Temperature drives caste-specific morphological clines in ants.

The morphology of organisms relates to most aspects of their life history and autecology. As such, elucidating the drivers of morphological variation along environmental gradients might give insight into processes limiting species distributions. In eusocial organisms, the concept of morphology is more complex than in solitary organisms. Eusocial insects such as ants exhibit drastic morphological differences between reproductive and worker castes. How environmental selection operates on the morphology of each caste, and whether caste-specific selection has fitness consequences is largely unknown, but is potentially crucial to understand what limits ant species' distributions. Here we aimed to examine whether ant shape and body size covaries with climate at the scale of an entire continent, and whether such relationship might be caste specific. We used 26,472 georeferenced morphometric measurements from 2,206 individual ants belonging to 32 closely related North American species in the genus Formica to assess how ant morphology relates to geographic variation in the abiotic environment. Although precipitation and seasonality explained some of the geographic variation in morphology, temperature was the best predictor. Specifically, geographic variation in body size was positively related to temperature, meaning that ants are smaller in cold than in warm environments. Moreover, the strength of the relationship between size and temperature was stronger for the reproductive castes (i.e. queens and males) than for the worker caste. The shape of workers and males also varied along these large-scale abiotic gradients. Specifically, the relative length of workers' legs, thoraxes and antennae positively related to temperature, meaning that they had shorter appendages in cold environments. In contrast, males had smaller heads, but larger thoraxes in more seasonal environments. Overall, our results suggest that geographic variation in ambient temperature influences the morphology of ants, but that the strength of this effect is caste specific. In conclusion, whereas ant ecology has traditionally focused on workers, our study shows that considering the ecology of the reproductive castes is imperative to move forward in this field.