A few unusual molt patterns are found in northern seabirds. The ivory gull _(Pagophila eburnea)_ has its major annual wing and body molt immediately before it breeds. In several other species such as the glaucous gull _(Larus hyperboreus)_ and Cassin's auklet _(Ptychoramphus aleuticus)_ the molt almost completely overlaps the reproductive cycle (Johnston 1961; Payne 1965). Potts (1971) documented a molt pattern in shags _(Phalacrocorax aristotelis)_ which is more typical of tropical seabirds. Several cycles of wing molt take place simultaneously, each lasting more than a year, and molt ceases in winter. By the time breeding age is reached, each flight feather is replaced once a year.
Within these broad categories of molt pattern there are sometimes variations according to age, sex, and even subspecies (Stresemann and Stresemann 1966). Male common eiders _(Somateria mollissima)_ molt directly after mating, when their reproductive role is completed, whereas females molt only after they have taken their young to sea. Nonbreeders and failed breeders frequently begin molt while other adults are still raising young and not molting--e.g., many alcids, gulls, storm-petrels, and fulmars (Stresemann and Stresemann 1966; Ingolfsson 1970; Harris 1971; Harris 1974; Sealy 1975_b_). In ivory gulls, which molt just before reproduction, and in Sabine's gulls, which complete molt just before breeding, nonbreeders may extend wing feather growth into the breeding season (Stresemann and Stresemann 1966).
There is little information on the energetic cost of molt, although there are indications of at least some expense. Belopol'skii (1961) showed that nonmolting seabird species tended to gain weight after reproduction, whereas those that immediately started molt tended to lose weight. Among other birds, however, it is common for individuals to gain weight just before, and even during molt (Payne 1972). The BMR is known to rise in molting birds (Blackmore 1969; Lustick 1970; Payne 1972), from as little as 5% to as much as 34% above nonmolting levels. In one study, about 35-40% of the increased BMR represented extra thermoregulatory costs incurred by lessening of insulation and increase in heat loss from well-vascularized new feathers; the rest of the increase represented the energetic cost of growing feathers (Gavrilov 1974). The fact that molt rarely overlaps with breeding suggests that the energetic cost, even if slight, may be incompatible with the already high costs of reproduction (Payne 1972). Cassin's auklets, which do molt while breeding, may cease molt while feeding large young (Payne 1965), and certain species interrupt molt during migration (Stresemann and Stresemann 1966). Doubtless a rapid simultaneous molt is more costly than a long gradual one.