132. The majority cross in October or November and return in February, March. Some few may remain, (some if not all young ones certainly do) on Uvea all the year round, young possibly mating and crossing with others in their second year. This movement has been caused I believe by general scantiness of food on the island, this having been reduced by a hurricane which, though destroying food, did not destroy adult birds, who ranged wide in search of food and reached Lifu. Next season they returned to Uvea to breed and later when the lean season arrived, rather than struggle through it they again left for Lifu. The annual movement having been more or less continued to this time. Even so Uvea is a most unsuitable island for more than a few Columba and could never carry a large population under present conditions and it seems strange that there should be any urge for the birds to return there from Lifu. Only a strong urge to return to the place of birth or pre- vious nesting can account for such movement. It is very noticeable on Uvea that this species eats fruits on that island, which it does not eat elsewhere. Young birds, according to native reports, eating fruits which seem entirely unsuitable and which in other areas I have only rarely heard of the species eating at such times as hurricanes have destroyed all their usual foods and ap- parently only doing so in desperation from hunger. Food during the seasons when the birds are more plentiful on Uvea is abundant and varied and could at that season carry a much larger population than is present, though such sup- plies are subject to sudden or total depletion by hurricanes.