Environmental influence on life-history traits: growth, survival, and fecundity in Black Brant

Ecology, Dec, 1995 by James S. Sedinger, Paul L. Flint, Mark S. Lindberg

For individual females that were known to have survived (i.e., observed at least 1 yr after hatching), we examined the relationship between their size as goslings and their likelihood of breeding as follows. Individual females were scored as having nested if they were flushed from a nest, observed with a brood during brood rearing, or captured during banding with a brood patch. Our assessment of breeding, therefore, included nesting, whether or not females were successful. We restricted analysis to females from the 1986 and 1987 cohorts, because only females from these two cohorts had at least two opportunities to nest within the years (1986-1990) considered in our analyses; most female geese do not begin nesting until 2 or 3 yr of age (Finney and Cooke 1978, Rockwell et al. 1983, J. S. Sedinger, personal observation). We restricted breeding records in this analysis to those in the 2nd or 3rd yr so that females from both cohorts had an opportunity to be detected as breeders in exactly 2 yr. We divided the sample of 108 females who had been measured as goslings and were resighted at least 1 yr later into three groups of approximately equal size based on their size as goslings. We then used [[Chi].sup.2] contingency table analysis to test the hypothesis that females of different size varied in their likelihood of breeding, given they had survived. We also performed an alternative analysis to examine the relationship between gosling size and later breeding propensity in which we used logistic regression to examine the relationship between breeding (breed = 1, not breed = 0) and PC1 scores.

We examined the relationship between gosling size and adult size by regressing body sizes of individuals recaptured at 1 yr of age or older against body sizes of the same individuals as goslings. To determine whether it was necessary to control for adult age in our analysis of the relationship between gosling and adult body size we used ANOVA to examine the relationship between mean size measures (mass, culmen, and tarsus) and age (1, 2, 3 yr). Because mass is potentially seasonally dynamic, we used analysis of covariance (ANCOVA), with age in years as a fixed factor and days following peak of hatch as the covariate in analysis of mass and adult age. We also examined age-related variation in the PC1 scores of adults using one-way ANOVA with age (1, 2, 3 yr) as a fixed factor.

We examined the relationship between body size of adults and their investment in eggs using ANCOVA. We used clutch volume as the dependent variable for this analysis, because nutrient reserves of female geese actually regulate mass (as indexed by volume) of eggs produced. It is important to distinguish between a physiological hypothesis of regulation of egg production by nutrient reserves, in which body size plays a role, and variation in the life history trait, clutch size. Clutch mass reflects nutrient investment in the clutch, which may be regulated by nutrient reserves. Clutch mass is clearly correlated with clutch size, but female geese vary substantially in egg size and failure to account for egg size in analyses of regulation of clutch size will fail to account for substantial variation in investment in the clutch. In both snow geese (Ankney and Bisset 1976) and brant (Flint and Sedinger 1992), females with the same total investment in eggs can produce different clutch sizes because of variation in egg size. Therefore, using clutch volume in our analysis controls for variation in egg size, which varies by 78% among female brant (Flint and Sedinger 1992). We used adult body size as the covariate and age as a fixed factor in these analyses. We controlled for female age in our analysis of the relationship between clutch size (and volume) and female size, because clutch size may increase with female age in geese (Rockwell et al. 1983, Forslund and Larsson 1992). Most females used in analyses of the relationship between adult female size and clutch size were not initially color-banded as goslings. They were not, therefore, of known age. For these females, we estimated their minimum age by assuming they were 2-yr-old when first captured with adult plumage. Uncertainty about female age likely reduced our power to detect a relationship between female age and clutch size (or volume). Our principal goal in this and other analyses described below, however, was to assess the relationship between female size and investment in a clutch. To the extent that estimated female age reduced unexplained variation in these analyses, inclusion of female age increased our power to detect a relationship between body size and clutch size or clutch volume. To directly assess the relationship between fecundity and adult size, we performed an alternative analysis (ANCOVA) using clutch size as the dependent variable and PC1 score and mean egg volume as covariares. Age was again a fixed factor in this analysis. We also examined the relationships between clutch volume (and clutch size) produced by adult females and size of the same individuals when they were goslings, by using ANCOVA, with clutch volume (or clutch size) as the dependent variable, gosling PC1 score as the covariate, and age as a fixed factor. This analysis provided a direct assessment of the relationship between early growth and subsequent fecundity.