524 Professor John W. Nicholson [May 2, 



attenuated towards the part of the slit opposite the thick end of the 

 wedge, where the absorption of light is greatest, and the image 

 ceases to be strong enongh to affect the plate beyond a certain 

 specific height which depends on the original intensity, in the beam 

 from the source, of this particular wave-length. 



The photograph thus consists, not of the usual spectrum with all 

 lines or slit-images of the same length, but of a spectrum in which 

 all the lines are cut down to specific heights depending on the 

 original intensities, and thus it gives a simultaneous record of all 

 the intensities in the spectrum at any one instant. All spectrum 

 lines have a breadth, due to Doppler effect of the atomic motions in 

 the kinetic theory, and to other agencies. The shape of one of the 

 truncated lines depends on the original law of intensity across the 

 line. This shape may be a wedge, or may be bounded by a more or 

 less rounded curve, from whose nature, if the boundary can be 

 defined sharply, we can deduce mathematically the law of intensity 

 across the original line. Sharp changes of intensity, such as occur 

 when the line has several close components overlapping one another, 

 are detected as peaks or kinks in this bounding curve. The original 

 photograph can be enlarged with considerable magnifying power, and 

 if the bounding curve on this enlargement is sharply defined, we can 

 obtain its mathematical shape very accurately, and deduce an esti- 

 mate of the intensity in any part of the line with a great degree of 

 precision. We have been able to show\ for instance, that in most of 

 our experiments, such accuracy as one part in 100 has been reached, 

 and it could be increased readily if desired by the use of greater 

 magnification of the original photograph. 



The determination of the exact boundary of *a patch of dark on 

 a white ground is a matter in which '' personal equation " is impor- 

 tant. We overcame this difficulty by enlarging positives, prepared 

 from the negatives, on to bromide paper through a ruled " process " 

 screen. The resulting photograph consists in this way af an assem- 

 blage of very minute dots, fading away towards the boundary into 

 invisibility. It is a simple matter to prick out the last dots visible all 

 round the contour, and in this way personal equation can apparently 

 be entirely eliminated. We adopted usually about 100 dots to the 

 inch on the final photograph. If comparisons of different lines with 

 the another are required, only the central heights of the figures are 

 necessary, and the topmost dot can be seen at once. 



The first application of the method was to the intensity distri- 

 bution in the lines of the hydrogen spectrum when a condensed 

 discharge was passed through the exciting tube. It was known that 

 with a condensed discharge the lines always appeared much broader, 

 and we concluded that the l)est method of obtaining information as 

 to the source of the effect was to examine the intensity distribution 

 across the lines. Some remarkable contours were obtained, showing 

 at once a clear distinction between this source of broadening and that 



