Twenty male and 20 female parental oysters, originating from four sites located along the French Atlantic coast, were crossed together. The 400 crosses were performed separately and then pooled to give a batch of larvae with a large genetic base. Successive sieving after 17 days at 23°C enabled the separation of the largest larvae from the batch. These larvae (i.e., sieving groups) were left to metamorphose and fix onto flat PVC collectors changed daily. Four groups of larvae representing distinctly different growth rates were successively separated during the attachment and metamorphosis of the whole population. This lasted 12 days, from the 17th day to the 29th after fertilisation. A part of each sieving group was settled separately as replicates on cultch. The post-metamorphic height was recorded weekly on 100 oysters per sieving group generating 400 growth curves. The oysters were removed from the collectors and weighed. The effect of the date of settlement (i.e., developmental rate) and sieving group (i.e., larval growth rate) affected (P<0.0001) the spat growth rate significantly. The correlation between sieving groups (i.e., larval growth) and spat growth rate was positive. After 11 months of growth under intensive conditions, the sieving group still had a significant effect on the total weight of juveniles settled on cultch (P<0.0001). These results justify the size-grading of larvae in bivalve hatcheries and show the great importance of early growth on growth in later stages in Crassostrea gigas.