<p>In most cooperative animal societies, interactions between groups are hostile or antagonistic. However, in a few species, groups that encounter each other can fuse together to form a new functional collective. In some ants and termites, such fusion events can be followed by changes in investment in soldiers and reproductives, for reasons that are poorly understood. Here we use a theoretical model to investigate optimal allocation to soldiers following colony fusion. Each colony can invest a proportion of its membership into the soldier caste, which provides a public benefit for the whole two-colony collective but comes at a cost to the investing colony by limiting colony growth and therefore reproduction. Our model predicts that soldier investment will typically decrease following colony fusion. However, the reported post-fusion soldier investment from empirical studies could be explained if the cost of soldier production falls following fusion. The model predicts that a colony’s reproductive output will typically decline as a result of fusion, unless the other colony makes a greater investment, or the cost of soldiers decreases sufficiently. Finally, our model predicts that smaller colonies will produce fewer soldiers than larger colonies after fusion because their lower average relatedness to the rest of the collective makes each soldier less beneficial; this means that smaller groups might, under certain conditions, gain from fusing with larger groups. Our results demonstrate two potential pathways to the evolution of group fusion in the lower termites – exploitation of additional resources or exploitation of additional soldiers (or potentially other workers). Within-group conflict, not considered in our model, may explain some discrepancies between our model and empirical reports.</p>