Priority effects, i.e., effects of an early cohort on the performance of a later cohort, are generally studied between, and not within, species. The paucity of intraspecific assessments does not reflect a lack of ecological importance, but the technical problem associated with differentiating between conspecific cohorts. Here, we examine priority and density-dependent effects on larval Salamandra salamandra infraimmaculata. Larvae deposited by their mother early in the season have increased risk of desiccation, as rains at the beginning of the season are less frequent and unpredictable. However, breeding later may incur a high cost through conspecific priority effects, including cannibalism and competition. In an outdoor artificial pool experiment, we established densities of 0, 1, 2, 4 or 6 newly born larvae per pool ( approximately 30 l), and 40 days later, added a second cohort of three newly born larvae to each pool. We differentiated between cohorts using natural individual-specific markings. For the early cohort, increasing density decreased survival and size at metamorphosis, and increased time to metamorphosis. For the late cohort, survival was 100% in pools without early-cohort larvae, but ranged between 13 and 33% in the presence of early-cohort larvae. Time to metamorphosis was significantly longer in the presence of low vs high densities of early-cohort larvae. Results suggest that early-cohort larvae are mainly subjected to exploitative competition and cannibalism mediated by food limitation, and that late-cohort larvae are subjected to cannibalism and interference due to size asymmetry between cohorts. The strong priority effects suggest that Salamandra females could increase their fitness by adjusting the number of larvae they deposit in specific pools to avoid cannibalism and intraspecific competition.