Second Spectrum of Hydrogen. 343 



is also not suited to satisfy the requirements of spectroscopic 

 science, since in the absence of numbers, or even of a scale, 

 not even approximate wave-length determinations of the few 

 lines given can be obtained from it. The measurements given 

 below and the drawing constructed from them (PL X. fig. 1) 

 will supply this deficiency, at least to some extent. 



The spectroscope which I employed is of the usual Steinheil 

 model, and when a powerful Rutherfurd prism is employed, 

 shows the spectrum in great detail. 



The dispersion, which was somewhat greater than that of 

 two ordinary flint-glass prisms of 60°, might, with the means 

 at my command of rendering the gas luminous, have been 

 somewhat increased ; but little would have been gained by 

 that, since the rest of the spectrum was not of sufficient inten- 

 sity to bear a higher dispersion. The magnifying-power of 

 9*3 employed in the first series of observations was increased 

 to 17*0 with the remaining series, since the brightness of the 

 spectrum proved to be sufficient to permit such an increase. 



The wave-lengths of the lines were determined by micro- 

 metric measurement with reference to suitable metallic lines 

 chosen from Thalen's tables. These lines were always so 

 chosen that the errors of interpolation might influence as little 

 as possible the reliability of the wave-lengths determined. It 

 would certainly have been an advantage if the lines of the 

 gas could have been directly determined from those of the 

 solar spectrum ; but with a spectrum of such feeble intensity 

 as the one in question, this method is attended with consider- 

 able difficulties, and may, in my opinion, even introduce 

 greater errors than those which are to be apprehended from 

 the employment of the metallic spectra as references. 



It is well known that a Geissler tube, placed immediately 

 in front of the slit of the spectroscope, acts as a cylindrical 

 lens, and every small displacement produces a greater or less 

 apparent change in place of the lines of the spectrum. This 

 circumstance in the present case, when, in order to observe the 

 comparison-spectrum, the tube had to be pushed on one side, 

 would have given rise to extremely prejudicial irregularities 

 in the measurements ; and I have therefore, in order to avoid 

 this source of error, employed an achromatic lens of short 

 focus, so as to throw a sharp image of the tube or of the spark 

 used for comparison upon the slit. In this way, the projecting 

 lens being fixed, the spectrum remains altogether fixed in the 

 field of view, and the source of error mentioned is avoided. 



The source of electricity employed was, in all cases, a large 

 Ruhmkorff's coil, made by Keyser and Schmidt, of Berlin, 

 excited by a battery of six large Bunsen cells. Commonly 



