place, its power of resistance is very seriously reduced; and unless the utmost care and attention be observed, collapse with all its attendant calamities is sure to be the result. The thickness of plates, the diameter of the tubes, and the position as well as the form of the flues, are all points of consideration; but unfortunately we have no precise data on which to establish formulæ for calculating the strengths of those parts, and in the present state of our knowledge we are left to guess at the strengths and the capabilities of sustaining resistance in those positions to which we have referred. It is more than probable that the strengths will be for the same thickness of plates as the diameters; but we are entirely at a loss to determine what diameter of flue and what thickness of plates may be necessary to give uniformity of strength in every part of a boiler so constructed. In the absence of these data, I must reserve the investigation of the question till I have an opportunity of establishing by actual experiment the laws of resistance to collapse under different forms of construction. As soon as those facts have been ascertained, we shall then be able to effect constructions on principles calculated to ensure uniformity of strength under the combined forces of tension and compression. In the mean time, I shall avail myself of a Table of Strengths showing the safe working-pressure and ultimate powers of resistance in boilers of different dimensions. * This Table is computed from my own experiments on the strength of iron plates, and is introduced to the public with the following observations by Mr. Coburn: It has been found by actual experiment that good English forged iron will bear a strain of 25 tons to the square inch; that is, a bar of 1 inch square, or a plate of * These remarks, true when originally written, no longer apply. The data required have been determined, and will be found in this edition at page 44. wrought iron containing 1 inch of sectional area, will require 25 tons, or 56,000 lbs., to wrench it asunder: but Mr. Fairbairn states, that "plates when riveted together are reduced in strength, from the fact that nearly one-third of the material is punched out for the reception of the rivets," and therefore he takes 34,000 lbs. as equal to the strength of riveted plates containing 1 inch of sectional area. The following Tables are deduced from the assumption that 34,000 lbs. per square inch is the tensile Diameters Working-pressure Bursting-pressure Working-pressure Bursting-pressure of boilers for th-inch plates for th-inch plates for 3-inch plates for 4-inch plates Rule for th-inch Plates.-Divide 4250 by the diameter of the boiler in inches, the quotient is the working-pressure, being one-sixth the strength of the joints. Rule for -inch Plates.-Divide 5666.6 by the diameter of the boiler in inches, and the quotient will be the greatest pressure that the boiler should work at when new, that is, at one-sixth the actual strength of the punched iron. resistance of wrought-iron plates. The Tables show at a glance what different diameters and thicknesses of plates are required for a safe working pressure, viz. one-sixth of their actual strength, when made of good material and workmanship. It should be observed, that the ends, if properly made and stayed, have only half the pressure exerted upon them that the diameter has, so that they have only to resist one-twelfth of their strength.' Having supplied sufficient data in connection with construction, we now proceed to the consideration of other circumstances closely allied to those of security and economy. In pursuing these enquiries, it will, however, be necessary for us to follow the same consecutive system that we have adopted on former occasions, that is, by dividing the subject into sections, as follows: : 1. On the proportion or relative value of the surfaces of the furnace to the absorbent surfaces as the recipients of heat. 2. On the safety valves and other adjuncts calculated to insure safety. 3. On high-pressure steam worked expansively. 4. On management. 1. On the proportion or relative value of the surfaces of the furnace to the absorbent surfaces as the recipients of heat. I have to observe that, on this question, there is diversity of opinion, as much depends upon the quality of the fuel used, and the rate at which it is consumed. We have at present no fixed rule for finding the proper proportion of the surface of the grate-bars to that of the boiler exposed to the action of heat. On these points a series of well-conducted experiments are much wanted, in order to determine not only the relative proportions, but also to ascertain the quantities of heat absorbed by the surfaces surrounding the furnace, and at different distances, as these surfaces recede from the immediate source of heat. On this subject we are quite in the dark. Every man on these points is his own doctor, and the wildest theories are constantly promulgated for the good of the public, and, as some would have it, for the advancement of science. It is now some years since I made an attempt to rectify these discrepancies; but I met with so many different forms, and such widely different proportions, as almost induced me to abandon the enquiry as a hopeless task. I however persevered; and taking the mean of fifteen boilers examined, I found that the ratio of grate-bar surface to that of the boiler surface should be as 1 to 11 nearly. This ratio varied from 1 to 9 up to 1 to 13, the mean being, as before observed, as 1 for the gratebar surface to 11 for the recipient surface. This investigation took place thirteen or fourteen years ago, before the introduction of the tubular system of boilers. On comparing the stationary, multi-tubular, marine and locomotive boilers, we find the following ratios of furnace to absorbent surface: These proportions, although differing from those in use by some engineers, may however be taken as approximately correct, and such as appear to be generally in use for obtaining the best results. I do not, however, give these proportions as conclusive; on the contrary, I entertain doubts of their accuracy, as the ratios are founded upon no fixed law, but taken from crude observations; they can only be considered as the mean results of general practice, which have yet. to be confirmed by actual experiment. On this question we must therefore not be led astray with the impression that the multi-tubular is either the best or the most economical in a commercial point of view. It may or it may not be so; I contend for it more upon the principle of diffusion and a saving of space than on any other property it may possess over the flue boiler; and for this simple reason, that we find a multi-tubular stationary boiler, 24 feet long and 7 feet diameter, generate as much steam as a flue boiler of one-fourth greater capacity. It presents nearly double the absorbent surface, but it does not from that cause follow that less fuel is consumed. Several competent judges contend that the expenditure is greater, and amongst them is Mr. C. Wye Williams, -no mean authority, who maintains, that so far as regards general efficiency, the flue system is capable of supplying all that is required, while it is free from the anomalies incidental to the multi-tubular plan. He states that when large quantities of steam are required for larger engines, this can be best obtained, not by additional tiers of tubes, but by extending the areas and length of run, thus increasing the number of units of time, distance, and surface along which the heat-transmitting influence may be exerted.' To prove this, Mr. Williams gives an example in the original boilers built for the Great Western steamer, having been replaced by others on the multi-tubular principle; and although the relative proportions of the heating surfaces were as 3860 to 7150 square feet, the former was the better and more efficient of the two.* Now, the only way to remove these doubts and differences, and to clear the question of all ambiguity, is to appeal to experiment. * See note to present edition, p. 73. |