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



551 



et al., 1987). In series three, the mean depths of oc- 

 currence of eggs and larvae decreased to between 68 

 and 110 m, and the mean length of larvae increased 

 to 4.4 mm, indicating that growth of some pollock 

 larvae had occurred (SD of length was 0.18 in series 

 three as opposed to 0.09 in series two). The standard 

 deviations of depth of larval occurrence in series three 

 (83 to 91 m) were larger than in any other series, 

 suggesting that these larvae were in transition from 

 the deep hatching environment to shallower levels. 

 Although larvae respond positively to light within 

 24 hours of hatching, their negative geo/barotaxis 

 may enable them to reach the upper layers since in- 

 sufficient light for response penetrates to hatching 

 depths (see Olla and Davis, 1990). The relatively 

 shallow mean depth of eggs in series three also may 

 account for the large variation in larval depth dur- 

 ing that series. Older larvae from eggs at the depths 

 observed during series two (>150 m) could have 

 mixed with larvae hatching from eggs found at the 

 depths observed during series three (<110 m). The 

 larger standard deviation of larval length in series 

 three compared with series two supports this expla- 

 nation. In later series the mean depth of occurrence 

 of the larvae was less than 60 m, and the standard 

 deviation generally was less than 20 m. Once they 

 reach the upper layers, vertical movements increase 

 as larvae develop. No significant diel migrations were 

 noted in series four, as opposed to the pattern seen 

 in series nine. The mean length of larvae in series 

 four was 5.3 mm; in series nine it was 7.9 mm. The 

 larvae sampled by Kendall et al. (1987) in late May 

 were 11.0 mm long and demonstrated a pattern of 

 vertical distribution similar to the larvae collected 

 here in series nine. 



During series nine, larvae followed a diel (crepus- 

 cular) pattern of vertical movements in which they 

 ranged deepest at noon, shallowest at dusk, and pro- 

 gressively deeper through the following noon. Al- 

 though this pattern was observed on both days, the 

 amplitude of movements were reduced on the sec- 

 ond day. However, the wind had markedly increased 

 by evening of the second day Larvae may have been 

 avoiding the turbulent surface on the second day 

 when their mean depths were deeper. Olla and Davis 

 ( 1990) found that pollock larvae avoid turbulence in 

 the laboratory. 



The relationships of larval fish feeding, growth, 

 and survival to storms and turbulence have been the 

 subject of numerous studies (e.g. see Sundby and 

 Fossum, 1990; Maillet and Checkley, 1991). Both 

 positive and negative effects have been postulated 

 and observed. Positive effects of increased turbulence 

 include hypothesized enhanced encounter rates be- 

 tween larvae and their prey (Rothschild and Osborn, 



1988), and enhanced primary production after mix- 

 ing has ceased owing to infusion of nutrients from 

 below the photic zone. Negative impacts include di- 

 lution of vertically enriched layers of prey to levels 

 below successful feeding thresholds and reduced 

 naupliar production in lower phytoplankton concen- 

 trations (Lasker, 1978). There is evidence that in 

 Shelikof Strait, below-average walleye pollock pro- 

 duction may result if strong wind events occur when 

 larvae are at the first-feeding stage. 5 Incze et al. ( 1990 ) 

 found that naupliar concentrations remained above 

 feeding threshold levels during the passage of a 

 storm, but this was a relatively transient phenom- 

 enon. The present study indicates that larvae may 

 avoid upper layer turbulence by moving deeper in 

 the water column. If so, they might experience prey 

 densities or light levels too low for optimal feeding. 



In a 24-hour study of the vertical distribution of 

 pollock larvae in Auke Bay, Alaska (average depth 

 60 m), Pritchett and Haldorson (1989) found larvae 

 congregated at 10 m at noon, at 5 m at dawn and 

 afternoon, and at 15-20 m at night. At twilight (0.3 

 hour before sunrise, 1.5 hours before sunset), larvae 

 were more dispersed, seen mostly at 10 m near sun- 

 rise and at 15 m near sunset. The vertical extent of 

 diel migration increased with larval length. In the 

 present study, depths of occurrence were greater at 

 all times than those reported by Pritchett and 

 Haldorson ( 1989), and noon depths of larger larvae 

 were greater than the night depths. However, in both 

 studies, larvae were found to be deeper at noon than 

 at dawn and dusk, and a relationship between verti- 

 cal migration and larval length was seen. The depth 

 distribution of copepod nauplii in Auke Bay usually 

 centered around 5-10 m (Paul et al., 1991). Inzce 

 et al. ( 1990) reported maximum densities of copepod 

 nauplii in the upper 30 m of Shelikof Strait when 

 pollock larvae are abundant. Since nauplii are the 

 primary prey for pollock larvae, the larvae may well 

 adjust their daytime feeding depths to correspond to 

 those of the nauplii. Alternatively, the greater day- 

 time depth of pollock larvae in Shelikof Strait may 

 be related to the greater depth of light penetration 

 in Shelikof Strait compared with Auke Bay (Zeimann 

 etal., 1990). 



Light is frequently cited as a factor controlling the 

 depth of occurrence for fish larvae. Larvae of some 

 species follow the common trend of rising toward the 

 surface at night and of remaining deeper during the 

 day (Smith et al., 1978; Kendall and Naplin, 1981; 

 Davis et al., 1990). Other species follow an opposite 

 pattern, ranging deeper at night than by day (Boeh- 



5 Bailey, K. M., AFSC, and S. A. Macklin, Pacific Marine Envi- 

 ronmental Laboratory. 7600 Sand Point Way NE., Seattle, Wash- 

 ington 98115-0070. Personal commun., February 1993. 



