by the opening of the stop valve or by the automatic action of a float connected to the throttle valve by a chain or rope.

Pulsometer Pumps.-The very first steam pump, which was invented by Thomas Savery in 1698, had no working parts except the valves. This type has been revived for certain kinds of work in pumps of the pulsometer class, of which Bailey's "Aqua Thruster" is a good example. It consists of two long chambers, in each of which there is a valve opening upwards at the bottom, and one opening outwards at the side. At the top junction between these two chambers there is a flap valve which can put either in communication with a steam pipe while the other is shut off therefrom. Now, suppose the right-hand chamber to be full of water while the left one is full of steam, and that the upper valve is in the position shown. Steam will enter the right-hand chamber and force the water out through the delivery valve at the side. At the same time the steam in the left compartment will be condensing, and water will therefore rise into it through the bottom valve, provided the apparatus be not too far above the free surface of the water. The inertia of this water will cause it to continue in motion after all the steam is condensed, and it will therefore compress the air that remains to a sufficient extent to shift over the valve to the other side.* If there is no air, then the water itself will strike the valve and knock it over to the other side. The conditions of the chambers are now interchanged. Water will be forced out from the left one, and fresh water will rise into the other, and the process begins again.

A large loss occurs in this kind of pump through the condensation of steam during the down stroke of the water, and also owing to the fact that the steam is used non-expansively. To reduce the former loss little cocks open into the top of the chambers and admit a little air during the time there is a vacuum inside. This air prevents the steam from coming so quickly into contact with the water as it otherwise would do, and thus reduces the loss during admission. A slight escape of steam takes place

This is not the common explanation of the working of the pulsometer valve. It is-"As soon as the water is lowered below the upper surface of the delivery valve, steam blows through with some violence and causes a commotion and a rapid condensation in the chamber. The valve is then drawn to the right-hand side." This is quite wrong. The valve can only be shifted by being pushed, owing to the pressure on the closed side becoming greater than that on the other side, and it is difficult to see how the pressure in a chamber in direct communication with the boiler can become less than that in one where the steam is already all condensed, and where the pressure is considerably below that of the atmosphere. Besides, it is probable that in steady working the water never gets as low as the delivery valves.

through these cocks, as they are always kept open; but as their bore is so very small, the loss is less than the gain. To diminish the latter loss an extra self-acting valve, called the "grel," has been added to some forms of pulsometer with the intention of cutting off the steam earlier, and then using it expansively.

Pumps of this class are exceedingly handy for dealing with dirty water and for temporary purposes owing to their simplicity, few working parts, and the ease with which they can be erected. It is sufficient to suspend them by a chain and connect them by a pipe to a portable boiler. A suction pipe projects down below the water surface, and a flexible hose pipe will carry off the discharged water. The full page illustration shows the "Aqua Thruster" in use for pumping water from a dock during its construction.

Roots' Blower.-A form of rotary pressure pump, known as Roots' Blower, is used for obtaining a blast of air at a moderate

pressure, and for pumping liquids. Two vanes rotate inside a closed casing, and sweep the fluid round with them. They are connected by spur wheels outside so as to be always at right angles to each other, and they have such a shape that practically nothing is carried backwards at the central part of the machine. They can produce a higher pressure than an ordinary blowing fan, and are handier for many purposes than a blower of the cylinder and piston type. The student should note that this is not a centrifugal pump or fan, although there are no reciprocating parts, but simply a rotary form of pressure pump.

If a fluid be forced through this machine, then it will cause the vanes or teeth to rotate; hence it will work as a motor, and there

fore it is a reversible machine. Many ingenious attempts have been made to produce economical steam engines on this principle; but largely owing to the difficulty of keeping them tight, they have not been so successful as their sanguine inventors expected.* Bramah's Hydraulic Press.-This useful machine was invented by Pascal, but he could not make the moving parts water-tight. Bramah, about the year 1796, discovered a means by which this difficulty was effectually overcome; and thus the instrument has been handed down to us under his name. As may be seen from the following figure, it consists of a single-acting force pump in connection with a strong cylinder containing a plunger or ram, which is forced outwards from the cylinder through a tight collar by the pressure of the water delivered into the cylinder from the force pump.

After what has been written about force pumps, we need not particularise about this part of the machine, except to say that the suction and delivery valve boxes at SV and DV can be disconnected from the pump, and the valve cover-checks removed at any time for the purpose of examining the parts, or of regrinding the valves into their seats. The pump plunger PP, extends through a stuffing-box and gland filled with hemp packing, and is guided by a centrally bored bracket bolted to the top flange of the

See Lecture XXXV. for Centrifugal Pumps.