Mutualism is often a complex interaction among multiple species, each of which may respond to temperature differently. We studied the thermal ecology of the mutualism between the fishhook barrel cactus, Ferocactus wislizeni, and its common ant defenders, asking (1) how temperature affects the frequency of ant attendance and (2) how temperature affects the interspecific distribution of ants on the plants. Plants tended by Crematogaster opuntiae, Solenopsis aurea, and S. xyloni have higher fruit-set than plants tended by Forelius pruinosus. We measured surface temperature and the number of ants foraging (a measure of partner effectiveness) on 12 plants, each occupied by a single ant species. We tested for an interspecific difference in the optimal temperature (Topt; at 100% of the maximum activity), the thermal mean (average temperature at which ants are active), and the distribution of ants on plants in response to average plant temperature. To investigate the thermal tolerance of each ant species, we conducted critical thermal maximum (CT_max; the temperature above which ants cannot forage) experiments using a heated circulator and water bath to determine the temperature at which 50% of the ants die. 

Results/Conclusions:

Ant activity on the plant was significantly correlated with surface temperature, with a maximum peak in activity at moderate temperatures and a decline at colder and hotter temperatures. There was no evidence for an interspecific difference in Topt (ANOVA, P = 0.3064). There was a significant difference in the average temperature at which the ants were active; the order, from higher temperatures to lower temperatures, is F. pruinosus > S. xyloni > C. opuntiae >S. aurea (Tukey-Kramer HSD). There is an interspecific difference in the distribution of ants on plants in response to average plant temperature (ANOVA, P < 0.0001). F. pruinosus occupies hotter plants than C. opuntiae and S. aurea, but not S. xyloni (Tukey-Kramer HSD). Although the thermal performance curves suggest that F. pruinosus, the lowest-quality ant mutualist, has a higher Tmax than the other species, it was not possible to distinguish the Tmax from these observations for comparison. These results may be sufficient to explain observed differences in ant attendance and the effect of temperature on partner effectiveness. Extremely little is known about the thermal ecology of species interactions; further research is essential to understand how they will respond to rising temperatures worldwide.