or water pipes to accommodate themselves to the irregularities of underground roadways. A ring or cylindrical casting, K, is made with internal projections, F, at each end, of such size as to pass freely over the outside of the pipe B B. These projections are about or of an inch, and allow the ends of the pipe to move to the inside of the ring at J J, and by so doing the pipe, if 9 feet long,

Fig. 141.-SKETCH SHOWING ARRANGEMENT OF STEAM PIPES IN A SHAFT WHERE THE USE OF
EXPANSION JOINTS IS DISPENSED WITH.

can be placed from 9 inches to 18 inches out of a straight line, as shown in dotted lines at A A, without at all affecting the security of the joints at D D. The joints D D are made by placing india-rubber rings over each end of the pipe B B, and to ensure their having a thoroughly tight grip of the pipe, the india-rubber rings must be smaller than the external part of the pipe B B. The ends of the cylindrical ring K are grooved out at F, and so also are the inside of the loose flanges E E, so that the india-rubber rings C C are kept firmly in place and cannot be blown out by pressure.

After all parts are put in their places, the through-bolts G G are screwed tight, when the joint is completed, ready for use.

C

Any pipe may be used for putting on the expansion-joint by cutting 5 or 6 inches out of the middle, or two old or broken pipes may be taken if one flange of each is good.

Fig. 142.-STEAM-PIPE EXPANSION JOINT
FOR UNDERGROUND ROADWAYS.

The joint being simple is quickly made, as it requires no special connecting-pipes when joining new pipes, and when replacing an old one it need not be exactly the same length as the one removed; the length of the pipe may be varied a few inches by pushing it in or drawing it out of the expansion-ring K.

The feed-water should be conveyed to the boilers hot. The temperature of the water in its passage along the pipes is not high enough to cause inconvenience from expansion and contraction of the pipes, and no expansionjoints are needed. The exhaust steam is often used to heat the feed-water before the latter passes into the boilers, but unless this is done without allowing the steam to mix with the feed-water it is objectionable, as the grease and dirt pass from the cylinders into the boilers and so into the cylinders again and again. There are various methods of utilising the heat in the exhaust steam, most of which consist of passing it through pipes surrounded by the water-tank out of which the feed is supplied or by reversing the arrangement. Formerly feed-pumps were used to force the water into the boilers, and this was frequently done without raising the temperature of the water before doing so. The water entered the boiler through a pipe at the top and extended downwards a little, terminating in an open-mouthed pipe. Where the water has to be brought from a distance or from any point below the level of the boilers, pumps worked by small engines may still do most useful work, but they are not much used to pump the water direct into the boilers. The usual practice is for the pump to deliver the water into a tank placed a little above the level of the boilers, unless the water will reach the tank by gravitation. After the temperature of the water has been raised in the tank, it is conveyed by means of injectors to the boilers; or, if an exhaust injector is used, the temperature of the water is not raised in the tank, but enters this form of injector at 90° F. or lower.

The injector is lighter, occupies less space, and absorbs less power than a feed-pump, and is quite as reliable in action, with the further advantage of coming into action only when required.

Injectors may be divided into two classes, viz., live-steam injectors and exhaust

steam injectors. The action of both is somewhat strange, and has puzzled many engineers. In the former the water is forced into the boiler by the pressure of live steam direct from the boiler acting on the surface of the water in the injector, under certain conditions. In the latter the same object is attained by the use of exhaust steam, also under certain conditions.

M. Giffard invented the injector many years ago, and it was introduced into this country by Messrs. Sharp, Stewart & Co., whose works were formerly at Manchester, but have been removed to Glasgow. This form of injector is still the most generally used, but it is not applicable to all purposes. Although Giffard's injectors are stated to draw water from the supply, they all, but more especially the small sizes, work better when placed below it, so that the water may run to the injectors by gravitation. The pressure of steam to work the injector must not be under 5 lbs., nor must the pressure against the water being forced into the boiler be more than twice the pressure working the injector. It cannot be employed when the temperature of the feed-water is above 135° F. for low pressures and 105° F. for the highest pressures, as a part of its efficacy arises from the condensation of the steam. If therefore these temperatures are exceeded, a large volume of water is required to condense the steam, and the effect of this is to reduce the velocity of the steam in driving forward the large body of water. To work efficiently under the conditions named, the regulators require very careful adjustment.

The following particulars are furnished by the Patent Exhaust Steam Injector Co., Limited:

"The Giffard injector was the first ever made, and the improved form of the original instrument, shown in Fig. 143, has been proved by experience to be a trustworthy apparatus, simple to work, and not liable to get out of order. Having self-contained steam and water regulators, it will work at a greater range of steam pressure than any other class. It may be fixed either above or below the level of its supply water, and the fixing can be done by any local engineer or mechanic.

"Among its advantages are the following:

66

1. Small first cost, the absence of all working parts in motion, and the consequent small repairs.

2. It supplies water to the boiler without the necessity of working the steam engine, and in a continuous stream instead of intermittently as with a pump.

3.-The steam admitted to the injector is condensed and re-enters the boiler with the feed-water, thereby raising its temperature, so that only hot water enters the boiler, and the straining of plates consequent on unequal expansion when pumping in water at a low temperature is avoided.

4. The consequent saving of the power required to work pumps, as well as of their wear and tear, which, in the case of donkey-engines especially, is considerable.

5.-The freedom from risk of damage or stoppage by frost.

Very many thousands of these improved injectors are at work all over the world, and the makers confidently recommend them as being in all respects most complete and efficient, and certain to give the greatest satisfaction to steam users in all countries. Having all the progressive number books in their possession, exact duplicate injectors of these and all other classes made by Sharp, Stewart & Co., Limited, can only be obtained from their sole successors, The Patent Exhaust Steam Injector Co., Limited, 4, St. Ann's Square, Manchester. "Instructions for Fixing.-The injectors may be placed horizontally or vertically, with connecting pipes. There must be a continuous supply of dry steam, with a cock or valve on the steam-pipe to the injector. The water-supply must

be continuous, and should not be hotter than 135° Fahrenheit with low pressure, or 105° Fahrenheit with the highest pressures. The delivery to boiler is of course

much hotter than this.

Fig. 143. THE GIFFARD INJECTOR.-CLASS A.

The injector may be placed either above or below the level of the water-supply; the small sizes will draw from 2 feet to 5 feet, and the

larger ones 12 feet. There must be a back-pressure valve on the delivery pipe to the boiler. The water-supply pipe must be perfectly air-tight. A cock or valve is only required on this pipe when the water flows to the injector, and has to be shut off when not working.

"Instructions for Working.-Open the cocks on the steam- and water-supply pipes. Open the water regulator and slightly open the steam regulator until water issues from the overflow, after which, open it fully and the water will be forced into the boiler. If all the overflow is not taken up when the steam regulator is fully opened, adjust the water regulator to stop this. The water regulator should be turned to its full extent occasionally, to prevent its setting fast. The quantity of water delivered to the boiler may be reduced when the injector is working, by turning down the steam regulator until there is an overflow, and taking this up as usual by the water regulator.

"Steam from the boiler passes through the top branch pipe where shown on the drawing, and is admitted to the injector through a conical nozzle. The lower end of a vertical spindle fits accurately into the nozzle, and an ascending or descending motion may be given to this spindle by means of the screw actuated by a hand-wheel at the top. The admission of steam may be thus regulated. The feed-water enters the injector on the opposite side from the steam and flows through a branch pipe at a slightly lower level to the steam-pipe branch, as shown in the Fig. It passes round the outside of the conical nozzle to meet the steam which rushes in with great velocity; the steam is condensed and a partial vacuum formed, into which the following steam rushes with great velocity, to be in its turn condensed, but the water is carried along with it, and the energy is so great that the water is forced into the boiler notwithstanding the pressure against it. The water supply is regulated by the hand-wheel shown at the side. It works a small pinion inside the injector, and, on being turned, moves a tube up and down. Provision is made by means of a branch pipe placed below the level of the feed-water branch, for the overflow water. A back-pressure valve is placed at the bottom of the injector, and on the side of the boiler not shown in the Fig. a check-valve is placed, which prevents the escape of water when the injector is not in action."

The difficulty of understanding the action of the injector arises from the fact that the steam is able to drive the water it meets in the steam-cone into the boiler against a pressure equal to or greater than its own.

The explanation given in Prof. Jamieson's Text-book on Steam and SteamEngines is as follows:-If two jets issue under the same pressure, one of steam and the other of water, the velocity of the former will be many times greater than that of the latter, and if steam, in its passage from a boiler, is condensed to water, but not reduced in velocity to that of water issuing under the same pressure, it is then able to overcome the pressure of water in its own boiler. In the Giffard injector, the meeting of the steam and water in the nozzle causes condensation of the steam without its losing its velocity, except that due to the friction of the passages. The partial vacuum, caused by the condensation, is followed by more steam, which rushes in with great velocity to meet more water in the nozzle; condensation follows, and the feed-water in advance is carried on by the force of the condensed steam-jet into the boiler. The velocity of the steam-jet is necessarily reduced by imparting a high velocity to this volume of water, but it is not reduced nearly so low as that of a jet of water issuing under the same pressure, and hence it is capable of overpowering and forcing back the water in the boiler.

There are many forms of live steam injectors, in which improvements have been attempted, the greatest of which is for the apparatus to work when placed above the level of the water-supply. The most that has been attained hitherto in this direction does not exceed 20 feet.