Nearly all societies share food beyond the household. Yet foods are not shared randomly: certain foods are more likely to be shared after acquisition than others. Typically such foods include the meat from large game animals or highly valued resources such as honey, while foods that come in rather small packages or are acquired with little variability tend to be rarely shared. One prominent hypothesis to explain variability in sharing patterns according to food type focuses on risk reduction, proposing that resources that are unpredictably acquired are more likely to be shared and more widely shared than those that are predictably acquired. Among the Meriam, we found that measures of risk did very poorly in predicting differences in the frequency of food sharing. We found significant differences in sharing frequency between hunt types that were not due to the sex of the acquirer or differences in harvest size. Among Martu, hunters (both male and female) who acquired low-variance goanna kept more for their families to eat after sharing distributions than did hunters who acquired kangaroo.  Among Martu, provisioning of public goods is argued to be a political strategy, whereby younger individuals gain the prestige needed to rise in the ritual hierarchy.