We studied food limitation as a proximate determinant of resource allocation in social insect colonies by repeating a field experiment that had previously produced negative results. A series of plots was supplemented with food in Vermont, USA, and others were left as controls. After two years of the food supplementation, all plots were excavated and nests of Leptothorax longispinosus ants were collected. We found significant effects of the food treatment on reproductive output, male allocation ratios, and the proportion of biomass allocated to reproduction versus growth. These results contradict a previous study on the same population of these ants, and we attribute the different outcomes to annual variation in the selective regime experienced by the ants. Thus interpretation of field experiments should be couched in terms of their temporal dimensions.