The number and identities of species coexist-ing at a particular locality are not fixed for-ever, but result from dynamic interplay be-tween local extinctions and immigrations. While this dynamic structure must apply to any local community, it is particularly con-venient to study on islands, because of their sharp boundaries. In their well-known theory of island biogeography MacArthur and Wil-son (1963, 1967) proposed that the number of species on an island approaches an equilibrium between extinctions and immigrations, and that populations are subject to turnover. It is likely that large differences in turnover rates will be found if one compares the same taxo-nomic groups on islands of different area, iso-lation, latitude, and habitat, or if one compares different species on the same island. The measurement and understanding of these dif-ferences is beginning to emerge as a major empirical and theoretical problem, as a problem in assessing the importance of group selection and understand-ing the evolution of social behavior (Levins 1975, Wilson 1975:1X-116), and as a major practical problem in conservation strat-egy (Terborgh 1974a, 197413, Diamond 1975, 1976, Wilson and Willis 1975, Sullivan and Shaffer 1975). A convenient situation for studying these problems is provided by the land and fresh-water breeding birds of the Channel Islands off the coast of southern California. The eight islands of this group vary in area (A) from 2.6 to 249 km2, in distance from the mainland (d) from 20 to 98 km, and in number of land and non-swimming water bird species breed-ing in a given year (S,,,,) from 7 to 39 (see Philbrick 1967, Power 1972, Johnson 1972, and Jones 1975 for maps and table of areas and dis-tances). As of 1917, when A. B. Howell' s monograph Avifauna of the Islands off the Coast of Southern California was published, 170 species of birds had been recorded from the Channel Islands. By 1976 this number had increased to 325. All but five of these 155 additions were of non-breeding visitors, re-corded principally by us and by other observ-ers since 1968. The reason is that most of the early ornithologists who visited these islands were collectors and oologists mainly interested in obtaining specimens or egg sets of the breeding birds, some of which are considered endemic or nearly endemic subspecies. Conse-quently, the breeding avifauna of all the Chan-nel Islands was more thoroughly documented by 1917 than was the non-breeding avifauna (see appendix p. 545 for discussion). In 1968 one of us (J. D.) visited each of the Channel Islands l-3 times, plus the nearby Mexican island group Los Coronados, to sur-vey the breeding avifaunas for comparison with the surveys summarized by Howell (1917). On each island he found that there had been turnover, reflected both in dis-appearances of some former breeding popula-tions and in breeding presence of some for-merly absent or non-breeding species. A brief account of the results, and of estimated turn-over rates in relation to A, d, and Serl, was pub-lished (Diamond 1969). One of Diamond' s findings was that some breeding populations had immigrated and become extinct several times on the same island between the early 1909' s and 1968. Thus, surveys separated by many decades must have underestimated turn-over rates and might only have revealed the tip of the iceberg of community dynamics (see Diamond and May in press for examples). For organisms as mobile as birds, and islands as close to the mainland as the Channel Is-lands, it is likely that there will always be a long waiting list of potential immigrants; many more unsuccessful attempts at colonization than attempts that succeed even briefly; and many more brief successes and rapid failures than foundings of a colonist population that survives a long time. Hence repeated surveys at one-year intervals were clearly required to reveal the dynamic structure of the avifauna more accurately. Such studies of turnover at one-year intervals have the further advantage of reducing interpretative problems associated with habitat changes and effects of man which develop over longer intervals.