analyzed are defined in Table 1. Analyses were 

 confined to feeding-stage larvae of marine fishes 

 and to a few anadromous species. Salmonid 

 fishes, which lack a typical larval stage, were not 

 included. 



FISHERY BULLETIN: VOL. 87, NO. 3, 1989 



rate (G) and dependent upon species-specific 

 weight at metamorphosis. Stage duration was 

 calculated from the literature-derived data on G, 

 dry weights at hatch (Wq), and dry weights at 

 metamorphosis {W,„) 



Table 1. — Relationships that were analyzed and abbrevia- 

 tions that are used in text. 



D 



(In W,„ - In Wo) 



T 

 G 

 D 



Z 



I 



A/mBt 



Temperature, °C 



Weight-specific growth rate, d"' 



Stage duration, defined as days from hatching 



to metamorphosis, d 

 Instantaneious daily mortality coefficient, d"' 

 Gross growth efficiency, G/l 

 Weight-specific ingestion rate, d^^ 

 Weight-specific oxygen uptake, 



jiL 02/mg/h 

 Net survivorship; fraction of a cohort expected 



to survive from hatching to metamorphosis 



Growth 



Weight-specific growth rates (G) of larvae 

 were obtained from laboratory and field studies. 

 Growth coefficients were taken directly from 

 published work, when available, or calculated 

 from the data. In a few cases, length-weight 

 relationships were used to convert gi'owth-in- 

 length data to gi'owth-in-weight. Weight-specific 

 growth data for larvae of some commonly 

 studied species were unavailable and those 

 species could not be included in analyses. 

 Weight-specific growth rates are 



G 



(In W, - In Wq) 

 t 



where V^o and Wi are dry weights of larvae at 

 hatch and at the end of a growth period of t days' 

 duration. 



A linear regression was fitted to express the 

 relationship of weight-specific gi'owth rates on 

 temperature. Because both G and T for a species 

 sometimes had a considerable range reported in 

 the literature, the midpoints of reported G and T 

 values were selected arbitrarily as the data for 

 each species in the regression analysis. A Qio 

 was estimated from predicted values of G in the 

 range 5°-30°C. 



Stage Duration 



Stage duration (D) was defined as days from 

 hatching to metamorphosis. As such, it is in- 

 versely proportional to weight-specific growth 



A power regression was fitted, in which the 

 geometric midpoint, i.e., antilog [(logu, low D + 

 logio high D) * V2] of the estimated range of Z) for 

 each species was regressed on the estimated 

 midpoint of temperature, to describe the rela- 

 tionship between stage duration and tempera- 

 ture. 



Mortality 



Instantaneous daily mortality coefficients (Z) 

 of post-yolk-sac larvae were obtained from pub- 

 lished field studies in which the rates had been 

 determined or could be calculated from the 

 authors' estimated abundances-at-age. Much of 

 the mortality data was obtained from summaries 

 and references in papers by Dahlberg (1979) and 

 McGurk (1986), supplemented with Z estimates 

 from additional and more recent sources. 



A linear regression of the midpoints of the 

 estimated range of Z on the midpoints of T for 

 each species was fitted to describe the relation- 

 ship between Zand T. 



Relationship Between Mortality and 

 Growth Rates 



Solution of the equation for growth rate on 

 temperature in terms of temperature and its 

 substitution into the equation for mortality rates 

 on temperature yielded an expression relating 

 mortality rates to growth rates. 



Net Survivorship 



Based on the estimates of growth rates, stage 

 duration, and mortality rates predicted from the 

 regressions, the proportion of a cohort expected 

 to survive to metamorphosis (A'^met) was calcu- 

 lated for 10°, 20°, and 30°C. The percentage 

 cohort survivorship at each of the temperatures 

 was calculated from the exponential relationship 



100 A^met = e"^^ . 

 Effects on survivorship of decreases in G and 



472 



