Macy et al,: Metabolic rate of Brevoortia tyrannus 



283 



weight/d. The acclimation tanks were cleaned daily. 

 During the holding and experimental periods, a 12 

 h:12 h lightidark photoperiod was provided by over- 

 head fluorescent lights, and sufficient current was 

 provided to cause the fish to orient head into the cur- 

 rent. Experiments took place 3-14 December 1989. 

 Vital statistics for the fish are given in the top sec- 

 tion of Table 1. 



Respiration measurements were made on a small 

 school of menhaden because studies of schooling 

 fishes have shown that respiration rates are higher 

 and growth rates lower in isolated fish than in those 

 kept in groups (Skazhina, 1975; Kanda and Itazawa, 

 1978; see also Ross and Backman, 1992). Experi- 

 ments were begun with the 15°C trials on 3 and 4 

 Dec, followed by the 10°C trials on 7 and 8 Dec. The 

 20°C trials were conducted on 14 Dec. Thirteen fish 

 were used for the 10°C and 15°C experiments. For 

 the 20°C trials only 10 fish were used. Digestion is 

 rapid in menhaden (Durbin and Durbin, 1981), and 

 metabolism quickly returns to preceding levels after 

 a meal (Durbin et al., 1981). Thus a starvation pe- 

 riod of 24 h was considered sufficient to remove the 

 effect of the previous meal on metabolic-rate mea- 

 surements. Handling and transfer protocols were 

 similarly chosen to reduce bias due to stress-induced 

 factors, such as those identified by Waring et al. 

 (1992) in flounder and salmon. 



A sealed, toroidal fiberglass flume tank (Fig. lA), 

 modified from Hettler ( 1976), was used as a respirom- 

 eter. The flume (2.75 m inside diameter, 0.51 m depth, 

 and 0.39 m wide annular swimming channel) was 



large enough to hold a small school of adult menha- 

 den. The flume was painted with a white, nontoxic, 

 polyester gel coat. Black radial stripes 5.1 cm wide, 

 spaced 21.8 cm apart (at the inside diameter of the 

 flume), were painted on the walls and bottom of the 

 flume, to provide reference marks for later determi- 

 nation of swimming speeds and to provide visual cues 

 for the fish to sense displacement by the flow. To pro- 

 duce the desired flow rates, a high-pressure, 3.7-kW, 

 multi-stage electric pump (Fig. lA) was connected 

 to coils of polyethylene tubing (6 mm inside diam- 

 eter) arrayed along the walls of the flume (Fig. IB). 

 Discharge holes were punched at an angle of about 

 45° downstream and 20° downward from the hori- 

 zontal with a hypodermic needle along the length of 

 tubing to form small jets creating a counterclockwise 

 flow in the flume. A valve placed at the pump outlet 

 (Fig. lA) controlled the speed of water in the flume. 

 Return water to the pump flowed through a drain 

 line located at the bottom of the tank and thence 

 through a bank of 4 fiber-wound cartridge filters (3 

 |am). An integral reservoir (12. 9-L capacity) was used 

 to replenish sample water removed during experi- 

 ments. Cooling coils were installed along the inner 

 wall of the flume and around the pump housing (Fig. 

 IB) to remove heat generated by the pump and to 

 maintain a constant temperature during trials. Wa- 

 ter temperature was monitored with a YSI Model 

 2100 Tele-thermometer. 



Flow rates within the flume were measured with 

 a General Oceanics Inc. flowmeter connected to a 

 proprietary digital display. The transducer head was 



