Both by the overflow from the receiver-pipe B, and by subsequent condensation in the tube d, tar and ammoniacal liquor collect progressively in the cistern or pit under C, by which mingled liquids the lower orifice of the vertical tube f is closed, so that the gas cannot escape into the empty space of this cistern. These liquids flow over the edges of the inner vessel when it is full, and may, from time to time, be drawn off by the stopcock at the bottom of the cistern.

Though the gas has, in its progress hitherto, deposited a good deal of its tarry and ammoniacal vapours, yet, in consequence of its high temperature, it still retains a considerable portion of them, which must be immediately abstracted, otherwise the tar would pollute the lime in the vessel E, and interfere with its purification. On this account the gas should, at this period of the process, be cooled as much as possible, in order to condense these vapours, and to favour the action of the lime in the purifier E, upon the sulphuretted hydrogen, which is more energetic the lower the temperature of the gas. The coal gas passes, therefore, from the tube f into the tube h of the condenser D, which is placed in an iron chest g filled with water, and it deposits more tar and ammoniacal liquor in the under part of the cistern at t, t. When these liquids have risen to a certain level, they overflow into the tar-pit, as shown in the figure, to be drawn off by the stopcock as occasion may require.

The refrigerated gas is now conducted into the purifier E, which is filled with milk of lime, made by mixing one part of slaked lime with 25 parts of water. The gas, as it enters by the pipe l, depresses the water in the wide cylinder n, thence passes under the perforated disc in the under part of that cylinder, and rising up through innumerable small holes is distributed throughout the lime liquid in the vessel m. By contact with the lime on this extended surface, the gas is stripped of its sulphuretted hydrogen and carbonic acid, which are condensed into the hydro-sulphuret and carbonate of lime; it now enters the gasometer F in a purified state, through the pipe p t, and occupies the space q. The gasometer, pressing with a small unbalanced force over the counterweight s, expels it through the main u u, in communication with the pipes of distribution through the buildings or streets to be illuminated.

The parts A B C D E and F, of which this apparatus consists, are essential constituents of every good coal-gas work. Their construction rests upon peculiar principles, is susceptible of certain modifications, and therefore deserves to be considered in detail.

The Retorts.—These are generally made of cast iron, though they have occasionally been made of baked clay, like common earthenware retorts. The original form was a cylinder, which was changed to an ellipse, with the long axis in a horizontal direction, then into the shape of the letter D with the straight line undermost, and lastly into a semi-cylinder, with its horizontal diameter 22 inches, and its vertical varying from 9 to 12. The kidney form was at one time preferred, but it has been little used of late.

The form of retort represented in [fig. 483.] has been found to yield the largest quantity of good gas in the shortest time, and with the least quantity of firing. The length is 712, and the transverse area, from one foot to a foot and a half square. The arrows show the direction of the flame and draught in this excellent bench of retorts, as mounted by Messrs. Barlow.

The charge of coals is most conveniently introduced in a tray of sheet iron, made somewhat like a grocer’s scoop, adapted to the size of the retort, which is pushed home to its further end, inverted so as to turn out the contents, and then immediately withdrawn.

The duration of the process, or the time of completing a distillation, depends upon the nature of the coal and the form of the retort. With cylindrical retorts it cannot be finished in less than 6 hours, but with elliptical and semi-cylindrical retorts, it may be completed in 4 or 5 hours. If the distillation be continued in the former for 8 hours, and in the latter for 6, gas will continue to be obtained, but during the latter period of the operation, of indifferent quality.