and their Records, and on, Inspection as a Means of Prevention, by Edward B. Marten, mem. inst. m.e. a.i.c.e., excerpt Minutes of Proceedings of the Meeting of the Institution of Mechanical Engineers, at Manchester, 1st August, 1866, Joseph Whitworth, Esq., President, in the Chair. By permission of the Council.

---

The subject of Steam Boiler Explosions, which was brought before this Institution in June, 1848, in a paper by the late Mr. William Smith of Dudley in reference to an explosion near that place, and again in 1859 in a paper by Mr. Longridge on the economy and durability of stationary boilers, is one of great importance and is now attracting increased attention. The first public notice of the subject was by a parliamentary committee in 1817, which was appointed in consequence of a very fatal boiler explosion in London in 1815; evidence was then collected as to steamboats, and many boiler explosions were referred to. That committee recommended among other things that boilers should be made of wrought iron, instead of cast iron or copper, which had been the materials mainly used previously; that they should be inspected and tested; and that there should be two safety valves, each loaded to one third of the test pressure, under penalties for any excess. A great part of the information now existing upon the subject, especially in regard to the earlier explosions, is to be found in the records of inquests after fatal cases; and some of the careful reports of eminent engineers on those occasions have materially assisted in the formation of correct views as to the causes of explosion. Latterly also the printed reports of the inspectors of mines, and more especially the reports of the explosions of locomotives, illustrated by diagrams by the inspectors of railways, have furnished very valuable information. Since the subject has been taken up by private associations for the prevention of explosions, many more records have been published, although their usefulness is much impaired by their not containing the names of the places whereby the explosions could be identified.

When the writer's attention was first directed to this subject, he met with great difficulty in obtaining correct records of boiler explosions, from which to arrive at the results of past experience; and wishing to base his own opinion on facts, rather than on the inferences of others however reliable, he followed the example of the Franklin Institute in their elaborate investigation of the subject, and collected all the records he could find; and by way of facilitating reference, arranged an index, a manuscript copy of which is presented with the present paper to the Library of this Institution. All must be agreed as to the importance of reliable information on such accidents as boiler explosions; and the writer would suggest that this Institution may materially aid in obtaining the desired records and placing them within easy access, by becoming the depository of reports on explosions, and by inducing those who have the opportunity to forward copies of reports, that these may be arranged so as to be easily found and consulted. It is very desirable that these reports should as far as possible be illustrated by sketches, as aids to the description; and also by slight models like those now shown to the meeting, by which the whole matter may be seen at a glance. So few persons comparatively have the opportunity of examining boilers after explosion, that the most erroneous ideas have prevailed, and theories have been advanced which would soon be dissipated by practical experience or by reading accurate reports. It would also very much aid in the understanding of published matter on the subject, if full descriptions of each case alluded to in illustration could be obtained. These records are as useful to the engineer as the "precedents" or "cases" to the lawyer or the surgeon. After any serious explosion, the newspapers of the neighbourhood in which it has occurred contain voluminous articles describing the disastrous result and the damage done, which, although useful as far as they go, do not in the least assist in arriving at the cause of explosion. The really important particulars, such as the description and construction of the boiler, its dimensions, and the pressure at which it worked, are in most cases omitted altogether.

The record of explosions presented to the Institution contains a list of the boiler explosions in each year of the present century, as far as known to the writer, with the names of the places, and the description and sizes of the boilers, and the supposed cause of explosion, together with references to the books or papers from which further information may be obtained. Of course many of the explosions have to be put down as uncertain in some of the particulars; but every year improves the record, as fresh information is obtained, and with the assistance of the members of this Institution it might be made far more perfect and extensive.

---

The total number of explosions here recorded is 1046, and they caused the death of 4076 persons and the injury of 2903. The causes assigned for the several explosions are very numerous, and are no doubt incorrect in many cases; but they may be generally stated as follows:

397	are too uncertain to place under any heading; but of the rest
145	were from the boilers being worn out, or from corrosion, or from deteriorated plates or rivets.
137	from over pressure, from safety valves being wedged or overweighted, in some cases intentionally, or from other acts of carelessness.
125	from faulty construction of boiler or fittings, want of stays, or neglect of timely repair.
119	from collapse of internal tubes, generally from insufficient strength.
114	from shortness of water, or from scurf preventing the proper contact of the water with the plates; or from improper setting so as to expose the sides of the boiler to the flame above the water line.
9	from extraneous causes, such as effect of lightning striking down the stacks upon the boilers, or from fire in the building or explosion of gas in the flues.
1046	total number of explosions.

The exploded boilers were of the following descriptions:—

232	are not sufficiently described to place under any head; but of the rest
320	were Marine boilers of various forms.
141	were Cornish, Lancashire, or other boilers internally fired.
120	were Locomotive, or other multitubular boilers.
116	were plain Cylindrical boilers, externally fired.
64	were Balloon or haystack, wagon, Butterley, British-tube, elephant, or Trevithick boilers.
29	were Portable, agricultural, upright, or crane boilers.
14	were Heating apparatus or kitchen boilers.
10	were Upright boilers attached to puddling or mill furnaces at ironworks.
1046	total number of explosions.

The theories as to the causes of explosion have been numerous. In the early days of the steam engine, when the steam was used only as a condensing medium and the pressure in the boiler was frequently allowed to get below atmospheric pressure, many boilers were destroyed by the excess of the external atmospheric pressure becoming too great, causing them to be collapsed or crumpled up; and this led to the use of the atmospheric valve still found on old boilers. Even so lately as last year, 1865, a boiler in the neighbourhood of Bury, Lancashire, has suffered in this way by collapse from external pressure; its appearance after the accident is shown in Fig. 1, which is copied from a photograph. The early explosions were so palpably due to the weakness of the boilers, which compared with those of the present day were most ill constructed, that no one thought of any other cause than the insufficient strength of the vessel to bear the expansive force of the steam contained in it. When the advantages of high-pressure steam became recognized, and the boilers were improved so as to bear the increased strain, the tremendous havoc caused by an explosion led many to think that something more must be required than the expansive force of the steam to produce such an effect; and they appear to have attributed to steam under certain conditions a detonating force, or a sudden access of expansive power that overcame all resistance. To support this somewhat natural supposition, it was asserted that the steam became partially decomposed into its constituent gases, forming an explosive mixture within the boiler. That this belief is still sometimes entertained is seen from the verdict of a jury even in the present year, 1866, in the case of the explosion of a plain cylindrical boiler at Leicester, shown in Fig. 2, the real cause of which appears to have been that the shell of the boiler was weakened by the manhole. It seems hardly necessary to point out the fallacy of imagining decomposition and recomposition of the steam to take place in succession in the same vessel without the introduction of any new element for causing a change of chemical combination; but it is necessary to refer to this supposition, as the idea is shown to be not yet extinct.

Again it has been asserted that the steam when remaining quite still in the boiler becomes heated much beyond the temperature due to the pressure; and that therefore when it is stirred or mixed or brought more in contact with the water by the opening of a valve or other cause, the water evaporates so rapidly as to produce an excessive pressure by accumulation of steam. In support of this view the frequency of explosions upon the starting of the engine after a short stand is adduced; but it is very doubtful whether by this means a sufficient extra pressure could be produced to cause an explosion, unless the boiler had been previously working up to within a very small margin of its strength. Explosions are seldom caused by a sudden increase of pressure, but rather by the pressure gradually mounting to the bursting point, when of course the effect is sudden enough. Nor is it necessary in many cases to look for much increase of pressure as the cause of explosion; for it is far more often the case that the strength of the boiler has gradually degenerated by wear or corrosion, until unable to bear even the ordinary working pressure. It is so very easy, when examining the scene of an explosion, for the first cause of rupture to be confounded with the causes of the subsequent mischief, that in many cases erroneous conclusions have been arrived at in this way.

The most important points to find out in connection with any explosion are the condition of the boiler and all belonging to it immediately before the explosion, together with the locality of the first rent, the direction of the line of rupture, and the nature of the fracture; as everything occurring after the instant of the first rent is an effect and not a cause of explosion. As soon as the first rent has taken place, the balance of strain in the fabric is disturbed, and therefore the internal pressure has greatly increased power in continuing the rupture; and also the pressure being then removed from the surface of the water, which is already heated to the temperature of the steam, the whole body of the water gives out its heat in the form of steam at a considerable pressure, and thus supplies the volume of steam for carrying on the work of destruction. When thus quickly generated, the steam perhaps carries part of the water with it in the same way that it does in ordinary priming; and it has been thought by some that the impact of the water is thus added to that of the steam, to aid in the shock given to all surrounding obstacles.

It is seldom that one out of a bed of boilers explodes without more or less injury to the others on either side of it; but sometimes two boilers in one bed, or three, or even five, have exploded simultaneously.

---

The causes of boiler explosions may be considered under the two general heads of—

Firstly, faults in the fabric of the boiler itself as originally constructed, such as bad shape, want of stays, bad material, defective workmanship, or injudicious setting:—and

Secondly, mischief arising during working, either from wear and tear, or from overheating through shortness of water or accumulation of scurf; or from corrosion, in its several forms of general thinning, pitting, furrowing, or channelling of the plates; or from flaws or fractures in the material, or injury by the effect of repeated strain; or from undue pressure through want of adequate arrangements for escape of surplus steam.

There is no doubt that many of the early explosions were from faults of construction. The stronger materials now used were then found so difficult to manipulate that others easier to work were chosen, and often the shape of the boiler was only selected as the one easiest to make. The early boilers were made of copper or cast iron, with leaden or even wooden tops, and of the weakest possible shape. Such was the boiler used by Savery, shown in Fig. 3, and the Tun Boiler and Flange boiler, Fig. 4 and Fig. 5. The very fatal explosion in London in 1815, referred to by the parliamentary commission previously named, was of a cast-iron boiler, which failed because one side was too thin to bear the pressure, as the casting was of irregular thickness. The steam being at that time used only at or below atmospheric pressure as a means of obtaining a vacuum by condensation for working by the external pressure of the atmosphere, so little was pressure of the steam thought of, that boilers were proposed and it is believed were actually constructed with hooped wooden shells, like barrels, and internal fireplaces and flues of copper; and even a stone chamber was named as being a suitable shell for a boiler, with internal fireplace and copper flue passing three times the length of the inside and out at the top, like an ordinary stove and piping. These boilers must have been something like the sketches given in Fig. 6 and Fig. 7, and were intended to be exposed only to the external pressure of the atmosphere.

Cast iron was frequently used for the shell of boilers, with an internal fireplace and tubes of wrought iron, as shown in Fig. 8., and boilers of this construction are still to be found in use at some of the older works at the present day. As the outside shell and front plate are 1½ inch thick and are not exposed to any wear at all, these boilers are sufficiently strong. A duplicate front plate with set of tubes attached is always kept on hand in case of need. Another form of cast-iron boiler is shown in Fig. 9., made in several parts put together with flange joints, with an internal fireplace and flue also made of cast iron. When cast iron was used for the parts exposed to the fire in boilers intended for high pressure, it was sometimes employed in the form of tubes of small diameter and proportionately thinner; as in Woolf's boiler, so much spoken of in the evidence before the parliamentary committee of 1817. This boiler, shown in Fig. 10., consisted of nine cast-iron pipes, about 1 foot diameter and 9 feet long, set in brickwork so that the flame played all round them. These small tubes were connected with another of larger size placed transversely above them, forming a steam receiver, and this again with a still larger one, which formed a steam chamber. No details of any explosions of the three last mentioned boilers have been obtained; but it is known that the cast iron was found a most treacherous material, especially when exposed to the action of the fire; and that the effect of explosion was very disastrous, because the boiler burst at once into many pieces, each of which was driven out with great velocity, and the danger was not mitigated by the circumstance of large masses holding together, as is found to be the case with wrought-iron boilers when exploded.

---

When wrought-iron boilers came into use the shapes were most varied, and the dimensions much larger than before. One of the earliest was the Wagon boiler, shown in Fig. 11., with round top and plain flat sides, which could only be made to bear even the smallest pressure by being strengthened with numerous stays. In most cases of explosion of this class of boiler the bottom was torn off, owing to the angle iron round it being weakened by the alternate bending backwards and forwards under each variation of pressure, as all the sides and the bottom must be constantly springing when at work. Such was the explosion at Chester in 1822, and many others. This shape was soon improved in its steam generating powers by making the sides concave instead of flat, as shown in Fig. 12., so that the heating surface was greater and also in a better position to receive the heat from the flame in the flues. This shape was further elaborated by rounding the ends as in Fig. 13., and in some cases making the bottom convex to correspond with the top, as in Fig. 14. All these forms however still required numerous stays to retain them in shape, the safety of the boiler being dependent upon the stays; and numerous explosions show the weakness of these boilers. They generally gave way at the bottom, as in an explosion that occurred at Manchester in 1842, where the boiler had been weakened by frequent patching; they also sometimes exploded through the failure of the stays.

A very early improvement in the right direction consisted in making the shell circular; and some few large boilers still exist that were made completely spherical, as show in Fig. 15., so that the whole of the iron was exposed to tension only, and required no assistance from stays, and the boiler had no tendency to alter its shape under varying pressure. This shape however had the great disadvantage of possessing the least amount of heating surface for its size or cubic contents; and also it was very liable to injury from sediment on the bottom, which accumulated on the most central spot. The spherical form was therefore soon modified into the shape shown in Fig. 16, by making the bottom more shallow, although still convex; and afterwards by putting flat or concave sides and a flat or concave bottom, with the angle constructed either of bent plates or angle iron, as in Fig. 17 and Fig. 18, which represent the forms known so well in the Staffordshire district as the common Balloon or Haystack boiler. Many of these have been made of very great size, measuring as much as 20 feet diameter, and containing so much water and steam as to be most formidable magazines for explosion. Perhaps no form of boiler has exploded more than this, partly because of the great number that have been used, but chiefly because of the inherent weakness of the shape. The records however have not been obtained of the great majority of these explosions, because they seldom caused sufficient damage or loss of life to attract much attention, as these boilers generally worked in isolated positions at collieries. The bottom is only prevented from blowing down into the fireplace by numerous stays from the top, and the angle iron round the bottom of the sides is much tried by the constant springing of the plates under every alternation of pressure; and the weakness thus occasioned is increased by the angle resting on the brickwork and being exposed to corrosion. The effect of this continued alternation of strain is well shown by the elastic model exhibited.

Notwithstanding the dependence of these boilers upon stays for their strength, many have been made as large as 12 and 15 feet diameter without stays; and explosion sooner or later has been the consequence. Such was an explosion that took place at Smethwick in 1862, which is shown in Fig. 19. As the force of the explosion was only slight, the effect of the bottom giving way, and the consequent rolling over caused by the reaction of the issuing steam and water, is clearly seen. Another example that occurred at Wednesbury in 1862 is shown in Fig. 20, where the explosion was rather more violent, the bottom of the boiler being torn off all round and left upon the firegrate, and rent nearly into two pieces; while the top and sides were thrown some height in one mass, and were only put out of shape by the fall. The weakness of this boiler had been further increased by making the bottom angle of angle iron, as shown enlarged in Fig. 21, with a ring of flat plate A interposed between the angle-iron ring and the concave bottom of the boiler; so that all the effect of the springing of the bottom, as shown by the dotted lines, was thrown upon the angle iron, which was accordingly found cut off all round. Had the concave bottom been made to rise direct from the angle iron, as in Fig. 22, the springing could not have been so great, and the angle iron would only have had to stand the shearing strain of retaining in its place the rigid bottom; but as about one foot all round the bottom was flat, and the concavity was only in the central part, the angle-iron ring had to bear an up and down strain, as shown by the dotted lines in Fig. 21, and the bending action was more severe than it would have been if the bottom had even been made quite flat all over.

A further form of the Balloon boiler is shown in Fig. 23, where the heating surface of the bottom is increased by an internal central dome-shaped fireplace, with an arched and curved flue conducting the flame through one revolution within the boiler before passing again round the outside. This construction however must necessarily have diminished the strength of the boiler greatly. In the drawing the top of the boiler, as indicated by the dotted lines, is removed to show the interior.

The desire to add to the strength of boilers by lessening the diameter of the shell led to the construction of the Plain Cylindrical boilers. They were made first with flat ends of cast iron, which frequently cracked and gave way when exposed to the fire, as described in many of the early American explosions. The flat ends when made of wrought iron, as shown in Fig. 24, are exposed to the same strain as the bottom of the balloon and wagon boilers, and are constantly springing with variation of pressure like drum heads, causing injury to the angle-iron joint. They also require long stays through them to hold in the ends, and these are subject to so much vibration that they seldom continue sound for long together, especially when joined with forked ends and cotters.

---

As the flat ends of such boilers are always being sprung by each alternation of pressure into a more or less spherical shape, as shown by the elastic model exhibited, this consideration no doubt led to the ends being made hemispherical, as shown in Fig. 25; and plain cylindrical boilers with these hemispherical ends are now so commonly used that they far outnumber any other form of boiler. Their shape renders them very strong, as the whole of the iron is in simple tension, and internal pressure has no tendency to alter the shape, as is shown by the elastic model exhibited. There is one circumstance very much in favour of the plain cylindrical boilers, and that is that they can be so easily cleaned and repaired, as a man can stand properly at his work at every part and the whole of the interior surface is exposed equally to view. They are of course exposed to all the evils of boilers externally fired, the part under greatest strain being weakened by the action of the fire; and the bottom is also exposed to injury from accumulation of mud and chips of scurf, which cannot be prevented from falling there, and lying upon the part exposed to the direct action of the fire. When made of great length, such as 70 or 80 feet, as is the practice for applying the waste gas from blast furnaces, these boilers are also liable to seam-rips or "broken backs," owing to the greater expansion of the bottom exposed to the fierce flame for its whole length, than of the top which is kept cooler by exposure to the air; and it would therefore be better to have a succession of short boilers, rather than only a single one, where great length is required.

One boiler has been seen by the writer where extreme length was avoided by curling the boiler round until the ends met forming a Ring or Annular boiler. This boiler is shown in Fig. 26, and is 5 feet diameter with 25 feet external diameter of the ring, or a mean length of about 63 feet; it has been found to work well for some years, although exposed to the heat of six puddling furnaces.

Explosions of plain cylindrical boilers have been very frequent indeed, although they have not caused a proportionate number of deaths, because they work usually in isolated positions at colliery and mine engines. The sketch shown in Fig. 27, represents an explosion that occurred at Darlaston in 1863, and illustrates the way in which these boilers usually explode. They generally open first at a longitudinal seam over the fire, which has become deteriorated by accumulations of scurf preventing proper contact of the water, so that the plates become overheated, their quality injured, their edges cracked or burnt, and the rivets drawn or loosened. The rent generally continues in the longitudinal direction to the sound seam beyond the bridge at the one end, and at the other end to the seam joining the front end to the shell; and then runs up each of the transverse seams, allowing the rent part of the shell to open out flat on both sides, and liberating both ends of the boiler, which fly off in opposite directions. Of course it is seldom that an explosion is quite so simple as this, as the direction of the flight of the various pieces is so much influenced by the last part that held in contact with the main body of the boiler. The want of due observation of this point has often led to erroneous conclusions.

In the explosion shown in Fig. 28, and in the model exhibited, which occurred at Westbromwich in 1864, the lower part of the side of an upright boiler was blown out; and the liberated part was also divided into two pieces, each of which fell some distance behind the boiler, in an opposite direction to the side from which they came. The explanation of this became obvious on examination, as the cause of the rupture had been the corrosion of the bottom, and the rent had run up the seams until it met the angle iron of the side tubes, round which it ran to the first seam above. This seam acted as a hinge on which the ruptured pieces turned, and they swung round so violently that they were wrenched off, but not before they had pulled the boiler over and received the diverting force which gave them their direction, for they flew off at a tangent, to the circle in which they had swung round on the sound upper seam as upon a hinge.

In order to avoid having a large diameter for plain cylindrical boilers, especially where exposed to the fire, boilers have been used that have supplied the required steam power by a combination of several cylinders of small diameter. One of these known as the Elephant boiler, has been so much used in France that it is sometimes called the French boiler; it is shown in Fig. 29, and consists of two cylinders of small diameter connected by upright conical tubes to a large cylinder above. Another form called the Retort Boiler, shown in Fig. 30, has been described at a previous meeting of this Institution (see Proceedings Inst. M. E. 1855 page 191). The disadvantages of these two combinations of plain cylinders are that they are not easy to clean or examine internally, and also there is not free exit for the steam, which has to find its way along small channels, and carries the water away with it, causing priming, and also retarding the generation of steam and endangering the boiler plates. With a view to strengthen the plain cylinder made of wrought-iron plates, the seams are sometimes made to run diagonally, as shown in Fig. 31, on the principle that, as the longitudinal is the weakest seam and the transverse the strongest, a diagonal between them gives the greatest amount of strength to the boiler as a whole.

---

Plain cylindrical and wagon boilers have for many years been made with internal tubes of various shapes and arrangement, through which the flame passes to add to the heating surface. These are shown in dotted lines on the previous drawings of wagon boilers, Fig. 11 and Fig. 12. They are also shown in Fig. 32, where a tube passes from over the fire to the front of a plain cylindrical boiler; in Fig. 33 two tubes pass from the sides to the front: in Fig. 34 the tube passes from the back, but returns over the fire and passes again to the back: and in Fig. 35 a tube from the back passes out through a cross tube in each side. The boilers in all these cases are fired externally. This addition of tubes has tended very much to increase the size of these boilers in order to make room for the tubes. These boilers are now found of 9, 10, and even 11 feet diameter; and this large shell being fired externally is exposed to the same dangers as those described in the plain cylindrical boiler, while it is not so easy to keep clean on account of the obstruction offered by the internal flues. When the flame has passed under the whole length of the bottom of these large boilers before going through any tube, it is doubtful whether the heating surface of the tube helps much in the generation of steam; but the tube is of use in reducing the quantity of water in the boiler, as it occupies a considerable space.

Explosions of these boilers have sometimes taken place by collapse of the tubes, but much more generally by the failure of the shell over the fire, as shown in the sketch Fig. 36, representing an explosion that occurred at Wolverhampton in 1865, in which the first rent took place in a seam over the fire where frequent repair had led to a considerable length of longitudinal seam being in one continuous line. The four plates over the fire parted and opened out until they had ripped two seams completely round the boiler; and the plates were thrown in one flat piece, as shown, upon a bank behind. The main body of the boiler with the tubes was turned over, and the front end blown away.

A modification or amalgamation of several of the forms of boilers already mentioned led to the construction known as the Butterley boiler, shown in Fig. 37, with a wagon-shaped end over the fire, continued in a single tube within a plain cylindrical shell beyond. This boiler has been found to generate steam very rapidly; but the extreme weakness of the construction over the fire and along the tube, especially at the part where the front end of the tube widens out in a bell mouth to meet the wagon-topped fireplace, has led to so many explosions that few boilers are now made of this form. A very early explosion that occurred at Edinburgh in 1821 was of a boiler somewhat of this shape, only that the wagon-topped fireplace was much longer. Other explosions of this form of boiler occurred at Ashton-under-Lyne in 1845, at Wolverhampton in 1854, and at Tipton in 1856.

The desire to economise fuel led to placing the fire inside the boiler, in a tube running from end to end, as shown in Fig. 38, and the great number of boilers of this form used in Cornwall gave it the name of the Cornish boiler. The exceedingly good duty performed by these boilers led many to believe them the most perfect for economy and durability; but the great number of explosions, or more properly of collapsed flues, that have happened, have altered this opinion, and led to the double-flue boiler shown in Fig. 39, in which not only is the heating surface increased but the strength also, by having two tubes of smaller diameter in the same shell. There are a great many varieties of the two-tube boiler, which have been made for the purpose of obtaining various particular results. In some cases the two tubes have been made to unite into a single tube immediately behind the fires, forming what is known as the Breeches-tube boiler, as shown in Fig. 40, and in other instances the outside shell of the boiler has been made oval, as shown in Fig. 41, with the two tubes continued through from end to end. The heating surface has also been increased, and the strength of the main tubes, by placing smaller transverse tubes across them at right angles; but these advantages are gained by increased complication, leading of course to greater difficulty in examination and repair.

---

The frequent failure of tubes by collapse when used for high pressures, and also the results of careful experiments, led to the simple addition of strengthening rings of different makes around the exterior of these tubes, by which the shell and the tubes are rendered of equal strength. It has taken considerable time for the belief in the weakness of large tubes when exposed to external pressure to become general, and a great many boilers are still made and used having even large tubes without the strengthening rings; and in some districts such boilers are used in great numbers and at far higher pressures than can be considered judicious. In more than one bed of boilers, one boiler after another has exploded by the collapse of the tube from the want of strengthening rings, and yet these have still been believed unnecessary; and the cases of isolated boilers of this construction where the large tubes have collapsed are extremely numerous, yet any other reason than the weakness of the tube has been considered more probable as the cause of explosion. A sketch of a boiler with collapsed flue is given in Fig. 42, which exploded at Burton-on-Trent in 1865; and it is selected from many others because it was a new boiler, well made and mounted, and was a good example of the weakness of a large tube to resist high external pressure when made of great length without the support of strengthening rings.

There are a great many advantages in the tubular boiler internally fired. The shell which is exposed to the greatest tension is not also exposed to the first action of the fire. The fire is in the midst of the water, so that the greatest effect is obtained from it; and the heating surface immediately over the fire, from which most steam is generated, has not so great a depth of water above it for the steam to pass through as in the externally fired boilers heated from the bottom. The tubes also act as stays to the ends; and the mud in the water falls off the tubes, where it would do mischief, and settles on the bottom, where it is comparatively harmless.

These tubular boilers are however subject to disadvantages peculiarly their own. It is not so easy to move about within them for cleaning and examination as in the plain cylindrical boiler, as the tubes fill up the space so much. The difference of expansion between the highly heated tube and the comparatively cool shell produces a strain, which causes the ends to bulge out; or if the ends are made rigid, the strain sets up a contortion in the tube, which causes furrowing of the plates by making the iron softer or more susceptible of corrosion in certain lines of strain. Notwithstanding these drawbacks however this form of boiler is an excellent one.

Many modifications in the forms of boilers have been made to enable the manufacturers to use the waste heat from various processes, especially from the making of iron. The plain cylindrical boiler has been used in this way, with sometimes as many as eight puddling furnaces made to work upon one boiler. One of the earliest special arrangements for this purpose was the Upright boiler with central tube, shown in Fig. 43, which was originally made for two furnaces; and about 7 feet diameter and 16 feet high. The size has since been increased to 10 feet diameter and 28 feet high, as shown in Fig. 44. These boilers are made for one, two, three, or four puddling furnaces; and consist of a cylinder with spherical ends, standing upright, with a central tube from the bottom to about half the height, into which the side tubes join. The heat from each furnace plays over a portion of the shell, and then passes through the side tubes and down the centre tube into the underground flue to the chimney.

These boilers have many good points: there is great heating surface; and the shell being heated all round does not strain the plates and seams by unequal expansion so much as in the horizontal plain cylindrical boiler heated only at the bottom; and as both ends are spherical there is no alteration of shape under internal pressure. Moreover in consequence of the upright position of the boiler a safe depth of water can easily be maintained, and yet the steam is taken off so high above its surface that there is little priming; and every part can most easily be cleaned and examined, as a man can stand upright both in the boiler and in the flues. But the great drawback to this class of boilers is that they must stand in the midst of the workmen; so that, although they are not more liable to explode than any other form of boiler, yet when they do burst they necessarily endanger more lives than is usually the case with other boilers that can be placed more away from the men employed at the works. Should anything arise with the boiler to make it desirable to withdraw the fire, this cannot be done without much delay, as the furnaces have to be stopped and the iron run out. Also an explosion can hardly happen without some of the melted iron being scattered among the men at work.

Some of the most fatal explosions of these boilers have arisen from careless construction. Such was the case in an explosion at Dudley in 1862, shown in Fig. 45, where the crown plate forming the top of the centre tube was attached to the sides of the tube by so slight an angle iron, as shown enlarged in Fig. 46, that the pressure of steam on the flat crown plate fairly sheared the angle iron through, and allowed the plate to be blown down the centre tube into the chimney flue, whereupon the boiler was violently thrown off its seating by the reaction of the issuing steam and water thus liberated.

---

The double-tube horizontal boiler is also used in connection with iron-making furnaces in many places, one furnace working into each tube. Although by this arrangement the boiler can be placed a little further from the workmen, some very fatal explosions have happened to such boilers, as at Masborough in 1862.

---

Single-furnace boilers have been much used in the form of a single-tube boiler standing on end, as shown in Fig. 47, with the flame passing up the tube, which is continued in the form of a chimney on the top of the boiler. The tube passes through the steam at the top, so that the plate is not protected from overheating by contact with water; and this has caused explosion in some instances, although the tube has been lined on the inside with firebrick to shield the plate from the flame. Another great disadvantage of this Chimney boiler is that the space between the tube and the shell is so narrow that it is almost impossible to examine or clean it internally.

A further arrangement for a single-furnace boiler is the Elbow boiler, shown in Fig. 48, where the two difficulties mentioned in the previous boiler are avoided.

Many internally fired upright boilers of various shapes have been constructed to suit various purposes. One of a large size that has been at work many years is shown in Fig. 49, with an internal fireplace and a suspended cone and cross tube for increasing the heating surface. This boiler is set in brickwork in such a way that the heat passes through the side tubes and round the exterior shell before going off to the chimney.

A very fatal explosion at Stoke-upon-Trent, in 1863, resulted from an attempt to work a boiler of somewhat the same general form, but without the same careful attention to the details of construction. This boiler is shown in Fig. 50; the internal fireplace is of conical shape, 4 ft. 6 ins. diameter on the top and 6 ft. 10 ins. at bottom, and was joined to the external shell by a flat annular bottom. Almost the first time it was worked at high pressure the conical fireplace collapsed, breaking off at the seam at the top of the cone, and blowing down upon the grate, as shown in Fig. 51. The flat bottom was then left without the support of the cone and side tubes, and gave way all round the outside angle iron; and the top flew up a great height into the air, and fell a crumpled heap, as shown in the sketch. In this case the only wonder is that a boiler of such weak construction worked at all without explosion.

---

There yet remains to be noticed a very large and varied class of boilers that have been designed with the express object of avoiding explosion. Some of these, made of cast-iron pipes of small diameter, have already been referred to. When steam carriages were first constructed, boilers were tried made of a cluster of small pipes, set both upright and horizontally, connected with a general receiver and with each other by still smaller pipes. These were found to have such small circulation of water that they very soon burnt out, and also led to much priming. Afterwards, narrow chambers made of corrugated plates set like the cells of a battery were tried, but without much success. The multitubular boilers of the locomotive type soon superseded all others as quick steam generators, and until lately they have been considered as almost absolutely safe from explosion. It is found however that the barrel of these boilers is peculiarly liable to furrowing, owing to the strain weakening the iron in certain lines. Perhaps no boiler shows more clearly than the locomotive how necessary it is that every part should be open to examination; and also how unwise it would be to use for stationary purposes small cramped up boilers, only intended to meet the necessities of locomotion. Many explosions of locomotive boilers have taken place; but it is not necessary to give details in this paper, as they are fully given in the published official reports of the government inspectors.

Among the form of boilers designed to obtain very rapid generation of steam, combined with increased safety from explosion, may be specially named that consisting of a system of small pipes within a shell with an artificial circulation of water, and also the boiler consisting of a cluster of cast-iron spheres, both of which have been described at previous meetings of the Institution (see Proceedings Inst. M. E. 1861 page 30, and 1864 page 61); but neither has been much used in this country at present. The boilers also which consist chiefly of small tubes hanging down into the fire, with smaller tubes or other arrangements within them for securing a natural circulation, deserve mention, as they appear successfully to accomplish that end.

The principle of all these small boilers appears to be that only a small quantity of water should be contained in them, so that there should not be a reservoir of danger in the shape of a mass of highly heated water ready to be converted into steam if a rupture takes place: and it cannot be denied that this is an advantage. But on the other hand these boilers of small capacity, which evaporate their whole contents in a few minutes, are subject to new dangers from that very cause; and although admirably adapted for purposes where steam is wanted quickly on a sudden emergency, as in the case of fire engines, or where the generating power required varies each moment, as in the locomotive, they are for the most part ill adapted for ordinary stationary purposes, such as the mill or the colliery. They require constant firing and vigilant attention to the feed, and cannot be left for a time with safety like the ordinary stationary boilers. It has to be borne in mind also that the very reservoir of danger so much dreaded is also a reservoir of power, which assists in the steady maintenance of the machinery in motion. The large mass of water heated to the evaporating point, the heated brickwork of the flues, and the large fireplace, are so many assistances to regularity, and enable the man in charge to attend to his other duties without the risk of spoiling the boiler or letting down the steam by a few minutes' absence from the stoke hole. Steam employers are found at present to prefer the known dangers of the large boilers to the supposed safety of small boilers, which they fear are troublesome in practice.

---

Many of the early boilers were rendered weak by the injudicious manner of arranging the seams. The longitudinal seams were made in a continuous line from end to end, as shown in Fig. 24, page 20, with the transverse seams also continued completely round the boiler, so that at the corner of each plate there were four thicknesses of iron. The crossing of the seams, as in Fig. 25, page 21, adds much to the strength, and also often prevents a rent from continuing forward to a dangerous extent.

It is scarcely requisite to mention the necessity of good material and workmanship to secure strength in a boiler, however perfect the design. If the plates are of weak and brittle iron, or imperfectly manufactured, they will never make a good boiler. Apart from the strain upon the boiler when at work, the iron has to undergo the strain of the necessary manipulation, shaping, and punching, during the construction of the boiler. If the plates forming the boiler are not well fitted to their places before the rivet holes are made, the errors have to be partially rectified by using the drift in the holes to an unwarrantable extent, and then using imperfect rivets to fill up the holes that do not correspond with each other; and the mischief is too frequently increased afterwards by excessive caulking, in the endeavour to stop the leaking which is sure to show itself. In this way a boiler is often exposed to most unequal internal strain between its several parts before it is set to work at all; and when the heat is applied to it, the mere expansion causes undue contortion, and leads to seam rips, and ultimately to disaster. Several specimens of faulty rivetting and caulking were exhibited to the meeting, and a sketch of one of them is shown in Fig. 52.

The strength of a boiler is often very much lessened by the injudicious manner in which the mountings are fixed upon the boiler, and many explosions are the consequence of this defect. Not only are a great many holes for fittings cut out of the boiler in one line, but these holes are made needlessly large. Steam domes are often so placed as greatly to weaken the shell of the boiler, the hole cut out of the plate being made the full diameter of the dome; and in some cases the domes or steam chests have been made square or rectangular, so as to weaken the shell still more, as shown in Fig. 53.

Manholes are often a source of danger, if not properly arranged and duly strengthened. Even in very small boilers they are often placed with the longest diameter in the longitudinal direction of the boiler, so that the shell is greatly weakened, as in the sketch, Fig. 54, of an exploded boiler at Walsall in 1865. This boiler was 5 ft. 3 ins. long and 2 ft. 6 ins. diameter, and yet the manhole was 18 inches by 13 inches, and placed within a few inches of one end. The end was fastened in by angle iron, which was not welded, and consequently there was so little strength at the small portion of the shell between the end and the manhole that it gave way and liberated the end and the manhole lid, after which the main body of the boiler was thrown by the reaction across several streets to a great distance.

A somewhat similar injudicious arrangement of the manhole is shown in Fig. 55, where a manhole 17 inches by 14 inches was cut out of the flat top of a steam dome only 2 ft. 6 ins. diameter, without any strengthening ring to compensate for it. The repeated strain of screwing up the manhole lid, combined with the pressure of the steam, caused the lid to force its way out through the plate and blow away. This explosion occurred at Birmingham in 1865.

The preceding examples have shown how explosions often result from faults in the construction of boilers; and the following instances illustrate the explosions caused by mischief arising during working. A boiler perhaps more than any other structure is subject to wear and tear; and let it be worked ever so carefully, it will seriously deteriorate. The wonder is, considering the work they have to perform, that so many boilers are found which have worked twenty, thirty, or even fifty years without explosion. The terms wear and tear however are too vague for this subject, and the mischief met with must be considered under distinct heads.

There is no doubt that the thing most to be dreaded for boilers is corrosion; because when the plate is once thinned, it cannot be strengthened again, but must remain permanently weakened. Corrosion the more deserves attention because it is easily detected by moderate vigilance, and can generally be prevented by moderate care, or by the boilers being so arranged that they can be readily examined in every part. Corrosion has been the direct and unmistakeable cause of a very large proportion of the explosions that have happened: it occurs both inside and outside the boiler, according to circumstances, and attacks the iron in various ways and in different places.

Internal corrosion sometimes takes place from bad feed water, and its effects are different in extent in the different parts of the same boiler. It very seldom thins the plate over a large surface regularly, but attacks the iron in spots, pitting it in a number of holes. These are sometimes large, as if gradually increasing from a centre of action; and sometimes small, but so close together as to leave very little more space whole than that which is attacked. A very curious example of the latter was exhibited to the meeting, and shown in Fig. 56 and Fig. 57, cut from the lower part of the shell of a large tubular boiler externally fired. The corrosion was greatest along that part of the shell most exposed to heat, and was so extensive that two boilers exploded simultaneously. The boilers had been at work sixteen years, but the corrosion commenced about eight years before the explosion, when the feed water was rendered corrosive by being obtained from some iron mines. This explosion occurred at Aberaman in 1864. The corrosion had been seen going on for years, and was not considered sufficient to cause danger; but the depth to which it extended through the thickness of the metal is seen in the half size section, Fig. 57. Another sample equally curious was exhibited to the meeting, and shown in Fig. 58 and Fig. 59, taken from the sweep plate over the fire in a plain cylindrical boiler which had worked about ten years.

The feed water was occasionally bad, and attacked the iron over the area DDD, where unprotected by scale. The protection afforded by scale against occasional corrosive feed water is worthy of notice. In the two specimens exhibited it is seen that the protection has been perfect where the scale has not been chipped off; and the edge of the sound part projects over the hollow, as seen in the half size sections, Fig. 57 and Fig. 59, the corrosive water having eaten away a larger area beneath than that through which it first entered the surface of the iron.

Internal corrosion is frequently observed where boilers are fed from canals or streams in the neighbourhood of chemical works from which corrosive matter is discharged at intervals into the water. The corrosion takes place in isolated spots, but causes deep holes; which seems to be accounted for on the supposition that the scale previously upon the plate cracks during the cooling of the boiler for cleaning, and forms a blister, so that a piece of about 2 inches area is raised slightly from the iron. When the boiler is again put to work, this blister becomes filled with the corrosive water, which is held there without circulation and causes corrosion. When the boiler is again emptied these blisters may be seen, and if broken show the blackened water and the injured surface. In future working each of these blisters forms a constant unprotected point for attack. It is frequently seen further that such corrosion is arrested if water be used which deposits scurf; but fresh blisters and renewed corrosion will result from a return to the use of the bad water.

The internal corrosion called furrowing has proved a frequent cause of explosion, especially in locomotive boilers. It differs from other corrosion by being in deep narrow continuous lines with abrupt edges. It will sometimes go completely through a plate; and is found where a sudden change of thickness occurs, either along the lines of the seams, or opposite the edge of angle-iron attachments. This effect is supposed to be due to the alternate springing of the plates under each variation of the pressure or temperature, causing the line of least resistance to receive a strain somewhat similar to that produced by bending a piece of iron backwards and forwards for the purpose of breaking it. This line of injury is exposed to constant attack from corrosion, because the scurf is always thrown off from it.

---

External corrosion is a far more frequent cause of explosion in stationary boilers; and it arises from many causes. The most frequent cause, although the most easily detected, is leakage from the joints of the fittings on the top of the boiler, which are too frequently attached by bolts instead of rivets. This evil is much increased when the boilers are covered with brickwork, which holds the water against the plates, and hides the mischief from observation. It is astonishing to find how much damage is allowed in this way to go on without attention, until the tops of boilers are corroded so thin that little holes burst through. These are sometimes found stopped with wooden pegs or covered by screwed patches of plate, either of which cause leakage that hastens the mischief, as shown by the sample exhibited. Boilers exposed to the weather will of course become corroded like anything else made of iron and not painted; and yet so much mischief is sometimes caused by leakage beneath improper covering that exposure may almost be said to be the smaller evil of the two, as it is better to see what is going on than to rest in false security. No covering will be found cheaper, or better, in the long run, than a roof, which prevents the loss of heat by exposure, and yet allows free access to all the fittings and joints on the top of the boiler.

Some examples of the evils of covering can be given that have come under the writer's observation. A set of boilers had been well covered by arches of brickwork, so built as to keep out all water, and also set so as to touch the boilers only at intervals, leaving a space generally of a few inches. After about seven years' working, the whole of the tops of the boilers were discovered to be dangerously thin, and had to be renewed. The cause was leakage from the joints of fittings and seams of the boilers, and the issuing steam had been drawn along the space between the boilers and the arches, and had escaped at a place where it had not attracted notice. In another case, a somewhat similar set of boilers were covered with ashes, to prevent the loss of heat by radiation; and the rain and the leakage beneath the ashes, in conjunction with the corrosive matter from the ashes themselves, thinned the tops of the boilers to a dangerous extent in less than two years. A sketch of the corrosion caused in this instance by covering with ashes is shown in Fig. 60 and Fig. 61.

Similar mischief has been noticed in boilers covered with sand, as shown in the sketches Fig. 62 and Fig. 63, which represent an instance of corrosion after eight years' working; although nothing forms a better covering than sand for preventing loss of heat by radiation. In both these examples it will be seen that the corrosion has continued until the thickness of the plate has been so eaten away that a hole has been burst out at SS. A very good covering is formed by brickwork in cement; or various cements made for the purpose, which adhere to the surface of the plate and yet show leakage; or such materials as sacking or felt; or sheet-iron casing, leaving about 6 inches of air space all round the boiler. But all these have the great objection that they hide the boiler from inspection, except by the expensive process of removing the covering; and in this way dangers that have caused explosion have remained hidden from observation.

Explosions have also taken place from general corrosion of the surface of the boilers in the flues. A new boiler which was set on sidewalls built upon a foundation of porous rock was found to have become corroded all along the bottom in less than two years, owing to the dampness which rose from the foundations causing a constant presence of vapour. The corrosion was peculiar, and more like that found on old iron left for a long time in a damp place; for the iron plate fell to pieces when touched, and large flakes could be raised from the surface, and the greater part of the thickness of the plate could be removed with the fingers. Somewhat similar corrosion had taken place in a boiler which exploded at Loughborough in 1863; the bottom of the shell became rent at the corroded part, and as the fracture continued spirally round the boiler several times, nearly all the shell was peeled off in the curious manner shown in Fig. 64. The explosion shown in Fig. 65, which occurred at Leeds in 1866, also arose from corrosion of the bottom of the boiler.

The greater part of the corrosion found in the side flues of boilers is caused by the leakage of seams. Many boilers are emptied for cleaning as soon as work is over on Saturday night, and long before the brickwork of the fireplaces and flues has cooled; and consequently, the boiler, having no water in it, is made much hotter than it ever is in working, and the seams are injured and sprung and the rivets loosened by the extra expansion so caused. This is sometimes done intentionally, in order to loosen the scale by the greater expansion of the iron than of the scurf. When the boiler is again set to work, the seams and rivets leak and cause that corrosion which is called channelling. This has been observed to occur to such an extent that all the seams in a boiler have been seen thus corroded; and the same has sometimes been found in all the boilers in a large manufactory. Specimens of this channelling were exhibited to the meeting. One in particular, shown in Fig. 66 and Fig. 67, deserves attention, as it shows the effect of a jet of steam and water from the leaking rivet R, in cutting a series of channels into the plate along the course of the dotted lines EEE, and producing a hole in the plate at S. This corrosion had been going on for about four years, but was in a part of the boiler seldom seen in ordinary examination. Many explosions have resulted from this form of corrosion; for when a rent is once made, the fracture continues along the thinned channel of the plate.

The corrosion most to be dreaded, because most difficult to detect, is that which takes place where the boiler is in contact with brickwork; and it is found alike in all forms of boilers set in brickwork. When found at the part where the side flues are gathered in at the top against the boiler, it is usually occasioned by the leaking of fittings or feed pipes, or by rain being allowed to run between the boiler and the brickwork. More than one explosion has been caused by the droppings from a roof being allowed to fall upon the tops of the flues. When the corrosion is found at the point where the bottom flue walls touch the boiler, it is frequently caused by the leaking of seams that have been strained by the weight of the boiler; and this often arises from want of care to replace the brickwork, after repair of the boiler or flues, in such a position as to take again its proper proportion of the weight of the boiler. Cases have been met with where the shape of the bottom of large boilers has been quite altered by such means. The brackets on the sides of heavy boilers have not only been strained so that the rivets or bolts have leaked and caused corrosion, but they have also bent or cracked the side plates of the boiler. The bracket shown at B in Fig. 53, page 40, made of only an angle iron with a piece of plate attached, is especially liable to cause injury if the brickwork is not rebuilt close up to the angle iron, as the leverage is so great. This is avoided by the better form of bracket shown at C, consisting of an elbow of flat bar-iron rivetted at top and bottom to the boiler.

In the old balloon and wagon boilers, the angle where the bottom joined the sides scarcely ever remained sound for long when in contact with the brickwork, and many of those that exploded have been found almost corroded through where they stood upon the brickwork. The explosion before alluded to and shown in Fig. 7, was caused by corrosion of the bottom of the boiler where it was set on the brickwork. Many boilers are so set that the brickwork of the flues is made to follow the shape of the boiler, with as little space between as possible; but the slight advantage gained in increased heating effect is far outweighed by the impossibility of getting into the flues for examination. It is only by having the flues sufficiently roomy that proper examination can be made, and that the indications on the brickwork of leaking can be seen and remedied, and corrosion arrested. A remarkable case of corrosion occurred in a boiler with an oval shell, set upon a middle wall. The flues were too narrow for a man to enter, and a leak in the bottom was only discovered by the boiler nearly running empty while the engine pumps were standing for a short time. It was subsequently found that the whole bottom where it rested on the wall was extensively corroded in a continuous line, and that explosion was only prevented by the numerous stays across the bottom to compensate for the oval shape. Fig. 68, shows the position and extent of the corrosion, and the plate was completely in holes at the parts indicated by the black marks. This corrosion was supposed to have been going on for about three years.

It is sometimes asserted that corrosion cannot be the cause of an explosion, because the corroded place would simply give way and let off the steam harmlessly, or at least the boiler would not be displaced from its seating. When the corrosion is only local, and surrounded by sound plates of sufficient strength to arrest the extension of the fracture, this may be the case, as in an explosion at Sheffield in 1865, shown in Fig. 69, where a piece of plate was blown out on one side of the boiler, allowing the steam and water to escape without displacing the boiler; the thickness of the plate at that part had been reduced to 1/8th inch by corrosion in about 1½ years, which had been caused by leakage at the seams from inefficient repair with bolts instead of rivets, and also from the moisture having been allowed to be kept against the plate by the brickwork. But even under such circumstances, if the piece blown out should be from the bottom, the whole boiler may be thrown a great distance by the reaction of the issuing steam, as in an explosion at Leeds in 1865, shown in Fig. 70. If the corrosion extends for any length, the first rent is almost sure to continue until a complete explosion is the result. Several of the small models exhibited to the meeting showed the line of fracture in various cases of explosion. One showed the appearance of a plain cylindrical boiler after explosion caused by corrosion along the whole length where it rested on brickwork; this explosion occurred at Wigan in 1865, and a sketch of it is given in Fig. 71.

Many explosions of boilers have been caused by accumulation of scurf. The mischief is not so much from scurf being gradually deposited all over the interior of the boiler to a dangerous thickness as from the chips off the sides falling in heaps on the bottom. The plate beneath this accumulation becomes overheated, because not in contact with the water, and softens and sinks down into a "pocket," which if unnoticed will soon burn quite through. If the scurf that has caused the mischief is thick and hard enough to resist the pressure for a little time, the hole enlarges, until the scurf suddenly gives way and allows the contents to issue so violently as to disturb the boiler, or at least to blow the fire out of the grate. Such was an explosion at Bilston in 1863, where a large plain cylindrical boiler, 9 ft. diameter, was heated by three large fires placed side by side along the bottom; and a large "pocket" burst out over the third grate, and scalded the attendant to death. A similar pocket in a boiler, 4ft. 6 ins. diameter, which exploded at Dudley in 1864, after having been at work six weeks without cleaning, is shown in the transverse section, Fig. 72. In this case the scurf had filled up the circle of the boiler to a depth of 3 inches at the bottom, as shown in the drawing, and was of a very hard description; and the boiler plate was bent out in a gradual curve, and thinned to about 1/16 inch, the original thickness being ½ inch.

The whole bottom of a boiler is sometimes injured, and the plates buckled and the seams sprung, from the accumulation of mud. One case may be mentioned where the water was very full of mud, and the boilers were worked day and night during the week but stopped for several hours on Sunday, during which time the deposit of mud was so thick that it did not get thoroughly disengaged again from the bottom when the boiler was set to work, but hardened into a mass. Although many of these pockets and injuries to the plates may occur without serious damage, they sometimes cause that first rent which destroys the equilibrium of the structure and leads to explosion. Some of the specimens of scurf exhibited to the Meeting show that their thickness is made up of small chips, carelessly left after cleaning or fallen from the sides of the boiler, as seen in Fig. 73, or from cotton waste or other matter left in the boiler and forming a nucleus for the scurf to accumulate upon. Other specimens show that foreign matter must have been put into the boiler to stop leaking.

Accumulations of scurf in the feed pipes at the point of entrance into the boiler have also caused explosion by stopping the supply of water. The same result is caused by the freezing of the water in the pipes which are exposed, and each winter one or two boilers are injured or exploded from this cause, especially small household boilers placed behind kitchen grates. Scurf cannot be considered so great an evil as corrosion, since it can be removed, and if this is done in time, the boiler is restored to its original condition.

The advantage of a pure water, which does not deposit scurf, is so great for the supply of boilers that it is always worth while to go to considerable expense for obtaining it; or to take some steps for purifying the feed water as much as possible. If it is only mud mechanically suspended, which would deposit by gravity on the bottom of the boiler, frequent use should be made of the blow-off apparatus. If the impurity is light enough to be carried to the surface in the form of scum, the blow-off apparatus should discharge from the surface of the water as well as from the bottom. If the impurity is chemically suspended in the water, some one of the many substances which form the refuse from various manufactures, and which may contain suitable ingredients, should be used to counteract the effect of the impurity. Common soda will answer the purpose perhaps better than anything else. It must not be forgotten however that the blow-off apparatus must afterwards be used more frequently, to rid the boiler of the foreign matter, or the mischief will be increased. In marine boilers, constant attention is necessary to get rid of the saline deposit; and in stationary boilers using impure water an equally systematic attention is needed to get rid of the earthy deposit.

---

Perhaps no cause of explosion is oftener mentioned than shortness of water, and this is not unfrequently coupled with turning on the feed suddenly into an overheated boiler. Many explosions have been attributed to this cause, when closer investigation would have revealed some far more probable reason. For instance, shortness of water was stated as the cause of the explosion, at Abercarn in 1865, of a single-tube boiler with a very large flue tube, which collapsed upwards from the bottom. The top of the tube and the sides of the shell had not the slightest mark of overheating, although exposed to the flame of three furnaces, one of which worked through the tube, and the others on each side of the shell. In this case the cause of explosion was clearly the weakness of the tube, and not shortness of water. It is erroneous to suppose that if a boiler runs dry, or if the feed is turned into a red-hot boiler, there must necessarily be an explosion. If a boiler unconnected with any other runs rapidly empty, from the breaking of the blow-off pipe or any such cause, it will simply get red-hot and sink out of shape upon the fire, as may often be seen, but no explosion would happen. If the water only falls gradually, as it would if the feed were turned off and evaporation continued, the parts exposed to the fire would get overheated as the water left them. If the subsidence of the water were very slow, those parts might get red-hot, and so much softened and weakened as to be incapable of bearing the pressure, when an explosion would take place, as at Smethwick, in the present year, where the flues were set above the water line, as shown in Fig. 74.

If however the water were turned on again before the overheating had gone so far, and the feed pipe were, as usual, carried down to nearly the bottom of the boiler, the water would gradually creep up the heated sides and cool the plates, the heat of which would not be sufficient to cause greater evaporation than the ordinary safety valves would carry off. The danger would not arise so much from the excess of steam generated by the heat accumulated in the heated plates of the boiler, as from the injury and strain that would be caused to the plates by the undue expansion and sudden contraction, especially as this action would take place on only a portion of the boiler. A singular case, bearing on this point, may be mentioned. A four-furnace upright boiler, like that shown in Fig. 44, happened to run so nearly empty, through the accidental sticking of the self-acting feed apparatus, that the level of the water sank to the top of the hemispherical end forming the bottom of the boiler. The feed apparatus then became released of itself, and, the feed being turned full on, the water gradually rose until the whole occurrence was only discovered by the leaking at the seams that had been sprung, which caused so much steam in the flues as to stop the working of the furnaces. The overheating had been sufficient to buckle the plates, and in one place a rupture had almost commenced; but there was no explosion. By way of direct experiment upon this point, boilers have been purposely made red-hot and then filled with cold water, without causing explosion.

It has been supposed that boilers sometimes explode from overheating without the water level being below the usual point, or without the accumulation of scurf previously alluded to, but simply by the rapidity of the evaporation from an intensely heated surface causing such a continuous current of steam as to prevent the proper contact of the water with the heated plate. Such has been the cause assigned for the explosion of a three-furnace upright boiler at Birmingham in 1865, shown in Fig. 75. A piece of plate about 3 ft. by 1½ ft. was blown out of the side, at a place where an enormous flame impinged continually. The plates had first bulged out, and then given way in the centre of the bulge, each edge being doubled back and broken off. There was no positive evidence as to the water supply; but the crown of the centre tube, which was much above the bottom of the part blown out, remained uninjured.

A somewhat similar case was that of a large horizontal boiler at Kidderminster, the tube of which collapsed in 1865, as shown in Fig. 76. It was heated by four furnaces, one of which worked into the tube, one under the bottom, and one on each side; and all the furnaces worked into the same end of the boiler. The tube was found to have partially collapsed at that end, and the top had dropped 11 inches. This was repaired in the first instance, but was afterwards again found injured by overheating, although not so seriously. It is very probable that the extremely rapid ebullition from the sides and bottom, from which the steam had to pass up the narrow space between the tube and the shell, produced such a foaming that very little solid water could reach the top of the tube where it was exposed to extreme heat.

Many explosions have been attributed to deterioration of the iron through long use, as in an explosion at Durham in 1864, and another at Haswell, near Sunderland, 1865, where the boilers had worked constantly for 25 and 30 years respectively. When an explosion arises from the failure of a plate which has not been properly welded in rolling, there is no question that it was unsound when put in, and escaped notice; but when the plate that fails is found to be brittle and of bad iron, the fault is rather attributed to the effect of working than to original bad quality. Of course this is not always the case, as the injury done to plates by overheating has been already explained. Pieces of plate have in some cases been erroneously pronounced to be deteriorated by work, which have been taken from situations in the boilers where they were not exposed to any action of fire that could cause overheating; and therefore in reality the injury could only have taken place when the boiler was being made, by burning the iron in bending it to the required shape. A frequent cause of fatal injury to boilers is injudicious repair, whereby the crossing of the seams is destroyed, as in the explosion at Wolverhampton in 1865, previously referred to and shown in Fig. 36. Moreover the edges of the old plates, already tried by the first rivetting and the subsequent cutting out of the rivets, are frequently strained again by the use of the drift to draw them up to the strong new plates; and many a seam rip is thus started which ultimately causes explosion.

Many explosions have been caused by the want of proper apparatus for enabling the attendant to tell the height of the water and the pressure of the steam, and also by the want of sufficient apparatus for supply of feed water and escape of steam, or by the failure of one or other of these; but such explosions can only be referred to generally in the present paper. The mountings on a boiler are usually so open to observation, and the importance of having them good and efficient is so universally acknowledged, that much remark is not needed. Mention has already been made of the sticking of self-acting feed apparatus as a cause of mischief, and similar failures of floats and gauges have constantly happened; but this should by no means be considered to condemn self-acting apparatus, either for assisting in the steadiness of working, or for giving warning of danger. The apparatus however should be relied on for assistance only; and an attendant cannot be called careful who leaves a boiler dependent on such apparatus without watching. The self-acting principle has been seen by the writer applied in a novel and useful way in a recording pressure gauge, which proved the more interesting as it had shown the actual pressure of steam at the time of the explosion of one of the boilers with which it was connected.

Among the numerous boiler explosions that have been attributed to over-pressure through deficient arrangements for escape of steam, in many cases the safety valves have been placed on the steam pipes in such a manner that the communication with them was cut off whenever the steam stop-valve was shut, which is just the time when the safety valves are most wanted. Safety valves are too often found needlessly overweighted; and it is believed that many boilers are constantly worked with safety valves so imprudently arranged and weighted, that they could not carry off all the steam the boilers would generate without a very great increase of pressure.

---

It is concluded that enough has now been said to show that boiler explosions do not arise from mysterious causes, but generally from some defect which could have been remedied if it had been known to exist. It only remains therefore to consider what is the most ready and efficient way to discover the true condition of a boiler. It has been maintained that this end is best accomplished by what is called the hydraulic test, in which a pressure of water is maintained in the boiler for a given time at a certain excess above the working pressure. This test is undoubtedly useful so far as it goes, and is perhaps the only one that can be applied to boilers with small internal spaces, such as locomotive boilers, not admitting of personal inspection over the whole of the interior; and it is also admirable for testing the workmanship of a new boiler. But on the other hand the conditions of a boiler at work are so different from those which exist during the hydraulic test, that this alone cannot be depended on; for old boilers have been known to stand this test to double their working pressure without apparent injury, although known to be dangerously corroded. The difficulty also of seeing or measuring the effect of the hydraulic test upon large boilers set in elaborate brickwork is so great that little practical benefit has resulted in many cases.

It is believed by the writer that the surest way to ascertain the true condition of a boiler is to examine it at frequent intervals in every part, both inside and outside; and as this can only be done when both the boilers and the flues can be readily entered, it is specially important that facility for examination should be made a consideration in selecting a construction of boiler. Permanent safety should be considered as an element of economy, in addition to its still higher importance in reference to the preservation of life.

---