Aquatic Wtt 



83 



dial) with another, half its diameter, 

 mounted on the hour hand spindle. This 

 idle gear should carry a disc one-half of 

 which should be composed of brass, 

 while the other should be of ebonite or 

 other insulating substance, A brush 

 bearing on the periphery of this disc 

 connected to one side of the circuit and 

 the other connected to the stationary con- 



•WlRlttCi DIAGRAM 



worth of this method was unfortunately 

 not possible owing to unexpected neces- 

 sity for dismantling the room used in 

 the experiment. It was in progress on 

 about ioo young Barbus conchonius and 

 vitatus in a 4-foot tank. A control tank 

 like the first in all respects and contain- 

 ing similar fish, but feed by hand, failed 

 to show the same growth although the 



tacts would cause every revolution of the 

 hour hand (once in 12 hours), and con- 

 sequently one-half a revolution of the 

 disc to alternate twelve hours of closed 

 circuit with twelve open, approximating 

 the day light hours close enough for most 

 purposes. In this manner if an eight 

 day clock is used, the device need be 

 thought of but once a week. 



A complete and positive test of the 



priod in which the device was in use 

 lasted little over a month. It was found 

 that the amount consumed was consid- 

 erably greater in the machine-fed tank, 

 which is the most significant fact to be 

 noted. About four times as much was 

 taken with by thus giving a little at a 

 time all through the day, instead of a 

 comparatively large quantity once a day. 

 The writer believes that it would pay an 



