Landing calls of known individual male gannets were recorded over a period of a month. The calls were subsequently analyzed using a sonograph with an amplitude display attachment. Changes in the amplitude envelope in time appeared to be the only measure which readily exhibited consistent individual features. Each sequence of landing calls which was given on a single occasion was reduced by averaging to a single 'profile' or pattern of numbers. Three such profiles were obtained for each of 10 males. One of each male's profiles was set into a computer as one of 10 reference sources; the remaining profiles were tested against them. The matching procedure consisted in cross-correlating the sample profiles with the reference patterns and identifying the sample with the reference which gave the highest correlation. The identification was usually correct. The same procedure was repeated using progressively less information from the sounds to see how much of a sound was necessary for correct identification. It was found that sometimes as little as one quarter of the sound profile was sufficient. The limitations of the technique and the implications of the results are discussed.