Persist or Give Up? Social Insect Decision-Making When Faced With Problems

Abstract

A social insect colony is a self-organized system with a few reproductive queens and multiple sterile workers performing various tasks cooperatively. As a result, the colony as a whole is a unit of selection. Most social insects are clustered within the order Hymenoptera. Haplodiploidy within this order (and beyond) leads to unusual relatedness among sisters, and this has been hypothesized to promote the evolution of cooperation. Workers can use different communication systems to coordinate recruitment to food sources and to perform various other tasks. Social insect colonies can face various problems, such as environmental perturbations, internal errors, and competition from other colonies. My thesis aims to understand how social insects respond to these problems and to understand whether haplodiploidy drives the evolution of eusociality. Specifically, my thesis accomplishes the following: (1) I applied a robustness mechanism categorization framework from systems biology to understand how individual Solenopsis xyloni workers respond to a perturbation of an already established pheromone trail using a classic T-maze setup. I found that S. xyloni workers abandoned the sucrose feeder when the already established pheromone trail was perturbed. Ants that managed to make a choice with the disrupted trail made random choices and did not reinforce the pheromone trail. However, when the sucrose concentration in the feeder was doubled, ants were less likely to abandon the trail, but continued to make random choices, and did not reinforce the pheromone trail. This study suggests a way forward to systematically study individual and collective responses to different perturbations. (2) I designed an agent-based model in Netlogo to examine how different types of errors affect collective foraging outcomes, and how different communication systems modify these outcomes. Overall, I found that errors led to decreased exploration and exploitation. Errors that had a high probability of occurring (false positive and forgetting errors) were the most detrimental to collective foraging performance. Communication overall improved resource collection rate and efficiency, but surprisingly led to an increased exploration for resources. In addition, communication presence mitigated the effects of errors rather than exacerbating them (i.e., there were no misinformation cascades). The results from my model suggest that the evolution of communication systems may also have been shaped by selection on exploration and robustness against different error types, as well as selection on efficient foraging. (3) I examined whether individual Temnothorax rugatulus ants can assess each other’s fighting abilities during a dyadic contest. I used contest duration as the response variable, and dry weight (proxy for body size) and head width (proxy for weaponry) as proxies of fighting abilities. I found that neither proxy of fighting abilities predicted the contest duration, indicating an absence of assessment. However, since I did not provide any resource for the ants to fight over, the ants may not have perceived this interaction as a contest. My study shows that individual contests are not identical to colony-level contests, and needs further investigation. (4) I investigated whether the evolution of eusociality is correlated with the evolution of haplodiploidy using a family-level hexapod phylogeny. I used three phylogenetic comparative methods (Pagel’s test, phylogenetic logistic regression, and D-test) to tackle this question. Two of three tests (Pagel’s test and D-test) show a clear support for haplodiploidy driving the evolution of eusociality. Experimental manipulations of data reveal that the non-significance of the logistic regression was due to the absence of haplodiploidy in aphids and termites. Overall, my study provides mixed support for the haplodiploidy hypothesis.