Larval and early post-larval growth has been investigated in Hyas araneus L. (Majidae) reared in the laboratory. Growth was measured as dry weight (DW), ash-free dry weight (AFDW), carbon (C), nitrogen (N), hydrogen (H), gross biochemical constituents (protein, lipid, carbohydrate, chitin, ash) and energy (calculated separately from carbon and biochemical composition). During larval development, i. e. from freshly hatched zoea-1 to late megalopa, all these cr~teria of b~omass increase by factors ranging between 5 and 14; carbohydrate shows the lowest, chitin the highest Increment. There are indications of loss in organic body weight during the latest period preceding metamorphosis to the crab stage. When no food is offered during this time, megalopae lose significantly more biomass than control larvae. This suggests that food is still required, but feeding activity is reduced to a level below maintenance ingestion rate. Follow~ng metamorphosis, the juvenile crab accumulates biomass at a far higher absolute rate (expressed as pg d-') than all larval stages. This alteration in growth pattern is attributed to lack of further morphological changes. Among the biochemical constituents, protein is most prominent (50 to 68 % AFDW), followed by lipid (I? to 30 %) and chitin (7 to 17 76). Carbohydrates play a minor role contributing only 1 6 to 3.7 % of AFDW. The percentage of protein increases during zoeal development; later it decreases. This pattern is inversely related to changes in ash content, which ranges from 19 to 30 % DW. Lipid (% AFDW) shows a decreasing tendency during larval development. Separate calculations of energy content from C and from biochemical composition yield systematic differences, the former usually being lower than the latter. Only in late megalopae is this pattern Inverted, presumably due to inorganic C in the increasingly calcified cuticle of this stage. Non-linear regression equations are given for conversions of C to protein, lipid, and carbohydrate, N to protein, and H to lipid. The relationship between N and protein suggests that non-protein, non-chitin N may also be important, and calculation of protein by multiplication of total N X 6.25 results in an overestimation of protein.