local corrosion, caused, perhaps, by galvanic action. In their stead light iron cover plates, filled with cement, are bolted over these seams, as shown in fig. 75. They were very efficient in stopping leaks where caulking was unavailing. Of course if the shell plate is seriously weakened it is necessary either to fit an efficient doubling plate, or the defective part must be cut out, and a proper patch put in its place; but it is very seldom that this occurs, except perhaps with donkey boilers, whose bottoms cannot be properly protected from moisture.

It is different with the end plates of main boilers; they are exposed at one end to the hot and moist ashes, and at the other to leakages from flanges and manholes, &c. As these plates are originally very thick, and more than sufficiently supported, their wasting away is not so much a danger as a nuisance. Most of the patches to these parts are therefore simply doubling plates over the thinnest places, and it is only when no other remedy is thought effective that the bad parts are cut out.

The following are a few sketches of repairs which may occasionally be necessary. Figs. 76, 77 show a case in which the

circumferential seams at the front ends of the shell and the furnace have wasted away and have been repaired by a covering patch. As it forms a ridge on the inside of the furnace, against which the rake will always be knocking, it is better to carry out this repair as shown in figs. 78, 79. Part of the furnace bottom is cut

If only the front plate is wasted away, and if this is only between the two furnaces, an internal patch can be fitted, as shown in figs. 80, 81. Generally a manhole or a few large stays will be found here, and as it is impossible to get a sufficiently large patch into the boiler to cover them, it will be necessary to fit it from the outside.

When the lower front seams both of the furnaces and the shell are in a bad condition, it may be necessary partly to renew the front plate,

and to make its flanges sufficiently deep to take in another row of rivets. The joints should be placed where there is sufficient room for working at them. Lap joints, as shown in fig. 83, are very difficult to

make, for if looked at from the back (fig. 82), it will be noticed that the patch has to be fitted under the remaining plate, which cannot be done efficiently.

A simpler plan is to butt the plates at these points, and to fit flanged butt straps internally (fig. 84), or, where the space is too narrow, a solid block (fig. 85). In either case this part has to be carefully fitted after all the other seams have been riveted up. Of course with the solid

block it will be necessary to use screwed studs and bolts, except perhaps for those points marked c, for which through rivets may be used, but only if the holes come opposite each other. The flanges should

not be cut away parallel to the axis of the boiler and furnace, but should slant downwards (see dotted line, fig. 86); this will enable the

plate edges to be driven tight together.

Before deciding which plan to adopt, a sketch should be made to see whether it is possible to insert the various rivets. Much space is gained if the rivet holes are countersunk inside so as to do away with the heads, and outside, so as to be able to use short rivets (see fig. 87).

Evaporation and Circulation. Some information on this subject, particularly on the distribution of evaporation in a boiler and its efficiency, will be found in the chapter on 'Transmission of Heat.' Here attention will be drawn, not so much to the economical side of the question as to its practical bearings on the life of a boiler.

FIG. 87

Formerly, when there were almost more boiler explosions on steamers than ashore, while neither iron nor boiler-makers would admit that it was their fault, speculations about retarded ebullition and similar subjects were freely indulged in. The summary of about twenty papers, dating from 1786 to 1864, is contained in C. Tomlinson's paper (Phil. Mag., 1869). The Franklin Institution made some experiments to show that the sudden opening of the stop valve of an overstrained boiler led to its destruction. Mr. L. E. Fletcher made experiments which were published as a Report on a Series of Red-hot Furnace Crown Experiments by the Manchester Steam Users' Association, Manchester, 1889 (Charles Sever), which showed that showering cold water on red-hot furnace crowns did not lead to fractures, nor did it lead to sudden generation of steam, a danger which until then was believed to be real. A related subject is that known as the action of water hammers, which is met with chiefly in steam pipes, and was evidently the cause of the explosions on the Elbe' and at Deptford. A considerable amount of information on this subject will be found in Clark and Colburn's paper, Am. M. E.,' vol. iv. p. 404.

Priming. Information on this subject is as yet almost non-existent. The generally accepted views are that low-pressure boilers prime far more than high-pressure ones, and therefore require more steam space, or perhaps more steam height; also, that throttling the steam at the main stop valve reduces or stops priming, and that the injection of oil, particularly mineral oil, is a still more efficient remedy. Soda and salt seem to increase priming, probably by forming soaps with the oils. The action of mineral oils may be compared to the action of oil on troubled waters: it prevents the formation of light bubbles. That the partial closing of the boiler stop valve has an effect in reducing priming, while linking up the engine, so as to use the same reduced amount of steam, has not, clearly points to the necessity of imparting violent motion to the steam, or rather to the as yet uninjured froth, so as to burst the bubbles.

J. T. Thornycroft (C. E.,' 1890, vol. xcix. p. 41) argues this point very clearly, and has effectively demonstrated that a very small steam space will suffice, if only the mixture of steam and water is

A

properly guided. In his water-tube boiler he admits this mixture at the upper circumference of the dome (fig. 88), and dashes it against an internal baffle plate, whereupon the water falls to the level W and the steam is carried off through the pipe A. This small dome separated sufficient steam for 774 I.H.P., equal to about 10 cubic ft. of steam per second, while the internal diameter was only 26 ins., and the clear height from water level to crown of baffle plate only 15 ins. An ordinary double-ended marine boiler of the same power would require at least 48 ins. of steam-space height. It must, however, be admitted that in his more recent practice the steam drums are larger than they used to be.

FIG. 88

That this large space is necessary need not be wondered at; for, in spite of the wide water space between the nests of tubes, which are primarily intended to act as downcast shafts, facilitating circulation, they are very far removed from doing their work properly. Perhaps

о

FIG. 89

FIG. 90

more than one-third of the total evaporation takes place at the furnace crowns, but instead of allowing the steam to ascend to the water level as freely as possible, by removing a central row of tubes (fig. 89), it has to rise as best it can (fig. 90), and naturally struggles towards the water spaces, where it comes in conflict with the downward current. By closing up all but the wing water spaces, this struggle can be made to grow so severe that the imprisoned steam raises the upper water

E

level, as, for instance, in locomotive and navy boilers. The consequence is that certain parts of the heating surface are more often exposed to steam than to water, and even though they may not get burnt, are occasionally ineffective.

FIG. 91

A half-hearted attempt to direct the current of steam and water is sometimes made by bolting a few plates horizontally to the lower row of steam-space stays; but to be really effective they ought to cover the whole water level, with the exception of the wings, and no downward currents should be allowed at the centre water spaces.

In some boilers flat sheet-iron tubes are fitted to wings, reaching to the boiler bottoms, but it is to be feared that the inducement for the water to circulate through them, while there are wider water spaces on either side, is not a very strong one. Undoubtedly, however, the water which does enter them must fall below the furnaces.

Improvements in circulation, though not a remedy against priming, could sometimes be effected by removing a central row of tubes from each nest, and securing vertical plates to the outside rows of tubes, which would prevent any steam from entering the water spaces. By placing horizontal plates over each nest of tubes (fig. 90) the separation of steam and water would also be effected; but all loose parts in a boiler are a nuisance.

The notions as to what takes place inside a boiler are exceedingly vague, and it may be of interest to draw attention to a few points.

Take, for instance, a boiler 15 ft. diameter and 11 ft. long. Its steam space will be about 4 ft. high and have a capacity of 450 cubic ft. It will be capable of evaporating about 10,000 lbs. of water per hour under natural, and double that quantity under forced draught. The total area of the water level is about 150 sq. ft., so that under forced-draught conditions every square foot of this water level emits 133 lbs. of steam per hour, or s lb. per second. At 180 lbs. working pressure this is equal to a volume of 1 cubic ft., or 144 cubic ins., per second, while at atmospheric pressure the volume would be more than 10 times as great. But even 144 cubic ins. of steam per square foot of water surface is a large quantity, amounting to about two bubbles 1 in. in diameter per second per square inch of surface, or 2,000 per second if only 1 in. in diameter. No wonder, then, if the water contains any frothing substances, that the bubbles will not burst till they are carried into the steam pipe, particularly as it takes less than forty seconds to remove all the steam contained in the steam space.

It is well known that it is quite impossible to make soap bubbles in an electrified atmosphere, and it is very surprising that in the days of low pressures no attempts were ever made to turn this knowledge to practical account.

Attempts to measure the amount of priming water were first made