CHAPTEE IV. 
THE ORTHOSILICATES OF DYAD BASES. 
Although the orthosilicates of the dyad metals are presumably sim- 
pler than those of aluminum, the problem of their constitution, studied 
in the light of mineralogical evidence, is peculiarly difficult. Starting 
points exist, in the salts of magnesium, iron, manganese, zinc, and 
glacinum, but the derivatives are fewer than in the case of aluminum, 
and the evidence upon which to base argument is proportionally limited. 
Expressed in the simplest terms, the normal orthosilicates of this 
group are represented by the general formula E 2 Si0 4 . To this type 
the following minerals correspond: 
Forsterite Mg 2 Si0 4 
Fayalite Fe 2 Si0 4 
Tephroite Mn 2 SiG 4 
Willemite Zn 2 SiG 4 
Phenakite Gl 2 Si0 4 
Monticellite CaMgSi0 4 
Knebelite MnFeSiO, 
Between these minerals there are many intermediate species or varie- 
ties, which may be either isomorphous mixtures or double salts repre- 
senting polymers of the fundamental type. Thus, chrysolite or olivine 
may be a mixture of forsterite and fayalite, or, in the case of hyalosider- 
ite, a salt of the formula Mg 4 Fe 2 (Si0 4 ) 3 . So also, allied to knebelite, we 
have igelstromite, Fe 4 Mn 2 (Si0 4 ) 3 : and in trimerite we find the salt 
Gl 3 Mn 2 Ca(Si0 4 ) 3 . 
Upon studying closely the derivatives of these normal salts, the 
assumption of polymerization seems to be necessary. If the theory of 
substitution is valid, then the existence of polymers must be taken for 
granted ; and upon this basis it becomes possible to develop a system 
of formulae which satisfies all the conditions imposed by the evidence 
now at hand. For some of the species already mentioned the degree 
of polymerization is difficult to determine, and synthetic investigations 
seem to be needed. In other cases the problem is comparatively simple, 
and the indications as to the true formulae are apparently clear. For 
instance, a good example is furnished by the chondrodite group, for 
68 
