Parrots live in smaller groups on islands

Species living on islands are predictably different from their mainland counterparts in morphology and behaviour, but the source of these differences is still debated. Islands, in particular, are characterized by depauperate predator communities. Relaxed predation pressure might explain why living in groups, a common anti-predator adaptation in animals, is considered less likely on islands. However, the empirical evidence for this effect is scant and alternative explanations have been overlooked. For instance, smaller groups might be more common because island species are less studied, because more stable food supply associated with benign climate on islands favours territoriality or because population density is too low to allow the formation of larger groups. I examined the determinants of foraging group size in parrots, a large worldwide avian order with many island populations. Using a multivariable phylogenetic framework, I found that foraging group size was smaller on islands than on the continents controlling for ecological variables known to influence group size such as diet and body size. In addition, the island effect persisted after controlling for research effort, climate variables, and population density, suggesting that impoverished predator communities are an important driver of group size on islands.

Methods
I searched the literature for studies reporting foraging group size in parrots. I excluded extinct species or species extinct in the wild. As nearly all parrots are diurnal, I excluded nocturnal species because timing of activity can influence group size. To obtain the minimum and maximum group sizes for each species, I started with two general monographs on parrots and updated these values by carrying out searches from primary sources on Google Scholar combining each species name and the key words flock or group. I did not consider group sizes associated with flying or roosting groups when that distinction was made. I also did not consider group sizes coming from studies conducted in cities or with species outside their native range because of possible anthropogenic effects on group size. In addition to foraging group sizes, I determined whether the data for a particular study came from an island or the continent as some species of parrots are known to live on islands as well as the continent. Other traits were taken from the published literature.