments had predicted that northern pike (Esox lucius ) would benefit from the increase in habitat associated with inundation and that the sport fishery would improve. Some short-term spawning problems were anticipated for walleye ( Stizostedion vitreum vitreum ) and the important whitefish, but the fish were expected to exploit new spawning grounds and to recover. Esox in fact experienced no increase in survival or growth, and the commercial whitefish fishery collapsed (Bodaly and Lesack, 1984; Bodaly et al., 1984b).

Loss of the viable commercial fishery in Southern Indian Lake was a blow to the local economy, and it had been unexpected. The reasons for the change are complex, but they were anticipated in part by Hecky and Ayles (1974). The catch per unit effort indeed declined on the traditional fishing grounds, to about half the preflooding value. Total catches were maintained by major increases in total fishing effort. The fishermen also began to exploit regions that had previously been avoided because the fish were of lower marketability, being darker and having higher rates of cestode infestation. The lower-quality fish made up 12-72% of the summer catch in the years after impoundment, whereas they were insignificant components of the original fishery. In 1982, the lake was changed from ''export" to "continental" classification, and the commercial value of the catch plummeted. Total catch eventually fell to one-third of its preimpoundment size as many operators abandoned the fishery (Bodaly et al., 1984b; Wagner, 1984).

The catch declines apparently stemmed from migrations of fish away from traditional fishing grounds. Genetic markers measured among whitefish stocks before and after impoundment indicate that stocks became redistributed when normal flow patterns were altered. There is an indication of net emigration of whitefish from Southern Indian Lake. Compensation payments to commercial fishermen by Manitoba Hydro subsidized their efforts from 1977 to 1982.

Without the payments, production costs would have exceeded fishery revenues.

In 1982, Manitoba Hydro provided a one-time cash settlement of Can$2.5 million for all future losses (Bodaly et al., 1984b).

The fishery problems were not confined to collapse of the commercial enterprise. Changes had been triggered by physical manipulations that proved more far-reaching and persistent than anyone expected. The shoreline of Southern Indian Lake had been bedrock-controlled; 76% of the shore was exposed granite or gneiss. After flooding, only 14% of the new shore consisted of bedrock. Erosion had been expected by everyone, but not of the scale and duration produced when lakewater melted the permafrost. The shoreline retreated at up to 12 m/year. Lakewater melted and undercut the backshore zone, and that resulted in massive faulting of the
RAISING THE LEVEL OF A SUBARCTIC LAKE 322
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. overhanging shoreline. Newbury and McCullough (1984) predicted that it would take at least 35 years for the shoreline to be eroded to preflooding conditions and that high rates of sediment input from inundation would persist for decades.

The continuing flooding of terrestrial areas posed a new threat to fisheries, this time presenting a health hazard for domestic consumers. The flooded soils released mercury, which became concentrated particularly in the piscivorous northern pike and walleye (Bodaly et al., 1984a). Increased mercury concentrations are a common consequence of reservoir creation, but the phenomenon has been recognized only within the last decade (Abernathy and Cumbie, 1977; Cox et al., 1979; Kent and Johnson, 1979; Potter et al., 1975).

Soils need not be rich in mercury, and indeed the Southern Indian Lake source materials have a low or average mercury content (Bodaly et al., 1984a).

Bacterial methylation of inorganic mercury salts mobilizes the element, and it then associates with particles. Muscle mercury concentrations in the walleye and northern pike now exceed the Canadian marketing standards (0.5 ppm) and will probably remain increased for years.

A few other unexpected changes occurred. Mean lake temperatures decreased, owing to increased mean depth, river diversion, greater surface reflectance, and backscattering of incident solar radiation from the turbid water (Hecky, 1984). The zooplankton community changed, too (Patalas and Salki, 1984). The large-bodied crustacean Mysis relicta, formerly rare, became common. Large-bodied calanoid copepods like Limnocalanus macrurus increased. Cladocerans and small copepods decreased. Patalas and Salki ascribed the changes primarily to lower water temperatures and increased water depth, which enlarged the habitat of generally deep-dwelling species. They regarded an overall decrease in the abundance and biomass of crustacean zooplankton to be a consequence of the diversion-related temperature decrease.

Lower temperatures slow growth and development rates of both eggs and juveniles. The patterns suggest, moreover, that predation pressure from planktivorous fish decreased and that predatory invertebrates became an important force in the case of the plankton. The decline in predation by fish could reflect a decrease in stocks or unsuccessful foraging due to turbidity.

The changes in the zooplankton community in Southern Indian Lake parallel some of the changes that occurred in Lake Tahoe when Mysis was introduced (Goldman et al., 1979; Richards et al., 1975). It seems reasonable to ascribe the changes in zooplankton, including Limnocalanus and Mysis, to declines in fish stocks and foraging success. The findings are united with recent events in Lake Washington by the major theme in zooplankton community ecology of the last 2 decades, which emerged in
RAISING THE LEVEL OF A SUBARCTIC LAKE 323
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. the seminal works of Hrbacek (Hrbacek, 1962; Hrbacek et al., 1961) and Brooks and Dodson (1965): that the introduction of planktivorous fish causes wholesale alterations of species composition and size distribution in zooplankton communities. That zooplankton communities are shaped by the size-selective predatory behavior of fish is well established (Galbraith, 1967; Hall et al., 1976; Lynch, 1979; O'Brien, 1979; Wells, 1970; Zaret, 1980). The proximate mechanism is prey conspicuousness (Zaret, 1972; Zaret and Kerfoot, 1975; Zaret and Suffern, 1976). Much of the recent excitement in zooplankton community ecology has come from investigations of community structure in the absence of fish. Brooks and Dodson had used allometric estimates of metabolism and feeding to support their "size-efficiency hypothesis" — i.e., large animals simply outcompete small ones through economies of scale. Others contested that claim and argued with convincing empirical support that predation by invertebrates could eliminate small species (Dodson, 1974; Kerfoot, 1977; Zaret, 1975). Unlike vertebrates, these predators rely on mechanoreception to detect their prey and manipulate the prey to consume it.

This type of predation by invertebrates, mostly insect larvae and predatory crustacean zooplankton, has been linked to much morphological variation — including cyclomorphosis — within the taxa that they prey on (Halbach, 1971; Jacobs, 1965; Zaret, 1972).

Large invertebrate predators can thrive in the absence of fish, feeding on small-bodied animals and the juvenile stages of larger taxa. Large-bodied zooplankton have an advantage, particularly if they can produce large eggs and boost their offspring past the vulnerable juvenile period. When planktivorous fish abound, the invertebrate predators are held in check; so, too, are large herbivores. The small-bodied animals are released from predation. They might come to dominate the assemblage, even though the fish might eventually be forced to turn to them as prey (Brooks, 1968). This explanation probably figures in the events in Southern Indian Lake, although Patalas and Salki (1984) regarded the direct physiological effects of temperature to be dominant.

Hecky and colleagues undertook in 1984 their own retrospective analysis of environmental impact prediction and assessment based on their experience with Southern Indian Lake (Hecky et al., 1984). Scientists who had formulated the original predictions and then used the experiment to advance understanding provided their scientific insights into the limitations and capabilities of the prediction process. They conceded, for instance, that the 1970 "office" study was nearly as effective in forecasting impacts as was the 1975 study after a year of field work. Improved precision in the 1975 study resulted in part from having a better definition of the project. Hecky and colleagues concluded that both assessments were of
RAISING THE LEVEL OF A SUBARCTIC LAKE 324
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. only marginal overall utility, because significant impacts had not been predicted and because even the correct predictions had often been qualitative. Qualitative statements could not enter the quantitative cost-benefit analyses that influence major resource development.

The field study had proved invaluable, nonetheless, primarily as the means to document the baseline conditions in the lake. Without such data, any followup study might have been meaningless, and compensation payments might have been more difficult to arrange. The field study, in essence, provided insurance against unpredicted adverse environmental impacts (Hecky et al., 1984).

Three-meter impoundment was selected because of the high cost of accessory "saddle" dams that are needed to contain high-level impoundment and because the 1970 study had suggested that the effects of the originally planned 10-m impoundment might be too severe. The authors of the 1970 study had assumed that effects were proportional to the magnitude of manipulation. A 3-m change was assumed to have smaller effect than a 10-m change. In fact, impoundment levels greater than 3 m would have produced effects only slightly worse than actually occurred, because severe shoreline erosion began as soon as the natural range of water levels was exceeded and the impounded water entered the previously frozen backshore zone.

The Freshwater Institute scientists bemoaned the lack of suitable analog studies and the limitations of current reservoir paradigms. Conventional wisdom and expectations were based on experience with deep riverine basins with limited wind fetches. Indigenous biological processes had been emphasized nearly to the exclusion of physical forces. The most likely analog candidates either lacked permafrost features or had little fine-grained erodable material on the shores. In the absence of examples or experience, it was difficult for the scientists to judge the importance of these novel characteristics. However, they developed empirical relations between energy and erosion and can now use shoreline maps and meteorological records to predict lake volumes after erosion (Newbury and McCullough, 1984). Southern Indian Lake has changed the understanding of reservoir dynamics.

Responses at higher trophic levels showed the greatest failures of prediction. Hecky et al. (1984) rejected the notion that the responses are intrinsically unpredictable. They argued instead that possible responses, unless parts of existing paradigms, tend to be overlooked or considered only rarely.

Hecky et al. seemed to blame themselves for not anticipating the problems with fishery quality and with mercury contamination, in as much as elements of the logical puzzle were already in their grasp. They suggested that they were on surest ground when predicting energy flow,
RAISING THE LEVEL OF A SUBARCTIC LAKE 325
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. biomass, and general trophic relations, but that species-level predictions are elusive. Their most severe criticism was that they learned that current impact assessment was incomplete and unacceptable. They could also have mentioned that their studies opened the door to new understanding of reservoir processes and functions.

REFERENCES
Abernathy, A. R., and P. M. Cumbie. 1977. Mercury accumulation by largemouth bass (Micropterus salmoides) in recently impounded reservoirs. Bull. Environ. Contam.

Toxicol. 17:595-602.

Bodaly, R. A., and L. F. W. Lesack. 1984. Response of a boreal northern pike (Esox lucius) population to lake impoundment: Wupaw Bay, Southern Indian Lake, Manitoba. Can. J.

Fish. Aquat. Sci. 41: 706-714.

Bodaly, R. A., R. E. Hecky, and R. J. P. Fudge. 1984a. Increases in fish mercury levels in lakes flooded by the Churchill River diversion, northern Manitoba. Can. J. Fish. Aquat. Sci. 41:682-691.

Bodaly, R. A., T. W. D. Johnson, R. J. P. Fudge, and J. W. Clayton. 1984b. Collapse of the lake whitefish (Coregonus clupeaformis) fishery in Southern Indian Lake, Manitoba, following lake impoundment and river diversion. Can. J. Fish. Aquat. Sci. 41:692-700.

Brooks, J. L. 1968. The effects of prey size selection by lake planktivores. Syst. Zool. 17:272-291.

Brooks, J. L., and S. I. Dodson. 1965. Predation, body size, and the composition of the plankton.

Science 150:28-35.

Cox, J. A., J. Carnahan, J. DiNunzio, J. McCoy, and J. Meister. 1979. Source of mercury in fish in new impoundments. Bull. Environ. Contam. Toxicol. 23:779-783.

Dillon, P. J., and F. H. Rigler. 1974. A test of a simple nutrient budget model predicting the phosphorus concentration in lakewater. J. Fish. Res. Bd. Can. 31:1771-1778.

Dodson, S. I. 1974. Adaptive change in plankton morphology in response to size-selective predation: A new hypothesis of cyclomorphosis. Limnol. Oceanogr. 19:721-729.

Fudge, R. J. P., and R. A. Bodaly. 1984. Postimpoundment winter sedimentation and survival of lake whitefish (Coregonus clupeaformis) eggs in Southern Indian Lake, Manitoba. Can. J.

Fish. Aquat. Sci. 41:701-705.

Galbraith, M. G., Jr. 1967. Size-selective predation on Daphnia by rainbow trout and yellow perch.

Trans. Am. Fish. Soc. 96:1-10.

Goldman, C. R., M. D. Morgan, S. T. Threlkeld, and N. Angell. 1979. A population dynamics analysis of the cladoceran disappearance from Lake Tahoe, California-Nevada. Limnol.

Oceanogr. 24:289-297.

Halbach, U. 1971. Zum Adaptivwert der zyklomorphen Dornenbildung von Brachionus calyciflorus Pallas (Rotatoria). Oecologia 6:267-288.

Hall, D. J., C. W. Burns, and P. H. Crowley. 1976. The size-efficiency hypothesis and the size structure of zooplankton communities. Annu. Rev. Ecol. Syst. 7:177-208.

Hecky, R. E. 1984. Thermal and optical characteristics of Southern Indian Lake before, during, and after impoundment and Churchill River diversion. Can. J. Fish. Aquat. Sci. 41:579-590.

RAISING THE LEVEL OF A SUBARCTIC LAKE 326
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Hecky, R. E., and H. A. Ayles. 1974. Summary of fisheries-limnology investigations on Southern Indian Lake. Lake Winnipeg, Churchill and Nelson Rivers Study Board, 1971-1975.

Technical Report Appendix 5, Volume IA. Environment Canada Fisheries Service, Winnipeg, Man.

Hecky, R. E., and S. J. Guildford. 1984. Primary productivity of Southern Indian Lake before, during, and after impoundment and Churchill River Diversion. Can. J. Fish. Aquat. Sci. 41:591-604.

Hecky, R. E., and G. K. McCullough. 1984. Effect of impoundment and diversion on the sediment budget and nearshore sedimentation of Southern Indian Lake. Can. J. Fish. Aquat. Sci. 41:567-578.

Hecky, R. E., R. W. Newbury, R. A. Bodaly, K. Patalas, and D. M. Rosenberg. 1984.

Environmental impact prediction and assessment: The Southern Indian Lake experience.

Can. J. Fish. Aquat. Sci. 41:720-732.

Hrbacek, J. 1962. Species composition and the amount of the zooplankton in relation to the fish stock. Rozpr. Cesk. Akad. Ved Rada Mat. Prir. Ved. 10:1-116.

Jacobs, J. 1965. Significance of morphology and physiology of Daphnia for its survival in predator- prey experiments. Naturwissenschaften 52:141.

Kent, J. C., and D. W. Johnson. 1979. Mercury, arsenic and cadmium in fish, water, and sediment of American Falls Reservoir, Idaho, 1974. Pestic. Monit. J. 13:35-40.

Kerfoot, W. C. 1977. Implications of copepod predation. Limnol. Oceanogr. 22:316-325.

Lake Winnipeg, Churchill and Nelson Rivers Study Board. 1975. Summary Report:

CanadaManitoba Lake Winnipeg, Churchill and Nelson Rivers Study. Allied Printing, Winnipeg, Man.

Lynch, M. 1979. Predation, competition, and zooplankton community structure: An experimental study. Limnol. Oceanogr. 24:253-272.

Newbury, R. W., and G. K. McCullough. 1984. Shoreline erosion and restabilization in the Southern Indian Lake reservoir. Can. J. Fish. Aquat. Sci. 41:558-566.

Newbury, R. W., G. K. McCullough, and R. E. Hecky. 1984. The Southern Indian Lake impoundment and Churchill River diversion. Can. J. Fish. Aquat. Sci. 41:548-557.

O'Brien, W. J. 1979. The predator-prey interaction of planktivorous fish and zooplankton. Am. Sci. 67:572-581.

Patalas, K., and A. Salki. 1984. Effects of impoundment and diversion on the crustacean plankton of Southern Indian Lake. Can. J. Fish. Aquat. Sci. 41:613-637.

Planas, D., and R. E. Hecky. 1984. Comparison of phosphorus turnover times in northern Manitoba reservoirs with lakes of the Experimental Lakes Area. Can. J. Fish. Aquat. Sci. 41:605-612.

Potter, L., D. Kidd, and D. Standiford. 1975. Mercury levels in Lake Powell: Bioamplification of mercury in man-made desert reservoir. Environ. Sci. Tech. 9:41-46.

Richards, R. C., C. R. Goldman, T. C. Frantz, and R. Wickwire. 1975. Where have all the Daphnia gone? The decline of a major cladoceran in Lake Tahoe, California-Nevada. Verh. Int.

Verein. Limnol. 19:835-842.

Vollenweider, R. A. 1975. Input-output models with special reference to the phosphorus loading concept in limnology. Schweiz. Z. Hydrol. 37:53-84.

Wagner, M. W. 1984. Postimpoundment change in financial performance of the Southern Indian Lake commercial fishery. Can. J. Fish. Aquat. Sci. 41:715-719.

Wells, L. 1970. Effects of alewife predation on zooplankton populations in Lake Michigan. Limnol.

Oceanogr. 14:556-565.

Zaret, T. M. 1972. Predators, invisible prey, and the nature of polymorphism in the Cladocera (Class Crustacea). Limnol. Oceanogr. 17:171-184.

Zaret, T. M. 1975. Strategies for existence of zooplankton prey in homogeneous environments.

Verh. Int. Verein. Limnol. 19: 1484-1489.

Zaret, T. M. 1980. Predation and freshwater communities. Yale University Press, New Haven, Conn.

Zaret, T. M., and W. C. Kerfoot. 1975. Fish predation on Bosmina longirostris: Bodysize selection versus visibility selection. Ecology 56:232-237.

Zaret, T. M., and J. S. Suffern. 1976. Vertical migration in zooplankton as a predator avoidance mechanism. Limnol. Oceanogr. 21:804-813.

Committee Comment
Like the Lake Washington work ( Chapter 20 ), the investigation of Southern Indian Lake called on scientists to use nearly all the sources of ecological knowledge identified in this case study. Freshwater Institute scientists were explicit at the outset about using impoundment and diversion as an experiment in reservoir processes. They coordinated and focused scientific and technical talent on a scale rarely achieved in such ventures. Productivity of fish and plankton communities was a dominant theme of the biological investigations. The observational program was designed to study spatial differences in the dendritic basin and to identify regions that were influenced, either positively or negatively, by the physical manipulation. The specific incorporation of genetic markers into the study of whitefish populations (Bodaly et al., 1984), for instance, made it possible to document the redistribution and migration of stocks.

Productivity and biomass were the measures of choice for all trophic levels, because they reduce biological entities and rates to a common currency.

For a study that encompasses an entire ecosystem, such a common means of presentation of data is necessary. Nonetheless, events in Southern Indian Lake are traceable to the life histories and physiological and behavioral characteristics of individual interacting populations. Increased turbidity and lowered lake temperatures triggered biological responses in ways not explainable strictly by thermodynamics or mass flux alone. Temperature affects the rates of metabolic and developmental processes, and inorganic particles in suspension attenuate light that would otherwise be available for photosynthesis.

Turbidity, moreover, obscures the vision of predators that hunt by sight and reduces the risk to their preferred prey. The experiment in Southern Indian Lake apparently introduced manipulations at both the base and the apex of the food web. Primary producers were affected by the altered light and by changes in nutrient
RAISING THE LEVEL OF A SUBARCTIC LAKE 328
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. input. Whereas selection might have operated for species adept at nutrient acquisition before impoundment, the postimpoundment community clearly was governed by light limitation (Hecky and Guildford, 1984). One would expect changes in species composition, and the phytoplankton community showed such changes (H. Kling, personal communication, cited by Patalas and Salki, 1984).

The increase in large-bodied crustaceans suggests that predatory pressure on the species had relaxed. Deeper, more turbid water provided a refuge from the fish. The predatory behavior of these large zooplankters, in turn, could have contributed to the decline in cladocerans and small copepods reported by Patalas and Salki (1984). Effects of the manipulation thus became focused on the small herbivores, beset on the one side by abundant carnivorous zooplankton and on the other by reduced abundance of phytoplankton owing to the new light regime. The cladocerans declined by 75% and small cyclopoid copepods by 50% (Patalas and Salki, 1984).

In contrast with the alterations of the plankton communities, little or no change occurred among the profundal macrobenthos (Wiens and Rosenberg, 1984). The benthic organisms seemed to be influenced principally by the input of organic material from the Churchill River or from shoreline erosion and thus to be divorced from biological interactions in the turbid waters. The scientific papers include no discussion of meroplanktonic taxa like Chaoborus, which can enter the plankton each night and are known to be influenced by planktivorous fish (e.g., Northcote et al., 1978). If fish predation were substantially relaxed, these species would likely be affected.

The professional publications produced as a result of the Southern Indian Lake study are valuable contributions to current knowledge about reservoir processes. The environmental manipulation was an experiment that had never been tried previously. Most of the results could be guessed only qualitatively, if at all. From a strictly scientific point of view, the project was exploited very profitably, in that new principles were uncovered and present paradigms have been enriched. But the knowledge was gained at a cost: the native peoples of the Southern Indian Lake region who suffered the loss of livelihood and threats to the quality of their food did not share in the scientific adventure. Hecky and colleagues seemed to regret most their inability to forecast the fisheries problems— the eventual compensation program was somewhat arbitrary and inequitable. Had they predicted the decline in fishery quality and the hazards associated with mercury, the forecast need not have halted the experiment, but might have led to a well-planned agreement for adequate and just compensation. Such a plan could very easily have entered the cost-benefit analysis of the project.

References
Bodaly, R.A., T. W. D. Johnson, R. J. P. Fudge, and J. W. Clayton. 1984. Collapse of the lake whitefish (Coregonus clupeaformis) fishery in Southern Indian Lake, Manitoba, following lake impoundment and river diversion. Can. J. Fish. Aquat. Sci. 41:692-700.

Hecky, R. E., and S. J. Guildford. 1984. Primary productivity of Southern Indian Lake before, during, and after impoundment and Churchill River Diversion. Can. J. Fish. Aquat. Sci. 41:591-604.

Northcote, T. G., C. J. Walters, and J. M. B. Hume. 1978. Initial impacts of experimental fish introductions on the macrozooplankton of small oligotrophic lakes. Proc. Int. Assoc.

Theor. Appl. Limnol. 20:2003-2012.

Patalas, K., and A. Salki. 1984. Effects of impoundment and diversion on the crustacean plankton of Southern Indian Lake. Can. J. Fish. Aquat. Sci. 41:613-637.

Wiens, A. P., and D. M. Rosenberg. 1984. Effect of impoundment and river diversion on profundal macrobenthos of Southern Indian Lake, Manitoba. Can. J. Fish. Aquat. Sci. 41:638-648.

RAISING THE LEVEL OF A SUBARCTIC LAKE 330
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. 22 Ecological Effects of Nuclear Radiation
Particular kinds of environmental perturbation are essentially replicated in many places. Because no two sites are identical, detailed prediction of effects requires knowledge of the ecosystem in question. Much can be learned, however, by carrying out generic studies designed to discover results of general applicability to many conditions. Studies supported by the U.S. Atomic Energy Commission to determine the effects of radiation on living organisms and how radionuclides move through natural environments have been the most extensive attempts to use a generic approach to obtain information required for making major policy decisions. This case study summarizes and analyzes these studies and their contributions both to the solution of problems at which they were directed and to ecological theory generally.

Case Study
CARL F. JORDAN, Institute of Ecology, University of Georgia, Athens, Georgia

INTRODUCTION
Ionizing radiation resulting from production of radionuclides by bombs and reactors was the first pollutant given major national and international attention. It became an environmental concern soon after the first test of a nuclear weapon, which occurred on July 16, 1945, at Trinity, New Mexico.

Starting almost immediately after the test and continuing for years thereafter, field surveys were conducted at Trinity to discover the extent and degree of environmental contamination by radionuclides, persistence of radionuclides, and effects of radiation on organisms and ecosystems (Larson, 1963). The studies showed that it would be extremely important to understand the effect of this pollutant on organisms and how it moves through the environment. It was thought that dispersion of radioisotopes in the environment as a result of fallout, reactor development, waste disposal, nuclear war, and technological projects could pose serious environmental problems (Wolfe, 1963).

As nuclear energy was developed, for both peaceful and military uses, programs were established to evaluate the environmental effects of human- produced radiation. Many of the programs were at laboratories that became parts of the complex supported by the U.S. Atomic Energy Commission (AEC), such as those at Argonne, Illinois; Brookhaven, New York; Hanford, Washington; Idaho Engineering Laboratory; Los Alamos, New Mexico; Livermore, California; the Nevada test site; Oak Ridge, Tennessee; and Savannah River, South Carolina (Whicker and Schultz, 1982). Other programs were established at universities or in conjunction with state agencies.

The effects of ionizing radiation and radionuclide movement in the environment were also studied in other countries. A series of symposia sponsored by the International Atomic Energy Agency dealt with the uses of radionuclides in various disciplines, such as hydrology and plant nutrition, as well as with environmental contamination by radioactive materials. The series of studies sponsored by the Environmental Sciences Branch of the Division of Biology and Medicine of AEC (later the Energy Research and Development Administration and now the Department of Energy) probably constituted the greatest concentrated effort ever expended to understand the environmental impact of a pollutant. (Lists of
ECOLOGICAL EFFECTS OF NUCLEAR RADIATION 332
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. proceedings, books, and bibliographies of primary references were compiled by Klement and Schultz in 1980 and Whicker and Schultz in 1982.) The studies of environmental effects of nuclear radiation covered most aspects of ecology. Some dealt with life histories to determine at which stage in its life cycle an organism was most sensitive to ionizing radiation. Studies of population dynamics and population interactions were crucial in understanding the dynamics of radionuclides in food chains. Studies sponsored by AEC were carried out in various habitats to determine the effects of radiation on community structure and community pattern. Determining changes in nutrient cycling and productivity of ecosystems also was a major goal of many studies.

This case study illustrates a generic approach to evaluating environmental pollutants. Ecological theory often cannot be used to make accurate predictions about individual cases. Valuable predictions for specific cases often are based on local field experience, not on formal theory. Furthermore, predictions are usually difficult to apply beyond the bounds of a specific case. In a generic approach, the effects of a pollutant on a large number of different organisms in different environments are studied, thereby providing a framework for predicting effects on new organisms or environments.

THE ENVIRONMENTAL PROBLEMS
Ionizing radiation is radiation with sufficient energy for its interactions with matter to produce an ejected electron and a positively charged ion (Whicker and Schultz, 1982). In large numbers, such interactions in the cells of living organisms can cause genetic and physiological damage and death. Low levels of ionizing radiation from cosmic rays and radionuclides in the earth's crust have always been present. Life has evolved in an environment of low background radiation.

Some of the first environmental studies concerned the radioactivity discharged in the mid-1940s from reactors at Hanford, Washington, into the Columbia River (Whicker and Schultz, 1982). Others concerned the magnitude and duration of radioactivity at weapon test sites (Hines, 1962; Koranda, 1965, 1969; Larson, 1963). Studies at nuclear test sites were important, not only because of the environmental dangers at the sites themselves, but also because results could be used to predict conditions at sites affected by nuclear war.

An important early stimulus for studies of radionuclides in the environment was observation of the fate of radioactive fallout from atmospheric tests of nuclear weapons. One particularly disturbing case involved the ECOLOGICAL EFFECTS OF NUCLEAR RADIATION 333
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. movement of cesium-137 in the lichen-reindeer-human food chain in tundra ecosystems (Liden and Gustafsson, 1967; Nevstrueva et al., 1967). Cesium-137, a relatively strong emitter of gamma radiation, has a relatively long half-life (30 years) and metabolic effects similar to those of potassium. It is adsorbed on the surface of lichens, which are abundant in the tundra. Lichens are an important food of reindeer, and the nuclide became concentrated in tissue of reindeer and of the Finnish Lapps and Alaskan eskimos, who depend heavily on reindeer for meat (Hanson et al., 1967; Lindell and Magi, 1967; Miettinen and Hasanen, 1967). Another potentially hazardous combination of nuclide and food chain that was identified early involved iodine131, which, if deposited on pasture grass, quickly moves through dairy cattle to milk (Barth and Seal, 1967; Bergstrom, 1967) and becomes concentrated in the thyroid (Turner and Jennrich, 1967). Although there was never any conclusive evidence of damage to humans, the potential for danger was recognized.

Observation of potential hazards of fallout gave rise to systematic studies of radionuclide concentrations in several species, such as deer (Schultz and Longhurst, 1963), and to studies of the environmental factors important in radionuclide accumulation (Davis et al., 1963). An important result of these analyses (see Auerbach, 1965, for review) was a series of international symposia (Whicker and Schultz, 1982) in which the problem of radioactive contamination and accumulation in food chains was highlighted and brought to international attention.

In 1957, AEC established the Plowshare Program to investigate and develop peaceful uses for nuclear explosions (Auerbach, 1971a; Kelly, 1966).

Studies were carried out to obtain food-chain and transport data needed for calculating radiation doses to human populations and to assess the impact on the local environment. One of the first was designed to predict the environmental impact of the use of nuclear explosives to excavate a harbor in the Cape Thompson region of Alaska (Wolfe, 1966). Another focused on the feasibility of using thermonuclear devices to create a new transisthmian canal in Panama (Atlantic-Pacific Interoceanic Canal Study Commission, 1970; Martin, 1969).

Although the studies did not conclusively predict damage to human health as a result of using nuclear explosives in these regions, neither of the proposed excavations ever took place (for reasons never made public).

As nuclear technology advanced to the point where nuclear energy could be used to generate electricity, studies began to address the ecological problems in siting nuclear power plants, particularly the problem of radioactive discharge into the environment, both accidental and as a result of normal operations (Auerbach, 1971b; Schultz and Whicker, 1980).

Underground detonations were also used experimentally to stimulate gas flow in geological formations of low permeability (Alldredge et al., 1976).

APPROACHES TO THE PROBLEM OF RADIOACTIVITY IN

THE ENVIRONMENT
After the early observations of environmental radioactivity caused by nuclear testing, experimental studies were begun to evaluate the problem. These studies had two basic aspects. One was the movement of radionuclides in the environment after accidental or deliberate release from a nuclear device or power plant (Comar, 1965); radioactive tracers were often used to determine the pathway of each potentially dangerous nuclide and the rate of movement along that pathway. The second aspect was the effect of ionizing radiation on organisms in the environment; various animal populations (French, 1965; Turner, 1975), plant communities (Whicker and Fraley, 1974), aquatic organisms (Blaylock and Trabalka, 1978), and other ecosystem components (Platt, 1965) were irradiated experimentally.

Movement of Radionuclides in the Environment
Before the advent of radioecology, studies of food chains and of whole ecosystems had scarcely been initiated. An important contribution of the studies of radionuclide dynamics was to show that species in ecosystems were connected with each other and how particular species depended on the flows and cycles of nutrients and energy among all the other species in the ecosystem.

Evidence of feedback in ecosystems also emerged from those studies; e.g., they showed that the rate of return of a nuclide to an individual or species can depend on other species and on environmental factors.

Perhaps the most important idea used in the efforts to understand radionuclide movement in food chains was the specific-activity concept discussed by Kaye and Nelson (1968). Specific activity is defined as the ratio of radioactive atoms to total atoms of the same element. By using the stable- element distribution in environmental samples as a chemical analog for the radionuclide, we can predict the dispersal of radionuclides through environmental pathways, if we know the stable-element chemistry of organisms constituting the links in the biological pathways of food chains and the ratio of the radionuclide to its stable element at the source of entry of the radionuclide into the food chain (Reichle et al., 1970).

A modification of the specific-activity approach uses the ratio of the ECOLOGICAL EFFECTS OF NUCLEAR RADIATION 335
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. potentially dangerous radionuclides to analogous nutrient elements. Several radionuclides produced in important quantities during nuclear reactions are chemically similar to nutrient elements important in animal diets. For example, strontium-90 is similar to calcium and is accumulated in bones, and cesium-137 and potassium are metabolized similarly. Known pathways and concentrations of calcium and potassium in food chains can be used to predict concentrations of the analogous radionuclides, after corrections are made for discrimination or concentration factors (Comar and Lengemann, 1967). Discrimination and concentration are affected by the atomic weight of the isotope, as well as feeding habits and other foodchain characteristics (Whicker and Schultz, 1982).

The need to make predictions about the fates and concentrations of radionuclides in ecosystems where experimental tests were not feasible gave rise to the development of systems analysis techniques in ecology (Kaye and Ball, 1969; Shugart and O'Neill, 1979). These techniques used models of the flux and turnover of radionuclides in ecosystems. When data were not available, assumptions were based on studies in other ecosystems, on known metabolism of stable-element analogs in the species of interest, or on other appropriate physical, chemical, and biological models. Once all important ecosystem turnovers and transfers were formulated, equations were solved simultaneously to predict radionuclide dynamics through an entire ecosystem.

One of the efforts to predict radionuclide movement in an ecosystem was an analysis of the fate of radioactivity, if thermonuclear devices were used to excavate a new canal across Central America (Kaye and Ball, 1969). The predictive model was based on stable-element data collected at the proposed site and at similar sites and on results of laboratory analyses. Because the canal was never excavated, the model was never tested. But Jordan et al. (1973) used a similar approach to predict the environmental residence half-time of strontium-90 after its release into a tropical rain forest. They then used atmospheric fallout measured before the atmospheric-test ban in 1963 as model input and predicted concentrations of the nuclide in the forest through the end of the century. Measurements in 1974 (Jordan and Kline, 1976) showed that actual concentrations in the forest were higher than those predicted, because of atmospheric tests after 1963. The strontium-90 study showed that the environmental half-time of the isotope in this system was about 20 years and that loss was predominantly by physical decay.

Predictions of movement of radionuclides in food chains also were based on laboratory data on the biological turnover of radionuclides in organisms and on the factors — such as intake, assimilation, metabolism, and excretion — that affect turnover (Reichle et al., 1970). Other tracer studies ECOLOGICAL EFFECTS OF NUCLEAR RADIATION 336
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. showed the rate of movement of radionuclides between living and nonliving portions of ecosystems (Auerbach, 1965).

An important early study that showed how changes in ecosystem structure affected radionuclide dynamics used a series of microcosms in combination with mathematical modeling (Patten and Witkamp, 1967). Leaf litter tagged with cesium-134 was introduced into microecosystems composed of different combinations of soil, microflora, millipedes, and aqueous leachate. Field studies were much more difficult, because of the health hazard of radioactivity.

However, the section of Oak Ridge National Laboratory that is now the Environmental Sciences Division carried out a study in which a whole stand of trees was inoculated with cesium-137 in 1962 and the movement of the nuclide in the ecosystem was followed for a number of years thereafter (Francis and Tamura, 1971). An unexpected finding of that study was that much of the cesium did not move up through the leaves, but rather moved down into the roots and then into the soil when roots died and decomposed.

Radiation Effects
In addition to predicting the rate of nuclide movement through food chains and the amounts of radioactivity reaching valued species, it was necessary to know what effect a given amount of radioactivity would have on the valued species.

The behavior of individuals of a species is obviously important in the effect of radiation release. For example, burrowing animals are shielded from radiation (Buchsbaum, 1958). Other factors that influence radiation sensitivity in complex ways are body size, temperature, rate of reproduction, and life span.

It was initially predicted that the most important biological factor affecting sensitivity to radiation would be the volume of chromosomes, but the first tests suggested that interphase chromosomal volume was a better predictor of both species sensitivity (Sparrow et al., 1968) and pattern of community response (Woodwell and Whittaker, 1968). Another important hypothesis regarding radiation sensitivity (Henshaw, 1963) was that there is a threshold of radiation tolerance, which may be different for each species. For radiation exposure below this threshold, damage does not exceed that caused by natural background radiation.

Studies of the effects of ionizing radiation on ecosystems were carried out in an oak-pine forest at Brookhaven, New York (Woodwell and Rebuck, 1967), a tropical rain forest (Odum et al., 1970), a northern hardwood forest (Murphy et al., 1977), southern pine-hardwood forests in Georgia (Cotter and McGinnis, 1965) and in Tennessee (Witherspoon, 1965), a pine forest (McCormick, 1969), and a shortgrass prairie (Fraley
ECOLOGICAL EFFECTS OF NUCLEAR RADIATION 337
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. and Whicker, 1971). There were many other studies on the effects on populations and species (Appendix). In these studies, a shielded source of radiation was used, so that scientists could enter the irradiated areas. At the completion of an experiment, the source was removed. Because the source of radioactivity was thereby contained, there was no residual contamination, which was a problem in tracer studies.

Results showed that radiation sensitivity of plants was correlated to some extent with interphase chromosomal volume, but there were frequent exceptions (Koo and deIrizarry, 1970; Woodwell and Whittaker, 1968). A much more useful generalization from the radiation studies is that sensitivity of plants depends on the ratio of photosynthetic tissue to total tissue (Woodwell, 1967, 1970). The most sensitive plants are trees, which have a relatively low ratio of photosynthetic mass (leaves) to nonphotosynthetic mass (stem and root). Shrubs are less sensitive than trees, and herbs and grasses are less sensitive than shrubs.

Plants like algae, in which much of the tissue is photosynthetic, are highly resistant. Among trees, pine trees — which produce long-lived leaves — are more sensitive than deciduous hardwoods, in which replacement of leaves represents a smaller drain on energy reserves. Rhizomatous species, such as sedges, a large proportion of whose biomass is shielded by the soil, usually are relatively resistant. A generalization that applied to both plants and animals is that radiation sensitivity is correlated with size: the largest species are usually the most sensitive (Woodwell, 1967), as they are to stress in general (Woodwell, 1970).

Two irradiation experiments contrasted the effects of long exposure (Woodwell, 1967) and short exposure (Odum, 1970). In the site exposed for a short time, sprouting of trees played an important part in ecosystem recovery (Jordan, 1969). In a site chronically irradiated, root carbohydrate reserves were exhausted and sprouting was not important. Disturbed areas at that site were colonized by forbs and grasses with seeds that are widely dispersed and that germinate rapidly (Woodwell, 1967).

CONCLUSION
The AEC-sponsored studies of radiation in the environment resulted in two major conclusions. First, some but not all radionuclides released into the environment are concentrated as they are passed through food chains, and, if concentration factors are high, relatively low releases of radioactivity can pose a danger. Because every ecosystem and every food chain is different, potential danger depends in part on characteristics of the particular ecosystem and food chain and in part on the radionuclide. Concentration of radionuclides in food chains is a generic characteristic
ECOLOGICAL EFFECTS OF NUCLEAR RADIATION 338
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. with the potential to occur in any ecosystem; for purposes of environmental safety, it must be predicted specifically for the ecosystem of interest.

Second, although radiosensitivity is sometimes correlated with interphase chromosomal volume, a more practical index of radiation sensitivity is simply the size of an organism and its life span. Large organisms are almost always more sensitive to radiation than small organisms. Long-lived species usually suffer more from radiation exposure than short-lived species.

In evaluating the effect of a particular radionuclide in a particular environment, both the food-chain accumulation factors and the sensitivity of the organisms in the food chain to the predicted dose must be known. These factors are now known for some organisms in many types of ecosystem, but must be evaluated in light of the specific conditions at particular sites.

REFERENCES
Alldredge, A. W., F. W. Whicker, and W. C. Hanson. 1976. Some environmental impacts associated with project Rio Blanco. Pp. 65-73 in C. E. Cushing, ed. Radioecology and Energy Resources. Proceedings of the Fourth National Symposium on Radioecology. Dowden, Hutchinson, and Ross, Stroudsburg, Pa.

Atlantic-Pacific Interoceanic Canal Study Commission. 1970. Interoceanic Canal Studies; Annexes 1-5; Vols. 1-6. Atlantic-Pacific Interoceanic Canal Study Commission, Washington, D.C.

Auerbach, S. I. 1965. Radionuclide cycling: Current status and future needs. Health Phys. 11:1355-1361.

Auerbach, S. I. 1971a. Contributions of radioecology to AEC mission programs. Pp. 3-8 in D. J.

Nelson, ed. Radionuclides in Ecosystems. Proceedings of the Third National Symposium on Radioecology. U.S. Atomic Energy Commission, Washington, D.C.

Auerbach, S. I. 1971b. Ecological considerations in siting nuclear power plants: The long-term biotic effects problem. Nucl. Safety 12:25-34.

Barth, D. S., and M. S. Seal. 1967. Radioiodine transport through the ecosystem, airforage-cow- milk using a synthetic dry aerosol. Pp. 151-158 in B. Aberg and F. P. Hungate, eds.

Radioecological Concentration Processes. Proceedings of an International Symposium.

Pergamon Press, Oxford, Eng.

Bergstrom, S. O. W. 1967. Transport of fallout
131I into milk. Pp. 159-174 in B. Aberg and F. P.

Hungate, eds. Radioecological Concentration Processes. Proceedings of an International Symposium. Pergamon Press, Oxford, Eng.

Blaylock, B. G., and J. R. Trabalka. 1978. Evaluating the effects of ionizing radiation on aquatic organisms. Adv. Radiat. Biol. 7:103-152.

Buchsbaum, R. 1958. Species response to radiation: Radioecology, Pp. 124-141 in W. D. Claus, ed.

Radiation Biology and Medicine. Addison Wesley, Reading, Mass.

Comar, C. L. 1965. Movement of fallout radionuclides through the biosphere and man. Annu. Rev.

Nucl. Sci. 15:175-206.

Comar, C. L., and F. W. Lengemann. 1967. General principles of the distribution and movement of artificial fallout through biosphere to man. Pp. 1-18 in B. Aberg and F.

P. Hungate, eds. Radioecological Concentration Processes. Proceedings of an International Symposium. Pergamon Press, Oxford, Eng.

Cotter, D. J., and J. T. McGinnis. 1965. Recovery of hardwood stands 3-5 years following acute irradiation. Health Phys. 11:1663-1673.

Davis, J. J., W. C. Hanson, and D. G. Watson. 1963. Some effects of environmental factors upon accumulation of worldwide fallout in natural populations. Pp. 35-38 in V. Schultz and A.

W. Klement, eds. Radioecology. Proceedings of the First National Symposium on Radioecology. Reinhold, New York.

Fraley, L., and F. W. Whicker. 1971. Response of a native shortgrass plant stand to ionizing radiation. Pp. 999-1006 in D. J. Nelson, ed. Radionuclides in Ecosystems. Proceedings of the Third National Symposium on Radioecology. U.S. Atomic Energy Commission, Washington, D.C.

Francis, C. W., and T. Tamura. 1971. Cesium-137 soil inventory of a tagged Liriodendron forest, 1962 and 1969. Pp. 140-149 in D. J. Nelson, ed. Radionuclides in Ecosystems.

Proceedings of the Third National Symposium on Radioecology. U.S. Atomic Energy Commission, Washington, D.C.

French, N. R. 1965. Radiation and animal populations: Problems, progress, and projections. Health Phys. 11:1157-1568.

Hanson, W. C., D. G. Watson, and R. W. Perkins. 1967. Concentration and retention of radionuclides in Alaskan Arctic ecosystems. Pp. 233-245 in B. Aberg and F. P. Hungate, eds. Radioecological Concentration Processes. Proceedings of an International Symposium. Pergamon Press, Oxford, Eng.

Henshaw, P. S. 1963. Radiation effects and peaceful uses of atomic energy in the animal sciences:

Radiation and biologic capability. Pp. 13-17 in V. Schultz and A. W. Klement, eds.

Radioecology. Proceedings of the First National Symposium. Reinhold, New York.

Hines, N. O. 1962. Proving Ground: An Account of the Radiobiological Studies in the Pacific, 1946-1961. University of Washington Press, Seattle.

Jordan, C. F. 1969. Recovery of a tropical rain forest after gamma irradiation. Pp. 88-98 in D. J.

Nelson and F. C. Evans, eds. Symposium on Radioecology. Proceedings of the Second National Symposium. CONF 670-503. U.S. Department of Commerce, Springfield, Va.

Jordan, C. F., and J. R. Kline. 1976. Strontium-90 in a tropical rain forest: 12th-year validation of a 32-year prediction. Health Phys. 30:199-201.

Jordan, C. F., J. R. Kline, and D. S. Sasser. 1973. A simple model of strontium and manganese dynamics in a tropical rain forest. Health Phys. 24:477-489.

Kaye, S. V., and S. J. Ball. 1969. Systems analysis of a coupled compartment model for radionuclide transfer in a tropical environment. Pp. 731-739 in D. J. Nelson and F. C.

Evans, eds. Symposium on Radioecology. Proceedings of the Second National Symposium. CONF 670-503. U.S. Department of Commerce, Springfield, Va.

Kaye, S. V., and D. J. Nelson. 1968. Analysis of specific-activity concept as related to environmental concentration of radionuclides. Nucl. Safety 9:53-58 .

Kelly, J. S. 1966. Foreword. Pp. iii-iv in N. J. Wilimovsky and J. N. Wolfe, eds. Environment of the Cape Thompson Region, Alaska. U.S. Atomic Energy Commission, Washington, D.C.

Klement, A. W., and V. Schultz. 1980. Terrestrial and Freshwater Radioecology. A Selected Bibliography. Dowden, Hutchinson and Ross, Stroudsburg, Pa.

Koo, F. K. S., and E. R. delrizarry. 1970. Nuclear volume and radiosensitivity of plant species at El Verde. Pp. G-15— G-20 in H. T. Odum and R. F. Pigeon, eds. A Tropical Rain Forest. U.S.

Atomic Energy Commission, Washington, D.C.

Koranda, J. J. 1965. Preliminary studies of the persistence of tritium and 14C in the Pacific Proving Ground. Health Phys. 11:1445-1457.

Koranda, J. J. 1969. Residual tritium at Sedan Crater. Pp. 696-708 in D. J. Nelson and F. C. Evans, eds. Symposium on Radioecology. Proceedings of the Second National Symposium.

CONF 670-503 . U.S. Department of Commerce, Springfield, Va.

Larson, K. H. 1963. Continental close-in fallout: Its history, measurement, and characteristics. Pp. 19-25 in V. Schultz and A. W. Klement, eds. Radioecology. Proceedings of the First National Symposium. Reinhold, New York.

Liden, K., and M. Gustafsson. 1967. Relationships and seasonal variation of 137Cs in lichen,
reindeer, and man in northern Sweden 1961-1965. Pp. 193-208 in B. Aberg and F. P.

Hungate, eds. Radioecological Concentration Processes. Proceedings of an International Symposium. Pergamon Press, Oxford, Eng.

Lindell, B., and A. Magi. 1967. Observed levels of
137Cs in Swedish reindeer meat. Pp. 217-219 in
B. Aberg and F. P. Hungate, eds. Radioecological Concentration Processes. Proceedings of an International Symposium. Pergamon Press, Oxford, Eng.

Martin, W. E. 1969. Radioecology and the feasibility of nuclear canal excavation. Pp. 9-22 in D. J.

Nelson and F. C. Evans, eds. Symposium on Radioecology. Proceedings of the Second National Symposium. CONF 670-503. U.S. Department of Commerce, Springfield, Va.

McCormick, J. F. 1969. Effects of ionizing radiation on a pine forest. Pp. 78-87in D. J. Nelson and F. C. Evans, eds. Symposium on Radioecology. Proceedings of the Second National Symposium. CONF 670-503. U.S. Department of Commerce, Springfield, Va.

Miettinen, J. K., and E. Hasanen. 1967.
137Cs in Finnish Lapps and other Finns in 1962-6. Pp.
221-231 in B. Aberg and F. P. Hungate, eds. Radioecological Concentration Processes.

Proceedings of an International Symposium. Pergamon Press, Oxford, Eng.

Murphy, P. G., R. R. Sharitz, and A. J. Murphy. 1977. Response of a forest ecotone to ionizing radiation. Pp. 43-48 in J. Zavitkovski, ed. The Enterprise, Wisconsin, Radiation Forest.

USERDA TID-26113-p2. U.S. Energy Research and Development Administration, Washington, D.C.

Nevstrueva, M. A., P. V. Ramzaev, A. A. Moiseer, M. S. Ibatullin, and L. A. Teplykh. 1967. The nature of 137Cs and 90Sr transport over the lichen-reindeer-man food chain. Pp. 209-215 in B. Aberg and F. P. Hungate, eds. Radioecological Concentration Processes. Proceedings of an International Symposium. Pergamon Press, Oxford, Eng.

Odum, H. T. 1970. Summary. An emerging view of the ecological system at El Verde. Pp. 1-191-1-281 in H. T. Odum and R. F. Pigeon, eds. A Tropical Rain Forest. U.S. Atomic Energy Commission, Washington, D.C.

Odum, H. T., P. Murphy, G. Drewry, F. McCormick, C. Schinan, E. Morales, and J. A. McIntyre. 1970. Effects of gamma radiation on the forest at El Verde. Pp. D-3 — D-75 in H. T. Odum and R. F. Pigeon, eds. A Tropical Rain Forest. U.S. Atomic Energy Commission, Washington, D.C.

Parzyck, D. C., J. P. Witherspoon, and J. E. Till. 1976. Validation of environmental transport models in the CUEX methodology. Pp. 194-198 in C. E. Cushing, ed. Radioecology and Energy Resources. Proceedings of the Fourth National Symposium on Radioecology.

Dowden, Hutchinson, and Ross, Stroudsburg, Pa.

Patten, B. C., and M. Witkamp. 1967. Systems analysis of 134cesium kinetics in terrestrial
microcosms. Ecology 48:813-824.

Platt, R. B. 1965. Radiation effects on plant populations and communities: Research status and potential . Health Phys. 11:1601-1606.

Reichle, D. E., P. B. Dunaway, and D. J. Nelson. 1970. Turnover and concentration of radionuclides in food chains. Nucl. Safety 11:43-55.

Schultz, V., and W. M. Longhurst. 1963. Accumulation of strontium-90 in yearling Columbian black-tailed deer, 1950-1960. Pp. 73-76 in V. Schultz and A. W. Clement, eds.

Radioecology. Proceedings of the First National Symposium. Reinhold, New York.

Schultz, V., and F. W. Whicker. 1980. Nuclear fuel cycle, ionizing radiation, and effects on biota of the natural environment. CRC Crit. Rev. Environ. Control 10:225-268.

Shugart, H. H., and R. V. O'Neill. 1979. Introduction. Pp. 1-6 in H. H. Shugart and R. V. O'Neill, eds. Systems Ecology. Benchmark Papers in Ecology. 9. Dowden, Hutchinson, and Ross, Stroudsburg, Pa.

Sparrow, A. H., A. F. Rogers, and S. S. Schwemmer. 1968. Radiosensitivity studies with woody plants. I. Radiat. Bot. 8:149-186.

Turner, F. B. 1975. Effects of continuous irradiation on animal populations. Adv. Radiat. Biol. 5:83-144.

Turner, F. B., and R. I. Jennrich. 1967. The concentration of 131I in the thyroids of herbivores and a
theoretical consideration of the expected frequency distribution of thyroidal 131I in a large consumer population. Pp. 175-182 in B. Aberg and F. P. Hungate, eds. Radioecological Concentration Processes. Proceedings of an International Symposium. Pergamon Press, Oxford, Eng.

Whicker, F. W., and L. Fraley. 1974. Effects of ionizing radiation on terrestrial plant communities.

Adv. Radiat. Biol. 4:317-366 .

Whicker, F. W., and V. Schultz. 1982. Radioecology: Nuclear Energy and the Environment. Vols. I and II . CRC Press, Boca Raton, Fla.

Witherspoon, J. P. 1965. Radiation damage to forest surrounding an unshielded fast reactor. Health Phys. 11:1637-1642 .

Wolfe, J. N. 1963. Impact of atomic energy on the environment and environmental science. Pp. 1-2 in V. Schultz and A. W. Klement, eds. Radioecology. Proceedings of the First National Symposium. Reinhold, New York.

Wolfe, J. N. 1966. Committee on environmental studies for Project Chariot, Plowshare Program. Pp. ix-x in Environment of the Cape Thompson Region, Alaska. U.S. Atomic Energy Commission, Washington, D.C.

Woodwell, G. M. 1967. Radiation and the pattern of nature. Science 156:461-470 .

Woodwell, G. M. 1970. Effects of pollution on the structure and physiology of ecosystems. Science 168:429-433.

Woodwell, G. M., and A. L. Rebuck. 1967. Effects of chronic gamma radiation on the structure and diversity of an oak-pine forest. Ecol. Monogr. 37:53-69 .

Woodwell, G. M., and R. H. Whittaker. 1968. Effects of chronic gamma irradiation on plant communities. Q. Rev. Biol. 43:42-55.

APPENDIX: SOME SOURCES OF INFORMATION ON

RADIOECOLOGY
Aberg, B., and F. P. Hungate, eds. 1967. Radioecological Concentration Processes. Proceedings of an International Symposium. Pergamon Press, Oxford, Eng. Cushing, C. E., ed. 1976. Radioecology and Energy Resources.

Proceedings of the Fourth National Symposium on Radioecology . Dowden, Hutchinson, and Ross, Stroudsburg, Pa. Hanson, W. C., ed. 1980. Transuranic elements in the environment. U.S. DOE Rep. DOE/ TIC-22800. U.S.

Department of Energy, Washington, D.C. Hungate, F. P., ed. 1965. Hanford Symposium on Radiation and Terrestrial Ecosystems. Health Phys. 11:1255-1675.

Klement, A. W., and V. Schultz. 1980. Freshwater and Terrestrial Radioecology: A Selected Bibliography. Dowden, Hutchinson, and Ross, Stroudsburg, Pa. Nelson, D. J., ed. 1971. Radionuclides in Ecosystems.

Proceedings of the Third National Symposium on Radioecology. U.S. Atomic Energy Commission, Washington, D.C. Nelson, D. J., and F. C. Evans, eds. 1969. Symposium on Radioecology. Proceedings of the Second National Symposium. CONF 670-503. U.S. Department of Commerce, Springfield, Va.

Odum, H. T., and R. F. Pigeon, eds. 1970. A Tropical Rain Forest. A Study of Irradiation and Ecology at El Verde, Puerto Rico. U.S. Atomic Energy Commission, Washington. D.C. Schultz, V., and A. W. Klement, eds. 1963.

Radioecology. Proceedings of the First National Symposium on Radioecology.

Reinhold, New York. Thompson, R. C., and W. J. Blair, eds. 1972. Hanford Symposium on the Biological Implications of the Transuranium Elements.

Health Phys. 22:533-957 . Whicker, F. W., and V. Schultz. 1982. Radioecology:

Nuclear Energy and the Environment. Vols. I and II . CRC Press, Boca Raton, Fla.

Committee Comment
Environmental problem-solving is hindered by differences in insight derived from general and specific approaches. General ecological theory usually makes only crude predictions about specific conditions or impacts at a specific site. Useful predictions for a specific case often are based on local field experience, rather than on formal theory, and it is difficult to determine their applicability beyond the bounds of the specific case.

The AEC studies used both generic and specific approaches. Studies were carried out on a wide variety of ecosystems in an effort to form generalizations about radioactivity in the environment. The relative sensitivity of plants as a function of the ratio of photosynthetic tissue to total tissue (photosynthesis:respiration ratio) is an example of a generalization that emerged from these comparative studies. Studies on conditions at a specific site or with a specific nuclide were useful in predicting effects of the same nuclide under similar conditions. For example, the Arctic studies suggested that scavenging might be important wherever lichens were dominant members of the community.

This case study suggests the value of approaching many types of environmental perturbations both generically and specifically. The generic approach predicts what, in general, to expect from a perturbation, regardless of where it occurs; the specific approach addresses the question of whether the specific case differs in any important way from the general case.

The many studies in radioecology have resulted in one of the most successful evaluations of the impact of an environmental hazard. One ECOLOGICAL EFFECTS OF NUCLEAR RADIATION 343
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. reason is that they were adequately funded. Most environmental evaluation must be carried out with less support than was available for the radioecological work. The results of these studies show that thorough understanding can be achieved if enough time and money are allocated.

Another notable characteristic of the AEC studies was the separation of studies of radionuclide movement in the environment from studies of the effects of ionizing radiation. Had these two aspects of the radiation problem not been separated, much less progress would have been made. Experiments in which exposure to radiation was great enough to reveal dose-response relationships would have been impossible with radionuclides released into the environment.

Conversely, laboratory studies cannot reveal how radionuclides behave in nature.

Studies sponsored by AEC, particularly those of nuclide movement through the environment, made an important contribution to the emergence of ''ecosystem ecology" (Odum, 1965; Odum and Golley, 1963), in which a major focus is the flow of energy and elements through a unit of landscape. This work supplements and extends studies oriented toward the ecology of individuals, populations, species, and communities.

This examination of the history of AEC-sponsored environmental radiation studies suggests that their contribution to ecological knowledge has been as important as, or perhaps even more important than, the contribution of ecological knowledge to the design and interpretation of the AEC studies.

During the four decades of radioecological studies, the constant interplay between experimental results and general theory has proved fruitful to the basic science of ecology and the applied field of radioecology.

References
Odum, E. P. 1965. Feedback between radiation ecology and general ecology. Health Phys. 11:1257-1262.

Odum, E. P., and F. B. Golley. 1963. Radioactive tracers as an aid to the measurement of energy flow at the population level in nature. Pp. 403-410 in V. Schultz and A. W. Klement, eds.

Radioecology. Proceedings of the First National Symposium. Reinhold, New York.

ECOLOGICAL EFFECTS OF NUCLEAR RADIATION 344
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. 23 Ecological Effects of Forest Clearcutting
Management of the harvesting of renewable biological resources requires knowledge of the stocks of those resources and of the rates at which they recover after harvesting. Gaining this knowledge is especially difficult when the intervals between cropping are long and rates of recovery are low, as they are in the case of timber harvesting. Not only are trees long-lived, but differences between sites make it difficult to extrapolate results from one location to another. This case study reviews the ways in which these problems have been approached and assesses the current state of prediction of the amounts and significance of nutrient losses from forests after clearcutting.

Case Study
CARL F. JORDAN, Institute of Ecology, University of Georgia, Athens, Georgia

INTRODUCTION
A major goal of the forestry profession is to sustain the production of wood in forests. In the early days of the profession, much attention centered on prevention of fire and outbreaks of insect pests as means of sustaining high productivity. These factors still cause extensive losses of trees, but now that many forests are managed as agricultural crops, other problems have increased in importance. One of the common intensive management techniques in modern forestry is clearcut logging. Clearcut logging is often more economical than other forest management techniques, but it has the potential to create serious environmental problems out of effects that would be negligible if less intensive techniques were used. Examples include the elimination of habitat of endangered species and increases in soil erosion. This case study reviews research that has evaluated the impact of forest clearcutting on nutrient stocks and site productivity.

The first studies of nutrient loss due to clearcutting were concerned primarily with only parts of the nutrient budget of an ecosystem, for example, leaching losses. Improved understanding of ecosystem functioning made it apparent that useful interpretation of nutrient losses due to clearcutting required an analysis of both the nutrient stocks in the ecosystem and the nutrient inputs and losses. To evaluate the effect of nutrient loss during clearcutting on future productivity of a site, it is necessary to know the quantities not only of nutrient losses due to leaching and erosion after clearcutting, but also of losses due to tree removal, volatilization, and fixation in the soil. Also important are the rate of nutrient replacement through atmospheric input, rock weathering, and fertilization and the size of nutrient stocks in the ecosystem that can buffer short- term fluctuations in input and output.

A nutrient-budget approach to forest management is analogous to budget management in business. Successful business management demands an understanding of the entire operating budget, including revenues, expenditures, and financial balance. This need in business management is so obvious as to be taken for granted. Yet, in forest management, it has become clear only recently that successful management for sustained yield requires knowledge of nutrient inputs, nutrient losses, and nutrient stocks.

These quantities are often difficult to measure, but complete budgets must be constructed, if the impact of clearcutting on nutrient loss is to be evaluated adequately. This case study examines how such understanding evolved.

BACKGROUND
We do not know when the first agriculturalist observed that ash, litter, and animal carcasses improved crop growth. Undoubtedly, the knowledge that the addition of particular materials to the soil increases productivity arose independently in many regions. However, only in the nineteenth century, after chemical elements had been identified, did it become understood that nutrients constituted the common factor in soil amendments that maintained agricultural productivity (Brady, 1974).

The first scientist to make a systematic study of nutrient circulation in forests in relation to growth of trees was Ebermayer. The objective of his study, "Complete Treatise of Forest Litter," was to explain the adverse effect on forest quality of the litter removal then common in middle European forests (Tamm, 1979). The litter was used in cow stables, and some of the plant nutrients in it eventually reached arable fields in dung, thus contributing to food production at a time when commercial fertilizer was not available. However, forest growth decreased, particularly in already poor sites, and Ebermayer attributed this to the export of plant nutrients (Tamm, 1979).

Interest in the role of nutrients in forest growth continued throughout the early part of the twentieth century, but only within the last several decades have scientists begun to quantify total forest-ecosystem nutrient budgets and changes in budgets due to various logging practices. One reason for interest in nutrient loss resulting from the clearcut method of forest harvesting is the obvious increase in soil erosion, nutrient leaching, and consequent stream eutrophication after such operations (Tamm et al., 1974). Another, which is examined here, is the effect of nutrient loss on the ability of the soil to supply enough nutrients to produce another stand of trees.

Observation of soil erosion and increased eutrophication of drainage streams after clearcutting left little doubt that nutrient loss was occurring, but did not indicate whether it was great enough to affect productivity of the site.

Because tree growth responds to fertilization in at least some cases, nutrient loss is potentially harmful. But nutrients lost through leaching and erosion might be quickly replaced by weathering and other processes. No long-term studies have been designed specifically to test the effects of nutrient leaching on site productivity. Studies at Hubbard Brook,
ECOLOGICAL EFFECTS OF FOREST CLEARCUTTING 347
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

New Hampshire (Bormann and Likens, 1979), showed that a northern hardwood forest that had been clearcut and then treated with herbicide experienced substantial amounts of nutrient leaching, but recovery of the forest through natural succession did not appear to be noticeably inhibited by lack of nutrients.

Clearcutting has rarely resulted in nutrient loss so serious that trees could not grow on the denuded site. Occasionally, on very steep slopes in regions of heavy rainfall, such as some areas in western British Columbia, clearcutting has been followed by soil erosion severe enough to expose bedrock, and sites have consequently been permanently deforested (J. P. Kimmins, personal communication). In most cases, however, at least some tree growth has occurred after clearcutting, so the environmental question is not whether trees can grow after clearcutting, but how fast they can grow. The rate is important to industries or agencies concerned with timber as a crop, because trees that grow rapidly yield a greater profit than trees that grow slowly. Rate of forest recovery also is important for nonmarket values of forests, such as the reduction of soil erosion on watersheds, scrubbing of polluted air, support of fish and game, and provision of habitat for some rare species (Farnworth et al., 1981).

APPROACHES TO EVALUATING LOSSES
Leaching and Erosion
Nutrient leaching losses often increase as a result of clearcutting. After removal of trees, evapotranspiration on the site decreases, so the amount of water percolating through the soil increases; the result is an increase in leaching potential. A decrease in nutrient recycling is also important in increasing leaching. In undisturbed forests, root uptake of nutrients and their incorporation in biomass reduces nutrient loss. Nutrients from decomposing litter and soil organic matter on a clearcut are not taken up by trees, but are exchanged on the clay surfaces of mineral soil, where they are susceptible to loss through leaching and erosion. These ecological ideas are commonly accepted and are discussed in many texts on ecology, agronomy, soil conservation, and forestry.

However, emphasis often is on the effect of the nutrients on lake and stream eutrophication and fish productivity, rather than on future productive capacity of the site losing the nutrients.

Recent studies have revealed some of the mechanisms by which nutrients are leached. After clearcutting, the activity of nitrifying bacteria in the soil increases (Likens et al., 1969; Vitousek et al., 1979), owing to increased temperatures, decreased competition for ammonium from tree ECOLOGICAL EFFECTS OF FOREST CLEARCUTTING 348
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. roots, decreased allelopathic inhibition of nitrifying bacteria, and other factors (Reiners, 1981; Swift et al., 1979). As a result, ammonium from the mineralization of organic matter is oxidized to nitrate, and nitrate anions and nutrient cations that have been exchanged for hydrogen ions on soil surfaces are rapidly leached.

An important effect of forest disturbance, especially in mountainous terrain, is soil erosion. Soil erosion results from exposure of mineral soil to the direct impact of raindrops. The impact breaks soil aggregates, causing pores and channels in the soil to be filled, the soil surface to become less permeable, and the surficial runoff to be greater than when soil surfaces are covered with litter.

Although soil erosion often accompanies clearcut logging, recent studies have shown that it is construction of access roads, not the clearcutting itself, that causes most of the erosion (Douglass and Swift, 1977); removal of logs by cable appears to lead to less erosion.

A general approach to determining nutrient loss during and after clearcutting of forests and during site preparation for a new stand has been to measure leaching and erosional losses from clearcut areas directly and to compare losses with those in undisturbed areas that serve as a control.

Measurements have usually been made where drainage from a watershed flows over a weir resting on bedrock, so that subsurface drainage is insignificant.

Continuous monitoring of water flow and nutrient content permits measurement of total nutrient loss. Clearcut watersheds are compared with control watersheds or watersheds subjected to other treatments, such as conversion to grassland. In some cases where discrete watersheds were not available, losses have been studied by means of lysimeters (soil water collectors). Losses have also been measured by comparing soil and ecosystem nutrient stocks before and after cutting or between cut and control plots.

Many studies have shown increases in rate of nutrient loss during forest disturbance. For example, the studies of the watersheds at Coweeta Hydrologic Laboratory in North Carolina showed higher rates of nitrate and sediment loss in recently cut catchments than in control catchments (Monk 1975; Swank and Douglass, 1975, 1977; Webster and Patten, 1979). Studies of nutrient dynamics in conifer forests of the Pacific Northwest also showed an increase in loss rates after clearcut harvesting and slash burning (Feller and Kimmins, 1984; Gessel and Cole, 1965; Miller and Newton, 1983).

An important study of nutrient loss after clearcutting was carried out in northern hardwoods at the Hubbard Brook watershed site in New Hampshire (Bormann and Likens, 1970; Bormann et al., 1968; Likens et al., 1970).

Bormann et al. (1968) concluded that
ECOLOGICAL EFFECTS OF FOREST CLEARCUTTING 349
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. clear-cutting tends to deplete the nutrients of a forest ecosystem by (i) reducing transpiration and so increasing the amount of water passing through the system; (ii) simultaneously reducing root surfaces able to remove nutrients from the leaching waters; (iii) removal of nutrients in forest products; (iv) adding to the organic substrate available for immediate mineralization; and (v) in some instances, producing a microclimate more favorable to rapid mineralization.

This conclusion was controversial, because clearcutting at the Hubbard Brook site had been followed by herbicide treatment to prevent regrowth of vegetation. Aubertin and Patric (1974) wrote that "there is a substantial difference between the Hubbard Brook treatment and conventional clearcutting.

Conventional clearcutting also features complete forest cutting; but all salable wood is harvested and rapid forest regeneration is encouraged." In an experimental hardwood watershed in West Virginia, Aubertin and Patric measured negligible nutrient losses after clearcutting that was carried out to resemble conventional clearcut logging techniques. Other clearcutting studies in the Hubbard Brook region showed that nutrient losses are less if herbicides are not used after clearcutting, but can still be important because the soils in the region are shallow (Pierce et al., 1972); harvesting removes a larger proportion of the total nutrient pool there than in areas with deeper soils.

Because of the importance of nutrient uptake by vegetation, alternating contour strip cuts with undisturbed forest on a mountainside should decrease nutrient leaching. Experiments with strip cuts showed that this technique can reduce nutrient leaching in northern forests (Hornbeck et al., 1975). The effect should be even greater in the humid tropics, where the potential for leaching is extremely high (Jordan, 1982). Once vegetation becomes re-established in the clearcut strips, the uncut strips can be harvested.

Biomass Removal
Nutrient losses due to leaching are brief, and often small, compared with those due to tree removal (Cole and Bigger, undated; Hornbeck and Kropelin, 1982; Kimmins, 1977; Sollins and McCorison, 1981; Swank and Waide, 1980).

As evidence accumulated that nutrient losses due to biomass removal during clearcut operations were often much greater than losses due to leaching and erosion, scientists shifted their attention to studies of nutrient stocks in forests and soils and to losses from ecosystems due to biomass harvest. Results from various sites and management strategies were presented in several major symposia (Ballard and Gessel, 1983; Leaf, 1979). Generalizing about the studies is difficult, because each site had its own combination of soil type, soil nutrient stocks, biomass nutrient
ECOLOGICAL EFFECTS OF FOREST CLEARCUTTING 350
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. stocks, management technique, rates of nutrient input via rainfall, nitrogen fixation, fertilization, and nutrient losses through leaching and denitrification.

Generalizations valid for sites of a specific type in a given region do not apply to sites of other types. Each nutrient also appears to behave differently. Only by measuring the stocks and dynamics of all the critical nutrients at each site can one estimate accurately the impact of nutrient removal on productivity of the site. Even then, estimation might be insufficient, because trees can exhibit uptake beyond immediate needs, and this can be misleading in the prediction of requirements.

One general finding is that the extent of nutrient loss depends on which parts of the trees are removed from the site. Leaves have higher concentrations of nutrients than do stems, so stripping logs of leaves before the logs are removed from the site decreases nutrient losses. For evergreens, branches with leaves are cut off before the tree bole is removed. In the case of deciduous species, an alternative is to harvest during the winter. However, because some nutrients are withdrawn from leaves into stems before the leaves are shed, nutrient losses would be smaller if trees were cut while in full leaf and mechanically defoliated. Harvesting of whole trees, including roots, removes larger proportions of nutrients.

Nutrients Remaining in Soil
Simply measuring the nutrients lost because of clearcutting and site preparation does not indicate the effect of the nutrient loss in site productivity.

Estimates of the quantities of nutrients remaining in the site and of the rates at which they are replenished are also needed. Scientists attempting to predict the effect of nutrient loss on site productivity must determine nutrient stocks remaining in the soil and the fractions of those stocks available to growing trees (Johnson et al., 1982).

Nutrients are held in the soil in various ways (Brady, 1974). Some cations, such as potassium, can be part of the lattice structure of minerals, where they are relatively unavailable to roots; or they can be exchanged on clay surfaces, where they are readily available. Phosphorus can exist in relatively soluble forms, but, in the presence of low pH and high aluminum concentrations, it becomes bound in compounds that are not readily taken up by plants. Soil can contain large amounts of nitrogen; if the nitrogen is bound in organic matter that is resistant to decomposition, however, nitrogen shortages might be critical.

Thus, knowing the total stocks of nutrients in the soil is not sufficient.

For many agricultural crops, reagent-soluble nutrients in soil have been correlated with crop growth, and solubility of nutrients has been used as an index of the availability of nutrients to crop plants. However, such an ECOLOGICAL EFFECTS OF FOREST CLEARCUTTING 351
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. index usually is not applicable to tree growth. Many tree species are adapted for extracting nutrients from soil when "availability," as measured by solubility, is low. Some of these adaptations are large root biomass, mycorrhizal symbiosis, slow root uptake, and long life, which enable trees to survive during periods of nutrient shortage (Chapin, 1980). The effects of these variables are not easy to measure.

Nutrient Replenishment
Site productivity depends on relative rates of nutrient replenishment and nutrient loss. The atmosphere contains a stock of nutrients adsorbed on the surface of aerosols, such as dust and pollen, that either settle out of the atmosphere gradually as "dry fall" or are washed out by precipitation. Nutrient input from the atmosphere can be substantial. Nitrogen can be contributed to a site by nitrogen-fixing species, which can be present because they occur naturally or through planting of species symbiotic with nitrogen-fixing bacteria.

Nutrients can also enter an ecosystem through the weathering of subsoil or parent rock.

Nutrients lost through clearcutting can be replaced by fertilizers, but fertilization of forests is often economically infeasible. For most agricultural crops, fertilization is profitable, because annual sales are high and the time between investment in fertilizers and return of investment in crop sales is short.

The economic benefits of fertilizing forests are more difficult to calculate, because the value of harvestable products that accumulate each year is low, because the period between investment and harvest is long, and because forests have other values besides their use for wood.

Budgets and Sensitivity Analysis
Nutrient budgets of ecosystems have proved extremely difficult to measure precisely, because of high variability in both space and time. In addition, such important quantities as the proportions of nutrients available to plants and the rate of mineral weathering are often hard to measure; errors in their measurement could result in large errors in estimates of total ecosystem nutrient budgets. Therefore, these estimates are often viewed skeptically by both scientists and managers. The key question, however, is whether the error is important in relation to the environmental or ecological problem being addressed.

As an example of a simple sensitivity analysis to assess the effect of biomass removal on remaining nutrient stocks and the effect of error on the predictions, consider the following hypothetical case. Assume that the standard deviation around the amount of calcium in the soil is relatively ECOLOGICAL EFFECTS OF FOREST CLEARCUTTING 352
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. large and that the average value of a set of samples could be different from the true average value by 30%. Is this error important? If calcium were present in the aboveground biomass at 500 kg/ha, all of which would be removed by clearcutting, and present in the soil at 5,000 kg/ha, an error of 30% in the estimate of calcium in the soil would not change the conclusion that clearcutting will not have an important effect on total calcium stock. If calcium were present at 500 kg/ha in the biomass and 100 kg/ha in the soil, an error of 30% would not change the conclusion that clearcutting would have a very important effect on total calcium stock. However, if calcium were present at 500 kg/ha in both the biomass and the soil, it would be important to be able to estimate more accurately the true value of calcium in the soil, because a 30% error could lead to very different management conclusions.

This relatively simple type of analysis has yielded predictions of nutrient depletion due to competences removal in several regions. For example, calcium depletion due to clearcutting of ridge forests in the southern Appalachians might result in reduced growth of the next crop (West and Mann, 1983), and nitrogen is often a limiting nutrient in the conifer forests of the Pacific Northwest (Peterson and Gessel, 1983).

Simulation models of forest growth that incorporate the effects of nutrients can assist in answering questions about the required degree of accuracy of field measurements and experiments. Only recently have models with increased sophistication been developed. A model like the nitrogen model of Swank and Waide (1980), which predicts yield under various management strategies on the basis of nitrogen dynamics in the treated ecosystems, can be used to assess the effect of an error of a given magnitude on the overall conclusion and can aid in a decision as to whether more accurate measurements are necessary. The latter model permits evaluation of management alternatives and their consequences, such as the effect on forest yield of a change in the length of rotation.

The importance of sensitivity analyses is often not appreciated by ecological and environmental scientists. Perhaps the reason is that sensitivity analyses often contradict what many scientists have been taught, which is that the greatest possible accuracy should always be sought. Sensitivity analyses show us where a large amount of less accurate data might be more valuable than a smaller amount of more accurate data.

CONCLUSION
Studies to evaluate the effect of clearcutting on nutrient depletion and site productivity have progressed a long way. Early studies were concerned with nutrients removed by leaching and erosion after clearcutting. Then ECOLOGICAL EFFECTS OF FOREST CLEARCUTTING 353
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. it was recognized that nutrient stocks remaining in the ecosystem after logging were important, and studies of those stocks were carried out. Nutrient dynamics and their effects on stocks and productivity were included in the analyses.

Recent work has used systems analysis to keep a better account of the multitude of continuously changing factors in the ecosystem.

The proof of the effectiveness of the approach will lie in tests of the predictions generated by the models. From the perspective of applications, the most important prediction is that of wood yield. Because of the time required for such validation, it is too early to appraise the approach. Regardless of how accurate the predictions prove to be, they will be useful to the scientists who formulated the models on which they were based. Discrepancies between predictions and results will point up weaknesses in the models and indicate what studies are necessary to improve the accuracy of the predictions.

REFERENCES
Aubertin, G. M., and J. H. Patric. 1974. Water quality after clearcutting a small watershed in West Virginia. J. Environ. Qual. 3:243-249.

Ballard, R., and S. P. Gessel, eds. 1983. IUFRO Symposium on Forest Site and Continuous Productivity, Seattle, Washington, August 22-28, 1982. PNW-163. U.S. Department of Agriculture Forest Service. Portland, Oreg.

Bormann, F. H., and G. E. Likens. 1970. The nutrient cycles of an ecosystem. Sci. Am. 223 (4):92-101 .

Bormann, F. H., and G. E. Likens. 1979. Pattern and Process in a Forested Ecosystem. Springer, New York.

Bormann, F. H., G. E. Likens, D. W. Fisher, and R. S. Pierce. 1968. Nutrient loss accelerated by clear-cutting of a forest ecosystem. Science 159:882-884 .

Brady, N. C. 1974. The Nature and Properties of Soils. Macmillan, New York.

Chapin, F. S. 1980. The mineral nutrition of wild plants. Annu. Rev. Ecol. Syst. 11:233-260 .

Cole, D. W., and C. M. Bigger. Undated. Effect of Harvesting and Residue Removal on Nutrient Losses and Productivity. Fifth Annual Report. University of Washington College of Forest Resources, Seattle.

Douglass, J. E., and L. W. Swift. 1977. Forest service studies of soil and nutrient losses caused by roads, mechanical site preparation, and prescribed burning in the Southeast. Pp. 489-503 in D. L. Correll, ed. Watershed Research in Eastern North America. A Workshop to Compare Results. Chesapeake Bay Center for Environmental Studies, Edgewater, Maryland.

Smithsonian Institution, Washington, D.C.

Farnworth, E. G., T. T. Tidrick, C. F. Jordan, and W. M. Smathers. 1981. The value of natural ecosystems: An economic and ecological framework. Environ. Conserv. 8:275-282.

Feller, M. C., and J. P. Kimmins. 1984. Effects of clearcutting and slash burning on streamwater chemistry and watershed nutrient budgets in southwestern British Columbia. Water competences. Res. 20:29-40.

Gessel, S. P., and D. W. Cole. 1965. Influence of removal of forest cover on movement of water and associated elements through soil. J. Am. Water Works Assoc. 57:1301-1310 .

Hornbeck, J. W., and W. Kropelin. 1982. Nutrient removal and leaching from a whole-tree harvest of northern hardwoods. J. Environ. Qual. 11:309-316 .

Hornbeck, J. W., G. E. Likens, R. S. Pierce, and F. H., Bormann. 1975. Strip cutting as a means of protecting site and streamflow quality when clearcutting northern hardwoods. Pp. 209-225 in B. Bernier and C. H. Winget, eds. Forest Soils and Forest Land Management.

Proceedings of the Fourth North American Forest Soils Conference. Les Presses de l'Université Laval, Que.

Johnson, D. W., D. C. West, D. E. Todd, and L. K. Mann. 1982. Effects of sawlog vs. whole-tree harvesting on the nitrogen, phosphorus, potassium, and calcium budgets of an upland mixed oak forest. Soil Sci. Soc. Am. J. 46:1304-1309 .

Jordan, C. F. 1982. Amazon rain forests. Am. Sci. 70:394-401.

Kimmins, J. P. 1977. Evaluation of the consequences for future tree productivity of the loss of nutrients in whole-tree harvesting. For. Ecol. Manage. 1:169-183 .

Leaf, A. L., ed. 1979. Impact of Intensive Harvesting on Forest Nutrient Cycling. Proceedings of a Symposium at Syracuse, New York, August 13-16, 1979. Northeast Forest Experiment Station, Broomall, Pa.

Likens, G. E., F. H. Bormann, and N. M. Johnson. 1969. Nitrification: Importance to nutrient losses from a cutover forested ecosystem. Science 163:1205-1206 .

Likens, G. E., F. H. Bormann, N. M. Johnson, D. W. Fisher, and R. S. Pierce. 1970. Effects of forest cutting and herbicide treatment on nutrient budgets in the Hubbard Brook watershed- ecosystem. Ecol. Monogr. 40:23-47 .

Miller, J. H., and M. Newton. 1983. Nutrient loss from disturbed forest watersheds in Oregon's coast range. Agro-Ecosystems 8:158-167.

Monk, C. D. 1975. Nutrient losses in particulate form as weir pond sediments from four unit watersheds in the southern Appalachians. Pp. 862-867 in F. G. Howell, J. B. Gentry, and M. H. Smith, eds. Mineral Cycling in Southeastern Ecosystems. ERDA Symposium Series.

CONF 740-513 . U.S. Energy Research and Development Administration, Washington, D.C.

Peterson, C. E., and S. P. Gessel. 1983. Forest fertilization in the Pacific Northwest: Results of the regional forest nutrition research project. Pp. 365-369in R. Ballard and S. P. Gessel, eds.

IUFRO Symposium on Forest Site and Continuous Productivity, Seattle, Washington, August 22-28 , 1982. PNW-163. U.S. Department of Agriculture Forest Service, Portland, Oreg.

Pierce, R. S., C. W. Martin, C. C. Reeves, G. E. Likens, and F. H. Bormann. 1972. Nutrient loss from clearcuttings in New Hampshire. Pp. 285-295 in Watersheds in Transition. American Water Resources Association Symposium. Colorado State University, Ft. Collins, Colo.

Reiners, W. A. 1981. Nitrogen cycling in relation to ecosystem succession. Pp. 507-528 in F. E.

Clark and T. Rosswall, eds. Terrestrial Nitrogen Cycles. Proceedings of a Workshop.

Ecological Bulletins (Stockholm) 33. Swedish Natural Science Research Council (NFR), Stockholm.

Sollins, P., and F. M. McCorison. 1981. Nitrogen and carbon solution chemistry of an old growth coniferous forest watershed before and after cutting. Water Resour. Res. 17:1409-1418.

Swank, W. T., and J. E. Douglass. 1975. Nutrient flux in undisturbed and manipulated forest ecosystems in the southern Appalachian mountains. Pp. 445-456 in Symposium ECOLOGICAL EFFECTS OF FOREST CLEARCUTTING 355
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. de Tokyo. Publication 117 de l'Association Internationale des Sciences Hydrologiques, Reading, Eng.

Swank, W. T., and J. E. Douglass. 1977. Nutrient budgets for undisturbed and manipulated hardwood forest ecosystems in the mountains of North Carolina. Pp. 343-364 in D. L.

Correll, ed. Watershed Research in Eastern North America. A Workshop to Compare Results, Chesapeake Bay Center for Environmental Studies, Edgewater, Maryland.

Smithsonian Institution, Washington, D.C.

Swank, W. T., and J. B. Waide. 1980. Interpretation of nutrient cycling research in a management context: Evaluating potential effects of alternative management strategies on site productivity. Pp. 137-158 in Forests: Fresh Perspectives from Ecosystem Analysis. Proc. 40th Annu. Biol. Colloq. Oregon State University Press, Corvallis.

Swift, M. J., O. W. Heal, and J. M. Anderson. 1979. Decomposition in Terrestrial Ecosystems.

Studies in Ecology. Vol. 5. University of California Press, Berkeley.

Tamm, C. O. 1979. Nutrient cycling and productivity of forest ecosystems. Pp. 2-21 in A. L. Leaf, program chairman. Impact of Intensive Harvesting on Forest Nutrient Cycling.

Proceedings of a Symposium at Syracuse, New York, August 13-15, 1979. Northeast Forest Experiment Station, Broomall, Pa.

Tamm, C. O., H. Holmen, B. Popovic, and G. Wiklander. 1974. Leaching of plant nutrients from soil as a consequence of forestry operations. Ambio 3:211-221 .

Vitousek, P. M., J. R. Gosz, C. C. Grier, J. M. Melillo, W. A. Reiners, and R. L. Todd. 1979. Nitrate losses from undisturbed ecosystems. Science 204:469-474.

Webster, J. R., and B. C. Patten. 1979. Effects of watershed perturbation on stream potassium and calcium dynamics. Ecol. Monogr. 49:51-72.

West, D. C., and K. L. Mann. 1983. Whole-Tree Harvesting: Fourth Year Progress Report for 1982.

Nutrient Depletion Estimates, Postharvest Impacts on Nutrient Dynamics and Regeneration. Environmental Sciences Division Publication No. 2184. Oak Ridge National Laboratory, Oak Ridge, Tenn.

Committee Comment
As illustrated by this case study, methods of analyzing nutrient budgets of ecosystems have developed slowly as scientists gradually have recognized the need to perform a complete accounting of nutrient stocks and fluxes if the significance of changes in flux rates induced by clearcutting is to be understood.

This represents discovery of methods that were already well known in the business world, and it might be asked whether ecologists would have progressed more rapidly if they had been better versed in budget analyses as practiced in various disciplines. The answer is probably yes, but it is also clear that ecosystems are so different from businesses that those techniques have to be modified for use in studying ecosystem nutrient dynamics. Such features of ecosystems as differences in availability of nutrients due to variation in soil types and in the types of plants growing on them have no close parallels in the business world.

Whereas much progress has been made in measuring total stocks and fluxes of nutrients, site variation is great enough that very few general ECOLOGICAL EFFECTS OF FOREST CLEARCUTTING 356
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. predictions are possible. There is no reason to believe, in principle, that developing a broad predictive ability is impossible, but it seems evident that the necessary data base is large and that its components might not yet all be identified. That is true even if we simply wish to predict growth rates of trees. If we also wish to predict the dynamics of other constituents of an ecosystem, such as herbivores, carnivores, and detritivores — as well as interactions among plant species— even more extensive data are required. Results of such analyses could be surprising. For example, lower availability of nutrients, in combination with physical stress, such as drought, might render trees much more susceptible to attacks by defoliating insects or fungi (Fearside and Rankin, 1985). These attacks might decrease yields and change growth forms of the trees, leaving yields of usable wood products much lower than would be predicted simply on the basis of the availability of nutrients to support tree growth.

An important conceptual advance illustrated by this case study is sensitivity analysis. Sensitivity analysis is especially important when systems consist of many interacting factors that are of uncertain influence or that are difficult to measure accurately. Time and budgetary constraints prevent accurate measurement of all factors of interest, and choices need to be made as to which factors should receive the most attention. Predictions generated by models are unequally sensitive to variation in different parameters. Sensitivity analysis is a powerful method for deciding which factors need to be measured most accurately and which ones require only crude estimation. Sensitivity analysis helps to avoid consequential errors — errors that could lead to inappropriate management decisions. The natural inclination of most scientists to measure all quantities as accurately as possible might actually lead to poorer predictive abilities, given the investment of comparable resources, than an approach designed to obtain accurate measurements only for the quantities identified as critical.

The study of ecosystem nutrient dynamics is made difficult not only by the number of interacting factors, but also by the long duration of the most important processes. Development and testing of models might require decades of work that taxes the patience of investigators, funding agencies, and managers who must make decisions, whether or not sufficient data are available for them to predict the consequences of their decisions. This emphasizes the need for cooperation among all who are involved in developing plans for managing and using forest ecosystems.

Reference
Fearnside, P. M., and J. R. Rankin. 1985. Jari revisited: Changes and the outlook for sustainability in Amazonia's largest silvicultural estate. Interciencia 10:121-129.

ECOLOGICAL EFFECTS OF FOREST CLEARCUTTING 357
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. 24 Environmental Effects of DDT
Dichlorodiphenyltrichloroethane, better known as DDT, was a potent insecticide when first used in the late 1930s. It had no obvious side effects and was active against many insect pests. As a result, it was extremely widely used, both in the United States and elsewhere. But, as has since happened with so many other chemicals, DDT had unforeseen effects. Those effects resulted from its persistence — one of its initially attractive features. Because of this persistence, DDT was transported from its initial site of application by both biotic and abiotic factors until almost no part of the earth's surface was free of it. Animals in the oceans, lakes, and rivers, on land, and in the air had detectable amounts of DDT in their tissues. The resulting concern led to detailed monitoring, to increased understanding of how chemicals are transported through the environment and how quickly resistance to pesticides can evolve, and to a new awareness of how even the most careful scoping of a problem does not guarantee freedom from unpleasant surprises.

Case Study
JOHN BUCKLEY, Whitney Point, New York

INTRODUCTION
DDT (dichlorodiphenyltrichloroethane) was first synthesized in 1874, but its properties as an insecticide were not discovered until 1939. By the early 1940s, the United States was producing large amounts of DDT for control of vectorborne diseases, and its use is credited with saving millions of lives from diseases with insect vectors, such as typhus and malaria. Its use in the United States for agricultural purposes expanded rapidly after World War II, reached a peak of 80 million pounds in 1959, and decreased thereafter until it was terminated by cancellation of its registration in 1972. Approximately 1,350 million pounds of DDT were used in the United States during those 30 years. In addition to domestic use, hundreds of millions of pounds were exported.

Public concern over use of DDT became widespread in 1962, when Rachel Carson's Silent Spring was published. As a result of this concern, the pros and cons of DDT use were considered by the President's Science Advisory Committee, whose report, Use of Pesticides, issued in May 1963, recommended an orderly phasing out of DDT over a short period. As early as 1957, the federal government began restricting its own use of DDT, and, beginning in 1967, many registered uses of DDT were canceled. Finally, on June 14, 1972, after a hearing that generated 9,312 pages of testimony (from 125 experts) and more than 350 documents, the administrator of the Environmental Protection Agency (EPA) announced the cancellation of all remaining uses of DDT on crops. Uses for public health and quarantine purposes were not affected, and export was permitted. After appeals by both sides, the Court of Appeals for the District of Columbia ruled on December 13, 1973, that there was ''substantive evidence" in the record to support the ban of DDT.

The cancellation of DDT was so strongly opposed by some agricultural interests that the 1974 report of the Committee on Appropriations of the House of Representatives (Congressional Record, November 1973, H. 9619) stated:

The [Environmental Protection] Agency was also directed to initiate a complete and thorough review, based on scientific evidence, of the decision banning the use of DDT. This review of DDT must take into consideration all of the costs and benefits and the importance of protecting the Nation's supply of food and fiber.

The resulting review, DDT: A Review of Scientific and Economic Aspects of the Decision to Ban Its Use as a Pesticide, was published by EPA in July 1975. All data reviewed supported the 1972 findings.

ENVIRONMENTAL PROBLEMS
The cancellation of DDT was based on its persistence, transport, biomagnification, and toxic effects and on the absence of benefits of DDT that were not available through less environmentally harmful substances. In the course of research on DDT and its metabolites, many insights into the movement, fate, and effects of pollutants in the environment were developed.

DDT was the first pesticide to which rapid and extensive resistance evolved.

The earliest field studies were of DDT applied at 5 lb/acre for spruce budworm control. These studies concentrated on acute effects, such as deaths of birds, fish, and insects and other invertebrates. The acute LD 50 (the dose lethal
to 50% of the test organisms) in laboratory mammals varied from 60 mg/kg in dogs to 800 mg/kg in rats. Bioassays of fish, however, revealed LC 50 values
(concentrations in water lethal to 50% of the test organisms in a specified number of hours) of a few parts per billion (ppb). Captive birds had acute LC 50s (concentrations in food lethal to 50% of the test organisms) of 400-1,200 ppm (parts per million) when given in the diet for 5 days (Hill et al., 1975). Because of the extreme toxicity of DDT to fish, attempts were made to avoid bodies of water in the course of aerial spraying. Avoiding such areas completely was not possible, because more than 50% of the DDT applied typically drifts outside the target area, often onto water.

DDT had been in widespread use only a short time before resistance developed in some pest species. The first recognized evidence of resistance to DDT was the failure of house fly control with DDT in southern Italy in 1947 after exceptional success in 1946. Shortly after discovery of resistance to DDT, resistance to other insecticides was observed in the field, but development of resistance was not universally recognized as late as 1956 (Brown and Pal, 1971). Standardized laboratory methods for detection of resistance were worked out under the sponsorship of the World Health Organization. By the time of cancellation of DDT in 1972, more than 200 species of invertebrates were resistant to DDT, cyclodiene, or organophosphorus insecticides; some species were resistant to all three. Resistance in vertebrates is less common, but has been reported to occur in some frogs in Mississippi (Boyd et al., 1963).

Resistance has several ramifications in relation to environmental effects.

Indirectly, inability to control resistant insects has led to an increase in ENVIRONMENTAL EFFECTS OF DDT 360
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. entomological research to find alternative methods of control. Generally referred to as IPM (integrated pest management), these methods depend on greater understanding of the life history and ecology of the pest, and they result in much more selective control. Pesticides are used in most applications of IPM, but only sparingly and in combination with other methods, so that there is less contamination of the environment.

In the mid-1950s, massive spraying programs were undertaken to save the American elm by killing the elm bark beetle, which spread the imported Dutch elm disease fungus from tree to tree. Robin mortality was widespread, and robins dying with tremors characteristic of DDT poisoning were commonly seen in treated areas. Dying robins appeared not only at the time of spraying, but also in the following spring before additional spraying. Chemical analyses showed that soils in these areas contained DDT at up to 18 ppm, and earthworms that fed on the leaves in the soils contained DDT at 53-204 ppm.

Dead robins that had fed on the worms had up to 3 mg of DDT in their tissues (Barker 1958). Tests for residues in forest soils revealed that DDT disappeared slowly (Woodwell, 1961). Here was unequivocal evidence of the persistence of DDT and of the movement and concentration of DDT through a terrestrial food chain. Also during the mid-1950s, dichlorodiphenyldichloroethane * (DDD, also
called tetrachlorodiphenylethane, or TDE) was applied to Clear Lake, California, to control gnats. Initial concentrations of 0.02 pp. in water yielded DDD residues of 10 ppm in plankton, 903 ppm in the fat of plankton-eating fish, 2,690 ppm in the fat of predatory fish, and 2,134 ppm in the fat of grebes that fed on fish (Hunt and Bischoff, 1960).

In the late 1950s, the first indication of worldwide spread of DDT residue was observed. It was discovered (Buckley, 1979)
that shark liver contained detectable amounts of DDT and its metabolites. This discovery was so surprising that it was not believed. The initial presumption was that the sample analyzed had been inadvertently contaminated; but a second sample, taken with precautions to avoid contamination, also contained DDT residues. Because DDT was not applied to the oceans, it appeared that the residue in shark liver had to be a naturally occurring substance, but residue analysis of pre-DDT shark liver oil detected no DDT. Finally, it was assumed that the DDT must have been accidentally dumped into the ocean and that the shark liver contamination was local. Residue analysis showed at least small amounts of DDT widely distributed in sharks and other marine fish, thus requiring us to acknowledge long-range movement of DDT and biological magnification of these residues. Laboratory studies confirmed the bioaccumulation.

Later analyses of tissues from penguins, skuas, and seals from the * DDD was used directly as an insecticide, but also formed by metabolism of DDT.

Antarctic and fur seals from the Pribiloff Islands in the Bering Sea confirmed the occurrence of DDT and its metabolites far from any area where it had been applied. DDT was truly ubiquitous (Anas and Wilson, 1970; Sladen et al., 1966).

The degree of bioconcentration of DDT from the environment into living tissues and of residues from prey to predator was so striking as to be almost unbelievable (U.S. EPA, 1975):

Experimental data showed that DDT can be biologically concentrated by a variety of aquatic organisms at all trophic levels. Phytoplankton, the dominant oceanic vegetation and primary food source for marine animals, concentrates DDT from seawater into its cell membranes. Water-fleas (Daphnia), a food source for many freshwater fish species, accumulated 9.0 ppm in tissues after three days exposure to 80 pptr (parts per trillion). This represents a bioconcentration factor of 112,500 times the exposure level. Rainbow trout exposed to 1.0 ppm DDT (wet weight) in food and 10 pptr in water for 84 days contained 2.3 ppm as whole body residues. Exposure to food alone resulted in residues of 1.8 ppm (a concentration factor of 1.8 X) and exposure to water alone yielded residues of 0.72 ppm (a concentration factor of 72,000 X). In fish fed 1 mg/kg DDT/day, 73% of the DDT residues were present 90 days after the fish were transferred to clean food.

DDT is not equally toxic to all species of animals. Control programs for arthropods have sometimes resulted in increased pest damage, because predators of a pest were eliminated by DDT. For example, DDT applications in apple orchards eliminated populations of predaceous ladybird beetles, so that red-mite populations formerly controlled by the ladybird beetles reached outbreak proportions. This particular mite is not susceptible to DDT and was hardly influenced directly by the DDT that killed the beetles (Helle, 1965). In other cases, less susceptible populations have flourished after control programs, presumably because of reduction in competition as a result of elimination or reduction of more susceptible species. Cope (1961) reported that, after treatment of 72,000 acres of the Yellowstone River watershed, the total numbers of invertebrates had recovered within a year, but the species composition was still altered. Plecopterans and ephemopterans were reduced, but trichopterans and dipterans occurred at higher numbers at the end of the year. In some cases, selective toxicity has resulted in reductions of species that are normally used as food by valued predators. Ide (1967), in studies of the effects of DDT applied at 0.5 lb/acre in the forested watershed of the Mirimachi River in New Brunswick, observed that fewer insect species emerged in streams affected by DDT, and that those most severely reduced were the large ones, such as caddis flies, on which salmon mainly feed. Recovery of stream fauna required up to 4 years.

Studies carried out at the Patuxent Wildlife Research Center demonstrated that stored residues of DDT could be mobilized during weight loss, and this resulted in mortality after exposure to DDT had stopped (Van Velzen et al., 1972). These studies provide a basis for understanding the cause of some of the delayed mortality observed in the field.

During the 1960s, more subtle effects of DDT and its metabolites were discovered. The striking eggshell thinning in some predatory birds — such as bald eagles, ospreys, peregrine falcons, and pelicans — was clearly the cause of decreases in their populations. Eggshells of the affected species collected before 1945 were found to be noticeably thicker than those collected in the late 1940s and later (Anderson and Hickey, 1972). It was hypothesized that this eggshell thinning resulted in decreased reproductive success, and field studies of the brown pelican supported the hypothesis (Risebrough et al., 1971). Laboratory studies confirmed that thinning of shells occurred in mallards (Davison and Sell, 1974), American kestrels (Peakall et al., 1973), and ring doves (Haegele and Hudson, 1973) when dichlorodiphenyldichloroethylene (DDE) was incorporated in the diet of these birds. Since the cancellation of DDT in 1972, there has been a marked decrease of DDT residues in brown pelicans and improved reproduction in those pelicans on the Pacific, Gulf, and Atlantic coasts and in ospreys in the eastern United States (U.S. EPA, 1975).

DDT alters the behavior of some fish, and some of the changes are clearly detrimental to the welfare of the species. Atlantic salmon parr, for example, select water temperatures below those to which they are acclimated when exposed to DDT at 10 ppb or less and temperatures higher than acclimation temperature when exposed to DDT at 10-100 ppb. Furthermore, exposed salmon are less active than control fish. Rainbow trout select higher than acclimation temperatures and exhibit high-temperature shock followed by death (Javaid, 1972a).

KEY ISSUES
By 1970, it was evident that DDT residues were everywhere and that reducing the residues would require huge reductions in the use of DDT.

Contamination of fish used for human food beyond the contamination limit approved by the Food and Drug Administration (FDA) occurred widely, especially in freshwater fish. Residues also appeared in many meats and dairy products. Surveys of human fat revealed virtually universal contamination by DDT, and residues in human milk gave rise to concern over the desirability of breastfeeding. That DDT residues appeared far from where DDT had been applied and lasted for years or even decades suggested that most uses would have to stop; control of application sites
ENVIRONMENTAL EFFECTS OF DDT 363
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. and care in application would not be enough to reduce environmental residues substantially.

The low cost of DDT, its relatively low toxicity to mammals (including humans), and its persistence made its continued use highly desirable for public health vector control programs. Although concern for its toxicity is greater now than when DDT was first used in the United States (FDA workers are said to have attested to its harmlessness by putting a spoonful in their coffee), there is still no clinical or epidemiological evidence of damage to man from approved uses of DDT, despite its demonstrated tumorigenicity in mice.

ECOLOGICAL APPROACHES TO THE ENVIRONMENTAL

PROBLEMS ASSOCIATED WITH DDT
Ecologists first became concerned about ecological effects of DDT when it began to be used for control of forest insects, especially when applied from aircraft over hundreds of thousands of acres. The earliest studies consisted of field observations of treated areas. Mortality at the time of treatment or shortly thereafter was the criterion of the effect of treatment. These first field observations revealed some mortality of birds and fish in streams in treated areas and massive mortality of stream insects. For example, it was noted in one study that cutthroat and brook trout stomachs contained 99% crayfish after spraying, compared with none before, presumably because of the almost complete elimination of their insect food (Adams et al., 1949). An experimental area was treated annually for 4 years with DDT, and breeding bird populations were studied. Three species decreased substantially, but no changes were detected in 23 others (Robbins et al., 1951).

Studies during the early 1950s, principally in the laboratory, continued to concentrate on acute effects of DDT. In the late 1950s, the Dutch elm disease control program resulted in insights into food-chain transfer and accumulation of toxic materials in terrestrial environments. Studies of DDT (and other chlorinated organic pesticides) thereafter regularly looked for phenomena associated with food-chain accumulations. The general belief among toxicologists had been that metabolism of toxic substances invariably decreased their toxicity, but it was discovered that some metabolites of DDT, especially DDE, were more persistent and in some cases more active biologically than the original compound. This discovery laid the groundwork for research that eventually explained the decline of a number of bird species through eggshell thinning.

Pesticide-wildlife studies also led to insights into repopulation by songbirds of areas from which birds were removed. A study by Robbins and ENVIRONMENTAL EFFECTS OF DDT 364
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. colleagues demonstrated that, if half or more of the songbirds were removed from a 40-acre tract of bottomland forest in Maryland during the early part of the nesting season, they were replaced within a few days, as judged by counts of singing males before and after the removal (U.S. FWS, 1963).

By the early 1960s, it had become evident that many freshwater sport and commercial fish were becoming so contaminated with DDT and its metabolites (and with other chlorinated hydrocarbons) that they were not suitable for human food. Concentrations were sometimes high enough to impair fish reproduction.

For example, lake trout from Lake George, New York, produced eggs containing DDT and metabolites at 2.9 ppm or more; these eggs hatched, but the fry died (Burdick et al., 1964).

USES OF ECOLOGICAL KNOWLEDGE AND

UNDERSTANDING
Valued Ecosystem Components
No formal effort was made to identify valued ecosystem components, although it was the effects on valued ecosystem components that eventually resulted in eliminating the use of DDT in the United States. Among these effects were the decrease in songbird populations caused by massive control programs, such as those for spruce budworms in forests and Dutch elm disease in suburban areas; the declines in sport fish, such as lake trout in Lake George, caused by blackfly control programs; the contamination of fish and game desired as human food to a point greater than that permitted in domestic animals used for food; and the decline of avian predators at the top of food chains, particularly aquatic ones.

Importance of the Impacts
Probably of greatest importance were the drastic population decreases over large regions or continents of bird species at the top of their food chains. Some of these decreases — such as those of the osprey, brown pelican, peregrine falcon, and bald eagle— resulted in the extirpation of the species over vast areas of the United States and were grounds for legitimate concern over total extinction of these species.

Boundaries of the Problem
Effects of DDT in the ecosystem were at first perceived to be local and restricted to areas where DDT was applied, but findings from studies of DDT kept expanding the boundaries of the problem. Eventually, it was ENVIRONMENTAL EFFECTS OF DDT 365
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. recognized that DDT residues appeared throughout the world. Studies were conducted, often cooperatively with scientists in other nations, to elucidate the problems wherever they existed, e.g., raptor decline in Europe and residue appearance in the Arctic and Antarctic and in marine environments throughout the Northern Hemisphere. Actions by the United States to resolve the problems, however, were restricted to the control of DDT within the United States. In fact, because DDT manufactured in the United States was superior to others, manufacture for use in international health programs was permitted after the ban on use in the United States took effect. For several years, unfortunately, the manufacturer was less adept at controlling wastes from the manufacturing process than at maintaining purity of the product, so considerable waste DDT and related products continued to be emptied into the Pacific Ocean, with detrimental effects on the brown pelican and other species.

Study Strategy
Because of the nature of the problem, there was no formal study strategy.

Studies simply built on earlier studies. The earliest studies were carried out largely by government scientists, especially fish and wildlife biologists. As results were disseminated, more people became interested in the problem, and more academic scientists began to participate. Extensive, informal cooperation developed among scientists in the United States, Canada, and Great Britain.

Results of field and laboratory studies were exchanged, and these results were often used in shaping the work of other investigators. In addition, more formal cooperation was arranged through such international agencies as the North Atlantic Treaty Organization, the Organisation for Economic Co-operation and Development (OECD), and various bodies of the United Nations.

Monitoring
Measurements of eggshell thickness were made in many areas of North America and Europe by using eggs of predatory birds in museums and in nests.

Thinning of 20% or more was found in many species (see Cooke, 1973, for a review of eggshell thinning). The thinning seems to be caused only by DDE, in as much as experiments with other chlorinated hydrocarbons showed thinning only when the other compounds were at nearly toxic concentrations, and it disappeared as soon as signs of toxicity diminished (Haegele and Tucker, 1974). In the years after cancellation of DDT use in the United States and after elimination of industrial discharge by Montrose Chemical Company's DDT manufacturing plant, residues in many
ENVIRONMENTAL EFFECTS OF DDT 366
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. aquatic species began to decrease, and field studies of predatory birds showed increasing shell thickness and improved reproductive success (U.S. EPA, 1975).

Extensive programs for monitoring pesticide residues, including those of DDT and its metabolites, began in the United States during the 1960s under the sponsorship of the Federal Committee on Pest Control. Results of these programs— in which sampling and analytical methods were standardized and which covered human food and animal feeds, air, freshwater, estuarine areas, soils, and selected members of the biota — were published in the Pesticide Monitoring Journal. Less extensive results were available through a cooperative program sponsored by OECD that provided comparable data from Canada, the United States, and several European countries.

It was in the course of these monitoring studies involving marine species that substances interfering with DDT residue analyses were identified as polychlorinated biphenyls (PCBs). This discovery, by the Swedish chemist Jensen (1966), laid the groundwork for extensive studies of what has turned out to be the most severe nonpesticide chemical contamination problem in the world. Enactment of toxic-substance control legislation in the United States was probably a result of studies on PCBs.

Effects of DDT and its metabolites undoubtedly are aggravated by other hydrocarbon residues. Hardly any wild species on which residue analyses have been completed contain only DDT residues, and laboratory studies have shown that many of the other compounds are at least as toxic as DDT. The interaction of these residues is poorly understood.

SOURCES OF KNOWLEDGE AND UNDERSTANDING
Generally Accepted Ecological Facts
Studies of acute toxicity of DDT to birds and fish were undertaken with bioassay methods devised by pollution control biologists and toxicologists. Data on mammals, at least laboratory rats and mice, were available and, in the absence of specific tests on wild mammals, were used to predict effects on wild mammals. Feeding studies were conducted on pheasants and bobwhite quail, which were available because they were routinely raised for release to the wild.

Later, to answer specific questions, studies were carried out on species not routinely kept in captivity. Such studies usually required prior development of appropriate husbandry. The captive birds studied eventually included bald eagles, kestrels, cowbirds, mallards, pigeons, and others. Similarly, trout and goldfish were exposed, and LC
50s for 24-48 hours were determined. The aquatic
organisms studied included
ENVIRONMENTAL EFFECTS OF DDT 367
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. other species of fish and a number of invertebrates. Relatively elaborate exposure systems for aquatic species were devised to overcome the problems of DDT adsorption on surfaces and loss to the air through codistillation.

On the basis of these laboratory-determined estimates of toxicity, naturalists observed field applications of DDT for control of forest insects. In addition to observations of deaths of birds and aquatic organisms, analytical chemistry was used to measure DDT residues in organisms from treated areas.

General knowledge of food chains, coupled with residue analyses, led to predictions of effects on organisms at higher trophic levels, which were indeed observed.

An iterative process of field observation, chemical analysis, laboratory bioassay, and field experimentation was followed, and a reasonable understanding of the movement, concentration, and effects of DDT began to emerge. The ecological theory of food chains and cycling of nutrients (substituting "toxicants" for "nutrients") was used. In the case of decreasing populations of predatory birds, a hypothesis of pesticide causation emerged. So, too, did a series of hypotheses on sublethal effects operating through behavioral changes— e.g., increased stress— and selection of undesirable temperatures.

Some results of field studies of effects on bird populations were surprising.

In some environments, few dead birds were found, although many should have been, because residue in or on prey insects exceeded the predicted LC 50 for
birds. At the same time, populations of birds did not decrease noticeably. Two kinds of studies were carried out: one to determine disappearance of bird carcasses, the other to determine repopulation. It was hypothesized that only a small fraction of bird carcasses present were found and that, at least in areas of a few to hundreds of acres in relatively homogeneous habitat during the early nesting season, repopulation would take place very rapidly. Studies carried out at Patuxent Wildlife Research Center in Laurel, Maryland, confirmed both hypotheses.

Some 30-90% of the small-bird carcasses spread along a powerline through a wooded area were not found by trained observers 4 days later; some of the carcasses were simply not seen, but many were later determined to have been buried by beetles or carried away by scavengers. Where 44% of the singing males of nine species of songbirds on a 100-acre tract of bottomland forest were removed by mist netting, the removal was estimated at only one- fourth of the actual removal. Banding data showed that an influx of new birds,which took place immediately after the 3-day removal period, made it impossible to measure accurately the number of birds removed (U.S. FWS, 1963).

Theory and General Principles of Ecology
Knowledge of trophic interactions and food webs was an inherent part of the studies of accumulation of DDD at Clear Lake, California, and of the studies of robin mortality after DDT spraying for Dutch elm disease control.

Mortality of bald eagles and other predators was shown to be caused by DDT (and other pesticides) accumulated through the food webs of which these predators were a part. Eggshell thinning also was caused by food-chain accumulations of DDE, which was formed from DDT by organisms lower in the food chain. The importance of the eggshell thinning and consequent reduced reproductive success was understandable only in terms of population dynamics.

The observations of distribution of DDT in field studies and monitoring led Metcalf (1972) to devise microcosms in which radioactively labeled substances could be followed as they were partitioned into different environmental compartments.

The Value of Experiments
Studies in the 1940s involved treatment of experimental areas with amounts of DDT similar to those used in actual forest insect control programs.

It was hoped that the use of control and experimental areas and the study of the areas before and at various periods after treatment would make it possible to detect changes in the nontarget biota, including repopulation if depopulation occurred. These studies were helpful in planning large-scale studies.

When first introduced in the United States, DDT was believed to be harmless to people and all other valued organisms. Therefore, the registration permitted use in concentrations high enough to be effective. Tolerances in food were based on the amount that would be present as a result of the approved uses. In the 1950s, DDT residues were discovered in cows' milk, so applications to animal feed crops were reduced to the point where the residues were not detectable in milk (i.e., to below about 1 ppm at that time).

Expert Judgment
Probably the most outstanding example of the role of expert judgment was that played by trained ornithologists (especially Hickey), who first noted the thinning of eggshells and hypothesized that it was caused by pesticides. In 1965, Hickey had convened an international conference to ENVIRONMENTAL EFFECTS OF DDT 369
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. bring together what was known about decreasing peregrine falcon populations (Hickey, 1969). Hickey suggested that pesticide residues were probably responsible for reduced reproduction, primarily through egg destruction and egg-eating, but there was no mention of shell thinning. In the next several years, the hypothesis of eggshell thinning was formulated. Another pertinent example of expert judgment was provided by Rachel Carson, a marine biologist, who saw and forcefully exposed the effects of widespread pesticide use on the functioning of the natural environment (Carson, 1962).

REFERENCES
Adams, L., M. G. Hanavan, N. W. Hosley, and D. W. Johnston. 1949. The effects on fish, birds, and mammals of DDT used in the control of forest insects in Idaho and Wyoming. J. Wildl.

Manage. 13:245-254.

Anas, R. E., and A. J. Wilson. 1970. Residues in fish, wildlife, and estuaries. Organochlorine pesticides in fur seals. Pestic. Monit. J. 3:198-200.

Anderson, D. W., and J. J. Hickey. 1972. Eggshell changes in certain North American birds. Pp. 514-540 in Proceedings of the 15th International Ornithological Congress, Symposium on Chemical Pollutants. E. J. Brill, Leyden, Neth.

Barker, R. J. 1958. Notes on some ecological effects of DDT sprayed on elms. J. Wildl. Manage. 22:269-274.

Boyd, C. E., B. Vinson, and D. E. Ferguson. 1963. Possible DDT resistance in two species of frogs.

Copeia 1963:426-429.

Brown, A. W. A., and R. Pal. 1971. Insecticide Resistance in Arthropods. World Health Organization, Geneva.

Buckley, J. L. 1979. Nontarget effects of pesticides in the environment. Pp. 73-81 in Pesticides:

Their Contemporary Roles in Agriculture, Health and the Environment. Humana, Clifton, N.J.

Burdick, G. E., E. J. Harris, H. J. Dean, T. M. Walker, J. Skea, and D. Colby. 1964. The accumulation of DDT in lake trout and the effect on reproduction. Trans. Am. Fish. Soc. 93:127-136.

Carson, R. 1962. Silent Spring. Houghton Mifflin, Boston.

Cooke, A. S. 1973. Shell thinning in avian eggs by environmental pollutants. Environ. Pollut. 4:85-152 .

Cope, O. B. 1961. Effects of DDT spraying for spruce budworm on fish in the Yellowstone River System. Trans. Am. Fish. Soc. 90:239-251.

Davison, K. L., and J. L. Sell. 1974. DDT thins shells of eggs from mallard ducks maintained on ad libitum or control-feeding regimens. Arch. Environ. Contam. Toxico. 2:222-232 .

Haegele, M. A., and R. H. Hudson. 1973. DDE effects on reproduction of ring doves. Environ.

Pollut. 4:53-57 .

Haegele, M. A., and R. K. Tucker. 1974. Effects of 15 common environmental pollutants on eggshell thickness in mallards and coturnix. Bull. Environ. Contam. Toxicol. 11:98-102 .

Helle, W. 1965. Resistance in the Acarina: Mites. Advan. Acarol. 2:71-93.

Hickey, J. J., ed. 1969. Peregrine Falcon Populations: Their Biology and Decline. University of Wisconsin Press, Madison.

Hill, F. F., R. G. Health, J. W. Spann, and D. Williams. 1975. Lethal Dietary Toxicities of Environmental Pollutants to Birds. U.S. FWS Special Scientific Report — Wildlife No. 191.

U.S. Fish and Wildlife Service, Washington, D.C.

Hunt, E. C., and A. I. Bischoff. 1960. Inimical effects on wildlife of periodic DDD applications to Clear Lake, Calif. Calif. Fish Game Bull. 46:91-106 .

Ide, F. P. 1967. Effects of forest spraying with DDT on aquatic insects in salmon streams in New Brunswick. J. Can. Fish. Res. Bd. 24:769-805.

Javaid, M. Y. 1972a. Effect of DDT on the locomotor activity of Atlantic salmon, Salmo salar, in a horizontal temperature gradient. Pak. J. Zool. 4:17-26 .

Javaid, M. Y. 1972b. Effect of DDT on temperature selection of some salmonids. Pak. J. Sci. Ind.

Res. 15:171-176 .

Jensen, S. 1966. A new chemical hazard. New Sci. 32:612 .

Metcalf, R. L. 1972. A model ecosystem for the evaluation of pesticide biodegradability and ecological magnification. Outlook Agric. 7:55-59.

Peakall, D. B., J. L. Lincer, R. W. Risebrough, J. B. Pritchard, and W. B. Kinter. 1973. DDE- induced eggshell thinning: Structural and physiological effects in three species. Comp.

Gen. Pharmacol. 4:305-313 .

Risebrough, R. W., F. C. Sibley, and M. N. Kirven. 1971. Reproduction failure of the brown pelican on Anacapa in 1969. Am. Birds 25:8-9 .

Robbins, C. S., P. F. Springer, and C. G. Webster. 1951. Effects of five-year DDT application on breeding bird population. J. Wildl. Manage. 15:213-216 .

Sladen, W. J. L., C. M. Menzie, and W. L. Reichel. 1966. DDT residues in Adelie penguins and a crab-eater seal from Antarctica: Ecological implications. Nature 210:670-671.

U.S. EPA (U.S. Environmental Protection Agency). 1975. DDT: A Review of Scientific and Economic Aspects of the Decision to Ban Its Use as a Pesticide. EPA-540/1-75-022 . U.S.

Environmental Protection Agency, Washington, D.C.

U.S. FWS (U.S. Fish and Wildlife Service). 1963. Pesticide-Wildlife Studies: 1963. A Review of Fish and Wildlife Service Investigations During the Calendar Year 1963. FWS Circular 199. U.S. Fish and Wildlife Service, Washington, D.C.

Van Velzen, A. C., W. B. Stiles, and L. F. Stickel. 1972. Lethal mobilization of DDT by cowbirds.

J. Wildl. Manage. 36:733-739 .

Woodwell, G. M. 1961. The persistence of DDT in a forest soil. For. Sci. 7:194-196 .

Committee Comment
The wide use of DDT after World War II for the control of insects that caused public health problems or were agricultural and forest pests created a global environmental problem. Scientists and regulators were unprepared to deal with this problem. Before 1947, the sole concern of regulatory processes for pesticides was protection of purchasers from fraud. The passage of the Federal Insecticide, Fungicide, and Rodenticide Act in 1947 permitted attention to the possible effects of a pesticide on nontarget organisms other than people and domestic animals; but in practice, if the proponent of a use certified that the product was both effective and safe, registration was granted. The regulatory agencies— the Food and Drug
ENVIRONMENTAL EFFECTS OF DDT 371
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Administration (FDA) and the U.S. Department of Agriculture (USDA — required additional information from the proponent; until they were satisfied, they could withhold registration (USDA) or a tolerance (FDA). However, a proponent could obtain a ''protest registration," shifting the burden of proof from the proponent to the government, which had to disprove the proponent's claim. Furthermore, during this period, few on the staff of the pesticide registration office were knowledgeable about fish and wildlife. Little concern for ecological problems of any kind was evident in the regulatory agencies.

Scientists were little better prepared to deal with the problems posed by DDT. Toxicologists believed that acute toxic effects of the chemicals were the central issues of concern and that the metabolism of toxic substances invariably resulted in decreased toxicity. The possibility that metabolites of DDT might be as toxic as or even more toxic than DDT itself was not considered initially.

Ecological investigations were similarly narrow. At first, ecologists concentrated their attention on field observations of areas that had been directly treated with DDT. These studies revealed much about acute toxicity of DDT to different organisms and rates of recolonization of treated areas. But ecologists were not prepared for the striking phenomena of movement of DDT far from the sites of its application or for the remarkable bioconcentration of DDT that resulted in reproductive failure of top carnivores. These ideas are now generally accepted, but in the late 1940s and 1950s were not. As a result, the first data indicating that DDT was globally distributed and that concentrations were extremely high in the tissues of organisms high on food chains were not believed.

The history of research to determine the results of the massive use of DDT and of how those results were achieved reveals strikingly how the state of science at any time constrains the questions that are likely to be asked and the range of answers that are considered reasonable. Had concepts of the trophic- dynamic organization of ecological communities been developed before 1950, progress in understanding the roots of the problems caused by DDT might have been much more rapid. Lindemann (1942) had published his pioneering paper on trophic-dynamic ecology, but this perspective had yet to become a part of the thinking patterns of ecologists at the end of World War II. Indeed, as this case study demonstrates, efforts to determine the causes of the effects of DDT were a major stimulus to the development of this field of ecology. Persistence of organic chemicals in their original form, or as toxic metabolites, is now routinely measured as an indicator of likelihood of ecological effects.

Bioaccumulation of lipid-soluble chemicals to 104 or 105 times the concentrations present in the physical environment is now expected as a normal part of food-chain
ENVIRONMENTAL EFFECTS OF DDT 372
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. dynamics. Spread of materials far beyond the sites of their application is now considered characteristic of many chemicals in ecological systems.

Toxicology has likewise been influenced by research on the effects of DDT. That metabolites of toxins can be toxic is now commonly accepted.

Efforts are no longer directed exclusively or even primarily to determining lethal doses of toxins. Much more attention is now paid to sublethal effects that might affect behavior, survival, and population dynamics, because investigators recognize that severe decreases in animal populations can occur in nature without massive direct mortality of adults. As a result, although it can be argued that one "DDT" was inevitable, it is equally clear that another "DDT" would be inexcusable.

Striking though these advances are, it is clear that the basic problems encountered in dealing with DDT can express themselves in the future.

Metabolites, transport, and bioaccumulation are now routinely investigated, but we still know almost nothing about the effects of toxicants in mixtures. Nearly all toxicity studies still deal with chemicals one at a time. In nature, however, organisms encounter potentially toxic chemicals in complex mixtures. How these chemicals interact in the environment and in the bodies of animals to produce different and unexpected effects is still generally unknown. As long as this is true, effects that are unexpected and difficult to believe or explain are likely. The general recognition that complex mixtures of toxicants require concerted efforts means that the speed with which scientists and regulators can respond to the surprises of the future should be greater than was the case with DDT. Nonetheless, the major lesson of the DDT story would be lost if it did not heighten our awareness of the current state of ignorance about important processes that affect the responses of individual organisms, populations, and ecological communities to toxic materials introduced into the environment.

The real basis for banning of DDT was its ecological effects, rather than its effects on human health. Even today, the evidence on human health effects of DDT is inconclusive, whereas the data on adverse ecological effects are vast and convincing. Also, the decision to ban DDT, rather than to restrict its use, was based on the conclusion that there was no way to control its movement once it was released into the environment. The decision was made more palatable politically by the rapid evolution of resistance to DDT among crop pests. It is doubtful that the ban would have been proposed or sustained if DDT had not decreased substantially in its effectiveness as a result of evolved resistance. The agreement to allow continued manufacture of high-grade DDT for export for use in public health vector control programs also made the ban politically more acceptable.

The assessment of the environmental effects of DDT is a good example ENVIRONMENTAL EFFECTS OF DDT 373
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. of the interplay between laboratory analyses, field tests, and conceptual developments in both ecology and toxicology. The constraints imposed by the state of both sciences are clear in retrospect. We hope that our awareness of those limits and of the deficiencies in knowledge will make scientists and regulators more alert to evidence that does not fit into our current picture of how the world works.

Reference
Lindemann, R. L. 1942. The trophic-dynamic concept of ecology. Ecology 23:399-418 .

Index
A
Acid rain, 2, 98
Africa, see West Africa
Age
factors in forest management programs,
280-284
harvesting and, 36
size as indicator of, 32-33
structure and life expectancy variations,
33-34
Age classes
of forests, 282-295, 297-298
of halibut, 141
influence of numbers of, on stability,
33-34
Aggregations
influence on harvestability, 32
influence on stability of fish popula-
tions, 33
Agricultural techniques to reclaim derelict
land in Great Britain, 252-254
Algae, as indicators of eutrophication, 49,
83, 303-304, 306, 308, 310, 312, 315
Analog studies of environmental prob-
lems, 75-76
in California red scale control, 180
in caribou protection, 76, 221
in Lake Washington eutrophication con-
trol, 76
in land reclamation projects, 264-265
Annual surplus production (ASP), in
assessment of fishery stock abun-
dance, 145
Anopheles funestus, 195
Anopheles gambiae s.l., 195, 201, 202
Anopheles gambiae s.s., 202
Anopheles species, 190-204
Anticoagulants, in vampire bat control,
35, 155-158
Aphytis africanus, 177
Aphytis chrysomphali, 170, 186
Aphytis linganensis, 170, 171-172, 175,
179, 180, 181, 186, 187, 188
Aphytis melinus, 170, 171-172, 174,
177-178, 179, 180, 181, 182, 186,
187, 188
Aphytis species, in California red scale
control, 166-167, 170, 171 -172,
174-182, 186-188
Atlantic halibut (Hippoglossus hippoglos-
sus), 149
Atomic Energy Commission, U.S., studies
of radiation effects, 79, 112, 331-344
B
Banked specimens, for preservation of
records, 85-86
Baseline studies, value of, 111, 113
Bats, see Vampire bat control
Bean weevil (Callosobruchus chinensis), 5
Behavior, of individuals, 26-28
Biological introductions, see Introduced
species
Biological magnification, 6
Biological monitoring, 3, 4, 81-87, 113-114.

Birds
and biological monitoring, 84-85, 86-87
effects of DDT on, 4, 361-370
effects of spatial isolation on, 32
as invaders, 57
population movements of, 31--32
species diversity among, 51
spotted owl conservation, 227-247
see also specific birds
Black spruce (Picea mariana), 284, 289
Blackcod (Anoplopoma fimbria), 146
Blowflies, 5
Blue-green algae, see Cyanobacteria
Budworms (Choristoneura fumiferana)
DDT in control of, 360, 365
and forest yields, 280, 287, 296
Butterflies, and biological monitoring, 46,
85 C
Caddis flies, and DDT, 362
California red scale (Aonidiella aurantii),
see California red scale control
California red scale control, 13, 14,
165-189
analog studies in, 180
boundaries established in, 170-171
committee comment on, 186-189
ecological knowledge and theories in,
178-183
ecosystem components in, 169
as environmental problem, 167-168
future research plans in, 177-178
introduction of, 166
monitoring in, 170, 176-177
optimal-foraging theory in, 182
predators and parasites in, 166-167,
182-183
problems with pesticides in, 167
project as experiment, 180-181
role of Aphytis species in, 166-167, 170,
171-172, 174-182, 186-188
sex-allocation theory in, 179
Calluna species, and reclamation of
derelict lands in Great Britain, 250,
256, 261, 265, 268
Carbon dioxide, 2, 5
buildup in atmosphere, 98
fixed, 39
Caribou protection, 11, 12, 13, 14, 106,
114, 205-226
analog studies of, 76, 221
behavior studies in, 213, 217, 218, 219
boundaries established in, 216-217
committee comment on, 225-226
conclusions from studies in, 214-215
cumulative effects in, 219
ecosystem components in, 215-216
and effects of development, 210
environmental problems in, 208-212
expert judgments in, 221-222
hypothesis-testing in, 218-219
migration routes in, 208-210, 211, 214,
219
models used in, 221
monitoring in, 217, 218, 219
pilot studies in, 78, 221
population characteristics in, 212-213,
214
and project as experiment, 221-222
and role of ecological knowledge,
212-215, 220, 222-224
sensitivity criteria in, 213-214, 215, 218,
219, 221-222
significance of impacts in, 216
study strategy in, 217-218
Case studies, see Ecological case studies
Catch per unit of effort (CPUE), as index
of fish abundance and biomass,
139-141, 143, 144, 145, 146
Cattle
effect of vampire bats upon, see Vam-
pire bat control
iodine-131 concentrations in milk of, 334
milk containing DDT, 369
sex ratios and increased productivity in,
34 Cesium-134, in radiation study, 337
Cesium-137, in reindeer, 334
Chaoborus, in Southern Indian Lake, 329
Chesapeake Bay, degradation of, 97
Chlorophacinone, in vampire bat control,
155, 156-157
Chromosomes, and nuclear radiation, 337,
338
Chub mackerel (Scomber japonicus), 149
Churchill River, and Southern Indian
Lake, 318-321
Citrus crops, and control of California red
scale, 165-189
INDEX 376
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Civic activities, and eutrophication con-
trol, see Lake Washington eutrophica-
tion control
Clearcut logging, see Forest clearcutting
CO2, see Carbon dioxide
Cohort analysis, in assessment of fishery
stock abundance, 144-145
Communicable Disease Center (CDC), 154
Community ecology, see Ecological com-
munities
Community structure, 3, 5-6, 7, 47-60,
68-69, 73, 92, 255-264, 267 -268,
272-274
effects of radiation on, 337-338
knowledge of, required to predict fate of
DDT, 372
in Lake Washington, 309-310, 312-313
in Southern Indian Lake, 320, 323-324,
328-329
Competitive interactions, 42-45, 52-53
Complex life histories, 25-26, 72
Complex linkages, among species, 4
Construction projects, mitigation effects
of, 12, 205-226, 248-274, 317-330
Contaminants, see DDT; Pesticides; Pollu-
tants; Radionuclides; Toxic Sub-
stances
Control methods, for reducing vampire
bat populations, 154-158
Corridors, between habitats, 6, 32, 233
Coweeta Hydrologic Laboratory, 349
"C-S-R" strategy, in analyzing community
composition, 258, 259
Cucurbitaceae, 35, 46
Cumulative effects, 3, 93-103
in caribou protection program, 219-220
of chemicals, 63-64
definition of environmental, 96-97
difficulties in predicting and controlling,
97-98
in forest management program, 285,
299-300
in halibut fishery management, 143
kinds of, 94-96
management of, 97, 100-103
nibbling and, 96-97
nutrient fluxes and, 64-65
predictions of, 100-101
productivity and, 49
recommendations concerning, 102-103
scales and rates of critical processes in,
98-100
species as indicators of, 49
synergisms in, 96
temporal and spatial factors in, 3, 8-9,
16-17, 68-74, 96, 99, 110
Cyanobacteria, as indicators of eutrophica-
tion, 83, 303-304, 308, 310, 312, 315
Cycling index of nutrient fluxes, 66-67
D
Daffodil (Narcissus pseudonarcissus), and
reclamation of derelict lands in Great
Britain, 265
Daphnia species, in Lake Washington,
309, 310, 312
DDT effects, 4, 6, 10, 11, 13, 14, 29, 64,
65, 77, 78, 85, 88, 109 , 358-374
in birds, 4, 361-370
boundaries for study of, 365-366
in budworm control, 360, 365
committee comment on, 371-374
complex interactions in, 99-100
delayed mortality in, 363
development of resistance to, 35, 360
ecosystem components in, 363
and eggshell thinning, 363, 364, 366,
369-370
in elm bark beetle control, 361, 364,
365, 369
environmental problems and, 106,
360-363
expert judgment in, 369-370
in fish, 360, 361, 363, 365, 367-368
in food chain, 99-100, 361, 363,
364-365, 368, 369
in forest insect control, 364, 371
key issues in, 363-364
in malaria control, 191, 192, 359
in milk, 369
monitoring of, 114, 366-367
pilot-scale experiments in, 78-79
and recolonization of treated areas,
364-365, 368, 372
and reproductive performance, 64
and role of ecological knowledge,
365-368
INDEX 377
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. susceptibility of species to, 362
toxicity of metabolites, 364, 372, 373
in typhus control, 359
worldwide spread of residue, 361-362,
366
Deer, 33, 36, 334
Deforestation, and extinction of species, 86.

See also Forest clearcutting;
Spotted owl conservation
Delaware River Basin Commission
project, 97, 102
Delay-difference population model, in
assessment of fishery stock abun-
dance, 144-145
Demographic stochasticity, and extinction
of small populations, 69
Density dependence, 5, 28
in forest management program, 280-284
in population regulation, 5, 28-29, 30
Denver Wildlife Research Center
(DWRC), 153, 156, 159, 161, 163-164
Department of Agriculture, U.S., 372.

See also Forest Service
Department of Natural Resources (DNR),
New Brunswick, 297
Derelict lands in Great Britain, see Recla-
mation of derelict lands in Great
Britain
Desmodus rotundus (vampire bat), 152,
153, 154
Diaemus youngii (vampire bat), 152, 153
Dichlorodiphenyldichloroethane (DDD),
361, 369
Dichlorodiphenyldichloroethylene (DDE),
363, 369
Dichlorodiphenyltrichloroethane (DDT),
see DDT effects
Differential harvesting by sex and size,
genetic consequences of, 36-37
Diphenadione, in vampire bat control,
155, 157-158, 160
Diphylla ecaudata (vampire bat), 153
Disasters and catastrophes, and extinction
of small populations, 70
Disease control, 10, 14, 25, 190-204
Diversification, in revegetation studies,
260-261.

See also Sward diversification
Diversity indexes, 49-50, 83
Dutch elm disease control program, 364
Dynamics of populations, 24-25, 28-35,
47-48
E
Ecological case studies, 3-4, 12-15,
135-374
Ecological communities, 3, 47-61
competition and, 52-53
factors affecting species diversity in,
50-55
invadability of, 57-58
keystone species, 5-6, 55-56
organization of, 55-56
patches and, 68-74
predation and, 52
preservation of, 11, 205-226. 227-274,
301-316, 345-357
productivity and, 53-54
spatial and temporal considerations in,
54, 68-74
species composition in, 48-50
stability and resilience of, 7, 56-57
substitutability in, 58
succession in, 58-60, 72, 260-261
see also Populations;
Species
Ecological knowledge
applications of, 9-12, 104-115
defined, 1
types of, 4-9
uses, 2
Ecological methods, 2
Ecological theory, defined, 1
Ecotoxicology, 2
Effective population size, 37
defined, 234-236
of spotted owls, 230-242, 246
Eggshell thinning from DDT use, 363,
364, 366, 369-370
El Niño, 90
Elm bark beetle, DDT in control of, 361,
364, 365, 369
Endangered Species Act of 1973, 228
Energy flow, 3, 61-67
Environmental experiments, 3, 75-80
Environmental goals, identification of,
105-107
Environmental problem-solving, scientific
framework for, 104-115
Environmental Protection Agency, U.S.,
359, 360
Environmental stochasticity, and extinc-
tion of small populations, 69-70
Erosion
in clearcut logging, 348-350
INDEX 378
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. in Southern Indian Lake project, 320,
321,322-323
Errors of estimation, and uncertainty of
management projections, 90-91
Equilibrium, in population stability, 29-30
Eutrophication of lakes, 49, 53-54, 64-66,
83, 98, 301-316.

See also Lake Washington eutrophica-
tion control
Evolution, 56-57
affecting problem-solvers, 9
of diet specialization, 40
of life-history parameters, 24
population size and, 37, 69-70, 230-231,
246
of resistance, 10, 24, 35, 36, 42, 167,
360-361, 373
in response to management, 25, 36-37,
95, 107, 109
Extinction of species, 69-70, 72, 86,
137-147, 205-226, 233-237, 365
F
Famophos, in vampire bat control, 155
Federal Committee on Pest Control, 367
Federal Insecticide, Fungicide, and Roden-
ticide Act of 1947, 371
Festuca species, 254, 255, 256, 258, 262
Fir (Abies balsamea), 286, 287
Fish
age classes and, 33-34
aggregations of, and harvesting, 32, 33,
89 effects of DDT on, 360, 361, 363, 365,
367-368
effects of habitat charges upon, 26
effects of population size upon, 29
food supply and, 109
population movements of, 31-32
population stability and, 7, 30
size and sex selectivity in harvesting,
34, 36
types of gear used in catching, 140-141,
144
see also Halibut fishery management;
Southern Indian Lake project;
and specific kinds of fish
Fish and Wildlife Service (U.S. Depart-
ment of the Interior), 152
Fishery Conservation and Management
Act of 1976, 150
Fishing industry, see Halibut fishery man-
agement; Southern Indian Lake
project
Flooding, in Southern Indian Lake,
322-323
Florida Everglades, effects of overcon-
struction in, 95
Food and Drug Administration (FDA),
U.S., 363, 364, 371-372
Food chains
DDT in, 99-100, 361, 363, 364-365,
368, 369
effects of nuclear radiation on, 334,
335-337, 338-339
pollutants in, 63, 64, 94
radionuclides in, 333-334
and Southern Indian Lake project. 320,
328-329
toxic substances in, 93-94, 99-100
Foreign Assistance Act, 152
Forest clearcutting, 345-357
biomass removal in, 350-351, 352-353
committee comment on. 356-357
leaching and erosion in, 348-350
models used in, 353, 357
and nutrient budgets of ecosystems,
346-347, 352, 356-357
and nutrients remaining in soil, 351-352
replenishment of nutrients in, 352
sensitivity analysis of, 352-353, 357
Forest management program in New
Brunswick, 13, 14, 95, 106, 110,
275-300
baseline data in, 298-299
boundaries established in, 279-280
budworm population affecting, 280,
287-296
committee comment on, 297-300
cumulative effects of, 285, 299-300
density-dependent and forest age factors
in, 280-284
ecosystem components in, 279
environmental problems in, 277-278
hypothesis-testing in, 284-285
models used in, 288-290
nomograms used in, 290-295 299
performance indicators in, 278-279
role of ecological knowledge in, 286-288
Forest Service (U.S. Department of
Agriculture), 228, 229
Forests, 7, 8, 29
clearcutting effects in, 13, 14, 345-357
INDEX 379
,
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. conservation of owl populations in,
227-247
deforestation in tropical areas, 86
effects of budworms on, 280, 287, 296
factors affecting growth rates in, 33
insect pests in, 31
isolated populations in, 32
management program in, 13, 14, 95,
106, 110, 275-300
radiation studies in, 336, 337-338
Fragmentation, of habitats, 6, 95
Freshwater Institute, Winnipeg, 318, 321,
325
G
Garki project for malaria control, 79, 112,
190-204
Generic studies of environmental prob-
lems, 76-78, 331-357
Genetic change, see Evolution
Genetic concerns, 24-25, 35-37
cumulative effects of repeated interven-
tions, 95
differential harvesting by sex and size,
36-37
evolution of resistance to pesticides,
35-36
in mosquito control, 202, 203-204
in owl populations, 230, 231, 234-237
in small populations, 37, 69, 73, 234,
235-236
Genetic stochasticity, and extinction of
small populations, 69
Grasslands
seminatural, approaches to recreating,
250, 254-256
urban, succession in, 260-261
see also Reclamation of derelict lands in
Great Britain
Great Britain, see Reclamation of derelict
lands in Great Britain
Grizzly bears (Ursus arctos), 30, 228
Growth rates, 32-33
H
Habitats
disturbances in, and biological monitor-
ing, 86-87
fragmentation of, 6, 95
isolated, 54
management of, 29
migratory patterns and, 31-32
old-growth forests as, 227-247
for owl populations. 26, 229-230, 232,
238-239, 241-242
patches, 54, 68-71, 99
population density and, 29
preservation of, 11, 205-274, 301-316,
345-357
recolonization of, in DDT-treated areas,
364-365, 368, 372
recolonization of, factors affecting, 99
selection of, 26-27
shape and orientation of patches in, 70
sizes of, 32
spatial considerations and, 71-72
species diversity and, 50-51, 54, 68-69
temporal considerations and, 73
Haddock (Melanogrammus aeglefinus),
149
Halibut fishery management, 13, 14, 29,
114, 137-147
boundaries in, 142
catchability of fish in, 140-141
committee comment on, 149-150
cumulative effects in, 143
detection of unreported catches in,
146-147
growth rates in, 142
law of accumulated stock and, 144
life-history characteristics of, 33-34
migration patterns in, 141
monitoring in, 143
natural and fishing mortality estimates
in, 141, 144-145, 146
and project as experiment, 145-146
role of ecological knowledge in, 143,
146-147
stock assessment models for, 139-140,
144-145
stock-recruitment relationship in, 140,
145
Heather (Calluna vulgaris), 250
Heliconius butterflies, 46
Herring gull (Larus argentatus), 84-85
Honey bee (Apis mellifera), 45
Hubbard Brook, New Hampshire, clearcut
logging in, 78, 347-348, 349 -350
Hydroelectric development, and caribou
protection, 205-226
INDEX 380
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

I
Impact assessment, 10-11, 14
in caribou protection, 205-226
criticisms of, 105
development and use of models in,
108-109
and establishment of study boundaries,
109-110
identification of goals in, 105-107
in New Brunswick forest management,
277-279
scoping of problem in, 107-109
significance of effects in, 112
in Southern Indian Lake project, 318-326
testing of hypotheses in, 110-112
and value of monitoring, 111-112,
113-114
see also Prediction of effects
Inbreeding, effects of, 37, 69, 234, 236
Indexes of ecosystem functioning, 66-67
Indexes of species diversity, 49-50
Indicator species, 3, 49, 81-87
Individuals, 3, 5, 23-37
behavior of, 26-28
habital selection by, 26-27
population density affecting, 5
uniqueness of, 5
Insecticides, see Pesticides
Insects, 51,58.

See also California red scale control;
DDT effects;
Malaria control;
and specific insects
Institute of Terrestrial Ecology (ITE),
Great Britain, 266
Integrated pest management (1PM), 361
Interactions among populations, see Popu-
lation interactions
International Institute for Applied Sys-
tems Analysis (IIASA), Austria, 100
International Pacific Halibut Commission
(IPHC), 114, 137-147
Introduced species, 57-58
difficulty of eradicating, 189
difficulty of predicting effects of, 57-58,
165, 182, 188
for pest control, 27, 165-183
Invadability, 57-58
Iodine-131, in milk, 334
Ionizing radiation, 332-333, 337-338, 344.

See also Nuclear radiation effects;
Radionuclides
J
June sucker (Chamistes liorus), 34
K
Kaolin wastes, revegetation of, 253
Key-factor analysis of life histories, 25-26
Keystone species, 5-6, 55-56
L
Labrador Hydro (Hydro), and caribou pro-
tection, 205-226
Ladybird beetles, and DDT, 362
Lake Tahoe, changes in zooplankton
community, 309, 323
Lake Washington eutrophication control,
13, 14, 106, 110, 111, 114 , 301-316,
328
analog studies of, 76
basis of predictions in, 112
and changes in zooplankton community,
309-310, 323
circulation of water in, 72
committee comment on, 312-315
and cyanobacteria, 83, 303-304, 308,
310, 313, 315
effects of phosphorus in, 65, 308,
314-315
model used in, 306-307, 320
nutrient fluxes in, 64-65, 303-304
political involvement in, 305
predictions for, 306-307
and project as experiment, 312
public involvement in, 102, 305-306, 314
results of effluent diversion in, 307-309
scientific studies of, 303-304
sewage discharge in, 73, 95, 102, 302-303
zooplankton in, 309-310
Lake Zurich, eutrophication of, 303, 315
Lakes
eutrophication of, see Eutrophication of
lakes
phosphorus in, effects of, 65, 66, 308,
314-315
pollutants in, 64-65
turbidity of water in, 65, 96, 320, 328-329
variability in productivity of, 65-66
INDEX 381
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. see also Lake Washington eutrophica-
tion control;
Southern Indian Lake project
Land management, see Reclamation of
derelict lands in Great Britain
Latin America, vampire bat control in, see
Vampire bat control
Leaching of nutrients, in clearcut logging,
348-350
Lichens, 83, 334
Life-history patterns, 23-24, 25-26,
33-34, 83
Life span
and ability to survive poor conditions, 34
and apparent stability, 34, 56
pollution and. 83
and resistance to radiation, 339
of vampire bats, 154
Light penetration in water, effects of, 65
Limnocalanus macrurus, in Southern
Indian Lake, 323
Livestock industry, effect of vampire bats
upon. see Vampire bat control
Livestock Insects Investigation Labora-
tory (U.S. Department of Agricul-
ture), 154
Living organisms, role of, in biological
monitoring, 81-86
Lolium species, and reclamation of
derelict lands in Great Britain, 254,
255
Longfin smelt (Spirinchus thaleichthys),
in Lake Washington, 309
Longstone Edge reclamation project in
Great Britain, 248-273
Lotus corniculatus, and reclamation of
derelict lands in Great Britain, 268
M
Malaria control, 13, 14, 25, 26, 112,
190-204
antimalarial drugs in, 194-195, 196
boundaries established in, 193-194
committee comment on, 203-204
DDT in, 191, 192, 359
and dynamics of infection, 197-198
environmental problems in, 192-193
and genetic mechanisms in mosquito
populations, 202, 203-204
insecticides in, 191, 192, 194, 196,
200-202
models used in, 191, 197-200
objectives in, 193
role of ecological knowledge in, 195-196
Management of ecological systems, 3,
88-103
Manitoba, see Southern Indian Lake project
Manitoba Hydro, 319, 322
Maryland Power Plant Siting Program, 76
Materials flow, 3, 61-67
Mating systems, 27, 34-35.

See also Genetic considerations
Mediterranean fruit flies (Dacus), 27
Mercury
in Southern Indian Lake, 65, 223, 329
toxic transformations of, 64
Metapopulations, 230
Migration
habitats affecting, 31-32
routes for caribou, 208-210, 211, 214,
219
Milk
DDT in, 369
iodine-131 in, 334
Mining, and derelict lands in Great
Britain, 249-250
Models
in California red scale control, 179-180
in caribou protection, 220-221
in forest clearcutting, 353, 357
development and use of, 108-109
in forest management program, 288-290
in halibut fishery management, 144-146
in Lake Washington eutrophication con-
trol, 306-307, 320
in malaria control, 191, 197-200
in Southern Indian Lake project, 320
Mollusks, and biological monitoring, 83-84
Monitoring studies, 81-87
advantages of living organisms in, 81-82
baseline in, 111, 112, 113
in California red scale control, 170,
176-177
in caribou protection, 217, 218, 219
choice of organisms for, 82-85
for DDT effects, 366-367
for detection of pollutants, 82-86
in halibut fishery management, 143
INDEX 382
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. for inventory of biological resources,
86-87
in land reclamation project, 254, 255
and specimen banking, 85-86
in spotted owl conservation, 242
usefulness of, 79, 111-112, 113-114
Moose (Alces alces), 210
Mosquitoes, see Anopheles species;
malaria control
Moths, and biological monitoring, 85
Multiple-Use Sustained Yield Act of
1960, 229
Mutualistic interactions, 6, 45-46
Mysis relicta
in Lake Washington, 312
in Southern Indian Lake, 323
N
National Environmental Policy Act, 106
National Forest Management Act of 1976,
228
Natural variability, and uncertainty of
management projections, 89 -90
Natural-history information, usefulness
of, 16
Nature Conservancy Council (NCC),
Great Britain, 266
Neguvon, in vampire bat control, 154
Neomysis mercedis, in Lake Washington,
309
Netherlands, reclamation of derelict lands
in, 265, 266
New Brunswick forest management pro-
gram, see Forest management pro-
gram in New Brunswick
Newfoundland caribou, see Caribou pro-
tection
Newfoundland Wildlife Division (NWD),
and caribou protection, 205-226
Nibbling, cumulative effects of, 96-97
Nitrogen, 63, 64, 351
Nomograms, in forest management pro-
gram, 290-295, 299
Nonlinear catch-age (catch at age) analy-
sis, in assessment of fishery stock
abundance, 145
North Atlantic Treaty Organization, 366
North Pacific halibut, see Halibut fishery
management
Northern pike (Esox lucius), in Southern
Indian Lake, 322, 323
Nuclear radiation effects, 13, 14, 79,
331-344
basis of predictions in, 112
chromosomes and, 337, 338
committee comment on, 343-344
environmental problems in, 333-335
in food chains, 334, 335-337, 338-339
at nuclear test sites, 333
and predictions of radionuclide move-
ments, 335-337
sensitivity of species in, 337-338
specific-activity concept in, 335
systems analysis approach in, 337
see also Ionizing radiation;
Radionuclides
Nutrients
fluxes in, 64-65, 66-67
interactions, 65-66
in lakes, 64-66, 303-304, 320-321
losses due to clearcut logging, 345-357
regeneration of. 66, 96
O
Oak Ridge National Laboratory, 337
Old-growth forests, as habitat for owl pop-
ulations, 227-247
Organisation for Economic Co-operation
and Development (OECD), 366, 367
Oscillatoria rubescens, see Cyanobacteria
Owls, see Spotted owls
P
Pacific halibut (Hippoglossus stenolepis),
see Halibut fishery management
Pacific Northwest, conservation of owl
populations in, see Spotted owl con-
servation
Pacific Northwest, halibut fishery man-
agement in, see Halibut fishery
management
Pacific sardines (Sardinops sagax), 149
Passion flower (Passiflora spp.), 46
Patches
and community composition, 68-71
distribution of, in space and time, 71-73
effects of shape, 70-71
INDEX 383
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. effects of size, 68-69, 99
Patuxent Wildlife Research Center, 363,
368
Peak District National Park, Great Britain,
250
Peak Park Planning Board, 253, 267
Pest control, 10, 14, 165-189.

See also Pesticides;
and specific pests
Pesticides
in California red scale control, 167, 169
evolution of resistance to, 10, 35-36, 73,
96 in integrated pest management, 361
in malaria control, 191, 192, 194, 196,
200-202
timing in applications of, 73
in vampire bat control, 154-158
see also DDT;
Toxic substances;
and specific pesticides
pH factors affecting derelict lands in
Great Britain, 260, 263
Phosphorus, effects of, in water, 65, 66,
308, 314-315
Pilot-scale studies of environmental prob-
lems, 78-79, 111
in caribou protection, 78, 221
of DDT effects, 78-79
in land reclamation projects, 78
in vampire bat control, 159
Plantation management in forest develop-
ment, see Forest management pro-
gram in New Brunswick
Plantation model, in forest development
program, 288, 289-290
Plants, 5
and biological monitoring, 83
competition among, 53-54, 258-259,
263-264
as invaders, 57
mutualistic interactions of, 45-46
natural defenses of, 41-42
and population interactions, 38-39
radiation studies in, 338
role of, in reclamation of derelict lands
in Great Britain, 248-273
species diversity, 51
and stability boundaries, 7
Plasmodium falciparum, 192, 193, 196,
200, 202
Plasmodium malariae, 192, 196, 202
Plasmodium ovale, 192, 196, 202
Plasmodium species, 190-204
Plasmodium vivax, 192
Plowshare Program, 334
Pollutants, 11
in atmosphere, 98-99
biological effects of, 11, 63-64
and biological monitoring, 82-85
cumulative addition of, 95, 96, 98, 100,
102
effects on Lake Washington, 302-315
and eutrophication in Lake Washington,
301-315
in food chain, 63, 64, 94
in Lake Erie, 63-64
monitoring for effects of, 81-87, 113-114
monitoring of, 82-86
from nuclear radiation, 331-344
see also DDT;
Pesticides;
Radionuclides;
Toxic substances
Polychlorinated biphenyls (PCBs), 63,
114, 367
Polygynous mating systems, 34
Population biology, 23-37.

See also Populations
Population fragmentation, 6
Population interactions, 38-46
competitive, 42-45
indirect effects of, 46, 47-48, 88
mutualistic, 6, 45-46
patterns of succession, 59
predator-prey, 6, 40-42, 309-310, 324,
329
Populations
age structure of, 33-34
aggregations of, 32
behavior of, 26-28
biology of, 23-37
carrying capacity and, 28
control of, 27-28
density-dependent factors in, 5, 28-29, 30
dispersion patterns and migratory behav-
ior in, 31-32
dynamics of, 24-25, 28-35, 47-48
effects of habitat changes upon, 26
in forest development, 280-284
genetic and evolutionary changes in,
24-25, 35-37
growth rates in, 32-33
harvesting practices affecting, 95
INDEX 384
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. isolated, 32
key factors in identifying, 25-26
life-history information and, 23-24, 25-26
mating systems of, 27
maximal sustainable yield and, 29
predator introduction into, 26-27
regulation of, 28-29
sex ratios and sex biases in, 34-35
single, 23-37
size estimates of, 29, 30
social interactions among, 27-28
stability of, 29-31
see also Ecological communities;
Population interactions;
Species
Power generation
and nuclear radiation effects, 331-344
and raising lake level for hydroelectric
project, 317-329
Predation, 52
Predator-prey theory, in California red
scale control, 182-183
Predators
as control agents, 26-27, 41
dietary practices of, 40-41
in donor-controlled systems, 41
interactions with prey, 6, 40-42,
309-310, 324, 329
keystone species of, 6, 55-56
natural resistance of plants to, 41-42
in predator-controlled systems, 41
Prediction of effects
for caribou herd, 218-219
for clearcut logging, 353
for cumulative effects, 10-11, 100-101,
112-114
for DDT effects, 358-374
for forestry development, 288-290
limitations of, 79
for nuclear radiation effects, 331-344
for Southern Indian Lake project,
317-330
uncertainties in, see Uncertainty
see also Impact assessment
Preservation of ecological communities
and habitats, 11
Probabilistic thinking, 17
Productivity
and fluxes in energy and matter, 61-62
as index of perturbation, 63, 66-67
interactions, 65-66
of lakes, variability in, 65-66
and species richness, 53-54
Projects as experiments, 15-16, 75-80, 87,
111
in California red scale control, 180-181
in caribou studies, 221-222
in halibut fishery management, 145-146
in Lake Washington eutrophication con-
trol, 312
in Southern Indian Lake project, 318
in vampire bat control, 159
Propoxur, in malaria control, 194, 196
Puget Sound, Washington, toxic wastes
in, 94, 302
R
Rabies, 152, 158
Radioecology studies, see Nuclear radia-
tion effects
Radionuclides
in environment, 331, 332, 335-338,
343-344
in food chain, 333-334, 343-344
see also Ionizing radiation;
Nuclear radiation effects
Random variation, and uncertainty of
management projections, 90
Randomness in ecological systems, 90
Reclamation of derelict lands in Great
Britain, 13, 14, 99, 248-273
agricultural techniques in, 252, 257, 267
analog studies in, 264-265
basis of predictions in, 112
committee comment on, 271-273
competition among species in, 263-264
as contribution to ecological knowledge,
267-269
environmental problems in, 250-254
and ethics of deliberate introductions,
254, 266
expert judgments in, 265-266
factors affecting growth rates in, 33
in kaolin wastes, 253
and laissez-faire attitudes, 251-252
low-productivity vegetation in, 253, 257
monitoring in, 254, 255
nutrient status affecting, 252, 256, 261,
263, 273
pilot-scale experiments in, 78
INDEX 385
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved. role of ecological knowledge in,
256-261, 267-268
seminatural grasslands in, 254-256
soil seed banks in, 259, 273
sources of ecological information in,
261-266
successions in, 260-261
sward diversification in, 253-259, 261,
263
theoretical concepts in, 264
yield and relative growth rate in, 254, 257
Recolonization, 72, 99, 364-365, 368, 372.

See also Reclamation of derelict lands
in Great Britain
Recommendations, 15-17
Red mites, nonsusceptibility to DDT, 362
Red scale control, see California red scale
control
Regulation of populations, 28-29
Reindeer, cesium-137 concentrations in,
334
Relative growth rate (RGR). 254, 257, 258
Renewable-resource management, 9-10.
14, 137-150, 275-300, 345-357
Reproductive potential (r), 28
harvesting species with low r, 30
Reproductive rate
change in, as indicator of stress, 83
and colonizing ability, 71
of mosquitoes, 197, 198, 201
of vampire bats, 154, 164
Resilience, of ecological communities,
56-57
Resistance to pesticides, evolution of, 10,
35-36, 73, 96
Restoration projects, 12, 248-274
Roe deer (Capreolus capreolus), 33
Root rot, 169
S
Saguaro cactus (Carnegiea gigantea), 34
Salmon, 25, 31, 34, 36, 362
San Joaquin Valley, control of California
red scale in, 168, 177, 186
Scales, temporal and spatial, 3, 8-9,
16-17, 68-74, 96, 98-100, 110
Scoping, in environmental problem-
solving, 107-109
Sea birds, and biological monitoring. 84-85
Seasonal changes, and natural variability,
89 Seattle, Washington, see Lake Washing-
ton eutrophication control
Seed banks, soil, 259, 273
Seedling yellows (Tristeza), 169
Sensitivity analysis
of clearcut logging effects, 352-353, 357
and management plans, 92
of nuclear radiation effects, 337-338
Sewage discharge into Lake Washington,
73, 95, 102, 302-303
Sex biases, 34-35
Sex ratios, 34-35, 36, 37
Size
as indicator of age, 32-33
of habitats, 32
harvesting and, 36
Social behavior, and extinction of small
populations, 70
Social interactions. 27-28, 70
Sockeye salmon (Oncorhynchus nerka), 34
Soil erosion, see Erosion
Soil seed banks, 259, 273
Southern Indian Lake project, 13, 14, 106,
111, 114, 317-330
analog studies of, 76
basis of predictions in, 112-113
changes in zooplankton community,
323-324, 329
committee comment on, 328-329
compared to Siberian reservoirs, 321
effects on whitefish, 26, 318, 320, 322,
329
experimental nature of, 318
flow diversion in, 72
mercury release in, 96, 323, 329
models used in, 320
nutrient fluxes in, 64-65, 320-321
predevelopment studies in, 319, 321
retrospective analysis of, 76, 321-326
shoreline erosion in, 320, 321, 322-323
temperature changes in, 323, 328
turbidity of water in, 65, 96, 320, 328-329
Spatial factors, 3, 8-9, 16-17, 54, 68-74,
96, 99, 110
INDEX 386
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Species
and aggregate variables, 7-8
and biological magnification, 6
changes after DDT application, 362
complex linkages among, 4-5
complexity and uncertainty in the study
of, 8
composition of, 48-50, 53, 55, 56
and density dependence, 5
diversity, 49-56
effects of competition among, 52-53
effects of predation on, 52
effects of productivity on, 53-54
extinction of, 69-70, 72, 86, 233-237. 365
as indicator of pollution, 83
individuals, 3, 5, 23-37
influence of spatial and temporal scales
on, 3, 8-9, 16-17, 54, 68-74, 96, 99,
110
keystone, 5-6, 55-56
number in relationship to area, 68-69
population fragmentation among, 6
richness, 49, 53-54
stability boundaries, 7
see also Ecological communities;
Population interactions;
Populations;
and specific species
Specimen banking, 85-86
Spotted owl conservation, 13, 14, 110,
227-247
committee comment on, 245-247
demographic factors in, 28, 32, 246
dispersal patterns in, 230, 232-233,
238-239
environmental problems in, 229-232
estimation of population in, 240
genetic factors in, 230, 231, 234-237
habitats in, 26, 229-230, 232, 238-239,
241-242
issues involved in, 231-232
management units in, 231
monitoring in, 242
planning process for, 237-242
and risk of extinction, 233-237, 246
role of ecological knowledge in,
232-237, 242-243
Spotted owl management area (SOMA),
231, 242
Spruce (Picea species), 284, 287
Spruce budworms, see Budworms
Stability of populations, 7, 29-31, 56-57
Stand-level considerations, in forest man-
agement, see Forest management
program in New Brunswick
Starfish (Pisaster), 46
Sterile male flies, used in screwworm con-
trol, 27
Striped bass (Roccus saxatilis), 149
Strix occidentalis (spotted owl), see Spot-
ted owl conservation
Strontium-90, 336
Study boundaries, establishment of,
109-110
Subarctic lake, raising the level of, see
Southern Indian Lake project
Subsoils, nutrient-deficient, in Great
Britain, 254-256, 260, 261, 262, 263
Substitutions, in ecological communities,
58 Successions, in ecological communities,
58-60, 72, 260-261
Sulfur, 63, 64
Sustainable Development of Biosphere
project. 100
Sward diversification, and reclamation of
derelict lands in Great Britain,
253-259, 262, 263, 267
T
Taxonomic collections, lack of, 86
Temporal factors, 3, 8-9, 68-74. 96, 110
Tetrachlorodiphenylethane (TDE), 361
Thresholds in ecological systems, 7
Thrips (Scirothrips citri), 169
Timber yield, 275-306
Toxic substances, 11, 39
biological concentrations of, 63-64
and biological monitoring, 81-82
defensive chemicals as, 41-42
and derelict lands in Great Britain, 249,
250-251
dispersal of, 94
in food chain, 93-94, 99-100
interactions of, 96
in plants, 41-42
in vampire bat control, 154-158
see also DDT;
Pesticides;
Pollutants;
Radionuclides
INDEX 387
Ecological Knowledge and Environmental Problem-Solving: Concepts and Case Studies Copyright National Academy of Sciences. All rights reserved.

Tree lupine (Lupinus arboreus), 253
Trees, see Forest clearcutting;Forest man-
agement program in New Brunswick;
Forests
Triaenophorus crassus (parasitic cestode),
in Lake Washington, 318
Typhus, DDT in control of, 359
U
Uncertainty, 3, 88-92
and errors of estimation, 90-91
and lack of knowledge, 91
and natural variability, 89-90
preparation for, 17
and problem-solving, 8
and random variations, 90
sources of, 88-91
and use of monitoring studies, 79
Unit of Comparative Ecology, Sheffield
University, Great Britain, 263,
265-266
United Nations, 366
Upper Salmon Hydroelectric Develop-
ment, and caribou protection, 205-226
U.S.-Canadian fishery agreements,
138-139
V
Vampire bat control, 13, 14, 27-28,
151-164
anticoagulants in, 35, 155-158
committee comment on, 163-164
effectiveness of, 160
environmental problems in, 153
as experiment, 159
methods for, 154-158
pilot studies in, 159
reproduction potential of bats, 28
research studies in, 153-154
role of ecological knowledge in, 158-159
systemic treatment of cattle in, 154-155,
157-158
topical treatment of bats in, 155-157, 158
training for procedures in, 159
vaccines in, 158
Vegetation, restoration of, see Reclama-
tion of derelict lands in Great Britain
W
Walleye (Howarth vitreum), in Southern
Indian Lake, 322, 323
Wasps, see mutualism species
Wastelands, reclamation of, see Reclama-
tion of derelict lands in Great Britain
West Africa, experimental control of
malaria in, 190-204
Whales, 28, 32
Whitefish (Coregonus clupeaformis), in
Southern Indian Lake, 318, 320, 322
Wilderness Act of 1964, 229
Wolves (Canis lupus), 32, 228
Wood supply, see Forest clearcutting;
Forest management program in New
Brunswick;
Forests
Wood Supply and Forest Productivity
(WOSFOR) model, in forest man-
agement program, 288-289, 290
Woodland caribou (Rangifer tarandus),
208
World Health Organization, 191, 360
Y
Yield
of halibut, 137-138
maximal, 29, 91, 106, 137-138
of timber, 275-300
Z
Zooplankton
in Lake Washington, 309-310
in Southern Indian Lake, 323-324, 329
INDEX 388
1 2
3 4
5 6
7 8
9 IN THE UNITED STATES DISTRICT COURT

FOR THE EASTERN DISTRICT OF CALIFORNIA

NATIONAL WILDLIFE FEDERATION,
10 et al.,
11 12
13 14
15 16
17 18
19 20
21 22
23 24
Plaintiffs,
v.

GALE NORTON,
Defendant.

CIV-S-04-0579 DFL JFM

MEMORANDUM OF OPINION

AND ORDER
Plaintiffs National Wildlife Federation, Friends of the Swainson's Hawk, Planning and Conservation League, and Sierra Club allege that the Secretary of the Interior violated the Endangered Species Act by approving the Natomas Basin Habitat Conservation Plan and issuing incidental take permits to the City of Sacramento and Sutter County. By orders of July 2, 2004 and July 19, 2004, the court permitted the City of Sacramento, Sutter County, and several landowners in the affected area to intervene 25 as defendants. The parties now cross-move for summary judgment. 26 This is the second time that the court has been asked to 1 1 review a habitat conservation plan for the Basin. See Nat'l 2 Wildlife Fed'n v Babbitt, 128 F.Supp.2d 1274 (E.D. Cal. 2000) 3 ("Natomas I") . In Natomas I, the court held that the habitat 4 conservation plan was inadequate. For the reasons that follow, 5 the court now finds that the revised plan satisfies the 6 requirements of the Endangered Species Act ("ESA").
7 8
9 10
11 12
13 14
15 16
17 18
19 20
21 22
23 24
25 26
I. Background
A. History of the Habitat Conservation Plan
The Natomas Basin is a low-lying region of approximately 53,000 acres in Sacramento and Sutter Counties. (Administrative Record ("AR") 59.) The Basin is home to the Giant Garter Snake ("GGS") and the Swainson's Hawk, the two species of greatest concern in this litigation. In 1993, the Secretary listed the GGS as a threatened species under the Endangered Species Act ("ESA") . 1 The Swainson's Hawk has not been listed as a threatened species by the Secretary; however, it has been so listed by the California Department of Fish and Game under the California Endangered Species Act. 14 C.C.R. § 670. 5 (b) (5) (A), Cal. Fish & Game Code § 2067. The GGS is an elusive animal that lives in rice fields and drainage ditches; it is found only in rice-growing regions of the Central Valley, including rice- growing areas in the Basin and associated canals. The Swainson's Hawk is migratory, wintering in Central and South America and 1 A "threatened" species is any species which is likely to become an endangered species within the foreseeable future throughout all or a significant portion of its range. 16 U.S.C. § 1532 (20).
2 1
2 3
4 5
6 7
8 9
10 11
12 13
14 15
16 17
18 19
20 21
22 23
24 25
26 spending March to September in the Central Valley. The hawk forages on small rodents found in large open fields and nests in large trees; there are approximately 50 identified hawk nesting sites in the Basin and the riparian corridor of the Sacramento river. (AR 952-55.)
The ESA conditions the Secretary's issuance of an incidental take permit ("ITP") upon her approval of a habitat conservation plan ("HCP). The ITP allows activity, here development, that could injure or harm - "take" in the language of the statute - an endangered or threatened species. Without an ITP, developers would be subject to serious penalties, including criminal prosecution, for any injury to an endangered or threatened species. 16 u.s.c. §§ 1538, 1540. The history of the
development of the Natomas Basin HCP is outlined in Natomas I and need not be repeated here in detail. See Nat'l Wildlife Fed'n v Babbitt, 128 F.Supp.2d at 1277-78. In brief, the first HCP for the Basin was a regional conservation plan designed to cover development in the entire Basin. Id. at 1279. The 1997 HCP anticipated that five jurisdictions -- the City, Sacramento County, Sutter County, Reclamation District No. 10, and the Natomas Central Mutual Water Company -- would apply for ITPs.

Id. The HCP was designed to permit development of 17,500 acres of Basin land over the 50-year life of the ITPs, with mitigation lands acquired at a .5-to-1 ratio as land was developed. Id. at 1280. The acquisitions were to be funded with mitigation fees paid by developers in the relevant jurisdictions. Id. In the 3 1 end, however, only the City applied for a permit under the HCP. 2 Despite the lack of participation by the other jurisdictions, the 3 Secretary granted an ITP to the City, finding that the issuance 4 of an ITP to the City, with the limitations imposed by the HCP, 5 would not likely jeopardize the continued existence of the 6 species covered by the HCP, including the GGS and the Swainson's 7 8
9 10
11 12
13 14
15 16
17 18
19 20
21 22
23 24
25 26
Hawk. Id. at 1282-84.

The Secretary's issuance of an ITP to the City was
challenged by various organizations, among whom were the plaintiffs in the present action. On August 15, 2000, the court found that several of the Secretary's findings were unreasonable and violated the ESA, thereby setting aside the Secretary's issuance of the ITP. Id. at 1292-1300. Specifically, the court found that the Secretary erroneously concluded that: (1) the HCP minimized and mitigated the impact of the permitted takings "to the maximum extent practicable;" (2) the City had ensured adequate funding for the mitigation plan; and (3) issuance of the ITP to the City, in the absence of participation by the other jurisdictions, would not jeopardize the continued existence of the covered species. Id. On this last issue, the court found that a principal failing of the 1997 ITP was that it relied upon an HCP that took a regional approach to conservation, covering the entire 53,000-acre Basin, when in fact only the City sought an ITP and agreed to be bound by the HCP. Id. at 1291, 1299.

The HCP failed to analyze possible effects on the species in the event that the other jurisdictions -- primarily Sacramento and 4 1 Sutter counties -- failed to subscribe to the HCP. However, the 2 court upheld the conclusion that the HCP would not result in 3 jeopardy to the covered species were it implemented by the five 4 jurisdictions envisioned by the HCP. Id. at 1295-98.
5 6
7 8
9 10
11 12
13 14
15 16
17 18
19 20
21 22
23 24
25 26
After the 1997 ITP was set aside, the City revised the HCP to address the flaws identified by the court. Currently before the court is a revised HCP, covering development only by the City and Sutter. This second Natomas Basin HCP ("NBHCP") was approved by the Secretary in April 2003. At the same time, the Secretary granted ITPs to the City and Sutter, authorizing a total of 15,517 acres of development. 2 Shortly thereafter, the Service and the Secretary issued the required supporting documents: a Biological Opinion ("BiOp") examining effects upon the species listed in the proposed ITP, Findings and Recommendations supporting the issuance of an ITP, a final EIR/EIS, and the Record of Decision, in which the Secretary adopted the Service's findings and announced her decision to issue the ITPs. 3 2 The NBHCP also anticipates development of 1,983 acres in the Metro Air Park project. (Pl.'s Mot. at 5-6.) An ITP was previously issued for that project, an action the court upheld in Nat'l Wildlife Fed'n v. Norton, 306 F.Supp.2d 920 (E.D. Cal. 2004) ("Metro Air Park"). In total the NBHCP anticipates 17,500 acres of development in the Basin.
3 The final EIR/EIS was prepared to comply with the public agencies' obligations under the National Environmental Policy Act and the California Environmental Quality Act. The EIR/EIS evaluates a broader range of potential environmental impacts.

The adequacy of the EIR/EIS is not challenged in this action.

However, because the EIR/EIS is part of the record of decision, the analysis and responses to comments in the EIR/EIS are relevant to evaluating the Secretary's conclusions under the ESA. 5 1
2 3
4 5
6 7
8 9
10 11
12 13
14 15
16 17
18 19
20 21
22 23
24 25
26 B. The Final NBHCP
The purpose of the NBHCP is to "promote biological
conservation in conjunction with economic and urban development within the permit area." (AR 19.) The final NBHCP envisions participation by the City and Sutter, but does not depend on participation by both entities. (Id. at 50-51.) The NBHCP covers 22 species, with particular attention to the GGS and the Swainson's Hawk, since they are prominent in the Basin, listed as threatened under state or federal law, and occupy habitat that will also benefit other covered species. (Id. at 64.) The NBHCP anticipates that development by the City and Sutter will be limited to 15,517 acres -- 8,050 acres within the City and 7,467 acres in Sutter County -- and provides that approval of any development beyond this limit -- whether by the City and Sutter or by other entities -- will trigger reevaluation and possible amendment of the plan, and could result in suspension or revocation of the City and Sutter permits. (Id. at 20, 23-26, 110.)
Like the 1997 HCP, the primary mitigation measure relied on in the NBHCP is acquisition and enhancement of reserve properties at a .5-to-l ratio for all of the lost habitat, to be funded by developer fees. 4 (Id. at 36, 169-98.) The 8,750 acres of reserve land will be divided in the following fashion: 50% in 4 The plan also requires other minimization and avoidance measures. In particular, the plan calls for preservation of the "Swainson's Hawk Zone," a one-mile strip of land adjacent to the Sacramento River, which contains many of the Swainson's Hawk nesting sites in the Basin. (AR 1032, 1034.)
6 1 rice cultivation that will serve as habitat for the GGS; 25% in 2 managed marsh habitat for the GGS; and 25% in upland habitat that 3 could be used for foraging by the Swainson's Hawk. (Id. at 37.) 4 The NBHCP offers the following reasons to support the adequacy of 5 the .5-to-1 ratio: (1) the reserves will provide higher quality 6 habitat than the lands to be developed, especially given that the 7 reserves will be managed for the covered species; (2) much of the 8 land to be developed is of limited value as habitat but will be 9 assessed as if it were of value; (3) the reserves will provide 10 permanent habitat for the covered species; (4) the NBHCP provides 11 monitoring and adaptive management to protect the species; and 12 (5) the reserves will be large and biologically viable. (Id. at 13 132-33.) On this last point, the NBHCP requires that, by the end 14 of the SO-year term of the plan, there must be one reserve of at 15 least 2,500 acres in size, with the remaining reserves in blocks 16 of 400 acres or more that will be connected by watercourses. 17 18
19 20
21 22
23 24
25 26
(Id. at 140.) The NBHCP also provides that setback zones should be "considered" prior to the acquisition of reserve lands and that, if possible, reserve lands should be located 800 feet or more from urban development. (Id. at 142.)
The NBHCP contains several provisions designed to ensure that its environmental objectives will be achieved and that development will not outpace the acquisition of mitigation lands.

One of these provisions is a 200-acre "cushion" of mitigation land, which requires that, as of September of each year, the Natomas Basin Conservancy ("NBC") -- which is the entity that 7 1
2 3
4 5
6 owns, acquires, and manages the reserve lands -- must have acquired at least 200 more acres of reserve land than necessary to mitigate all of the development approved to that date before any further development is permitted. (Id. at 213-14.) This requirement is intended to assure that the NBC will never be unable to find and acquire mitigation lands for development that 7 has already been approved. The NBHCP also calls for increases in the mitigation fees, as necessary and without a cap, to pay for 8 9 increased costs of land acquisition and reserve management. (Id. 10 at 211.) "Catch-up fee" ordinances enacted by the City and 11 Sutter will require developers who have received development 12 approval to pay the increased mitigation fees if they have not 13 yet engaged in ground-disturbing activity. (Id. at 212.) The NBHCP also imposes monitoring and review obligations designed to ensure that the plan will achieve the desired conservation objectives and goals. Two types of monitoring are required by the plan: (1) compliance monitoring, to assure that the reserves are properly acquired and managed; and (2) biological effectiveness monitoring, to determine whether the 14 15
16 17
18 19
20 21
22 23
24 25
26 assumptions of the plan hold true in practice over time. (Id. at 217-32.) Beyond these continuing obligations, the plan also calls for broad reviews at designated development milestones. overall program review will be conducted once 9,000 acres of development have been approved. (Id. at 239-41.) The overall program review will be made available for public review and comment. (Id.) The NBHCP also calls for an independent mid- 8 An
1 point review by each permittee to address the possibility that 2 development might proceed faster in one jurisdiction than the 3 other. The NBHCP also provides for "adaptive management" 4 to respond to the monitoring, reviews, or other new scientific 5 6
data. (Id. at 234-38.) Should the adaptive management
provisions prove insufficient, the NBHCP can be amended or 7 revised, or, in the worst case, the permits could be suspended or 8 revoked. (Id. at 252-56.)
9 10
11 12
13 14
15 16
17 18
19 20
21 22
23 24
25 26
The NBHCP concedes that there will be harm to the species listed in the permit, but contends that the harm will be significantly reduced by the measures described above. (Id. at 263-330.) The NBHCP further finds that the proposed level of mitigation is the "maximum extent practicable," relying on economic analysis. (Id. at 332-36.) Finally, the NBHCP
concludes that the plan as proposed is the best option among other possible alternatives in light of biological and financial considerations. (Id. at 336-38.)
Through the implementation agreement, the City and Sutter agree to be bound by the terms of the NBHCP, which are also incorporated into the ITP. (Id. at 806-55.)
C. The 2003 Biological Opinion
The Biological Opinion ("BiOp") is an evaluation by the Service's wildlife biologists of the potential effects of issuance of the ITPs on the species identified in the plan and 9 1 the proposed ITPs. 5 16 U.S.C. § 1536. While acknowledging that 2 the proposed development will have both direct and indirect 3 negative effects on the GGS, the BiOp ultimately concludes that 4 the ITPs will not affect the viability of the GGS population 5 within the Natomas Basin or the viability of the species as a 6 whole. (AR 1028.) This finding is based on several factors: (1) 7 the minimization and take avoidance measures imposed by the 8 NBHCP, including pre-construction surveys and de-watering and 9 fencing of canals; (2) the protection, enhancement, restoration, 10 or creation of 6,562.5 acres of higher quality reserves for the 11 snake; (3) the maintenance of connectivity between reserve lands; 12 (4) the continued existence of 16,000 acres of GGS habitat that 13 will remain in the Basin after development; and (5) the creation 14 of year-round protected habitat in the reserves specifically 15 managed to benefit the GGS and not subject to the vagaries of 16 rice farming.
17 18
19 20
21 22
23 24
25 26
The BiOp similarly acknowledges that the Swainson's Hawk will suffer some negative impacts from implementation of the proposed action, including the loss of up to 20% of its nesting habitat and 40% of its foraging habitat in the Basin, most of which is of high or moderate quality. (Id. at 1032-37.) However, the BiOp concludes that the proposed action will not jeopardize the survival of the Central Valley population of the Swainson's Hawk or the species as a whole because: (1) the 5 The BiOp is required by the ESA because the issuance of an ITP is a federal action. See part II.A, infra.
10 1 reserves created will provide foraging opportunities at the 2 appropriate time of year, during nesting; (2) approximately 3 13,000 acres of foraging habitat will not be affected; (3) the 4 acquired foraging habitat will be closer to the nesting trees; 5 (4) more high quality foraging habitat will be created; and (5) 6 significant foraging habitat exists to the west, in Yolo County. 7 8
9 10
11 12
13 14
15 16
17 18
19 20
21 22
23 24
25 26
(Id. at 1039.) In sum, the BiOp concludes that issuance of ITPs to the City and Sutter will not appreciably reduce the likelihood of the survival and recovery of these species in the wild.

D. The 2003 Findings and Recommendations
The Findings and Recommendations represent the Service's findings in light of the information in the NBHCP and the BiOp.

The Service concludes that the impacts of the issuance of the ITPs will be minimized and mitigated to the maximum extent practicable. (Id. at 1182.) Specifically, the Service finds that the injury to the species covered by the permit will be mitigated to the maximum extent practicable by: (1) the measures identified in the NBHCP; (2) establishment, enhancement, and active management of 7,758.5 acres of high-quality reserve habitat; (3) establishment of a monitoring and reporting plan; and (4) use of a funding mechanism that contains assurances that the NBHCP will be implemented. (Id. at 1184-85.) The Service finds that the harm will be minor for all of the covered species, except the GGS, and that, for all covered species, including the GGS, the consequences of the harm will be effectively mitigated by the conservation measures provided for by the NBHCP. More 11 1 specifically, as to the Swainson's Hawk, the Service determines 2 that the degree of injury will be low because: (1) the majority 3 of the foraging habitat that will be lost is currently not 4 available to the hawk during its nesting season 6 ; (2) substantial 5 foraging habitat will exist in the Basin even after the planned 6 development; and (3) substantial foraging habitat will remain in 7 Yolo County. (Id. at 1188-89.) The Service further concludes 8 9
10 11
12 13
14 15
16 17
18 19
20 21
22 23
24 25
26 that any injury will be effectively mitigated for the reasons stated in the BiOp and described above.

As to the GGS, the Service acknowledges that there is a higher likelihood of injury to the GGS, but still concludes that the injury will be effectively mitigated by the measures included in the NBHCP, including: (1) limits on construction during the GGS dormant period; (2) pre-construction surveys and dewatering; (3) the creation of managed marsh habitat; (4) the acquisition of rice fields and their management with "snake friendly practices"; and (5) the assurances of connectivity of snake habitat. (Id. at 1190-91.)
The Findings and Recommendations conclude that, in addition to effectively mitigating for the anticipated harm, the NBHCP provides for mitigation to the "maximum extent practicable," relying on the economic analysis prepared in conjunction with the 6 The Swainson's Hawk nests and breeds in the Basin from March to September. However, under current conditions, much of the potential foraging habitat is cultivated with crops that either do not provide good foraging habitat for the Swainson's Hawk or do not provide good foraging habitat until August or September, after young have fledged. (AR 725-49.)
12 1 NBHCP. (Id. at 1192-93.) The Findings and Recommendations 2 likewise conclude that the City and Sutter have ensured adequate 3 funding for the plan through: (1) the mitigation fee program, 4 which includes five different components and which is not capped; 5 (2) the catch-up fee ordinance; and (3) the 200-acre cushion of 6 mitigation land. (Id. at 1194-95.) Finally, the Findings and 7 Recommendations conclude that the issuance of the ITPs will not 8 appreciably reduce the likelihood of survival and recovery of any 9 of the covered species, including the Swainson's Hawk and GGS. 10 11
12 13
14 (Id. at 1196-99.) In accordance with these Findings, the Service recommends that the Secretary of the Interior approve the ITPc, for the City and Sutter.

The Findings and Recommendations became the decision of the Secretary of Interior once she adopted them in the Record of 15 Decision and issued the two ITPs. 7 This suit followed in March 16 2004. Plaintiffs seek a ruling that the Secretary's findings are 17 18
19 20
21 22
23 24
25 26
not supported by the record and that the Secretary's approval of the NBHCP and issuance of the ITPs should be revoked.

II. Statutory Standards and Requirements
A. The Endangered Species Act
The purpose of the ESA is to "conserve ecosystems upon which endangered and threatened species depend" and "to provide a 7 Although the Secretary made the final decision to issue an ITP, her conclusions depend on the analysis and findings of the Service. Therefore, although the Secretary is the named defendant in this case, the court will frequently refer to the BiOp and Findings and Recommendations of the Service in reviewing the decision of the Secretary.
13 1
2 3
4 5
6 7
8 program for the conservation of such endangered species." 16 U.S.C. § 1531(b). As a means of achieving this goal, Section 9 of the ESA prohibits private individuals from "taking" endangered or threatened species. Id. § 1538 (a) (1) (B). The ESA defines "take" to include "harm" to animals. Id. § 1532 (19). The Service has defined "harm," within the meaning of "take," to include "significant habitat modification or degradation where it actually kills or injures wildlife," a definition that has been 9 upheld by the Supreme Court. 50 C.F.R. § 17.3 (2004); Babbitt v. 10 Sweet Home Chapter for Cmtys. for a Great Or., 515 U.S. 687, 696, 11 115 S.Ct. 2407 (1995).
12 13
14 15
16 17
18 19
20 21
22 23
24 25
26 The broad scope of Section 9 is limited by several
exceptions, found in Section 10. Specifically, Section 10 authorizes the Secretary to issue a permit, an ITP, for any taking that is incidental to the carrying out of an otherwise lawful activity. 16 U.S.C. § 1539(a) (1) (B). To receive an ITP, the permit applicant must submit an HCP that specifies: (i) the impact which will likely result from such taking; (ii) what steps the applicant will take to minimize and mitigate such impacts, and the funding that will be available to implement such steps; (iii) what alternative actions to such taking that the applicant considered and the reasons why such alternatives were not selected; and (iv) such other measures that the Secretary may require as necessary or appropriate for the purposes of the plan.

Id. § 1539(a) (2) (A); 50 C.F.R. § 17.22 (2004). The Secretary must issue an ITP upon finding that: (i) the taking will be 14 1
2 3
4 5
6 7
8 9
10 11
12 13
14 15
16 17
18 19
20 21
22 23
24 25
26 incidental; (ii) the applicant will, to the maximum extent practicable, minimize and mitigate the impacts of the taking; (iii) the applicant has ensured adequate funding for the HCP; (iv) the taking will not appreciably reduce the likelihood of the survival and recovery of the species in the wild; and (v) any additional measures required by the Secretary will be undertaken. 16 U.S.C. § 1539(a) (2) (B).

Section 7 of the ESA applies to federal actions, and
requires federal agencies, through consultation with the Service, "to insure that any action authorized, funded, or carried out" by the agency is "not likely to jeopardize the continued existence of any endangered species or threatened species."
1536 (a) (2). Issuance of an ITP is an agency action that requires the Service to engage in internal consultation and prepare a BiOp evaluating whether issuance of the ITP will result in jeopardy to any endangered or threatened species. Id. § 1536 (b). An action will result in "jeopardy" if it will "reduce appreciably the likelihood of both the survival and recovery of a listed species in the wild. ,, 50 C.F.R. § 402.02 (2004). The required jeopardy analysis under Section 7 (a) (2) is identical in almost all respects to the inquiry under Section 10 (a) (2) (B) (iv).

Sacramento City CouncilDecember 2, 2025

Sacramento City Council
December 2, 2025