JOHNSON and BARNETT: CORRELATION BETWEEN CHARACTERS AND FOOD SUPPLY 



tance of two principal dangers to larval survival: 

 starvation versus predation. 



We hypothesize that selection on mid-water fish 

 populations inhabiting areas of low productivity 

 has favored mechanisms tending to offset the 

 danger of larval starvation, and that these 

 populations will exhibit: 1) larger average egg size, 

 2) lower fecundity, and 3) larger average larval 

 size at hatching and at comparable stages of 

 development than populations living in areas of 

 higher productivity. Advantages that might 

 accrue to larger larvae in areas of lower produc- 

 tivity include increased mobility, a wider possible 

 search volume, increased diversity of potential 

 prey organisms, and a longer period of sur- 

 vivorship solely on yolk reserves. 



The danger of starvation is presumably lower in 

 areas of higher productivity but the danger of 

 predation, resulting from presumed higher densi- 

 ties of potential predators on fish larvae, may be 

 greater. Here selection may have favored 

 increased fecundity tending to offset the danger 

 of increased predation on larvae. We believe that 

 in areas of higher productivity populations of 

 mid-water fishes will exhibit: 1) smaller average 

 egg size, 2) increased fecundity, and 3) smaller 

 average larval size at hatching and at comparable 

 stages of development than populations living in 

 areas of lower productivity. 



In developing this hypothesis we have largely 

 followed Hempel's (1965) explanation for varia- 

 tions in egg size and fecundity between popula- 

 tions of herring in the eastern North Atlantic and 

 North Sea. We note that there exists no evidence 

 for increased predation pressure on larval popula- 

 tions of mid-water fishes in areas of higher 

 productivity. It is possible to retain the main fea- 

 tures of our hypothesis without including preda- 

 tion pressure by relating variation in fecundity 

 and egg size solely to food density requirements. 

 By definition, selection will favor maximizing 

 reproductive output, thereby favoring fewer 

 larger larvae in areas of low productivity where 

 the danger of larval starvation is greater, and 

 favoring higher fecundity (with the concomitant 

 of smaller eggs and larvae) in areas where the 

 danger of starvation is lessened. 



There exists limited available evidence to sup- 

 port these predictions. Ahlstrom and Counts (1958) 

 showed that egg size and size at hatching in Vin- 

 ciguerria lucetia are directly related, the smallest 

 larvae (at a defined stage of development) are 

 found in areas where average egg diameters are 



least. They also showed that mean values of ver- 

 tebral and IPVALA photophore counts are lowest 

 in those areas where egg and larval size is least 

 and where metamorphosis occurs earliest (i.e. at 

 smallest size). Although no small larvae were 

 available for this study, we were able to compare 

 development in prejuvenile specimens of V. nim- 

 baria from the central North Pacific with 

 specimens from the central equatorial Pacific. In 

 V. nimbaria the last four VAL photophores are 

 late-forming, are laid down serially from anterior 

 to posterior, usually the left member of a pair of 

 VAL photophores develops just before the right, 

 and the number yet to develop can be determined 

 uniquely from the one to one correspondence with 

 the posterior photophores in the V AV segment. In 

 Figure 5 standard lengths of all available 

 prejuvenile specimens of V. nimbaria from the 

 central North Pacific and central equatorial Pacific 

 are plotted against the number of VAL pho- 

 tophores left to appear. If this character can be 

 used as an index to comparable stages of develop- 

 ment, then at comparable stages of development 

 the larvae from the area of lower productivity are 

 the larger, as predicted. 



We believe that the correlation between meris- 

 tics and productivity results from a correlation 

 between meristics and egg size, and that egg size 

 and, hence, size at hatching are genetically deter- 

 mined features reflecting adaptation to produc- 

 tivity conditions. A number of authors have stated 

 or suggested that such a correlation exists (Ahl- 

 strom and Counts 1958; Lindsey 1958, 1961; Garside 



Figure 5.— Size of larvae of Vinciguerria nimbaria from the 

 central North Pacific (CNP) and the central equatorial Pacific 

 (CEP) at comparable stages of development. Ordinate: standard 

 length in millimeters. Abscissa: number of VAL photophores yet 

 to form (determined from VAV count) on each side of each 

 specimen. Lines fitted by eye. 



295 



