286 



Fishery Bulletin 97(2), 1999 



Results 



The menhaden formed a loose school in the upper 

 central portion of the flume. The fish usually oriented 

 head into the current and moved slowly upstream 

 (positive ground speed). Occasionally a fish would 

 break away and swim downstream with the flow, but 

 within a few seconds reoriented itself to swim against 

 the flow. These brief periods of downstream swim- 

 ming occurred most frequently at slow current speeds 

 and at the warmest temperature, 20°C. The fre- 

 quency of occurrence of downstream coasting or 

 swimming during respiration trials was as follows: 

 of 11 trials at 10°C: 2 of 11 at 15°C (2.2'7f of obser- 

 vations during those 2 trials); and 6 of 10 at 20°C 

 (6.4% of observations in those 6 trials). When cor- 

 rected for current speed, swimming speeds of fish 

 moving downstream were comparable to those offish 

 moving upstream. Overall, the incidence of downstream 

 swimming did not significantly affect the results. 



In our study, menhaden swimming into a water 

 current maintained a nearly constant ground speed 

 at 10°, 15°, and 20°C (6.8, 5.3, and 4.7 cm/s respec- 

 tively, or about 0.2 BL/s) (Fig. 2), whereas their true 

 swimming speeds ranged from 20.5 to 64.6 cm/s 

 (10°C), 19.2 to 65.7 cm/s (15 C), and 15.5 to 66.0 cnVs 

 (20°C). The mean ground speed was not signifi- 

 cantly different at the three temperatures {x± 

 SD=6.8 ±5.8 cm/s at 10°C; 5.3 ±4.8 cm/s at 15°C; and 

 4.7 ±8.2 cm/s at 20°C). Because ground speed re- 

 mained nearly constant, actual swimming speed in- 

 creased linearly with increasing current speed in the 

 flume (Fig. 2). 



The distribution of individual swimming velocities 

 within the school was unimodal but was skewed to- 

 wards a speed slightly faster than the flow rate in 

 the flume (Fig. 3 1. The coefficient of variation in 

 swimming speed declined curvilinearly as velocity 

 increased, indicating that fish behavior grew less 

 variable at higher speeds (Fig. 4). The pattern of 

 decreasing variability with increasing swimming 

 speed was similar at all three temperatures. 



Respiration rates increased exponentially with 

 increasing swimming speed (Fig. 5, Table 1). Analy- 

 sis of covariance revealed that the relationships be- 

 tween respiration rate and swimming speed differed 

 significantly at the three temperatures. The regres- 

 sion slope, or the rate of increase in oxygen consump- 

 tion per unit swimming speed, was greatest at 10 'C, 

 significantly higher than at 15' or 20°C (Fig. 6). The 

 15° and 20 C curves differed in elevation but not in 

 slope, indicating that the overall level of metabolism 

 was higher at 20°C than at 15 C, but the rate of in- 

 crease in metabolic rate with increasing swimming 

 speed was similar at the two temperatures. The cost 



of swimming was therefore higher at 10°C (metabo- 

 lism increased by a factor of 2.32 per 1 BL increment 

 in swimming speed) than at 15° or 20°C (metabo- 

 lism increased by 1.75 and 1.65, respectively) (Table 

 2 ). At the higher test speeds, metabolic rates at 10°C 

 were equal to or greater than those at 15°and 20° 

 (Fig. 6). 



Standard metabolism, as indicated by they-inter- 

 cepts of the equations given in Table 1, increased with 

 temperature. At 10°, 15°, and 20°C, standard me- 

 tabolism was equal to 0.040, 0.073, and 0.087 mg 

 0./(g wet wt • h) respectively, or 0.131, 0.238, and 

 0.284 mg 02/(g dry wt • h). The Qjg for standard me- 

 tabolism (Prosser, 1973) was higher over the inter- 

 val 10-15°C (3.3) than for 15-20°C ( 1.4). The Qj^ over 

 the interval 10-20°C was intermediate (2.2) (Table 3). 



Discussion 



Previous work has shown that menhaden swimming 

 behavior reflects environmental conditions and food 

 availability. At 20°C, adult menhaden of about 300 g 

 swim at a characteristic speed of about 12.2 cm/s (0.5 

 BL/s) in still water in the absence of food (Durbin et 

 al., 1981). During feeding, swimming speed and 

 metabolic rate increase hyperbolically with increas- 

 ing food concentration (Durbin et al., 1981). Hettler 

 ( 1976) found that routine swimming speed in larval 

 and juvenile menhaden (up to 80 g) increased with 

 temperature but decreased with increasing salinity, 

 starvation, and in the dark. 



Ground speeds in the present experiments were 

 remarkably stable considering the tested tempera- 

 ture range and the different energy expenditures at 

 the different current speeds. The decline in individual 

 variability in swimming speed as mean swimming 

 speed increased (Fig. 4) was very similar to that ob- 

 served by Durbin et al. (1981). This decline in vari- 

 ability may reflect decreased excitability as fish swim 

 faster as well as a transition to steadier swimming 

 with lower energy costs (reviewed in Brett and 

 Groves, 1979; Boisclair and Tang, 1993). 



The nearly constant ground speed in all trials in- 

 dicated that menhaden did not approach their maxi- 

 mum swimming capacity at any of the tested speeds 

 and, as a result, were probably not fatigued at the 

 end of a day's trials despite the lack of rest between 

 trials. The maximum observed metabolic rate offish 

 in our study, 0.356 mgOy(gwet wt • h) at 20°C, while 

 fish were swimming at 63.2 cm/s, was considerably 

 lower than that associated with filter feeding at a 

 slower speed (0.538 mg O.^ /(g wet wt • h) at 20°C and 

 43.4 cm/s: Durbin et al., 1981 ), further indicating that 

 the fish were not performing near their physiologi- 



