which shall express its percentage composition [C=12, O=16, H=1], is H,CO; the other possible formulæ are H,C,O,, HC,O,, HC,O, &c., &c. Now if we tabulate the observed. densities of acetic acid vapour at different temperatures, and compare them with the densities of the hypothetical compounds H,CO &c. we have this result.

Density of vapour of acetic acid at 760 mm.

The progress of the change represented by the varying value of the density is very gradual; we cannot therefore suppose that each temperature on the table is marked by the presence of molecules of a definite weight, and by the presence of these molecules only. The process represented by these numbers is analogous to the processes of dissociation which we have considered; the explanation in terms of the kinetic theory of gases has already been given (par. 101, pp. 207-8), and has, I think, been shewn to be fairly satisfactory'. If that explanation is accepted it follows that the density of any chemically homogeneous vapour obtained by heating a liquid substance should not attain a constant value until the gas has been raised some degrees above the boiling point of the liquid. The temperature-interval through which the gas must be raised will vary, according to the nature of the gas, and of the liquid from which it is obtained. This theoretical deduction has been verified by experiments so far as these

1 See in connection with this subject O. E. Meyer's Die Kinetische Theorie der Gase, pp. 76–82.

have yet extended'. Numbers have been given (par. 101, p. 209) which shew that constant values are obtained for the densities of the easily gasifiable elements chlorine and bromine, only at 150° or 200° above the boiling points of these bodies.

The numbers representing the density of the gas obtained by heating phosphorus pentachloride which are given in par. 101 (p. 204) shew, that even at about 30° above the boiling point of this compound the observed density is approximately 30 per cent. less than that calculated from the formula PC,. Wurtz' diffused the vapour of this compound, at temperatures a little below its boiling point, into a flask containing air; by determining the weight of the mixed gases (air and vapour of phosphorus pentachloride) and the volume of each, it was possible to calculate the partial pressure to which the latter was subjected. Some of Wurtz's results are given in tabular form; the volumes are stated in cc. and are reduced to 0° and 760 mm.

The density of the gas obtained by vapourising phosphorus pentachloride under these conditions is only about 8 per cent. less than that required by the formula PCI,.

Wurtz then diffused the vapour of the pentachloride into a flask containing a known amount of phosphorus trichloride vapour; from the sum of the weights of the two gases, and from analyses of the contents of the flask, he was able to calculate the volume and weight of the vapour from the pentachloride, and the pressure to which that vapour was subjected in the flask. As the mean of 12 experiments, at temperatures ranging from 160° to 175°, and pressures varying from 168 to 413 mm., Wurtz obtained the number 723 as representing the density of the vapour of phosphorus penta

1 See, for illustrations, Naumann's Thermochemie, pp. 155-8.

2 Compt. rend. 76. 601.

chloride, obtained by gasifying this substance into an atmosphere of phosphorus trichloride.

Hence there seems to be little doubt that the so-called anomalous vapour density of phosphorus pentachloride is indicative of the dissociation of molecules of PCI, into molecules of PCI, and Cl.

There has been a great deal of discussion within recent years as to the action of heat on chloral hydrate. Does the vapour obtained by heating this substance contain chloral and water, or is it composed of chloral hydrate? The following results were obtained by Naumann'.

Hence chloral hydrate appears to undergo complete dissociation into chloral and water at 78°, and under so small a pressure as 162 mm.

These numbers were not accepted by Troost, Berthelot, and others, as conclusive, because these chemists are opposed to the conceptions of modern chemistry which are founded on the distinction between atoms and molecules. As this distinction is an outcome of the application of Avogadro's law to chemical processes, it is evident that, could this law be overthrown, the distinction in question would appear to be less well grounded. Now if it could be proved that the density of a homogeneous gas is only half as great as the number calculated by the use of Avogadro's law, an important step would be made towards overthrowing the law in question.

It has been shewn that the vapour obtained by heating chloral hydrate diffuses as a mixture of chloral vapour and water and not as a homogeneous gas2; further that chloral vapour is not hydrated in the vapour obtained by heating chloral hydrate, provided the pressure of the former is greater

1 Ber. 9. 822.

2 Wiedemann and Schulze, Wied. Ann. 6. 293.

than the equilibrium-pressure of the latter at the temperature of experiment1; and also that the vapour in question behaves towards a hydrated or dehydrated salt in the same way as a gaseous mixture containing water. Moreover Engel and Moitessier3 have shewn that when chloral hydrate is distilled with chloroform at 60o, the distillate contains both water and chloral. Naumann has also proved that when chloral hydrate is distilled alone, the distillate contains much chloral and a little water, and the residue much water and a little chloral*. Finally the densities of the gases obtained by heating chloralalcoholate and butylchloralhydrate shew that these compounds, which are analogous to chloral hydrate both in composition and function, undergo dissociation when heated. Hence there can be no doubt that the compound CCI,.CH(OH), is dissociated, even at low temperatures and small pressures, into CCl2.COH + H„O2.

5

171. The characteristic features of dissociation, as contrasted with decomposition, have already been summarised (par. 169). It is not possible however to define the term dissociation. There are many changes which present more or less close resemblances to well-marked processes of dissociation. Sometimes the resemblance is so close that we have little hesitation in classing the actions under the heading of dissociation; sometimes it is impossible to place the phenomenon entirely in the class of dissociation or in that of decomposition. Thus when hydrogen and water-vapour are passed over a mixture of iron and magnetic oxide of iron, it is found that a state of equilibrium is maintained for any given temperature, and that this equilibrium is independent of the relative quantities of the two solids present, and is conditioned only by the pressures of the water-vapour and

1 Naumann, Thermochemie, 136.

2 Wurtz, Compt. rend. 84. 977: 86. 1170: 90. 118, 337, 572.

3 Compt. rend. 88. 285: 90. 97.

4 Ber. 12. 738.

5 Wurtz, Compt. rend. 85. 49.

6 Engel and Moitessier, Compt. rend. 90. 1075.

7 For a fuller discussion of the dissociation of chloralhydrate, see Naumann's Thermochemie, 134—137.

hydrogen, which pressures are always in the same ratio to each other as long as the temperature remains unchanged1.

Although there is here a more complex series of changes than in the cases of dissociation hitherto studied, yet because the change is brought about by raising the temperature, because the amount and direction of the change are dependent only on the temperature and the pressure of the gaseous components of the changing system, and because the change is reversible, we are entitled to class the reaction in question as a dissociation-phenomenon.

172. It has been experimentally proved that many salts— e.g. ammonium salts, many of the alums, cobalt chloride, sodium sulphate, &c.—are partially resolved into their constituents when dissolved in much water. Further, it has been shewn that the amount of this decomposition is dependent on the temperature and the relative masses of water and salt; and finally that the change can be, at least partially, reversed by lowering the temperature of the solutions'.

The action of the water in these changes has been compared to the action of the pressure of the gases produced in a process of dissociation. Dilution may thus be regarded as analogous to decreased pressure. But the analogy is misleading. In many if not all cases of so-called dissociation in solution, water is itself one of the components of the original substance; hence, judging from the analogy of gaseous dissociation, we should expect that as this product of the change accumulates the process would become slower and would eventually stop. But we find that increasing the quantity of water acts in the same way as decreasing gaseous pressure. The water probably exerts two actions; one, which may be called physical, whereby an increase in the quantity of water gives greater freedom of motion to the particles of the dissolved substance, and also lessens the chances of combination between the separated components of this substance; and

1 Deville, Compt. rend. 70. 1105: 71. 30.

2 Examples will be found in Watts's Dict. Supplt. 2. 292 et seq. See also for the case of iron sulphate, chloride and nitrate, G. Wiedemann, Pogg. Ann. 126. I: 135. 177.