FISHERY BULLETIN: VOL. 76, NO. 4 



questionable, because of differences in experimen- 

 tal techniques." Preexperimental feeding regimes 

 differ significantly between studies. Cherry et al. 

 (1975) discontinued feeding 2.5 to 3 days before 

 testing; in contrast, Reynolds and Thomson ( 1974) 

 fed fish approximately 2 h prior to experimenta- 

 tion. The time that the fish were allowed to adjust 

 to the experimental chamber before initiation of 

 data collection was also variable. Tat'yankin 

 (1972) used a habituation period of only 0.5 h, 

 while McCauley and Read ( 1973) used a period of 2 

 days. Length of time between observations was 

 not consistent between studies. Ferguson (1958) 

 observed fish hourly, but Norris (1963) recorded 

 fish position every 6 s. Although most investiga- 

 tions used organisms once, Javaid and Anderson 

 (1967) reused fish between successive experi- 

 ments. Some of the differences in experimental 

 techniques may have been due to species-specific 

 problems; however, in many studies the primary 

 concern was simply to provide a thermal gradient, 

 to place test organisms in it, and to measure their 

 response. Often, little attention was paid to factors 

 relevant to the well-being of the fish (e.g., shock 

 from capture and handling, nutritive condition) 

 and to simulating natural intensities and quality 

 of light. 



General procedural recommendations have 

 been made by Richards et al. ( 1977). Our methods 

 build on techniques from past studies, including 

 their best attributes. Some of the differences in 

 experimental techniques are subtle, but impor- 

 tant, especially since this investigation concerns 

 marine species, on which little work has been 

 done. The methods, procedures, and physical con- 

 ditions of collection, handling, and experimenta- 

 tion in this study were chosen to minimize trauma 

 to test organisms. The following methods and 

 rationales that we employed are provided as 

 suggestions toward standardization: 



1) Our preferred methods of obtaining test or- 

 ganisms involved collecting by means of lift 

 nets or traps or to rear them from eggs. These 

 minimized damage to the fishes. 



2) Fish feeding was established in the labora- 

 tory. We tested only fish that fed, as this indi- 

 cated that they were most likely not in shock 

 from collection. 



3) We brought fish to desired acclimation tem- 

 peratures, from that at which they were col- 

 lected, at a rate of l°C/day. The test specimens 

 were held for at least 2 wk prior to testing. The 



length of the holding period was particularly 

 important to assure acclimation to cold tem- 

 peratures (Brett 1970). 



4) We did not use anaesthetics in collections or 

 preexperimental handling; Goddard et al. 

 (1974) have shown that MS-2224 can 

 influence thermal behavioral responses for 

 several weeks after treatment. 



5) We fed the fish ad libitum just prior to their 

 placement in the gradient. This standardized 

 the feeding history, which has been shown to 

 alter temperature selection (Ivlev and Leize- 

 rovich 1960; Javaid and Anderson 1967). 



6) We placed the fish in the experimental 

 chamber, adjusted to their acclimation tem- 

 perature, on the evening preceding testing. 

 This allowed adjustment to the new sur- 

 roundings prior to experimentation. No fish 

 were reused. 



7) Testing fish in groups allowed us to assess the 

 effects of temperature on populations, but we 

 could only use this methodology to study gre- 

 garious species that did not display agonistic 

 behavior. 



8) We established the temperature gradient 

 about the fish. Introduction of test organisms 

 to an established gradient, even at the loca- 

 tion of their acclimation temperature, can re- 

 sult in the fish darting to another part of the 

 tank and experiencing a temperature shock. 



9) We shifted isotherms during an experiment, 

 and the hot and cold ends were reversed be- 

 tween experiments. This allowed us to at- 

 tempt to partition any tendencies of the fish to 

 select a particular compartment independent 

 of temperature. 



10) We shielded the test chamber from external 

 light and observed the fish from above down- 

 ward-directed experimental lights to insure 

 that the fish would not respond to presence of 

 the experimenter. 



11) We used low levels of lighting, during the 

 experiment, based on minimum intensities 

 for schooling and larval feeding (Blaxter 

 1970). These levels of illumination did not 

 appear to disturb the fish as was sometimes 

 the case with brighter light. Additionally, 

 Sullivan and Fisher ( 1954) reported increased 

 precision of temperature selection at low light 



''Reference to trade names does not imply endorsement by the 

 National Marine Fisheries Service, NOAA, or by Occidental 

 College. 



838 



