Abstract Abundance of juvenile plaice was quantified with a non-selective drop trap during two consecutive years (1991–1992) from time of settlement in late spring until start of emigration in late summer in a shallow (0–1 m) bay in the Gullmar Fjord, Sweden. Individual lengths were recorded and otolith sub-samples were examined to determine age distributions (days after metamorphosis). Peak abundance reached 1.4 ind·m −2 in 1991 and 10.0 ind·m −2 in 1992. Settlement in 1991 occurred later than usual, possibly due to offshore winds during spring. Mean lengths at approximately one month after settlements were similar in the two years. Later in the season, however, lengths at age were lower in 1992, implying density-dependent growth. In addition, analyses of individual cohorts demonstrated reduced growth of late settlers in 1992. Instantaneous mortality rates (d −1) did not differ between the two years. Control of population number in 1992 must have been governed by density-independent factors ( e.g. larval supply) shadowing or decoupling a density-dependent regulation. Biomass of one major epibenthic predator Crangon crangon was equal in the two years, but that of another Carcinus maenas was higher in 1992 than in 1991. Thus, a functional response and/or other sources of mortality must be assumed to account for the higher numbers of plaice eaten in 1992.