198 MESSES. A. SCHUSTER AND G. HEMSALECH ON 



3. Method of Conducting Experiments and Measurement of Photographs. 



The method of conducting an experiment is obvious from the preceding descrip- 

 tions. The apparatus being in adjustment, a photograph was taken of the spark 

 spectrum, on the stationary disc. This is for the sake of reference, chiefly for the 

 purpose of identifying the lines. The disc was then set spinning and when the desired 

 speed was attained and found to be uniform the Wimshurst machine was set going by 

 hand until about six sparks had passed. We thus obtained a number of images on 

 each film, some of which sometimes were found accidentally to overlap. The images 

 are not found to be equal in clearness, the spark not always passing parallel to the 

 slit ; the two best were selected for measurement. For this purpose the film, after the 

 images had been developed and fixed, was cut up, each image being carefully marked 

 for future reference. 



The measurements were made by means of a " comparator," a very beautiful 

 instrument, made by ZEISS, of Jena. The instrument consists essentially of two 

 microscopes at a fixed distance from each other. A sliding table underneath serves 

 to carry the photograph to be measured under one microscope and a scale under the 

 other. A photograph having been fixed to the table by means of soft wax, a fine 

 adjustment screw allows a certain relative displacement between it and the scale, 

 so that when any desired line is in the centre of the field of view of one microscope 

 the scale may be read with the other, and adjusted to give any desired reading. 

 This adjustment is of great convenience in the comparison of different spectra with 

 each other, as some air line, common to all, may always be placed so that the reading 

 is in every case the same. The scale is divided into \ millims., and a micrometer 

 eye-piece allows readings to 7^5^ millim. 



Speaking for convenience sake of the slit images as " vertical," the photograph 

 must be placed so that when the sliding table is moved the trace of the centre of the 

 field of view on the photograph is horizontal, and the measurements consist in 

 measuring the horizontal displacements of each spectrum line at different distances 

 from the edges of the spectrum. 



Our first method of conducting the measurement was to take contact prints on 

 glass of the selected images. On these copies horizontal equidistant lines were ruled 

 by a fine needle point fixed to a dividing engine. The displacements could then 

 be measured along each of these lines. The disadvantage of this method consists in 

 the labour of taking copies and ruling lines, and also in the loss of definition, which, 

 however small, was always noticeable in the contact print. At a later stage, there- 

 fore, the method of measurement was changed to the following : A brass frame was 

 made in which the film could be clamped so that it lay perfectly flat under the 

 microscope. Underneath the photograph was placed a transparent scale, made of a 

 portion of a film from which the sensitive layer had been removed, and on which 

 horizontal lines at a distance of about '55 millim. had been ruled. These lines could 



