Gannets were studied on the Bass Rock (Scotland) between 1960 and 1963, and particularly in a colour ringed group (the observation colony). The account is in two parts (general breeding biology and behaviour) which are amalgamated in this brief summary. The entire Gannet population of the Bass was estimated in June 1962 and a considerable increase over the 1949 population (last count) shown (about 4,800 pairs in 1949 to about 7,000 pairs in 1962, excluding 'club' birds). The increase in the observation colony was mapped in detail, for the period of the study. The mid-cliff regions of the Rock are the first to be re-populated each year. In the observation colony (and presumably in all areas) old pairs return before newer ones, and males before females. The maximum seasonal stay at the breeding Colony is from late January to early November. During this period both sexes actively defend the nest site by fierce fighting, where necessary, and a special aggressively-motivated 'ownership' display, bowing (analysed in detail). Sites are guarded continuously and males gather nest material throughout (females only after egg laying and for a much shorter period). Nest sites and mates are usually permanent, though females show less attachment to the site than do males (an attempt to separate the two effects is made). New sites (which may be on cliff ledges of various sizes or on flatted ground) are established only by males usually four years old (beginning in April) and are held a year before breeding is attempted. Birds tend to return to the same small area of the Colony from which they originated. Site establishment normally involves fighting (discussed in detail). Pair formation takes place only at the site; the male performs an advertising display - a modified form of bowing and females 'prospect' for such males. Males are conspicuously aggressive to females - especially in new pairs but also throughout life - and females show a high tolerance of male attack and an appeasement posture (facing-away) restricted to this situation. Males bite their mates whenever they meet on the site, and the pair then perform a prolonged 'friendly' meeting ceremony (mutual fencing). This, like male advertising, is a modified form of bowing. Laying begins late March or early April and continues (by first-time breeders) until late June or exception- ally the first half of July. Older females lay earlier and produce heavier eggs. Greater density also probably causes earlier and more closely synchronised laying. Incubation behaviour in both sexes was, in some cases, released by donated eggs, but others were refused close to the laying date of the pairs concerned. Males take slightly longer incubation stints than females. No two-egg clutches (except the product of different females) were found, though Gannets usually replace a lost egg in 6 - 32 days (first-time breeders significantly less often than experienced females). However two eggs are incubated as successfully as one. The incubation period, 43.6 days, is the same for new and experienced birds. Eggs lose 10 - 13% in weight during incubation (under- foot). Eggshells are not systematically removed. Hatching success for birds breeding at least the third time was 86% and for first-time breeders 62.5%. The nidicolous young may be fed immediately on hatching. Feeding, by incomplete regurgitation, continues for the entire pre-fledging period (no starvation period). The length of attendance spells drops sharply after hatching. Even during the phase of maximum chick growth, the pair spend some 15% of daylight hours together at the nest. The young reach a maximum of 150% of the adult weight. Starvation among chicks was never found during the study period. Juveniles leave the nest at about 90 days (both from new and experienced breeders). However the former lost more small chicks and had much lower overall success (49% of eggs laid gave fledged young in new pairs, 82% in experienced pairs). The growth of young was followed in detail. Parents do not discriminate in favour of their own young and will accept substitutes, probably at any stage in the chick's growth, and even when natural and foster chicks differ markedly in age. However adults repel wandering chicks and even attack unguarded ones. Furthermore, chicks normally stay strictly on their nests. These factors prevent doubling up. Artificially twinned nests revealed that if the chicks were of about the same age both survived and were adequately fed, though fledging at about 94 days and growing slightly slower than singles. A significant age difference (three or more days) however led to the persecution of the younger by the older and hence its starvation. Excluding such cases, pairs with twins gained an 80% reproductive advantage in 1962, the year of the experiment. The implications of the twinning results are discussed together with other factors (deferred maturity, non-breeding population, etc.) affecting recruitment rate. Return of colour ringed adults over the three years of the study gave a 6% annual adult mortality and thus a life expectancy of 16.2 years. Mortality between fledging and returning to breed is calculated to be about 80%. The ontogeny of behaviour in chicks is described. Juveniles show characteristic pre-leaving behaviour and fly well at the first attempt, though cannot rise for some days having alighted. They are not accompanied by parents and do not return to the nest. Adults tend to attack them on the sea. They may return to the breeding Colony in their first year, but do not normally do so until two or three years old. Immature plumage stages are described and illustrated. Body maintenance activity (preening, sleeping, plumage shaking, etc.) is described. Rotary head shaking is a response to peripheral tactile stimulation and occurs as a probable displacement reaction in some fear situations. The ordinary sideways head shake is shown to be a simple movement which has been incorporated into several complex displays and also occurs alone, in ritualised form, in at least one signal situation. Social behaviour away from the nesting site is discussed and contrasted (in complexity) with breeding behaviour. A special posture (sky-pointing) which precedes and accompanies movement away from the site is discussed here. Its comparative occurrence within the Sulidae is also discussed. Finally adaptations to cliff nesting in the Gannet and Kittiwake are compared. The two show many similarities, evolved convergently, but also several important differences (apart from those inevitably resulting from dissimilar phylogeny). The general discussion centres round the importance of the site and associated aggression in the Gannet's breeding biology.