We wish to dedicate this paper to the memory of Professor A. J. Marshall whose untimely death occurred while it was in press. Professor Marshall was formerly the head of the Department of Zoology, St. Bartholomew's Hospital Medical College, London, and subsequently held the Foundation Chair of Zoology and Comparative Physiology, Monash University, Australia. Probably the most valuable of his many contributions to the field of avian physiology was his ability to relate laboratory experimentation to the natural environment. Much of the thought contained in the present paper has been stimulated by his approach and pioneer studies.