August 6, 1909] 



SCIENCE 



167 



and to isomorphous mixtures; when these 

 sources of doubt are eliminated, simple 

 empirical formula can in most cases be de- 

 duced. To obtain such formuloe is obvi- 

 ously the first step in the general investiga- 

 tion ; a task which, fortunately, had already 

 been largely performed by others. 



Empirical formulae, however, do not solve 

 the problem of constitution. That has to 

 be attacked along other lines. The natural 

 associations of the silicates, their isomor- 

 phous relations and their alterations all 

 supply evidence, which can be supple- 

 mented by experiments in the laboratory. 

 How does a mineral come into existence? 

 How does it decompose ? What character- 

 istic reactions are possible with it? To 

 each of these questions answers have been 

 sought in the laboratories of the survey, 

 and important data have been obtained. 

 The synthetic work of Day and Allen on 

 the feldspars was done in the Geological 

 Survey, and similar researches are now 

 being undertaken in the Carnegie Institu- 

 tion of Washington, to which that class of 

 investigations has been transferred. The 

 geophysical laboratory of the Carnegie In- 

 stitution is a direct outgrowth from the 

 survey, in which its work began. 



The decomposition and alteration of sili- 

 cates can be studied both in their natural 

 occurrences and in the laboratory. Wlien 

 one mineral alters into another, a direct 

 relation is established between the two, 

 which the constitutional formulae ought to 

 sjonbolize. In the laboratory, such altera- 

 tions may be often brought about artifi- 

 cially, as was done long ago by Lemberg 

 and other investigators. The decomposi- 

 tion of minerals by heating is even more 

 easily studied, and two or three reactions 

 of this kind have been established in the 

 survey. For example, talc, H2Mg3Si40i2, 

 was commonly interpreted as an acid meta- 

 silicate, but Groth regarded it as a basic 



salt of metadisilicic acid, HoSijOj. In the 

 latter case heating should of course elim- 

 inate water, but there could be no further 

 breaking down. Clarke and Schneider, 

 however, have shown that when talc is 

 sharply ignited, one fourth of the silica is 

 split off in the free state ; a reaction which 

 is intelligible only on the basis of a meta- 

 silicate formula. A similar reaction is 

 furnished by pectolite, HNaCaoSiaO,,, which, 

 as G. Steiger found, gives up one sixth of 

 its silica upon ignition. In this instance, 

 as in the ease of talc, the liberated silica is 

 proportional to the acid hydrogen in the 

 initial substance. How far such reactions 

 can be trusted in discriminating between 

 metasilicates and the salts of other silicic 

 acids is yet undetermined. 



Similar experiments with serpentine gave 

 even more important results. Daubree had 

 shown, qualitatively, that when serpentine 

 is fused the residue is a mixture of enstatite 

 and olivine. Olivine is easily soluble in 

 dilute acids, enstatite is insoluble and it 

 was therefore easy to ignite serpentine, to 

 dissolve out the olivine, and so to prove 

 that the reaction is quantitative. Now for 

 its application. Tschermak had proposed a 

 theory of the chloritic minerals, in which 

 they were treated as mixtures of two end 

 products, serpentine and amesite, the latter 

 being an extremely basic salt to which the 

 formula H4Mg2Al2SiO(, was assigned. If 

 this were true, a chlorite like clinochlore 

 should yield on ignition an insoluble resi- 

 due containing the enstatite end of the 

 serpentinous decomposition. In fact, a 

 clinochlore which ought to have given 18 

 per cent, of enstatite gave none at all, but 

 a residue consisting of spinel, MgAljO^. 

 That is, it failed to give the decomposition 

 products of serpentine, therefore serpentine 

 was absent, and the Tschermak theory feU 

 to the ground. Other allied minerals 

 furnished a similar spinel reaction, and in 



