is now so well known. Three years later, a machine, working on the compression system and using sulphur dioxide as a refrigerating agent, was introduced by Raoul Pictet of Geneva, and is still in use to a considerable extent abroad. Pictet discarded lubrication in the compressor, and was probably the first to use a water-jacket to absorb some of the heat of compression, not only the compressor itself, but also the piston and piston-rod, being cooled in this way.

The seventies, more particularly in the United States, were pregnant with patents and improvements in refrigerating machinery. It would require too much space to describe the various men and machines at length, but, to select a few out of many, mention must be made of David Boyle, whose career has been ably described by Mr Skinkle, and who is said to have perfected his designs during the siege of Vicksburg; Charles Ball, who considerably developed the absorption machine; Ferguson, who designed the first machines for the Frick Company; and John De la Vergne, who introduced oil-injection, and whose name is now world-known.

THE BINARY SYSTEM.

Yet another system of refrigeration remains to be described. This is known as the Binary or Double Agent System, and the refrigerating effect is brought about either by the use of a liquid, being a compound of two agents, one of which will liquefy at a low pressure and then absorb the other, or by the use of a simple agent which is liquefied partly by compression and partly by absorption.

The principal agent used in the former system is a mixture of ether and sulphur dioxide, it being found, after numerous experiments with other mixtures, that this

combination was the most suitable, as the ether possesses at ordinary temperatures an absorbing power for sulphur dioxide of some 300 times its own volume, while at 60° F. the vapour tension of the compound was below atmospheric pressure. A machine using this compound has been used in the United States by De Mota and Rossi.

A machine using a cycle of the second class was introduced in 1869 in Sydney, N.S.W., by Mort and Nicolle, who have been previously referred to. In this ammonia was used with water as an absorbent. The refrigerator contained a strong solution of ammonia and water, and a pump was used to reduce the pressure in this vessel, and so vaporize the ammonia. The gas discharged from the pump was brought into contact with the weak liquor remaining, and was again absorbed, the resulting strong liquor being once more forced to the refrigerator. The machine, however, did not come into general use.

About the year 1884 Pictet, the inventor of the sulphurdioxide machine, invented a compound agent which is universally known as Pictet's liquid. It was composed of certain proportions of carbon dioxide (CO2) and sulphur dioxide (SO2), the complete formula for the mixture being CO,S. The liquid resulting from this combination has a vapour tension which is much less than CO2 at all temperatures, and for temperatures above 78° F. is below sulphur dioxide itself. The liquid, though it attracted considerable attention at the time of its discovery, has not come into general use as a refrigerating agent.

2

Quite a number of agents have from time to time been proposed, and some used experimentally for refrigerating machines. Among these may be mentioned chloride of methyl, ethyl chloride, ethylic ether, ethylo-sulphurous

dioxide (Mota and Rossi's mixture), and an obscure agent called cryogene. Recently acetylene (C2H2) has been proposed as a refrigerating agent; it is said to require a pressure of 720 pounds on the square inch to produce liquefaction at 32° F. Its intense inflammability would certainly militate against it. Other agents, such as naphtha, gasoline, etc., have been experimented with, more particularly in the United States, but there, as in England and elsewhere, the only two agents which have any extended use and which require consideration are anhydrous ammonia (NH ̧) and carbon dioxide (CO2).

CHAPTER IV.

THE PROPERTIES OF AMMONIA AND CO,.

FOR the convenience of comparison the critical data of the three principal refrigerating agents were given in the Introduction. It will, however, perhaps be useful to repeat those for ammonia as follows:

Ammonia, the agent which, with the exception of carbon dioxide (CO) is now almost exclusively used in refrigerating machines, is a combination of nitrogen and hydrogen corresponding to the formula NH-that is to say, one part of nitrogen in combination with three parts. of hydrogen, or, if by weight, 82-4 per cent. of nitrogen and 17.6 per cent. of hydrogen. It is an alkali, and in its natural state is a gas. At the pressure of the atmosphere it will not liquefy until it is reduced to a temperature of -30° F. Its liquefaction at higher temperatures is, in accordance with the law of gases, merely a question of pressure.

Anhydrous ammonia, when pure, is a colourless liquid with a peculiar caustic taste. It is manufactured from sulphate of ammonia treated with lime and from other bye-products, but chiefly from the ammoniacal liquor given off in the preparation of coal gas. It is also prepared from aqua ammonia of 26° Beaumé, or containing 38 per cent. of ammonia. In the early days it was made directly from its mineral salts, and was absolutely anhydrous. The price of the liquid manufactured in this way is, however, prohibitive.

The process of manufacture is very simple, and is generally similar to the cycle employed by the absorption machine, which, indeed, is sometimes used for the purpose. The operation is as follows:

The material from which the anhydrous ammonia is to be manufactured is placed in a cast-iron vessel termed a still, which is fitted with a steam-coil. These are marked respectively A and B in the diagram (fig. 18). Steam is admitted to the coil, and the heat drives off the NH, leaving the residue behind. The ammonia escapes through the pipe at the top of the still into the vessel C. This is called the separator, and is filled with cast-iron plates. It somewhat resembles the analyzer of the absorption machine, to which, indeed, its function is similar. Any water carried over with the ammonia gas is caught by these plates and falls back into the still A. The ammonia gas proceeds onwards to a condenser worm marked D. This is surrounded by water, which condenses. the gas and also any watery vapour that may have escaped as steam past the cast-iron trays of the separator. The mixture of liquid ammonia and water is expanded into two vessels E and F, commonly called Wolffs' bottles, where it turns into gas and is further purified. These