Components of Variance in Measurements of Nestling European Starlings (Sturnus vulgaris) in Southeastern Pennsylvania

Abstract

Abstract I report the results of experiments designed to distinguish factors affecting within-population variability in measurements of nestling European Starlings (Sturnus vulgaris). Individuals were switched among pairs of nests at the beginning of the incubation period and at the beginning of the nestling period. Variables were analyzed in a replicated (nested), three-way factorial analysis of variance to distinguish the contributions of factors associated with (1) the genotype of the embryo or composition of the egg, (2) the incubation period, and (3) the nestling period. In addition, I analyzed the correlations among growth variables within the sample as a whole and among main effects in the experimental design to search for patterns of genotypic and phenotypic interrelationship. I also related growth variables of nestlings to the size and composition of sibling eggs and to the length of the incubation period. The present experiment did not reveal any genetic or egg-quality component to within-population variation in rate of mass increase and mass asymptote achieved, length of the tarsus of fully grown nestlings, or rates of growth of the wing and outer primary feather. Similarly, the nest in which the individual was incubated had no effect on postnatal growth. The nest in which the individual was reared significantly contributed between 19 and 29% of the sums of squares in rate of mass increase, length of the tarsus, rate of increase in length of the wing, and the maximum length of the sheathed portion of the outer primary feather. Rearing nest did not have a significant effect on the mass plateau of the chick. These results, particularly the absence of some effects, are difficult to interpret, because a large proportion of the variation in several variables occurred between pairs of switched nests. These differences undoubtedly included some of the variation that might have been attributable to effects within experiments. Growth variables were weakly correlated over the entire sample, revealing little pattern of variation in postnatal growth. Over certain of the effects in the experimental design, however, particular groups of variables were strongly correlated, indicating interrelated responses of some of the growth parameters to environmental and, perhaps, genetic factors. The size of the egg had a small effect only on the mass plateau and the length of the tarsus of the chick. The composition of a sibling egg influenced only the maximum length of the sheath of the outer primary feather. The present study was somewhat weakened by small sample size and an inexplicable between-experiment effect that reduced its power to distinguish among effects of interest. The experimental design has the potential, however, to disentangle many classes of factors associated with genotype and parental care that contribute to within-population variation in phenotypic measurements.