480 



Fishery Bulletin 88(3)^ 1990 



There appears to be a positive correlation between 

 egg size and latitude in walleye pollock (Table 1). This 

 correlation was also noted by Gorbunova (1954). In this 

 study, this correlation was seen both among widely 

 separated areas such as Puget Sound and the Gulf of 

 Alaska, and within the Gulf of Alaska itself. Latitudinal 

 clines in egg size have been seen by Demir (1963), 

 Chiechomski (1973), and others, and have been dis- 

 cussed by Rass (1941), Marshall (1953), and Thresher 

 (1988). As noted by Thresher, "Latitudinal variation 

 in life histories of marine organisms typically are at- 

 tributed to selection acting on local populations along 

 a latitudinal environmental gradient. . ." 



Several studies have noted an inverse relationship 

 between egg size and temperature (Southward and 

 Demir 1974, Ware 1977, Marsh 1984, Houghton et al. 

 1985, Imai and Tanaka 1987). Imai and Tanaka (1987), 

 working with Japanese anchovy Engraulis japonicn, 

 found yearly changes in egg size which correlated with 

 temperature, and were able, experimentally, to alter 

 egg size during the spawning cycle of anchovy held in 

 the laboratory by changing water temperature. Ware 

 (1977) noted that peak spawning of Scotnher scomhrus 

 occurs later if water temperatures warm more slowly 

 than usual in the spring. 



Temperature may be the latitudinal environmental 

 gradient along which selection for egg size acts on local 

 populations. The relationship between latitude and egg 

 size appears to be correlated to temperature in the dif- 

 ferent spawning regions over the range of walleye 

 pollock (Table 1). In the laboratory experiments on 

 spawning done for this study, ambient water tempera- 

 ture increased over the spawning period, while egg size 

 declined significantly for each female over this period. 

 Mean monthly egg size and mean monthly water 

 temperatui'e at the depth where eggs are spawned (150 

 m to the bottom) within Shelikof Strait, however, show 

 no apparent inverse correlation. There was also no in- 

 verse correlation between mean egg size in April and 

 mean April water temperature at the depth of spawn- 

 ing over the years examined (J.D. Schumacher, Pacific 

 Mar. Environ. Lab., Seattle, WA 98115-0070, unpubl. 

 data). It may be that temperature acts on the parent 

 stock at some earlier date, for example while the stock 

 is migrating to the spawning grounds. Data are insuf- 

 ficient to investigate this question. 



Consequences of egg size 

 variation for larval survival 



A correlation between egg size and larval size has also 

 been noted in other species {Clupea harengus, Blaxter 

 and Hempel 1963; Salvelinus alpinus, Wallace and 

 Aasjord 1984; Etheostoma spectabile, Marsh 1986; 

 Gadus morhua, Solemdal 1970, Knutsen and Tilseth 



1985). Only a few studies have reported no correlation 

 (Zonova 1973, Reagan and Conley 1977, Lagomarsino 

 et al. 1988). 



Egg size has been correlated with other larval fac- 

 tors. Growth rate, for example, has been seen to in- 

 crease with increased egg size (Blaxter and Hempel 

 1963, Bagenal 1969, Wallace and Aasjord 1984, Moodie 

 et al. 1989). This may be related to increased feeding 

 success due to larger mouth sizes or increased search 

 area (from increased reactive distance or swimming 

 speed). Mouth size has been positively correlated with 

 egg size (Blaxter and Hempel 1963, Shirota 1970, 

 Knutsen and Tilseth 1985). Larvae with larger mouths 

 are able to take larger or more varied sizes of prey 

 items. Length of time from hatch to starvation (due 

 to increased endogenous yolk reserves; Blaxter and 

 Hempel 1963, Bagenal 1969, Theilacker 1981, Marsh 

 1986, Wallace and Aasjord 1984), and survival (Blax- 

 ter and Hempel 1963, Bagenal 1969, Pitman 1979, 

 Small 1979, Moodie et al. 1989) have also been positive- 

 ly correlated with egg size. If larval survival is affected 

 by larval size at hatch (Miller et al. 1988) and other fac- 

 tors, then variation in egg size in walleye pollock may 

 cause differences in larval mortality rates. 



It has been proposed that changes in egg size may 

 be an adaptation to the timing of the production cycle, 

 to ensure that larvae are produced that are able to take 

 advantage of the available food supply (Gushing 1967; 

 Hempel and Blaxter 1967; Bagenal 1971; Jones and 

 Hall 1974; Ware 1975, 1977). The reproduction of zoo- 

 plankton (such as copepods, e.g., Pseudocalanus and 

 Oithona spp., whose naupliar stages constitute the 

 most common food of first-feeding walleye pollock lar- 

 vae in the Bering Sea and the Gulf of Alaska; Clarke 

 1978. 1984; Nishiyamaand Hirano 1983; Kendall etal. 

 1987) is temperature and food related (Checkley 

 1980a,b; Durbin et al. 1983; Runge 1985; Corkett and 

 McLaren 1978; Landry 1976). This implies that tem- 

 perature would provide information on the timing of 

 zooplankton reproduction, and therefore on the size of 

 the available larval food supply (Ware 1977). Informa- 

 tion on the changes in species and size composition of 

 larval food sources in the western Gulf of Alaska is not 

 presently available. The observed trends in egg size 

 would be explained in terms of larval food supplies if 

 the abundance of prey items increased and the size of 

 food particles decreased and became more uniform as 

 the season progressed. 



The observed trends in egg size may also be a result 

 of an adaptation to relieve predation pressure on eggs 

 and larvae. Seasonal changes in egg size could result 

 in changes in the predator-prey size ratio for eggs and 

 larvae, affecting predation rates (Bailey and Houde 

 1 989). This may not be important if egg predators are 

 mainly planktivorous fishes. With zooplankton pred- 



