Patterns of Morphology and Ecology in Grassland and Shrubsteppe Bird Populations

We analyze patterns of morphological variation in a guild of 12 species of grassland/shrubsteppe birds in relation to various theories expressing ecomorphological correlations. Variability in morphology over the six geographic locations sampled was moderate for most species, but several morphological traits exhibited significant variation between years at given locations. Sexual variation (dimorphism) was significant for most traits in most species; it was most apparent in polygynously mating forms, but occurred as well in several monogamous species. Dimorphism was generally greater for traits related in a general way to body size than for bill size traits, suggesting a stronger role for sexual selection than for intersexual trophic niche differentiation in driving dimorphism in these species. In nonsexual dimensions of variation, variability in bill features and body measurements tended to be similar. Migratory species did not differ from resident species in the degree of morphological variability. There was heterogeneity as well in the patterns of covariation among morphological traits. Principal component analysis indicated different clusterings or "suites" of morphological traits in component space for the different species, although in several the first principal component appeared to represent a general "body size" set. In these species there was thus substantial morphological variation that occurred in no readily discernible patterns in space or time, suggesting that simple characterizations of the morphology of these species from museum specimens or from literature values are at best only general approximations, and may be inappropriate for testing ecomorphological theories at the level of local populations and communities. An analysis of relations between morphology and diet revealed few clear associations in the directions predicted by theory within species populations. Bill size and prey size in individual birds were generally uncorrelated, and the few correlations between prey size and suites of covarying morphological traits were quite weak. Variability in bill dimensions was unrelated to dietary diversity or variety; populations in which bill sizes were more variable did not consume more variable—sized prey than less variable populations of a species. Similarly, there was no relationship between sexual dimorphism and diet variety. Between species, on the other hand, there were significant relationships between body length or bill length and prey size. Principal component analysis revealed that over the entire assemblage of species considered, bigger species had bigger bills and larger territories; they tended to eat bigger things that were more variable in size, and they were primarily carnivorous. The sizes of birds in this assemblage were distributed in a discontinuous fashion, however, with a gap in the body size spectrum between 30—40 g species and those weighing over 100 g. Removal of the large species from the morphology—diet analysis destroyed the clear patterns. The agreements with theoretical predictions thus hold only at a general level, but do not appear when the several small (10—40 g) species are considered as a group. Examination of the frequency distributions of prey sizes taken by coexisting guild members at several locations revealed a wide array of patterns, ranging from orderly size—related separations or rankings through complete overlap in prey size distributions to inversions, in which small—sized species consumed larger prey than large—sized species. It is apparent that any predictions of dietary characteristics from morphology are tenuous indeed. Body size was significantly related to territory size over the entire spectrum of bird species, but there was substantial variation in territory sizes within any given species, and the relationship did not obtain within the set of small species. Variations in abundance of populations were not associated with changes in the variation of morphological traits; more abundant populations were not necessarily more variable. Overall, our analysis casts doubt on the applicability of much current ecomorphological theory to the bird guilds we studied. We believe that close associations between morphology and ecology are obscured by the substantial variability in both. This variability may reflect the fluctuating selection pressures characteristic of arid and semiarid environments, where environmental stresses (and the attendant tight restriction of population variation in morphology and ecology) are sporadic in their occurrence.