Frances C. James

Geographic Size Variation in Birds and Its Relationship to Climate

There is a high degree of concordance among the patterns of geographic size variation in birds in the eastern and central United States. This is demonstrated for 12 species by assuming that wing length measurements are an indicator of body size on the intraspecific level, and by arranging the data in the form of a grid of means of wing lengths for sample areas. Maps giving isophenetic lines for wing length indicate gradually increasing size dines northward and westward from Florida in the Hairy Woodpecker (Dendrocopos villosus), Downy Woodpecker (Dendrocopos pubescens), Blue Jay (Cyanocitta cristata), Carolina Chickadee (Parus carolinensis), White-breasted Nuthatch (Sitta carolinensis), and Eastern Meadowlark (Sturnella magna). In each case there is a trend for larger (or longer-winged) birds to extend southward in the Appalachian Mountains and for smaller (or shorter-winged) birds to extend northward in the Mississippi River valley. Maps made by a computer and automatic plotter using contour intervals of 0.5 mm of mean wing length for the Downy Woodpecker, for male White-breasted Nuthatches, and for female Blue Jays show that, in addition to the pattern just mentioned, relatively longer-winged birds extend southward in the interior highlands of Arkansas, and relatively shorter-winged birds extend northward up other river valleys. These subtle relationships between intraspecific size variation and topo- graphic features suggest that the link between the two phenomena may be precise adaptations to even minor climatic gradients. The relationship between these findings and the subspecies concept is discussed. Correlation coefficients for the pattern of variation in the Downy Woodpecker with seasonal and annual wet-bulb temperature, vapor pressure, and absolute humidity were all either equal to or higher than correlations with dry-bulb temperature. Since these variables reflect the combined effects of temperature and humidity, the obvious indication is that size variation is more closely related to this combination than to temperature alone. Additional correlations using the mean wing length data for seven other species confirmed that wet-bulb temperature patterns are more closely related to bird size than either dry-bulb temperature patterns or latitude. These relationships can be expressed numerically as regressions of mean wing length on either annual wet-bulb temperature or mean annual total heat per pound of air. Since increased evaporation at high altitudes and in arid areas accentuates the depression of a wet- bulb thermometer, my hypothesis may partially account for several cases of size variation in birds cited by others as disturbing exceptions to Bergmanns ecogeographic rule. Sections of a translation of Bergmanns paper published in 1847 are given. The biological mechanisms by which these relationships are maintained are unknown, and the wide range of tolerance by birds to diurnal and seasonal temperature variations tends to mask them. If the well-established inverse relationship between weight and metabolic rate per gram of homeotherms is operative on the intraspecific level, the relationships can be discussed in terms of avenues of heat loss and of energy budget equations.

RELATIONSHIPS BETWEEN TEMPERATE FOREST BIRD COMMUNITIES AND VEGETATION STRUCTURE

General patterns of density, species richness, and relative abundance of breeding birds are examined in a wide variety of North American forests. Taking data from censuses published in American Birds, we use rarefaction to ordinate the species richness of communities in terms of samples of equal numbers of individuals. By this criterion young forests in secondary succession and mature deciduous forests can be equally rich in bird species; coniferous forests and dense successional stands having only one or two species of trees have the fewest species of birds. For our data set the density of birds is higher in mature deciduous forests than in successional stands. The species/area relationship (sla) is a function of both this species/individuals relationship (s/n) and the individuals/ area relationship (n/a). According to the Jonckheere-Terpstra statistic there is a statistically significant pattern to the order in which bird communities fall with regard to s/n and n/a. To see these relationships in terms of the structure of the vegetation, we present the positions of 56 stands in a graphic space determined by principal components. Tree species richness and canopy cover dominate the first axis, and variation in canopy height the second. Tree density has a somewhat independent pattern of variation and is expressed by the third axis. In the bivariate space of principal components one and three, a SYMAP contour diagram of the number of bird species expected in 10 ha (sla) shows maximal values in mature deciduous forests, but not in those that have the highest tree species richness, canopy height, or tree density. The number of individual birds/10 ha (n/a) shows a similar pattern except that maximal density occurs at maximal values of tree species richness and canopy height. By this criterion both bird species richness and density are minimal in coniferous forests characterized by high tree density, low canopy, and few species of trees. These patterns are not discernible by the classic bird species diversity/foliage height diversity method proposed by Mac Arthur and Mac Arthur in 1961. They are masked by correlation coefficients, partly because the coefficients are insensitive to nonlinear relationships. We recommend rarefaction, principal components analysis, and contour diagrams to display relationships among communities.

Environmental Component of Morphological Differentiation in Birds

Geographic character variation in birds is usually attributed to natural selection for phenotypes that reflect locally adapted genetic differences. However, experimental transplants of red-winged blackbird eggs between nests in northern and southern Florida, and from Colorado to Minnesota, show that in this species a significant proportion of the regional differences in nestling development is nongenetic. If natural selection is maintaining the clines of character variation that are observed in adult phenotypes, the genetic and nongenetic components of phenotypic variation must covary.

New Approaches to the Analysis of Population Trends in Land Birds

We thank Drs. Link and Sauer for their comment (Link and Sauer 1997b) on our article (James et al. 1996) and welcome this chance to expand further on the reasoning behind the methodology we presented. We are especially gratified that there is general agreement about what the data say, even if not on how to analyze them. Since the beginning, when we proposed our nonparametric regression methods (James et al. 1990), our philosophy has been that flexible description is likely to yield the most information from the data. This philosophy was opposed to the methods employed at the time, in which linear relationships were fitted on the log scale to measure trend and the emphasis was mainly on hypothesis testing (Geissler and Noon 1981, Geissler and Link 1988, Geissler and Sauer 1990). We feel now, as we did then, that it would take a number of different approaches to explore different questions and different aspects of the data fully. We also note that Peterjohn et al. (1995) used LOESS (Cleveland and Devlin 1988), as does the North American Breeding Bird Survey (BBS) web site (Sauer et al. 1997) housed at the Patuxent Wildlife Research Center, following our lead in fitting flexible responses through time. Conversely, and partially in response to suggestions by Link and Sauer, we have incorporated the ability to test hypotheses, to include observer effects, and to include more routes in our analyses. We now address each of their points in turn.

THE GRINNELLIAN NICHE OF THE WOOD THRUSH

Our results suggest that the limits of the breeding range of the wood thrush and its relative abundance within its range are not highly related to the presence of ecologically similar species. These parameters are better accounted for by variables such as species-specific nesting and foraging requirements, which in turn covary with the vegetation structure of the eastern deciduous forest. Studies of single-species geographical ecology should precede studies of assemblages. The Grinnellian model is more likely than the Hutchinsonian model to provide sound information on factors regulating the distribution and abundance of animals. Nevertheless, without controlled experiments both the Hutchinsonian and the Grinnellian models are descriptive, and inferences about processes underlying patterns will continue to be weak. The neo-Grinnellian approach uses some of the same analytical techniques as the Hutchinsonian approach, but it employs multiple comparisons to study the geographical ecology of single species, and it is cautious about interpretation. Comparisons with congeners and species of similar ecology are still of interest, including the extent to which species differences result from interspecific niche shifts. An important problem that will always plague niche analysis based an observational data is whether regional differences in resource use are a reflection of (1) biological differences among the populations; (2) measurement of resources that are irrelevant to the species; or (3) interspecific niche shifts. In cases in which controlled experiments are not possible, this problem can be addressed as single-species, large, multifactor studies in geographical ecology. These studies need not be cast in terms of community theory. Future studies should try to clarify the relationship between habitat selection and both Hutchinsonian and Grinnellian niche analysis. Careful design of sampling procedures with multiple comparisons could permit evaluation of the relative power of proposed mechanisms to explain observed patterns.

Data Analysis and the Design of Experiments in Ornithology

One familiar statement one hears about data analysis in ornithology is that traditional ornithologists accumulated facts, but did not make generalizations or formulate causal hypotheses (Emlen, 1981; Wiens, 1980). The approach of modern ornithologists, on the other hand, is said to be more theoretical. Modern ornithologists formulate hypotheses, make predictions, check the predictions with new data sets, perform experiments, and do statistical tests. Proponents of the modern approach frequently state their position in a condescending tone, and this is guaranteed to arouse the ire of those more committed to empirical research (Olson, 1981). In this essay, we hope to lessen the differences of opinion on this subject as they were recently expressed in the commentary section of The Auk (Austin et al., 1981). We agree with Popper (1959, 1972, 1983) that the goal of science is to develop better explanatory theories about processes in nature, but it does not follow that advances are made only by hypothesis testing.

ECOSYSTEM MANAGEMENT AND THE NICHE GESTALT OF THE RED‐COCKADED WOODPECKER IN LONGLEAF PINE FORESTS

We use the term “optimal niche gestalt” to refer to the concept that there are structural features of the environment that allow a species to thrive over and above those that allow it to persist. Analyses of the covariation between demographic and habitat features can reveal a trajectory toward this optimal state. To help identify new criteria for foraging-habitat guidelines for the Red-cockaded Woodpecker (Picoides borealis) in the Apalachicola National Forest, we examine seven years of demographic data for the woodpecker population and habitat in core stands, the naturally regenerated prime habitat in the centers of their territories. For both districts of the forest, two compound habitat variables are highly related to the average number of adult birds per social group, the average number of young fledged per group, and the density of groups. These variables are, first, the difference between the density of trees >35 cm dbh and that of trees 15–25 cm dbh and, second, the difference, in the ground cover...

SEASONAL PREVALENCE OF A HAEMATOZOAN PARASITE OF RED-BELLIED WOODPECKERS (MELANERPES CAROLINUS) AND ITS ASSOCIATION WITH HOST CONDITION AND OVERWINTER SURVIVAL

Abstract We examined seasonal prevalence of a haematozoan parasite (Haemoproteus velans) of the Red-bellied Woodpecker (Melanerpes carolinus) in the Apalachicola National Forest, northern Florida. We also investigated how infection with H. velans was associated with host mass, body condition, and overwinter survival. Analysis of blood smears taken from individual woodpeckers between May 2000 and July 2001 indicated that prevalence of H. velans peaked in July 2000, at ∼80% of individuals sampled, decreased to 0% in January and February 2001, and peaked again in July 2001, at ∼50% of individuals. Infection with H. velans was associated with low mass and poor body condition in males. Infection showed no association with female mass. In addition, infection with H. velans showed no relationship with overwinter survival. Our data reemphasize the importance of considering seasonal variation in parasite prevalence during testing for haematozoa. In addition, our data suggest that, although infection with H. velans is associated with poorer host condition, it does not negatively affect host survival.

Developmental Plasticity in the Shell of the Queen Conch Strombus Gigas

To evaluate the developmental plasticity of shell morphology in the queen conch Strombus gigas, we conducted transplant experiments at five sites in the Exuma Cays, Bahamas. Juvenile conchs from a sixth site were placed in cages at three sites, in an open cage at a fourth site, and in a naturally confined area at a fifth site. All animals spent 8 mo in the field. At the end of the experiments, the morphology of their shells was quantified by means of univariate and multivariate analyses of size and shape as advocated by Mosimann and James (1979). Comparisons were made within and among experimental groups at the first three sites, between groups at the fourth and fifth sites, and between experimental groups and naturally occurring animals at all six sites. Although the experiments extended over only 20% of the juvenile growth period, anal- yses showed that experimental animals at three of the sites differed significantly in the shapes of their shells. The shapes of these shells were also significantly different from those of the long-spined, bulky shells of naturally occurring animals of similar age at the source site. The greatest change in shape and growth rate occurred at the fifth site, where the rapidly growing uncaged experimental animals developed relatively slender, short-spined shells with large apertures and short spires. These changes were clearly in the direction of the size and shape of unmanipulated animals in the foster population. The results of these experiments, when combined with those of others, suggest that, although variation in the morphology of marine gastropods can have a genetic basis and can be influenced by natural selection, direct environmental induction can be the dominant process underlying patterns of local among-population differentiation. We recommend (1) that the morphometric meth- ods used here to investigate the complex consequences of variation in the size and shape of organisms be applied in other taxa and (2) that the extent to which among-population patterns of geographic variation are directly induced by the environment be given more attention.

Diet of the Red-Cockaded Woodpecker in the Apalachicola National Forest

Most territories of the largest population of red-cockaded woodpeckers (Picoides borealis; RCWs) in the western district of the Apalachicola National Forest (ANF) do not meet the current foraging-habitat requirements in the recovery plan for this species. We suspect that more information about the relation between the diet of the RCW and its habitat is needed and would permit development of better guidelines for the ANF and other populations. We studied diets of adult and nestling RCWs and their relation to variation in habitat. We found that ants were the dominant prey item by biomass (58%). Additional major items in stomachs of adult RCWs were other arthropods, fruits and seeds, and wood. Of the 4 species of ants present in samples, 74% of the biomass consisted of eggs, larvae, pupae, and adults of the arboreal ant Crematogaster ashmeadi. Territories burned more frequently had pine trees (Pinus spp.) with a higher diversity of ant species and lower proportions of trees occupied by C. ashmeadi than did territories burned less frequently. Also, stomach samples from adult male woodpeckers at frequently burned sites had lower proportions of ants, including C. ashmeadi. More information is needed about the dynamics of interspecific relations within arthropod communities of longleaf pine trees (P. palustris), the importance of ants to the ecology of the RCW, and the relevance of variation in habitat to diets of RCWs.