on the use of Carbon bisulphide in Prisms. 273 
heat, the prism did not appear to feel the effect for half an 
hour. Then the lines ‘began to move toward the red end o 
the spectrum and continued to do so for more than five hours. 
On turning off the heat, the lines began to recede in half an 
hour; but although it was again turned on, they did not be- 
come stationary foran hour. On the 28d, the even temperature 
box was enlarged and a secovd box constructed within it, the 
space between the two being filled with cotton. On the 29tb, 
the regulator was started at 7 A.M. At1P. M. the temperature 
became stationary at 70° F. and the spectrum remained at rest 
or two hours. On the morning of January Ist, 1882, the 
temperature of the box was 54° F., having cooled from 72° F. 
over night, and the sodium lines had moved a distance of 1°8 
inches. From these experiments, the conclusion was reached 
controlled the next day. There were fluctuations of course ; 
ut they were not enough to injure the definition if the expos- 
ure was not over 15 or 20 minutes. On the 25th of January 
some changes were made in the heating tube by placing a 
funnel upon its lower end, by continuing it up through the top 
When the cireait was closed. This relay acted in concert with the 
electromagnet above mentioned, closing the tube when the yas 
was turned on and opening it when it was turned off. On light- 
Ing the gas jet below the funnel, the heated air rises into the 
tube and passes into the box through the lateral opening in the 
tube. When the temperature is reached for which the bars are 
set, they come in contact and close the circuit. The electro- 
became admirable in a few moments after starting the stirrer, 
On the 30th, the first systematic test was made with it. The 
