When reproductive failure becomes chronic as observed on peregrine falcons (Falco peregrinus) by Hickey (1969) and in ospreys (Pandion haliaetus) by Ames and Mersereau (1964), the population of adults may hold on for a number of years without evident decline. Damage to the structure of the whole population may be serious before any numerical results are evident.
• Although there may not be intensive competition for food in the habitat away from breeding colonies, there is intense competition for food and breeding sites at and around the colonies. Hence age and previous experience in seabirds assume importance in establishing territory and in breeding success. Associated with this is the tendency for immature birds to delay breeding until they are several years old and for the immatures to remain on feeding grounds at some distance from the colonies. In some cases young birds may "hang around" breeding colonies and even feed some of the young. When young birds do first breed they usually lay smaller clutches and raise fewer young than do older birds. The importance of age and experience upon breeding success has been well documented for kittiwakes (Coulson 1966) and red-billed gulls (Mills 1973).
The fundamental biological importance of this delayed maturity seems to be emphasized by the persistence for several years of immature plumages, so clearly identifiable that even a human observer can recognize the age of an individual. One assumes such an evident feature must have adaptive significance.
Wintering Grounds
When colonial nesting seabirds leave their breeding islands for their wintering grounds, their identification with that island is lost as far as population effects are concerned, because birds from many colonies mingle on the wintering grounds. Major mortality takes place on the wintering grounds and must therefore act on the species population as a whole rather than differentially on individuals associated with especially dense colonies. Such a direct relation between colony density and mortality would be necessary for density-dependent mortality to regulate the number of birds on a breeding colony. Conversely, one cannot expect that all colonies will decrease equally because mortality should be equally distributed if all the population gathers on a common wintering ground. Thus density-dependence acts only in a very general way upon the sum of animals considered as an abstract entity—the population.
In fact, on the wintering grounds, as shown by a graph of numbers of gulls reported on Christmas Counts on Cape Cod, Massachusetts (Kadlec and Drury 1968), herring gulls are very responsive to local conditions and move several tens of miles to gather at favorable feeding sites. An aerial survey of the gulls on the East Coast of the United States (Kadlec and Drury 1968) showed that more than half of the gulls were gathered near major food sources in large metropolitan districts. Most of the remainder were gathered near small fishing ports. Very few were scattered along the shoreline in what one assumes is the traditional gull habitat. Later analyses of the relation between the distribution of banding recoveries of birds in their first winter and the distribution of immatures as found on this winter census (Drury and Nisbet 1972) suggested that proportionately more first-year gulls died in those areas where the birds were sparsely distributed than died in the crowded metropolitan areas.
These results suggest both that there is not a direct feedback between reproductive rate and mortality, and that mortality may even be inversely density-dependent on wintering grounds. This last runs counter to traditional ecological ideas that density causes a change in mortality rate. The idea that individuals gather where "living is easy" and mortalities are low is consistent with the theory of natural selection. One would not expect the food of the gulls to be evenly distributed, and one would expect individuals to move away from areas where food is scarce and mortality is high.
Differences in Breeding Success Between Colonies
Breeding success has been shown to vary among individual pairs of gulls (Drost et al. 1961). Certain groups of individuals nesting in patches within a single colony have greater breeding success than do others (Coulson 1968; Drury and Nisbet, in preparation). Differences in breeding success also occur between colonies (Frazer-Darling 1938; Kadlec and Drury 1968; Drury and Nisbet 1972). Some colonies reproduce consistently better than others—for example, the gull colonies close to fishing ports and metropolitan areas. Other colonies produce consistently fewer young, such as the colonies on the outer islands in the Gulf of Maine (Drury 1963; Kadlec and Drury 1968; Hunt 1972). The populations of successful colonies grow while the numbers of unsuccessful colonies decline, even during a period of general population increase (Kadlec and Drury 1968).
The difference between success and failure, growth and decline, appears to lie in the food available. Colonies increase where breeding success is high and decrease where breeding success is low. One important reason seems to be that adult gulls may move to a more productive colony even after they have nested with another colony (Drury and Nisbet 1972; Kadlec 1971). Such adaptations can be viewed as adjustments by which individuals meet the requirements of an environment in which the availability of food and other necessities is patchy and shifting.