284 



Fishery Bulletin 97(2), 1999 



2 75m inside 



discharge manifolds 



filter bank 



pressure gauge 



flow adjusting valves 



cooling coil 



pump 

 flow meter transducer 



B 



inlet discharge lines 



0.51m 



0.39m 



flow balancing valves 



Figure 1 



(Al Schematic view of the toroidal flume respirometer. The direction of flow is 

 indicated by arrows within lank. (B) Cross-sectional view of flume tank show- 

 ing arrangement of water and coolant flow. A typical cross section of the gen- 

 erated flow field is also shown. Speeds are cm/s. The grid spacing is 10.8 cm 

 vertically by 8.3 cm horizontally. 



suspended in the flume by a stain- 

 less steel rod (Fig. IB). During experi- 

 ments the flowmeter head was fixed 

 radially in the middle of the flume, 

 near the top of the water column. 

 However, sensor depth and radial 

 position could be varied to measure 

 flow rate in different regions of the 

 flume. By varying the number of dis- 

 charge holes across the vertical ar- 

 ray of inlet coils and by regulating the 

 flow to the inner and outer inlet coils, 

 it was possible to produce a relatively 

 uniform cross-sectional flow field. 

 Typical flow rates (cm/s) determined 

 at maximum pump pressure (1.38 x 

 lO'^kPa ), on a grid 10.8 cm vertically 

 by 8.3 cm horizontally measured ver- 

 tically from the flume bottom, are 

 shown in Figure IB. Preliminary tri- 

 als established that the relationship 

 between pressure of the inflowing 

 water from the pump and the flow 

 rate generated inside the flume was 

 linear up to the maximum pump pres- 

 sure, 1.38 X lO'^ kPa, corresponding 

 to a flume speed of 64 cm/s. Blank 

 trials without fish demonstrated that 

 the system was airtight and that 

 changes in oxygen concentration due 

 to microbial respiration were negligible 

 over periods comparable to those in 

 the experiments. 



Each experiment comprised a se- 

 ries of trials over a 2-d period. Men- 

 haden were anesthetized with MS- 

 222 and transferred to the flume 24 

 h before an experiment to permit re- 

 covery from handling stress (Barton 

 and Schreck, 1987; Waring et al., 

 1992). The menhaden remained in 

 the flume until the conclusion of the 

 experiment and were fed in the late 

 afternoon of each day. In the morn- 

 ing, fecal material and debris was si- 

 phoned off the bottom of the flume, 

 the bottom and interior walls of the 

 flume were scrubbed with a polyeth- 

 ylene scouring pad, and the flume 

 was flushed with filtered seawater. 

 All experiments were conducted dur- 

 ing daylight in the afternoon, 24 h 

 after the last feeding. The fish were 

 accustomed to this procedure and 

 quietly circuited the flume. A clear 



