Reciprocal altruism, vampire bats, and relatedness

Vampire bats (Desmodus rotundus) are social mammals, which naturally determine kinship differences, form dyadic relationships that remain consistent over time, and practice food sharing with roost-mates who were unable to feed. These dyadic relationships occur between both kin and non-kin. Vampire bats can die if unable to feed for more than 70 hours. Roost-mates may regurgitate food into the mouth of the unfed roost-mate.

An experiment was conducted on 20 vampire bats (11M, 9F) over a 2-year period to determine whether this food sharing behavior was altruistic. The bats were housed in a flight cage, large enough to allow the bats to freely associate, for more than 2 years before the beginning of the study. The bats were removed, fasted for 24 hours, re-introduced into the roost, and their social interactions were recorded for two hours after reintroduction. None of the bats tested showed signs of parasite infestation. Researchers examined food sharing (licking the roost-mate’s mouth for more than 5 seconds with > 5 seconds between bouts of licking) and subsequent allogrooming (licking locations other than the roost-mate’s mouth). The bat’s behavior was monitored to determine subsequent interactions. The coefficient of determination values for food donated, donor sex, allogrooming, and relatedness are shown in Figure 1. Sharing happened most often between females; males were not recorded sharing food with other males. Donors initiated food sharing more often than fasted bats and three donors, on average, fed each of the fasted bats. There was no relation found between recipient (age or size) and donations or allogrooming received. Some donors rejected certain begging (harassing) recipients; however, some fasted subjects rejected food sharing from certain perspective donors.

Figure 1: Relative variation in food donated; 38% of the variation in the amount of food donated can be explained by this model. Food donated (p < 0.0002), donor sex (p < 0.0002), and allogrooming (p < 0.0002) received were all significant; however, relatedness was not (p = 0.16).