Food limitation does not mediate fatal nurserymate aggression in a phytotelm-breeding frog
Any time the amount of care parents provide to a brood falls short of offspring demands, this resource limitation automatically generates conflicts-of-interest among family members. At the proximate level, such conflict can play out as aggression among siblings, aggression that is often fatal. When the loss of one or more brood members negatively influences parental fitness, parents should take action to prevent this outcome, for example by delivering more food. One proposed explanation for the evolution of parental feeding (in particular trophic egg feeding) is that this care reduces competition and curtails cannibalism. We tested this hypothesis in a phytotelm-breeding poison frog (Oophaga pumilio) where i) mothers feed tadpoles with trophic eggs, ii) the number of tadpoles sharing a nursery is variable, and iii) it is rare for more than one nurserymate to survive to metamorphosis. We staged intrusions (i.e., the addition of a new tadpole to an occupied rearing site) after manipulating the hunger level of resident tadpoles. If food mediates aggression, this would support the cannibalism-suppression hypothesis for the evolution of trophic egg provisioning, and suggest that parents have a mechanism by which to influence proximate competition within the nursery.
Resident hunger had no influence on the outcome of staged intrusions: aggression and the death of one tadpole was common and evenly distributed between treatments. Surprisingly, we found no evidence for cannibalism, a benefit of aggression that is perhaps too often assumed in similar systems. Two factors seem likely to drive resource-independent aggression in this species. First, the fitness benefits of trophic eggs to tadpoles accrue linearly, and there is no means by which an individual tadpole can monopolize parental deliveries if a nurserymate is present. Second, the oldest tadpoles were vulnerable to attack by younger intruders, so dispatching a nurserymate early may be the best defense against later falling victim to it. These results have broad implications for the ecology of phytotelm-breeding frogs. Complete offspring control of how many tadpoles can be supported in a single rearing site will exacerbate any abiotic or other biotic factors limiting rearing site availability, and thus ultimately shape habitat use, distribution, and abundance of this and similar frogs.