that are concentrated below the skin along the 

 lateral midline and become thicker posteriorly 

 (see also Kishinouye 1923, fig. 16; Braekkan 1959, 

 fig. 1). To ensure that the tip of the red muscle 

 thermocouple would remain in place, the wire was 

 passed from near the second dorsal fin obliquely 

 through white muscle and then into the thin red 

 muscle band. Once inserted, its position was easily 

 verified by gentle fingertip probing. 



To facilitate positioning of the two muscle ther- 

 mocouples, 3-4 cm deep holes were tapped with a 

 20-gage hypodermic needle. The heart ther- 

 mocouple was passed into the pericardial cavity 

 through a 17-gage needle that was subsequently 

 withdrawn. All wires were anchored in place by 

 skin sutures. Wire leads ( 1 m long) to the recorder 

 were lap wound together, passed posteriorly, and 

 sutured to the dorsal midline near the finlets to 

 prevent tangling around the tail. Implanting re- 

 quired about 15 min after which the fish was 

 transferred to the respirometer swimming tube 

 where aerated water was circulated over the gills 

 by the driving impeller at a slow speed. 



Two hours recovery from anesthesia and a brief 

 period of swim training was required before a fish 

 could maintain station in the tube and regulate 

 swimming speed in response to water flow. This 

 time delay also allowed stabilization of tissue 

 temperature at ambient conditions following 

 surgery. 



Adaptation to the swimming chamber was car- 

 ried out at a basal swimming speed which is 1.5 

 BL/s (body lengths per second) for S. Japan icus 

 (Magnuson 1973). This speed is alsojustabove the 

 velocity required for sustained ram gill ventila- 

 tion(Roberts 1975). Flow rates in the respirometer 

 were calibrated with a ducted flowmeter (Marine 

 Advisors, Inc. model B-7C ) and controlled by alter- 

 ing the applied armature voltage to the impeller 

 pump motor. Eight fish were used for excess tem- 

 perature measurements and seven were used to 

 monitor EMG (4) and EGG (3) patterns. 



Calibration Procedures 



Thermocouples were made by soldering to- 

 gether the twisted bared tips of the copper and 

 constantan wires and sealing them with epoxy 

 cement. The three tissue thermocouples and a ref- 

 erence thermocouple (for respirometer water 

 temperature) were each connected in series (con- 

 stantan leads) to an ice-bath reference couple (0°G) 

 and to an RS Beckman Dynograph (copper leads) 



through a high-quality, shorting rotary-switch. 

 This arrangement permitted rapid switching be- 

 tween thermocouples without opening the recorder 

 circuit. Thermocouples were standardized in a 

 water bath at 20°±0.05°G before and after each 

 trial. 



Paired electrodes for recording ECG's and 

 EMG's were prepared and implanted (in the same 

 sites used for thermocouples) as described by 

 Roberts (1975). The EGG and EMG signals were 

 preamplified using high impedence, probe am- 

 plifiers (Grass, P511DR) to improve the frequency 

 response of the RS Dynograph. 



Seawater was kept continuously flowing 

 through the respirometer tube and ambient tem- 

 perature was maintained within 2.0°G in each ex- 

 periment by mixing warm and cold seawater at 

 the outlet taps of the laboratory seawater system. 

 Over the 2-mo course of experiments, respirome- 

 ter temperatures ranged from 16° to 22°G. 



Results 



Changes in excess tissue temperatures that ac- 

 company increased swimming speed in the mack- 

 erel are best seen in a particularly successful trial 

 with fish number 6 (Figure 1). Similar, but some- 

 what variable records of heart, and red and white 

 muscle temperatures were obtained for all fish 

 (Table 1). 



While cruising at low speeds, excess tempera- 

 tures reached a maximum of about 0.3°G in the red 

 and white muscles, but doubled within 3 min 

 swimming at enforced higher speeds (3.2-4.5 

 BL/s). Excess temperatures recorded in the heart 

 averaged about one-half of the excess developed in 

 muscles at all swimming velocities. When swim- 

 ming speeds were reduced once again to slow 

 cruising, excess temperatures returned to pre- 

 burst levels within 8-15 min. 



During bouts of prolonged high-speed swim- 

 ming (5-6 min), water in the swimming tunnel was 

 warmed about 1°C due to frictional heating even 

 though a continuous exchange of seawater was 

 maintained from the supply tap (about 15 1/min). 

 This thermal error was minimized by rapidly ac- 

 celerating the fish from slow cruising to its pre- 

 determined, burst-swimming velocity. In Figure 1 

 for example, the fish was accelerated from 1.4 to 

 3.9 BL/s in about 5 s followed by sustained swim- 

 ming for 3 min, and then rapidly decelerated to 1.4 

 BL/s. Equilibration of tissue thermal excess (i.e., 

 generation minus dissipation) occurred in most 



862 



