HOBSON ET AL.; CREPUSCULAR AND NOCTURNAL ACTIVITIES OF CALIFORNIA FISHES 



tends to be diffused over a greater range of visual 

 circumstances. Bioluminescence is more constant: 

 the spectral compositions of moonlight and star- 

 light change with water depth and atmospheric 

 conditions, but the spectral composition of 

 bioluminescence is independent of these vari- 

 ables. And, of course, there is less light from moon 

 or stars with increased water depth, which, again, 

 is untrue of a bioluminescent emission. So it is not 

 surprising that the narrow range of Xrnax positions 

 in visual pigments of Californian fishes more 

 closely matches bioluminescence than it does 

 moonlight or starlight (Figure 16). 



Undoubtedly moonlight and starlight have 

 strong influences on nocturnal relationships 



MOONLIGHT 



BEST x„„ 



400 



500 600 



WAVELENGTH (nm) 



70 



Figure 16. — Relationships between the spectral distributions of 

 moonlight, starlight, and bioluminescence (as produced by Noc- 

 tiluca miliaris) in seawater typical of southern California and 

 the spectral sensitivities of fishes that live there. The five curves 

 for each type of light represent the spectral distribution expected 

 to reach a fish at the distance (in meters) indicated by the 

 number that accompemies each curve. The curves depict moon- 

 light and starlight off a flat reflector at a depth of 3 m (using 

 values from Munz and McFarland 1977), and the luminescence of 

 N. miliaris (using values from Nicol 1958), in water equivalent 

 to typical conditions at Santa Catalina Island (Coastal Type 1 of 

 Jerlov 1968). The solid circle on each spectral curve identifies the 

 wavelength to best match with a visual pigment for maximum 

 photosensitivity, and the stippled column represents the spectral 

 range of maximum photosensitivity in scotopic pigments of 

 Californian fishes tXmax^- Note that these coincide only with 

 bioluminescence. 



among predators and prey. But probably both help 

 fishes more on offense than on defense. Predators 

 can position themselves, and time their attacks, to 

 play both types of light to their advantage, and to 

 the prey's disadvantage. By charging at prey from 

 below, for example, predators view their targets 

 against the water's relatively light surface, while 

 their own movements are masked by the sur- 

 rounding gloom {Hobson 1966). Certainly attacks 

 that so often spring from the shadows would 

 greatly dilute a defensive advantage prey might 

 gain with spectral sensitivities that match moon- 

 light or starlight. Under these circumstances prey 

 face a broad range of threats that calls for a more 

 generalized response. So we can understand why 

 many smaller reef fishes that habitually range 

 into the water column at night stay closer to shel- 

 ter under moonlight (Hobson 1968a). 



Despite the offensive advantage that certain 

 predators likely gain from moonlight (or star- 

 light), their scotopic visual pigments tend to be 

 better matched to bioluminescence (or twilight), 

 probably because this answers a more pressing 

 need on defense. So it would seem that even those 

 species that have special tactics to use moonlight 

 or starlight to better see their prey must com- 

 promise with visual pigments less than optimal 

 for this task. Included are those nocturnal plank- 

 tivores, like subadult Sebastes serranoides, that 

 characteristically hover tail-down in the water 

 column, where apparently prey are visible to them 

 against moonlight or starlight from above. In- 

 cluded, too, are those predominantly diurnal 

 fishes, like Paralabrax clathratus and Cymatogas- 

 ter aggregata , that apparently are able to hunt at 

 night close to sand where light levels are elevated 

 by reflected moonlight and starlight. We suggest 

 not that their visual pigments are unsuited to see 

 prey by moonlight or starlight, but rather that 

 these pigments simply could have better spectral 

 sensitivities for this particular job (Figure 16). 



These arguments, favoring bioluminescence 

 over moonlight and starlight as a selective force in 

 determining Xj^ax position, would also favor 

 bioluminescence over twilight. But there is an im- 

 portant difference in this last comparison. Moon- 

 light and starlight would select for spectral posi- 

 tions different from that selected for by 

 bioluminescence, and so a conflict would exist. 

 Twilight, on the other hand, would select for essen- 

 tially the same spectral position as biolumines- 

 cence, so that the two would act in concert (see 

 below). 



27 



