brushes (fig. 20) are used for the soot and ashes, and the split scrapers (fig. 21) to remove salt. The latter tool is of little use at sea, because a tube which is once plugged up with salt closes up again after half an hour's steaming, and need therefore not have been cleaned. The brush should be strong enough to remove all the dust. These tools are hinged at their centres for convenience of handling. The sweepings of each tube naturally fall into the combustion chamber, but the draught at once carries the greater part into the adjoining tubes and up the funnel. In order not to get smothered the firemen keep the damper open, and very light fires on the grates, just sufficient to produce a draught. Except in large steamers, or in vessels fitted with several funnels, it is usual to clean the tubes of all the boilers at

FIG. 22

one operation, the whole watch being sent below for the purpose; for, as the draught is nearly non-existing during this time, little steam can be generated, and the sooner the operation is completed the better.

The efficiency of a boiler is seriously affected by accumulation of soot on the heating surfaces, so much so that all land economisers are fitted with scrapers which are constantly moved up and down the tubes by machinery. It is found that whenever this motion ceases the efficiency of the economisers is much reduced. Soot is also a great inconvenience to water-tube boilers, and may one day lead to similar appliances being employed on them: at present their tubes are mostly cleaned by steam blasts, which can of course only remove the loose soot, whereas it is believed that the hard tarry deposit nearest the metal is a most efficient non-conductor.

Furnace Bridges. The object of fitting bridges in furnaces is partly to keep the fuel from falling into the combustion chambers, and partly to reduce the air channel, so that, by imparting a momentarily high

FIG. 23

velocity to the flame, a thorough mixture of its gases is effected, and the combustion completed. The same result is aimed at in Mr. Holt's boiler (fig. 22), in which the bottom of the combustion chamber is raised above the centre of the furnace, and the two are connected by a comparatively small air tube. One disadvantage of this arrangement is, that no alteration can be made in the size of this air passage, and if the draught is insufficient to overcome the various resistances,the consumption cannot be increased. Brick fire-bridges are arranged as

FIG. 24

FIG. 25

FIG. 26

shown in figs. 23-26. The space at the back of the bridge shown in fig. 23 is usually quite full of ashes, even though a door is fitted at the bottom by which they could be removed.

C

The heat which the firebricks collect may occasionally do a considerable amount of harm to the plates against which they rest if, as sometimes happens, the water is let out of the boiler before it is cold.

Viewed from the front, the bricks are generally placed horizontally; but where much trouble is experienced by their falling out of place, which happens when clinkers are allowed to adhere to them and have to be broken, or when the fires are forced, or when the furnaces are large, the bricks are sometimes arranged in the form of an inverted arch (fig. 27). In double-ended boilers, with through combustion

FIG. 27

FIG. 28

FIG. 29

chambers, the same plan has been tried on the central partition wall; but no advantage seems to have been gained, and where it is desired to carry the wall high up, it is safer to bolt two angle irons down the sides of the combustion chamber.

In cases where the furnace saddles have been cropped, or where the tube plate has been flanged to meet the furnace, exposing the back end seam to the direct action of the flames, it is sometimes necessary to protect it by a firebrick arch (fig. 28). Here, too, it is an advantage if the bridge is curved, as it gives a better support to the arch. (See fig. 29.)

The Influence of Heat and Cold on Boiler Plates. In the chapter on Strength of Materials' will be found information on the effect of repeated coolings and heatings of metals, which seems to show that this treatment can make iron and steel brittle. It also seems to be a fact that it tends to alter the shape of structures. For both reasons it is desirable to guard against such changes as much as possible, which means that the periods during which the fire doors are kept open should be few and short.

Dampers. An alternative is to have a damper fitted in the smoke box for each combustion chamber, and while opening any fire door the respective damper could be partially closed. They can be arranged as in fig. 30. Something of this sort, but stronger, is desirable with forced draught in closed stokeholds, for the pressure is there much greater than with natural draught, and an enormous amount of air escapes when the doors are opened. But there are also strong objections against fitting any dampers in these cases.

If the forced draught is used with closed ashpits, dampers are also necessary, for the suction above the bars is not always equal to the pressure which is transmitted from below. In fact, when firing, it is customary to restrict the blast, because otherwise so much gas is

generated that the funnel draught is unable to carry it off, and the fumes would enter the stokehold.

Flames in Stokeholds.-The most serious mishaps may be caused if steam is suddenly generated in the combustion chambers, which is a powerful reason why fusible plugs should not be fitted. With some boilers in the Navy sad loss of life has several times been caused by sudden leakages of the back ends of the tubes. It has been suggested that this happened with H.M.S. 'Barracouta,' but a similar effect might be produced by stopping the fan in one of the stokeholds, the flames being at once driven into it over the bridge by the fan in the other stokehold, while the funnel is unable to draw them away. The difference of pressure would no doubt be increased by men anxious to render assistance blocking the stokehold passages. Similar accidents have also happened in single-ended and in navy-type boilers, in which this action is impossible.

FIG. 30

Water Gauges. Of the various boiler mountings, one of the most important is the gauge glass, and it ought always to be in good working order, and capable of being tested at any time. The material and construction should be of the very best. Asbestos packed cocks are a great advantage, but even with them, if the gun metal is bad, leakages cannot be prevented. Iron and steel plates are attacked by acids, whereas copper alloys suffer severely when in contact with caustic soda; care should therefore be taken never to make the boiler water too alkaline.

In some cases the gauges are fitted direct to the boiler plating, but this is not a good plan; not only does the panting of the plates lead to breakages of the glass tubes, but the scum cannot be kept out of the glass, and the oscillations due to the rolling of the vessel and to irregularities of ebullition make it difficult to obtain good readings.

The usual plan is to connect the gauge glass to the top and bottom of the boiler by means of long pipes, the one reaching well up into the steam space, and the lower one into the water, where little

circulation is expected to be found. These pipes can be fitted either externally or internally. The latter arrangement is not always possible, on account of the smoke box, and some engineers object to it in toto, on the principle that no copper pipes should be fitted inside a boiler. From the following table it will also be seen that, if not often blown through, gauges with long connections may indicate a different water level from that actually existing in the boiler. When fitted externally these pipes have the effect of cooling the water contained in them, and as its density is thereby very much increased, the indicated water level is lower than that in the boiler. Matters are still further complicated by the accumulation of distilled water in the upper parts of the connecting pipe, which, if there is much salt in the boiler, has the effect of making the water level appear higher than it really is. The possible differences caused by a connecting pipe 10 ft. long are shown in the following tables, which contain the corrections which should be applied to the water-gauge readings under the various conditions.

The first of these tables has been calculated on the assumption that the rate of expansion of water is independent of the pressure; but even if not strictly true, both tables show that a serious difference may exist between the gauge-glass indication and the water level, unless the cold or distilled water in the connecting pipe has first been blown out. Few gauges are so fitted that this can be readily done (see fig. 31), for if the water is low, the opening of the cock D, while B is closed, will not necessarily clear the pipe C D of its water, or at any rate not as quickly as if the cock A had been closed. This being inconvenient, some gauge stands are fitted with a large cock at F.

No double bends, capable of collecting water, should be allowed to exist in the gauge steam pipe, as they would cause the glass to show a wrong level.

Test Cocks are intended to indicate the boiler water level. They are often fitted to the gauge-glass stand, but their proper position is on the boiler plating. One should be at or near the ordinary water level, a second a few inches higher, and a third on a level with the highest heating surface; for, should the water have been lost out of the gauge glass, it is very important to know whether a mishap can still be averted by putting on all available feed, or whether it is necessary to draw the fires. It is not always well to check the fires, for the