132 



Fishery Bulletin 101(1) 



to the nearest 0.1 mg and their egg contents counted with 

 the aid of a dissecting microscope. The mean number of 

 eggs per gram from each fish was then expanded to the 

 total ovary weight to estimate annual fecundity (total eggs 

 per individual female). 



To estimate population weight-specific fecundity, first 

 the length and weight data were fitted to the power func- 

 tion and the bias-corrected regression equation was used 

 to estimate the weight of every fish that was aged. Next, 

 for each sex, growth equations were obtained by fitting the 

 weight and age "data" to the von Bertalanffy growth model 

 ( Ricker, 1975 ). Maturity was quantified by fitting the logis- 

 tic equation to the proportion of females that were mature 

 within 5-mm-TL intervals, and fecundity was estimated 

 by fitting fecundity and female weight data to the power 

 function, with appropriate bias-correction. These various 

 functional relationships were then combined in a life table 

 analysis to determine the expected biomass per female 

 recruit and the expected lifetime lai-val production per fe- 

 male recruit. The ratio of these quantities is the estimated 

 equilibrium cohort weight-specific fecundity (i.e. Oflarvae/ 

 gl ) of female shortbelly rockfish. Given an estimate of total 

 age-0 larvae (N,,), the female biomass responsible for the 

 observed larval production can be estimated as A^q^"'. 

 Finally, the combined sex biomass can be determined by 

 expanding female biomass using weight-based cohort sex 

 ratio estimates from the life table analysis. 



Ichthyoplankton sampling 



The primary set of ichthyoplankton samples used in our 

 study was obtained by using bongo nets during a cruise of 

 the NOAA RV David Starr Jordan (DSJ-9102) conducted 

 in the winter of 1991. Sampling began at 1500 h on 8 Feb- 

 ruary and ended at 0230 h on 15 February. During that 

 6V2-day period 150 stations were occupied in the region 

 bounded from lat. 36°.30'N to lat. 38°00'N and offshore to 

 a maximum distance of 130 km (Fig. 1). The study area 

 included Pioneer and Ascension Canyons — two features in 

 the continental slope known to harbor large numbers of 

 adults (Lenarz, 1980; MacGregor, 1986; Chess et al., 1988). 

 At these sites the sampling density was increased. 



Field and laboratory processing of the bongo net samples 

 followed prescribed CalCOFI guidelines (i.e. Kramer et 

 al., 1972; Smith and Richardson, 1977), with minor modi- 

 fication. For example, at every fifth sampling station, the 

 bongo frame was deployed with 333-pm and 505-)im mesh 

 nets to determine the extent of extrusion of small larvae 

 in the standard 505-|im mesh (Lenarz, 1972; Somerton 

 and Kobayashi, 1989). After the nets were washed down, 

 samples from both mesh sizes were preserved in 80% 

 EtOH. At all other stations, the net frame was deployed 

 with two 505-pm mesh nets; one sample was preserved in 

 80'/^ EtOH (to allow later age determinations from larval 

 otoliths), and the other was preserved in lO'/f buffered for- 

 malin. In addition, because Sehastes larvae were believed 

 to occur only in the upper mixed layer (Ahlstrom, 1959), 

 the maximum amount of wire deployed was 200 m, result- 

 ing in a maximum depth fished ctiual to 140 m. Following 

 splitting, sorting, identification, and enumeration of the 



larvae in the laboratory, abundance was expressed as the 

 number of shortbelly rockfish per 10 m^ of sea surface. 



To estimate the age composition of the larval popula- 

 tion, the sorted shortbelly rockfish larvae from each of the 

 150 EtOH-preserved bongo hauls were randomly subsam- 

 pled for otolith microstructure examination. To determine 

 the size of an age subsample (N^), based upon the total 

 number of larvae occurring in a haul (A^^, ), we applied the 

 following rule: 1) for A^,, less than or equal to 10, N^ = N/^, 

 2); for A^^, greater than 10 but less than or equal to 410, N^ 

 = 10-1- 0.10 [AT; -10]; and 3) for A^;, greater than 410, N^ = 5o! 

 Otoliths were extracted from each specimen in the haul 

 subsample and individual ages determined by methods 

 outlined in Laidig et al. (1991) and Ralston et al. (1996). 



The age composition of the larvae in each bongo sample 

 was then estimated by expanding the percent age-fre- 

 quency obtained from the subsample to the haul total (N/,). 

 The estimated numbers-at-age of larvae in each haul were 

 standardized to the number per 10 m^ of sea surface irij.^ 

 for age T and haul i) by application of standard haul fac- 

 tors (Kramer et al., 1972; Smith and Richardson, 1977). 



We expanded the n-j.^ to the entire survey area by using 

 the method of Sette and Ahlstrom ( 1948). The Sette-Ahl- 

 strom estimate is calculated by the following equation (see 

 Kendall and Picquelle, 1990): 



^T=Y,^^'^T" 



where for each of k hauls A, = the area that haul ( repre- 

 sents (units of 10 m'-); and 

 Nj, = the total abundance of lar- 

 vae of age T in the entire 

 survey area. 



The area for a haul (A, ) is defined as the area circumscribed 

 by a polygon containing all points in space closer to a 

 haul's location than to the location of any other haul. With 

 this definition, we were able to write a simple computer 

 program to calculate Sette-Alilstrom weights by dividing 

 the study area into a fine grid and assigning each grid 

 point to a haul. Note that this definition and procedure for 

 obtaining Sette-Ahlstrom areas is equivalent to construct- 

 ing polygons manually using perpendicular bisectors and 

 measuring their areas (Sette and Ahlstrom, 1948). 



The mean daily larval production rate during the cruise 

 was estimated as the bias-corrected antilogarithm of the 

 y-intercept of the ordinary least-squares linear regression 

 of log.lA^J against larval age. Moreover, the regression 

 slope provides an estimate of the total instantaneous mor- 

 tality rate of the larvae (Z |/d|). This calculation implicitly 

 assumes that the age distribution of larvae was stationary 

 throughout the 6V2-day period of the cruise. 



To determine if shortbelly rockfish larvae occur deeper 

 in the water column than 140 m, a series of 1-m^ multiple- 

 opening-closing-net with environmental sensing system 

 (MOCNESS) tows was conducted aboard the RV David 

 Starr Jordan (cruise l)SJ-9203i during the 1992 spawn- 

 ing season. At that time, 21 tows were made in the area of 



