MEPS 317:237-244 (2006)  -  doi:10.3354/meps317237

ABSTRACT: A cohort of the fast-growing sprat Spratelloides gracilis was sampled during late-larval, juvenile and adult life-history phases using light traps on the North West Shelf of Western Australia. Otoliths from 154 larvae, juveniles and adults that hatched during a 20 d window were analysed to produce back-calculated daily records of size-at-age and growth rate. These traits were compared among sequential samples using repeated-measures MANOVAs (multivariate analyses of variance) to determine whether selective mortality occurred in the cohort. Late-stage larvae in our catches averaged 23 ± 3 d of age and were 22 ± 2 mm standard length (SL), while juveniles averaged 47 ± 6d of age and 36 ± 6 mm SL. We found that individuals that survived the larval stage to become juveniles underwent strong selective mortality. This selective mortality acted to preferentially remove fish that were slow-growing and/or relatively small members of the cohort. The size variation on which this selection acted was present at hatching and propagated by growth during the larval stage. Size at hatching is principally determined by egg size, implying that maternal contributions had an important influence on the outcome of selective events. We found no evidence of selective mortality operating during the transition of juvenile sprats to adulthood. Adults averaged 78 ± 6 d of age and 44 ± 5 mm SL. Log-linear analyses indicated that the cohort underwent 8.6% mortality mm^–1 SL and had a daily mortality rate of 3.7% between larval and adult stages. Given an average linear growth rate of 0.96 mm d^–1 during the larval phase, this suggests that selective mortality based on size (bigger-is-better) was approximately twice as important as mortality due to age differences (stage duration) among members of the cohort.

KEY WORDS: Larval fish · Otoliths · Size selection · Growth rate · Selective mortality · Predation · Bigger-is-better · Stage duration