Critical periods and growth-dependent survival of larvae of an estuarine fish, the rainbow smelt Osmerus mordax

We investigated the Link between hatch date, growth trajectory, estuarine retention and survival of rainbow smelt larvae Osmerus mordax in the St. Lawrence Middle Estuary. Young smelt were collected over their entire distributional range across the salinity gradient of the Middle Estuary (0 to 25 psu) during 2 growing seasons. Individual growth trajectories and hatch-date distributions were reconstructed from otolith microstructure. High abundances of larvae were observed in the upstream region (0 to 10 psu) corresponding to the estuarine turbidity maximum (ETM). Larval rainbow smelt hatched in May during a period of 26 to 27 d, and mortality rates were variable among hatch dates. These variations were related to the feeding conditions experienced by first-feeding larvae (3 to 5 d after hatching) that fluctuated from day to day according to a predictable cycle of 15 d associated with photoperiod and tide. Comparison of growth trajectories of larvae at various ages indicated that older individuals, i.e. survivors, grew faster. Furthermore, smelt larvae collected within the ETM exhibited significantly higher growth rates than those sampled in the downstream area of the Middle Estuary, suggesting that larvae collected in the latter region represented dispersal from the ETM. We concluded that inter-annual variations in year-class strength might not be generally affected by early starvation, as larvae will experience recurrently good, medium and bad first-feeding conditions each year because of the length of the hatching period. However, any factor that reduces growth rate during the larval stage, including parasitism, has the potential to affect interannual variability in year-class strength.