FISHERY BULLETIN: VOL. 79, NO. 1 



If scotopic visual pigments with spectral posi- 

 tions slightly above 500 nm are optimally located 

 to detect bioluminescence in Californian coastal 

 waters, one can see why such pigments occur there 

 in fishes exposed to nocturnal predators. But ques- 

 tions remain concerning why fishes apparently 

 less threatened are consistent in having scotopic 

 pigments positioned slightly under 500 nm. Why 

 are these pigments not loosely positioned above, as 

 well as below, the optimum location? The answer 

 to this question might lie in ancestral relation- 

 ships. In grouping the Californian fishes according 

 to whether their geographical affinities are tropi- 

 cal or temperate (Table 4), our concern was with 

 current relationships. In fact, all these fishes be- 

 long to groups that stem from tropical origins 

 (Berg 1940). The radiation of acanthopterygian^ 

 fishes from a relatively few ancestral forms early 

 in the Cenozoic (Patterson 1964; Romer 1966) has 

 been related (by Hobson 1974) to the concurrent 

 development of modern coral reef communities 

 (Newell 1971). And we have seen that conditions 

 under which coral reefs flourish, compared with 

 conditions in temperate Californian waters, favor 

 in fishes' scotopic visual pigments that are more 

 sensitive to slightly shorter wavelengths. The 

 mean k^^^ of 492 nm that Munz and McFarland 

 (1973) found in the scotopic pigments of coral-reef 

 fishes, compared with the values around 500 nm 

 that characterize Californian fishes, is consistent 

 with the fact that water around coral reefs is gen- 

 erally clearer and more transparent to blue light 

 than water around Santa Catalina (Figure 2). The 

 extent that k^^ values of scotopic pigments in 

 Californian coastal fishes have shifted toward the 

 green from what may have been ancestral posi- 

 tions near 490 nm, and, especially, have become 

 located slightly above 500 nm (the optimal posi- 

 tion in California), may roughly measure the rela- 

 tive strength of nocturnal or crepuscular preda- 

 tion pressures on each species in these greener 

 waters. 



The argument that scotopic pigments may be 

 positioned to detect bioluminescence can be ex- 

 tended to coral reef fishes. Our calculations show 

 that visual pigments with different Amax values 

 would trap percentages of the light from Noctiluca 

 miliaris as follows: P450 nm = 62%, P475 nm = 



^All fishes considered in this paper have been included among 

 the acanthopterygians, or spiny-finned teleosts, although one, 

 Atherinops a f finis, would be relegated by some systematists (e.g., 

 Greenwood et al. 1966) to another group. 



94%, P490 nm = 100%, P500 nm = 99%, P525 nm 

 = 93.5%, P550 nm = 79%. Although the peak is 

 broad, the central wavelength for maximum ab- 

 sorption at zero range would be near 495 nm (Fig- 

 ure 14), which is close to the k^^^ of 492 nm in the 

 scotopic pigments of coral-reef fishes. And as dis- 

 tance from the source increases the match be- 

 comes even better (Figure 15). 



Twilight or Bioluminescence? 



Whether the clustering and spectral position of 

 visual pigments in warm-temperate and tropical 

 reef fishes is the result of natural selection, or is 

 simply fortuitous, is a complex question. If, as we 

 believe, these features have been refined by in- I 

 tense selection pressures from predators, it re- | 

 mains problematical whether the advantage lies 

 in detecting bioluminescence, or in enhancing 

 photoabsorption during twilight — as suggested 

 by the Twilight Hypothesis. Indeed, both may be 

 important. The scotopic pigments have spectral 

 positions that would be effective in both functions, 

 and the benefits of one would complement the 

 other. Unquestionably, there has been ample time 

 to influence evolution. Bioluminescent plankton 

 have existed since before the first fishes (Seliger 

 1975), and so has twilight's unique spectrum. Be- 

 cause fishes have experienced these features 

 throughout their history, the slightest favorable 

 adjustment could have been adaptive. Although a 

 5-10 nm shift in k^^^ position would improve 

 photoabsorption by no more than a few percentage 

 units, even this could have been meaningful. And 

 if selection pressures to detect bioluminescence 

 have, in fact, acted in concert with those to en- 

 hance crepuscular vision, which to us seems likely, 

 then their combined impact certainly would have 

 been a powerful evolutionary force. 



ACKNOWLEDGMENTS 



We thank Russell Zimmer and his staff at the 

 Catalina Marine Science Center, University of 

 California, especially Robert Given and Larry i 

 Loper, for making facilities available and assist- ' 

 ing us in many ways. For constructive criticism of 

 the manuscript we thank Alfred Ebeling, Univer- 

 sity of California, Santa Barbara; Gene Helfman, 

 University of Georgia; and Richard Rosenblatt, 

 Scripps Institution of Oceanography. Dolores 

 Fussy, Southwest Fisheries Center Tiburon 

 Laboratory, National Marine Fisheries Service 



28 



