Integral Weights in Chemistry. 323 



mative specific gravities got with water at 4° (or better at 15°) 

 as an available basis for fixing the integral weights of most 

 solid species. 



9. We are thus led to the doctrine of high integral weights 

 and of polymerism for liquid and solid species, suggested by 

 Favre and Silbermann in 1847 from their thermo-chemical 

 studies, and by Graham in 1849 from his researches in diffu- 

 sion. In 1.853 this view of polymerism was further discussed 

 by the present writer, who then proposed for the carbon spars, 

 regarded as polycarbonates, provisional integral weights, vary- 

 ing with their densities from about 1500 to 2500. The subject 

 has much more recently engaged the attention of Guthrie, 

 of Spencer Pickering, and especially of Louis Henry in his 

 inquiry into the Polymerization of Metallic Oxyds*. Studies 

 of the ammonio-cobalt bases, the formulas of which require 

 integral weights of not less than from 500 to 2500, have 

 thrown further light upon the subject ; and especially the late 

 researches of Wolcott Gibbs on the salts of what he has de- 

 signated " The Complex Inorganic Acids," among which he 

 has made known polytungstates the simplest formulas of which 

 lead to integral weights of not less than 5000, and in one case 

 of 20058. The high numbers thus deduced for the last two 

 classes of bodies represent, however, the simplest possible 

 integral weights, like that for water-vapour ; while the species 

 themselves, as known to us, are derived from these theoretical 

 species by what is called polymerization, or, more correctly, 

 by homogeneous integration, being, like water, and like ice, 

 examples of metamorphosis by condensation. 



10. As we calculate the integral weights of all gaseous 

 species by a comparison of their densities with that of hy- 

 drogen gas, so we may calculate those of liquid and solid 

 species, either by comparing their densities directly with that 

 of hydrogen or else with that of water at 100° ; remembering 

 that this has an integral weight of 29244, and that these den- 

 sities are subject to corrections for expansion, as already set 

 forth. The integral weights of these species being thus 

 known, and also those of the simpler species from which they 

 are formed by homogeneous integration, it is obvious that by 

 dividing the first by the second we get the coefficient of con- 

 densation. Thus, calcium carbonate, CCa0 3 = 99*89 ; and as 



* This paper, which in an English translation is published in the 

 Philosophical Magazine for August 1885, first appeared in 1879 as "Etudes 

 de Chimie Moleculaire, l re partie, les Oxydes Metalliques," in the Annales 

 de la Societe Scientifique de JBruxelles, 



