for use in the examination of Minerals. 131 



the passage of the tubes D, T, and K ; by this means all extra- 

 neous light is prevented from reaching the instrument. 



The point of novelty in this spectroscope is, in addition to 

 some minor mechanical details, the reflection, by means of the 

 prism g, of the bundle of rays transmitted by the slit at right 

 angles to their former path, with a minimum loss of light and 

 without the annoyance of secondary spectrum. The advantage 

 gained is an important practical one. While using but one 

 prism for dispersion, the eyepiece of the observing-telescope and 

 the slit S are brought close together, so that the analyst, having his 

 eye at the former, still has the slit, the gas-flame, and the plati- 

 num wire carrying the substance to be examined perfectly under 

 his own control. The approximate position of the slit in the 

 ordinary instruments is shown by the dotted line H, in which case 

 it is scarcely within easy reach if the apparatus be of large size. 



By thus increasing the facility with which the spectroscope may 

 be used in the examination of minerals, I hope it may induce 

 many mineral-chemists to employ it in their investigations, and 

 thus render their observations on the constituents of minerals 

 more accurate and in every way satisfactory. 



I may conclude with a few hints as to the method which expe- 

 rience has shown me is most satisfactory in testing minerals with 

 the spectroscope when the colourless flame of the Bunsen burner 

 is used as the source of heat : — 



1. Moisten the finely powdered mineral with pure water, take 

 up some of the paste on a loop of platinum wire and slowly in- 

 troduce into the flame, observing the spectrum produced, if any. 



2. Moisten the same portion, or a fresh quantity if the first 

 be lost, with a drop of pure hydrochloric acid ; again subject the 

 test-specimen to the heat of the flame, and observe. Most sili- 

 cates yield up their bases easily by this treatment. 



3. To the same test add a drop of a mixture of hydrofluoric 

 and hydrochloric acids, allow the liquid to act for a few minutes, 

 and then heat as before. 



By adopting this systematic treatment, no substance which 

 gives a characteristic spectrum under these circumstances can 

 escape detection ; and if the three tests be applied to the same 

 specimen, the first one, by causing the volatilization of the greater 

 part of the most volatile constituents, renders the succeeding 

 stages more clear and precise. 



Finally, in examining sulphides, it is necessary to heat a por- 

 tion of the substance in a small closed tube ; in the sulphur sub- 

 limate thallium is easily detected if it be present in the ore ; and 

 the residue needs but to be moistened with hydrochloric acid, and 

 heated, for the detection of the presence of such metals as admit 

 of easy recognition by the method of spectrum- analysis. 



K2 



