DISTRIBUTION OF INTENSITY IN BROADENED SPECTRUM LINES. 
4G5 
spectrum lines produced under these conditions are of such a breadth that the dispersion 
of a single prism spectrograph was found to be sufficient. The spectrograph consisted 
of a large model constant deviation spectroscope by Hilger with a camera attachment. 
The neutral wedge, which was supplied by the same firm, was fastened on to a dia¬ 
phragm which was fitted in grooves immediately in front of the slit. The slit width 
used was 0‘025 mm. The width of the slit might necessitate a correction in the 
interpretation of the photographs, but since in the present case the greatest correction 
would be less than one per cent, of the width of the base of the wedge, it may for the 
present purpose be neglected. The curvature of the spectrum lines obtained with 
prism spectrographs has also been considered as a possible source of error, but this also 
was found to be negligible over the length of slit in use. 
The greatest care was taken to ensure the even illumination of the slit, since this 
might give rise to serious errors. In the case of vacuum tubes there is no difficulty 
in ensuring this, and in every case the even illumination of the slit was verified before 
the wedge was put into position. In the case of the spark discharges through 
hydrogen at atmospheric pressure, the light was concentrated on to the slit by means 
of a sphero-cylindrical condenser, and in some cases a piece of ground glass was 
interposed as a further precaution. Wratten and Wain weight's panchromatic 
were used, and were in some cases intensified with mercuric bromide and sodium 
sulphide after fixing. Enlargements were then made on to bromide paper with an 
enlarging apparatus provided with a Zeiss-Tessar lens, which gives no appreciable 
distortion over the field required. Since, however, it is in general more easy to work 
with a negative than a positive, the enlargements were usually carried out in two 
stages, the first enlargement being made on a Wratten “ process ” plate or a Paget 
“ half-toneplate, and the second enlargement on bromide paper. 
There remains the personal error in picking out the points of equal density. It is 
believed that this has been almost entirely eliminated by enlarging the plates through 
a ruled process screen, the resulting image thus consisting of fine black dots on a 
white ground. In this way measurements may be easily made by pricking out the 
last dot visible at a large number of points on the wedge, and subsequently drawing 
a curve through these points. The extreme dot visible is a very definite point and 
the drawing of the curves is thus reduced to an almost mechanical process, whereas 
the recognition of points of equal density on photographs, which have not been 
prepared in this way, is a matter requiring considerable practice and would undoubtedly 
be a source of error. 
For investigations of a more qualitative nature of the spectrum of lithium, a concave 
grating spectrograph was used. This instrument consisted of a concave grating of 
four feet radius of curvature and 20,000 lines to the inch and was mounted according 
to the arrangement described by Eagle, # the dispersion being about 10 A. IT. per 
millimetre. 
* ‘ Astrophys. Journ.,’ 31, p. 120, 1910. 
