Kendall et at.: Vertical distribution of eggs and larvae of Theragra chalcogramma 



541 



abiotic factors (Sclafani et al., 19931. Vertical migra- 

 tions in ichthyoplankton usually take place within the 

 upper water column and may be con-elated with the 

 vertical distribution of larval fish prey (Munk et al., 

 1989; Pritchett and Haldorson, 1989). In some species 

 the vertical extent of diel migrations increases as lar- 

 vae growl Ellertsonet al., 1981;Yamashitaet al., 1985). 



In early April a large population of walleye pol- 

 lock, Theragra chalcogramma, spawns in a restricted 

 region of Shelikof Strait, Gulf of Alaska. The resulting 

 planktonic eggs and larvae can be found, often in large 

 patches, through April and May. Recruitment varia- 

 tion of this population has been examined, particularly 

 those factors affecting interannual fluctuations in egg 

 and larval mortality (Schumacher and Kendall, 1991). 



Previous studies of the vertical distribution of 

 walleye pollock eggs and larvae in Shelikof Strait 

 have relied upon unmonitored discrete depth sam- 

 pling (Kendall et al., 1987; Kendall and Kim, 1989). 

 They have suggested that eggs occur generally be- 

 low 150 m, and systematically change depth during 

 their ~ 14-day incubation. Most larvae in these stud- 

 ies were found in the upper water column between 

 about 10 and 50 m, but they migrated vertically. They 

 seemed to congregate near the upper part of this 

 range around sunrise and sunset, were situated 

 somewhat deeper during midday, and were distrib- 

 uted more uniformly at night. A similar pattern for 

 larval walleye pollock has been observed in other 

 parts of the North Pacific (Kamba, 1977; Pritchett 

 and Haldorson, 1989; Walline 1 ). 



The present study describes the vertical distribu- 

 tion of walleye pollock eggs and larvae in relation to 

 changing environmental factors, measured on sev- 

 eral dates in Shelikof Strait. This study is based on 

 the analysis of 36 MOCNESS tows (Multiple Open- 

 ing-Closing Net and Environmental Sensing System; 

 Weibeet al., 1976) made over a period of three years. 

 We examine ontogenetic changes in vertical distri- 

 bution of eggs and larvae, and relate observed pat- 

 terns to estimated ambient light and to measured tem- 

 perature, salinity, and seawater density. Effects of ver- 

 tical distribution of larval prey and changes in mixed 

 layer depth due to wind events are also considered. 



Materials and methods 



MOCNESS tows were made from the NOAA ship 

 Miller Freeman during spring 1986, 1987, and 1988, 

 in areas where high abundance of eggs or larvae were 



1 Walline, PD. 1981. Hatching dates of walleye pollock (Theragra 

 chalcogramma) and vertical distribution of ichthyoplankton 

 from the eastern Bering Sea, June-July 1979. NWAFC Pro- 

 cessed Rep. 81-05. Northwest and Alaska Fish. Cent., NMFS, 

 NOAA, Seattle, WA 98115-0070, 22 p. 



expected ( Kendall and Picquelle, 1990 ) and had been 

 detected by exploratory sampling. Details of sampling 

 during these cruises are contained in Incze et al. 

 (1987), Proctor (1989), and Lawrence et al. (1991). 

 Some of the tows reported here (8-11 May 1986, se- 

 ries four) were also used in a study of the response of 

 zooplankton to the passage of a storm (Incze et al., 

 1990). On each MOCNESS tow, seven or eight nets 

 were deployed. The nets were towed obliquely at 0.8- 

 1.0 m/second through selected depth intervals and 

 opened and closed in sequence. Net depth, flowmeter 

 readings, temperature, and salinity ( 1987 and 1988) 

 were displayed in real-time and digitally recorded. 

 The nets had a nominal mouth area of 1 m 2 . Volumes 

 filtered per tow ranged from 34 to 980 m 3 (mean=167 

 m 3 ). Net mesh was 0. 153, 0.333, or 0.505 mm, depend- 

 ing on the size range of the larvae expected as well as 

 other sampling objectives. 



Depth intervals were based on previous studies 

 indicating that the eggs occurred primarily below 

 150 m and the larvae primarily in the upper 50 m 

 (Kendall et al., 1987; Kendall and Kim, 1989). Early 

 in the season when eggs were predominant, we fo- 

 cussed on the lower water column (>150 m), which 

 was subdivided into 20-m sampling strata. Later in 

 the season, we subdivided the upper water column 

 (<90 m) into 15-m sampling strata. 



Samples were fixed in 4-5% formalin and shipped 

 to the Polish Plankton Sorting Center in Szczecin, 

 Poland, where fish eggs and larvae were separated, 

 and walleye pollock eggs and larvae were identified 

 and counted. Stage of development of eggs was de- 

 termined in each tow taken early in the season ac- 

 cording to Blood et al. (1994). When more than 100 

 eggs were present in a sample, a subsample of 100 

 was staged. The egg stage data then were compressed 

 into six stage groups as in Kendall and Kim ( 1989). 

 Standard length (SLXtoO.l mm) of the larvae in each 

 sample was measured. A subsample of 50 larvae was 

 measured when more than 50 larvae were present 

 in a sample. Catches of eggs and larvae per depth 

 interval are reported as numbers per 1,000 m 3 of 

 water based on volume filtered as determined from 

 digital flowmeter records and net-frame angle. Mean 

 and standard deviations (SD) of numbers per 1,000 m 3 

 for each tow were computed from the weighted aver- 

 age of the numbers per 1,000 m 3 from each net within 

 the tow, by using the length of the depth interval for 

 each net as the weight. Estimation of egg and larval 

 mean depth, larval mean length, and their standard 

 deviations is based on cluster sampling where each 

 tow represents a cluster, each net subsamples eggs 

 and larvae from the cluster, and each egg or larva is 

 an element within the cluster ( Equations 8. 1 and 8.2 

 in Scheaffer et al., 1986). The observation associated 



