to increase a little in numbers. Yet at non-breeding season on breeding islands the percentage of immatures increases enormously and adults or nearly adults are almost unprocur- able. Only a close study over some time in two areas- a breeding and a non-breeding area at the same time by sepa- rate observers - would show any definite results or enable one to arrive at definite conclusions with any degree of certainty. But I have an idea something such as the follow- ing is the case. Adult and second year birds (probably skull 3/4 or more ossified) arrive in a breeding area about October. Mating and nesting takes place. Probably one strong male may mate with several females and there may be no definite selection of mates and this is most likely the case. Juveniles of the previous year have to be ejected and driven away and this causes great stirs in the colony. One egg is laid, most birds laying in November but some earlier possi- bly and a number later. (Infertile eggs and addled eggs from previous year which are very common are ejected on to the floor of the cave.) As soon as the young which is attended by only one parent is able to fly adults all leave the area and do not return till the following year to breed again. This movement away takes place about April or May. A missionary on Lifu told me that early in June at the mission station at Chepenebe on Lifu great crowds of Collocalia gathered about the mission (even settling on a large tree a most unusual thing for Collocalia and after a day or two of much noise and calling seemed to disappear and he thought there were much fewer about later. I only got immature there