Fishery Bulletin 93(2), 1995 



ity between otolith growth and fish growth (Fraser- 

 Lee, simple regression), will accurately estimate fish 

 length if somatic growth is relatively fast (Secor and 

 Dean, 1989, 1992; Campana, 1990). If somatic growth 

 is slow, otoliths of slow-growing fish may grow rela- 

 tively fast, causing the relationship between fish and 

 otolith growth to be nonlinear and thus bias back- 

 calculated growth-rates (Geffen, 1982; Reznick et al., 

 1989; Secor and Dean, 1989, 1992; Campana, 1990). 

 To minimize biased estimates of back-calculated 

 lengths at age for 1988 Potomac River larvae, when 

 growth effects were significant and body size and 

 otolith-size relationships were nonlinear, we fol- 

 lowed the recommendation of Campana and Jones 

 (1992) and log^-transformed fish lengths and 

 otolith radii before applying the biological inter- 

 cept method. 



Larval cohort mortality rates were estimated 

 from the exponential declines in cohort abun- 

 dances at age obtained from 60-cm sampler 

 catches. In 1989, larval mortality rates in the 

 Potomac River and Upper Bay may have been 

 overestimated when only larvae collected in the 

 60-cm sampler were considered. Striped bass 

 spawned later in 1989 than in 1987 or 1988, and 

 abundances of cohorts hatched in late May prob- 

 ably were under-represented in 60-cm sampler 

 collections that were completed by 1 June. Better 

 estimates of cohort mortality rates in 1989 were 

 obtained by combining larval collections from the 

 60-cm sampler with those from the 2-m 2 Tucker 

 trawl made from late May to mid-June. 



A "Pareto" model, which assumes that mortal- 

 ity rate declines in relation to age (Lo, 1986), was 

 compared with the linear model, and the model 

 with the lowest residual sum of squares was se- 

 lected to estimate each cohort's mortality rate. The 

 ratio of instantaneous growth-in-weight rate to in- 

 stantaneous mortality rate (GIZ), an index of lar- 

 val production (biomass), was estimated for each 

 cohort. Instantaneous growth-in-weight rates were 

 derived from back-calculated growth-in-length 

 rates by using a length-weight relationship ( Houde 

 and Lubbers, 1986) for striped bass larvae. Abun- 

 dances of cohorts at 8 mm standard length (SL) 

 were estimated to index their potential contribu- 

 tions to recruitment. At 8-10 mm, striped bass 

 larvae reach the postfinfold larval stage (Fritsche 

 and Johnson, 1980; Olney et al., 1983), when de- 

 velopment of fins and increased swimming ability 

 facilitates feeding as well as ability to avoid preda- 

 tors and can result in a decline in mortality rate. 

 Abundances, and growth and mortality rates of 

 larval cohorts were analyzed in relation to envi- 

 ronmental variables by using stepwise multiple 



regression analysis (SAS, 1988) to identify factors 

 associated with variable recruitments. 



Results 



Spawning, egg and larval abundances, and 

 zooplankton densities 



Striped bass spawned during April and May, gener- 

 ally during periods of rising temperatures (Figs. 2 

 and 3). The major spawning peak in the Potomac 



2000 



1500- 



1000- 



c 

 o 



B 



0) 



o 



c 

 ro 

 T3 

 C 

 3 

 JO 

 (0 



O) 

 O) 



UJ 



<D 



3 

 ■o 



B) 



C 



O 



Figure 2 



Riverwide egg abundances (millions) of striped bass, Morone 

 saxatilis, estimated on each survey date (bars) during the three 

 years of sampling effort in the Potomac River, 1987-89. Wa- 

 ter temperatures recorded at Wilson Bridge during the spawn- 

 ing season also are given (open diamonds). The 12°C critical 

 low temperature, at which 100<£ egg and larval mortality may 

 occur, is indicated by the dotted line. Note that the Y-axis scales 

 change among years. 



