HUNTER and ZWEIFEL: SWIMMING SPEED AND TAIL BEAT FREQUENCY 



APPARATUS 



Swimming speeds were measured in an ac- 

 tivity chamber (Figure 1) built after that of 

 Beamish ( 1966) . A fiber glass tube 230 cm long 

 and 41 cm in diameter was immersed in an open 

 bath. An 80-cm section of tube was the swim- 

 ming compartment. The compartment had 

 metal screens at the ends and a transparent 

 acrylic plastic hatch which conformed to the 

 contours of the tube. The walls of the tube 

 within the swimming compartment were white 

 and had black stripes spaced at 5.0-cm intervals 

 to provide a visual reference for the swimming 

 fish. Water velocity in the chamber was reg- 

 ulated by the speed of a 39-cm propeller driven 

 by a variable speed 50-hp motor and by changing 

 the screens at the two ends of the swimming 

 compartment. Water was drawn into the com- 

 partment from the bath over deflectors, and 

 through baffles and screens. The screens, bafiles, 

 and deflectors i-educed tui-bulence and provided 

 water of relatively uniform velocity throughout 

 the swimming compartment. Their arrange- 

 ment and design were determined empirically 

 by measurement of the horizontal and vertical 

 distribution of flow in the chamber. 



The speed range of the apparatus was 12 to 

 212 cm/sec. The full range was obtained by 

 changing the screens at the ends of the swim- 

 ming compartment. A velocity range of 12 to 

 69 cm/sec was obtained when screens of 39% 

 open area were used, one of 15 to 139 cm/sec for 

 screens of 56 Sr open area, and one of 19 to 212 

 cm 'sec for screens of 15'','r open area. 



A digital voltmeter measured to the nearest 

 millivolt the voltage produced by a voltage gen- 

 erator attached to the propeller shaft. The volt- 

 age produced by the generator was proportional 

 to propeller revolutions and was used to regulate 

 them. An impeller flowmeter (Marine Advisers 

 Inc., Model B-7C)" was used to relate propeller 

 revolutions in volts to water flow in the appa- 

 ratus. The meter sampled an area 7.6 cm in 

 diameter and had an accuracy of ± 2.5 cm/sec 

 when moved through static water at a known 



^ Reference to commercial products does not imply 

 endorsement. 



speed (for a description of the meter and a cal- 

 ibration curve see Olson, 1967). 



Propeller revolution was related to water 

 speed in the chamber by three series of cali- 

 brations, one for each of three screen types used. 

 Nine to 19 difl'erent speed levels were measured 

 in each series of calibrations. More levels were 

 required for slow speed ranges than for fast 

 ones because the response of the flowmeter was 

 nonlinear at low speeds. At each level water 

 speed was measured at 12 different radial po- 

 sitions midway in the swimming compartment. 

 The speed of the water at a given level was the 

 average of the 12 measurements, adjusted for 

 the extent of the area sampled by the meter 

 (Tranter and Smith, 1968). Variation among 

 the 12 sampling points did not exceed ±10% 

 of the mean speed and was usually much lower. 

 The relationship between mean water speed in 

 the chamber and propeller revolutions was li- 

 near, and the error in estimating the mean water 

 speed from revolutions did not exceed ± 0.2 

 cm/sec. Thus, the principal sources of error 

 in estimating the speed of the water in which 

 a fish swam were the possible 10 '^r variability 

 in flow within the chamber and the ± 2.5 cm/sec 

 accuracy of the flowmeter. 



Fish swimming in the compartment were as- 

 sumed to be swimming at the mean speed of 

 the water in the compartment. They were 

 photographed from above with a 16 mm high- 

 speed motion picture camera at speeds of 64 to 

 200 fps. Camera speed was adjusted to pro- 

 vide about 10 frames per complete tail beat. A 

 viewing box floated on the water surface above 

 the swimming compartment to eliminate distor- 

 tion in the photographs caused by ripples. 



METHODS 



Film was analyzed by use of a coordinate 

 reader and digitizer (Hunter, 1966) and a com- 

 puter program was used to calculate tail beat 

 frequency and tail beat amplitude from the 

 digitized information. One tail beat was one 

 complete oscillation of the tail, and the tail beat 

 frequency was the number of beats per second. 

 Tail beat amplitude was the distance in centi- 

 meters between the lateral most excursion of 



255 



