376 PROCEEDINGS OF THE AMERICAN ACADEMY. 



apparatus could be placed -with the axis of the runway directly over 

 the normal axis of the diagram on the table and so oriented that its 

 anterior end was directed either north or south at will and coincided 

 with the directive axis. The animal to be tested was then dropped 

 into the cylinder and made to move about until at length it entered 

 the runway. When it came out at the other end of the runway, it was 

 properly oriented, with one light at its right, the other at its left. 

 Some of the most active individuals would run through this passage 

 and emerge without stopping ; but there was much individual variation 

 in their behavior in this respect. By far the larger number of them 

 stopped in the runway as soon as their antennae projected from the 

 farther end, and there they Avould remain, waving the antennae about 

 or cleaning them by drawing them through their jaws, but refusing to 

 go farther without being stimulated from the rear. Such individuals 

 were stimulated by pushing a small block of wood into the runway 

 behind them until it came against the cerci or the hind legs. The 

 block was made nearly as broad as the runway in order that it might 

 strike the animal evenly and not stimulate one side more than the 

 other. Usually the slightest touch was sufficient to start the insect 

 from the runway, but some individuals seemed very reluctant to leave, 

 and in a few instances, immediately they were outside, they turned 

 about and regained the shelter they had just left. Most of them, how- 

 ever, as soon as outside, ran ahead steadily and rapidly, either going 

 nearly straight or swerving to one side or the other ; others would 

 hesitate or even stop altogether, and then make sharp turns, often 

 going off finally in a direction entirely different from that which they 

 first took when they came out of the runway. The records indicate 

 the points at which they crossed the outer circle — the one having a 

 radius of 15 centimeters. 



Table VI summarizes the results of 98 trials with the large light 

 alone and 200 each with the small hght alone and with the two lights 

 used simultaneously. As in the preceding experiments, Figure 13 

 shows the records in graphic form. 



These records show that Periplaneta is decidedly negative to light 

 from one side, — 42.8 per cent of the reactions may be so regarded in the 

 case of the large light and 55 per cent in that of the small light. 

 Here the peculiarity will be noticed that, as in the case of Oniscus, there 

 was a larger per cent of reactions away from the small light than 

 away from the large one. Moreover, in the cockroach, as may be seen 

 from the lower third of Table VI and from Figure 18, C, there were 

 more turnings from the small light than from the large one, when both 

 were used. This excess is very small, it is true, — only 9 reactions in 



