272 Experiments made by the late Dr. Henry Draper 
the Thollon prism as arranged with the stirrer. On December 
2d, using sunlight, the relative dispersion of the two was found 
to be very nearly equal, the flint train giving perhaps one- 
eighth more; but the amount of light transmitted by the bisul- 
phide prism was at least four times and in the region about @ 
eight times as great as that obtained from the Steinheil prisms. 
But now another source of error was developed. Although 
when the propeller was running, the definition of the bisul- 
change of position in the lines due to temperature-change 
wonld absolutely destroy the definition on the photographic 
experiment, made by placing a thermometer in the bisulphide, 
showed that for a change of 3°75° F. a change took place in the 
the spectrum a distance of 1:°375 inches. On turning on te 
* These measurements were made by Mr. D. 0. Chapman, who at that time was 
acting as Dr. Draper’s assistant. 
