Quantitative genetics of life-history traits in a long-lived wild mammal.

Quantitative genetic studies of life-history traits in wild populations are very rare, yet variance/covariance estimates of these traits are crucial to understanding the evolution of reproductive strategies. We estimated heritabilities (h2) of several life-history traits (longevity, age and mass at primiparity, and reproductive traits) in two bighorn sheep (Ovis canadensis) populations, and both phenotypic (rP) and genetic (rA) correlations between life-history traits in one population. We included adult mass in our analyses because it is related to several life-history traits. We used the mother-daughter regression method and resampling tests based on data from long-term monitoring of marked females. Contrary to the theoretical prediction of low heritability for fitness-related traits, heritability estimates in the Ram Mountain population ranged from 0.02 to 0.81 (mean of 0.52), and several were different from zero. Coefficients of variation tend to support the hypothesis of a higher environmental influence on life-history traits. In contrast, at Sheep River we found low heritabilities of life-history traits. Phenotypic correlations varied between -0.09 and 0.95. Several genetic correlations were strong, particularly for different reproductive traits that are functionally related, and ranged from -0.34 to 1.71. Overall, genetic and phenotypic correlations between the same variables were similar in magnitude and direction. We found no phenotypic or genetic correlations suggesting trade-offs among life-history traits. Bighorn sheep may not form the large, outbred populations at equilibrium that are assumed by both Fisher's fundamental theorem and by theories predicting antagonistic pleiotropy between life-history traits. Alternatively, the absence of negative genetic correlations may result from genetic variation in ability to acquire resources or from novel environmental conditions existing during the study period.