FISHERY BULLETIN: VOL. 76, NO 2 



thoroughly mixed with seawater of 36%o salinity 

 until a desired salinity was obtained in the tank. 

 Frequently, schools of fish in the field were not 

 caught in water of identical salinity to that used 

 experimentally. Before placing these fish in the 

 experimental tank, freshwater slowly ran into the 

 container of water in which the fish were trans- 

 ported to the laboratory. When the desired lower 

 salinity was reached by dilution and overflow of 

 the water in the container (this required about 60 

 min to accomplish), the fish were transferred by 

 dip net to the experimental tank. The experimen- 

 tal tank had a water temperature within 1° or 2°C 

 of the water in which the fish were caught and 

 transported or in which dilution occurred. 



Within 30-60 min after being placed in the tank, 

 schools and individual mullet swam throughout 

 the tank at relatively uniform speeds. Single and 

 grouped individuals "grazed" along the sides and 

 bottom of the screen in the tank. The behavior of 

 the mullet at this time appeared to be similar to 

 the behavior of undisturbed fish observed in the 

 field. 



The water in the tank was cooled and heated 

 simultaneously after the fish demonstrated what 

 appeared to be "normal" schooling and grazing 

 behavior. At half-hour intervals the temperature 

 at various levels was recorded and the behavior 

 and depth and temperature distribution of the 

 mullet were noted. The observations required 

 about 1 or 2 min to complete. Observations con- 

 tinued until the vertical distribution of the mullet 

 did not change with respect to specific water tem- 

 peratures during two or three consecutive observa- 

 tions. This occurred between 4 and 8 h after com- 

 mencing the experiments, when the maximum 

 water temperature in the tank was between 36.0° 

 and 40.8°C, and the minimum temperature was 

 between 13.0° and 19. 0°C. Upon termination of the 

 observations, the heating and chilling equipment 

 was turned off, water samples were collected, and 

 one of the sides of the enclosure around the tank 

 was slowly removed. The lights were dimmed 

 slowly and turned off by a rheostat. The fish were 

 left in the tank overnight, exposed to natural 

 twilight conditions in the evening and morning, 

 the room having a number of large windows. Dur- 

 ing the evening and overnight periods the tank 

 temperature gradually became more uniform. 



Well before a second series of observations was 

 to be made during the next day, the side of the 

 tank's enclosure was replaced, the lights having 

 been turned on during the twilight period by a 



302 



rheostat. The tank was also oxygenated for ap- 

 proximately 30 min, followed by at least an addi- 

 tional 30-min period during which the tank was 

 not disturbed. The second day of observations 

 served to check on the experimental procedures 

 and results obtained the first day, and to achieve 

 higher tank temperatures than reached during 

 the first day. Statistical comparisons between first 

 and second day activity were not significant 

 (P&0.05), indicating that the fish did not change 

 their depth or temperature distributions. 



RESULTS 



Experimental 



Prior to the onset of heating and chilling, fish 

 generally <50 mm SL (Table 1), with one excep- 

 tion, initially concentrated near the surface or in 

 the upper half (40-80 cm) of the tank. Fish in the 

 30- to 50-mm SL size ranges in 0%o salinity were 

 initially concentrated in the lower half ( 0-40 cm) of 

 the tank. Fish generally ^50 mm SL initially con- 

 centrated near or on the bottom, or in the lower 

 half of the tank. However, fish of all sizes continu- 

 ously moved throughout the 80-cm deep tank, in- 

 dicating that the light and pressure gradients and 

 the depth of water were not limiting. When left for 

 hours in the tank at constant temperature condi- 

 tions at each test salinity, the mullet exhibited the 

 same behavior, distribution patterns, and move- 

 ments as those observed in mullet before heating 

 and chilling were initiated during the experi- 

 ments. These observations served as a control for 

 the distributional patterns of mullet observed 

 under test conditions. 



Experimental results were grouped and ana- 

 lyzed according to test salinities (0, 15, and 34%o) 

 and mullet size ranges (20-30, 30-50, and 2^50 mm 

 SL) (Table 1). There was overlap and occasional 

 fish outside of specific size ranges in some experi- 

 ments due to the size composition of the individual 

 mullet caught in schools in the field and sub- 

 sequently put into the experimental tank. The 

 experimental time period was divided for analysis 

 into five observation intervals, each consisting of 

 three observations (i.e., 1 complete hour of obser- 

 vation with observation being made each half 

 hour). Only data for observation intervals 1,3, and 

 5 are presented in Figure 3. Each graph consists of 

 open histograms for the temperature and depth 

 distributions for fish of specified size ranges, test 

 salinity, and observation interval. Observations of 



