The ability to determine the sex of Adélie penguins by discriminant analysis of morphometric parameters was examined for breeding birds of known sex at Bechervaise Island near Mawson Station, Antarctica (latitude 67°36’S, longitude 62°53’E). It was found that 89 % of birds could be correctly sexed by comparing the discriminant score D = 0.582 (bill length) + 1.12 (bill depth) + 0.219 (flipper width, cubitocarpal) with the mean discriminant score of 55.39. Males have a higher score and females a lower one. The sex of juvenile birds (identified by their white chins) cannot be determined by this method. The method is sensitive only where very accurate measurements are taken and it is recommended that the mean discriminant score be determined for a single operator and at each breeding colony to reduce these sources of variation. Further, it should be applied only to breeding birds. The probability of correctly identifying the sex of the birds can be increased by avoiding those with a discriminant score close to the mean. However, to be 100% certain of the sex of birds, 80% of the total measured sample would have to be discarded. In addition to the increased work load, this leads to serious bias in the resultant sexed samples. The usefulness of discriminant analysis for sex determination of Adélie penguins for applied monitoring studies is questioned.
WG-CEMP-91/5 Rev. 1