We compared the use of fatty acid signature analysis with results from traditional faecal methods using milk samples and enemas concurrently collected from 8 female Antarctic fur seals Arctocephalus gazella foraging from Bird Island, South Georgia. The seals were serially sampled throughout the 1998/1999 breeding season, with collections taken after every foraging trip when possible. The total lipid content of the milk increased throughout the season. A general linear model using 4 principal components showed that it was the point within the breeding season at which a milk sample was collected that had the biggest influence on its fatty acid composition. There were also significant differences in the milk composition of individual seals. However, the results of faecal analysis, which are themselves subject to potential biases, were not correlated with the relative amounts of individual fatty acids. It has been suggested that seasonal changes in the milk fatty acids, previously observed in the same population of seals, are a result of increased consumption of fish. We found no evidence of this in the faecal material and suggest that the seasonal variability may be a result of changes in the fatty acid composition of the seals’ main prey (krill), or because of the changing needs of the developing pups. This study highlights the need for comprehensive testing of milk fatty acid signature analysis before the full potential of this technique can be realised.