468 A. M. Banta 



allel planes, 2.3 cm. apart, upon a base of slate 46 cm. long by 6 

 cm. wide. The slate afforded a fairly good foothold for the ani- 

 mals which crawl, but do not swim. This base of sufficient 

 width to prevent an easy overturning of the trough. The upper 

 end of the trough, into which the water flowed, was closed by a 

 piece of tin to within 4 mm. of the top; the lower end was closed 

 also by a piece of tin which, however, rose only to within i cm. of 

 the top. Closely fitting partitions of copper gauze 16 meshes to 

 the centimeter were used to confine the animals within the more 

 nearly central part of the trough, where the current was uniform. 

 Across the tank over each of these partitions was laid a piece of 

 glass to prevent the animals from crawling over the parti fon. 



The trough when in use was supported upon a small stand. A 

 glass jar 31 cm. high was placed close to the upper end of the tank 

 and under a faucet. The jar was kept filled to overflowing with 

 water during the experiments. Siphons of various sizes, depending 

 upon the strengh of current desired within the trough were used 

 to conduct the wate from the jar to the trough. The gauze par- 

 titions were placed at 10.5 cm. from the upper end and 3.5 cm, 

 from the lower end, this making the functional part of the trough, 

 within which the animals were confined, 32 cm. long. 



The siphon used introduced the water near the extreme upper 

 end of the trough and below the level of the water. There were 

 considerable irregularities in the upper portion of the current. 

 In order to break up these irregularities and reduce the current 

 to uniformity, three partitions of wire netting, 5 meshes to the 

 centimeter, were interposed at intervals of about 2 cm. between the 

 place where the siphon discharged into the trough and the upper 

 partition of copper gauze. This arrangement secured a practically 

 uniform current throughout that portion of the trough in which 

 the annuals were placed. 



During some of the experiments one end of the tank was dark- 

 ened while the other end was illuminated by direct sunlight, dif- 

 fuse day light or bv a 6-glower Nernst lamp. The part of the 

 trough darkened was covered with a closely fitted black cloth; 

 when the upper end was covered the cloth was held about the 

 siphon by a stout rubber band, and when the lower end was dark- 



