252 SCIENCE PROGRESS. 



may, in fact, be regarded as substitution derivatives of simple 

 normal salts. 



According to this view the mica group is derived from 

 the normal aluminium silicate Al 4 (Si0 4 ) 3 , in which one or 

 more atoms of aluminium are replaceable by other radicles. 



Now Clarke has found that when serpentine is exposed 

 to the action of dry hydrochloric acid gas, one-third of its 

 magnesia is given up ; from the fact that olivine is un- 

 attacked, while serpentine is attacked proportionally to the 

 excess of oxygen over the orthosilicate ratio, he concludes 

 that in the magnesian silicates, only that part of the mag- 

 nesium is removed by gaseous hydrogen chloride which is 

 present as the univalent group MgOH ; serpentine then is 

 regarded as Mg 2 (Si0 4 ) 2 H 3 (MgOH), which is a substitution 

 derivative from the normal olivine orthosilicate Mg 4 (Si0 4 ) 2 . 



Ripidolite gives up 1 3 per cent, of magnesia ; if this be 

 represented by MgOH the constitution falls under the type 

 R" 2 (Si0 4 ) 2 R' 4 , which is again an olivine silicate and brings 

 the chlorites into the same series with serpentine. 



On the other hand phlogopite is unaffected by gaseous 

 hydrogen chloride, though completely decomposed by the 

 aqueous acid ; this is in accordance with the view that it 

 is Al(Si0 4 ) 3 Mg 3 R' 3 , a substitution derivative of normal 

 aluminium orthosilicate. 



The strength of Clarke's position as compared with 

 that of other mineral chemists is that his arguments are 

 based upon experimental evidence in addition to the 

 ordinary analytical results. There is scarcely a single 

 group of silicates for which some such evidence is not 

 needed. 



As another instance take the scapolite group ; the 

 nature of the scapolite series of minerals is evidently to be 

 explained by the isomorphous intermixture of two or more 

 substances ; but the only available theory, that recently pro- 

 pounded by Tschermak (8), according to which they are 

 mixtures of Ca 4 Al Si fl O 2 r, (meionite) and Na 4 Al 3 Si y 24 Cl 

 (marialite), is considerably weakened by the fact that the 

 latter compound is purely hypothetical ; the artificial pro- 

 duction of this compound would of course immensely 



