MEPS 193:157-166 (2000)  -  doi:10.3354/meps193157

Osmotic condition, buoyancy change and mortality in larval cod Gadus morhua. A bioassay for assessing near-term mortality

Matthew Sclafani*, Gray Stirling, William C. Leggett**

McGill University, Department of Biology, 1205 Dr. Penfield Ave., Montreal, Quebec H3A 1B1, Canada
Present addresses: *New York State Department of Environmental Conservation, Division of Fish, Wildlife and Marine Resources, 205 North Belle Meade Rd., Suite 1, East Setauket, New York 11733, USA. E-mail:
**Queen's University, Department of Biology, Kingston, Ontario K7L 3N6, Canada

ABSTRACT: In larvae of marine fish, density differences are known to be associated with nutritional state and osmotic failure. We hypothesized that these density measures can be used to predict the near-term mortality of larvae. Specifically, we predicted increased larval mortality to be positively and continuously associated with increased larval density. Under controlled conditions, we measured (1) the average density (g cm-3) and daily percent mortality of fed and starved larval cod Gadus morhua L. in culture tanks and (2) the relationship between larval density and percent mortality of both fed and starved larvae whose survival was monitored for a 24 h period following density determination. We observed a strong, positive linear relationship between average larval density and average percent mortality in both experiments (r2 = 0.79 and r2 = 0.78 respectively). The regression slopes did not differ significantly between experiments. A linear regression on the pooled data yielded a highly significant relationship (r2 = 0.77, p < 0.0001). A second-order polynomial regression was also significant and improved the fit (R2 = 0.80, p < 0.0001). While the second-order polynomial regression better described the functional relationship between average density and average mortality, the simpler first-order model was judged preferable for predicting near-term mortality. The consistency and strength of the linear regressions suggests that larval density can be a reliable predictor of larval mortality. We determined the density of larvae associated with 50% mortality (potential 'critical' density value) using linear, polynomial and logistic regressions. All 3 produced similar 'critical' densities and we used the linear model to estimate this value. Larval cod having densities >1.0338 g cm-3 have a 50% probability of dying within 24 h. The potential for using larval density to predict the probability of near-term mortality in laboratory and field studies is discussed.


KEY WORDS: Larval fish · Buoyancy · Condition indices · Mortality · Bioassay · Cod


Full text in pdf format