84 THOMA'S STERRY HUNT ON A NATURAL SYSTEM IN 



state ; and if it should hereafter be found that its density corresponds to six times the above 

 formula, it would follow that liquid acetic acid, whose density differs bvit slightly from 

 that of fused stéarine, must have a formula and an equivalent weight about one hundred 

 times that which we deduce from the density of acetic acid vapor, CjHjOj." 



§ 114. In the papers of 1853 the equivalent or molecular weights assigned by me to 

 mineral species were confessedly arbitrary. Taking as a term of comparison the formulas 

 of such well known species as the alums, the orthophosphates of soda, ferrocyanid of 

 potassium, lactose, the compound of glucose and sodium-chlorid, (the empirical equivalents 

 of which were doubled to compare with that of piperine,) minimum molecular weights 

 were adopted for the various polycarbonates and polysilicates which it was sought to 

 compare in volume with the species just named. Thus it was that in the notation of the 

 time the formulas and molecular weights of the rhombohedral carbon-spars, smithsonite 

 and calcite were represented respectively by CJZn^^Oy^, = 2500, and C.j,,Cai,,0., ,, = 1500 ; and 

 those of the polysilicates, spodumene and diopside, by si^i(aL,4Li,Na,,)09o = 1400, and 

 si52(Cai3Mg,3)0-H = 1404. The question of the relations of these provisional formulas and 

 provisional weights to the still higher and as yet unknown values which belong to these 

 species was left untouched. These silicates and carbon-spars are, so far as known, incap- 

 able of any chemical combinations or decompositions but such as affect their more or less 

 complete resolution, and at the same time afford no guide to the molecular constitution 

 of the species. 



§ 115. We have already noticed the partial adoption of the notion of polysilicates 

 in 1860 by Ad. Wurtz, who did not, in any way consider the great problem of their 

 molecular weights, or their specific gravities. Meanwhile, however, the arguments 

 adduced by me in 1853 in favor of high molecular weights have been strengthened by 

 discoveries which serve to show the wide application of the analogies then drawn from 

 the chemistry of the carbon series. First among them may be noticed the various artifi- 

 cial crystallized cobalt compounds, such as the potassio-cobaltous nitrite, to which was 

 assigned a formula with CoJv,; and N,., and a unit-weight of 958. More remarkable 

 still are the ammonio-cobalt bases, studied by Frémy and other chemists, but made more 

 completely known by the elaborate researches of Wolcott Gribbs and F. A. Genth in 185'7. 

 Among these may be noted the chlorid of purpureo-cobalt of the latter chemists, tetragonal 

 in form, with a specific gra^dty of 1.802, which includes in its formula, with Co^, not less 

 than Clg and N,„, and has a unit-weight of 501 ; the chlorid of luteo-cobalt of Frémy, 

 clinorhombic in form, with specific gravity 1.101, which with the same proportions of 

 cobalt and chlorine contains not less than Nio, and has a lanit-weight of 535 ; and, 

 finally, the orthometaphosphate of luteo-cobalt of Braun, for which is deduced a formula 

 including Co,, and N^c, with P,,,, giving a unit-weight of 2540.' These numbers, like 

 those deduced for alums, ferrocyanids, sugars, and alkaloids, represent the weight of the 

 chemical units of which the species are supposed to be in all cases polymères, the molecular 

 weights of which are as yet unknown. 



§ 116. Still farther light has, however, been thrown on the subject of polybasic 

 salts of great complexity by the studies of the compounds of tungsten and molybdenum. 



' See Gibbs and Genth, Smithsonian Contril)utions, ix. 1S57, and Amer. ,Iimr. Siicme, xxiii-xxiv; also farther 

 Gmelin-Kraut, Handbucli, iii. sub met " Kobalt." 



