FISHERY BULLETIN: VOL. 73, NO. 2 



this case. (2) The data for Vinciguerria nimharia 

 may indicate the existance of separable popula- 

 tions in our different study areas. This is sug- 

 gested by the results for the Antipodes transect 

 (Table 9) in which is found no clear evidence for an 

 onshore to offshore trend toward higher IPVALA 

 counts, despite the fact that the productivity 

 measures are higher inshore and decrease 

 (rapidly) to seaward (Reid 1962; Koblentz-Mishke 

 et al. 1970). Mean values of IPVALA photophore 

 counts for specimens from each of the 11 pairs of 

 Philippine Sea stations are significantly higher 

 than the mean value for South China Sea 

 specimens. This may suggest that genetically dis- 

 tinct, separable populations of V. nimharia are 

 found in each area. Gill raker counts for V. nim- 

 haria (Table 10) apparently support this sugges- 

 tion in that counts of gill rakers are discordant 

 with counts of IPVALA photophores (Table 6) and 

 vertebrae (Table 7). For the four Pacific areas the 

 counts of vertebrae and IPVALA photophores for 

 V. nimharia agree in perfect rank-order with the 

 IPVALA photophore counts for D. taenia (Table 

 10). That this is not true for gill raker counts may 

 indicate the existence of separable populations of 

 V. nimharia in the South China Sea, central 

 equatorial Pacific, and the North Pacific central 

 gyral areas (Philippine Sea, central North Pacific). 

 (3) The ecophenotypic explanation implies that in 

 areas of low productivity elevated meristic counts 

 result from retardation of growth and that this 

 retardation is the result of the average survivor 

 being underfed compared to larvae in areas of 

 higher productivity. As year class strength in 

 pelagic fish populations is probably largely deter- 

 mined in early stages of larval life and not by the 

 total number of eggs produced or mortality during 



Table \0.— Vinciguerria nimharia, comparison of gill raker 

 counts with vertebral and longitudinal photophore row counts. 



v. nimbaria, total gill rakers on first gill arch. 



Area 17 18 19 20 21 22 23 24 25 26 n Mean±95% limits 



GG ______ 3 6 10 1 20 24.45 ± 0.386 



SOS 4 7 3i______i5 18.07 ±0.489 



CEP - _ _ 4 13 13 1 - - - 31 21.35 ±0.277 



PS - 2 26 54 31 2 - - - - 115 20.04 ± 0.147 



CNA -4 8---_--_12 18.67 ±0.313 



CNP -- 3 6 3----- 12 20.00 ± 0.469 



V. nimbaria and Diplophos taenia, comparison of counts, given 

 as mean (rank. 



Vinciguerria nimbaria Diplophos taenia 



Area Gill rakers IPVALA Vertebrae IPVALA 



SCS 18.1 (1 66.5 (1 39.4 (1 143.6 (1 



CEP 21.4(4 68.0(2 40.2(2 145.5(2 



PS 20.0(2.5 69.6(3 41.2(3 149.2(3 



CNP 20.0(2.5 70.3(4 41.8(4 151.7(4 



advanced prerecruit stages (Hempel 1965), it 

 seems likely that selection would strongly favor 

 any mechanisms that tended to protect the larvae 

 of mid-water fishes occurring in areas of low 

 productivity against starvation. The possible 

 materials on which this selection might operate 

 and the possible consequences on meristic 

 characters form the basis for a second explanation 

 of the observed meristic variation, that it is the 

 by-product of genetic differences between 

 separable populations in areas of low and high 

 productivity. 



Hempel (1965), Blaxter (1965), and others, con- 

 cerned mainly with pelagic clupeoid fishes, have 

 developed strong evidence that under normal cir- 

 cumstances the main restriction on the success of a 

 year class occurs within a short period of larval 

 life, the critical period of Hjort (1914, 1926) and 

 others (e.g. Marr 1956; Schumann 1965). Selection 

 has apparently resulted in adaptive mechanisms 

 tending to balance the two main dangers to larval 

 survival: the danger of starvation and the danger 

 of predation (Blaxter and Hempel 1963; Hempel 

 1965). 



Blaxter (1965), Hunter (1972), and others, have 

 shown that at the onset of feeding, just before or 

 at the time of yolk-sac absorption, surprisingly 

 small differences in size can significantly affect 

 the probability of larval survival. Hunter (1972) 

 has shown for northern anchovy Engraulis mor- 

 dax Girard, larvae that slight increments in size 

 are associated with highly increased searching 

 abilities, highly increased success of attempted 

 feeding acts, and vastly diminished minimum prey 

 density requirements for survival. Blaxter (1965) 

 discusses the significance of the greater spectrum 

 of particle sizes available to larger larvae in terms 

 of increased diversity of available prey organisms. 

 Similar findings have been reported for other fish 

 larvae (e.g. Arthur 1956; Einsele 1965). Size at 

 hatching, at least for Atlantic herring, Clupea 

 harengus Linnaeus, is a direct function of egg 

 size; larger larvae hatch from larger eggs. 

 Fecundity is inversely proportional to average egg 

 size (Baxter 1959; Blaxter and Hempel 1963; 

 Blaxter 1969). 



We believe that the meristic variation between 

 populations occurring in areas differing in 

 productivity values is the result of adaptations in- 

 volving the adjustment of egg and larval size to 

 the productivity regime. We believe that these 

 adaptations reflect differences between areas of 

 low and high productivity in the relative impor- 



294 



