A previously published fluorometric method for vitamin A in milk, involving saponification in centrifuge tubes and extraction with hexane, was automated using conventional automated equipment including a filter fluorometer. The extraction and separation steps were affected by the presence of fat, so standards were prepared in milk. The peaks obtained with samples were 95% of steady state values. The coefficients of variation for 10 replicate analyses were 1.9% for unfortified milk and 1.3% for fortified milk. In tests on 19 different milk samples, the results of the automated and manual fluorometric methods differed by less than 10%; a colorimetric method using SbCl3 gave more erratic results and was more lengthy and laborious.