Two methods are presented for estimating instantaneous daily mortality rates of larval fishes with otolith-increment data. The first involves construction of a histogram of the numbers of daily increments per larva from a sample of larvae, and estimation of mortality from the decline in the loge-transformed increment-frequency curve. The second approach is to convert larva lengths to number of increments with a growth curve derived from otolith data, and then to estimate mortality from the descending limb of a loge-transformed increment-frequency histogram. Counts of daily increments in otoliths of larval anadromous alewives Alosa pseudoharengus collected from a coastal Massachusetts pond were used to compare results from these two methods with results from a third method that uses the decline in catch per unit effort of larvae through time. The estimate of instantaneous daily mortality for larval alewives from otolith methods was 0.122, which was an average rate for larvae 6.0 to 13.9 mm total length over a 1-month sampling period. The three methods generally produced similar mortality estimates over large sampling periods; however, otolith-derived mortality estimates from single sampling dates were highly variable. Major advantages of otolith increment methods over traditional catch-curve methods for estimating larva mortality are elimination of the need for fitting a growth curve to estimate mortality and the ability to obtain better estimates of mortality for individual cohorts of larvae.