TEMPERATURE SELECTION BY YOUNG TOPSMELT 529 



show shari^er avoidance of water near their upper than nesir their 

 lower lethal temperature. This more precise avoidance of undesirably 

 warm water is probably caused by their higher level of activity in 

 warm temperatures. 



The temperatures selected by topsmelt 15 min after the 

 formation of a temperature gradient may provide a more meaningful 

 insight into their thermal preference because of their tendency to 

 become positioned within the gradient. In nine out of ten experi- 

 ments, this initial selected temperature (1ST) was in the range of 19.2 

 to 23.1° C (Table 1) and was independent of acclimation tempera- 

 tures (r = 0.28, n - 10, P > 0.05). Additional evidence that the 1ST is 

 the best indicator of the topsmelt's thermal preference was obtained 

 from the field investigation. 



Atherine larvae were collected in every sampling period from 

 January through October, with the highest densities obtained from 

 March through June, but the fish were not identified below the 

 family level. It appears, however, that this group of specimens 

 consisted of topsmelt and grunion (Leuresthes tenuis) Icirvae and 

 possibly a few individuals of the third atherine species, jacksmelt 

 (Atherinopsis calif or niensis). Topsmelt and grunion have been found 

 to occur together during their larval and juvenile stages (Ehrlich and 

 McGowen, unpublished data). These determinations were based on 

 the field collection and subsequent rearing of larvae to a size when 

 they could be speciated. The temperature-specific frequency distribu- 

 tions of the larval atherines are shown in Table 2. Temperature 

 occurrences for the fish density calculations were considered only for 

 samples containing atherines. Although the range of 22 to 23° C 

 occurred only 15.4% of the time, 40.3% of the fish were collected at 

 these temperatures (Table 2). Similarly, 19 to 23°C occurred 40.4% 

 of the time, but 59.3% of the total fish population sampled was 

 found at these temperatures. Furthermore, temperatures 27° C and 

 higher accounted for 23.1% of all occurrences, but only 9% of the 

 larvae were found in this range. This indicates a strong avoidance 

 response of these temperatures. Ehrlich (1977) demonstrated that 

 the laboratory-determined preferred- temperature range of juvenile 

 grunion was 22 to 23° C; this was also the temperature range that 

 maximized larval yolk utilization efficiency. The secondary mode at 

 16° C (Table 2) may indicate that larvae were inadequately feeding; 

 Ehrlich (1977) showed that even 1 day of food deprivation results in 

 a decrease in the preferred temperatures of larval grunion and many 

 other species. Thus the laboratory and field data corroborate each 

 other and demonstrate that topsmelt have a eury thermal preference; 

 they showed little discrimination in the range from 19 to 23° C, and 



