The nozzles by which the water flows into the cones are jacketed with the fresh water flowing into the tanks, for the purpose of preventing them from being choked by the ice which would otherwise form there. The cones are closed by hinged lids, which are made perfectly air-tight by means of 2 indiarubber rings, with an annular space between them filled with water. To secure the easy disengagement of the mass of ice when completed, the cones are jacketed with a space into which steam can be admitted. The vacuum is produced and maintained by a compound air-pump. This pump can produce a vacuum of 1/2 mm. of mercury - that is, 1/1500 part of an atmosphere. The large cylinder is double-acting, and has a diameter of 32 in. and stroke of 20 in., hence a capacity each revolution of 875 pints. The exit valves are self-acting; the entrance valves are actuated mechanically. The small cylinder is single-acting, and has a diameter of 8.8 in. and stroke of 8.4 in.; hence a capacity of 14 pints. The normal speed of the pump is 54 revolutions per minute. In beginning to exhaust with this compound pump, very great force would be required, owing to the great size of the large cylinder in comparison with the small.

This inconvenience is avoided by opening a release valve giving communication between the two ends of the large cylinder. When a tolerable exhaustion is attained, this valve is closed, and the pump is worked compound, the large cylinder effecting compression of the residual air or vapour to a certain point, and the small cylinder completing the compression to the tension of the external air. But a small part of the water evaporated passes through the pump; the major portion is absorbed in a cylindrical vessel of boiler-plate. In this, revolves an agitator about a horizontal axis. This vessel is 23 ft. long, and 34 in. diameter, and should be about } full, then holding 5250 pints of acid. This vessel is surrounded by water, to prevent the temperature of the acid from rising too much, but for this purpose no very large supply of cold water is needed.

After 3 turns have been worked, the acid is so far diluted that it is well to change it for that which has been concentrated. The air having been admitted, the dilute acid is allowed to run out into a tank. On again exhausting the cylindrical vessel, fresh acid is drawn up from a chamber. The concentration of the acid is effected in a lead-lined vessel, in which is a coil of lead-piping heated by the admission of steam from the boiler. The pressure in this vessel is kept low by means of an ordinary air-pump. No acid pump is needed, as all the moving of acid is effected by the pressure of the air. A complete cycle of work would be approximately as follows: - Start at 12 o'clock, with acid in the reservoir, having a density of 60° B., that is an actual density 1.70, equivalent to 79 per cent. H2S04 to 21 per cent. water; about 40 minutes' pumping is required to produce the vacuum before beginning to run in the water; the water is then admitted in a regulated stream to the cylinders, the pump is kept at work, and the vacuum is maintained at about 3 mm. of mercury; about 3780 pints having entered the 6 cylinders, the flow is stopped; 792 lb. has evaporated, and 3960 lb. is formed into ice, and passed through the pump or to the acid. The valves between are now closed.

The lids of the cones are opened, and steam is introduced into the jacket. The 6 blocks, of 6 cwt. each, presently fall down. The lids are again closed, the exhaust valves are opened, and the air is pumped out. The process is repeated with a somewhat slower flow of water, as the absorption by the acid is less rapid. A third time is the air pumped out, and 36 cwt. of ice presently produced. But now the acid is becoming too dilute, its density is reduced to about 51.5° B., that is actually a density of 1.55, indicating 66 per cent. H2S04 and 34 per cent. water. It must be changed for that which has been undergoing concentration during the cycle of 3 turns. The cycle just described produces 108 cwt. of ice, and occupies from start to the next start, and allowing time to change acid, between 8 and 9 hours. Therefore the machine will produce between 15 and 16 tons of ice in 24 hours. The consumption of coals is very small, all told less than 8 per cent. of the weight of ice produced; the acid serves over and over again, indefinitely; the wear and tear of the machine so far appears to be, practically, nil.

The success of the machine, as regards durability, is largely due to the fact that the sulphuric acid is never so concentrated as to attack lead, and is never so dilute as to attack brass or iron. The condensed water, both from the vacuum pump and from the concentrator pump, did not reveal the slightest trace of acid. In the early days of the machine, the concentrator consisted of a rectangular cast-iron vessel, without any lead lining; it is astonishing how well the cast iron withstood the corrosive action of the hot sulphuric acid when ordinary air was absent. One would expect that there would be some difficulty in keeping the journal of the agitator of the absorber tight, exposed as it is on the interior to acid. This is effected by placing the external portion of the shaft under water. On a large scale, in continuous work, ice can be made at a less cost than 4s. a ton, allowing 10 per cent. on the plant for depreciation, 5 per cent. for interest, and appropriately for other dead charges and contingencies. Though a perfect pump is a necessary adjunct of the machine, the work is nearly all done by the absorption of the acid.

The indicator diagrams of the engine employed for driving the pumps gave the following results at the various stages: - With no load, 2*15; during the earlier stage of the exhaustion, with the release valve open, 10.1; after closing the release valve, during the process of freezing, 5 * 0 in maintaining the vacuum. As the power varies at different stages from 2 to 10 horse, it is clear that by judicious arrangement, 2 complete apparatus could be driven by an engine capable of indicating 12-horse power. A larger quantity of steam is required for heating the absorber from its coil - about 187 pints of water are condensed per hour. From this figure, and from the indicated horse-power, it is easy to see that the consumption of coal may be made very materially less than 8 per cent. (Dr. Hopkinson.)