20 feet in length, 18 feet in width, and 3 feet in depth, with a center seal, while our carbureted water gas was passing a set of two boxes, 20 by 20 by 6 feet, with the Morava valve system. In the latter we had been compelled at times, when the consumption of gas was at its maximum, to use a layer of lime in the upper tray; and it was the use of this material we desired most to avoid, next, of course, to the fact that we desired to supply a quality of gas which would cause

no just complaint from our consumers. In these deep purifiers we used three sets of trays, and above each of these we had a test-cock connection. During the past winter the four new boxes have been used for purifying our water gas, while the two deep boxes have been reserved for contingencies which might arise, being connected so that either coal or water gas can be turned into them at any time. When the last set of purifiers was erected, openings for extra 8-inch

connections were left in one bottom plate of each for revivification purposes; and as the other boxes were idle, from time to time openings were made in them with the same intention.

A blower was purchased from the Detroit Blower Company; alterations were made for the purpose intended, and connections made so air could be drawn through any one of the ten purifiers as required. Instead of employing valves for the purpose of shutting off the boxes when not undergoing revivification, water seals have been used between each purifier and the main connection with the blower. With valves a sediment invariably collects in the seats and prevents closing the gates down tightly. After the gas is turned into a purifier and revivification of the foul oxide of another is commenced, a leak of gas in sufficient quantities to be very dangerous is quite liable. The heated impure gases about the blower might cause an expensivè explosion, the damage resulting being such as to greatly inconvenience the plant for some time to come, as such an occurrence would be most apt to assert itself at a time when the consumption of gas was at a maximum and the purifying department doing its greatest work. With a water seal of the pattern I shall describe, better satisfaction is to be derived, as the slightest leak in the seal is easily detected and there is no possibility of the gas passing beyond the inlet of the seal and accumulating about the blower. It would be difficult to imagine a scheme more perfect and at the same time more simple, and the description which follows I have attempted to make as clearly comprehensible as possible.

An 8-inch perpendicular pipe drops from the bottom plate of each purifier through and 8 inches below the cover of the seal box, which is constructed of cast iron of an inch in thickness, the dimensions being 27 inches in length, 12 inches in width and 12 inches in depth. The distance between the inlet and outlet center is I foot, and the outlet merely screws into the cover of the seal. One foot above the cover a 1-inch wrought iron overflow screws into the outlet, extends horizontially 6 inches, bends downward 2 feet, and after running a few inches in a horizontal direction turns upward 2 feet 2 inches, and by a downward return overflows into a funnel connected by a 2-inch wrought iron pipe, with a sewer. The siphon described has a cock which can be shut off when the water has been run into the seal, or at any other time as required. The water inlet is a 14-inch pipe entering the box at the bottom of the end next the 8-inch outlet, and can be shut off or turned on without the workmen leaving the seal. From the bottom of the seal the water can be drawn off when revivification is commenced by opening a 2-inch cock on a line which screws into the bottom shell, and communicates with the 2-inch wrought iron pipe receiving the overflow. It can be readily seen by the drawing and description that this affords sufficient seal for all practical purposes and I am satisfied it is less dangerous than a valve system.