Division of Labor: from self-organization to social selection
1. Division of labor in harvester ant queen associations
In most harvester ant species, newly mated queens found nests as single individuals. However, we can force these queens to form artificial social groups by placing them together in laboratory nests. When we do, a spontaneous division of labor emerges, in which one queen tends to take over the task of excavation while the other one tends brood. The role a queen takes in these associations seems to be based on intrinsic differences in their sensitivity to different tasks. They bring these different task preferences with them into the association, but once there, these initially small differences are amplified into a quite strong system of task specialization. This emergent division of labor is interesting because (a) it seems to appear at the very beginning of social cooperation, and (b) it generates associated emergent fitness disparities. In the case of our ant associations, the queen who becomes the excavation specialist often suffers higher mortality.
How might division of labor as an emergent property affect social evolution? Our data show that levels of division of labor in evolved cooperative foundress associations are lower than the emergent division of labor in our forced associations. This suggests that emergent division of labor may actually act as a barrier to the evolution of cooperative sociality. Read more on this topic in: Fewell and Page (1999); Helms Cahan and Fewell (2004), Jeanson and Fewell (2008); Clark and Fewell (2014).
2. Division of labor in groups of normally solitary Sweat Bees
The genus Lasioglossum shows multiple transitions from solitary living to cooperative nest sharing. How is the transition to cooperation connected to the emergence of division of labor? Our data suggest that task specialization and division of labor emerge in artificial groups of normally solitary queens – even though they have no history of eusociality. Our work in this system reinforces the findings from our harvester ant studies: that division of labor can spontaneously self-organize in social groups, and that it appears even at the origins of sociality. Our comparisons of division of labor in solitary versus communal-cooperative species of Lasioglossum suggest that cooperation involves a transition from task specialization to task sharing; we see a similar trend in our harvester ant associations, again suggesting that specialization and division of labor generate potential costs for cooperative social groups. More on this topic can be found in Jeanson et al (2007); Holbrook et al (2013).