The kill of dolphins (Stenella attenuata and S. longirostris) in the eastern tropical Pacific tuna purse-seine fishery has been underestimated because of unobserved deaths of nursing calves due to separation from their mothers during fishing. Based on an analysis of dolphins killed from 1973 to 1990, and depending on the length at which calves are assumed to become independent, there was a deficit of calves relative to the number of lactating females killed in 24%–32% of 1,847 spotted-dolphin sets and in 13%–19% of 563 spinner-dolphin sets. We found a deficit of 0.31–0.45 spotted dolphin and 0.15–0.26 spinner dolphin calves per set. If these missing calves were added to the observed kill, it would represent an increase in the kill of 10%–15% for spotted dolphins and 6%–10% for spinner dolphins in the sets we examined. We did not attempt to estimate the actual number of unobserved calf deaths due to purse-seine fishing on dolphins, either in the sets we examined or in all dolphin sets. The actual number of unobserved calf deaths is likely to be higher than the calf deficit we found. Separation of dolphin mothers from calves could occur at any of several points in the fishing process, but most of these would be invisible to us and not produce a calf deficit. Estimation of the actual number of unobserved calf deaths would require further information on how frequently permanent mother-calf separations occur, the fraction of calves that survive after separation, the fraction of lactating females set upon that are carrying calves, and the fraction of calves killed that are actually related to lactating females killed in the same set. In any case, the observation of a calf deficit indicates that the reported dolphin kill fails to measure the full impact of purse-seine fishing on spotted and spinner dolphin populations.