Here, I present 4 manuscripts on the evolution of defense in Temnothorax longispinosus acorn ants, which are the target of parasitic raids by the acorn ant species Temnothorax americanus. Temnothorax americanus raid T. longispinosus to steal their brood, timing their raids to coincide with the period when pupae are available in host nests. In the first chapter, I collected T. longispinosus nests May-October, censused nest characteristics, observed workers to determine their behavioral caste, and dissected them to measure the volume of venom in their venom sacs, in order to determine the ecological, behavioral, and nest characteristics that contribute to venom investment. I found that nurses have more venom when the pupa to worker ratio in the nest is higher, and that all workers have more venom earlier in the season, which correlates qualitatively with seasonal raid risk. Additionally, I found that foragers generally had more venom than nurses, and that foragers and nurses in larger colonies were more different from each other in terms of their venom volume than in smaller colonies. In the second chapter, I conducted raids in the lab to determine whether T. longispinosus colonies would emigrate in response to experiencing and surviving a parasitic raid, presumably to protect themselves against further exploitation. I found that nests that experienced raids emigrated significantly more than nests that did not experience a raid. In the third chapter, I conducted raids in the lab to determine whether T. longispinosus colonies and workers that have experienced a raid have induced chemical defenses, and produce more venom in the following two weeks in order to reduce the fitness consequences of further raids. I found that colonies that had experienced raids had more venom per capita, and that foragers from raided colonies had significantly more venom than foragers and nurses from colonies matched for genetics and size that had not experienced a raid. In the fourth chapter, I investigated whether polydomy, or the splitting of a single colony among multiple nests, satisfies the criteria for bet hedging to protect against parasitic raids, and investigated the conditions under which we would expect polydomy to be selected for to protect against catastrophic losses. I found that polydomy does meet the criteria to be considered a bet-hedging strategy against parasitic raids for some colony sizes. Additionally, I found that polydomy is sometimes favorable, meaning that the reduction in mean annual fitness and variance in fitness leads to an increase in expected lifetime fitness when following colonies for 10 years. Polydomy is most favorable when raid risk is low and when the queen is relatively young. Additionally, expected nest size distributions predicted by our model if every colony is following their optimal polydomy strategy matched reasonably with nest size distributions from observational field work in 2021, lending some credibility to the model and support for the idea that polydomy could be selected for as a bet-hedging strategy against parasitic raids in T. longispinosus. Overall, I found that individual traits, colony traits, and seasonality were all correlated with venom volume, indicating that T. longispinosus individuals adjust their investment in venom with expected risk. I also found that experiencing a raid affects venom investment for foragers and nest emigration probability, and that colony size, queen age, and raid risk all affect the predicted optimal polydomy strategy for a colony.