524 



Fishery Bulletin 89(3). 1991 



came from larvae that were extruded early or late in 

 the season. Birthdate distributions were also compared 

 among species to reveal the extent of interspecific 

 covariation in annual survivorship. 



Methods 



Midwater trawl collections 



Annual 30-day cruises aboard NOAA's R/V David 

 Starr Jordan began in 1983 and were conducted dur- 

 ing late spring (May- June), a time when most pelagic- 

 stage juvenile rockfishes are identifiable to species, but 

 prior to their settling to nearshore and benthic habitats. 

 The sampling gear used was a midwater trawl with a 

 26 m headrope and 169 m 2 net opening, equipped with 

 a 0.945cm (3/8") inner liner. A standard haul consisted 

 of a 15-min nighttime tow at a depth of 30 m covering 

 a distance of roughly 1km. Additional tows were made 

 at other depths (i.e., 10 and 100m) as time and bottom 

 topography permitted. 



From 1983 to 1985, the survey was designed to 

 sample from nearshore waters out to a distance of 

 55km (30nmi) along transects located at 31km inter- 

 vals perpendicular to the coast. Sampling began at Pt. 

 Sur (36°18'N) and continued northward until comple- 

 tion of the cruise. In 1986, the sampling design was 

 modified to permit three "sweeps" through a study 

 area bounded by Cypress Pt. (36°35'N) and Pt. Reyes 

 (38°10'N) (Table 1). Four additional cruises were con- 

 ducted in April of 1985-88. These cruises ranged from 

 5 to 10 days in length and were limited to the Gulf of 

 the Farallones (1985-86), or covered one complete 

 sweep of the study area (1987-88). Trawls were con- 

 ducted at 5-7 stations per night along 7 transects dur- 

 ing a sweep. Wyllie Echeverria et al. (1990) discuss the 

 survey methodology in greater detail. 



As many as 24 species of pelagic-stage juvenile 

 rockfish were collected during a single cruise. All 

 specimens were tentatively identified and frozen at sea. 

 On rare occasions catches were subsampled when they 

 were too large to enumerate. After 1987, length com- 

 positions of large trawl catches were based upon ex- 

 pansion from measured subsamples of each species. 



Otolith preparation and analysis 



In the laboratory, species identifications were con- 

 firmed and fish were measured to the nearest 1.0 mm 

 standard length (SL). For this study, the most impor- 

 tant species were selected for age determination. A 

 subsample of juveniles, measured to the nearest 0.1 

 mm, was selected for daily age analysis based on the 

 size range encountered on each cruise. Specimens were 



Table 1 



Summary of sampling dates for juvenile rockfish cruises. 

 Asterisks indicate trawls limited to Gulf of the Farallones. 



Year 



Sweep 



Dates 



Stations 



selected at 0.5mm intervals from the smallest to 

 largest. The sagittal otoliths were removed, cleaned, 

 and adhered concave (lateral)-side-down onto glass 

 microslides using clear nail polish. To reveal the inner 

 microstructure, otoliths were ground with wet 300-600 

 grit sandpaper and etched with HC1 (Brothers et al. 



1976). 



A compound microscope with 40 and 100 x objectives 

 was used during otolith examination. Immersion oil was 

 applied to the otolith, which was illuminated with 

 polarized light. A closed circuit television camera was 

 mounted on top of the microscope and the image was 

 relayed to a video monitor. A digitizer with a precision 

 of 0.13Mm was interfaced with the monitor and a 

 microcomputer to accurately measure and record daily 

 increments (e.g., McGowan et al. 1987). In combina- 

 tion, the equipment yielded an 800-2000 x on-screen 

 magnification range. 



A computer program assisted in recording the dis- 

 tance from the nucleus to each increment, measured 

 along the growth axis running from the focus to the 

 postrostral margin. On occasion the periodic bipartite 

 microstructure of daily increments was difficult to 

 resolve within short (~50Mm) segments of the otolith. 

 In these cases their number and width were inter- 

 polated by averaging the widths of the rings imme- 

 diately preceding and following that segment, and 

 dividing the mean into the segment length. Usually 

 such averaging involved only about 10% of the total 

 increments observed. In most cases (~80%), only one 

 otolith was used unless it was damaged or destroyed 



