Comparing these results for nitrous oxide with those for carbonic acid found by Dr. Andrews, we find the compressibility of the two gases nearly the same at temperatures equidistant from their critical points. At the temperature of 25° 16, liquefaction begins under a pressure of 57.83 atmospheres; at 32°-28, the gas passes into the liquid state under a pressure of 67.45 atmospheres: at this point a great diminution of volume occurs, but not abruptly as in the case of carbonic acid; this must be ascribed to the presence of a greater quantity of a permanent gas in the nitrous oxide.

In the liquid state, nitrous oxide yields as much to pressure as carbonic acid; the rate of expansion by heat will be therefore very great. This is a confirmation of the results of Drion (Ann. de Chim. et de Phys. t. lvi. p. 37), that the coefficient of expansion of volatile liquids at a temperature still below the critical point grows equal to the coefficient of expansion of gases and increases further, till at the critical point it may attain to a value any number of times greater than that of air.

At temperatures above the critical point, the volume of nitrous oxide diminishes with tolerable regularity with increase of pressure, though much faster than according to the law of Boyle; the higher the temperature the more the compressibility approaches to that of a perfect gas. When the gas is reduced to the volume at which it might be expected to liquefy, no trace of liquid is to be seen, the whole mass of the gas remaining homogeneous; but a rapid diminution of volume occurs from a small increase of pressure: this diminution of volume is not abrupt as in the case of liquefaction, and diminishes greatly at higher temperatures.

The anomalies presented by nitrous oxide were :

1. Under a given pressure and temperature the volume of the compressed gas is variable, or vice versa. This anomaly is very obvious in that condition of matter where a rapid diminution of volume occurs at a small increase of pressure; under a given volume of the gas the difference of pressure can amount here to 2 atmospheres, in the other cases this difference is very slight, about 0.2 to 0-4 of an atmosphere. This appears from the following results::

2. The pressure required to liquefy the nitrous oxide and the volume of this gas at the beginning of liquefaction are variable.

The pressure required to liquefy the gas at 25°-15, recorded in Table I., is the mean of the following observations:

--:

The following series of experiments was performed in the course of a

At 32-2 the greatest difference of pressure amounted to 2 atmospheres, as appears from the next series of experiments.

3. After liquefaction has begun an increase of pressure of 16 atmospheres or more is required to liquefy the whole mass of the nitrous oxide; for at 25° 17 liquefaction began at a pressure of 57.83 atmospheres, whilst the whole was liquid at a pressure of 73.68 atmospheres. At 32°-2 I found the commencement of liquefaction at a pressure of 67.63 atmospheres, and the termination at a pressure of 84-09 atmospheres. For carbonic acid, that was mixed with to Too of air, the increase of pressure amounted to 1.5 atmosphere. Had the gas been pure no increase of pressure could have occurred. This shows that a greater quantity of a permanent gas must be mixed with the nitrous oxide; the variations of the volume of the gas under a given pressure and temperature result perhaps from its whole mass not being homogeneous, as the diminution of the volume is too fast to allow a perfect diffusion of the two gases.

The gas used for these experiments was prepared from pure nitrate of anmonium. The salt was carefully heated in a tin bath in order to prevent any decomposition of the liberated gas by a too irregular heating when directly exposed to a flame. It was washed by transmission through a strong solution of caustic potash and dried over sulphuric acid. The caustic potash decomposes any solid particles of the salt that might be carried over mechanically and retains the nitric acid, whilst the free ammonia is absorbed by the sulphuric acid. Purified in this manner, the gas was made to pass through the glass tube wherein it was to be compressed. A pressure of about 90 to 100 millims. of mercury was required to maintain a moderate current of gas through the capillary bore: this current was continued for five hours or more in order to ensure the complete removal of the air; the capillary end was then sealed and the other end introduced under mercury. As the experiments with the tube filled in this manner indicated always the presence of a permanent gas, I tried afterwards to remove the air by exhausting the tube with the air-pump and then to fill with the gas; this operation was successively repeated from twenty to thirty times, but with no other result.

As I could not get the gas pure by heating nitrate of ammonium, I tried to get it from liquid nitrous oxide as it is made in iron bottles in London; it was probable that the permanent gas would escape first and the nitrous oxide remain pure. This, however, did not occur, and I got nearly the same result as before.

In order to prevent diffusion as much as possible, all the caoutchouc joints were besmeared with a solution of tar and asphalt, and the current of gas issued under sulphuric acid. The amount per cent. of this permanent gas was determined in the following manner :-The absorption-tube of Bunsen's absorptiometer was partly filled under water with nitrous oxide and then left standing three days or longer. The whole of the gas was not absorbed: