In some cases isomorphism takes place to a greater or less extent between substances which are not chemically similar, and this brings about a variation in composition which at times has caused confusion. For instance, the mineral pyrrhotite has a composition which usually varies between Fe_{7}S_{8} and Fe_{11}S_{12}, and both these formulas have been assigned to it. It was recently shown by Allen, Crenshaw and Johnston in the Journal (=33=, 169, 1912) that this is a case where the compound FeS is capable of taking up various amounts of sulphur isomorphously. The idea of solid solution was advanced by van’t Hoff to explain the crystallization of mixtures, including cases of evident isomorphism. This view has been widely accepted, and it has been particularly useful in cases where isomorphism is not evident. Solid solution between metals has been found to be exceedingly common, many alloys being of this character. A case of this kind was observed by Cooke and described in the Journal (=20=, 222, 1855). He prepared two well-crystallized compounds of zinc and antimony to which he gave the formulas Zn_{3}Sb and Zn_{2}Sb, but he observed that excellent crystals of each could be obtained which varied largely in composition from these formulas. As the two compounds were dissimilar in their formulas and crystalline forms, Cooke assumed that isomorphism was impossible and concluded “that it is due to an actual perturbation of the law of definite proportions, produced by the influence of mass.” We should now regard this as a case of solid solution. _A Lack of Confidence in Avogadro’s Principle._—One reason why chemists were so slow in arriving at the correct atomic weights and formulas was a partial loss of confidence in Avogadro’s principle. About 1826 the young French chemist Dumas devised an excellent method for the determination of vapor densities at high temperatures, and his results and those of others showed some discrepancies in the expected densities. For example, the vapor density of sulphur was found to be about three times too great, that of phosphorus twice too great, that of mercury vapor and that of ammonium chloride only about half large enough to correspond to the values expected from analogy and other considerations. Thus, one volume of oxygen with two volumes of hydrogen make two volumes of steam, but only one third of a volume of sulphur vapor was found to unite with two volumes of hydrogen to make two volumes of hydrogen sulphide. Berzelius saw clearly that the results pointed to the existence of such molecules as S_{6}, P_{4}, and Hg_{1}, but it was not generally realized in those days that Avogadro’s rule is fundamentally reliable, and Berzelius himself appears to have lost confidence in it on account of these complications, for he did not apply Avogadro’s principle to decisions about atomic weights, except in the cases of substances gaseous at ordinary temperatures.