FISHERY BULLETIN: VOL 76, NO, 4 



normal frequency distribution (Table 1), but sub- 

 sequent food deprivation resulted in selection of 

 colder water (mean 18.7°Cl and a narrower dis- 

 tribution ( SD 3.3°C ). Selection of colder water dur- 

 ing food deprivation may be a mechanism of 

 energy conservation by poikilothermic organisms. 

 Other mechanisms of energy conservation, both 

 behavioral and biochemical, have been found dur- 

 ing starvation of herring and plaice larvae (Blax- 

 terandEhrlich 1974;Ehrlich 1974). The increased 

 precision of temperature selection during food de- 

 privation coincides with the findings of Reynolds 

 (1977) that stress can increase precision of tem- 

 perature selection. This pattern, however, of low- 

 ered selected temperature but increased precision 

 during food deprivation was not duplicated by all 

 species. 



Comparison of data from 2 consecutive days of 

 observation of shiner surfperch demonstrated that 

 the temperatures they selected, as well as their 

 tendency to school or aggregate, varied between 

 the 2 days. The selected temperatures decreased 

 on the second day (means 19.9° to 15.6°C and 

 modes 19° to 16°C), but unlike California grunion 

 the standard deviation of the mean increased (2.1° 

 to 2.6°C). School tightness also decreased as indi- 

 cated by a larger coefficient of dispersion (24 to 

 3T/( ). The major preexperimental difference be- 

 tween the first and second day runs was that the 

 fish were not fed before the second day of ex- 

 perimentation. The lowering of the preferred 

 temperature during food deprivation followed that 

 of the California grunion and brook trout (Javaid 

 and Anderson 1967). Furthermore, this selection 

 of cooler water during food deprivation agrees 

 with the findings of Brett et al. (1969) and Brett 

 and Higgs ( 1970) who showed that a limited food 

 supply resulted in a decrease in the optimal tem- 

 perature for growth, which is coincident with the 

 final preferendum for sockeye salmon (Brett 

 1971 ). The wider preferred temperature range and 

 reduced aggregation may reflect increased search- 

 ing by hungry fish. Hunger can cause fish to in- 

 crease the distance between individuals (Hunter 

 1965) as well as disrupt fish school integrity (Blax- 

 ter and Holliday 1958). 



Larval California grunion, topsmelt, rockpool 

 blennies, and painted greenlings selected water 

 that was often warmer than naturally available. 

 This behavior has several adaptive advantages 

 but also presents potential harmful consequences 

 resulting from man's alteration of their natural 

 habitat. Selection of warm water by larvae will 



reduce the duration of the highly vulnerable 

 yolk-sac and larval stages. In King Harbor it kept 

 many of these larvae in the back basins that con- 

 tained the largest concentration of food organisms 

 (McGowen'^). Furthermore, in genera such as 

 Hypsoblennliis that become demersal following 

 the planktonic larval stages, selection of warm 

 water will help the larvae to remain in the near- 

 shore environment where they must be when they 

 leave the planktonic community. Marshall ( 1966) 

 discussed similar mechanisms by which near- 

 shore demersal species maintain their population 

 integrity. Rockpool blenny larvae, however, that 

 entered water warmer than 28°C lost their 

 equilibrium, could not extricate themselves and 

 eventually died. This same behavioral response of 

 entering water above a lethal temperature and not 

 leaving it has been reported for other species (e.g., 

 Beitinger and Magnuson 1976). Similar behavior 

 in the field near sources of hot water such as power 

 plant effluents may significantly affect local popu- 

 lations. This could be magnified as a result of pre- 

 dation, for it has been shown that the vulnerabil- 

 ity of fishes to predation is increased by sublethal 

 heat shocks (Sylvester 1972; Coutant 1973; Yocum 

 and Edsall 1974). 



Various workers (Lowe and Heath 1969; Reyn- 

 olds and Thomson 1974; Reynolds and Casterlin 

 1976; Beitinger 1977) have reported that the final 

 preferendum is often close to the upper lethal 

 temperature. Frequency distributions skewed to 

 the left such as for topsmelt (Table 1, Figure 7) are 

 reflective of this behavior. The mode occurrence 

 for topsmelt was 26°C, and they sharply avoided 

 warmer water. Coutant (1975) pointed out that 

 the upper avoidance temperature is more sharply 

 defined than the lower one for most species. 

 Beitinger (1977) found bluegill, Lepomis mac- 

 rochin/s, tolerated less variation in avoidance 

 temperature near their lethal limits. 



Differences in group cohesion for individual 

 runs between nonschooling species such as adult 

 speckled sanddabs (Figure 5) and tightly aggre- 

 gating ones such as olive rockfish (Figure 6) are 

 illustrated by comparison of the standard devia- 

 tions about the run selected temperatures. 

 Although a low coefficient of dispersion may indi- 



^McGowen, G. E. 1977. Effects of thermal effluent from 

 Southern California Edison's Redondo Beach Steam Generating 

 Plant on the warm temperate fish fauna of King Harbor marina. 

 Ichthyoplankton study report for Phase II. Annual Report for 1 

 March 1975-29 February 1976. Unpubl. manuscr., 46 

 p. Southern California Edison Res. Contract No. U0654902. 



846 



