ORIEXTATIOX OF MIGRATING ANADROMOUS FISHES 



385 



T.\mK 4. — Kxnmplrs of chemical measurements maHe 

 ilurivij iiuliridual tests, June 11, 19!>0 



(Stream temperatures, 19° to 19.4° C] 



' COi, gaseous COj added; control, water immodifled; KsHCPOi), 

 KjHCPOO added. 

 ' Samples titrated with XaOH. 



when considering their physiological effects iipon 

 organisms is recognized. The ability of a gas to 

 diffuse through a membrane depends upon the 

 tension of the gas in solution. Therefore, the 

 tension of the gas is critical where, as in respira- 

 tion, an actual gaseous exchange is made. How- 

 ever, very iUtle is known of the sensory mecha- 

 nism by which fish detect differences of CO2 ; nor 

 is it known whether CO2 must actually permeate 

 a membrane to affect the sensory organs. Under 

 the circumstances of these experiments, where 

 botli channels receive water from the same source 

 under identical conditions, the CO2 tension would 

 be directly proportional to the amount of fi'ee 

 CO2 present. Therefore, the relative amounts of 

 CO2 in the two channels may be used as indexes 

 of tlie relative CO2 tensions. 



Oxygen 



Tlie amount of dissolved oxygen was deter- 

 mined by the standard Winkler method (Ameri- 

 can Public Healtli Association 1946) after pre- 

 liminary orientation tests (Ellis, Westfall. and 

 Ellis 1948) for the presence of interfering sub- 

 stances proved to be negative. 



Temperature 



Temperatures were measured to 0.1° C. by the 

 use of a mercury thermometer held horizontally 

 in the water with its long axis in the direction of 

 the current. The measurements were taken at 



the downstream end in each channel with the 

 center of tlie tiiermometer at a point 2 inches from 

 tiie bottom of the trough (the level at which the 

 lish usually swam) and 2 inches from the wall 

 whicli divided tiie trough into two channels. 



Velocity 



The velocity of tlie water flowing through the 

 trough was measured by means of an impeller- 

 fype current meter. Measurements were accu- 

 rate to within 0.2 feet per second, and represent 

 the average velocity of the water over a 90-second 

 period of time. The measurements were taken 

 at the downstream end of each channel at a point 

 halfway between the channel walls. 



ABERRATIONS IN EXPERIMENTAL CONTROL 



Throughout the experiments, measurements of 

 phj'sical and chemical water characteristics such 

 as temperature, velocity. //H, and amounts of 

 dissolved gases were imuh more juecise than 

 the experimental control of these characteristics. 

 Hydraulic conditions above the head of the ex- 

 ])erimental trough created a periodic eddying 

 (every 5 to 15 seconds) which resulted in a fluctua- 

 tion in the rate of flow alternately in each channel. 

 During tests in which water characteristics were 

 l)eing modified, eddying caused a periodic fluctua- 

 tion in the degree of modification. For example, 

 during tests in which the water was modified by 

 lieating. eddying resulted in a temperature fluctu- 

 ation of approximately 0.5° C. in the channel being 

 modified. 



Temperature measurements to within 0.1° C. 

 were made almost instantaneously, and the extent 

 of the temperature fluctuation was easily measured 

 by a comparison of maximum and minimum tem- 

 perature readings. However, measurements of 

 other water characteristics such as velocity, />H, 

 and amounts of dissolved gases were average 

 measurements. Velocity measurements (read in 

 terms of propeller revolutions per unit time) rep- 

 resented an average velocity for a 90-secoud period 

 of time. The manner in which the water samples 

 were taken for measurements of pH and amounts 

 of dissolved gases tended to collect a mixture of 

 the water flowing through a channel over a pei'iod 

 of time greater than the time for a complete cycle 

 of eddy fluctuation and so tiiese measurements also 

 represent measurements of average conditions. 



