The relative water voltage gradients plotted in Figure 2 were 

 determined with AC voltages applied to the electrodes. These gradients 

 were considered to approximate closely those which existed when pulsed 

 direct current was used. The measuring equipment consisted of a General 

 Radio Type 727-A high- impedance vacuum tube voltmeter and a specially 

 designed voltage gradient probe. 5/ All measurements were taken at 

 points halfway between the surfa'ce and the bottom along a line through 

 the centers of the electrodes. The ideally uniform voltage gradients 

 produced in the test tank were not duplicated in experiments conducted 

 in the field. The difference can be attributed to the confinement of 

 the test-tank electrode current to an insulated body of water whose 

 cross-sectional area was equal to the immersed electrode area. 



Field tests were conducted at Carp Creek and the Little Ocqueoc 

 River, Presque Isle County, Michigan. A mechanical vjeir trap of l/2-inch- 

 mesh hardware cloth (modified Milligan Creek type, Applegate and Smith, 

 1950) (selected because of its ready availability) was used as the posi- 

 tive electrode for the field tests. This trap had an opening approxi- 

 mately 18 inches square and was situated at the apex of a "V" formed by 

 2 frame-mounted pieces of cotton-cloth netting. Each piece of netting 

 was 3 feet long and together they functioned as a funnel. The cathode 

 was a section of l/2-inch-mesh hardware cloth 6 feet wide and 2 feet high. 

 The electrodes of the Carp Creek array were spaced 20 feet apart and 

 placed on a line almost directly across the stream. The Little Ocqueoc 

 River array was the same except that the distance between electrodes was 

 only 10 feet (Fig. 3). 



Field tests at the Carp Creek and Little Ocqueoc River sites 

 were conducted with the DC power supply and commutator unit previously 

 described in this section. This unit, mounted in a trailer, was powered 

 by a 5 KW 110/220 VAC, diesel-driven generator (Fig. U). 



Methods 



The initial tests were conducted in an attempt to find the best 

 combinations of variables to produce the desired leading effects. The cam 

 producing a duty cycle of 0.025 was installed and each repetition rate (l 

 through 5 cycles per second) was tested at a specific voltage level. The 

 voltage input was then increased and tests were again made at each repeti- 

 tion rate. Usually a series of tests started with a voltage input that had 

 little effect on the test fish and the voltage was increased in increments 

 of 10 volts until the test animals were electronarcotized. 



5/ See footnote 2. 



