BEHAVIOR OF AMPHIPODS WITH RESPECT TO LIGHT. 211 



and Shelf ord ('14). In the light grader the animals are kept 

 during experiments in a small rectangular tank having glass 

 sides. A false bottom in this tank allows running water to pass 

 through and thus keep the water the animals are in at constant 

 temperature. Midway between this tank and the nernst lamp 

 in the grader there is a partition having a triangular aperture and 

 this aperture is covered by a lens. By this means an intense 

 field of light is made to fall upon one end of the small tank when 

 the latter is placed at the focal point of the lens. This field of 

 intensity shades off to darkness in the opposite end of the tank 

 because of the triangular opening, thus making an intensity 

 gradient. The light which passes through the glass sides of the 

 small tank is reflected by mirrors to a dead black wall in another 

 part of the grader. 



When ready for the experiment the animals were placed in the 

 small tank, usually three at a time, and allowed to remain in 

 darkness for a short time to recover from the shock of handling. 

 Then the light was flashed upon the tank and immediately the 

 animals were released from the glass tube with which they had 

 been confined. Every thirty seconds the relative positions of 

 the animals in the tank were recorded. In part of the expei;i- 

 ments 40 readings each were taken, the first 10 of which were 

 discarded because of the excitement of the animals due to hand- 

 ling and to the flashing of the intense Hght upon them. In the 

 remaining experiments 25 readings each, with the first five dis- 

 carded, were found to be enough to give typical results. 



In each set of experiments the tank occupied by the animals 

 was placed in three different positions : position i , at right angles 

 to the direction of the light rays; position 2, at an angle of 45° 

 to the direction of rays with the dark end nearer to the lamp; 

 position 3, at an angle of 45° to the direction of rays with the 

 light end nearer to the lamp. The animals were first released 

 in the field of intense light and a series of readings taken; then 

 the same animals were released in the dark end, and oftentimes 

 also in the place of medium light, the readings being repeated 

 in each case. 



About 50 of the 257 experiments performed were eliminated 

 because of avoidable errors in preliminary work. The animals 



