272 C. 8. Hastings—Optical Constants of Glass. 
The next step was to rotate the great circle, and with it the 
prism, until the other face took exactly the same position ; the 
reading then would clearly be the supplement of the desired 
angle ; in practice, however, the angle was repeated by turning 
back the platform independently of the circle, the repetition 
being carried usually to six times, and then back over the same 
ground for verification. This method was well adapted to this 
work as the mass moved independently was small, and turned 
with little friction. It failed however, as might be expected, in 
measuring later the angles of deviation, for here the joints to be 
turned were much larger, and that limit of accuracy which is 
always reached so easily and quickly by the method of repeti- 
tion, was not below that of a single reading of the microscopes. 
f the adjustment of the prism for measuring the double 
angle of deviation for the different Fraunhofer lines, little need 
be said except that, besides the care taken in placing the sur- 
exist in every lens save for rays of a single wave length, or, in 
rare cases, of two wave lengths. : 
The position of the prism for minimum deviation of a single 
line was carefully determined on the small circle attached to 
the platform, by experiment, while those of the other lines 
observed wefe computed after an approximate measurement 
of the angle of deviation. 
point the telescope upon the particular Fraunhofer line which 
as to be ism in i 
etermined, with the prism in its first position of 
ation and in, the difference of readings being the 
double deviation required. With this observation was also 
put the temperature of the prism at the pido mepri 
thermometer reading to quarter degrees, estimated, however, to 
tenths, and then filled with water. The reading of the barom- 
ay 
