The experiments on these tubes do not at first sight appear to yield very satisfactory results. The first, GG, gave way with a pressure of 150 lbs. on the square inch, when it began to leak so much as to cause its removal from the vessel, to replace some of the rivets which were imperfect. After the necessary repairs, it was again subjected to experiment, when it gave way with a force of 146 lbs., showing how much it had been injured by the previous pressure. On comparing it with the mean results of all the other experiments, we find that it should have borne about 300 lbs.: it evidently failed at the rivets, and cannot be relied upon. The next experimented upon was a steel tube, of the same form and with similar rigid divisions to those of the iron one. This sustained 220 lbs. on the square inch, when it bulged in or collapsed in the middle division. The last was a plain tube of similar plates of iron, 143 inches in diameter, but without ribs. This collapsed with a pressure of 125 lbs. on the square inch; and this agrees nearly with the preceding experiments, as will be seen. Comparing Experiments 32 and 33, it would appear that the steel tube is not stronger than the iron; but we are not warranted in drawing general conclusions from a single experiment. The next experiments were of a different character, upon tubes of an elliptical form. Table X. gives the results. The two experiments on cylindrical tubes are appended for comparison. On comparing the elliptical tube Bb with the cylindrical tube X, which are of the same length and thickness of plates, and only about half a square inch different in sectional area, we have for the collapsing pressure of the former 127.5 lbs., and for that of the latter 420 lbs., where it will be observed there is a loss of about ths of the strength, in consequence solely of the flattening of the tube Bb, or in other words, a cylindrical tube will support nearly three times the pressure which would collapse an elliptical tube of the same weight when proportioned like tube Bb. A similar deficiency is observable in tube Aa, when compared with tube T. The change of form, from the cylinder to the ellipse, where the diameter was reduced to 1 inch in one direction and extended as much in another, reduced the bearing powers one-half. The comparative results obtained from the experiments on the thick tube are different from those on the thin one, the loss being much greater in the former than in the latter case, although the ratio of the diameters is about the same. Allowance must, however, be made for inaccuracies of construction, though we might reasonably have expected a nearer approximation in the ratios of the deficiency of strength. From these facts, however, it is obvious that in every construction, where tubes have to sustain a uniform external pressure, the cylindrical is the only form to be relied upon, and any departure from it is attended with danger. Resistance of Tubes to Internal Pressure. During the investigation on the comparative resisting powers of tubes to collapse, a question arose as to the relative powers of cylindrical tubes to resist an internal force acting uniformly over their surface. It has already been demonstrated that the resistance of cylindrical vessels to internal pressure varies inversely as the diameters, but what effect the length may have upon the strength has yet to be determined. We have already seen that a cylindrical tube, when subjected to external pressure, loses one-half its strength when the length is doubled, and so on in other cases; hence arose the inquiry, what effect, if any, will an increase of length have upon a tube exposed to internal pressure? To solve this problem, three tubes of precisely the same diameter and thickness of plates, but of different lengths, were prepared and submitted to experiment as seen in Table XI. Considerable discrepancies occur in the experiments on internal pressure, as in each case the tube gave way at the riveted joint. Every precaution was taken, by carefully brazing them, to render them as nearly uniform in strength as possible. The weakness of these joints was, however, very apparent, and the results are in accordance with those arrived at several years previously, when it was found that the strengths of riveted plates were as the numbers 100, for the solid plate; 70, for the double-riveted joint; This constant failure at the joints renders the experiments on internal pressure very unsatisfactory, as they do not exhibit the ultimate strength of the plate, but only the strength of the joint; and as boilers invariably present joints, these facts are probably of some significance when applied to them. On a careful examination of the fractures, that of the tube Ff appeared the most perfect. Ee was not so well soldered, and burst by tearing off the rivet-heads, and Dd was torn partly through the plates and partly through the rivets; the plate of which this tube was composed was, however, exceedingly brittle, and broke like cast-iron. Tube Gg was ruptured in the same way and in the same direction as the others; the rivets were torn through the plates, and the soldering (not very sound) was ripped up for 10 inches along the joint: this tube, as also the others, would have borne a greater pressure had the joints been more perfect and of sounder workmanship. Comparing the tube Cc, 1 foot long, with the tube Ff, 4 feet long, and assuming the joints to be equally perfect in each, it would appear that there is a slight loss of strength when the length is increased; and this again suggests the question, do the rigid ends in short tubes. increase the strength of the unsupported portion in proportion to the length of the tube? For example, let us take two tubes of any given diameter, the one 10 feet and the other 20 feet long; it would appear, primâ facie, that a 2 it was much easier to force the long tube into the form of a barrel, as at a, than it would be to produce the n Table XII., :— › tube Hh ruptured at the thin part of the metal, the bursting through a narrow slit; Jj ruptured simiand on measuring the expanded circumferences at roadest part, it was found that the metal of the · had elongated 14 inch, and that of the latter 11⁄2 se experiments seem to show pretty conclusively, appears RALISATION OF THE RESULTS OF THE EXPERI MENTS. the reduction of the experiments, I have, as on occasions, been ably assisted by my friend Mr. whose sound philosophical views and high mathel attainments are, from his numerous publications, ll known to the public. To that gentleman I am ed for many services, and among others for an elainquiry into the specific gravity and properties of which I hope will be shortly forthcoming as a new on to our knowledge, and that more particularly in TABLE XII. Resistance of Lead Tubes to Internal Pressure |