Coral reef fishes, like many marine organisms, have a complex life history that consists of a planktonic larvae stage and a benthic juvenile or adult stage. We used the growth records in the otoliths of a common damselfish to investigate the extent to which processes in the plankton determined the outcome of events after benthic settlement. Sequential samples of the same cohort showed that individuals that survived intense selective mortality 1-3 months after settlement were those fish that were the larger members of the cohort at hatching and grew faster during planktonic life. Such links between life history phases are likely to occur in reef fishes whenever there is selection for a trait that is cumulative, such as size. They may not only operate between life history stages in the same individuals, but even between those of different generations via maternal effects on size at hatching.

Abstract. While growth rates of pelagic larvae have been argued to be one of the principal determinants of the recruitment success of temperate marine fishes, it is not known if this is the case in the tropics. Here, we use larval growth histories derived from otoliths of a Caribbean reef fish to show that monthly variation in the intensity of settlement and recruitment of pelagic juveniles onto reefs is positively correlated with variation in growth rates 1-2 weeks after larvae begin feeding. Our results suggest that the processes thought to underlie recruitment of marine fishes in temperate regions may also operate in the tropics and contrasts with current research on the causes of recruitment variability in coral reef fishes, which emphasises the role of larval transport.

The late-stage larvae of the reef fish Pomacentrus coelestis were collected using light traps at stations on a cross-shelf transect near Northwest Cape, Western Australia, during the spring and summer months (October to February) of 1997-98 and 1998-99. Physical (water temperature, wind) and biological (chlorophyll a, zooplankton abundance) variables were measured concurrently at each station. In 1997-98, environmental conditions were characterised by intrusive upwelling onto the shelf, relatively cooler water temperatures and higher chlorophyll a and zooplankton biomass. During the 1998-99 summer, water temperatures were warmer, and chlorophyll a and zooplankton biomass were relatively low. Catches of P. coelestis were much lower in the first summer (197 fish) than the second (1483 fish). Records of planktonic growth were obtained from otoliths of subsets of larvae from both summers. Growth varied among months within each summer; however, on average, larvae grew more slowly in the 1997-98 summer than the 1998-99 summer (0.48 mm d-1 vs 0.53 mm d-1 respectively), despite the presumed food (copepods) being more abundant in the plankton. Partial correlation analysis showed that water temperature explained approximately 30% of the variation in growth of larval P. coelestis. In contrast, chlorophyll a and zooplankton abundance explained much less (4.1 and 3.5%, respectively).

We developed a model of back-calculation of fish size from otoliths that could accommodate both changes in the relationship between otolith and somatic growth that occur through time and variation in growth rates among individuals. We used this model to back-calculate estimates of size and growth from otoliths of three species of Diplodus from the Mediterranean Sea. The outcomes of our model were compared with those of three other models and with growth curves estimated directly from populations of fish in the field. We found that our new model produced estimates of size-at-age that were closer to those observed in the field than the biological intercept, time-varying growth, and body proportional models. Comparison of profiles of increment width from otoliths of newly settled and juvenile Diplodus puntazzo and Diplodus vulgaris showed that these species formed a settlement mark, where increment width declined at settlement. In contrast, a settlement mark was not evident in the otoliths of Diplodus sargus. However, settlement of all species coincided with a sharp decline in somatic growth rate. Thus, growth rate may provide a means of estimating the timing of settlement in species that do not display a marked change in increment width.

We investigated the size-selective mortality of Atlantic salmon (Salmo salar) fry during two consecutive summers that differed markedly in weather conditions. We sampled fry shortly after emergence in June and at the end of August to compare the distributions of back-calculated body size at hatching by examining otolith microstructure. Size-selective mortality was observed in both summers; however, the direction and strength of mortality differed. During the drought conditions of 1995, selective mortality was relatively weak and directed towards the smaller fry in the population. During the flood conditions of 1996, selective mortality was relatively strong and directed towards the larger fry of the same population. Interannual variability in size-selective mortality contributed to significant differences in the mean size of fry at the end of their first summer of life. Size-selective mortality rates estimated from the shifts in fish length at hatching observed during the first summer of life were comparable with published estimates of total mortality of Atlantic salmon fry, indicating that early mortality may be largely size selective. Mortality associated with hydroclimatic events can select against either small or large fish and is a key determinant of mean size attained by the end of the first summer of life.