Summary
A fundamental assumption of sexual selection theory is that the reproductive advantage of large size is balanced by a survival disadvantage. Previous studies of the sexually size-dimorphic red-winged blackbird (Agelaius phoeniceus) have indicated that the largest adult males have a survival advantage, suggesting that the limit to male size may be the cost of getting big rather than the cost of being big. If the cost of getting big limits male size, then starvation rates for male nestlings should exceed those of female nestlings. In addition, given high heritability of body size, larger parents should lose more nestlings, particularly males, to starvation. We tested these predictions for red-winged blackbirds using data on the sex of 1356 fledglings from 465 nests collected over 10 years. We found no disadvantage for male nestlings relative to females – 49% of fledglings were male and previous research had shown that 48% of hatchlings are male. We also found no disadvantage for male nestlings that would become large adults (i.e. those with larger parents) – partial brood loss and fledging sex ratios did not vary with mid-parent size. Given no apparent disadvantage to large size for males either as adults or as nestlings, this leaves only the period between fledging and adulthood during which natural selection might limit sexual size dimorphism, although other mechanisms might explain the failure to find a limit to male size.
Methodology
We conducted this study from 1985 to 1993 and in 1995 at the Queen’s University Biological Station in eastern Ontario (45∞37¢N, 76∞13¢W) in beaver pond and lakeshore marshes. Nest searches were conducted from May to the end of July, encompassing the entire nesting season. We checked nests daily or every second day until they either failed or the young fledged. Nestlings were individually marked on the first nest visit after they hatched. On each visit we weighed each nestling and measured its tarsus length. More detailed information on the study area and general field methods are available from Weatherhead (1995). We assigned nestling sex based on the bimodal distribution of body size as nestlings approach fledgling (Weatherhead & Dufour, 2000). Sex was initially assigned based on the most conservative guidelines collectively provided by Holcomb & Twiest (1970), Fiala (1981) and Westneat, Clark & Rambo (1995) as follows: Age 7, < 26 g = female, > 33 g = male, Age 8, < 31 g = female, > 33 g = male, Age 9 or older, < 33 g = female, > 35 g = male. When the sex of a given nestling fell between the size thresholds, we visually inspected plots of the mass vs. age relationship for all individuals in the brood. Often, the sex of the nestling in question was obvious when its pattern of weight gain was compared with that of its known-sex nest-mates. If not, the nestling was classified as ‘sex unknown’. Using this approach we were able to sex approximately 90% of all nestlings that fledged. Although natal philopatry is < 3% in our study population (Weatherhead & Dufour, 2000), 48 birds sexed as nestlings were later seen as adults and all had been sexed correctly. We designated a nestling as having fledged using several criteria in addition to observing the nestlings leaving the nest or being outside the nest. If nestlings were gone on or after the date they would have been at least 9 days old (the earliest age at which fledging was observed), and the nest was undisturbed, we assumed they had fledged. When nestlings hatched asynchronously, we assumed all nestlings had fledged if they had all disappeared synchronously, the nest was undisturbed and the oldest nestlings were at least 9 days old. Because we applied these criteria uniformly, any bias in our assumptions should not influence our conclusions with respect to the hypothesis being tested. Testing the prediction that large parents should fledge fewer sons requires knowing the true parentage of nestlings, because the prediction is based on body size being heritable. Approximately 25% of all nestlings in our study population are sired by extra-pair males, so paternity analyses were required. From 1986 to 1991 and in 1993 we collected blood samples from all nestlings when they were 6 days old, although we only conducted paternity analyses using samples from three of our study marshes. Details of the methods used are available from previous publications (Gibbs et al., 1990; Weatherhead & Boag, 1997; Dufour & Weatherhead, 1998). We determined parental size using the first component (PC1) from a correlation-based principal components analysis of wing chord (± 1 mm), tarsus length (± 0.1 mm) and bill length (± 0.1 mm), conducted separately for males and females. Mid-parent size was taken as the average size (PC1 score) of the two parents. We also repeated all analyses using wing chord alone as our index of adult size, but because the results were qualitatively unchanged, we only report results obtained using PCA.