The mean value of c-c, from the three figures of the last line, 0.0661, is very near to that found above, 0.0645, and the heat of vaporization 97.877 to 97.866.

The following table of the vapour-pressures of benzene is calculated from the values of x and y determined above.

Owing to the irregularities in the values of x, I took the arithmetical mean of three of its values in order to obtain a more reliable result. This may give rise to a mistrust in the figures subsequently determined. In order to test to what

*See Notes at end.

extent they deserve confidence, I recalculated the observations through intervals of 10°.

The minimum value of a lies between 80° and 90°. But as the values of a at these temperatures are nearly equal, it is best to take their arithmetical mean and to take the corresponding temperature as 85°. We then have

The value now found for x is near to that given above, and the values obtained for y by both methods can be considered as perfectly equal. The pressures calculated from these new values of x and y are given in the fourth column of the preceding table; they differ but little from those in the second column.

Bisulphide of Carbon, CS2.

12. 1st Method.

85°

10° 55° 70° 75° 80° 105° 2.4771 1-7183 1.6498 1.6206 1.6398 1-6515 1.8626

≈ has a minimum value 1·6206 at 75°. Hence

According to Hirn, at 80° c=0·25531; according to Regnault c=0.15956; hence c-c=0.09575. The latter figure is more than twice that found above-the difference is very considerable; nevertheless there can be no serious objection to such a discordance, for it has only to be admitted that there is an aggregate error of 0.05 in the specific heats, and the figures found for c-c1 will both closely agree. Conversely, calculating c1 from c and a we find that c=0.21266 instead of 0.15956.

According to equation (9) we find

68

K. D. Kraevitch on an Approximate Law of the According to Regnault it is 76-43. A certain disagreement proceeds from the fact that a considerable error might enter into the calculation of the amount of heat required to raise the temperature of the liquid from 0° to 85°, because Regnault's observations did not exceed 39°.5.

The mean value of c-c1 0·0425 and the value of ro 85·926 do not differ from those found by the 1st method, 0·04265 and 85.901.

...... 5.2217 6:4753 6.2296 6.5014 5.8202 6.8851 6-0289 6.2210

The figures as we see proceed irregularly, but there is hardly any doubt that the maximum lies at 10o.

It is impossible to compare this figure with the result of experiment, because the specific heat of the vapour of carbon tetrachloride is not known. But it is possible to conversely calculate this quantity, knowing c-c1 and c. According to Regnault c=0.19979, therefore c=0.11037*. By equation (9) we obtain

whence

y=4061.3,

c-c.......

53.295 52.865 52.467 52.063 51.654 51.264 50.870 0.430 0.398 0.404 0.409 0.390 0.394 0.0860 0.0796 0.0808 0.0818 0.0780

0.0788

* See Notes at end.

The values of c-c1 proved to be not quite similar to each other, or to the figure found by the 1st method, probably because the vapour-pressures, taken at rather low temperatures, are small and the errors they contain have a great influence on the calculation. The value of ro, 52-063, however, is similar to the former value.

In order to verify the results, I repeated the calculations by the 1st method through intervals of 10°, between 10° and 40°.

This table removes all doubt as to x attaining a maximum at 10°, hence

which is somewhat different from that found above, 0.08942. The same value is obtained for y as before, 4061.2. The vapour-pressures calculated from these data are given in the fourth column.

Ethylene Bromide, CH4Br2.

14. 1st Method.

In the case of ethylene bromide the calculations presented great irregularities, with which it was exceedingly difficult to deal. By means of numerous attempts and recalculations, I came to the conclusion that a has a minimum value of nearly zero, at 95°. The pressure at this point is equal to 245.51. Below this temperature, the errors contained in the vapourpressures are probably considerable compared to the true