JOHNSON and BARNETT: CORRELATION BETWEEN CHARACTERS AND FOOD SUPPLY 



Table 8.— IC photophore variation in three species of Chaxdiodus. (NIO, northern Indian 

 Ocean, TAA, TAB, areas of eastern tropical Atlantic discussed in text). 



Species Area 58 59 60 61 62 63 64 65 66 67 68 69 n Mean ± 95% limits 



C.pammelas^ NIO 1 4 12 4 —---____ 21 59.90 ± 0.350 



C. schmidtn TAA -- 17 32 14 ------ 54 62.09 it 0.186 



C.schmidtP TAB ---3 16 14 7 5----45 62.89 ± 0.335 



C.sloann SOS - - - - 1 16 69 91 25 2 - - 204 64.32 ±0.117 



C.sloani PS _______ 3244- 13 66.69 ±0.714 



C.sloann GNA -------- 363 1 13 67.15 ± 0.543 



C.sloani CNR --- — - — _ 2 97 2 — 20 66.45 ± 0.386 



iQibbs and Hurwitz 1967. 

 2Blache 1964. 

 3Ege 1948. 



China Sea. This agrees with results for other 

 species discussed in this paper. 



The only character diagnostically separating C. 

 schmidti from C. sloani is the lower number of 

 serial photophores in C. schmidti (Morrow 1961; 

 Blache 1964). Similarly Gibbs and Hurwitz (1967) 

 concluded that the only characters separating C. 

 pammelas from C. sloani were lower meristic 

 counts (IC, VAV, vertebrae) in C. pammelas and 

 greater development of the gill filaments in C. 

 pammelas, with filaments both longer and with a 

 greater number of lamellae per side. Gibbs and 

 Hurwitz (1967) noted that the greater gill filament 

 development of C. pammelas is correlated with a 

 well-marked oxygen minimum layer in the 

 northern Indian Ocean habitat of this species. Gill 

 filament length may vary intraspecifically in some 

 wide-ranging mid-water fish species (Johnson 

 1974). 



Both C. schmidti, inhabiting the eastern 

 tropical Atlantic, and C. pammelas, inhabiting the 

 northern Indian Ocean, are limited to areas of 

 high biological productivity (Ryther and Menzel 

 1965; Gibbs and Hurwitz 1967; Gushing 1971). Both 

 are distinguished from C. sloani by lower counts of 

 serial photophores (and vertebrae in C. pam- 

 melas), and essentially only by these lower counts. 

 In both cases the lower counts apparently agree 

 with our hypothesized relationship between 

 meristic counts and productivity. The counts for C. 

 schmidti are from specimens taken in two areas: 

 TAA, along the west African coast from lat. 03°56' 

 to 18°22'N, to the west and north of our Gulf of 

 Guinea study area; and TAB, along the west 

 African coast from lat. 01°20' to 17°53'S, to the 

 south of our Gulf of Guinea study area (Ege 1948; 

 Blache 1964). The counts for C pammelas are from 

 specimens taken between lat. 08° and 14°N, long. 

 58° to 66°E, in the Arabian Sea (Gibbs and Hur- 

 witz 1967). 



In view of other results presented in this paper, 



particularly those for Diplophos taenia, we sug- 

 gest that a reexamination of the status of both C. 

 pammelas and C. schmidti, with additional study 

 of meristic variation in C. sloani throughout the 

 range of this species, are in order. 



Results of the Antipodes Transect 



An essentially experimental opportunity to test 

 the hypothesized relationship between meristic 

 counts and productivity was afforded by fishes 

 taken by the Antipodes Expedition of the Scripps 

 Institution of Oceanography in 1970. On this ex- 

 pedition 22 mid-water trawl collections were taken 

 in the Philippine Sea and six mid-water trawl 

 collections were taken in the South China Sea 

 (Figure 4). Because of the 2,000 m or more sill 

 depth separating these geographically contiguous 

 areas, the upper water mass in both areas is the 

 same and differences in physical parameters are 

 minimal. Although the South China Sea is poorly 

 known, there is little doubt that at least nearshore 

 areas or areas over shelves of the South China Sea 

 are substantially more productive than offshore 

 areas in the Philippine Sea (Wyrtki, 1961; Sorokin 

 1973). 



We predicted: 1) that values of meristic counts 

 for species occurring in both the South China Sea 

 and Philippine Sea would be lower in specimens 

 from the South China Sea, and 2) that values of 

 meristic counts for species occurring in the 

 Philippine Sea would be lower in specimens taken 

 near land and increase with increasing distance 

 from shore. 



In all four cases thus far examined (Figure 4) 

 where differences exist in values of meristic 

 counts from the two areas, the counts are sig- 

 nificantly lower in specimens from the South 

 China Sea. This supports our first prediction. 



Vinciguerria nimbaria was the only species 

 taken in sufficient abundance to allow a test of our 



291 



