Life history and habitat do not mediate temporal changes in body size due to climate warming in rodents

Abstract

Temporal changes in body size have been documented in a number of vertebrate species, with different contested drivers being suggested to explain these changes. Among these are climate warming, resource availability, competition, predation risk, human population density, island effects and others. Both life history traits (intrinsic factors such as lifespan and reproductive rate) and habitat (extrinsic factors such as vegetation type, latitude and elevation) are expected to mediate the existence of a significant temporal response of body size to climate warming but neither have been widely investigated. Using examples of rodents, we predicted that both life history traits and habitat might explain the probability of temporal response using two tests of this hypothesis. Firstly, taking advantage of new data from museum collections spanning the last 106 years, we investigated geographical and temporal variation in cranial size (a proxy for body size) in six African rodent species of two murid subfamilies (Murinae and Gerbillinae) of varying life history, degree of commensality, range size, and habitat. Two species, the commensal Mastomys natalensis, and the non-commensal Otomys unisulcatus showed significant temporal changes in body size, with the former increasing and the latter decreasing, in relation with climate warming. Commensalism could explain the increase in size with time due to steadily increasing food availability through increased agricultural production. Apart from this, we found no general life history or habitat predictors of a temporal response in African rodents. Secondly, in order to further test this hypothesis, we incorporated our data into a meta-analysis based on published literature on temporal responses in rodents, resulting in a combined dataset for 50 species from seven families worldwide; among these, 29 species showed no significant change, eight showed a significant increase in size, and 13 showed a decline in size. Using a binomial logistic regression model for these metadata, we found that none of our chosen life history or habitat predictors could significantly explain the probability of a temporal response to climate warming, reinforcing our conclusion based on the more detailed data from the six African species.