FISHERY BULLETIN: VOL. 87, NO. 3, 1989 



large) or source (e.g., full-sibling groups) and the 

 analysis of the influence of egg size on ELH 

 traits was conducted among these groups. Cor- 

 relations, when reported, were based on the 

 average trait values in these groups. Such analy- 

 ses obscure variation in and correlation between 

 traits at the individual level — where natural se- 

 lection acts. 2) Simple correlations were calcu- 

 lated. These may reflect the response of two 

 traits to a third covarying trait rather than ap- 

 praise the influence implied. 3) Egg size was 

 confounded with other factors not explicitly con- 

 sidered (e.g., different source populations, 

 seasons of spawning, years, and ages of fe- 

 males). The influence attributed to egg size is, 

 therefore, equivocal. These are important limita- 

 tions. The estimates of relationships between 

 egg size and subsequent ELH traits so derived 

 cannot be presumed to be of the same magnitude 

 or statistical significance as those from within- 

 individual observations. The conclusions of 

 previous studies must be viewed as provisional 

 demonstrations of associations of egg size with 

 the ELH traits considered. 



The extent to which relationships exist at the 

 individual level between egg size, hatchhng size, 

 and prestarvation lifespan remains unanswered. 

 In this paper we analyze such ELH trait covaria- 

 tion in capelin, Mallotus villosus. We address 

 two questions: First, which pairs of ELH traits 

 covary when evaluated at the individual level 

 while (statistically) holding other traits con- 

 stant? Of particular interest is which, if any, of 

 these traits vary with egg quality. Second, is the 

 pattern of ELH trait variation attributable to 

 maternal influences? 



METHODS 



Spawning capelin were collected on 23 and 25 

 June 1988 at spawning sites in eastern Newfound- 

 land, Canada (lat. 49°39'N, long. 52°41'W). 

 These capelin were held in a 1,000 L flow-through 

 tank at the Marine Science Research Laboratory, 

 Logy Bay, Nfld. until 28 June when crosses were 

 performed. Eggs from each of 10 females were 

 fertilized with milt from one male. This provided 

 a range of eggs sizes and created the potential for 

 female effects on ELH traits of offspring while 

 minimizing male influence. 



Yolk diameters of 12-15 eggs per female were 

 measured at 20-28 h postfertilization at a magni- 

 fication of 40 X. These embryos were in the 

 blastula stage and their yolks remained approxi- 

 mately spherical. Embryos were then trans- 



ferred individually to separate wells of 12-well 

 tissue culture plates. These plates served to lim- 

 it possible interaction among incubating em- 

 bryos (e.g., bacterial growth on dead eggs) by 

 keeping embryos at least 3.5 cm apart. In total, 

 142 eggs were housed in 12 plates, each contain- 

 ing 1 or 2 fertilized eggs per female. Each set of 4 

 plates was submerged to a depth of 10 cm in a 

 separate plastic tank containing static seawater; 

 thus, all embryos in a tank were exposed to the 

 same seawater. This seawater was filtered to 1 

 p-m, sterilized with ultraviolet light, and 

 changed every 2 days to ensure sufficient levels 

 of oxygen for the developing embryos. All three 

 static seawater tanks were placed in a common 

 water table maintained at 9.5 ± 0.49°C (mean ± 

 SD). At the late eyed stage, embryos were 

 transferred individually to separate 125 mL 

 glass jars containing 100 mL of filtered, steril- 

 ized seawater. Embryos in jars occupied the 

 same location in the water table as they had 

 before transfer. Fluorescent lamps in the labora- 

 tory were set to 18 h light, 6 h dark. 



Embryos were inspected daily and their day of 

 hatching recorded. Hatchlings were anaesthe- 

 tized (MS-222, 25 mg/L), measured, revived, and 

 returned to their jars. Total length (TL), length 

 and height of yolk, and diameter of oil globule 

 were measured. Larvae were inspected daily 

 until their death, which was inferred if a larva 

 did not respond to gentle tapping of its holding 

 jar. Deformed hatchlings (3.5%) and those failing 

 to escape the egg membrane (1.4%) were ex- 

 cluded from analyses. 



Immediately after spawning, each female was 

 measured (standard length, SL) and frozen. Sub- 

 sequently, but within 2 months, females were 

 thawed and somatic weights determined. Con- 

 stant dry weights were achieved after oven dry- 

 ing at 60°C for 48 hours. These dried carcasses 

 were ground and their lipids were extracted by 

 flooding with diethyl ether and decanting 

 (Dobush et al. 1985). Constant lean, dry weights 

 were obtained after three cycles of lipid extrac- 

 tion. Indices of somatic condition (wet 

 weight/SL) and lipid condition (extractable lipid 

 weight/lean dry weight) were calculated for each 

 female. 



We analyzed six ELH traits: initial yolk vol- 

 ume (Yl), total length at hatching (TL), yolk 

 volume at hatching (Y2), oil globule volume at 

 hatching (G), age at hatching (H), and posthatch- 

 ing lifespan (L). Initial egg yolk and oil globule at 

 hatching were approximately spherical and we 

 estimated their volumes from their diameters. 



516 



