SAVOY AND CRECCO: MORTALITY OF AMERICAN SHAD 



Table 4— Predicted total adult recruitment of 

 American shad from the environmental dependent 

 stock-recruitment model (Equation (8)), 95% con- 

 fidence limits about the recruitment values and 

 juvenile abundance for the 1983-87 year classes. 



Table 5.— Estimates of the total (ZTotal,) and 

 postjuvenile (Z4,) instantaneous mortality rates 

 and their standard errors (SE) for American 

 shad from 1979 through 1987. 



Given that ZE represents total mortality through- 

 out the egg and prolarval period, daily egg mortal- 

 ity rates were determined by dividing ZE by the 

 average duration (Ds) of the egg and prolarval 

 stages of American shad in the Connecticut River. 

 Watson (1968) reported an inverse relationship be- 

 tv^^een the incubation period (D ) of shad eggs and 

 water temperature (T) by the expression: 



D = 120.95 exp(-0.154 • T). 



(10) 



To determine the duration {Ds) of the egg stage 

 from 1979 through 1987 we substituted the mean 

 June temperatures for the Connecticut River 

 (U.S.G.S. Annual Water Year Reports 1979-87) into 

 Equation (10). 



The total number of prolarvae (age 2 days) and 

 early larvae (age 10 days) for the 1979 through 1987 

 year classes was estimated by 



NE^ = Eggs, • exp(-Z£;,) (11) 



and 



NFt = Eggs, • exipi-ZEt - ZELi) (12) 



respectively, where NEt is the estimated number 

 of prolarvae larvae and NFt is the number of early 

 larvae. The abundance of older larvae and juveniles 

 from each year class was estimated by adding their 

 respective total mortality rates to Equation (12). To 

 determine the life stage(s) at which year-class 

 strength is established, we related the stage-specific 

 total mortality rates and abundance estimates to the 

 number of adult recruits (i?, or Rpt) from the 

 1979-87 year classes in several linear models. If 

 year-class strength is established early, there should 

 be a significant positive correlation between the 

 number of prolarvae (NE) and early larvae (NF) and 

 adult recruitment, and total mortality rates during 

 these early stages {ZE and ZEL ) should be inverse- 

 ly related to adult recruitment. 



To determine the extent of density-independent 

 mortality at the egg, larval, and juvenile stages, we 

 correlated the stage-specific total mortality rates 

 from 1979 through 1987 to mean May and June river 

 flows (m^/s) and water temperatures (°C) in sev- 

 eral linear models. May and June hydrographic 

 and meteorological parameters were used because 

 they coincide with egg, larval, and juvenile develop- 

 ment in the Connecticut River (Leggett 1977) and 

 were the only monthly abiotic variables that were 

 significantly linked to adult shad recruitment from 

 1966 to 1980 (Crecco and Savoy 1984, 1987c). May 

 and June water flows and temperatures were re- 

 corded by the U.S. Geological survey (U.S. Geo- 

 logical Survey 1966-1980) within the major spawn- 

 ing areas (river km 89) of American shad (Leggett 

 1977). 



Results 



The mean daily mortality rates from 1979 through 

 1987 declined by an order of magnitude from the 

 egg through the juvenile stages (Table 6). Total egg 

 mortality rates were relatively high (mean ZE = 

 2.584, cv = 18.5%) and were inversely correlated 

 (r = - 0.76, P < 0.03) with adult recruitment (Rt or 

 Rpt) from those year classes (Table 7). These data 

 indicate that shad eggs and prolarvae (age 1-2 days) 

 from 1979 through 1987 experienced high (85-96%) 

 mortality that was directly linked to year-class 

 success. 



The total mortality rates (ZEL) among early lar- 

 vae were slightly lower and more variable (mean 

 ZEL = 1.608, cv = 31.5%) than the egg mortality 

 rates (Table 6), and were also inversely related (r 

 = - 0.83, P < 0.01) to adult recruitment from 1979 

 through 1987. By contrast, the total mortality rates 

 of older larvae {ZML and ZLL) and juveniles (ZJ) 



475 



