Apparent heritability of parental care in Savannah Sparrows

Parental effort measures the total amount of time and energy allocated toward care of offspring, and it is often an obvious and variable component of an animal's life history (Winkler and Wilkinson 1988, Clutton-Brock 1991). Among birds, intraspecific variation in parental effort is more quantitative than qualitative. Individual adults differ in feeding rate (Yasukawa et al. 1990, Wright and Cuthill 1992), intensity of offspring defense (Winkler 1992), and tendency to engage in prolonged associations with young (Weatherhead and McRae 1990).

Like all quantitative traits, intraspecific variation in parental effort can be attributed to genetic and environmental causes. However, recent work has focused mostly on environmental sources of variation (Ketterson and Nolan 1994, Gowaty 1996a), and individual birds often are viewed as decision-making machines that respond adaptively to changes in their physical or social environments (Winkler and Wilkinson 1988). Indeed, birds modify parental effort in response to numerous proximate factors, including genetic parentage (Lifjeld et al. 1998), brood size (Ruusila and Poysa 1998), and offspring age and condition (Whittingham and Robertson 1993). In no case have environmental factors accounted completely for differences among individuals in the intensity of parental care, however, and some role for genetic or cultural determination seems likely.

For quantitative traits, genetic and environmental effects can be distinguished by examining the phenotypic resemblance of relatives. In cases of high narrow-sense heritability (h2), additive genetic effects contribute substantially to overall phenotypic variation, and relatives strongly covary (Falconer and Mackay 1996). Among vertebrates, estimates of heritability are lacking for measures of parental effort and for most quantitative behaviors in the field (Hailman 1986, Boag and van Noordwijk 1987, Mousseau and Roff 1987; but see Waser and Jones 1989). In general, heritabilities for behavioral traits are low (Lemon 1993, Berthold and Pulido 1994), suggesting that quantitative differences among individuals are largely driven by the environment.

Here, we examine the degree to which male and female feeding rates are heritable in an insular population of Savannah Sparrows (Passerculus sandwichensis). We show that quantitative differences in feeding rates among males can be attributed to the behavioral phenotypes of their fathers, suggesting that parenting strategies are substantially less flexible than commonly presumed.

Methods.-We studied the Savannah Sparrows breeding at the Bowdoin Scientific Station on Kent Island, New Brunswick, Canada (4435'N, 6646'W) from 1992 to 1995. Every adult in the population bore a unique combination of three colored leg bands. Daily censuses of the 7.5-ha study site allowed us to identify social pairings, the hatching date of all successful clutches, and the reproductive performance of adults. We quantified parental feeding rates during 1.5-h (1992) or 2-h (1993 to 1995) observation periods six days after the young hatched. To maximize the number of nests included in the study, we made no attempt to standardize observation periods for time of day or weather. Feeding rates were highly repeatable within years, suggesting that these short sampling periods were of sufficient duration to detect real behavioral differences among individuals (Wheelwright et al. 1992, Freeman-Gallant 1996). Additional details on field methodology are provided in Freeman-Gallant (1998).

Kent Island Savannah Sparrows are highly philopatric (Wheelwright and Mauck 1998), and the return of yearlings to the study site allowed us to compare the behavioral phenotypes of 24 adult offspring with their parents. Each of these offspring came from a different brood, so 24 families were available for analysis. In an independent data set, we also examined the repeatability of feeding rates across years for 19 females. In all cases, offspring and parents were observed by field technicians who had no knowledge of the genealogy or prior behavior of the sparrows.

Because 15 to 33% of male Savannah Sparrows on Kent Island are polygynous each year, and because males only rarely provision the nests of secondary females, we used feeding effort directed toward the oldest (first-hatched) nest only. Similarly, most female feeding rates were from social situations in which the female was the only mate of a monogamous male or the primary mate of a polygynous male (as determined by hatching date; Freeman-Gallant 1997). In four cases, limited data forced us to use feeding rates from secondary females. No results reported here are changed qualitatively by excluding these four cases.

Narrow-sense heritabilities were derived from simple linear regressions of offspring on male or female parents (slope =0.5h2; Falconer and Mackay 1996). Because sparrows did not mate assortatively (likelihood-ratio test of correlation between male and female feeding rates; r = 0.21, P > 0.20, n = 22), heritability estimates were uncorrected for the behavioral resemblance of mates. For two offspring, we had information on paternal feeding rates only. In all cases, residuals were approximately normal and homoscedastic, which justified the use of standard parametric tests.