168 



SCIENCE 



N. S. Vol. XXX. No. 762 



each case the amoimt of spinel formed was 

 quantitatively proportional to the excess of 

 oxygen in the original silicate over the 

 orthosilicate ratio. These reactions are, 

 furthermore, applicable to the analysis of 

 mixtures of talc and serpentine. It is only 

 necessary to ignite the mixture, and then, 

 with sodium carbonate to dissolve out the 

 liberated silica of the talc, and with dilute 

 hydrochloric acid the soluble olivine from 

 the serpentine. Such an analysis is of 

 course only approximate, but it is better 

 than none at all. 



In another group of experiments Clarke 

 and Steiger^^ attacked the silicate problem 

 by a different method. It was found that 

 certain silicates when heated with dry am- 

 monium chloride to its temperature of dis- 

 sociation reacted with it, forming deriva- 

 tives of unexpected stability. Analcite, 

 NaAlSiaOe-HjO, heated with the reagent in 

 a sealed tube, yielded sodium chloride, and 

 the silicate NH^AlSiaOe, which was stable 

 at 300°. Leucite, KAlSijOe, similarly 

 treated, gave the same derivative, thus 

 establishing a structural relationship be- 

 tween the two species. By means of this 

 reaction it became possible to determine, 

 quantitatively, the proportion of either 

 mineral in an igneous rock. The rock 

 powder was heated with ammonium chlo- 

 ride, and then leached with water. The 

 amount of ammonium fixed in the residue 

 gave a good estimate of the amount of 

 analcite or leucite in the rock. Natrolite 

 gave with ammonium chloride another sili- 

 cate, (NH4)„Al2Si30io, also quite stable, 

 and other zeolites were capable of partial 

 transformations. Going further, Mr. Stei- 

 ger, by fusing the same minerals with silver 

 nitrate or thallium nitrate, succeeded in 

 substituting the alkaline bases by the two 

 heavy metals, producing silver aluminum 

 and thallium aluminum silicates identical 



^= Bull. 207, U. S. Geological Survey. 



in type with the original compounds and 

 with the ammonium salts. These reactions 

 are effected with great ease and open up a 

 new line of attack upon the general prob- 

 lem of silicate constitution. In short, the 

 silicates have been found to be chemically 

 more plastic, that is, more open to meas- 

 urable transformations than they were 

 formerly supposed to be; a conclusion 

 which is evidently of considerable theoretic 

 importance. A new field of research has 

 been opened, but its full extent is yet un- 

 known. 



From data such as these, and from the 

 natural occurrences, associations and alter- 

 ations of minerals, some progress has been 

 made toward a theory of the silicates. At 

 least, some relationships are now estab- 

 lished, which can be rationally expressed 

 by constitutional formulae. Formulae of 

 that kind are easily written when one does 

 not go beyond simple empirical composi- 

 tion, but unfortunately they can be written 

 in several different ways. Each substance 

 must be studied in its relations to other 

 substances before a formula of real signifi- 

 cance can be devised. When that is done a 

 system of formulae develops which becomes 

 a useful tool in later investigations. 



I can not, in a paper of this scope, enter 

 into details. I can only give a brief indi- 

 cation of the theory which has gro\^Ti out 

 from the observed facts. First, the natural 

 silicates are considered as definite salts, nor- 

 mal, acid or basic, of relatively simple 

 silicic acids. Second, many silicates are 

 easily interpreted as substitution deriva- 

 tives of normal salts. For example, the 

 normal orthosilicate of aluminum is repre- 

 sented by the formula Al^C 8104)3. This 

 compound is not known to exist by itself in 

 nature, but many minerals are easilj^ inter- 

 preted, at least stoichiometrically, as de- 

 rivatives of it. Thus we have 



