192 F. W. Clarke — Tschermcbtf s Theory of the 



In several previous papers, some of them written in joint 

 authorship with others,* I have sought to establish the hypoth- 

 esis that the more complex silicates are merely substitution 

 derivatives of normal salts. For instance, many minerals are 

 easily and naturally represented as so derived from the normal 

 aluminum compound Al 4 (Si0 4 ) 3 ; successive aluminum atoms 

 being equivalently replaced by other atoms or groups. Among 

 these minerals, the micas seem to form a conspicuous family, 

 and with them are found the clintonite group and some ver- 

 miculites. To this family I shall recur later. 



In two quite recent papers, Schneider and I have tried, by 

 experimental methods, to obtain some positive evidence as to 

 the nature of the chlorites. That our results are not final, and 

 that Tschermak's criticism of them is perfectly legitimate, I 

 cheerfully concede, and yet we have found clues which are not 

 without meaning. The reaction between dry hydrochloric 

 acid and the lrydromagnesian silicates we are still investigating, 

 and its actual character and significance we hope to determine 

 in due time. So far, however, in every case of a true chlorite 

 examined by us, and in the case of serpentine also, we have 

 found that the chemical constitution may be represented by a 

 mixture of molecules of the general form Mg a (Si0 4 ) a B/ 4 , in 

 which B/ may be either A1H 2 2 , MgOH, or H, with of course 

 the usual equivalent replacements of aluminum and magne- 

 sium by ferric and ferrous iron. These molecules are evi- 

 dently substitution derivatives of the normal compound oli- 

 vine, when the ordinary formula of the latter is doubled and 

 made Mg 4 (Si0 4 ) 2 . Since we do not know the true molecular 

 weight of any metallic silicate, this doubling of the simplest 

 possible formula is perfectly allowable, and a reason for.it will 

 appear later. Here then, for a few chlorites at least, is a 

 scheme of interpretation quite unlike Tschermak's ; and we 

 have to determine whether it is equally exact and equally 

 general. • 



Taking Tschermak's paper as a basis for comparison, and 

 discussing at first only the analyses discussed by him, I find 

 that all of his orthochlorite series, except corundophilite and 

 amesite, reduce sharply to the above given type of formula. 

 Omitting amesite, which will be separately considered further 

 on, every orthochlorite may be written B/^SiOJ^R,^, with 

 various relations between the several components collectively 

 represented by R/ 4 . In fact, Tschermak's formulEe and mine, 

 although quite dissimilar, are curiously parallel ; as the follow- 

 ing scheme, based upon his classification of the orthochlorites, 

 plainly shows. His arrangement of the sub-group is this : 



*See U. S. Geological Survey Bulletins, numbers 27, 42, 60, 64 and 78. Also 

 this Journal for Sept., 1891. 



