In this study we describe the pattern of parental investment in the pink-footed goose (Anser brachyrhynchus) during the fledging period in Iceland, concentrating particularly on the analysis of vigilant behaviour as one important element of parental care. We quantify parental investment, and its cost to the parent, by comparing the behaviour of parents with that of 'pairs' of adults without young (most of which are probably failed breeders). Each partner's strategy of investment is not expected to be purely selfish in this long-term pair-bonding species, and the sex differences in parental care are examined in this light. The time budgets of parents and pairs differed, parents walking more, grazing more and preening less. Parents also spent more time in the extreme head up posture and less in the head low and head on back postures than pairs but time devoted to the head up posture was the same for both. Brood size had no effect on the time budget. Time spent extreme head up declined over the study period in parents but not in pairs. Spacing patterns and behaviour varied independently in non-breeding birds but families sat closer to other geese when the vigilance level of the parents was low (i.e. in the head on back or head low postures) than when it was high (the head up or extreme head up postures). All agonistic encounters between parents and non-breeders were both initiated and won by the parents. Tied encounters occurred between birds of equivalent status in terms of brood size or non-breeding group size. The potential sources of parental care are summarized (Table 6) and, after considering the evidence for each, it is concluded that (apart from brooding) two types of parental investment are made by parents: (1) enhancement of offspring feeding efficiency by reducing competition through agonistic behaviour, and perhaps avoidance; and (2) protection from predators by (a) active defence, (b) seeking proximity with other geese when resting, and (c) visual scanning for predators (mainly by the male) using the extreme head up posture. Parents paid for this investment by devoting less time to preening and sleeping. The male's investment in predator vigilance was made at the cost of a reduced feeding time and to compensate for this he pecked at a faster rate than his mate. These sex differences are explicable in terms of earlier differences during incubation. The adoption of unrelated goslings was observed and the implications of the phenomenon are discussed. For individuals in non-breeding groups the time spent extreme head up declined as group size increased. The functional significance of this finding is discussed and it is concluded that in sitting groups extreme head up is probably used to scan for predators.