3^36 AniCN'KY — PhtilognipJts of S/>nrk Sj)cc/r/i from the Rowland Spectrometer. 



The manganese spectrum by sparking a saturated solution of potassium 

 permanganate in a similar manner. 



The electrodes used with such solutions were made of six plates of ordinary 

 stout platinum foil, placed close together and fused into one end of a glass tube. 

 The glass tube containing the solution and carrying the lower electrode was of a 

 U shape, the open end being prolonged sufficiently above the closed end, to act 

 as a reservoir for the liquid, and to exert a slight pressure upon the electrode so 

 as to keep it fed continuously. The glass tube carrying the upper electrode was 

 straight, and was also filled with the solution. With a little experience, electrodes 

 of this form may be made to feed solutions of different strengths perfectly. 



The advantage of platinum electrodes over pure graphite ones lies in the fact 

 that during the necessarily long exposures they do not become shorter by combustion 

 or disintegration of the material. 



The specimen of gold was obtained from crystallized gold chloride, by preci- 

 pitation with pure oxalic acid, washing and redissolving in aqua regia, and again 

 precipitating with oxalic acid. The precipitated gold was finally fused under 

 borax in a clay crucible. 



The original photographs were taken from the first order of spectra ; and the 

 reproductions published with this communication are of the same size as the 

 originals. The definition of the lines of this order is extremely fine, and the 

 dispersion quite sufficient for most practical purposes. For these reasons it has 

 been deemed inexpedient to have enlarged rejiroductions prepared for publication. 



A further most important feature of this order is that the lines are the least 

 distorted by the astigmatism of the grating, so little indeed that the characteristics 

 of the component lines are quite apparent from the photographs, and the general 

 character of the spectra are wholly unaffected. By the characteristics of the com- 

 ponent lines is meant their peculiar features as observed in jjrism spectra, that is 

 to say, whether extended or not, continuous or discontinuous, well defined or 

 nebulous. 



On comparing these photographs with those obtained by Eder and Valenta, 

 and also with those of Exner and Haschek, important differences are to be 

 observed ; many lines photographed by those observers are absent ; differences are 

 also occasionally observable in the character of the lines. 



Some of these differences are possibly due to a greater degree of purity in the 

 elements and compounds examined, but the majority, however, are certainly to be 

 ascribed to the different methods of sparking employed by the several observers. 



Attention may be drawn to the marked effect which the sparking of the 

 solutions of potassium chromate and permanganate between platinum electrodes 

 has upon the character of some of the lines of platinum. 



Careful measurements of the wave-lengths of the lines in the photographed 



