passes away to the condensers. The dip pipe previously referred to "dips" into the hydraulic main to a certain distance below the level of the liquid contained therein, and is then said to have a certain "dip," or "seal,” the amount of the latter being usually from 1 to 2 inches.

Some few years ago it was suggested that it would be a decided advantage to dispense with the hydraulic seal, and for two reasons; firstly, in order to prevent the resulting pressure and oscillation which is one of the chief causes of the deposition of carbon on the interior surface of the retorts, and probably of the stoppages which occur in the ascension pipes; and secondly, in order to obviate the injurious effect due to the gas having to force its way through the tar seal in the hydraulic main, which was considered to have an injurious effect on the illuminating constituents of the gas. As a result of the investigations into the matter, numerous devices were brought forward to take the place of the hydraulic seal, such contrivances being known as anti-dips; but, after being thoroughly tested, the general conclusion arrived at was, that the employment of anti-dips was not advantageous, and that the old-fashioned dip was to be preferred, provided that the dip pipe is sealed in ammoniacal liquor, and that the tar as it accumulates is drawn off, as explained further on.

As already stated, the hydraulic main is made of various shapes in cross section, the one shown in the drawing being one of the most common.

The apparatus is now generally made of wrought-iron, on account of its lightness, and is usually constructed out of plate iron of from a quarter to five-sixteenths of an inch in thickness, and 3 × 3 × 3 inches angle iron. The cover, however, is usually made of cast-iron.

In modern establishments the hydraulic main is divided into separate lengths, corresponding to either one or two settings of retorts, each section being provided with its own

valve, and connected into a gas main running behind and parallel to the hydraulic main on the top of the retort bench, the pipe which conveys the gas from the hydraulic main to the gas main being provided with a weir valve, for regulating the level of the liquor in the hydraulic main as shown. In some works the hydraulic main is of such a length that it serves for six or seven settings, the gas in this case being taken off at the end of the hydraulic main. This plan dispenses with the gas main running along the back of the hydraulic main.

The hydraulic main should be of such width in proportion to the area of the dip pipes, and their distances apart, as to provide sufficient liquid for sealing them against the maximum back pressure which can reach them.

As previously remarked, the hydraulic main is usually supported by means of cast-iron standards, or brick piers, which rest upon the brickwork of the retort bench, but this method is not to be recommended, as owing to the unequal expansion of the brickwork when heated, the hydraulic main also suffers unequal expansion, with the result that it becomes out of level; and as it is absolutely necessary that the hydraulic main should remain perfectly level, this is a very serious defect, as it may cause the dip pipes to become unsealed, and thus draw in air. A better plan, and one which is now being generally adopted, is to place the hydraulic main on transverse girders, as shown in the drawing, the latter in turn being themselves held up by the buckstays which brace together the retort bench. This method possesses the further advantage that it prevents the heat of the retort bench from volatilizing the lighter portions of the tar in the hydraulic main, with its accompanying evil of the formation of pitch in the latter.

In connection with the subject of pitch in the hydraulic

main, it may be remarked that in the hydraulic main as usually met with, the liquid contents of the latter consist of a thick pitchy tar, and ammoniacal liquor. The lower portion of this tar is necessarily more or less stagnant, and is subjected at the bottom of the dip pipe to a continuous blast of gas fresh from the intensely hot sides of the retorts. The tar, therefore, is sufficiently heated to boil until every vestige of its fluidity is gone. The effect of this is, that in time the dip pipe is closed by a substance in a condition of viscousness and tenacity the very opposite of that which is required to give efficacy to its action. This throws pressure on the retort, causing loss through its pores, active deposition of carbon on its sides, and a constant choking of the ascension pipe, near the mouthpiece especially, but more or less throughout its length, occasionally even implicating the H-pipe.

What is required is to keep the tar in the hydraulic main in a fluid condition, and Mr. G. Livesey achieves this, by diminishing the bulk of the stagnant tar in the hydraulic main by placing in the latter only sufficient (say from four to six inches) to seal the dip to the required depth; consequently, as this quantity bears only a small proportion to the yield of tar from the gas, it is constantly being pushed forward to the overflow and renewed.

When a hydraulic main is first put to work, it is filled either with water or ammoniacal liquor, but in the ordinary way of working this is soon replaced by tar, generally of a high specific gravity, which, as previously shown, is injurious on two grounds: first, by reason of the tar abstracting a portion of the illuminating constituents of the gas; and, secondly, by causing a heavier seal, which is one of the causes of stopped ascension pipes. It is consequently desirable to provide means for the removal of the heavy tars from the hydraulic main as soon as they are deposited, and thus insure the dip pipes being sealed in liquor.

A method of doing this is shown in fig. 9, as devised by Messrs. R. and J. Dempster, of Manchester. This method possesses the further advantage that it enables the gas to travel to a great extent out of contact with the tar, for it has been observed that if the tar be allowed to travel forward with and cool with the gas, that the latter will suffer a loss of illuminating power by reason of the lighter hydrocarbons being absorbed by the heavier hydrocarbons

contained in the tar; the converse result being the case to some extent if the gas is kept in contact with the tar at a sufficiently high temperature.

With regard to the stoppages in ascension pipes, it is considered that the cause of such stoppages is due to high heats, and to the pressure caused by the hydraulic seal. The remedy is to keep the pipes cool, and to have as light a seal as possible, and the dip pipe to "dip" in liquor only. The ascension pipes may be kept cool by making the front walls of the settings a brick and a half thick,

so as to prevent undue radiation from the latter, and to arrange the pipes so that they project about eight inches from the front of the setting. Another very good plan is to allow a gentle stream of water to drip into a funnel fitted to the top of a goose-neck syphon, attached to that portion of the H-pipe which is connected to the ascension pipe, and a further preventive, and one which should never be omitted, is to use the bent auger, shown on Plate No. 1, before every draw.