clxxxii GENERAL SUMMARY OF SCIENTIFIC AND 



used in coinage by means of the spectroscope. These ex- 

 periments were made with the object of estimating the 

 practical value of a recent discovery of Mr. Lockyer, and 

 upon which he founded a theory of possible quantitative 

 spectrum analysis. Briefly stated, Mr. Lockyer's discovery 

 consists in the observation that when a powerful induction- 

 coil is employed, and the distance between the metallic elec- 

 trodes is gradually separated, certain of the lines in the 

 spectrum of the metal (or metals) under observation break 

 in the middle ; and that upon increasing the distance of the 

 electrodes, the hiatuses in the spectral lines increase propor- 

 tionately, but unequally, with different alloys. The experi- 

 ments of Mr. Outerbridge, which were quite thorough, seem 

 to establish, at least in the present state of spectroscopic 

 knowledge, the apparent paradox that the spectroscopic 

 method is in one respect far too sensitive and minute, and 

 in another far from being minute enough, to serve the uses 

 of assay. The problem simply stated is this to discern 

 differences in the lines of the spectra of different alloys of 

 gold and silver ; in other words, to utilize the spectroscope 

 as a means of quantitative as well as qualitative analysis. 

 We can not detail the construction of the apparatus nor the 

 conduct of the tests made therewith, but will simply state 

 the difficulties encountered and the conclusions drawn. The 

 author found, in the first place, that the method was not 

 sufficiently delicate to afford visible differences in two alloys 

 which could readily be estimated by the usual method. Va- 

 riations in the length and brightness of the lines, and the 

 confusion of the eye resulting from long-continued compar- 

 ison of minute differences, were likewise found to be inci- 

 dental to the process. Again, the quantity of metal vapor- 

 ized by a spark and giving the spectrum was found to-be 

 too infinitesimal (not more than the 1>ou ^ uuu of a grain) to 

 give safe results for a large melt, since the appearance of 

 the lines would be affected by the least want of homogeneity 

 in the metal. Another difficulty was found to reside in the 

 fact that, while the spectroscope is very sensitive to pure 

 metals, a comparatively large quantity of gold may be pres- 

 ent in an alloy, and the spectroscope not indicate its pres- 

 ence, and the same difficulty holds good for alloys of other 

 metals. 



