|| Checking for direct PDF access through Ovid
Despite more than 50 years of effort, the causes and mechanisms of small rodent population fluctuations remain unknown. The two major questions are as follows: (1) what is the cause of population decline and (2) what is the cause of cyclicity and its geographical variation? At present, no hypothesis can provide answers to both these questions. Recently, progress has been made by Boonstra (1994), who proposed the senescence hypothesis to explain the cause of cyclic decline in population numbers. Here, we tested the main prediction that voles in decline are older than in other phases of the cycle, by analysing changes in age structure in a fluctuating population of the bank vole (Clethrionomys glareolus). The results generally support this prediction; however, the differences in absolute age seem to be too small to explain the occurrence of senescent animals exclusively in declines. We propose a new model to explain changes in age structure and the mechanisms behind the decline and geographic variation in cyclicity. It is based on the idea that voles are oldest in declines, developed independently of Boonstra. However, it differs in three respects: (1) it is more general and thereby applicable to the whole cycle; (2) density-dependent changes in age structure are based on the bimodality in a female's age at first reproduction; and (3) it stresses developmental rather than physiological changes in the quality of decline of animals as being relevant to the rate of senescence. We propose that seasonality of the environment is a principal candidate to explain geographical variation in cyclicity. We present substantial theoretical and empirical evidence to indicate that in more seasonal environments with shortened vegetation periods, population dynamics is inevitably less stable due to increased variation in two critical parameters – age at first reproduction and the length of the breeding season – which determine population growth rates. Any external perturbation may then easily destabilize population numbers. The general applicability of the seasonality-senescence hypothesis to other mammalian species decreases with declining r and increasing life span. The hypothesis is falsifiable, and testable predictions are provided.