Macy et al.: Metabolic rate of Brevoortta tyrannus 



285 



Plexiglas lid, composed of interlocking sections, 

 was then carefully fitted over the surface of the 

 tank to provide an airtight seal. During respi- 

 ration trials, the flume was operated as a closed 

 system; between trials and at night, however, 

 the flume received a continuous inflow of tem- 

 perature-controlled filtered seawater, with a 

 flow rate of about 6 cm/s to orient the fish. 



During each 2-d experiment a series of trials 

 was carried out at flume speeds ranging from 

 12-13 (minimum) to 60-63 (maximum) cm/s. 

 Between four and six trials were conducted per 

 day, with each trial lasting 1 h. Test speeds were 

 selected randomly, and at the completion of a 

 trial the next test speed was set, followed by a 

 30-min interval during which flow rate became 

 stable again and fish were able to adjust to the 

 new speed. An interval of about 1 h was used to 

 reoxygenate the flume when necessary. Test 

 current speeds were set by regulating the dis- 

 charge pressure of the circulating pump (Fig. 

 lA) by using the predetermined pressure to 

 flow-rate calibration curve. Although we at- 

 tempted to use the same test current speeds at 

 all three temperatures, minor inaccuracies in the 

 calibration curve and gauge readings resulted in 

 only approximately equal test speeds (Fig. 2). 



Fish behavior was observed continuously 

 during experiments and recorded with a mono- 

 chrome MTI Dage M65 video camera suspended 

 directly above the center of the flume and con- 

 nected to a VHS format video recorder. Swim- 

 ming speed in relation to the bottom (ground 

 speed, GS) was determined from the video re- 

 cordings at 3-min intervals, by marking the 

 position of each fish on the video monitor screen 

 with a grease pencil and noting its subsequent 

 location 10 s later. The number of reference 

 marks crossed by each fish during this interval 

 was then multiplied by the distance between 

 marks (21.8 cm) to obtain ground speed. Fish 

 moving upstream were assigned positive values, 

 whereas those swimming downstream were scored 

 negatively. Actual swimming speed ( SS ) through the 

 water was then computed as the algebraic sum of 

 the ground speed (speed offish in relation to the bot- 

 tom) and the fiow rate in the flume. Mean swimming 

 speed during a trial was computed from the indi- 

 vidual swimming speeds, irrespective of direction. 



During experiments a YSI Model 51B dissolved oxy- 

 gen meter was used to monitor oxygen levels in real 

 time; oxygen content of the water was not permitted to 

 decline below 607f of saturation. At 15-min intervals 

 during a trial, duplicate water samples were siphoned 

 fi-om the collection port into 300-mL biological oxygen 



70 -| 



60- 



50- 



40- 



30 



20 



10 



1 

 o 



■d 



10°C 



• SS = 9.49 X F + 8.57. r = 0.93 

 O GS= -0.45 X F + 8.53, r^ = 0.03 



•^^i^^ii— cNr^ 



15°C 



• SS= 9.77xF + 603. r =099 

 O GS = -0.33 X F + 6.65. r = 0.14 



20°C 



• SS= 1.01 xF-^ 2 07. r" = 99 

 O GS = -0 49 X F -^ 6.73, r^ = 0.30 



-| r 



10 20 30 40 50 60 70 



Flume Speed, cm/s 



Figure 2 



Ground speeds (GS) and true swimming speeds (SSl of Atlantic 

 menhaden, Brevoortia tyrannus, swimming against water cur- 

 rents in a toroidal flume respirometer at 10', 15", and 20°C. Lower 

 ground speed error bars are not shown. 



demand (BOD) bottles that were allowed to overflow 

 twice their volume and that were then fixed for deter- 

 mination of oxygen content by Winkler titration ( Carritt 

 and Carpenter, 1966; Strickland and Parsons, 1972). 

 Analyses were performed on 5-6 50 niL subsamples 

 withdrawn fi-om the bottles with a volumetric pipette. 

 Dissolved oxygen declined linearly over time (overall 

 mean r^ ± 1 SD for all regressions=0.987 ±0.018). Meta- 

 bolic rates were therefore assumed to be unaffected by 

 declining oxygen content during the trials. Menhaden 

 respiration rates were computed fi-om the change in 

 dissolved oxygen content over time as determined by 

 least squares linear regression, corrected for the wet 

 weight of the fish and volume of the flume. 



