An ant outline comprised of zeros and ones symbolizes computer modeling of social insect behavior.
An ant outline comprised of zeros and ones symbolizes computer modeling of social insect behavior.

Division of Labor: Key to Evolution of Multicellular Life?

The cost of switching from one task to another, rather than increased efficiency through specialization, could be the driving force behind the evolution of division of labor, suggests a new study involving digital organisms.
Aug. 8, 2012
An ant community in the lab.
An ant community in the lab.

Dividing tasks among different individuals can be a more efficient way to get things done, whether you are an ant, a honeybee or a human.

A new study published in the Proceedings of the National Academy of Sciences (PDF) suggests that this efficiency may also explain a key transition in evolutionary history: from single-celled to multi-celled organisms.

The scientists found the cost of switching between different tasks gives rise to the evolution of division of labor in digital organisms. In human economies, these costs could be the mental shift or the travel time required to change from one activity to another. 
According to Anna Dornhaus, an associate professor in the Univeristy of Arizona department of ecology and evolutionary biology who collaborated with researchers at the BEACON Center for the Study of Evolution in Action at the University of Michigan in East Lansing, Mich., social insects often are thought to derive their evolutionary success from delegating tasks amongst highly specialized individuals, allowing the whole colony to be more competitive than groups lacking such organization. 
However, previous research in her lab involving more than 1,000 individually marked ants failed to support that assumption. Instead, it appears that the primary benefit of division of labor comes from avoiding the costs of task switching, a hypothesis proposed by economist Adam Smith in the 1700s but not tested in social insects until now.
Led by Heather Goldsby, now a postdoctoral researcher at the University of Washington, the team created a virtual ant colony of self-replicating computer programs and imposed a time cost on the digital organisms that had to complete various computational tasks to reap rewards. 
“More complex tasks received more rewards,” Goldsby said. “They evolved to perform these more efficiently by using the results of simpler tasks solved by neighboring organisms and sent to them in messages.” 
In this way, the organisms were breaking the tasks down into smaller computational problems and dividing them up among each other.
“Our idea here was to emulate a system with relatively simple individual workers in a computer, to see if the division of labor that evolved in social insects would emerge even given just a few assumptions,” Dornhaus said. “Indeed it turns out that gains in individual efficiency are not necessary for specialization to evolve.”
What’s more, the researchers discovered that even task allocation based on spatial position or communication signals, both strategies found in social insects, might evolve from a set of simple rules, such as, “If you find yourself in position X, do Y.” 
“Task allocation based on spatial position means every job is located somewhere, and if I happen to be standing at that point, I do it,” explained Dornhaus. “No one else will attempt to do it because they would bump into me. This ensures that not everyone tries to do the same thing, and if you have good ‘coverage’ of individuals, all tasks will get done.” 
The division of labor did not come about by bringing together individuals with different abilities – each member of a community was genetically identical, in the same way that all of the cells in a human body contain the same genetic material. Instead, the organisms had to have flexible behavior and a communication system that allowed them to coordinate tasks.
The authors said the most surprising result was that the organisms evolved to become dependent on each other.
“The organisms started expecting each other to be there, and when we tested them in isolation, they could no longer make copies of themselves,” said Charles Ofria, an associate professor of computer science and engineering at Michigan State who co-authored the study. 
The team’s findings have major implications for understanding the transition from single-celled to multi-cellular life forms. 
“In embryonic development as in the evolution of multicellularity, the initially identical cells divide up the tasks such that some become skin, others become muscle, et cetera,” Dornhaus said. “Not having to switch between these jobs makes them more successful, even if they potentially could do all of them since they all have the same genetic potential. Our result means that multicellularity might have evolved even if cells did not individually become more efficient.”