A certain number of substances are known which apparently do not conform to the laws which have been explained in the last chapter. For example, the compound of ammonia with hydrochloric acid, which has the formula NH4Cl, should have the density 26.75, f°r the atomic weights of the elements it contains are N= 14 ; H=i; CI =35.5 ; and the molecular weight is the sum of 14 4- 4 4- 35.5 = 53.5. But the found density is only one quarter of this number, viz., 13.375. It was at first imagined that this discrepancy was to be explained by abnormal expansion of the gas ; but with such a supposition, of course, Avogadro's law could not hold. Other substances which show the same " abnormal densities " are pentachloride of phosphorus and sulphuric acid. To explain this abnormality, Henri Saint-Claire Deville propounded the idea that such substances do not go into the state of gas as compounds, but that they split into simpler components, each of which has its usual density, and a mixture of the components will exhibit a mean density. Thus, if ammonium chloride be imagined to decompose into ammonia and hydrogen chloride on changing into gas, then the density of the supposed ammonium chloride gas will be the mean of the densities of its two constituents. Ammonia has the formula NH3, and hydrogen chloride, HCl; the former has the density 8.5, and the latter, 18.25 » an<^ tne mean of these two numbers is 13.375. Phosphoric chloride, which has the formula PCl5, splits in a similar manner into PCl3 and Cl2 ; and sulphuric acid, H2SO4, into water, H2O, and sulphuric anhydride, SO3. To this kind of decomposition, where the bodies which are decomposed by a rise of temperature re-unite on cooling to form the origin nal substance, Deville gave the name dissociation. It has been found possible, by taking advantage of the fact that light gases, like ammonia, pass out through an opening, or, as it is termed, " diffuse" more rapidly than heavier gases, like hydrogen chloride, to separate these gases, and thus to prove that they exist as such in the vapour of ammonium chloride; for compounds are not decomposed into their constituents by diffusion ; hydrogen chloride diffuses as such, and is not split into hydrogen and chlorine.

Let us look at this dissociation from another standpoint. We know that if 2 grams of hydrogen, or 32 grams of oxygen, or 28 grams of nitrogen, or, in fact, the molecular weight of any gas expressed in grams, be caused to occupy 22,380 cubic centimeters at o° C, the pressure exerted by the gas will be 76 centimeters of mercury. If the temperature is higher, the pressure will be increased proportionally to the increase in absolute temperature. Thus, suppose the temperature were 300° C, the pressure would be increased in the proportion 2730 Abs. : 573° Abs. :: 76cms. : 160cms. Now, if 53.5 grams of ammonium chloride were placed in a vacuous vessel of 22,380 cc. capacity, and the temperature were raised to 300° C, and if no dissociation were to take place, one would expect a pressure equal to that of 160 cms. of mercury. It has been found, however, that the actual pressure is twice that amount, or 320 cms. In order to account for the doubled pressure, the supposition that dissociation has taken place must again be made ; that is, in order that the pressure must be doubled, twice as many molecules must be present as one would have supposed from the weight taken. The fact of dissociation may accordingly be inferred either from a diminished density or from an increased pressure.