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



be a mechanism of density-dependent compensa- 

 tion. 



Growth rates of larval and juvenile anchovy 

 during El Nino periods are of particular inter- 

 est, because the growth rates during an environ- 

 mental perturbation may reveal details of the 

 processes of growth and survival during normal 

 conditions. During the 1982-83 El Nino event 

 off California, temperatures were elevated 

 (Lynn 1983; Simpson 1983) and zooplankton den- 

 sity was reduced in 1983 (McGowan 1985). If 

 planktivorous fish compete for food, competition 

 should be highest when zooplankton abundance 

 is lowest. Comparison of growth rates through- 

 out the larval and juvenile stages during periods 

 of low and high food abundance may show the 

 size range at which the larval or the juvenile 

 stage is most important. In this paper I examine 

 the interannual variability in growth rates of 

 anchovy larvae and juveniles during periods of 

 low and high food abundance. 



METHODS 



Samples of juveniles from the central popula- 

 tion of northern anchovy were obtained from 

 midwater trawl hauls taken in 1980-84 by the 

 Sea Survey Program of the California Depart- 

 ment of Fish and Game. Annual cruises to moni- 

 tor the strength of the incoming year class of 

 northern anchovy have been conducted during 

 September and October since 1976; Mais (1974) 

 has described the gear and sampling procedures. 

 The surveys extended along the Pacific coast 

 from central Baja California to Point Concep- 

 tion, CA, with the exception of 1982, when the 

 southern limit was the U.S. -Mexican border. 

 Trawl hauls were taken nearshore inside of a 75 

 fm contour where young of the year are most 

 abundant during the fall (Parrish et al. 1985). 



Twenty five juvenile northern anchovy were 

 randomly taken from each positive trawl haul 

 and frozen for age determination. A subsample 

 of 200 juvenile anchovy was randomly taken 

 from the 800 to 1,000 fish collected on each sur- 

 vey for analysis of growth rate using daily incre- 

 ments in the otoliths. Fish were measured from 

 the tip of the snout to the posterior edge of the 

 hypural plate (standard length), and otoliths 

 were removed and mounted on microscope slides 

 using Eukitt^ mounting medium. 



Growth rates and ages of juvenile northern 



'Reference to tradenames does not imply endorsement by 

 the National Marine Fisheries Service, NOAA. 



anchovy were determined by counting and 

 measuring daily increments in the otoliths 

 (Methot 1981). The increment widths were 

 measured and recorded along a transect from the 

 focus to the posterior margin of the otolith 

 (otolith radius), using a video-coordinate digi- 

 tizer connected to a microcomputer. The otoliths 

 were progressively polished between readings, 

 using 15 and 0.3 (x lapping film, to reveal incre- 

 ments along the entire transect. Data from dif- 

 ferent areas of the otoHth were collected as the 

 increments became visible. During each reading 

 groups of 10 or less (typically five) increments 

 were counted and measured. The number of 

 increments measured was reduced if width var- 

 ied among increments. 



ANALYSIS AND RESULTS 



Data from both otoliths and several replicate 

 transects per otoHth were combined to calculate 

 age and otohth increment widths (Methot 1981). 

 Mean increment width was calculated at each 

 point along the radius from the focus to the pos- 

 terior margin of the otolith. In some cases, it was 

 impossible to obtain measurements of all incre- 

 ment widths in the otoliths. In these cases, in- 

 crement widths were interpolated using linear 

 approximation (see Methot 1983 for details). 

 Data for increment widths and age estimates 

 were not used if more than 5% of age was calcu- 

 lated from interpolated increments. 



A direct comparison of the average size of 

 recruits from year to year is inappropriate be- 

 cause their ages may differ from year to year 

 owing to differences in the dates of spawning 

 (Methot 1983) or to dates of collection. To elimi- 

 nate this bias, the dates of hatching were deter- 

 mined from the dates of sampling and ages, and 

 fish were grouped by year and month of hatch- 

 ing. 



Interannual and seasonal differences in length 

 were tested, using an analysis of covariance 

 (Bartlett et al. 1984) with age as the covariate. 

 This analysis adjusted the mean length to the 

 grand mean age of 208 days, approximately the 

 time elapsed from the peak of spawning to the 

 date of capture. Since the relation of length to 

 age is typically nonhnear, some error is intro- 

 duced using a linear adjustment. The difference 

 in dates of sampling is, however, only about one 

 month and the relation of size to age is approxi- 

 mately linear from two to six months. 



Analysis of covariance of these data indicates 

 parallel lines with differences in means (Case 2c 



646 



