Fecundity of orange roughy in 1987-1989 adjusted for standard length (s.L.) varied significantly between New South Wales (42 787 eggs female-') South Australia (35 339 eggs female ~ I ) and east Tasmania (31 085 eggs female-'). Only 1&17% of the variability in fecundity of eastern Tasmania orange roughy was explained by S.L. in any year from 1987-1992. However, liver condition and age of the fish, in combination with s.L., explained 27% of the variation in fecundity. Fecundity declined in fish over 60 years old. It was also significantly correlated with lipid levels in the ovary, in particular, with triacylglycerol as a proportion of the total lipid fraction. Significant interannual changes in fecundity appeared to be related to the impact of fishing. From 1987-1992, the orange roughy stock off east Tasmania was reduced by 50% by the fishery, and mean fecundity increased 20% over that period. This compensatory increase in individual fecundity, combined with an apparent increase in the proportion of females spawning annually from 54 to 71%, limited the decline in the population's egg production over this period to approximately 15%. Q 1995 The Fisheries Society of the British Isles

This study was designed to determine whether orange roughy reproduce throughout their distribution in Australian waters, and if so, whether commencement of spawning and length at maturity vary within the region. We found that females from four widely separated sampling areas reproduced in 1988, but that the onset of spawning and length at maturity were not the same in all the areas. In New South Wales, ovulation was finished by mid-June, and 50% of females were mature at 28 cm standard length (s.L.). On the other hand, females from eastern and western Tasmania and South Australia did not spawn until mid-July, and 50% of females from Tasmania were not mature until 32 cm S.L. Females from all areas had oocytes with yolk granules at least 5 months before ovulation. However, in the few months before spawning, some adult females from all areas had not undergone vitellogenesis or were resorbing all yolky oocytes. The best estimates of the proportion of these non-reproductive females were made in March and April, when the single batch of eggs to be spawned was clearly distinguishable, and before spawning aggregations had formed. In 1990 in eastern Tasmania, at the site of a major spawning aggregation, the best estimate of the proportion of non-reproductive fish was 45%. We suggest that scarcity of food, coupled with the cost of joining a spawning aggregation, may result in intermittent spawning in this long-lived species.

Population dynamics of the symbiotic crab Djssodactylus mellitae Rathbun were studied by monthly collections from their sand dollar host Mellita quinquiesperforata (Leske) in North Inlet, South Carolina, USA. Size frequency analysis of carapace widths indicated a 12 to 15 mo life cycle for D. mellitae. Larvae were released during summer, grew to produce one or more clutches of eggs the next summer, and died shortly thereafter. Sex ratio was 1:l throughout the year; females attained a slightly larger maximum size. Fecundity did not vary over the reproductive season (April through September) and mean egg number was 206 ? 62 (S. d.) eggs per clutch. Most recruitment occurred from May through September. D, mellitae differs from other species in the family Plnnotheridae in that males and females are more similar in size and longevity, there are no soft carapace stages, clutch size is relatively small and the entire clutch is completely covered by the abdomen. Accessibility of sand dollars to settling D. mellitae and protection provided to post-metamorphic individuals may decrease mortality enough to balance the small clutches produced by this crab.