Most observations of stressor effects on marine crustaceans are made on individuals or even-aged cohorts. Results of these studies are difficult to translate into ecological predictions, either because life cycle models are incomplete, or because stressor effects on mixed age populations may differ from those observed in cohort studies. This problem is evident in several important environmental applications of population ecology, including ecological risk assessment of chemicals and futures analyses of ocean acidification. In particular, investigators have acknowledged the need for life cycle approaches in predicting adaptive responses to changing environments. To address this need, we developed an observational scheme using mixed age Americamysis bahia populations that allows inverse estimation of stage-specific vital rates. We used this system to examine compensatory demographic responses to resource limitation, which is an oft-cited complication in applied population ecology. We randomly assigned one of four feeding levels to each of 24 laboratory populations. Weekly length measurements from digital images were then compiled into a 10-wk time series of stage abundances for each population. For comparison, this study was preceded by a cohort-based study of resource limitation.

Results/Conclusions

Using inverse demographic analysis of these stage-structured time series, the most strongly supported models of resource limitation effects on mysid demography included opposing (i.e., compensatory) linear effects on two parameters (e.g., adult survival and juvenile maturation). The model with adult survival response only (i.e., no compensation) was also strongly supported. Moreover, the feeding effect on fecundity was negative in the compensatory models, partially offsetting positive logit-linear effects on adult survival. This contrasts with cohort-based results, where feeding effects on fecundity were strongly positive. Also, life table response analysis showed that resource effects on overall population fitness were dominated by effects on adult survival. These results suggest that emphasis in stressor-response studies on early life stages and even-aged cohorts may miss important demographic responses and should be augmented by observations of intact populations, especially as methods such as ours become more available.