EVALUATION OF BIOLOGICAL INDEXES FOR 3 EASTERN PACIFIC DOLPHIN SPECIES

To evaluate potential biological indices of population status (i.e., N relative to carrying capacity [K]), we tested for correlations betwee n estimates of several life history parameters and status for 3 dolphi n species in the eastern tropical Pacific between 1974 and 1992. On th e basis of a qualitative population response model, we predicted that the proportion of sexually mature females would be positively correlat ed with population status and that the proportion of simultaneously pr egnant and lactating females would be negatively correlated with statu s. Analyses were conducted by population, or management unit, for each species. The southern portion of western/southern pantropical spotted dolphin (Stenella, attenuata) and central short-beaked common dolphin (Delphinus delphis) populations declined in abundance from near K in 1974 and, therefore, were expected to exhibit compensatory responses i n life history parameters. Population status was correlated (P < 0.05) with greater-than-or-equal-to life history parameter, as predicted, f or both populations. Correlations (P < 0.05) for the northeastern pant ropical spotted dolphin also were found. No correlations were predicte d in this case, because although the population declined during the st udy, it was estimated to be 20% of K by 1974. We predicted no correlat ions for the eastern spinner dolphin (S. longirostris orientalis) and whitebelly spinner dolphin (S. 1. longirostris) because we detected no change in population during the study for either population. However, correlations (P < 0.05) between life history parameters and year were detected for both populations indicating they may have been respondin g to changes in population density. We conclude that regulatory mechan isms are likely to be operating between 20 and 100% of K and that life history parameters can be useful indices to distinguish between chang es in population status caused by exploitation or changes in K when re lationships between the parameters and status are known. Sample size r equirements for detecting changes in the parameters tested also are pr esented for time series of varying lengths.