Preferred temperature in the warmth of cities: Body size, sex and development stage matter more than urban climate in a ground-dwelling spider

Abstract

Most ectotherms rely on behavioural thermoregulation to maintain body temperatures close to their physiological optimum. Hence, ectotherms can drastically limit their exposure to thermal extremes by selecting a narrower range of temperatures, which includes their preferred temperature (Tpref). Despite evidence that behavioural thermoregulation can be adjusted by phenotypic plasticity or constrained by natural selection, intraspecific Tpref variations across environmental gradients remain overlooked as compared to other thermal traits like thermal tolerance. Here, we analyzed Tpref variation of spider populations found along a gradient of urban heat island (UHI) which displays large thermal variations over small distances. We measured two components of the thermal preference, namely the mean Tpref and the Tpref range (i.e., standard deviation) in 557 field-collected individuals of a common ground-dwelling spider (Pardosa saltans, Lycosidae) using a laboratory thermal gradient. We determined if Tpref values differed among ten populations from contrasting thermal zones. We showed that endogenous factors such as body size or sex primarily determine both mean Tpref and Tpref range. The Tpref range was also linked to the UHI intensity to a lesser extent, yet only in juveniles. The absence of relationship between Tpref metrics and UHI in adult spiders suggests a Bogert effect according to which the ability of individuals to detect and exploit optimal microclimates weakens the selection pressure of temperatures (here driven by UHI) on their thermal physiology. Alternatively, this lack of relationship could also indicate that temperature patterns occurring at the scale of the spiders' micro-habitat differ from measured ones. This study shows the importance of considering both inter-individual and inter-population variations of the Tpref range when conducting Tpref experiments, and supports Tpref range as being a relevant measure to inform on the strength of behavioural thermoregulation in a given population.