Ii0° -iO" 140° 150° 



FISHERY BULLETIN: VOL. 73, NO. 2 



PS 



scs 



PS 



scs 



PS 



scs 



-16 



-10 



-10 



-11 



PS 



scs 



-1 D taeni a 



P maul 



257 

 ■24 



V nimbaria 



6&46I 



IPVALA 



_L 



J , L 



J L I 1 L 



J_ 



J_ 



J I L 



-3 -2 -1 



-13 



-204 



7 8 9 



IC 



10 



64.63 



Figure 4.-Left: Antipodes Expedition station positions. Right: Comparison of IPVALA and IC photophore counts between specimens 

 from South China Sea (SCS) and Philippine Sea (PS). Data presented as in Figure 2 except that mean values for South China Sea 

 material of all four species have been set equal to zero and all other statistics are plotted as deviations from this zero point. Data for 

 Chauliodus sloani from SCS are from Ege (1948); the rest is original data. 



second prediction. In Table 9 mean IPVALA 

 counts for Philippine Sea specimens of V. nim- 

 baria are tested for relationship with distance of 

 site of collection from land (Japan, Ryukyu 

 Islands, Luzon, but not Bonin or Volcano islands). 

 While the highest mean counts were found in 

 specimens from the 4 stations most distant from 

 land, the data show no relationship between mean 

 counts and distance from land (tau^ = -0.273, 

 P >0.20). Mean IPVALA values for specimens 

 from each of the 11 pairs of Philippine Sea stations 

 are significantly higher than the mean IPVALA 

 values for specimens from the South China Sea. 



DISCUSSION 



One fact and two assumptions are prerequisite 

 to our discussion of the possible explanations for 

 the relationship between meristic counts and 

 measures of food supply. The fact: in Diplophos 

 taenia, Vinciguerria lucetia, and V. nimbaria, the 

 values of the meristic characters we have studied 

 are fixed at or before metamorphosis (Ahlstrom, 

 pers. commun., Ahlstrom and Counts 1958, Silas 

 and George 1971). This is probably also true for 

 Pollichthys mauli and Chauliodus sloani. This 



means that any explanation involves factors 

 operating on eggs and /or larvae. The assumptions: 

 1) that the meristic variation observed is not the 

 result of selection for certain absolute values of 

 the meristic counts, and 2) that the same basic 

 mechanism underlies the variation in counts for all 

 five species in the area studied (in this discussion 

 we ignore results for specimens of V. nimbaria 

 from the Gulf of Guinea). 



There are four possibilities: 1) that the observed 

 variation is ecophenotypic, i.e. nongenetic 

 modification of the phenotype resulting from the 

 effects of differing food availability conditions 

 upon early growth and development of meristic 

 characters; 2) that the observed variation is a by- 

 product and indicative of genetic differences 

 between populations in these eight areas, and that 

 these differences reflect differing selective pres- 

 sures resulting from differing conditions for early 

 growth; 3) that the observed variation is a com- 

 bination of ecophenotypy and genetic differences; 

 and 4) that the real explanation is none of these, 

 that a causal relationship between meristic 

 characters and productivity does not exist, and 

 that we have overlooked the real meaning of our 

 results. We are unable to deal with the third pos- 



292 



