MORSE: CATCHABILITY, GROWTH. AND MORTALITY OF LARVAL FISHES 



Mortality and Growth 



Survey timing has a profound effect on the 

 mortality estimates from either length- or age- 

 frequency curves. This is most evident if sam- 

 ples are not taken systematically during the 

 spawning season and surveys are not timed and 

 spaced evenly along the spawning curve (Hewitt 

 and Methot 1982; Morse and Hauser 1985; 

 Hauser et al. 1988). Mortality is overestimated 

 from samples taken when spawning or larval 

 production is increasing and underestimated 

 when spawning is decreasing. The magnitude of 

 the bias in mortality estimation shows a twofold 

 to threefold decrease with samples from the be- 

 ginning to the end of spawning. However, if sur- 

 vey samples are summed over the larval produc- 

 tion cycle and at least four surveys are spaced 

 evenly throughout the spawning cycle, the over- 

 and underestimates cancel out and the combined 

 data give a good estimate of mortality (Hewitt 

 and Methot 1982). The process of combining 

 surveys across years increases the number of 

 samples taken during the spawning cycle from 

 approximately 2 to 5 for each year to an observa- 

 tion every 4 to 10 days during the combined 

 production cycle for each taxa. The result of this 

 process is the calculation of an average length- 

 dependant larval mortality for the eight years 

 covered by this study. 



The relationship of larval mortality and water 

 temperature has some interesting implications 

 about larval growth rates. Mortality, as an 

 expression of the decrease in numbers over time 

 by substituting time (t,) for length (A',) in Equa- 

 tion 2, is directly related to larval growth rate by 

 the term t,. The assumption is that, as fish larvae 

 grow, mortality rate decreases (Ware 1975). 

 With constant predation rates, the amount of 

 time spent at a given size, commonly referred to 

 as "stage duration", will determine the number 

 of surviving larvae. The imphcation of this rela- 

 tionship of growth rate and mortality is that, 

 owing as water temperature increases, stage 

 duration will decrease to increased growth rate 

 and the shorter stage duration will decrease 

 mortality. This assumes that adequate food sup- 

 phes are available for the increased metabohc 

 demands of increased growth rates. The hnk be- 

 tween "stage duration" and particle-size depen- 

 dent mortality rates would appear to be valid for 

 most pelagic fish larvae given the rather small 

 size range of newly hatched fishes. However, if 

 larval mortality rates were, in fact, dependent 

 upon growth rates as outlined above, fishes 



spawning in warm waters would derive a signifi- 

 cant survival advantage over cold water 

 spawners and mortality would be inversely cor- 

 related with gi'owth rates. 



According to population dynamics theory, 

 mortality and growth rates must be positively 

 correlated with the ratio of the instantaneous 

 growth rate to instantaneous mortality rate, 

 averaging > 1 for the biomass of a cohort of fish 

 to increase (Beverton and Holt 1957; Ricker 

 1975; Ware 1975). If this were not the case, the 

 maximum biomass of a cohort would occur at the 

 egg stage. The results presented here (Fig. 5) 

 show that length-dependent larval mortality is 

 positively correlated with mean surface water 

 temperature, and it seems clear that estimated 

 mortahties are higher in the warm-water months 

 than during winter. The association of tempera- 

 ture and larval growth rate was determined 

 from a review of laboratory studies of larval 

 growth rates (Table 6). The relationship of in- 

 creasing growth rate with increasing tempera- 

 ture is not surprising, but the high coefficient of 

 determination (93%) is surprising, given the 

 variety of experimental procedures, prey 

 species, densities, and fish species utilized by the 

 experimenters. The data in Table 6 and the rela- 

 tionship of temperature to mortality confirm the 

 positive correlation of growth and mortality. 



Because expected ratios of instantaneous 

 growth and mortality rates have been shown to 

 be temperature-dependent, either rate is easily 

 determined if the other is known. This ability to 

 calculate either rate would have direct applica- 

 tions in modeling larval fish populations as it 

 relates to survivorship, predator-prey dynamics, 

 and the expected effect of environmental tem- 

 peratures. For this study, if growth rate is 

 known, then the mortality rates from field sam- 

 ples could be investigated to determine if net 

 avoidance is a serious bias as often speculated. 

 Growth rates in weight in Table 6 are rates per 

 day, but mortalities in Figure 5 are rates per mm 

 length interval. To convert lengths to ages {t = 

 days), the length (mm) to weight (jxg), relation- 

 ship and instantaneous gi'owth rate are dimen- 

 sioned in days and |xg where 



Weight = c * Length'' 



and the instantaneous growth rate (G„) 



G,, = {lnWui-lnWM.^i-t,). 



Thus 



441 



