The hunger season is the time of year when many farmers in developing countries run out of their previous year's harvest and have trouble purchasing staple foods because this time of year has the highest prices. This paper constructs a theoretical model to address such seasonality of food deprivation, and by using three years of weekly household panel data, empirically tests the extent to which farmers in rural Zambia can smooth their consumption from season to season, as well as from year to year, in response to income shocks. The theoretical model provides an explanation of how farmers try to smooth their consumption. This paper allows for heterogenous impacts of seasonal price changes on consumption. Some farmers buy staple food when prices are low and store it for the hunger season, while others run out of staple food, and so buy it when prices are high. Results indicate that the former group successfully smooths its consumption from season to season, as well as from year to year. In contrast, the latter group reduces consumption of non-food items and non-staple food items, especially relatively soon after harvest. These results are well explained by the theory. The theoretical model, combined with empirical evidence, depicts consumption seasonality in great detail.