The queenless ant Pristomyrmex punctatus has two types of worker: cooperators who do the colony's work, and cheaters who exploit the cooperative infrastructure without contributing. Without a queen to enforce division of labor, the colony faces a classic tragedy of the commons — cheaters free-ride on cooperators until the cooperative structure collapses.
Joy Das Bairagya and Sagar Chakraborty model these colonies as metapopulations with asymmetric migration. Cooperators and cheaters move between colonies at different rates, and colonies show territorial hostility to immigrants. The key finding: hostility to immigrants prevents the tragedy of the commons. By making the territory inhospitable to newcomers, colonies block the influx of cheaters that would overrun the cooperative workforce.
The counterintuitive detail is the gradient of hostility. The ants are more hostile to immigrants from nearby colonies than to those from far away. Kin selection predicts the opposite — nearby colonies share more genetic material, so neighbors should be welcomed. But the hostility gradient tracks exploitation risk, not kinship. Nearby colonies are close enough for their cheaters to migrate regularly. Distant colonies' cheaters rarely arrive. The hostility is proportional to the actual probability of defector immigration.
The through-claim the paper demonstrates but doesn't frame: the target of hostility reveals the nature of the threat. If the ants were defending against predators, the gradient would track predator density. If defending against disease, it would track pathogen exposure. The fact that it tracks proximity — and therefore defector migration probability — tells you that the primary threat to the colony is not external danger but internal exploitation by immigrants who look like workers but behave like parasites.
Hostility is usually framed as a cost — wasted energy, missed cooperation opportunities. Here it's structural insurance against a specific failure mode. The cost of hostility is the price of maintaining cooperation in a system where cooperation is exploitable and exploitation is mobile.