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



higher oxygen demand than larvae in colder 

 seas. If gross growth efficiency is constant in 

 relation to temperature, then assimilation effi- 

 ciency should decline as temperature increases, 

 while net growth efficiency should increase. 

 Based on the calculated energy budgets at 10°C, 

 20°, and 30°C (Table 8), predicted assimilation 

 efficiency dechned from 77.1% at 10°C to 59.8% 

 at 30°C. Average assimilation efficiencies of 

 larval fish generally are thought to be similar to 

 or less than the 80% m.ean assimilation efficiency 

 of juveniles (Brett and Groves 1979; Govoni et al. 

 1986). Net growth efficiency increased from 

 37.2% at 10°C to 48.1% at 36°C. The range of 

 predicted values of larval net gross efficiency lies 

 just above the 36% mean value reported for 

 juvenile fishes (Brett and Groves 1979). Re- 

 cently, Wieser et al. (1988) estimated the net 

 growth efficiency for larvae of the freshwater 

 Ridilus rutilus and indicated that it was high 

 and independent of temperature in the 15°-20°C 

 range. Inspection of their table 4 data indicates 

 that the net growth efficiency did in fact increase 

 by approximately 4% in the 5°C range of their 

 experiments, a percentage similar to that pre- 

 dicted from the energy budget analysis for 

 marine fish larvae. 



The mean assimilation efficiency for marine 

 fish larvae decHned as temperature increased. 

 There also is good evidence that both assimila- 

 tion and gi'oss growth efficiency of an individual 

 species decline at high ingestion rates (Houde 

 and Schekter 1981; Checkley 1984; Boehlert and 

 Yoklavich 1984; Kiorboe et al. 1987; Theilacker 

 1987). The abihty of larvae to capture prey, the 

 feeding conditions, and environmental factors all 

 affect assimilation and growth, as well as their 

 variability. Nevertheless, the efficiencies pre- 

 dicted here and their relationships to tempera- 

 ture still are believed to describe important lati- 

 tudinal effects. 



Despite expected high mortality rates, aver- 

 age survivorship of larval cohorts at tropical 

 temperatures was predicted to exceed that of 

 cohorts at high latitude temperatures. This re- 

 sult is a consequence of the high gi'owth rates 

 and the relatively short stage durations that 

 tropical larvae experience. While net survivor- 

 ship to metamorphosis in tropical systems may 

 be relatively high, the daily probability of death 

 also is high. Unless larval food abundance is 

 higher or tropical larvae are better able than 

 high latitude larvae to feed on scarce prey, 

 neither of which has been demonstrated, starva- 

 tion or other "critical period" mortalities (sensu 



Hjort 1914) may be more probable in the tropics. 

 Cohort survivorship is sensitive to small changes 

 in mortality and growth rates in either tropical 

 or high latitude systems. But, larval cohorts 

 developing in high latitudes are likely to suffer 

 proportionally gi-eater declines from small de- 

 creases in growth rates or increases in mortality 

 rates, a consequence of their extended larval 

 stage duration. If even weak or sporadic den- 

 sity-dependent growth or mortality operates in 

 the larval stage (Rothschild 1986), its effect 

 could be substantial in high latitudes where the 

 larval stage is long. 



Based on this analysis, starvation of first-feed- 

 ing larvae is hypothesized to be more likely in 

 warm seas because of their relatively great 

 ingestion requirement combined with low assim- 

 ilation efficiency. If it were possible for fish lar- 

 vae to live on a maintenance diet, they would 

 face less risk of food-limitation. But, in labora- 

 tory experiments it has been observed that 

 slow-gi'owing larvae are less likely to survive 

 (Laurence 1977; Houde and Schekter 1980, 

 1981). Estimates here indicate that approxi- 

 mately 15 food particles of 0.25 ixg dry weight 

 are required to produce 1 |xg dry weight of larval 

 gi'owth. Tropical fish larvae that are gi'owing 

 2-3 times as fast as larvae from cold seas, and 

 which also have an elevated metabohsm, must 

 consume nearly three times as much prey to 

 achieve average growth at ambient tempera- 

 tures. 



Average relationships reported here indicate 

 that larval mortality rates exceed weight-spe- 

 cific growth rates. Morse (1989) also has ex- 

 amined the relationship among growth, mortal- 

 ity, and temperature for 26 species of North 

 Atlantic fish larvae. He found that both gi'owth 

 rates and mortality rates increased with tem- 

 perature. He concluded that the ratio of mortal- 

 ity rates to growth rates is less than 1.0 for most 

 of those species, and suggested that, when ratios 

 exceeded 1.0, gear avoidance is the possible 

 cause. In my analysis I have accepted the possi- 

 bility that mortality rates may exceed growth 

 rates for teleost larvae, implying that there is a 

 loss of biomass between the egg stage and meta- 

 morphosis in most species. The ratio of mortality 

 rates to gi-owth rates, based on the regression 

 coefficients in the mortality rates on tempera- 

 ture and growth rates on temperature regi-es- 

 sions is 1.31. Morse (1989) concluded that if 

 larval growth rate is known, then mortality 

 rates can be predicted because the two rates are 

 correlated. This conclusion supports the obser- 



488 



