Chemical and Physical Papers. 63 



THE SPECTRUM AND SPECTRUM SCALES. 



By J. T. LOVEWELL, Topeka, Kan. 



TN the ordinary laboratory spectroscope we may often observe a 

 -■- line which we wish to compare, for identification, with numbers 

 in the tables of Kirchhoff, or with wave-lengths as given in Ang- 

 strom's and similar charts. Perfect agreement of readings cannot 

 be expected in different refraction spectroscopes, where lines are 

 designated either by the angle of deviation or by the number of a 

 photographed scale, which is brought into the field of view with 

 the spectrum. Every prism has its own problem of refraction, and 

 every photographed scale must have imperfections. In the dia- 

 gram herewith presented use is made of curves, which are now 

 employed in the discussion of so many physical problems. The spec- 

 troscope with which observations were made was a Duboscq single- 

 prism instrument, with the ordinary photographed scale reflected 

 into the field of the observing telescope. The scale numbers were 

 from 1 to 200, and the first problem was to compare these with the 

 numbers of Bunsen's charts of flame spectra. On a sheet of cross- 

 section paper, ruled in inches and tenths, the lower margin was 

 marked with the Bunsen chart numbers, which were to be the 

 abscissas of the proposed curve. A few of the Fraunhofer lines of 

 the solar spectrum are drawn in the margin as given in Bunsen's 

 charts. For ordinates of the curve, readings of the spectroscope 

 were used ; for solar lines and for a great variety of flame spectral 

 lines, obtained from soluble salts of metals, brought into the flame 

 of a Bunsen lamp. The scale number of the observed line forms 

 the ordinate of the curve, and these numbers are placed on the left- 

 hand margin of the chart. The scale was adjusted so that the 

 number 50 coincided with the sodium line or the Fraunhofer D 

 line of the solar spectrum. The abscissas and ordinates of a large 

 number of points in all parts of the spectrum were next found, and 

 through the mean position of these points a curve was drawn which 

 proved to be practically a straight line. Had the spectroscope em- 

 ployed by Bunsen been exactly like the one used here this line 

 would have been at an angle of 45 degrees, but it was found to 

 have an angle less than that, measured from the horizontal. 



The Bunsen and Duboscq scales were supposed to be alike, as 

 were also the prisms, but 50, the number to which they were ad- 

 justed, is the only point of exact coincidence where the lines cross. 



