onomic groups: Medusae, Siphonophorae, 

 Ctenophora, Chaetognatha, Cladocera, Ostra- 

 coda, Amphipoda, Copepoda, Euphausiidae, 

 crustacean larvae, other invertebrate larvae, 

 pelagic MoUusca, Polychaeta, Tunicata, Radio- 

 laria, and Foraminifera. Anchovy eggs and 

 larvae, sized to the nearest 0.25 mm, were also 

 enumerated from these collections. Zooplankton 

 enumerations are given in detail in Alvariiio and 

 Kimbrell (1987). 



Larval Characteristics 



Larval Growth 



Growth of larvae was estimated from 224 lar- 

 vae collected in 15 tows at Site 1 and from 116 

 larvae collected in 30 tows at Site 2. Samples for 

 otolith ageing of larval anchovies were taken 

 from the portside net of bongo tows and were 

 preserved after removal of gelatinous zooplank- 

 ters in 80% ethanol buffered with 20 mM tris 

 [hydroxymethyl] aminomethane. Larvae were 

 sorted from the plankton and stored in the same 

 preservative. 



Standard lengths of larvae were measured 

 prior to removal of their otoliths. Preserved 

 standard lengths were converted to live lengths 

 using a correction factor for net shrinkage 

 (Theilacker 1980). At Site 1, tow duration was 

 six minutes and each sample was fixed within 

 five minutes of tow completion. Length of the 

 inshore larvae were corrected by a net shrinkage 

 factor of eight minutes. Tow duration at Site 2 

 was also six minutes, but because of the large 

 number of salps collected offshore, time before 

 preservation increased to about 10 minutes. 

 Lengths of offshore larvae were corrected for 13 

 minutes of net shrinkage. 



Daily increments in the otoliths were counted 

 using a compound microscope equipped with a 

 closed-circuit television system, a video coordi- 

 nate digitizer, and a microcomputer (Methot and 

 Kramer 1979). Age from hatching was deter- 

 mined from the number of daily rings in the 

 otoliths. Because the initial ring is deposited at 

 the time of first feeding, at about five days from 

 hatching, time since hatching is 5 days more than 

 the ring count (Brothers et al. 1976). 



Laird-Gompertz growth curves were fit to the 

 length at age using nonlinear regression. To 

 compare differences in growth rates between the 

 two sites, analysis of covariance was perfor;.._ ' 

 on segments of the data so that the assumption 

 of linearity was reasonable. 



FISHERY BULLETIN: VOL. 87, NO. 3, 1989 



Larval Production and Mortality 



Age-specific larval anchovy production rates 

 were computed from counts of larvae in 1 mm 

 size classes to 9 mm preserved standard length 

 (SL) from bongo tows and in 0.5 mm size classes 

 to 7 mm SL from CalVETs. Counts of larvae 

 were corrected for volume of water filtered, 

 for net avoidance, and for losses owing to ex- 

 trusion through the net meshes (Zweifel and 

 Smith 1981). Larval production at each age was 

 estimated by dividing size-specific larval abun- 

 dance by time spent at each size. Duration at 

 size was specfied from Laird-Gompertz growth 

 curves (Methot and Hewitt 1980; Lo 

 1983). 



Larval mortality curves were based on the 

 Pareto decay function (Lo 1985): 



z{t) = p/« 



where 2(0 is the instantaneous mortality rate 

 (IMR), (3 is the IMR coefficient, and t is age 

 (days) since spawning. Daily production is given 



by 



P, = Poit/tor^ to<t<20d 



where P, is daily larval production at age t, Pq is 

 initial larval production (at hatching), f,, is age at 

 hatching, specified from hatching time as a func- 

 tion of incubation temperature (Lo 1983). Thus, 

 as larvae grow older, the rate at which they 

 appear in the next age gi'oup is diminished by 

 the factor given by p. 



The above equation (referred to subsequently 

 as the nonlinear model) can be expressed in 

 linear form by taking the natural logarithm of 

 both sides: 



ln(P,) = In(Po) - P ln(^/^,) . 



Both the nonlinear and the log-linear regres- 

 sion models can be used to estimate parameters 

 Po and the IMR coefficient (3. The log-linear 

 equation is mainly used in this study to compare 

 P between sites. 



Production and mortality of anchovy larvae at 

 Site 1 were estimated from 48 bongo samples (20 

 using 333 |xm mesh nets and 28 using 505 p-m 

 mesh nets) and from 49 CalVETs using 150 |xm 

 mesh nets. At Site 2, these parameters were 

 estimated from 49 bongo samples using 505 \i.m 

 mesh nets and from 50 CalVETs using 150 jjim 

 mesh nets. 



676 



