101 Ox\ A TABLE OF STANDARD WAVE LENGTHS OF THE SPECTKAL LINES. 
made to overlap, so that, by adding a constant to each set, a continuous series 
cuiild be formed for the whole spectrum which would be proportional to the wave 
lcii<'-th except for some slight errors due to the working of the apparatus for 
keeping the focus constant. Making this series coincide with two standards at the 
ends, tlie wave lengths of all could be obtained by simply multiplying the wdiole 
Bcrles by one number and adding a constant. This usually gave the wave lengths 
of the whole spectmm within 0.1 dr 0.2 divisions of Angstrom. The differences 
of this series from the primary standards were then plotted, and a smooth curve 
drawn through the points thus found. The ordinates of this curve then gave the 
correction to be applied at any point. 
It is to be noted that the departure from the normal spectrum was very small, 
and the correction thus found was very certain. The cause of the departure was 
not apparent, but may have been the slight tilting of the spectrum, by which it 
was measured ^somewhat obliquely at places. 
The visible spectrum wms thus gone over five or more times in this manner, 
with several different gratings and in different orders of spectra. The results are 
given in Table X., Columns C, R, p, q, m, 0, e, h, i, etc. The spectrum from the 
green down to and including A was also observed on a large instrument for flat 
gratings, having lenses six and one half inches in diameter and of ei";ht feet focus. 
These latter observations are marked 0\ This rcL^on I intend at some futui 
time to observe further. 
D 
Ii was now required to observe the ultra violet to complete the series. For 
this purpose the coincidences of the 2d, 3d, 4th, 5th, and 6th spectra of a 7000, 
21! feet radius, grating were photographed. My instrument will take in photographic 
plates twenty inches long, but there will be a slight departure from a normal spec- 
trum in so long a plate. Hence plates ten inches long were mostly used for this 
special series. Before the camera was placed a revolving plate of metal about 
three sixteenths of an inch thick, and havincr a slit in it of the same width.* When 
flat side was pai 
camera plate, a strip of the spectrum three sixteenths 
of an inch wide fell on the plate. When turned ninety degrees, the plate shielded 
this portion and exposed the rest. Using absorbents, it was thus possible to photo 
a strip of say the 4th spectrum between two strins of the 5th. This ar- 
better than havincr 
ps 
only two edges come toi^ether. To correct any 
niovement of the apparatus during the time of exposure, I expose on one specti 
then on the other, and back a-ain on the first 
described 
