Wilson Bull., 94(2), 1982, pp. 156-175
AVIAN NEST DENSITIES AND NEST-SITE
SELECTION IN FARMSTEAD SHELTERBELTS
RICHARD H. YAHNER
Farmstead shelterbelts are man-made habitats consisting of parallel rows of trees and shrubs designed to protect farmsteads from damaging winds, drifting snow and blowing soil, and to moderate the effects of wind, thereby reducing heat loss from humans, domestic animals, and farm buildings (Smith and Scholten 1980). These habitats may be viewed as small (generally <l ha) “wooded islands” surrounded by extensive fields of crops, pastures and natural prairies; however, despite their small size and isolation from other wooded habitats, shelterbelts are a source of food and song perches as well as roosting and nesting sites for a number of bird species (Orendurff 1941, Martin 1978). A vian communities in shelterbelts have been examined during the breeding season (Weiser and Hlavinka 1956; Cassel et al. 1966, 1967; Emmerich 1978; Martin 1978). However, with the exception of studies dealing with Mourning Doves (Zen&&z ma- crouru) (see Nelson 1976 for review), Ring-necked Pheasants (Phasianus colchicus) (e.g., Olson and Flake 1975) and Ferruginous Hawks (Buteo regalis) (Lokemoen and Duebbert 1976), little is known about the nesting ecology of many bird species breeding in shelterbelts. A study by Field (1971) represents the only published report giving the total number of nests per bird species. Her study dealt only with a portion of one shelterbelt where tree/shrub species used and height above ground of 75 nests of four bird species were noted. As part of an investigation of breeding-bird di- versity in 69 shelterbelts, Martin (1978) examined 185 nests of eight species and noted tree/shrub species used, height above ground, lateral distance of nest from main stem, and distance of nest from top of tree/shrub. How- ever, his study did not attempt to locate total nests per shelterbelt nor total nests per season.
A determination of both the number of avian species nesting in farm- stead shelterbelts and the abundance of nests per species is an initial and important step in assessing whether or not these small, narrow, agricul- tural habitats are valuable nesting areas or instead perhaps function to attract predators, hence reducing fledging success (see Gates and Gysel 1978). Moreover, because farmstead shelterbelts are restricted in size, the availability of nest-sites conceivably could be a limiting resource to birds.
Thus, a knowledge of factors critical to nest-site selection would give a better understanding of how coexisting nesting species effectively exploit these “wooded islands” (after Pianka 1973, Schoener 1974). 156
Y&w. AVIAN NESTING ECOLOGY IN SHELTERBELTS 157
In this paper, I examine the suitability of farmstead shelterbelts as nesting habitats for birds by estimating the annual density of total nests per bird species. Further, I quantify patterns of nest-site selection in com- mon nesting bird species based on two components: (1) types of woody species used as nesting substrate and (2) structural characteristics of nest- sites. Results obtained by this study may provide valuable insight into avian nesting requirements so that effective procedures for future planting and maintenance of farmstead shelterbelts can be formulated to benefit nesting avifauna. Such information would be particularly helpful in the intensively-farmed areas of the Great Plains where wooded habitats are at a premium, comprising less than 3% of the total area in this region (Griffith 1976).
STUDY AREA AND METHODS
The study was conducted at the University of Minnesota’s Rosemount Agricultural Ex- periment Station, Dakota Co., Minnesota. Seven shelterbelts were studied ranging in size from 0.21479 ha and in age from 4-33 years. The number of rows of trees and shrubs per shelterbelt varied from 3-9; 30 species of coniferous and deciduous trees and shrubs were planted by landowners. Shelterbelts and land-use practices of the Experiment Station are representative of those in this agricultural region.
Active and inactive nests were located by systematically searching the ground level and all trees and shrubs in each shelterbelt; searches were conducted at least once per week from late June to late September in 1978 and from early March to late September in 1979.
If eggs or young were found in a nest that was known to have been constructed earlier in the year (beginning after late March) and subsequently either abandoned or used to rear young by the same or different species, the nest was considered as another nest. Thus, a nest (or nest-site) is defined to include new nests, nests used for renesting, and those used by one species but constructed by another. Nests were often detected by observations of the activ- ities of breeding adults.
Nests constructed prior to the 1978 breeding season were not considered in the analyses and could be distinguished easily from nests of the 1978 breeding season by degree of deterioration occurring over the previous winter. Because the study did not begin until June 1978, nearly 80% of the nests found this year were no longer in use. Species that had built nests subsequently abandoned (“inactive”) were determined (estimated 95% accuracy) by the design of the nest or by the presence of shell fragments and/or feathers in the nest (Harrison 1975). The identity of nests positioned higher than 2.5 m above ground was resolved by using 7 x 35 field glasses at ground level, by climbing to the height of the nest, or by inspection with a mirror and pole device (Parker 1972). The tree and shrub used as a nesting substrate was recorded for each active and inactive nest positioned above ground level; nests on the ground were noted. In addition, 11 structural characteristics of each active and inactive nest above ground level were measured (Table 1).
Densities of nests were computed each year per individual bird species, individual shel- terbelt, total bird species and total shelterbelts. I believe that virtually all nests in 1979 were located and correctly identified. However, because the 1978 search of nests began in June, an unknown fraction of the total nests constructed in 1978 likely was not found due to destruction by predators or climatic factors. Thus, a comparison of densities between 1978 and 1979 is probably not valid. There also are other problems inherent in comparisons of 158 THE WILSON BULLETIN * Vol. 94, No. 2, June 1982
TABLE 1
ABBREVIATIONS AND DESCRIPTIONS OF 11 STRUCTURAL CHARACTERISTICS MEASURED AT
EACH ACTIVE AND INACTIVE BIRD NEST LOCATED ABOVE GROUND LEVEL IN SEVEN FARMSTEAD SHELTERBELTS IN SOUTHERN MINNESOTA IN 1978 AND 1979 Abbre-
viation Description
HTTR
HTNT
DICA
DITR
DINT
PTGC
HTGC
STLE
STMO
COMP
CLOS
Height (m) of tree or shrub in which the nest was located.
Height (m) of the nest above ground.
Diameter (m) of the canopy of the tree or shrub at the height of the nest above ground.
Diameter (cm) of the largest woody stem touching and/or supporting the nest; measured at the nest-site.
Lateral distance (m) of the nest from the main trunk of the tree or shrub containing the nest.
Percentage of herbaceous vegetation within a l-m radius of a point directly below the nest at ground level.
Maximum height (cm) of herbaceous vegetation within a l-m radius of a point directly below the nest at ground level.
Number of woody stems <2.5 cm in diameter which was within a l-m radius of the nest but exclusive of the woody stem nearest the nest (see DITR above).
Number of woody stems S2.5 cm in diameter which was within a l-m radius of the nest but exclusive of the woody stem nearest the nest (see DITR above).
Compass direction (degrees) of the nest relative to the position of the main trunk of the tree or shrub containing the nest.
Extent of closure or conspicuousness of a nest based on the amount of woody and herbaceous vegetation (estimated to the nearest 10%) in an approximate l-m radius sphere surrounding the nest-site.
nest densities as true of most field studies (see Miller and Johnson 1978). However, I feel that a comparison of approximate densities has value when comparing among bird species or habitats if done with caution (e.g., see Gates and Gysel 1978).
Nest-site selection based on differential use of woody species by all bird species combined (N = 17) or by individual bird species (N = 5 most common bird species) was determined by comparing observed versus expected number of nests per woody species using tests for goodness of fit (Sokal and Rohlf 1969). Ratios of observed to expected numbers of nests of the five common bird species in each shrub or tree (“woody”) species were used to arbitrarily rank individual nesting substrates; a ratio al.50 was used to designate a woody species as a preferred substrate. Substrate nesting breadth, NB,,, or the degree of ecological special- ization in selection of woody species for nest-sites, was derived for each of the five common bird species. NB,, was based on actual availability of each woody species in the seven shelterbelts and was obtained by the equation:
and
NB,, = (hi)“?
log Ai = (X pijlog qj - Z pulog PU)(N)~
Yahner * AVIAN NESTING ECOLOGY IN SHELTERBELTS 159
For bird species i, Ni is the total number of nests, pu is the proportion of observed nests in the j”” woody species and qr is the proportion of expected nests in the j”” woody species.
Values of NB,, could range from O-l, with 1 indicating maximum generality in use of different woody species as nesting substrates (see Petraitis 1979). Substrate nesting overlap, NO,,, was calculated between pairs of the five common bird species by the sum of the proportions of nests in common by a pair that were located in each of the woody species (modified from Yeaton 1974).
Nest-site selection based on the 11 structural characteristics of each nest-site (Table 1) was examined by stepwise discriminant analysis (BMDP7M, Dixon and Brown 1979), where g = 5 groups (common bird species), P = 11 variables (structural characteristics, Table l), and N = 589 cases (total nests of the five common bird species) in the data matrix. Discrim- inant analysis is an appropriate multivariate statistical technique to detect differences in microhabitat use among species (see Dueser and Shugart 1978, 1979). A priori probability of membership for each case in a group was based on the proportion of total cases represented in that group. Normality and homogeneity of variances for the variables were ensured by appropriate transformations (Green 1979, Johnson 1981), and each were tested by graphical plots of observations or by F,,,,,- tests, respectively (Sokal and Rohlf 1969, Johnson 1981).
Characteristics were entered into the discriminant functions in decreasing order of their importance in discriminating nest-sites among bird species (after Raphael 1981). A charac- teristic was excluded from entry if the F-to-enter (statistic computed at each step from a single-classification analysis of covariance where covariates are previously entered charac- teristics) did not exceed the F-to-remove (statistic obtained for the characteristic using a single-classification ANOVA before entering any characteristic into the function) (Dixon and Brown 1979). An approximate F-statistic based on transformation of Wilks’ A was used to test for overall differences among group means based on the characteristics entered into the discriminant function; F values computed from Mahalanobis D* statistics tested between each pair of group means (Lachenbruch 1975:25, Dixon and Brown 1979). Eigenvalues, dis- criminating information and group means of orthogonal canonical variables (linear combi- nations of structural characteristics) were derived from between-group and within-group covariance matrices. The canonical variables were adjusted so that the (pooled) within-group variances were 1 and their overall mean was 0; the canonical variables then were evaluated at group means (Dixon and Brown 1979). The significance of discrimination afforded by each of the canonical variables (discriminant functions) was tested using Bartlett’s x2 approxi- mation tests based on Wilks’ A criterion computed as a function of the eigenvalues (Cooley and Lohnes 1971:249). A jackknifed validation procedure was used to reduce the bias of group classifications (each case is classified into a group based on a classification function which includes all cases except the one being classified [Dixon and Brown 19791).
For each individual case (nest), Mahalanobis D2 was derived as the distance to each group (species) mean. These distances were used to calculate a mean Mahalanobis D2 for each of the five common bird species. Structural nesting breadth, NB,,, then was estimated by obtaining a coefficient of variation based on mean values of Mahalanobis DZ for each of the five species (modified from Dueser and Shugart 1979). Th e coefficient of variation is a proper measure of variation when comparing sample means that are known to be statistically dif- ferent (Sokal and Rohlf 1%9) and can be an approximation of breadth in resource use when using discriminant analysis (Dueser and Shugart 1979). Structural nesting overlap, NO,,, of the five common bird species was determined by a plot of the two principal canonical vari- ables (Lachenbruch 1975) and the principal axes technique (see Sokal and Rohlf 1969, Dueser and Shugart 1979, Williams 1981). A plot of 95% confidence ellipses for species’ cases was made rather than a plot using ellipses for species means. NO,, between two species then was computed as the ratio of the area of overlap for the two species relative to the cumulative 160 THE WILSON BULLETIN - Vol. 94, No. 2, June 1982
area, for that pair of species (after Dueser and Shugart 1979). Thus, NB,, was obtained by simultaneously considering all canonical variables extracted, whereas NO,, was based only on the two canonical variables giving the most discriminating information among bird species.
RESULTS
Nest densities.-A total of 617 nests was found in seven shelterbelts; 272 (81.9 nests/ha) in 1978 and 345 (103.9 nests/ha) in 1979 (Table 2).
Densities of nests among individual shelterbelts in both years averaged
88.5 nests/ha and ranged from 28.8 nests/ha in a three-row, 5-year-old coniferous shelterbelt in 1978 to 186.4 nests/ha in a five-row, 32-year-old coniferous-deciduous shelterbelt in 1979.
Nine of 17 bird species nested in the shelterbelts in both 1978 and 1979, whereas eight additional species nested in either 1978 or 1979 only (Table 2). Nests of Common Grackle (Quiscalus quiscula), Mourning Dove, Amer- ican Robin (Turdus migratorius), Gray Catbird (Dumatella carolinensis) and Chipping Sparrow (Spizella passerina) comprised 95.5% (N = 589) of the total number of nests found in both years combined. Variation in density of nests among individual shelterbelts within or between years often was attributed to the presence of large numbers of grackles. For example, density of grackle nests nearly doubled from 1978 (36.4 nests/ ha) to 1979 (64.5 nests/ha); nests of this species occurred in all shelterbelts in 1979 versus five of seven shelterbelts in 1978. Ring-necked Pheasant, Downy Woodpecker (Picoides pubescens), Barn Swallow (Hirundo rustica), Black-capped Chickadee (Parus atricupillus), American Goldfinch (Car- duelis tristis) and Vesper Sparrow (Pooecetes gramineus) were observed in one or more shelterbelts during the breeding season of both years (see Yahner 1980a) but nested in other habitats at the Rosemount Station.
Selection of substrate.-A total of 3589 woody plants of 31 species was available as substrate for nests in the seven shelterbelts (Table 3). Num- bers of observed nests for all bird species combined were considerably different from those expected in several abundant (N > 30 individual plants) coniferous and deciduous species (P < 0.001). Preferred conifer- ous species included Colorado blue spruce (Picea pungens), white spruce (P. glauca) and Black Hills spruce (P. g. densata), whereas infrequently used conifers were white cedar (Thuju occidentalis), red pine (Pinus re- sinosa) and Douglas fir (Pseudotsugu menziesii). Deciduous species favored for nest-sites were eastern cottonwood (Populus deltoides), silver maple (Acer saccharinurn) and box elder (A. negundo). Fewer nests than expected were found in various deciduous trees and shrubs, such as green ash (Fraxinus pennsylvanica), caragana (Caragana arborescens) and amur ma- ple (A. ginnala). In contrast, numbers of observed compared to expected nests were nearly equal in ponderosa pine (P. ponderosa) and tartarian
162 THE WILSON BULLETIN - Vol. 94, No. 2, June 1982
TABLE 3
ABUNDANCE OF AND USE OF 31 WOODY SPECIES AS NESTING SUBSTRATE BY 17 BIRD SPECIES IN SEVEN FARMSTEAD SHELTERBELTS (3.32 HA) IN SOUTHERN MINNESOTA IN 1978 AND 1979
Woody species
Total Observed
no. plants
Expected Proportion total
no. nest8 “0. nestsa observed “eats
Colorado blue spruce (Piceu pungens)
White spruce (Piceu gluucu)
Cottonwood (Populus deltoides)
Silver maple (Acer saccharinurn)
Ponderosa pine (Pinus ponderosa)
Box elder (Acer negundo)
Black Hills spruce (Piceu ghca densutu)
Green ash (Fruxinus pennsylvunicu)
Tartarian honeysuckle (Lo&era taturicu)
White cedar (Thuju occident&s)
Jack pine (Pinus bunksiunu)
Caragana (Curugunu urborescens)
Amur maple (Acer ginnulu)
Norway spruce (Piceu &es)
Red pine (Pinus resinosu)
American elm (Ulmus americana)
Ural willow (Sulix purpureu)
Buffalo berry (Shepherdiu urgenteu)
Flowering crab (M&s jloribundu)
Common lilac (S+zgu vulgaris)
Russian olive (Elueugnus ungustifoliu)
Red-berried elder (Sumb~cus pubens)b
High-bush cranberry (Viiburnum trilobum)
River birch (Bet& nigru)
Honey locust (Gleditsiu tricuncunthos)
Chokecherry (Prunus virginiana)
Southern arrowwood (Viburnum dentutum)
Gray dogwood (Cornus rucemosu)
Red-osier dogwood (Cornus stoloniferu)
Laurel-leaved willow (S&x pentundra)
Douglas fir (Pseudotsugu menziesii)
MiscellaneousC
Total
113
242
201
197
220
33 37
421
111
195
72 423
153
88 171
112
220
17 7
251
25 10
9 1
99 11
24 18
10 9
89 -
3589
125
124
60 50
38 35
30 20
19 16
14 13
12 11
10 9
5 4
3 3
2 2
2 1
1 1
1 0
0 0
0 6
617
19.4 41.6
34.6 33.9
37.8 5.7
6.4 72.4
19.1 33.5
12.4 72.7
26.3 15.1
29.4 19.3
37.8 2.9
1.2 43.2
4.3 1.7
1.5 0.2
17.0 1.9
4.1 3.1
1.7 1.5
15.3 -
-
20.3 20.1
9.9 8.1
6.6 5.7
4.9 3.2
3.1 2.6
2.3 2.1
1.9 1.8
1.6 1.5
0.8 0.6
0.5 0.5
0.3 0.3
0.3 0.2
0.2 0.2
0.2 0.0
0.0 0.0
0.0 1.0
-
’ Expected number of nests WBS determined from the total number of plants and the total number of observed nests.
Expected number of nests is significantly different from observed number of nests k’ = 1227, df = 30, P < 0.001. Test for goodness of tit, Sokal and Rohlf (1%9).
b Woody species which colonized shelterbelts.
’ Refers te nests found on ground or in log and brush piles, Yahner . AVIAN NESTING ECOLOGY IN SHELTERBELTS 163
TABLE 4
TOTAL NUMBER OF OBSERVED NESTS (AND EXPECTED NESTS) OF FIVE COMMON BIRD SPECIES IN 26 WOODY SPECIES USED AS NESTING SUBSTRATE IN SEVEN FARMSTEAD SHELTERBELTS IN SOUTHERN MINNESOTA IN 1978 AND 1979
Total number of nests/bird speciesb
Woody species’
Common
GFl&le
Mourning American
DOE
Gray
Robin Catbird
Chipping
SParrOW
Colorado blue spruce 57 (10.55)2 35 (4.28)’
White spruce 56 (22.59)5 48 (9. 17)3
Cottonwood 50 (18.76)4 8 (7.62)
Silver maple 40 (18.39)6 5 (7.47)
Ponderosa pine 37 (20.53)’ 0 (8.34)
Box elder 29 (3.08)’ 0 (1.25)
Black Hills spruce 11 (3.45)3 9 (1.40)’
Green ash 10 (39.30) 4 (15.95)
Tartarian honeysuckle 0 (10.36) 3 (4.21)
White cedar 7 (18.20) 3 (7.39)
Jack pine 7 (6.72) 4 (2.73)
C aragana 12 (39.48) 0 (16.03)
Amur maple 0 (14.28) 10 (5.80)4
Norway spruce 4 (8.21) 0 (3.33)
Red pine 9 (15.96) 0 (6.48)
American elm 3 (10.45) 1 (4.24)
Ural willow 0 (20.53) 2 (8.34)
Buffalo berry 0 (1.59) 1 (0.64)
Flowering crab 1 (0.65) 0 (0.27)
Common lilac 2 (23.42) 0 (9.51)
Russian olive 0 (2.33) 1 (0.95)
Red-berried elder 0 (0.93) 0 (0.38)
High-bush cranberry 0 (0.84) 0 (0.34)
River birch 0 (0.09) 1 (0.04)
Honey locust 0 (9.24) 1 (3.75)
Chokecherry 0 (1.02) 0 (0.42)
No. woody species used 16
Total no. observed nests 335
Gstatisticse 782
Level of significance
Substrate nesting
breadth, NBsud
P < 0.001
16 16 8 4
136 87 17 14
314 174 35 39
P < 0.001 P < 0.001 P > 0.10 P > 0.01
0.44 0.31
25 (2.74)’
19 (5.87)6
3 (4.87)
4 (4.78)
1 (5.33)
4 (0.80)4
8 (0.90)3
6 (10.21)
2 (2.70)
5 (4.73)
0 (1.75)
0 (10.25)
2 (3.71)
4 (2.13)’
0 (4.15)
1 (2.71)
0 (5.33)
0 (0.41)
2 (0.17)’
0 (6.08)
1 (0.61)
0 (0.24)
0 (0.22)
0 (0.02)
0 (2.40)
1 (0.27)5
0.31 1 (o.54)7
0 (1.15)
0 (0.95)
0 (0.93)
0 (1.04)
0 (0.16)
1 (0.18)5
0 (1.99)
8 (0.53)4
0 (0.92)
0 (0.34)
0 (2.00)
0 (0.72)
0 (0.42)
0 (0.81)
1 (0.53)6
0 (1.04)
2 (0.08)2
0 (0.03)
l(l.19)
0 (0.12)
2 (0.05)’
1 (0.04)3
0 (0.00)
0 (0.47)
0 (0.05)
0.09 8 (0.44)’
0 (0.94)
0 (0.78)
0 (0.77)
0 (0.86)
0 (0.13)
1 (0.14)2
0 (1.64)
3 (0.43)3
0 (0.76)
0 (0.28)
0 (1.65)
0 (0.60)
2 (0.34)4
0 (0.67)
0 (0.44)
0 (0.86)
0 (0.07)
0 (0.03)
0 (0.98)
0 (0.10)
0 (0.01)
0 (0.04)
0 (0.00)
0 (0.39)
0 (0.04)
0.08 ’ Scientific names of woody species and total number of individual plants for each are given in Table 3. Woody species not used by at least one of five common bird species are excluded from table. ’ Rank of each woody species BS nesting substrate per bird species is based on the ratio of observed to expected number of nests. The greater the ratio, the higher the rank 88 indicated by a lower numeric subscript. A ratio 21.50 arbitrarily designated a preference as a nesting substrate; B ratio 41.50 was excluded from ranking. ’ Comparison of observed to expected numbers of nests per bird species (df = 30). Test for goodness of fit, Sokal and Rohlf (1969).
d See text for derivation.
164 THE WILSON BULLETIN - Vol. 94, No. 2, June 1982
TABLE 5
SUBSTRATE NESTING OVERLAP, NOsu, BETWEEN PAIRS OF FIVE COMMON BIRD SPECIES NESTING IN SEVEN FARMSTEAD SHELTERBELTS IN SOUTHERN MINNESOTA IN 1978 AND 1979
Bird species Gr NO... (%)”
Common Grackle-Mourning Dove 55
Common Grackle-American Robin 59
Common Grackle-Gray Catbird 11
Common Grackle-Chipping Sparrow 22
Mourning Dove-American Robin 78
Mourning Dove-Gray Catbird 15
Mourning Dove-Chipping Sparrow 35
American Robin-Gray Catbird 15
American Robin-Chipping Sparrow 43
Gray Catbird-Chipping Sparrow 33
’ NO,. (%) is calculated as the aurn of proportions of nests (Table 4) in ~~rnrn~n by a pair in each woody species used as a nesting substrate.
honeysuckle (Loniceru tatarica). Three of the five common nesting bird species showed a significant difference (P < 0.001) in observed versus expected use of nesting substrate (Table 4). All five common bird species preferred spruces (Picea spp.), yet each also displayed specific prefer- ences for other woody species.
The five common bird species varied in substrate nesting breadth, NB,, (Table 4). NB,, was highest in grackles, intermediate in doves and robins, and lowest in catbirds and sparrows. The greatest amount of substrate nesting overlap, NO,,, occurred with doves and robins (780/o, Table 5).
These two species exhibited similar NO,, with grackles (55-59%). In con- trast, catbirds had little joint use of nesting substrate with doves (15%), robins (15%) and grackles (11%). NO,, between sparrows and each of the four other species ranged from 22-43%.
Selection of structural characteristics.-Stepwise discriminant analysis based on 11 structural characteristics (Table 1) measured at nest-sites (N = 589) indicated that group means among the five common bird species were not equal (F = 25.11; df = 32, 2129; P < 0.001). In addition, the group mean of each bird species was significantly different from those of the other species in paired comparisons (F’s = 3.73-53.78; df’s = 8, 22- 462; P’s < 0.001). Segregation of nest-sites in discriminant space (after Dueser and Shugart 1979, Raphael 1981) was due only to 8 of 11 structural characteristics (Table 6); of the 11 characteristics, HTTR, DITR and COMP (see Table 1 for abbreviations and descriptions) were not selected for entry into the discriminant function equations. The percentage of total Y&w. AVIAN NESTING ECOLOGY IN SHELTERBELTS 165
TABLE 6
MEANS AND STANDARD DEVIATIONS OF 11 STRUCTURAL CHARACTERISTICS MEASURED AT
NEST-SITES OF FIVE COMMON BIRD SPECIES NESTING IN SEVEN FARMSTEAD SHELTER- BELTS IN SOUTHERN MINNESOTA IN 1978 AND 1979
Char- Common Grackle American Robin
acteristic”
Mourning Dove Gray Catbird Chipping Sparrow
(N = 335) (N = 136) (N = 87) (N = 17) (N = 14)
HTNT* 6.27 ~fr 3.51”“” 2.26 ? 1.40g 2.46 2 1.888 1.26 ? 0.64g 1.08 ? 0.66g DICA* 2.43 2 1.79d” 3.15 2 1.14g 2.93 ? 1.57g 3.01 2 1.18g 2.74 + 1.33 DINT* 0.23 2 0.62ds 0.51 2 0.62=” 0.27 2 0.50des 0.12 2 0.30d” 0.63 ? 0.66”” HTGC* 0.30 ? 0.32d’cs 0.39 ? 0.35’=” 0.39 2 0.35=” 1.05 2 0.78gd” 0.60 ? 0.54gd’” PTGC* 44.8 ? 38.3”” 47.6 ? 36.2’“” 64.3 ? 37.9=’ 77.9 ? 32.2gd 79.3 & 30.8pd STLE* 0.36 2 1.98” 0.65 ? 2.09”” 0.31 2 1.41= 9.88 2 5.55gd” 2.79 + 4.97pd’C STMO* 0.14 * 0.74c 0.15 ? 0.58’ 0.16 ? 0.57c 0.53 2 0.62pd” 0.07 ? 0.27” cLos* 39.3 ? 27.8d’CS 56.6 ? 26.2= 63.7 2 25.1gdcS 96.5 ? 4.9gd’ 94.2 2 5.4pd’ HTTR*” 10.07 ? 5.48”“” 7.16 ? 3.81=” 6.92 2 4.29’= 3.34 ? 1.84gd’ 4.26 * 2.87gd DITR*” 0.08 2 0.06d’CS 0.05 2 0.05= 0.06 2 O.Wdc” 0.03 ? 0.03g’ 0.02 * 0.029’ COMP” 218.7 ? 100.8 228.6 ? 88.2 225.9 2 76.5 233.1 2 84.6 225.0 r 58.5 = N = 589.
b Characteristics BE listed in order of decreasing importance in discriminating among hird species in stepwise discrim- inant function. Characteristics indicated by an asterisk me significantly different among species (P < 0.05, single-classi- fication ANOVA, Sokal and Rohlf [1%91); superscript n indicates not included in the discriminant function equations (see text). Superscripts, g, d, r, c and s indicate that a mean for a particular characteristic is different (P < 0.05, Student- Newman-Keuls test) from mean of Common Grackle, Mourning Dove, American Robin, Gray Catbird, OI Chipping Sparrow, respectively. Descriptions and abbreviations of characteristics BE from Table 1. nest-sites accurately classified to bird species averaged 51.1% and ranged from 0.0% in sparrows to 88.1% in grackles (Table 7).
Stepwise discriminant analysis extracted four canonical variables (CV) or linear combinations of the 11 structural characteristics measured at TABLE 7
JACKKNIFED CLASSIFICATION MATRIX GIVING THE PERCENTAGE OF NESTS CORRECTLY
CLASSIFIED TO BIRD SPECIES AND THE NUMBERS OF NESTS CATEGORIZED INTO EACH OF THE FIVE COMMON BIRD SPECIES BASED ON 11 STRUCTURAL CHARACTERISTICS MEASURED AT 589 NEST-SITES
Number of nests classified/bird species
Bird species
Percentage
colTect
Common
Grackle “E?g
American Gray
Robin Catbird
Chipping
sparrow
Common Grackle 88.1 295 19 17 4 0
Mourning Dove 62.5 28 85 18 5 0
American Robin 34.5 22 33 30 2 0
Gray Catbird 70.6 2 1 2 12 0
Chipping Sparrow 0.0 0 4 7 3 0
Total - 347 142 74 26 0
166 THE WILSON BULLETIN. Vol. 94, No. 2, June 1982
TABLE 8
EIGENVALUES, DISCRIMINATING INFORMATION, GROUP (BIRD SPECIES) CANONICAL COR-
RELATION COEFFICIENTS AND STANDARDIZED COEFFICIENTS OF EIGHT STRUCTURAL CHAR- ACTERISTICS FOR CANONICAL VARIABLES 1 AND 2 EXTRACTED FROM STEPWISE DISCRIMI-
NANT ANALYSIS OF BETWEEN-SPECIES AND WITHIN-SPECIES COVARIANCE MATRICES OF 11 STRUCTURAL CHARACTERISTICS MEASURED AT NEST-SITES (N = 589) OF FIVE COMMON
BIRD SPECIES
Canonical variable”
1 2
Eigenvalue
Percentage of discriminating information
Bird species
Common Grackle
Mourning Dove
American Robin
Gray Catbird
Chipping Sparrow
Structural characteristicb
HTNT
DICA
DINT
HTGC
PTGC
STLE
STMO
CLOS
1.038 0.474
65.1 29.7
-0.75 -0.30
0.71 0.67
0.72 0.69
3.73 -2.90
2.17 -0.02
-0.23 -0.25
0.24 0.12
0.16 0.29
0.75 -0.09
0.01 0.00
0.19 -0.40
-0.27 0.43
0.01 0.00
a canonial variables 1 and 2 are both significant (P 4 0.001) based on Wilks’ A criterion computed as a function of the eigenvalues using Bartlett’s x2 approximation (Cooley and Lohnes 1971:249). Wilks’ A = 0.31 and 0.63; ,yz = 685 and 272; df = 44 and 30, respectively.
h Descriptions and abbreviations of structural characteristics are from Table 1. Characteristics not included in discrim- inant equations (Tahle 6) are omitted.
nest-sites, which best segregated among the five common bird species.
Only CVl and CV2 afforded a significant (P < 0.001) degree of discrimi- nation (Table 8). CVl is a “herbaceous” variable because it principally includes the structural variable HTGC, although HTNT, DICA and STMO are important also, as indicated by relatively high standardized coefficients (after Raphael 1981, Williams 1981). CVl accounts for 65.1% of the dis- criminating information that is available in the 11 characteristics measured at nest-sites to separate bird species. CV2 is termed a “woody stem” variable because it mainly associates the variables STLE and STMO. CV2 explains 29.7% of the discriminating information. CV3 and CV4 are nei- ther biologically interpretable nor statistically significant (P > 0.05), and Yahner * AVIAN NESTING ECOLOGY IN SHELTERBELTS 167
cv2
6-
American Robin
-2-
-4-
-6 - Chipping Sparrow’
-8-
,w , Gray Catbird
CVl
-10 I
,
-4 -2 0 2 4 6 8 10
FIG. 1. Plot of group means of bird species for CVl and CV2 (Table 8) and 95% confi- dence ellipses for the species’ cases (nests) on CVl and CV2 using the principal axes tech- nique (Sokal and Rohlf 1969, Dueser and Shugart 1979). Group means: G = Common Grackle, D = Mourning Dove, R = American Robin, C = Gray Catbird and S = Chipping Sparrow.
account for only 3.7 and 1.5%, respectively, of the discriminating infor- mation.
A plot of the two principal canonical variables, CVl on CV2 (Fig. l), gives the best visual and ecologically interpretable illustration of the dif- ferences in the structural characteristics of nest-sites among the five com- mon bird species. The relatively narrow, elongated ellipses of grackles, catbirds and sparrows (Fig. 1) indicate that most variance could be ex- plained by CVl (Table 8) (see Sokal and Rohlf 1969). In contrast, the more circular ellipses of robins and doves suggest that CVl and CV2 were of nearly equal importance in discriminating nest-sites of these species.
Species with group means to the right in Fig. 1 along the “herbaceous” variable, such as the catbird and the sparrow, were characterized by well- hidden nests. Nests of these two species were located close to the ground 168 THE WILSON BULLETIN - Vol. 94, No. 2, June 1982
TABLE 9
STRUCTURAL NESTING OVERLAP, NOsT, BETWEEN PAIRS OF FIVE COMMON BIRD SPECIES NESTING IN SEVEN FARMSTEAD SHELTERBELTS IN SOUTHERN MINNESOTA IN 1978 AND 1979
Bird species pair No,, (%I
Common GrackleMourning Dove 33
Common Grackle-American Robin 33
Common Grackle-Gray Catbird 9
Common Grackle-Chipping Sparrow 8
Mourning DoveAmerican Robin 69
Mourning Dove-Gray Catbird 20
Mourning Dove-Chipping Sparrow 26
American Robin-Gray Catbird 14
American Robin-Chipping Sparrow 20
Gray Catbird-Chipping Sparrow 34
a NO,, is calculated from 95% confidence ellipses for species’ cases (nests) on canonical variables 1 and 2 (Fig. 1) by the ratio of the area of overlap for two species relative to the cumulative area for that pair of species (after Dueser and Shugart 1979).
and were well concealed by herbaceous vegetation (Table 6), compared to nests of other species. On the other hand, the mean of the grackle nests was located to the left in Fig. 1 because nests were often high above ground and well exposed. Means of dove and robin nests were intermediate in height above ground and percentage of herbaceous cover.
The position of group means along the “woody stem” variable essentially separates catbird nests from those of the other four species (Fig. l), based on the density of woody stems surrounding nests (Table 6). Further, cat- bird and grackle nests tended to be located near the main stem of the nesting substrate relative to the location of nests of the other three species.
Structural nesting breadth, NBSt, or the relative variability in the dis- tance of each species’ nest from its respective group mean in 4-dimen- sional hyperspace (corresponding to four canonical variables) was greatest in grackles (NB,t = 1.97), followed by those of robins (1.89), sparrows (1.12), doves (1.08) and catbirds (0.89). As with NO,, (Table 5), structural nesting overlap, NO,t, in 2-dimensional hyperspace (Fig. 1) was greatest between doves and robins (69%, Table 9). NOSt of catbirds with sparrows and NO,t of grackles with either doves or robins were nearly equal (33- 34%). The least amount of shared use of structural characteristics of nest- sites was shown between grackles and either catbirds or sparrows (8-9%).
DISCUSSION
Suitability of shelterbelts as nesting habitat.-Farmstead shelterbelts are valuable habitats to avifauna in regions of intensive agriculture in Yahner - AVIAN NESTING ECOLOGY IN SHELTERBELTS 169
terms of both numbers of breeding species and density of nests (Table 2).
The importance of these “wooded islands” is supported perhaps by com- parisons to studies conducted in other types of habitats in the midwest region (e.g., see Skinner 1975, George et al. 1979). Studies of avifauna associated with farmstead shelterbelts report large numbers of nests for certain species, particularly grackle, dove and robin nests (see Randall 1955, LaPointe 1958, Field 1971, Martin 1978). In my study, 50.5, 20.5 and 13.1% of total nests (N = 617) were of these three species, respec- tively (Table 2).
A variety of factors may contribute to the suitability of farmstead shel- terbelts as nesting habitat to different bird species in addition to the avail- ability of preferred nesting substrate (Tables 3 and 4), and/or the micro- habitat configuration afforded by the substrate (Tables 6 and 8, Fig. 1).
One factor may include the spacing pattern used by landowners in planting trees and shrubs within and between rows of shelterbelts and its concom- itant effect on vegetative structure. For example, the distance between permanent trees in rows after thinning may be as much as 5-7 m (see Smith and Scholten 1980). This planting practice probably creates favor- able nesting areas for bird species adapted to mixed-breeding habitats, i.e., habitats with relatively open overstory canopies and/or dense cover near the ground for nest-sites (Table 6, Fig. 1) (after Gates and Gysel 1978, Stauffer and Best 1980).
The manner in which landowners maintain shelterbelts may also affect their importance to birds as nesting areas (see Read 1964, Smith and Scholten 1980 for maintenance procedures). For instance, younger shel- terbelts (~5 years) often are mowed to remove undesirable vegetation which compete for moisture with tree and shrub seedlings (George 1943).
Mowing of dense undergrowth in a 5-year-old shelterbelt may have enabled an Eastern Meadowlark (Sturnella mugrza) to nest in 1978 (Table 2); meadowlark nests are typically characteristic of open, non-wooded habi- tats (George et al. 1979). Removal of snags is a practice that likely reduces the suitability of shelterbelts for primary and secondary cavity nesters (see Hardin and Evans 1977, Martin 1978, Stauffer and Best 1980). Although Black-capped Chickadees and Downy Woodpeckers were two cavity-nest- ing species observed in shelterbelts during spring (Yahner 1980a), neither nested in shelterbelts as a probable consequence of few snags and/or small size of shelterbelts. Two known cavities were present in the seven shel- terbelts, and those were used as nest-sites by a House Sparrow (Passer domesticus) and a Starling (Sturnus vulgaris) in 1979 (Table 2).
The attractiveness of farmstead shelterbelts as nesting habitat to a par- ticular bird species conceivably depends on many factors not considered in this study. Among these are type, quantity and availability of food (after 170 THE WILSON BULLETIN - Vol. 94, No. 2, June 1982
Franzreb 1978), including food items procured in shelterbelts or in contig- uous habitats around farmsteads. For example, the probability of an American Robin selecting a shelterbelt as a nesting area may be contingent not only on the presence of Picea as nesting substrate (Table 4) but also on proximity of a preferred foraging area near the nest-site. Perhaps a mowed lawn, used as a foraging site and adjacent to the shelterbelt, con- tains sufficient food resources that were otherwise unavailable or in low abundance within the shelterbelt for raising a brood of robins. Another factor determining use of shelterbelts may be the ability of a bird species to tolerate disturbances associated with farm operations (after Emlen 1974, Cooke 1980). Shelterbelts are usually planted within 30-35 m of homes and buildings (Smith and Scholten 1980), and the disturbances created by humans, domestic animals and farm machinery may disrupt nesting ac- tivities of some species more than others. Two additional factors may be that some shelterbelts are narrower than the minimum width of a habitat required by certain breeding species (Stauffer and Best 1980), or that a less aggressive species benefits by the presence of a coexisting, more aggressive species (e.g., catbird) which effectively defends nesting areas against predation and parasitism (Clark and Robertson 1979). Finally, the presence of mammalian nest predators, such as red squirrels (Z’umiasciu- rus hudsonicus), in certain types of shelterbelts (Nelson 1976, Yahner 1980b) may influence the overall suitability of these habitats as nesting areas.
Nest-site selection: substrate and structural considerations.-Farmstead shelterbelts represent an excellent habitat in which to examine patterns of nest-site selection for two reasons. First, the number of suitable nest- sites per bird species is presumably restricted by the limited number of trees and shrubs available (N = 3589 or 1089 plants/ha, this study). Sec- ond, nesting substrate among shelterbelts differed with regard to plant species (N = 31) as well as physiognomy.
The five common bird species preferred certain nesting substrate; use of specific woody species for nest-sites was not proportional to the actual number of individual plants per tree or shrub species in the seven shel- terbelts (Table 4). These results agree with other studies dealing with doves and grackles (e.g., McClure 1946, LaPointe 1958, Field 1971). More- over, grackle, dove and robin nests were each found in 16 different woody species (Table l), but the relative proportions of total nests per woody species were less variable in grackles compared to doves and robins. As a consequence, NB,, was highest in the grackle (Table 4). Using NB,, as an inverse index of specialization (Levins 1968, Dueser and Shugart 1979) in use of substrate, the grackle may be classified as a generalist or op- portunist along a generalist-specialist continuum. At the opposite pole of Y&w - AVIAN NESTING ECOLOGY IN SHELTERBELTS 171
the NB,, continuum are specialists (low NB,,), the catbird and the sparrow.
The dove and the robin are positioned toward the middle of the continuum with intermediate values of NB,,.
Variability in location of nest-sites, based on structural characteristics (Table 6) and expressed by values of NBSt, may be viewed as a generalist- specialist continuum in discriminant 4-space (corresponding to canonical variable l-4). Again, the grackle, followed by the robin, was less stereo- typed in nest-site selection compared to the other three common species.
Thus, both measures of nesting breadth were broad in the grackle and the robin but varied from intermediate to narrow with the other species, de- pending on the measure of breadth. For example, the sparrow was very specific in its choice of nesting substrate (NB,, = 0.08, Table 4) but was more flexible in selecting the location of nests relative to structural fea- tures (NB,t = 1.12) when compared to the dove or the catbird. Conversely, the dove was capable of using many woody plants (NB,, = 0.31) yet was stereotypic in terms of structural nesting breadth (NBSt = 1.08). In short, NB,, was not necessarily related to NBSt in comparisons within or among common bird species.
Because of the spatially-limited environment of shelterbelts, a certain amount of overlap in both use of specific nesting substrate and structural characteristics surrounding nest-sites is plausible among bird species (af- ter Raphael 1981). NO,, and NOSt was greatest between doves and robins (Tables 5 and 9), partially the result of the use of robin nests by doves (e.g., McClure 1946, Weeks 1980, pers. obs.). Further, nests of these two species were often indistinguishable (Table 7) on the basis of high NOSt (Table 9) and similar mean values for 8 of the 11 structural characteristics (Table 6). On the other extreme, ecological requirements of some species in choice of nest-sites were so different (Tables 4 and 6) that little overlap in substrate use of structural characteristics was expected. For example, NO,, and NO,t were only 11 and %, respectively, between grackles and catbirds (Tables 5 and 9). Only 4 (1.2%) of 335 grackle nests were inac- curately classified as catbird nests (Table 7). Although the group mean of sparrows was segregated from means of each of the other species (P < O.OOl), all nests (N = 14) were classified as being those of dove, robin or catbird (Table 7). This may be partially due to moderate amounts of cumulative overlap by sparrows with these three species in selection of microhabitat for total nest-sites (Table 6 and 9). Alternatively, a structural characteristic(s) not considered in this study may have given better dis- crimination between nest-sites of sparrows and those of the other three species, perhaps giving better accuracy in classification (see Raphael 1981 for additional discussion). Perhaps sample size was a factor in classifica- tion, yet 70.6% of catbird nests (N = 17) were categorized accurately (Ta- 172 THE WILSON BULLETIN - Vol. 94, No. 2, June 1982
ble 7). Yabner (in press) has shown that abundance of this sparrow in both spring and summer is correlated (P < 0.05) to the extent of the perimeter of shelterbelts; thus, “edge” of shelterbelts or nesting substrate (NB,, =
0.08, Table 4) may be relatively more important than structural features (Table 6) in choice of nest-sites by this species. In short, these small “wooded islands” were sufficiently heterogeneous to allow the five com- mon bird species to select distinct nest-sites (Table 8, Fig. l), thereby permitting high densities of nests compared to non-wooded habitats in the midwest. If more data were available to quantify patterns of nest-site selection by the 12 other species known to nest in these seven shelterbelts (Table 2), comparisons could be made to test whether or not the five common species were more abundant because of more generalized eco- logical requirements for nests.
Patterns of nest-site selection: management implications.-A knowledge of patterns of nest-site selection by bird species in shelterbelts may be helpful in formulating effective decisions pertaining to the future design of these farmstead habitats. For instance, based on NB,, (Table 4) and NBst the grackle is extremely opportunistic in selection of nest-sites. This species appears capable of exploiting a variety of tree and shrub species (Table 4), with perhaps the only requirement being that the shelterbelt have some tall trees for nest-sites well-above ground level (Table 6, Fig. 1). Because nesting overlap between doves and robins was relatively ex- tensive (Tables 5 and 9), management decisions for the design and the maintenance of shelterbelts may have similar effects on these two species.
Stauffer and Best (1980), for example, concluded that partial removal of woody canopy in closed-canopy riparian woodlands would benefit both doves and robins. Thus, the partially-opened canopy that often occurs in farmstead shelterbelts due to spacing and thinning of trees and shrubs likely has similar effects on both of these species. But when spruce (Picea spp.), the preferred nesting substrate for doves and robins (Table 4), are planted close together (e.g., 3-4 m), the lower branches typically lose needles if contact is made with contiguous trees or shrubs due to shading (Smith and Scholten 1980). Doves and robins rarely constructed nests in lower (<3 m above ground), defoliated branches of spruce, perhaps be- cause nest-sites of both species required a moderate amount of conceal- ment (Table 6, Fig. 1). Loss of needles in lower branches also can reduce the effectiveness of spruce as wind barriers. Thus, an approximate spacing of 5-6 m between adjacent trees may help retain foliage, thereby serving a dual function of providing a long-term wind barrier and a nesting sub- strate for two common bird species in shelterbelts (H. Scholten, pers. comm.).
Catbirds, and to some extent sparrows, exhibited relatively narrow nest- Yuhner * AVIAN NESTING ECOLOGY IN SHELTERBELTS 173
ing breadths and reduced overlap with those of other species (Tables 4, 5 and 9; Fig. 1). Because of stereotypic nest-site selection, an obvious management recommendation is to provide preferred nesting substrate (Table 4). Moreover, nests of these two species (particularly catbirds) were well concealed but close to ground level (Table 6, Fig. 1). Therefore, re- moval of herbaceous or woody vegetation (i.e., via mowing) should be restricted to areas between rows and attempts should be made to retain vegetation within a 0.5-m radius of branches of preferred shrubs (e.g., tartarian honeysuckle) or trees (e.g., Colorado blue spruce) (Table 4).
SUMMARY
Avian nest densities and nest-site selection were examined in seven Minnesota farmstead shelterbelts for two years. A total of 617 nests belonging to 17 bird species was noted.
Suitability of shelterbelts as nesting habitats was determined by nest densities which were greater than those reported for non-wooded habitats in the region. Choice of nesting substrate by the total avian community or by individual bird species was not random. Nests-sites were segregated among the five most common bird species based primarily on two linear combi- nations of 11 structural characteristics termed as “herbaceous” or “woody stem” variables.
The five common bird species were positioned along a generalist-specialist continuum in selection of nest-sites using two measures of nesting breadth. Measures of nesting overlap varied considerably between pairs of species.
Farmstead shelterbelts, despite their small size, are important nesting habitats for birds.
These “wooded islands” are sufficiently heterogenous to permit the coexistence of several nesting species at high densities, Information about nest-site choice should be used by designers of shelterbelts for the benefit of certain breeding birds in the midwest.
ACKNOWLEDGMENTS
I am grateful to H. Scholten of the College of Forestry, University of Minnesota, and C.
Wilcox of the Rosemount Agricultural Experiment Station for their cooperation with all aspects of the study. I also thank M. Weller and J. Cooper of the Department of Entomology, Fisheries and Wildlife, University of Minnesota, and H. Scholten for comments on an earlier draft of the manuscript. Several helpful comments were given also by J. C. Barlow, R. F.
Johnston and J. D. Rising. This study was supported by the Minnesota Agricultural Exper- iment Station, and by grants from the Graduate School of the University of Minnesota, the National Rifle Association of America and the Max McGraw Wildlife Foundation. This paper is No. 11,196 of the Minnesota Agricultural Experiment Station, Scientific Journal Series, University of Minnesota, St. Paul.
LITERATURE CITED
CASSEL, J. F., W. E. MOORE, L. STRAND, H. VOIGT, R. KROODSMA AND J. LARRY. 1966.
Breeding birds of some shelterbelts in Barnes County, North Dakota. Audubon Field Notes 20:649-655.
-, W. BURESH, J. CASSEL, D. FORTNEY AND G. MOEN. 1%7. Breeding birds of some North Dakota shelterbelt~II1. Audubon Field Notes 21:650-656.
CLARK, K. L. AND R. J. ROBERTSON. 1979. Spatial and temporal multi-species aggregations in birds as anti-parasite and anti-predator defenses. Behav. Ecol. Sociobiol. 5:359371. 174 THE WILSON BULLETIN * Vol. 94, No. 2, June 1982
COOKE, A. S. 1980. Observations on how close certain passerine species will tolerate an approaching human in rural and suburban areas. Biol. Conserv. 18:85-88.
COOLEY, W. W. AND P. R. LOHNES. 1971. Multivariate data analysis. John Wiley and Sons, New York, New York.
DIXON, W. J. AND M. B. BROWN. 1979. B iomedical computer programs P-series. Univ.
California Press, Berkeley, California.
DUESER, R. D. 1979. Niche pattern in a forest-floor small mammal fauna. Ecology60: 108-118.
AND H. H. SHUGART, JR. 1978. Microhabitats in a forest-floor small mammal fauna.
Ecology 59:89-98.
EMLEN, J. T. 1974. An urban bird community in Tucson, Arizona: derivation, structure, regulation. Condor 76:184-197.
EMMERICH, J. M. 1978. Bird utilization of woodland habitat in the eastern quarter of South Dakota. M.S. thesis. South Dakota State Univ., Brookings, South Dakota.
FIELD, N. H. 1971. Use of an eastern South Dakota shelterbelt by nesting birds. South Dakota Bird Notes 23:43-45.
FRANZREB, K. E. 1978. Tree species used by birds in logged and unlogged mixed-coniferous forests. Wilson Bull. 90:221-238.
GATES, J. E. AND L. W. GYSEL. 1978. Avian nest dispersion and fledging success in field- forest ecotones. Ecology 59:871-883.
GEORGE, E. J. 1943. Effects of cultivation and number of rows on survival and growth of trees in farm windbreaks on the northern Great Plains. J. Forest. 41:820-828.
GEORGE, R. R., A. L. FARRIS, C. C. SCHWARTZ, D. D. HUMBURG AND J. C. COFFEY. 1979.
Native prairie grass pastures as nest cover for upland birds. Wildl. Sot. Bull. 7:4-9.
GREEN, R. H. 1979. Sampling design and statistical methods for environmental biologists.
John Wiley, New York, New York.
GRIFFITH, P. 1976. Introduction of the problems. Pp. 3-7 in Shelterbelts on the Great Plains (R. W. Tinus, ed.). Great Plains Agric. Council Publ. 78.
HARDIN, K. I. AND K. E. EVANS. 1977. Cavity nesting bird habitat in the oak-hickory forests . . a review. USDA For. Serv. Tech. Rept. NC-30.
HARRISON, H. H. 1975. A field guide to birds’ nests. Houghton Mifflin Co., Boston, Mas- sachusetts.
JOHNSON, D. H. 1981. The use and misuse of statistics in wildlife habitat studies. Pp. ll- 19 in The use of multivariate statistics in studies of wildlife habitats (D. E. Capen, ed.).
USDA For. Serv. Gen. Tech. Rept. RM-87.
LACHENBRUCH, P. A. 1975. Discriminant analysis. Hafner Press, New York, New York.
LAPOINTE, D. F. 1958. Mourning Dove production in a central Nebraska shelterbelt. J.
Wildl. Manage. 22:439-l40.
LEVINS, R. 1968. Evolution in changing environments. Princeton Univ. Press, Princeton, New Jersey.
LOKEMOEN, J. T. AND H. F. DUEBBERT. 1976. Ferruginous Hawk nesting ecology and raptor populations in northern South Dakota. Condor 78:464+470.
MARTIN, T. E. 1978. Diversity and density of shelterbelt bird communities. M.S. thesis.
South Dakota State Univ., Brookings, South Dakota.
MCCLURE, H. E. 1946. Mourning Doves in Nebraska and the west. Auk 63:2H2.
MILLER, H. W. AND D. H. JOHNSON. 1978. Interpreting the results of nesting studies. J.
Wildl. Manage. 42:471476.
NELSON, G. 1976. Mourning Dove nesting success in artificial and natural nests in central Minnesota. Proc. Iowa Acad. Sci. 83:112-115.
OLSON, R. A. AND L. D. FLAKE. 1975. Nesting of Ring-necked Pheasants in eastern South Dakota. Proc. South Dakota Acad. Sci. 54:126-136.
Yuhner - AVIAN NESTING ECOLOGY IN SHELTERBELTS 175
ORENDURFF, C. F. 1941. The first wildlife inventory of Nebraska shelterbelts. Nebraska Bird Rev. 9z7-8.
PARKER, J. W. 1972. A mirror and pole device for examining high nests. Bird-Banding 43:216-218.
PETRAITIS, P. S. 1979. Likelihood measures of niche breadth and overlap. Ecology 60:703- 710.
PIANKA, E. R. 1973. The structure of lizard communities. Ann. Rev. Ecol. Syst. 4:53-74.
RANDALL, R. N. 1955. Mourning Dove production in south central North Dakota. J. Wildl.
Manage. 19:157-159.
RAPHAEL, M. G. 1981. Interspecific differences in nesting habitat of sympatric woodpeckers and nuthatches. Pp. 142-151 in The use of multivariate statistics in studies of wildlife habitat (D. E. Capen, ed.). USDA For. Serv. Gen. Tech. Rept. RM-87.
READ, R. A. 1964. Tree windbreaks for the central Great Plains. USDA For. Serv. Agric.
Handbook No. 250.
SCHOENER, T. W. 1974. Resource partitioning in ecological communities. Science 185:
27-39.
SKINNER, R. M. 1975. Grassland use patterns and prairie bird populations in Missouri. Pp. 171-180 in Prairie: a multiple view (M. K. Wali, ed.). North Dakota Press, Grand Forks, North Dakota.
SMITH, M. AND H. SCHOLTEN. 1980. Planting trees for farmstead shelter. Univ. Minnesota Agric. Ext. Serv. Bull. 196.
SOKAL, R. R. AND F. J. ROHLF. 1969. Biometry. W. H. Freeman, San Francisco, California.
STAUFFER, D. G. AND L. B. BEST. 1980. Habitat selection by birds of riparian communities:
evaluating effects of habitat alterations. J. Wildl. Manage. 44:1-15.
WEEKS, H. P., JR. 1980. Unusual egg deposition in Mourning Doves. Wilson Bull 92:258- 260.
WEISER, C. AND D. J. HLAVINKA. 1956. Comparisons of breeding-bird populations of three deciduous shelterbelts. Audubon Field Notes 10:416-419.
WILLIAMS, B. K. 1981. Discriminant analysis in wildlife research: theory and applications.
Pp. 59-71 in The use of multivariate statistics in studies of wildlife habitats (D. E.
Capen, ed.). USDA For. Serv. Gen. Tech. Rept. RM-87.
YAHNER, R. H. 1980a. Breeding-bird censuses 111-117. Am. Birds 34:71-73. -. 1980b. Burrow system use by red squirrels. Am. Midland Nat. 103:409411. -. In press. Seasonal dynamics, habitat relationships, and management of avifauna in farmstead shelterbelts. J. Wildl. Manage.
YEATON, R. I. 1974. An ecological analysis of chaparral and pine forest bird communities on Santa Cruz Island and mainland California. Ecology 55:959-973.
DEPT. ENTOMOLOGY, FISHERIES, AND WILDLIFE, UNIV. MINNESOTA, ST.
PAUL, MINNESOTA 55108. (PRESENT ADDRESS: FOREST RESOURCES LAB., SCHOOL OF FOREST RESOURCES, PENNSYLVANIA STATE UNIV., UNIVERSITY PARK, PENNSYLVANIA 16802). ACCEPTED 6 OCT. 1981.
A COMPARATIVE STUDY OF NESTING BIRDS IN A
FIVE-ACRE PARK
HOWARD YOUXG
T HIS paper presents quantitative data on the size and nesting success of a breeding population of birds in a five-acre park land area; it traces the changes in this population and its reproductive fortunes through a complete breeding season, and compares the breeding cycles, the nest density, and the success of the various species composing it.
The material is based on observations made during the spring and summer of 1947 at Ho-Nee-Urn Pond, a small portion of the University of Wisconsin Arbo- retum, in the vicinity of Madison, Wisconsin. By means of frequent searches (almost daily in spring, later about twice a week) a high percentage of the nests on the area was found, and it was possible to trace the history of most of these through the nesting cycle. Approximately 250 man-hours were spent in the field. Mean temperatures for the study period averaged 2.2 degrees below nor- mal during March through July. August was the hottest ever recorded for the Madison region, 8.8 degrees above normal, and September was slightly above normal. Precipitation for the study period was 2.89 inches above normal.
Acknowledgment for assistance in nest searching and examination is made to Carl Strelitzer, Ernest Boyce, Robert McCabe, James Hale, and Arnold Jackson, Jr. Acknowledgment is made to Professor John T. Emlen Jr. for helpful sug- gestions and for a critical reading of the manuscript.
DESCRIPTION OF THE AREA
The Ho-Nee-Urn Pond area (Fig. 1) is a low-lying park on the northwest shore of Lake Wingra. It is roughly trapezoidal in shape and has an area of 5.2 acres. The 2 main plant communities are a mowed lawn of blue grass (Pea spp.) covering about 4Oa/o of the total area, and numerous plantings of closely spaced arbor vitae (Thuja occidentalis), covering about 26% of the total area.
These plantings are arranged in irregular patterns, making for extensive en- vironmental edges with the grass area.
The arbor vitae varies in height from 5 to 30 feet, with the average tree about 15 feet high. Mixed with it are scattered patches of red osier dogwood (Cornus siolonifera) and staghorn sumac (Rhus typhina), and lesser amounts of white birch (Betula alba), elderberry (Sambucus canadensis), ninebark (Physocarpus opuli- jolius), box elder (Acer tiegundo), honeysuckle (Lonicera tatarica), hawthorn (Crataegus sp.), and highbush cranberry (Viburnum opulus). The ground cover under the plantings is mainly blue grass, nettle (Urtica sp.), thistle (Cirsium sp.), and burdock (Arctium se.). The vegetation beneath the arbor vitae was cut 36 COMPARATIVE NESTING STUDY 37
once, in May, with scythes. Late in the season there were solid beds of swamp milkweed (Asclepias incarnata) along some of the edges.
The northeastern corner of the area contains a small group of black locust (Robiniu pseudo-acacia). The lake shore is lined by occasional cottonwoods HO NE U-l POND
FIG. 1. Ho -Nee-Urn Pond Area and
WINGRA
LEGEND
$2 ARBOR VITAE
QOtIClDlJOUS TREES
-L MARSH AREA
ll BUILDING
5 ROAD
50’ 100
Vicinity
(Populus deltoides), and there is a thick clump of black willow saplings (Salix nigra) in the southwest corner. Near these is a small swampy pond of approxi- mately .3 acres, thickly grown to sedge (Curex sp.), with a few patches of cat- tail (Typhu lutijolfa) and reed grass (Phrugmites commufais). The grass contains liberal amounts of dandelion (Turuxucum ojicinale) and plantain (Plantago major).
38 WILSON BULLETIN
NEST DENSITY AND NESTING CHRONOLOGY
The nesting population of the area is summarized in Table 1. Unless other- wise designated all the data refer to “active” nests (those in which at least one egg was laid). In addition to the species listed, the Cowbird (Molothrus ater) bred as a parasite of the Alder Flycatcher, Cedar Waxwing, Yellow Warbler, Rose-breasted Grosbeak, and Song Sparrow. Figure 2 presents the same mate- rial in graphic form; in this case, however, only nests actually found were con- sidered.
The high density, ranging up to 9.6 nests per acre (May 19-May 20) is con- sidered of special interest. This is probably due to the great interspersion of plant types, and the large amount of “edge” as shown in Figure 1. There are approximately 5000 feet of arbor vitae-grass edge in the area.
Computed according to the usual manner, i.e., the total number of nestson the area during the season, the density was 164 nests, or 32.8 per acre. This is re- ferred to as “Total Nest Density” in Table 1. These nests represent approxi- mately 94 breeding pairs, a density of 18.5 per acre. Steinbacher (1942) found 111 pairs of birds nesting in the 19 acre Frankfort Zoo park. This density of
5.8 pairs per acre, while only one-third that of Ho-Nee-Urn, was considered to be especially high. There do not appear to be other comparable studies in areas of similar size for comparison.
The density data were further analyzed by comparing them from month to month. This showed extensive fluctuations as some species ended their nesting and others started (Figure 2). Following are the average daily densities in nests per acre for the months covered by this study: April-3.5, May-8.6, June-6.0, July-4.8, Aug.-2.1, Sept.--.6. The overall average was 4.3 nests per acre.
It would seem that “Highest Nest Density” (Table 1-D) has a particular significance. Other workers have usually computed what is shown in Table 1 as “Total Pair Density” (Col. F) and “Total Nest Density” (Col. C). These are valuable as indices to the population, but computing the highest nest density has the advantage of showing the greatest nesting population at any one time, and as such is better suited for investigations of the various problems related to social competition such as density tolerance, density-induced behavior, repro- ductive success, etc.
Inter-specific strife was very low even during the periods of highest density.
Of the two most abundant species, Bronzed Grackles and Robins, the Grackles exhibited no recognizable territorial behavior; Robin territories were poorly de- fined and defended with indifferent vigor (Young: 1947). The other species were not observed sufficiently to draw conclusions regarding territorial behavior.
In Figures 2 and 3, the numbers and stages of the nests shown were determined partly by daily examination, and partly by interpolation and extrapolation.
Nests in various stages of the cycle were followed, and averages based on all records were determined for the length of time in each phase. When a new nest
COMPARATIVE NESTING STUDY
TABLE 1
Species Nesting at Ho-Nee-Urn Park, 1947
A
SPECIES
Bronzed Grackle
Qz&xxlus quiscztla
Robin
Turdzts migratorizrs
Catbird
Dumetella curolinensis
Cedar Waxwing
Bolnbycilla cedrorum
Yellow Warbler
Dendroica p&&a
Goldfinch
Spinus tristis
Song Sparrow
Melospiza melodia
Mourning Dove
Zenaidura mawoura
Alder Flycatcher
Empidonax traillii
Mallard
Anas platyrhynchos
R. N. Pheasant
Phasianus colchicz~s
Killdeer
Charadrius z’ocijerws
Rose-Br. Grosbeak
Plteuticus ludovicianus
Chipping Sparrow
Spizella passerina
Warbling Vireo
Vireo gilvus
Yellowthroat
Geothlypis t&has
__~_________~_
Total.
Sets found
26 36
22 14
12 9
2 11
5 1
2 1
1 1
0 0
143
B
Esti-
mated
total of
nests
I-
C
.__
Total
nest
density
D
1 rota.1 no.
of pairs
(eiti-
mated)
Total
pair
density
28 5.6 4.2 21 4.2
37 7.4 3.8 19 3.8
22 4.4 2.2 11 2.2
16 3.2 1.6 8 1.6
12 2.4 7 1.4
9 1.8 6 1.2
15 3.0
2.2 1.0
.4
.4
.2
.2
1.4 1.2
.5 1.0
1.0 11 1.0
5 .6
2 .2
2 .2
1 .2
1 .2
1 .2 .2
.2 ~ .2
I
.2 I .2
5 5
3 2
2 1
1 1
1 1
.6
.4
.4
.2
.2
.2
.2
.2
1 1
___ ____ ~___ _-__ ____
164
I 32.8 1 9.6 , 94 / 18.5
39 C & F computed from B & E.
D-highest density at an) one time; computed from A.
C, D, & F computed on a per-acre basis.
40 WILSON BULLETIN
was found it was thus possible to determine fairly accurately the date on which it was started. For example, observations showed that Robins took about 7 days on an average to build their nest; it then remained empty for an average of 4 days, after which the eggs were laid at the rate of 1 a day. A Robin nest found to contain 2 eggs on April 20 was therefore tallied as having been started on April 7, since 13 days were necessary to bring it to the 2-egg phase of the cycle.
In most cases it was only necessary to extrapolate 3 or 4 days. About 35% of the material presented in the figures was thus computed.
The efficiency of nest searches was tested by comparing the extrapolated totals of any given day with the number of nests actually known on that day. On this basis it was found that the number known on any given day varied from 32% FIG. 2. Ho-Nee-Urn Nests, 1947
to 100% of the number actually present. This does not take into consideration nests of Song Sparrows, since time did not permit extensive searches for them, and only 2 were found. On the peak days, May 19-20, 88yo of the nests had been located. Comparison of the number of nests found with the estimated total of nests in Table 1 shows that about 87% of all nests were eventually found. In view of these facts the writer believes that Figure 2 represents a fairly accurate graph of the nesting. Errors in extrapolation would push indi- vidual peaks and depressions 1 or 2 days in either direction, but these would tend to compensate for each other, and the general picture would remain the same.
Three major peaks in nesting density are discernible in Figure 2; late May (mainly Robins and Grackles), most of June (Robin, Catbird, Yellow Warbler and Cedar Waxwing), and mid-July (Catbird, Waxwing, Goldfinch). The COMPARATIVE NESTING STUDY 41
meagerness of the data prevents extensive discussion, but a few interesting things may be pointed out. The overlapping nesting periods of Robins and Grackles, both using the same nesting sites (arbor vitae), makes them competitors. Cedar Waxwings also nested almost exclusively in arbor vitae, but did not start until the Grackles were gone and Robin nesting was much diminished. The Gold- HO NE LJM 1947
0 NESTS UNDERCCNSTRLICTJON
L2 NESTS WITH EGGS
n NESTS WITH lCl”NG
I4
CATBIRD
BRONZED
GRACKLE
ROBIN
FIG. 3. Comparative Nesting Cycles, 1947
finches and Yellow Warblers nested mainly in ninebark; there was about a two week overlap in their nesting periods. It is impossible to say how much compe- tition for nesting sites affected the density of the various species; as previously mentioned, little interspecific strife was observed.
Figure 3 compares the nesting cycles of the 3 most abundant species at Ho- 22 WILSON BULLETIN
Nee-Urn. It demonstrates the “Highest Nest Density” discussed before and makes possible comparisons, as between the Robins and Grackles. More Robins nested on the area than Grackles (Table l), but their nesting was spread over a much longer period, and at no time did they attain a density as high as that of the Grackle’s peak. The graph of the Grackles, which are single-brooded, lacks the long “tail” characteristic of the other species. More than half of the Grackle nests were broken up, and it would seem that the lack of a “tail” on their graph indicates that either 110 renesting attempt was made, or that the birds moved en- tirely off the area for the next attempt. Part of the sharp fall in the number of Robin nests after May can possibly be explained by the scarcity of hardwood trees on the area. Howell (1942: 549) found that 5707, of the Robin nests he studied were built in conifers during the first nesting period, but later, when the TABLE 2
Ho-Nee-Urn Nesting Data, 1947
SPECIES
Active
nests
Bronzed Grackle I- 26
Robin. 36
Catbird .’ 22
Cedar Waxwing. 14
Yellow Warbler, 12
Mourning Dove, 11
Totals and averages 121
45 5”:
57 42
1 :;
D computed from A; F computed from B.
B-Nests producing at least 1 young.
B C D
Eggs
Per
nest
4.5 3.3
3.1 3.9
3.5 1.8
72 61
65 54
::
61 Fledg
lings
Per nest
4.3 2.4
2.9 2.6
2.8 2.0
-
G
%
Em pro-
hcing
fledg-
lings
--
:;
if
33 20
40 H I
% of
You%
ikdg-
ing
AV.
UCWSS
,f pro-
uctive
nests
:: E 79 92
:; ::
66 103
66 81
hardwood trees had leafed out, the number of Robins found nesting in conifers fell to 38%.
KESTING SUCCESS AND CAUSES oI3 F.~ILURE
Table 2 compares the nesting data of the various species for which at least ten nests were found. Again the small number of nests precludes intensive analysis of the figures, but some basis for comparison of nesting success is available.
Comparison of the percent of eggs hatching (E) with the percent of young fledging (H) shows a differential effect of environment on the various species under consideration. The Robins suffered losses approximately equal during the egg stage and the nestling stage. Catbirds and Grackles were affected in di- rectly opposite fashion, the former having a heavier loss during the egg stage, the latter during the nestling stage. In the case of the Grackles, a bad storm when there were many small young in the nest accounted for much of the mor- tality.
COMPARATIVE NESTING STUDY 43
The same comparison also indicates that the species having the poorest suc- cess in hatching eggs generally had a proportionately greater success in fledging their young. It is interesting to note that in comparing success from year to year, Nice’s (1937: 141) figures for the Song Sparrow show the same general pat- tern within a single species. In her Table XVI, by dividing the number of fledg- lings by the number of hatched eggs, the following figures are obtained:
Year
1930
1931
1932
1933
1934
1935
% 0J Eggs IIalching 70 of Young Fledging
68% 63%
72% 63%
61% 61%
51% 37%
35% 78%
41% 71%
The non-conformity of 1933 is apparently due to plowing of the nesting area at a time when most of the nests contained small young. Very close correlation should not be expected, since there are many other factors besides hatching suc- cess which could influence the fledgling success.
A similar situation is suggested in a study of the Eastern Red-winged Black- bird (Agelaizts phoeniceus) by Smith (1943: 190). Of two nestingRed-wingpopu- lations, “A” hatched 74% of 563 eggs and fledged SO% of the young, “B” hatched 707c of 577 eggs and fledged 84% of the young. Much more data is needed for statistical testing, but it appears that there is possibly a compensatory interaction here following the general pattern that low egg survival results in high nestling survival, or conversely, that high nestling survival is associated with low egg survival. Errington (1946: 170) used Smith’s data to illustrate the effects of predation, showing that a high loss of eggs apparently resulted in re- duced vulnerability of the nestlings to predation. Another factor possibly in- volved is competition between nestlings, which Emlen (1942 : 151) considered a major factor in nestling survival in the Western Crow.
Comparison between species of reproductive success, based on the number of young fledged per pair or per nest, will always be affected by the varying clutch size among the different species, and the varying number of broods raised.
These factors can be eliminated by comparing the percent of eggs producing fledglings (Table 2--G). Considering each egg as a reproductive attempt, this shows what percent of the reproductive potential is attained. On this basis Catbirds were the most successful breeders at Ho-Nee-Urn, (51% of their eggs producing fledglings) while Mourning Doves were the least successful, (only 2O7c of their eggs producing fledglings). The average success of productive nests (Table 2-I) was computed by dividing the number of fledglings by the number of eggs. The Mourning Dove nests were either complete failures or total suc- cesses, reflecting the birds’ quickness to desert when molested. On the other hand, the Robins were able to bring off broods successfully after disturbances, though in only a few cases were their nests 100% successful. 44 WILSON BULLETIN
Reproductive success possibly has the same inverse correlation with density as shown for reproductive activity by Kendeigh (1934: 308). Smith (1943:
204) found no evidence of this in the Red-wing, but Errington (1945: 14) found an inverse ratio between spring density and summer population gain of the Northern Bob-white (Colinus virginianus). Now that nesting studies have be- come more common it should be possible to compile similar data on various species from several areas. The value of these would be enhanced if it were also possible to compare the densities of the different areas, but this information is seldom available. More intensive studies might show optimum densities for TABLE 3
Nest Success in Various Studies
Robin.
Catbird. ............
Cedar Waxwing. .....
Yellow Warbler. ......
Mourning Dove.
1 .iI I 501
136
REFERENCE
Kendeigh, 1942 Ohio
Howell, 1942 N. Y.
Koehler, 1945 Wise.
This study Wise.
Thomsen, 1944 Wise.
Kendeigh, 1942 Ohio
This study Wise.
Kendeigh, 1942 Ohio
This study Wise.
Lea, 1942 Mich.
Kendeigh, 1942 Ohio
Kendeigh, 1941 Iowa
This study Wise.
McClure, 1946 Iowa
McClure. 1946
Kendeigh, 1942
Nebr.
Ohio
This study Wk.
McClure, 1946 Calif.
lOCATION YEAR
193;-38
1945
1947
1944
1;47
lG47
1940-41
1;40
1947
1938-40
194133
A47
1944
the different species, such as has been shown for various invertebrates by Allee (1931: 161-180).
Nice (1937: 143) postulates 4Oye to 46% as typical success for open nests of passerines in the North Temperate Zone. At Ho-Nee-Urn the passerine species averaged 49.8oje successful in nesting attempts, a figure which agrees quite well with her estimate.
Table 3 compares the nesting success at Ho-Nee-Urn with that of the same species in other areas. No references were found for the Bronzed Grackles, and again the species are limited to those for which at least 10 nests were found at Ho-Nee-Urn. The data of Kendeigh (1942) were adjusted to make them con- form to “active” nests as defined in this study. Unfortunately, information on density was not given in the other studies, but the consistently low success at COMPARATIVE NESTING STUDY 45
Ho-Nee-Urn (excepting Cedar Waxwings) is possibly a reflection of its highnest- ing density. However, such a comparison suffers from the fact that the other studies sometimes covered several years, and were from many different area types. Omitting the Ho-Nee-Urn data, the nest success of the passerine species listed in Table 3 averages 74y0,, nearly double Nice’s estimate for open nesting of passerine birds.
TABLE 4
Nest Failures Ho-Nee-Urn, 1947
SPECIES
Nest failures
A Predation
B Desertion.
CWeather......
D Total.
Bronzed~ Grackle j Rob1n
Eggs not hatching
A Predation.
B Desertion.
C Weather.
D Unaccounted.
E Infertile/addled.
14 15
6 4
7 FTotal.. _. .I 33 / 46
Young not fledging
A Predation.
B Desertion.
C Weather.
D Unaccounted.
E Fell from nest.
FDiedinnest
G Cowbird parasitism.
H Total
0 10
9 3
li I s
51 7
:
1 0 I :
33 1 29
10 / 6 ( 7
20 8 11
2 5
: : 2 0 1
2 7 5
24 1 25 1 22
MOUlT-
ing Total
DOW?
6 29
2 25
1 10
9 I 64
0” 1: 0 24
Per
cent
2 16
100
52 21
;
1.5 21
15 34
18 7
:
100
All nest failures could be attributed to one of 3 causes: predation, desertion, or storm damage. The effect of these on the species studied, referring only to “ac- tive” nests, is shown in Table 4.
Two Robin nest desertions were directly due to activities of the observer.
Three Robin eggs, 1 Grackle egg, and 1 Yellow Warbler egg were accidentally broken; they were listed as lost due to predation. As far as could be determined the activities of the study had no other effect upon the species under considera- tion.
The predation was nearly 100% by an unknown avian form which punctured the eggs in the nest. Bronzed Grackles were suspected, but were never seen at 46 WILSON BULLETIN
the nests of other species, or carrying their eggs or young. The Grackles were common on the area from late March to late June, but egg losses continued at about the same rate after their departure. If a predator took 1 egg out of a clutch and the owner deserted, the remainder of the eggs were listed as lost due to desertion. Predation on adult birds was not observed, and its extent is not known. However, it probably does introduce an error in the records on de- serted nests, since some of the resident birds may have disappeared because of predation rather than because they deserted.
Predation was the chief cause of nest failure, operating most strongly during the egg stage. Weather of course would act differently on the various species from year to year, depending upon the time that bad storms happened to occur.
SUMMARY
Sesting birds were studied in a five-acre park area with arbor vitae and blue- grass lawn as the main cover types.
A total of at least 15 species (94 pairs) bred on the area during the period of study.
May and June were the months of highest nest density. On May 19-20 there were 9.6 active nests per acre. The average number of nests per acre for the season was 4.3. The total density for the season was 32.8 nests per acre.
The high density did not produce any noticeable interspecific strife.
Catbirds were the most successful breeders, producing fledglings from 51% of their eggs; Mourning Doves were the least successful, producing fledglings from 207o of their eggs.
Those species suffering the greatest loss of eggs in the nest generally appeared to be the most successful in raising nestlings.
Predation by an unknown avian form was the main cause of nest failures.
REFERENCES
ALLEE, W. C.
1931 Animal aggregations. University of Chicago Press, 431 pp.
EYLEN, J. T. JR.
1942 Notes on a nesting colony of Western Crows. Bird-Bending, 13: 143-154.
ERRINGTON, P. L.
1945 Some contributions of a fifteen-year local study of the Northern Bobwhite to a knowledge of population phenomena. Ed. Mon., 15: 3-34.
1946 Predation and vertebrate populations. Quart. Rev. Biol., 21: 144-177.
HOWELL, J. C.
1942 Notes on the nesting habits of the American Robin. Amer. &f&.% Nut., 28:
529-603.
KENDEIGH, S. C.
1934 The role of environment in the life of birds. Ecol. Mon., 4: 299-417. 1941 Birds of a prairie community. Condor 43: 165-174.
1942 Analysis of losses in the nesting of birds. Jown. WiZdZ. Man.. 6: 24-42.
KOEHLER, A. Mks. AND G. KOEHLER -
1945 A bird study in the Madison Cemetery. Passenger Pigeon, 7: 15-19.
LEA, R. B.
1942 A study of the nesting habits of the Cedar Waxwing. WiZs. Bull., 54: 226-236.
COMPARATIVE NESTING STUDY 47
MCCLURE, H. E.
1946 The Mourning Dove in Nebraska and the West. Auk, 63: 24-42.
NICE, M. M.
1937 Studies in the life history of the Song Sparrow, Part I. Tvans. Linn. Sot. N. Y. 4: 141-143.
SMITH, H. M.
1943 Reproductive success in the Eastern Redwing (Ageluizcs 1. phoeniceus) in relation to Darling’s hypothesis. Ecol. 24: 183-207.
STEINBACHER, G. _
1942 Die Siedlungsdichte in der Parklandschaft. Jo24m. jiir Ounitlz., 90: 342-359.
THOMSEN, H. F.
1944 Ornithological abstracts from an ecological survey of the University Bay Region.
Passenger Pigeon, 6: 3-13.
YOUNG, H. F.
1947 1947 Ho-Nee-Urn Robin study. Unpubl. MS., Zoology Dept., Univ. Wise., Madison, Wise.
ZOOLOGY DEPARTMENT, UNIV. OF WISCONSIN, MADISON, WISCONSIN.
THE WILSON BULLETIN PUBLICATION DATES
The actual dates of publication of the four numbers in 1948 were: March 22, June 22, October 13, December 28.
.