568 
MR. J. NORMAN LOCKYER ON SPECTRUM-ANALYSIS 
40 fell as near the places assigned by Cornu as could possibly be expected, considering 
the great difference in the scale employed, and the vast increase in the details obtained 
by the grating. The new place for 39, which lies at 38’992 on Cornu’s map, was finally 
settled as the start-point. The distance from 39 to 40 on the glass plate was carefully 
divided into 1000 parts; a scale was made in ink, and from the glass plate, which 
now contained both the spectrum and the scale, an enlargement on paper was obtained 
slightly greater than the map to be produced from it. 
I willingly here express my great obligation to M. Cornu’s work, and add a tribute 
of admiration zo its value and completeness. The way in which it bears the severe 
test put on it by the larger scale I have employed is marvellous. 
C. Construction of the Map. 
Being thus in possession of an enlarged photographic print on which are marked the 
absolute wave-lengths, the next thing done was to study it side by side with a refrac¬ 
tion photograph brought up to the same scale. Owing to the cause I have before 
referred to, it was found that, though there was no difficulty in recognising the chief 
lines in both, still in the diffraction photograph the smaller details were in many 
places quite different, and in many others very difficult to harmonise, the intensities of 
the lines having been greatly changed. Although, therefore, I was able to use the 
diffraction photograph for the positions of the chief lines, I had to depend on the 
refraction photograph for the detailed work and the intensity of the lines. 
A trial map on the scale adopted was then very carefully made by the aid of 
the photographs and the original standard refraction negative. An enlargement was 
o 
then made on three times the scale, =: twelve times the scale of Angstrom. 
I tried several methods of conveniently comparing the drawing with the photograph, 
after the chief lines had been inserted in positions indicated on the scaled negative. 
The following I found most satisfactory, and, indeed, it has been of great service in 
the construction of the map. 
The board on which the large-scale map was drawn was placed in an upright 
position, and at about 10 feet in front of it a small sighting aperture of 3 millims. 
diameter was adjusted. 
The diffraction photograph with its accompanying scale was supported between the 
board and this aperture, care being taken to keep the line joining the eye-hole and 
the centre of the map at right angles to the plane of the map, and the plane of the 
photograph parallel to the same plane. The position of the photograph was so adjusted 
that the lines on the photograph and the map, thus viewed from the observing aperture, 
were absolutely continuous. In this manner I was able to correct any error in the 
positions of the lines in the enlarged map with the greatest certainty. The details 
and intensities were added afterwards by mounting the refraction photograph in the 
same manner. Having thus corrected, checked, and finished the large map in ink, it 
