2140. For circulating water below the level of the boiler, it is necessary, in the first place, to raise the water, as it is heated, to as great a height above the boiler as it is proposed to circulate it below. This may be done in any closed boiler, with a tube proceeding from its cover; or in any boiler of a height above the fire, equal to the depth below it, to which it is intended to circulate the water. This mode will be very readily understood by a reference to the apparatus of Mr. Weeks, explained by the following diagram (fig. 634.), In this figure, a is a section of the fireplace, which, instead of having a common boiler over it, is surrounded by cast-iron tubes. The uppermost of these tubes communicates, by means of the upright tube b, with the open vessel c; and the lowermost is connected, in a similar manner, by means of the tube d, with one or any number of tubes, under the level of the boiler at e. The uppermost of the tubes, at e, is connected by the tube f with the open vessel c. Now, all these tubes being so connected as to admit of water circulating freely through them, when a fire is made in a, the heated water ascends by its rarefaction into the open vessel c; and its place in the tubes round the fire is supplied by the colder water from e, through d; the heated water descending to supply its place from the open vessel c, by the tube f. The limit of the depth to which the water will circulate below, is that of the height of the open vessel above. To produce this circulation, it is not necessary that the water should boil; for, as every heated particle will ascend to the open vessel c, its place must be supplied by a cold particle from d. When the fire is urged so as to raise the water in the open vessel nearly to the boiling point, the circulation goes on with the greatest rapidity. The substitution of tubes round the fire for a boiler over it, is by no means necessary for the success of this plan, though by tubes the rapidity of the circulation is greatly increased. Any close boiler with the tube b attached to its cover, and communicating with an open vessel fixed at any height, such as c, having another tube, similar to f, affixed to it, will circulate the heated water from such vessel to a point below the bottom of the boiler, nearly equal to the balance of atmospherical pressure, or say 30 ft. below it. Messrs. Cottam and Hallen, and Mr. Timothy Bramah, have both circulated hot water on a large scale upon this principle.