Median ef'ective doses and 95% effective doses of 9 commercial mosquito repellents were determined for the yellow fever mosquito, Aedes aegypti, in an in vitro bloodfeeding test system and on the human forearm. Results obtained, in the 2 test sytems did not differ significantly but, because of the inherent variability of repellent test data, did not always agree closely. Potential modifications of in vitro bloodfeedingtest systems for increased accuracy, precision, and reliability are discussed.