Pellets of Cape Cormorants (Phalacrocorax capensis) are cast daily and can be used to determine prey species through identification of otoliths and other indigestible parts. Experiments in captivity indicate that some otoliths may be partially or entirely digested, making it impossible to estimate the original size or number of fish ingested per meal or per day. Birds of various species regurgitate casts or pellets of indigestible material (Knight 1964). Pellets of cormorants could be useful for study- ing diets of these birds because sampling is non-destructive and large samples can be ob- tained with little disturbance to the birds (Ain- ley et al. 198 1). Crystals of fish eyes taken from pellets might be counted to estimate daily food intake (Jordan 1959, Schlatter and Moreno 1976) and otoliths could be useful in identi- fying the species and sizes of fish consumed by the birds. However, certain aspects of pellet produc- tion require investigation before pellets can be used for either quantitative or qualitative stud- ies. How frequently are pellets cast? Does a pellet (accurately) reflect a meal or a bird's daily consumption of food? Do otoliths remain unchanged while residing in a bird's stomach, so that otoliths can be used to estimate the sizes of the fish eaten? How long do otoliths persist in a bird's stomach? We report here on studies designed to answer these questions, us- ing information obtained from pellets pro- duced by captive Cape Cormorants (Phalacro- corax capensis). METHODS Seven adult Cape Cormorants were taken from Dyer Island (344 1 'S, 1915'E), Cape Province, South Africa, on 19 November 198 1. The birds were transported overnight by sea to Cape Town and placed in cages. They were kept in the cages for six weeks until their weights sta- bilized and they became accustomed to cap- tivity. Birds were grouped in two cages of two individuals and one cage of three, because they refused to eat if kept separately. During the experimental period, the birds were fed com- mercially obtained horse mackerel (Trachurus cupensis; 140 mm) and cape anchovy (En- gruulis capensis; 120 mm). Birds were fed twice daily (10:00 and 16:OO) by throwing them fish, until satiated. Cages were carefully searched for pellets at 08:00, after which the birds were weighed. Recovered pellets were measured to the nearest millimeter for maximum length and their volumes determined by displacement in water. Each pellet was then cut open and any otoliths and eye crystals extracted and count- ed. Three experiments were carried out: Experiment one. Does the composition of pellets accurately reflect daily food intake? We counted the number of otoliths and eye crystals in pellets in relation to the number of fish con- sumed. The size and number of fish fed to each bird in each cage were noted. The number of pellets produced one day later was recorded for each cage. We combined the number of fish fed and number of pellets produced within each cage, since we did not know which bird pro- duced each pellet within a cage. If pellets per- fectly reflected fish intake, and if pellets are produced daily, the number of otoliths and eye crystals should have been twice the number of fish consumed the day before. Experiment two. Can otolith size be used to