Johan Hjort's "critical period" hypothesis, which postulates that year-class strength is determined in the short period following the onset of exogenous feeding, has rarely been supported by empirical data. Instead, the current understanding is that recruitment is determined by cumulative mortality throughout early life. Recent studies relied on the measure of growth autocorrelation derived from otolith daily increment widths to test the link between growth rate achieved during the post-hatch period and during subsequent phases of the larval stage. Based on this approach, we revisit the role of larval growth in driving survival potential in three clupeoid species: Japanese sardine Sardinops melanostictus, Japanese anchovy Engraulis japonicus, and Pacific round herring Etrumeus micropus throughout the larval stage, using a combination of published datasets of otolith increment widths. Strong growth autocorrelation was detected for all three species throughout the larval stage, suggesting that initial growth determines to some extent growth rates achieved later in life. The extent of autocorrelation was reduced in sardine relative to anchovy and round herring at older ages. This interspecific difference could be attributed to differences in sensitivity to variability of environmental factors such as water temperature and food availability.The present findings suggest that the effect of early growth rate persists into later life stages for driving survival potential, which could reconcile the classic concept of "critical period" and the current "growth-survival" paradigm.anchovy, early life history, growth rate, larvae, otolith daily increment width, round herring, sardine
| INTRODUCTIONJohan Hjort's "critical period" hypothesis states that year-class strength of marine fishes is determined shortly after yolk absorption, when early larvae need to find sufficient numbers of suitable prey (Hjort, 1914). To test this hypothesis, numerous studies have examined the relationships linking annual recruitment to prey availability during the early larval stage or to larval mortality immediately after the onset of feeding. However, the "critical period" hypothesis has rarely been supported by empirical data (Houde, 2008;Leggett & Deblois, 1994). The current state of the art expanded from the "critical period" by proposing that variability in recruitment is determined by cumulative mortality sources experienced throughout the larval stage, rather than during the sole first-feeding period (Houde, 1987(Houde, , 2008Leggett & Deblois, 1994). Regarding survival to recruitment, the "growth-survival" paradigm posits that larger and/or faster growing individuals will have a higher probability of survival