FISHERY BULLETIN: VOL. 86, NO. 3 



Table 6.— Stage-specific total instantaneous mortality rates of egg, larval, and juvenile 

 American sfiad from 1979 to 1987. SE = stage-specific standard errors; cv = coeffi- 

 cient of variation, Days = duration of each stage. 



Table 7.— Correlations between stage-specific total mortalities and 

 adult recruitment and several abiotic factors from 1979 to 1987 for 

 American shad. P = probability levels. 



were highly variable, but showed no relationship to 

 adult recruitment. 



Given that high (94-99%) egg and early larval 

 mortality rates were inversely related to year-class 

 strength, the abundance of first-feeding larvae (age 

 10 days) was closely correlated (r = 0.84, P < 0.01) 

 with adult recruitment (Rpt) from 1979 through 

 1987 (Table 8). Because year-class strength from 

 1967 through 1987 was independent (r = - 0.15, P 

 < 0.46) of egg production (EggS(), our results sug- 

 gest that year-class strength of American shad is 

 determined by the number of shad that survive the 

 embryonic and early larval stages. 



Early larval mortality rates from 1979 to 1987 

 were positively correlated with mean June river 



flows. No such relationships were evident for any 

 other stage-specific mortalities (Table 7). The high 

 egg and early larval mortality rates (Table 6) and 

 relative failure of the 1982 and 1984 year classes 

 coincided vdth major storm events in mid- June 1982 

 and late May 1984 (Crecco and Savoy 1984) which 

 increased river flows and kept water temperatures 

 below 17°C until late June. The high river flows in 

 1982 and 1984 were also coupled with the highest 

 parent stocks on record since 1967 (Table 1). By con- 

 trast, the relatively low egg and early larval mor- 

 tality rates for the dominant 1980 and 1987 year 

 classes were associated with low June river flows, 

 a steady rise in June water temperatures (Crecco 

 and Savoy 1984) and low to moderate size parent 

 stocks (Table 1). These results suggest that domi- 

 nant year classes of American shad are most likely 

 to occur when lower than normal June flows are 

 coupled with relatively small spawning stocks. 



Prejuvenile Density-Dependent 

 Mortality: A Closer Look 



Analysis 



Having estimated that 82% of the density- 

 dependent mortality for American shad takes place 

 before the juvenile stage (Table 3), we attempted 

 to estimate the relative contribution of density- 

 dependent mortality during the egg, early, mid-, and 

 late larval and juvenile stages from 1979 through 

 1987. We related egg mortality rates (ZEi) from 



476 



