Dorsey et a\ Cohort abundances and mortality of Anchoa mitchilli 



265 



equivalent to a stage-specific rate. At 27"C, the mean 

 stage-specific mortality of bay anchovy yolk-sac larvae 

 (Z=1.27 1 is higher than the mortality predicted (Z=0.77) 

 from the temperature-dependent model of Pepin ( 1991 ) 

 for yolk-sac larvae of marine fishes. 



There was a high degree of spatial and temporal 

 variability in the abundances of bay anchovy eggs. 

 The initial abundances of eggs varied among days 

 (i.e. experiments) and probably among sites. Because 

 repeated experiments were carried out only at sites 

 4 and 5, it was not possible to critically analyze 

 among-site differences in egg abundances or their 

 mortality rates. Some of the variation may be ex- 

 plained by adult distributions and spawning strategy; 

 Luo and Musick ( 1991 ) hypothesized that bay anchovy 

 spawns preferentially in areas with plentiful adult food. 

 Our study supports their hypothesis because zooplank- 

 ton biovolumes were positively con-elated with initial 

 anchovy egg abundances. Additionally, the two lowest 

 initial egg abundances in our study occurred at site 4, 

 which also had lowest zooplankton biovolumes; the 

 highest egg abundance was observed at site 7, where 

 highest zooplankton biovolume occurred. 



Our estimates of relative adult biomasses may be 

 considerably higher than actual mean biomasses of 

 bay anchovy in Chesapeake Bay because we had se- 

 lected survey sites with high egg abundances. Our 

 criterion for selecting sampling sites was the obvi- 

 ous presence of bay anchovy eggs in test tows. This 

 criterion probably accounts for the high biomass es- 

 timates. Our mean biomass estimate, 1.16 g/m 3 ( 18.0 

 g/m 2 ), equates to approximately one adult anchovy 

 per m 3 . In a concurrent study of bay anchovy in the 

 mid and upper Chesapeake Bay, acoustics-adjusted 

 trawl abundances in July 1991 indicated a mean 

 adult biomass of 0.58 g/m 3 (Wang and Houde, 1995). 



Initial mean abundance of cohorts of yolk-sac lar- 

 vae, uncorrected for the extrusion of a fraction of lar- 

 vae through the 280-m net meshes, equaled 385.4/ 

 m 2 (24.6 [±6.0]/m 3 ). This unadjusted abundance was 

 approximately 17 times lower than the initial mean 

 abundance of egg cohorts. If yolk-sac larvae abun- 

 dances were adjusted upward by the factor (2.3) re- 

 quired to correct them ( MacGregor, 1994 ), the differ- 

 ence between initial egg and yolk-sac larvae abun- 

 dances reduces to a factor of 5.8. The observed mean 

 initial abundance of yolk-sac larvae, after adjust- 

 ment, based upon our mean egg mortality rate (0.066/ 

 h), was still only 50% of that expected. However, be- 

 cause eggs hatch at ca. 2000 h and the initial yolk- 

 sac larval abundance was estimated at ca. 0230 h, 

 continuing mortality would have caused further de- 

 clines in yolk-sac larval numbers for 6.5 h before 

 samples were taken. Applying the mean yolk-sac lar- 

 val mortality rate (0.053/h) to the 6.5-h period fur- 



ther reduced the mean discrepancy, although initial 

 numbers of yolk-sac larvae were still 30% lower than 

 expected. The cause of the remaining discrepancy, if 

 real, was not identified but could have resulted if a 

 fraction of the sampled eggs were infertile or dead. 



Estimated biovolumes of two gelatinous predators 

 of bay anchovy eggs and larvae, M. leidyi and C. quin- 

 quecirrha , were variable, probably patchy on tempo- 

 ral and spatial scales. Both have been demonstrated 

 to be major consumers of bay anchovy eggs and lar- 

 vae (Cowan and Houde, 1993; Purcell et al., 1994). 

 There was a significant correlation between mortal- 

 ity of yolk-sac larvae and C. quinquecirrha biovolume 

 (Table 3). Biovolumes of the two jellyfishes were high- 

 est at site 4, where lowest zooplankton biovolumes 

 and lowest anchovy egg and larval abundances oc- 

 curred. At the three southernmost sites (sites 1, 2, 

 and 3 in Table 1), neither M. leidyi nor C. quin- 

 quecirrha were present. Despite their absence, there 

 was no evidence that mortality rates of anchovy eggs 

 or yolk-sac larvae were lower there than at sites with 

 jellyfish present. 



Spawning intensity by bay anchovy, indexed by 

 initial abundances of egg cohorts, was lowest at site 

 4 where the gelatinous predator biovolumes were 

 highest, raising a possibility that anchovy adults 

 might avoid areas where zooplankton have been de- 

 pleted on account of consumption by gelatinous 

 predators. This suggests a mechanism that may in- 

 directly reduce baywide predation mortality on an- 

 chovy eggs and larvae; this, in turn, is supported by 

 the negative correlations between both egg and yolk- 

 sac larval abundances and jellyfish biovolumes. The 

 combined densities of C. quinquecirrha andM. leidyi 

 were extremely high at site 4, near the mouth of the 

 Potomac River, and possibly caused the depressed 

 zooplankton biovolumes. Purcell (1992) noted that 

 consumption by C. quinquecirrha depleted zooplank- 

 ton in tributaries of Chesapeake Bay, but she re- 

 ported smaller effects in the mainstem of the Bay. 



An important conclusion of our study is that bay 

 anchovy suffers highest mortality during its first two 

 days after being spawned; on average, 73% of all 

 spawned bay anchovy eggs in Chesapeake Bay died 

 before hatching. Of the surviving 279c, approximately 

 72% perished during the 24-h yolk-sac larval stage. 

 In July 1991 only 7.6%, on average, of each night's 

 egg production survived until the first-feeding stage, 

 which commences 2 days after hatching. Lasker 

 (1985b) and Smith (1985) pointed out that fluctua- 

 tions in yearly recruitment in engraulids may result 

 from variable stage-specific mortality rates, which 

 can be a consequence of interannual variability in 

 both biotic and abiotic factors. Results of our study 

 suggest that such factors also vary significantly on 



