88 THOMAS STERRY HUNT ON A NATURAL SYSTEM IN 



Uiiigstico-tuugstate ol' sodium has a specilic gravity of (J. 617, while ibr two allied tungstic 

 compounds, the one potassic and the other sodic, are given the numbers V.OO and '7.28, 

 showing that these are similar in condensation to the anhydrous calcic and ferrous tnng- 

 states, scheelite and wolfram. For the soluble and hydrated polytungstates, Scheibler 

 found the sjiecific gravity of -tWO^. Na.O + lOAq. to be o.98Y, while that of the corres- 

 ponding barium salt, with 9Aq is 4.298, and that of 14W0,. 6Na,0 + 32Aq. is 3.846.' The 

 density of the complex hydrous phospho-vanadio-tungstate of barium described by Gribbs 

 (§ 114) with a unit-weight of 20058, is unknown, but, if we assume for it the uiimber 

 found by Scheibler for the hydrous barytic metatuugstate. we have a unit-volume of not 

 less than 4666, or more than three times the volume of the cobalt salt. What greater 

 unit-volume than this can be determined remains for the chemist to discover, biit it 

 should be considered that such elevated unit-weights as that last mentioned are not 

 readily attained, save with a few elements of high atomic weight, such as tungsten, 

 vanadium and barium. Moreover, with elements of low^er atomic weight the difficulty 

 of fixing definite formulas for their compounds is notably increased, so that the remark- 

 able results obtained by G-ibbs may, for a long time to come, mark the limits of the 

 chemist's skill in this direction of research. 



§ 125. If then, as we have argued, the molecules of mineral species are so complex, 

 and their minimum molecular weight is so large that their volume may be represented by 

 a sum not less than about 4666, or more probably some multiple thereof, it follows that a 

 silicate like pyroxene, for the simplest atomic unit of which we have found a volume of 

 about 5.5, must include in its molecule not less than 84S such atomic units, and wollas- 

 tonite, with a corresponding volume of 6.6. about 700. But as our simplest atomic formula 

 for these species embraces three of these units, (m|Si2)o;,, it will be evident that with the 

 molecular volume of 4666, here assumed, the constitution of pyroxene may be represented 

 very nearly by 282(ni,si^,)()„ and that of wollastonite by 236(m,si;)o,. In like manner, 

 the atomic volume ot the feldspars, anorthite and albite, approximately re^Dresented by 6.3, 

 is contained in our assumed molecular volume 742 times. The simplest atomic formula 

 for anorthite being (ca,al.jsi,)o„ and for albite (na,al;jsi,2)o,„, the constitution of anorthite 

 may be represented by 92(ca|al.,si,)o, and that of albite by 46(na,alssi|^)0|,i ; while for ortho- 

 clase and microcline, with a unit-volume of 68, we deduce a formula of 42(kial.jSi,2)o,„. 



It will be evident that attempts like these at molecular formulas are of value only so 

 far as they serve for illustration, since the unit-volumes assigned to the various species 

 are but approximations, and the molecular volume, 4666, which has been assumed, is 

 based on a supposed specific gravity, and can only be conjectured to be not far from the 

 truth. A series of careful studies of the specific gravities of various salts of the complex 

 inorganic acids may furnish us wath more trustworthy data for similar calculations. 

 Meanwhile, it is to be repeated that the formulas here given for pyroxene, wollastonite, 

 and the feldspars, are of value only as they serve to illustrate our conception of the 

 complex constitution of these silicates. For the purposes of compari.son, and for the 

 elucidation of polymerism and homologies, the unit-volumes which we have calculated 

 for the preceding tables of species of the different tribes of silicates, serve every pur- 

 pose, and show in a simple manner the relative condensation of the molecule in the 



' See Constants of Nature, by F. W. Clarke, Part i. 83. 



