Temporal Variation in Survival of Mammals: A Case of Environmental Canalization?

In an age- or stage-structured population, a given change on a proportional scale of different fitness components has different effects on the population growth rate. Because earlier studies have shown that variability is selectively disadvantageous for long-lived iteroparous species, fitness components whose variation has greatest impact on the population growth rate are expected to be canalized against temporal variability. We present here a test for such a canalization of fitness components most influential for population growth. If canalization occurs, (1) within a given population the variance of canalized fitness components should be less than that of noncanalized components, and (2) among populations an inverse relationship should occur between the potential demographic impact of fitness components and their temporal variability. We tested these hypotheses using data on age-dependent survival of ungulates. As expected for long-lived vertebrates, elasticities and sensitivities of adult survival were consistently higher than for juvenile survival. In support of the canalization hypothesis, the variance of juvenile survival with low potential demographic impact on fitness was consistently higher than the variance of adult survival with high potential demographic impact on fitness, in all of 22 ungulate populations for which long-term demographic monitoring data were available. However, a negative covariation between potential impact on fitness and temporal variation could also be accounted for by alternative hypotheses, the lower resistance of growing individuals to environmental "insults," or reduced adult phenotypic variability due to selective mortality of juveniles. To differentiate between canalization of adult survival and frailty/selective mortality of juveniles, we included five populations of small mammals. We then found a negative relationship between the relative potential impact of age-dependent survival on fitness and the relative variability of age-dependent survival, especially for populations faced with high environmental variation. This remained true after correcting for differences in mean survival of juveniles and adults and was independent of phylogeny. Although more data are required to extend the hypothesis to other taxonomic groups, we conclude that available data support the idea that adult survival of ungulates is canalized against temporal variation. Therefore, contrasted responses of juvenile and prime-age survival to environ