212 SCIENCE PROGRESS. 



More recently, however, Sollas (7) has devised an in- 

 genious application of heavy liquids, which deserves to be 

 widely known. By suitable dilution a series of Klein's 

 solutions are prepared ranging in specific gravity, say, from 

 3'3 to 2*5. Equal portions of these are introduced in order, 

 without mixing, into a vertical tube of diameter three- 

 quarter inch or an inch. When this has stood undisturbed 

 for perhaps twenty-four hours, it is found that diffusion has 

 proceeded just far enough to give a column of almost 

 uniformly graduated density from bottom to top, and so 

 slow is the movement of the salt in solution that the 

 "diffusion-column" so obtained can be used for several 

 days. Into it is introduced the rock-powder, sifted, washed, 

 and dried in the usual manner, and the constituent grains 

 at once range themselves at various levels according to 

 their several densities. When thus separated, they can be 

 withdrawn by a pipette and examined. Not only is the 

 process of isolation thus greatly simplified, but, further, the 

 specific gravities of the several minerals are easily deter- 

 mined at the same time. For this purpose a series of 

 "indicators," of suitably selected and accurately known 

 specific gravities, are introduced into the column, where 

 they take up their proper positions, and afford a scale of 

 comparison. By using a graduated tube to contain the 

 column, the specific gravity of a suspended mineral frag- 

 ment can readily be interpolated between those of the two 

 nearest indicators. These indicators may be prepared by 

 fusing short lengths of fine platinum wire into small glass 

 beads of various colours. 



Mechanical methods of isolating the several constituents 

 of a rock may now be considered to have reached a high 

 degree of perfection, and indeed they have almost super- 

 seded the chemical methods, depending- upon the selective 

 action of various acids upon the different minerals, which 

 have from time to time been practised for the same purpose. 

 Some such chemical methods were tried and abandoned by 

 Cordier, before he elaborated his process of " mechanical 

 analysis," and the experience of later workers seems to 

 show that they are not perfectly satisfactory except in some 



