530 Dr. G. J. Burch. Ona Method of Measuring  {[Feb. 7, 
deviation instrument, with this added advantage, that by suitably selecting 
the refracting angles and the initial position of the pair of prisms, the 
distribution of wave-lengths in the spectrum can be varied between wide 
limits, so that an increased dispersion can be obtained in the red without 
unduly altering that of the violet. | 
I used two dense-flint prisms of 60° which happened to be available, but 
the best distribution of wave-lengths was given by a pair of prisms of 45° 
which were unfortunately too small for my purpose. 
Each prism was mounted on a revolving table just big enough to hold 
it, and carrying a projecting arm about 20 cm. long. These arms were 
connected by hnkwork in such a manner that any movement of one was 
associated with an equal and opposite movement of the other. It should be 
noted that for ordinary purposes no great accuracy of page is needed 
for this part of the apparatus. 
In order to get as much light as possible, and to cover a wide field of view 
with the slit, I adopted my usual plan of a short-focus collimator—about 
25 cm.—and a long focus telescope—about 120 cm. For the second spectrum 
I used an instrument of the ordinary constant deviation type, of Se 
dispersion to show the sodium lines easily separated. : 
Light.—For all but the extreme violet Nernst lamps proved sition 
and gave a steadier light than that of the electric arc. Save for economy in 
lamps, no advantage was gained by illuminating both spectroscopes with hght 
from ‘the same source. 
Altcrnation of Flashes—The stimulus was rendered intermittent by 
rotating discs with suitably arranged sectors. There are several positions 
in the path of the rays where such discs may be placed, and these were all 
tried. The final choice lies between two arrangements, each of which has 
certain advantages. In one, two discs, one to each spectroscope, are geared 
together by a simple arrangement of shafting that admits of the separate 
adjustment of each both as to position, so that they can be focussed, and 
as to the sequence and duration of the flashes given by them. In the other, 
the disc carries a mirror by which the lght of one spectroscope is reflected, 
that of the other being seen periodically through an aperture. 
Superposition of Images—As in photometry, much is gained by making 
the border between the spectrum under examination and the monochromatie 
flash-light as nearly as possible invisible. For superposing the spectrum on 
the monochromatic field I have used reflection from the unsilvered surface 
of a prism, and have also employed an ordinary small total reflecting prism. 
I was investigating the relative merits of these two methods, when my 
illness put an end to the work. 
