ill ■Hi B ■iBRARY ■ NIVERSITYpF CALIFORNIA. A cce vv v 7/ v ATv. 3/3 j/ # / .S’.S’ ^ > jT V *r \ 3 / 3 ^ f „ T-., . to . > Welsh. £ ot Blaina, No. 3, \ % ot WSS, No. 3, Shropshire. This mixture is a strong iron, and therefore well suited for beams. Fig. 15. mm 1-570 Experiment IX. Ratio of the ribs, 1 to 4 j> nearly. Distance between supports and depth, as before. Dimensions of section in inches. Area of top rib . = l-05x’31 = 0-357 Area of bottom rib = 3*08x-51 = Thickness of vertical part = *305 Area of section . =» 3-37 inches Weight of beam 4 If lbs. Breaking weight 10727 lbs. = 95 cwt. 87 lbs. It broke by tension, 4 inches from the middle, but slanting towards it; and there seemed to be a small Haw in the bottom rib at the place of fracture. Here t jwg inch (sec fig. 12). Hence strength per in ch of section = 1 ~ = 3! 83 lbs. H 1 O.J37 Comparing this with the result of Experiment X., gives 3183 — 2792—391 ft excess.. ... 13 soi , .-. gam in strengthnearly. Kemakjl Though this beam had a larger bottom rib, it nevertheless broke by tension, or by tearing the bottom part first, which was evident, as it had neither been crushed nor broken by a wedge. This I had noticed to be the case16 ON CAST-IRON BEAMS FOR in every experiment. There had been gained j in strength, above that of the common beam by the addition already made ; and it was probable we might add still more to the lower rib without danger of fracture by compression; for in no case, except of the common beam, which sometimes twisted before it broke, had there been the slightest appearance of over-compression. This idea will be pursued in our future experiments. Experiment X. Common beam, cast upside down, in the usual manner. This, like the rest, was from the same model as that in Experiment IV. Distance between supports as before. Dimensions of section in inches. (See section, Experiment IV.) ‘ Thickness at A— -29 “ B = -125 “ C = *46 “ FE=2*3 “ DE= -53 Area of section =3’16 inches. Weight of beam=40£ lbs. Breaking weight=8823 lbs. It broke 1$ inches from the middle. The form of fracture was nearly as in Experiment IV.; here bn=225 and t r='8 (see fig. 12). 8823 Hence strength per inch of section=——=2792 lbs.* * The castings in Experiments IX. and X. were broken at four feet distance between props, on account of defects near the end of the castings ; the weight, however, was laid on the middle, 3 inches being taken off each end. The real breaking weights ■were 12068 and 9926 respectively ; those given above being the reduced ones to a span of 4 feet 6 inches. From this cause the deflections are neglected.SUPPORTING THE FLOORS OF BUILDINGS. 17 In the following experiments, the bottom rib is consjdjer'7! ably increased, agreeably to remarks made on Experiment IX.; but lest the top rib should be overpowered, and by its compression the point of support be thrown lower down the beam, and consequently the beam weakened, the top rib was a little strengthened likewise. The bottom rib will continue to be increased by small degrees, till such time as the beam breaks by compression, or by the separation of a wedge; at which point, perhaps, we shall have arrived at nearly the strongest form of section, for the same depth of beam and quantity of section. Experiment XI. Beam from model of Experiment IX., only its top ancR bottom ribs altered as above. Ratio of ribs 1 to 4 nearly. Distance between supports and depth as before. Dimensions of section. Area of top rib = l‘6x '315=0*5 inches. Area of bottom rib=4*16x -53=2-2 „ Thickness of vertical parts=*38 „ Area of section=4'50 inches. Weight of bcam=57 lbs. Deflection with 11186 lbs. *4 inches. „ 12698 „ -45 „ „ 13706 „ '52 „ Breaking weights 14462 lbs. = 129 cwt. 14 lbs. o o It broke by tension 1 inch from the middle ; b n—2*5j inches (see fig. 10). Hence strength per inch of section=lli®£= 3214 lbsi 4'5 Comparing this with the result of Experiment XIII., which bore 2693 lbs. per inch— Frg. 16.18 ON CAST-IRON BEAMS FOR 3214—2693=521=excess. . •. Gain in strength=i—U Inearly. ° 2693 , 6 We may seek for the gain by comparing the weights of the two beams, and the quantities they bore :* thus, since in Experiment XIII. the weight of the beam was 41 lbs., and it broke with 8942 lbs.; and the weight of this beam 57 lbs., and,its breaking weight 14462 lbs.; hence 41 : 57 : : 8942 : 12431=weight this beam should have borne, according to the strength of the common beam; but it did bear 14462. I*, 14462—12431=2031=excess, and d&litfin strength—-03— I - nearly. ■ J 6 12431 6 J * Let ?y=the weight of a beam of uniform dimensions, w' = the weight of a cubic loot of iron, then we readily find from Ex. 3) p. 11. ■ € dG> " « P ' 144m.' pi W P » " d where the value of Cr determined by experiment for any particular form of beam, enables us to ascertain its comparative strength. If l and d are constant, then W W Similarly we have, yr Wro, C| Wjeo ■which expresses the comparative strength of any two beams of the same length and depth. Taking the above examnle. we have '~ ^ o i 406 •012 2 1574 •120 3 2470 •152 4 3366 •223 5 4262 •284 6 5158 •326 H I 5606 •400 8 6056 •470 9 6502 •528 10 6950 .571 11 7398 •620 12 7 846 •670 13 •8294 •735 14 8742 •780 15 8854 Broke. Ultimate ) deflection f •790 3* "42 ----^.5---- Bottom flanch . . l-0 Top flanch .... ‘20 Depth of beam . . 4 Bottom flanch, 2’5 X '42, area=I'05 inches. Top flanch 1'0X'20, area='20 inches. Depth of beam in the middle, 4 inches.48 ON COMPOUND OR TRUSSED In the first experiment one of the truss-rods broke after sustaining the weight, 5502 lbs., some seconds. The second and third broke, through the upper flange, by compression. Taking, however, the mean of the whole, we have the breaking-weight at 7433 lbs. Owing to the failure of one of the truss-rods, the experiment was twice repeated with f-inch rods, in order to produce fracture in the cast iron, before the rods could yield to tension. Experiment IV. The same beam, with the broad Jlanch hel m, and without truss-rods, 4' 6ff between supports. ' No. of Experiment. | Weight laid on in lbs. Deflection. in inches. 1 406 •70 2 1574 1*50 3 2022 2-86 4 2470 3-26 5 2918 1 4-20 6 3366 4-91 7 3814 5-60 8 4262 6-31 9 4710 6-82 10 5153 7-25 11 5382 7-50 12 5606 7-7 6 13 5830 Broke. Ultimate 1 deflection j 7.86 Fig. 26. CAST-IRON BEAMS OR GIRDERS. 49 Experiment V. The same beam, loith the broad flanch above and two truss-rods, three quarters of an inch diameter, 4 ft. 6 in. between the supports. No. of Experiment. Weight laid on in lbs. Deflection in inches. i 406 •03 2 1,574 •07 3 2,470 TO 4 3,366 •16 5 4,262 •20 6 5,158 •24 7 6,054 •30 8 6,950 •36 9 7,846 •47 10 8,742 •53 » 9,638 •61 12 10,086 •69 13 10,534 •77 14 10,972 •84 15 11,420 •91 16 00 Oi GO 1-02 17 12,316 Ultimate ) deflection, j 1-06 The beam in this position would have carried a much greater load if the truss rods had been stiffer, as the lower part of the beam was torn asunder from the increased deflection, before the top flange had arrived at its ultimate powers of resistance to compression. B50 ON COMPOUND OR TRUSSED Experiment VI. The same beam reversed, with the broad fianch uppermost and without truss-rods. 4: ft 6‘ Vft. (ktween supports. No. of Experiment. Weight laid on iu lbs. Deflection in inches. 1 406 *080 , 2 1,574 •132 3 11 2,022 *226 4 2,470 *362 5 2,918 •473 6 3,14*2*. •514 7 3,366 Broke. ullirriate ) deflection. } •55 The experiments above recorded present for consideration the strengths of beams, with the following varieties of forms and dispositions. 1st, The cast-iron beam, in its best form, with the broad fianch below, when assisted by a wrought-iron truss. 2d, the same beam reversed, with the broad fianch above, and supported by a truss. 3d, The same beam, broad flaach above, and without a truss. And, lastly, the same beam, in its strongest form and position, also without a truss. Now, all these conditions admit of certain degrees of comparison. In working them out in detail, we have the following results and ratios of strengths:—CAST-IRON BEAMS OR GIRDERS. 51 Summary of Results. Sketch of Beam. jwq) Mag m iffli Description of Beam. Breaking Ratio of weight in lbs. strength in lbs. Cast-iron beam, with the broad flange downwards The samp beam in the same-] position, with double \ truss-rods supporting the | middle.......................J Beam reversed; the broad fianch uppermost without truss-rods................... The same in the same position, with broad flanch supported by truss-rods Beam with double' truss, ) as before ; broad flanch > below........................\ The same, with double 1 truss; broad flanch up- >• permost..................\ Beam without truss-rods ; broad flanch uppermost . The same beam, broad flanch downwards . . 5,380 7,433 3,366 "j 12,316 j 7,433 12,316 3,366 ^ 5,830 100:138 100 : 365 100!165 100 : 17352 ON TRUSSED BEAMS. From the above summary we may draw the following conclusions:— 1st, That the advantage gained by adding truss-rods to a cast-iron beam of the strongest section, and placed in the best position for resisting a transverse strain, is as 100:138, being rather more than one-third of increase in strength.* 2d, That the simple beam reversed, with the small flanch downwards, loses about one-half of its strength, as compared with the same beam in its most favourable position, with the large flanch downwards, or, as 100 : 173. Again, let the beam be reversed and trussed, with the small flanch downwards, and its resisting powers are increased 3£ times in strength as compared with the same beam in the same position without the truss, or, as 100 :365. Lastly, that the same beam, being trussed in both instances, first, with the broad flanch downwards, and subsequently with the broad flanch upwards, gains above three-fourths in strength in the latter case; whilst the other is not materially increased beyond that of the simple beam entirely free from auxiliary support. We might multiply these comparisons to a much greater extent; but we have done sufficient to prove the fact, that under the most favourable circumstances, there is not much gained in the strength of cast-iron beams by the addition of malleable-iron truss-rods. When such auxiliaries become absolutely necessary, I would then recommend them to be attached to beams with a strong flanch on the upper side to resist compression, and the tension-rods so regulated and proportioned in strength as to cause them to act simultaneously with the rigid top in their resistance to fracture. * For an account of the dangerous nature of a beam trussed in this manner, see Appendix No. III.ON CAST-IRON BEAMS. 53 What is, however, infinitely preferable, is a well-constructed malleable iron beam, which may be made of almost any given strength, and of any span within the limits of 500 to 1000 feet. General Remarks relative to Cast-iron Beams. I have probably bestowed more attention to this part of the subject than may at first sight appear necessary. It must, however, be borne in mind, that many serious accidents have occurred from an ignorance of the principles on which cast-iron beams and girders are constructed; and I may perhaps be permitted to hope, that the present investigation will not be without its use in those constructions, which have for their object the economy of material and greater security in the construction of buildings demanding the most careful consideration on the part of the architect and the engineer. Should these objects be accomplished, and more correct views of the principles of construction be established, I shall consider myself amply recompensed for the time and trouble which I have expended upon the investigation of the subject. Before entering upon another important division of the inquiry, that of the strength of wrought-iron tubular girders, as applied to bridges, I shall shortly advert to the application of wrought-iron beams for the support of thefoors of buildings, and to other purposes requiring solidity and security from fire. To the investigation of this part of the subject we bring with us a large amount of existing knowledge of form and construction, which for the last twenty years have guided our efforts in this department of practical science. To Mr. Hodgkinson’s able and conclusive inquiry into the strength of cast-iron beams, we are in-54 CONCLUDING REMARKS debted for many useful formulae and other aids in the art of construction. Cast iron cannot, however, be depended upon, even in the best forms, for several reasons, — viz. unequal contraction in the. cooling of the metal, the brittle nature of the material, imperfections and flaws in the castings, and its liability to break without warning. As regards the first point, we labour under great uncertainty in consequence of the “ shrinkage ” or contraction of metals during the process of cooling. A casting, even when well proportioned, will suddenly “snap” without any apparent cause; or exposure to rain, or intense frost during the night, not unfrequently produces fracture; and on these occasions rupture*takes place with a loud noise, like the report of a pistol* On minute examination the injury is at once seen to have arisen from the presence of an immense tensile strain in the immediate vicinity of the fracture, which is generally found to be greatly enlarged, and an enormous force-isr required to bring the parts again into contact. This unequal and dangerous force of ten®k>n, existing within the casting itself, appears to me to be produced by one of two "causes,—-either from unequal rates in the time of cooling, whereby the crystalline process is seriously deranged, or from imperfect mixture of the metals, wrhence the “ shrinkage ” is greater in one part than in another, and from which would follow unequal degrees of tension of the parts. Great care should therefore be observed in castings ; it should be seen that the metals are w ell mixed, and that the moulds and patterns are so proportioned as to insure uniformity in the rate of cooling. These are practical operations of some importance, and the moulds, after running, should be covered closely up, and as much time as possible given for attaining, by an equal rate of cool-ON CAST-IRON BEAMS. 55 ing, a greater degree of perfection in crystalline structure.* The second cause of danger is, that all crystalline bodies are of a more brittle and uncertain character than those which are of a fibrous structure ; and as wrought iron possesses more ductility, and partakes in a greater degree of the latter quality, it is better qualified to sustain heavy weights and shocks than cast iron ; and its high powers of resistance to a tensile strain render its application in the constructive arts an object of primary importance to all those connected professionally or otherwise with the erection of buildings. The superiority of its resistance to tension is not, however, its only recommendation, as the new forms and conditions under which it can be manufactured and applied, in position and distribution, to resist compression, is another powerful recommendation of it as a safer and lighter substitute for cast iron. Another defect of cast iron is the impossibility of discovering imperfections which may lie concealed under the surface of a casting, and which frequently baffle the scrutiny of the keenest observer. These defects are by no means uncommon; and repeated instances have occurred wherein castings, presenting every appearance of perfection, have been found to contain the elements of destruction either in concealed air-bubbles, or in the infusion of scoriae, which had been run * Tredgold appears to have been perfectly aware of the dangerous nature of unequal cooling. In his remarks on the quality and appearance of the metals, in his treatise on the strength of cast iron, he states, p. 8, “thatthe utmost care should he employed to render the iron in each casting of a uniform quality, because in iron of different qualities the ‘shrinkage’ is different, which causes an unequal tension among the parts of the metal, impairs its strength, and renders it liable to sudden and unexpected failure.” At another part he observes, “ I must not omit to remark, that cast iron, when it fails, gives no warning of its approaching fracture, which is its chief defect when employed to sustain weights or moving forces.”56 CONCLUDING REMARKS into the moulds, and skimmed over by a smooth covering of apparently sound iron. Now, this can never occur in the wrought-iron beam, as the different processes of manufacture, such as puddling, forging, piling, and rolling, are sufficient to cauSS any imperfection, calculated to endanger the soundness of tile plates, to be detected. It will, however, sometifdtis occur, that minute particles of scoriae will be inserted between the laminae or bars from which the plates are rolled; but this does not materially affect the strength, excepting only in the case of boilers, where they form blisters when exposed to intense heat. In the formation of beams, these defects are of less consequence, as they do not seriously impair their strength. On the Influence of Time and Temperature on the Strength of Cast-iron Bars. Before closing our remarks on cast-iron beams, it may be advisable to introduce a few experimental facts in connexion with two very interesting questions on the strength of materials, involving considerations of great importance, viz. the influence of time and of temperature on cast-iron, or the extent to which the&e agencies respectively affect its power of resistance to a force tending to sever or rupture its parts. The$£;questions occupied my attention some years ago, and considering that they bear directly upon the question now before us, it may be, instructive if we give a few extracts from the espuiiwents,* which were then considered of great value, and,which, lgd to several important results. On the Influence of Time. It has always been a question of doubt how far a body, * For a more detailed description of the effects of time and temperature upon iron, inn reader is referred tQ jJrF report in the sixth volume of the TransaptiooiB of the British Association for the Advancement of Science.ON CAST-IRON BEAMS. 57 say cast iron, can be loaded without impairing its powers of resistance. This question is still imperfectly understood, as all bodies are surrounded with many causes of deterioration and disturbance, which affect their permanent condition, and which lead by slow but certain degrees to their ultimate destruction. Meteorological influence, temperature, and time, are all elements apparently at work in that direction; and it is curious to ascertain which of them has the greatest effect upon the stability of a material so extensively applied to the arts of construction aB tftfit iron. In order to solve this problem, I entered in the year 1837 upon a series of experiments, which extended through a course of eight years ; and as these experiments affect formerly acknowledged theory on the strength of materials, it may be essential not only to state the results, but to apply such deductions as were derived from the experiments as they were recorded from time to time. Effects of Time. In the report read before the British Association for the Advancement of Science, the following introductory observations on the effects of time on loaded bars occur. In former experiments on the transverse strength of cast iron, it has been assumed that the elasticity remained perfect to the extent of at least one-third of the breaking-weight, and that it should never be loaded with more than that weight. This assumption, which has been attempted to be proved by Tredgold, has gained considerable credence; but so far as I can perceive, there appears to be no ground for such an opinion. In some early experiments on the comparative, value of hot and cold blast iron, it was observed, that in most cases the elasticity was considerably injured with one-fifth or one-sixth of the breaking-weight.58 CONCLUDING REMARKS This fact was of such importance as to induce me to pay particular attention to the set, as indicated by the preceding experiments, and also to note the defects of elasticity in those that follow, up to the time the weights became permanent upon the bars. From the method thus pursued, it wl be seen that the value of the set has been given with the deflections, at regular intervals, from the commencement of the experiments to the time of fracture, and the connexions between the weights, deflections, and sets, will therefore in all probability be better observed. The early period at which the elasticity became injured caused an additional series of experiments to be instituted, for the purpose of ascertaining whether or not such injury to the elasticity would, with the load continued, ultimately break the bar. This important question could only be determined by experiment. The inquiry, therefore, resolved itself into the question, —to what extent cast iron could be loaded without endangering its security ? This was a question of great importance ; one which involved a consideration of much interest, such as the stability of bridges, warehouses, factories, and other constructions to which cast iron is applied. In computing the bearing powers of cast iron, it has always been considered unsafe to suppose it loaded beyond one,-third of its breaking weight, and in order to be on the safe side, this is not an ynwise precaution; but in some £&ses, su$h as bridges, beams for warehouses, &c., it is desirable to calculate only one-fifth or one-sixth of the breaking weight, as the flaaterial may be subjected to accidental strain, arising from the force of impact, or from any other force acting unfavorably upon it. Taking the experiments, however, as a criterion, it will be found that out of ten rec-ON CAST-IRON BEAMS. 59 tangular bars, each 1 inch square, of cold and hot Coed Talon iron, certain results were obtained after the bars had been subjected to the following loads as permanent weights suspended from the middle of each bar. Table I. Table of Coed Talon rectangular bars (4 ft. 6 in. between the supports) loaded with different weights for determining the changes, if any, luhich take place during indefinite periods of time, the mean of the breaking weights of each sort of iron having been previously ascertained to be, for the cold blast 508 lbs., and for the hot blast 484 lbs. No. of bars. Permanent load in lbs. Mean breaking weight in lbs. Ratio of breaking weight to load. Remarks. i 280 508 1 : 551 2 336 508 1 : 661 3 392 508 1 : 771 - Cold-blast iron. 4 448 508 1 : 881 5 448 508 1 : 881 6 280 484 1 : 578 7 336 484 1 : 694 8 392 484 1 : 805 > Hot-blast iron. 9 448 484 1 : 925 10 448 484 1 : 925 From the above will be seen the nature of the experiments which were instituted for the purpose of ascertaining, by an exceedingly minute scale, the increase of deflection which from time to time took place in the bars. If that increase was progressive, it was then inferred that rupture must at some time or other (however remote) take place ; if otherwise, that a new arrangement of the parts under strain had taken place, and that they had thus become fixed at a power of resistance equivalent to the load.60 CONCLUDING REMARKS The results from March 1837 to June 1842 were as follows: Table II. Results on bars No. 2 and No. 7, loaded with 336 lbs. Tempe- rature. Date of observation. Cold blast, deflection in inches. Hot blast, deflection in inches. Rates of increase. March 11, 1837 . 1-270 1-461 Previous to the time of taking —T 00 o June 3, 1838 . . 1-316 1-538 the deflection in November -T to o July 5, 1839 . 1-305 1-533 and April the hot blast bar 61° June 6, 1840 . 1-303 1-520 had been disturbed. O O iO Nov. 22, 1841 . 1-306 1-620 58° April 19, 1842 . 1-308 1-620 Mean . . . 1-301 1-548 The above experiments show a progressive increase in the deflections of the cold-blast bar during a period of five years of'031, and of *087 on the hot-blast bar. Table III. Results on bars No. 3 and No. 8, loaded with 392 lbs. Tempe- rature. Date of observation. Cold blast, deflection in inches. Hot blast, deflection in inches. Remarks. March 6, 1837 . 1-684 1-715 O 00 !>» June 23,1838 . 1-824 1-803 72° July 5, 1839 . 1-824 1-798 61° June 6, 1840 . 1-825 1-798 O O lO Nov. 22, 1841. . 1-829 1-804 58° April 19, 1842 . 1-828 1-812 53° Mean . . . 1-802 1-788 ON CAST-IRON BEAMS. 61 During five years the deflection in the cold-blast bars has been rather more than in the hot blast, the increase being as 1*802 to 1*788 ; whereas in the former table* with lighter weights, the increase was on the other side, or as 1*301 in the cold blast to 1*548 in the hot-blast bars. Nevertheless, the deflections in this case indicate, as before, a steady increase of *118 for the cold blast, and *073 for the hot blast. Table IV. Results on bars Nos. 4, 5, 9, and 10, loaded respectively 448 lbs. Tempe- rature. Date of observation, j Cold blast, Hot blast, deflection deflection in inches, in Remarks. March 6, 1837 . 1-410 78° June 23,1838 . . 1-457 1 . . Both the hot-blast bars broke at once with 448 lbs., and 72° July 5, 1839 . . 1-446 one of the cold-blast bars 61° June 6, 1840 . . 1-445 broke after sustaining the 50° Nov. 22,1841. . 1-449 weight thirty-seven days. 00 April 19, 1842 . 1-449 53° Mean . . . 1-442 The progressive increase in the deflections in this case is *032; which, it will be observed, is much less than that exhibited in the former table with weights of 392 lbs., &nd less than the deflections on the hot blast, which were *073 with the same weight continued as a permanent load. Viewing the whole of these experiments in relation to the solution of a problem affecting the laws which regulate the resistance of bodies to continuous strain, it is important to observe how admirably the cohesive powers of matter adapt themselves to circumstances, and with what tenacity they resist forces tending to dissever and rupture their parts.62 CONCLUDING REMARKS It is still a question for consideration how far this power extends, and whether or not bodies, when loaded within even part of their breaking weight, would sustain the load for ever, provided that no disturbing cause were present to produce fracture. I am strongly inclined to think that such would be the case, notwithstanding the fact that the whole of the loaded bars exhibit a progressive increase of deflection, which fact I am disposed to attribute to the vibrations continually going on in the building where the bars were fixed, and to those atmospheric changes, such as temperature, oxidation, &c., to which every description of material is subject. In the experiments just enumerated one important fact has been fully established, namely, that a continued and strictly permanent pressure, even wrhen about to produce fracture, ife very different from the law of defective elasticity caused by changes affecting the conditions of material, such as the increase and diminution of pressure, producing a disturbing force on all the parts under strain, and thus, by a continued series of alternations, eventually destroying the powers of resistance. In the former* case the load, however near it approximates to that which is necessary to fracture, remains permanently fixed; whereas in the latter, the changes, however minute, wiEL if continued long enough, ultimately lead to destruction. Mr. Ilodgkinson’s experiments, as well as my own, lead to this conclusion ; and I have no doubt that any load (however small) producing a permanent set upon a bar will, if taken off and replaced a sufficient number of times, eventually break it. For example, let us take the bars supporting the lightest weight, 280 lbs., and suppose that the load to the extent ofON CAST-IRON BEAMS. 63 200 lbs. is removed, and then replaced at intervals of thirty seconds, it is evident that this change, often repeated, will in the end destroy the cohesive powers of the bar either in the lower part of its crystalline extended section, or in that of the upper, where the crystals are compressed; or, what is more probable, the destructive process would be progressive in a given ratio in the upper and lower sections of the respective means of compression and extension. This stant movement or sliding of the atoms Or parts of crystalline, as well as fibrous bodies, 'is thefofore the cause of breakage; and—assuming a change to tal$|$place in the molecules of the body—however slight the strain may be when first applied in one direction and then changed to another direction, it only becomes a question of time, how long the body will bear theSfe continued repetitions before rupture takes place; sooner or later fracture must take] place. I am further confirmed in the truth of these obsemp-* tions by the experiments now in progreg§g which bear eO forcibly upon this peculiar aspect of the, strength of materials, that I am induced to notice them in this place. I do so in order that we may become fully aware of the principles on which the security as well as the’durability of the material depends. It is important that all these' facts should be generally known, as there:is a wide differ^ice between the bearing powers of a beam exposed to changes of pressure, and one which has to sustain a perfectly quiescent and permanent load. On the Effects of Temperature f The principles of the effects of temperature upon cast iron are pretty fully developed in my Report, published in the sixth volume of the Transactions of the British Associa-64 CONCLUDING REMARKS tion for the Advancement of Science. The experiments therein contained are very conclusive; and those on time would have been less satisfactory if those on temperature had been omitted. In that communication it is stated that Rondelet, in his 9 'aite de Bdlir. had given and collected results from ex-periments made by himself and others on the expansion of bodies from the effects of heat; but I believe I was the first to ascertain the strength of metallic substances at different temperatures; and although the effects of heat upon metals had not escaped the attention of philosophers, yet I am not aware that any writers on this subject have conducted their experiments in a way at all analogous to that under consideration. Had time permitted, it was my intention to have pursued the experiments on temperature under a much greater variety of form and change. For example, it might have been desirable not only to load the bars until they were broken, but also to charge them with different weights ; and by alternate heating and cooling to ascertain how the bars so charged were affected by the change. Such an extension of the experiments might have led to the development of some new law, and that more particularly as regards the changes produced by the alternate increase and decrease of temperature. On some future occasion I may perhaps have an opportunity of returning to these interesting inquiries; they involve considerations of great importance, as well in theoretical as in practical science, and I have no doubt that they would be found of great value in every case where the material is exposed to frequent changes of temperature. For the present, however, it will be sufficient to give the comparative results which were obtained from the experiments.ON CAST-IRON BEAMS. 65 Table V. Comparative strength and power to resist Impact of the Coed Talon hot and cold blast irons at various temperatures. Transverse Strengths. Temperature. Coed Talon cold blast. Coed Talon hot blast. Ratio. Falir. No. 2 Iron. No. 2 Iron. 26° 851- 823-1 . . . . 1000 : 967-2 32° | 940-7 ) >• Mean 949.6 958-5 ) (933-4) -< (-Mean 919-7 (90G-0) 1000 : 977-6 190° 743.1 . . . . 833-6 . . . . 1000 : 1108 Red in dark. 723-1 • • • . 839-7 . . . . No. 3 Iron. No. 8 Iron. 212° i 1 G00° || 905-0 ) y Mean 924-3 943-6 ) 909-3 ) y ,, 1033-1 1157-0) l j 818-4 . . . . ( 834-1 ) V Mean 875-8 917-5 ) 1000 : 885-4 1000 : 847-7 Power to resist Impact. Temperature. Coed Talon cold blast. Coed Talon hot blast. Ratio. Falir. No. 2 Iron. !zS o to hH 0 1 26° 349-8 .... 340-8 .... 1000: 974 2° ^ ! 360-3 ) VMean3°2-4 404-5 ) ( 406-9 ) •< [ Mean 395-0 ( 383-2) 1000 : 1032-9 190° 223-7 . . . • 298-9 . . . . 1000 : 1336 F66 CONCLUDING REMARKS Modulus of Elasticity in pounds for a base of 1 inch square. Temperature. Coed Talon, fchld blast. Coed Talon hot blast. Fahr. No. 2 Iron. No. 2 Iron. 2 ' ■■■ ^St,994,400 ...... 14,267,500 32s | 18*606,700) V Mean 14,327,450 15,143,200 ) ( jk?23,500) Uleatt 14,003,3501 ( 14,283,200) ‘ 1906 14,39S,6tfO . . . . * 13,^000* 1 In pursuing the investigations, it unfortunately happened that the stock of No. 2 iron became exhausted, a circumstance whic|i intercepts the comparison from 6 degrees below the freezing-point to that of melted lead. The No. 3 should have been broken at all degrees of temperature* in order to ascertain the loss of strength as the heat was increased. It was not, however, accomplished ; and from this circumstance the comparison only holds good between the two qualities No. 2 and No. 3 from the boiling-point of water, or 212°, up to 600° of Fahrenheit. In thfe No. 2 iron it will be observed that thet strength continued to diminish as the temperature increased ; whereas in No. 3 it increased, as shown in the Table, from 921*3 to 1023*4, which can only be accounted for from the irregularity and greater rigidity of that description of iron. On the whole, we may infer that cast iron of average quality loses strength when heated beyond a mean temperature of 120°, and that it becomes insecure at the freezing-point, or under 32° Fahrenheit. On the subject of the mixtures of the different kinds of British iron for the purpose of producing casting® suitable to the purpose for which they are intended, we have no fixed or determined rule from which we can obtain any thing like correct resultf ■ Every iron founder appears to exercise his own judgment in those matters; and it is very difficult toON CAST-IRON BEAMS. 67 obtain castings under conditions where the proportions are specified, unless prepared under a strict and rigid surveillance. Iron-founders, managers, and furnace-men, appear to work under the impression of the non-importance of attention to the mixture of metals; and hence arise all the anomalous conditions of soft and hard, strong and weak irons, and many other disqualifications which might easily be avoided by closer attention to the quality and due proportion of each particular iron, and the quantity of carbon and flux used in its liquefaction. All these are considerations of deep importance in the art of founding; we have yet much to learn in the preparation of metals, as well as in the manipulative art of moulding and the necessary process of ventilation, —a process which requires no inconsiderable degree of thought and skill. It has always been an opinion that cast iron is improved by mixing ; and doubtlee$ this is the case, as we have almost every variety in the pig,—soft, hard, ductile, rich, and poor, as well as in the white, blue, grey, &c. irons, all of which are adapted in combination to form almost any description of metal required in the useful arts. They are, moreover, calculated to effect great improvement in the quality of the castings, and form compounds which, with proper care, may be varied according to the uses for which the castings are intended. In the case of beams and bridges, a mixture of two-thirds strong Welsh No. 3, a portion of Scotch or Slalfordshire No. 2, and a proportion of old iron, will form a good mixture. Other strong irons may be used, such as the finely granulated Scotch or Staffordshire No. 3; a small portion of No. 4, and about one-fourth or one-fifth of Hematite, answers the same purpose. These compounds are of great value when it is essential to have strong metal;68 CONCLUDING REMARKS but in ordinary cases, the mixture may be left to the discretion of those accustomed to the management of the cupola and the furnace. In this stage of the inquiry it will be proper to notice a mixture which, above all others, gives indications of superior strength. In the summer of 1847, Mr. Owen, the Supervisor of Metal to the Admiralty, was kind enough to forward to me a copy of his experiments upon Mr. Morries Stirling’s toughened iron, whereby the tensile strength of cast iron (according to Mr. Morries Stirling’s statement) is nearly doubled. This process of toughening cast iron is by an admixture of wrought iron fused simultaneously along with pigs in the cupola or air furnace. Mr. Owen gives the results from both processes, and the experiments being somewhat analogous, I have taken the liberty of referring to those which bear more immediately upon the subject of the present inquiry. Mr. Owen’s experiments were made on a large scale. The girders were 17 feet long and 16 feet between the supports, as in the annexed section (taken through the middle, wh$re the weight was laid on). The girders were constructed upon Mr. Hodgkinson’s principle, and weighed about 15 cwt. each, and their computed breaking weight was 39£ tons. The girders were made of cast iron prepared according to Mr. Morries Stirling’s mixture, as follows : Cupola. cwt. Breaking weight Russell’s Hall hot-blast Ho, 2 Staffordshire . . . 15 ^ Prior Field No.. 2 Staffordshire..........20 v 50i tons, \Yrought-iron scrap.......................6 )ON CAST-IRON BEAMS. 69 Air Furnace. The mixture above given sustained 51 £ tons, giving an increase of strength of 2 per cent, in favour of the air furnace. The experiments with seven different e&stings made entirely from the Russell’s Hall No. 2, Madeley Wood No. 3, Colebrook Vale No. 3 (Welsh), and Calder No. 1 (Scotch), give an average of 33£ tons as the breaking weight. Comparing the results of these experiments, we find the value, as regards strength, to be as 33*25 to 5T5, being the ratio of 1 : 1*55. It will not be necessary to extend thole examples beyond the following list of comparative results, as obtained from the series of experiments made upon each kind of iron : Unprepared Cast Iron. Toug hened Cast Iron. Breaking Weights. Breaking Weights. 1 . . 30 tons. l . . . . . 52J tons. 2 . . 85 „ 2 . . . . *.. eoj „ 3 . . 33 „ 3 . . ... 48 4 . . 34 „ 4 . 5 . . 33£ „ . o . , . . . . 52J „ 6 ■ I S4J 1 6 . . . . . GOJ „ If . . Hr 1 . 8 . . 46J „ B . . . , . ft 9 . . 41' „ 9 . . ...... 56 „ 10 ■ l 47* „ 10 . . . . . 48£ „ 11 11 . . ... 52 V. \ 12 . . 36£ ,J|j 13 • ■ 38| * Mean tons . . 38-3 Mean tons . . 52*3 T he ratio given by the above table is lower than that given by the first, being as 1:1' 36; but it is nevertheless sufli- cient to show that the mixture i decidedly improves the bear- ing powers of a casting when subjected to a transverse strain.70 CONCLUDING REMARKS Almost simultaneously with, or shortly after, Mr. Owen’s experiments, I was called upon to witness some experiments on square bars, made by my friend Mr. Lillie, of Store Street, Manchester, which had an admixture of wrought iron chiefly from turnings. They appear to fuse and combine with cast iron in the cupola in variotis proportions; and having been present at sbme of th6' experiments, I can vouch for the great superiority of strength which Mr. Lillie’s mixture seemed to indicator Mr. Lillie’s experiments were made upon cast-iron bars 3 feet longfy 1 inch square, and 2 feet 10 inches between the supports* * Somber of Experi- ments. Quality of Iron. Breaking weights in lbfV-U DefleetEro in inches.. Remarks. 1 ( %k>mpoaed of ; \ \ Gartsherrie iron J .1^831 •625 2 ; j Comuos^doOlr.) | Lillie’s Mixture j 1S4&' * * -16 0 ( wrought iron ) -] bar, same fe>h; V ( 1 i nch squa re ,) 1008 -625 f With an additional weight the J bar bent and destroyed, j so as to render: it incapable l of bearing the load. .» * From the abovA Table it appears that the mixed or toughened iron, as prepaid by Mr. Lillie, is increased in its transverse strength more than one-third, as compared with cast iron, and nearly orj^eighth as compared with wrought iron. It is, however, to be regretted that Mr. Lillie did not extend hi8 experiments so as to give the comparative values of the different irons by reducing the bars experimented upon trfexactly one inch square. It would also be of great value to ascertain the relative tensile strain and crushing power of these mixtures, as there cannot exist a doubt as to the increased strength, either transverse or ten-ON CAST-IRON BEAMS. 71 sile, of cast iron, when fused with a due proportion of wrought-iron borings, turnings, or «Crap iron. Entertaining these views, it may be Worthy of consideration how far it may be advisable to institute a series of experiments on the subject of these mixtures, and not only to determine accurately tire comparative strengths of cast iron in combination with certain portions of wrought iron, but to extend our knowledge generally on a subject which is still exceedingly imperfect; in fact, such was our ignorance on this subject, that until Mr. Morriffi| Stirling exhibited these properties of combination, it was supposed that wrought iron would not fuse and combine with cast iron. Under circumstances whei^ the toughened iron is not used, the following mixture will be found of great value in castings, such as girded for bridges, beams for buildings, &c., where rigidity and strength are required. LoW Moor Yorkshire^ No. 8 . 30 Bmina or Yor ksliire, No. 2 „' I Shropshire or Derbyshire, No. 3 . ' . 25 And good old jeemp . . « . 20 100 The above mixture will make what may be considered castings of superior strength; but it seldom happens that this mixture can be obtained, on account of the high price of Low Moor iron; hence it is that we can rarely depend upon iron founders for the introduction of the eyact quantity necessary to produce the required results', TheM matters arc generally left to subordinates; and either through ignorance, or to save themselves trouble, they almost invariably take up what comes readiest to hand, and thus defeat not only the calculation of the mathematician, but the hopes of those who confide in them.72 CONCLUDING REMARKS There are other combinations or mixtures of iron which possess other properties besides those of strength, such as the Scotch and Staffordshire iron for light work and for castings for machinery, No. 1, or the richer description of iron, which is easily worked, and is more ductile than the strong Welsh irons. In practice these mixtures require great consideration, as the success of the particular manufacture in some measure depends upon the castings produced. It would occupy too much time to enter upon all the questions connected with this subject. Suffice it to observe, that, after having made the selection and determined upon the mixture, much even then depends upon the skill and care of the furnace-man, particularly in attending to the temperature of the furnace, and the degree of heat at which the metal is run into the mould. All these are considerations which it would be superfluous at this time to discuss, but which enter largely into the formation of the crystalline structure, and which on no account should be neglected in the production of castings having for their object the union of the properties of fusibility and strength. We may further observe that the anthracite is a strong description of iron, and will be found useful for mixing where rigidity and strength are regarded. Experiments on this iron will be found in the sixth volume of the second series of the Manchester Memoirs. The following Table, the result of many years’ labours, shows the transverse strengths, and many other properties, of nearly all the irons of the United Kingdom, and may be appropriately introduced as the conclusion of this part of the subject.72* EXPERIMENTAL INQUIRY ON THE STRENGTH &c. OF CAST-IRON BARS. General Summary of Results of Experiments on Rectangular Bars of Cast Iron, each Bar being reduced to exactly inch ‘square. one In the foMving abstract the transverse strength, which majibe taken as a criterion of the value of each Iron, is obtained from the mean of the experiments, first on the long bars, 4 feet 6 inches b^E^fflae supports, and next on those® half the length, or 2 feet 3 inches between supports. All the-other values are deejXmtfrom, the 4 feet 6 inch bars. scale of strength. Names of Iron. No. oa each. Specific yemmty. Modyhfeqf per scJ^P^wch, * or stiffness. Breaking weight Am lbs. of bars f4lft. 6 in. mmeen supports'. Breaking weight in lbsJjoibars 2 ft. 3 in. 4 ft. 6 in. between si&ybt!« Mean breaking weight in lbs. p9 Ultimate deflectiofK of 4 ft. 6 bats in parts of an inch. Power of nra 4 feet 6 inch bars resist impact. Colour. Quality. Ponkey, No. 3, cold blast . . . . 7-ijjSPJj 567 595 581 1-747 992 Whitish gray Hard. 2 Devon, |||. 3, hot blast* . . . . 2 ®7-251 £v$2,473,65$^ 537 537 1-090 589 | White . . . 3 OldEsly, No. 3, hot blast .... 5 '7-500 |yg^733,4o|^ 543 537 530 1-005 549 » * • • . V 4 Carron, No. blast* .... | 1-056 17,873,100 520 524 527 1-365 710 Whitish gray 5 E£cEn5 No.^from prepared cokef 6 . . 515 . . 515 1-460 751 Light gray . . Rather hard. 6 Beaufort, 3, Jjgjt blast .... 5 7069 ..A 6,89*2,000 505 517 1-599 807 Dullish gray . Hard. 7 ^Butterlev • 4 7-038 15,379,500 489 515 502 J&815 889 Dark gray^^H Soft. 8 Bute, No. 4 7-066 J|&63,000 495 487 491 1*764 872 Bluish gray 9 Wind Mill End, NdEH^^jjfblast. . 4 Tjfh 16,490,000 483 495 489 1-591 765 Dark gray . . Hard. 10 5 p|7-049 14,607,000 441 529 48fflB 1-621 718 Gray.... Soft. %I1 Beaufort, No.®g®§Jast . / .- 4 7-108 478 470 474 729 Roull gray . . Hard. 12 Low Moor, No. MBaatlijlast . . . 4 7-055 14,569,500 46$; jp 483 1-852 855 Dark gray . . Soft. 13 puffery, No. ^^^^^ast* 5 15,3,200 463 . . 463 ^1^50 721 gray Rather hard. 14 Brimbo, No. 2, cold blast .... 5 7-017 ti'^ll,6^ 466 453 ^Si9 1-748 815 Light gray . ., H Aped ale, ^jfo. 2, hot blast ... . 3 jjQgl.7 14,852,000 45 5 456 791 n Stiff. 16 Oldber ry,cold blast.... 4 7-059 fcMjflS 1-811 82p^fe Dark gray . . Rather soft. ll Pentwyn, No. 2 4 7-03? 15,193,000 438 473 455 : 1-484 65^‘Cb Bluish gray ^ . Hard. 18 nn 2 5 7-038 , 13,959,500 453 455 454 Jb-957 886 Dark gray . Rather soft. 19 Muirkirk. No. lJBBhi blast* . . . 4 7*113 " fej^os^Jo 443 464 453 1-734 77*, Bright gray FluiS 2» Adelphi, No. 2, cold blast . . . . 5 gjj7-080. 13,815,500 441 ^57 t449 1-759 777 jpg}1 sray • • Soft. 21 Blania, Ne^^yayj blast .... 5 7-159 ! 1^81,466 433 464 448 1-726 747 Bright gray Hard. H Devon, 3®>ld blast .... 4 7-285 1 [V^07,7G#i 448 . . 448 •790 353 Light gray . . 11 23 Gartsherriej blast . . . 5 yji7 13,8^0 427 467 447 1-557 998 V • • Soft. 24 4, common coke . . 6 . . 427 . . |1R7 1-870 79jM| Dull gray . . Rather hard. 2W- Brood, B. 2,*cold a 5 7*031 ’ a3,112,666 460 434 447 , iiS^ l 841 Light gg$y. . Open. 26 Lane End,Ssflp2 ... . . . . 3 7-TOifi 15,787,666 444 . . 444 1-414 y&9 Dark gray . . Soft. *2® Carron, No. 3, cold blast .... 5 16,246,966 444 443 443 1-336 593 Gray.... 11 28 S^^bd^van, 4 Hi 430 443 1-469 1^4 Dull gray . . Rather soft. E9 wlnipteg (marke^^b®) % yJ*- -«jM 5 7-038 , 440 444 442 1-887 830 Bjjpsh gray Fluid. 30 Corby ns Hall, No. 2 . . 5 ‘ 7-007 13,845,866 430 442 1-687 727 . Gray.... Soft. 31 Bio. 2 151 |§80 l 13,186&$'* 7 7 439 44fi*i 440 1-857 816 Dull Rather B2 K^^t^oky^o. 3 5 6-979 lpig5,766 432 449 440 1-443 62^^ pp5I||fit graj{A^"-v’ 'fllttk^hard. KB Milton, No. 3, hot blast .... . 4 -ffPw 42fP 449 438 ■ fs68 585 Gray.... 34 Buff^yl^fo. 1, hot blast* .... 3 6-998 13,730^0 43^1; . . 436 1-640 721 R5 Level, 1, hot blast .... . 5 F^52^0 461 403 43*| 1-516 699 |iLight gray . . " ^ 36 Pant, No. 2 5 6-975 15#$,900 408 455 431 jSsi: 511 Rather hard. 37 Level, No. 2, hot blastBMK£g|®H 6 &1 15,241,000 419 439 429 1-358 Dull gray . . Soft. 38 S., fl 2 ' 5 . 7-041 14,953,333 113 446 429 fcjll-339 554 Plight gray . . 11 39 oumtf y, No^jrot blast . . 4 14,211,000 408W; 446 427 1 618 Bluish gray . w 40 Elsicar, No. 2, cold |l|^KR^iM| 4 6-928 446 408 427 2-224 992 Gray.... 11 41 ■ Virteg, N 7859 1-07 •29 f Broke asunder by tearing at a joint on the bottom 827$ < side 11 inehai from the shackle, where the plate i X lilt deflection 1l*io (, was weakened. Here, by increasing the material at the top of the tube, we attained more than double the strength—thereby showing that fibrous bodies like wrought iron, being more ductile, are more susceptible of injury from compression than from extension. This law is further confirmed by the succeeding experiments. Experiment XV., July 31, *<1845.—Rectangular tube 18 feet 6 inches long, 9*6 inches square, and 17 feet 6 inches between the supports.STRENGTH, ETC., OF WROUGHT-IRON BEAMS. 87 Thickness of plates, top . . — =-0757 inch. 1*14 „ „ bottom —=T425 „ „ „ sides . =*0757 „ Weight of tube=255 lbs. Weight of shackle=988 lbs. Weight in lbs. Deflection in inches. Deflection, load removed. Remarks. 988 2108 TO •05 ( In this experiment great ■weakness is exhibited, as well | as in the former one. *45 3228 •80 •09 3788 f With this weight the top plate began to buckle 2 feet 6 J inches from the shackle on one side, and 6 inches from 1 it on the other. It appears to require stiffness in order to resist the tendency to pucker. .■. Ult. deflection=*94 This experiment was repeated with a strong plate 2 feet 7 inches long, 11 inches wide, and *11 inch thick, laid along the top, in order to stiffen it and throw the strain more upon the bottom plate. The results were, however, unimportant, until the tube was reversed, with the thick side upwards, when a very important change was effected, as shown in the succeeding experiment. Experiment XV. a, July 31, 1845.—Rectangular tube the lamp as before; tube reversed, with the thick side uppermost. 1*14 Thickness of plates, top . . —=*142 inch. „ „ bottom ^j5=-0757 „ „ „ sides . ^=-0757 Weight of tube=255 lbs. Weight of shackle=988 lbs. 88 EXPERIMENTAL INQUIRY ON THE Weight in lbs. Deflection in inches. Deflec* tion, load removed. Remarks. 988 2108 •17 •50 •07 I The droWBon, as well as the modulus of elasticity, are ■i; touch greater in this than in any of the former expe-( riments. 3228 •78 •14 3788 •90 •18 4348 1-05 •20 4908 1-21 •26 5468 1-37 •32 6028 1-54 •40 6588 . 1*75 1 *60 7148 .** TTIt deflection = r76 J If we compare the results of the last two experiments with those, contained in Experiments XIV. and XV.a, we shall find that the proportions of the top and sides are widely different. In both cases, however, when the tube was reversed, with the thick side uppermost, double, or nearly double, the strength is obtained. Hence it follows that, ip order to obtain the section of greatest strength, the top side of a tube, when submitted, to a transverse strain, must be .considerably thicker than its lower side. This fact is fully established in every succeeding experiment, as well as in those already recorded, for the tube almost constantly gave way to compression, unless secured by stronger plates on the top side. Experiment XVI., August 1, 1845.—Rectangular tube 18 feet 6 inches long, 18'25 deep, 9*25 wide, and 17 feet 6 inches between the supports.STRENGTH, ETC., OF WROUGIIT-IRON BEAMS. 89 Thickness of plates, top . . ——=^1^0 inch. „ „ bottom “^=,3€f^0 „ „ „ sides . ^-°=*0594 „ Weight of tube=317 lbs. Weight of shackle=988 lbs. Weight in lbs. Deflection in inches. Deflection, load removed. Remarks. 988 ■15 •30 2108 3228 •44 *05 4348 •60 •07 5468 •70 •10 6588 1-00 •22 6812 / With this weight the top plate began to rise 18 inches •j from the shackle, after sustaining the weight for ( about a minute. . •.Ult. deflections: l-03 Having, in this experiment, crippled the upper side of the tube, it was turned upside down after the injured part was straightened, and the experiment repeated. In most of the experiments the tendency to rupture was slow and progressive,—a property which seems to be inherent in sheet-iron tubes, particularly when they yield to compression. Under this species of strain destruction is never instantaneous, as in cast iron, but advances gradually, the material emitting during the process a crackling noise for some time before the experiment is complete and absolute rupture takes place.STRENGTH, ETC., OF WROUGIIT-IRON BEAMS. 89 Thickness of plates, top . . —--T490 inch. „ „ bottom ^-^=•2690 „ „ „ sides . ^-°=-0594 „ Weight of tube=317 lbs. Weight of shackle=988 lbs. Weight in lbs. Deflection in inches. Deflection, load removed. Remarks. 988 •15 2108 •30 3228 •44 *05 4348 •60 •07 J 5468 •70 •10 6588 1-00 •22 t With this weight the top plate began to rise 18 inches 6812 -! from the shackle, after sustaining the weight for ( about a minute. .-. Ult. 1 deflections 1’03 Having, in this experiment, crippled the upper side of the tube, it was turned upside down after the injured part was straightened, and the experiment repeated. In most of the experiments the tendency to rupture was slow and progressive,—a property which seems to be inherent in sheet-iron tubes, particularly when they yield to compression. Under this species of strain destruction is never instantaneous, as in cast iron, but advances gradually, the material emitting during the process a crackling noise for some time before the experiment is complete and absolute rupture takes place.90 EXPERIMENTAL INQUIRY ON THE Experiment XVI.a, August 1, 1845.—The preceding tube reversed, with the thick side uppermost Thickness of plates, top . * *^^—*2690 inch. „ „ bottom ^jp=T490 „ „ „ sides . ^ft-0594 „ Weight of tube=317 lbs. Weight of shackle=988 lbs. Weight in lbs. Deflection in inches. Deflection, load removed. 1 Remarks. 988 •08 2,108 •30 3,228 •42 •03 4,348 •55 •10 5,468 •70 •15 6,588 •80 •20 7,708 •92 •24 8,828 1-10 •30 9,948 1’30 •31 * August 3. The freight, 11,068 lbs., 'Was left on the ■fO,508 1‘35 *32 i tube from half-past 3 O’clock p.m. till the following day at half-past 9 a.m., when; the deflection in- 11,068* 1-40 •40 creased from 1'45 to t’Mt ll,068f 1-60 •50 •j- Experiment continued after sustaining the weight 11,628 1*66 >*53 13 hours. 12,188 With this weight the top side puckered. . •. Ult. deflection=T73 The tube failed with 12,188 lbs. at two of the joints on the top side, 3 feet from the shackles. This failure was accompanied by the sides bending inwards on one side, with a similar indentation on the other, and the top plate doubling at the joints, in the form of the letter S.STRENGTH, ETC., OF WROUGHT-IRON BEAMS. 91 Experiment XVI.&, September 20, 1845.—The top side still yielding to compression, a stronger plate was riveted on the top side; and in order to cause the bottom to give way to a tensile strain, a thicker plate was riveted over the joint on the bottom side, and the experiment repeated. Distance between the supports as before, 17 feet 6 inches ; weight of shackle » 960 lbs.. Weight in lbs. Deflection in inches. Deflection. load removed.. Remarks. 960 2,697 •09 4,426! •16 6,173 •25 7,859 •34 9,555 •42 11,262 •60 •07 12,107 *65 •14 12,9901 •72 *15 f Broke after sustaining the weight some minutes, by 13,867 •< tearing the rivets from the joints on the upper side ( 3 feet 8 inches from the shackle. Ult. deflection = '76 The great powers exhibited in the last experiment by the addition of a certain quantity of material to the upper side of the tube suggested a further extension of the experiments, with some slight modifications of form in order to render more conclusive the principle which the previous trial had indicated. For this purpose a hollow girder 25 feet long and 15 inches deep, of the following dimensions, was constructed and submitted to experiment. Experiment XVII., August 2, 1845.—Rectangular92 EXPERIMENTAL INQUIRY ON THE tube or girder 25 feet 11 inches long, 15 inches deep, 21 inches wide, and 24 feet between the supports. Thickness of plates, top . . - = -260 inch. O “ B bottom —m *260 “ 5 “ “ sides. = T31 “ Weight of tube = 788 lbs. Weight of shackle = 800 lbs. Weight in lbs. Deflection in inches. Deflection, load removed. 800 -07 ■ 1,920 •20 3,040 •33 4,160 •50 6,280 •60 6,400 •70 7,520 •83 8,640 *05 ^•07 9,760 1-20 •15 10,880 1-35 ' *20 12.008 1-50 *25 13,120 *** I Ult. deflection = 1-613 Remarks. The elasticity remains nearly perfect up to 8640 lbs. ! Broke by tearing through the solid plate on the bottom side, 7 inches from the shackle, as the weight was laid on. I A flaw having been discovered in the plate where the fracture took place from imperfect welding, a stronger plate, 14 in. long and one-fourth of an inch thick, was riveted over the crack, and the experiment repeated. Experiment XVII.a, August 4, 1845.—Rectangular tube the same as before.STRENGTH, ETC., OF WROUGHT-IRON BEAMS. 93 W eight in lbs. Deflection in inches. Deflection, load removed. Remarks. 5,280 •65 •08 6,460 •77 •15 7,520 •90 •18 8,640 1-05 •23 9,760 1-20 •30 10,880 1-31 •21 'I The defects of elasticity were not, from some unknown 12,000 1-46 •21 causes, clearly indicated, until the weight 14,240 L lbs. was removed, when the defects were found 13,120 1-60 •21 =•60. Probably this might have arisen from some 14,240 1-75 •60 J unequal tension. 14,800 2-11 •62 15,360 2-17 •62 15,920 2-28 •68 16,480 2-36 •74 17,040 2-38 •80 j With this weight the top plate gave way by com- 17,600 ( pression. .’. Ult. deflection = 2-66 As this description of beam indicated very considerable powers of resistance, it was deemed advisable still further to test its powers by allowing the weight 14,240 lbs. to remain suspended during the night. This was done during a period of seventeen hours, after which the load was removed. The deflection during that time had increased from T75 to 2,00=*25, and the loss in elasticity was '60 — *30 = *3. The beam during the last two experiments had suffered considerably from the severity of the strains to which it had been subjected; and it was considered that the anomalous condition of puckering, which had all along been present, might be avoided by reversing the girder with the broad flancli uppermost. This was accordingly done, the injured part having first been straightened, and a strong plate 1994 EXPERIMENTAL INQUIRY ON THE inches long having been riveted upon it, the experiment was again proceeded with, as follows: Experiment XVII.6, August 5, 1815.—Rectangular tube, same as before; the beam reversed, with the narrow flancli downwards. W eight in lbs. Deflection in inches. Deflection, load removed. Remarks. 9,7G0 1-40 •38 ( The deflection and permanent sets must be added to and 10,880 1*05 •50 ( subtracted from the res pective numbers 1 TO and '38. 12,000 1-83 •59 pa 13,120 2-03 •69 14,240 2-30 •84 15,360 2-49 •91 15,920 ... ... f Broke when the weight was laid on by extension, the v lower plate tearing asunder 6 inches from the centre [ of the shackle. /. Ult. deflection = 2,58 Owing to the broad flanch being placed uppermost, it was expected that the tube would yield to extension, which was the case ; but the plate gave way at the rivets of a joint at some distance from the centre. This joint, moreover, had been a good deal strained by the former experiments, which may account for its fracture by a comparatively less weight. Experiment XXV., September 20, 1815. — Having tested the powers of the larger description of girder in a variety of ways, the smaller one was treated in the same manner, as follows: Rectangular girder 12 feet long, 8 inches deep, 1 inch wide, and 11 feet between the supports. Thickness of plates, top . . ^-^-=•282 inch. „ „ bottom -y^-=’116 „STRENGTH, ETC., OF WROUGHT-IRON BEAMS. 95 Thickness of plates, sides . ~‘067 inch. Weight of tube = 125 lbs. Weight of shackles = 930 lbs. Deflec- Deflec- Remarks. in lbs. tion in tion, load inches. removed. 930 •06 1,780 •11 'When 9974 lbs. were laid on, the ends, from the ex- 2,030 treme thinness of the plates, were giving way at the •16 ... -< supports; two pieces of hard wood were, however, inserted between the sides, which kept them straight 3,516 and prevented them from twisting. •21 4,382 •26 5,214 •32 6,105 •37 •010 0,543 •41 •012 6,996 •44 •035 7,433 •47 •040 === 7,801 •51 •050 8,273 •54 •058 JL 8,693 *58 •075 9,107 •62 •097 9,545 •07 •118 9,974 •74 T50 10,380 •87 •243 10,827 1-06 •325 'The top plate doubled up with this weight 4 inches 11,254 ... ... from the shackle. Finding the top of the beam the weakest, it was turned upside down, and a weight Beam reversed. of 6113 lbs. laid on to straighten the injured part. 0,113 •51 •10 0,549 •60 •13 0,978 •74 ■23 7,140 ... With this weight the top plate was forced upwards. .•. Ult. deflection = -75 96 EXPERIMENTAL INQUIRY ON THE This girder, although extremely light on the sides* with a tolerably thick top* nevertheless gave way by compression. Its bearing powers Were very considerable with the thick side uppermost; and provided that part had contained a little more material, it would have carried upwards of 12,500 lbs. During the progress of the experiment I had frequent conferences with Mr. Stephenson; and having reported to him from time to time the results that were obtained, and the impression they made upon my mind, he suggested that it might be desirable to have a tube made of an entirely different form, in order, if possible, to throw the top side as well as the bottom of the tube into a state of tension. This suggestion was intended to obviate the anomalous condition of puckering, and to prevent as much as possible that tendency to “buckle,” which, in every instance, is more or less present in rolled sheet-iron plates.* It had a further object in view, namely, to relieve the strain on the centre of the tube, whether arising from the effects of its own weight or the load, by extending the length beyond the supports to a distance of half the span on each side. This additional weight, extending over the piers, was supposed to act as a counterpoise; aa (see plate XIII. fig. 18) being the fulcrum to that part of the tube in the middle at n, and it would also assist in the support of the load during the passage of the trains through the tubes. For these objects a tube was made of the form shown in plate XIII., fig. 18. * It is almost next to impossible to roll plates with all the parts in the same degree of tension. Almost every plate has more or less buckle, and it requires no inconsiderable degree of skill to 8HHter those parts of a plate wheAc tension is greatest, and also-to find oOt the parts which cause the “buckle.” m considerable poBH^H variable tension in. the earn position of plates is pffijably caused by unequal contraction in the process of cooling, and also by a difference of temperature in the blooms from which the plates are rolled.STRENGTH, ETC., OF WROUGIIT-IRON BEAMS. 97 Experiment XVIII.a, August 3, 1845.—Rectangular tube 37 feet 8 inches long, 13*25 inches deep in the middle, 7£ inches wide, with, the upper part raised to 17*25 inches, as at aa, and 18 feet between the supports. The width of the top and bottom plates as per section below. Thickness of plates, top . . • *&T 425 inch. “ “ bottom 11^ *= *1425 “ 8 “ I sides . *=*1127 « Weight of tube *^640 lbs. o Weight of shackle = 800 lbs. Weight in lbs. Deflection jn inches. Defloe- J tion, load removed. 800 •09 1,920 •20 •02 ■8,040 ■82: '05 J 4,160 •45 •09 i 5,280 ■59 •16 j 6,400 ; •11 •19 Y.520 ' ■84 ■22 1 8,640 •99 *27 9,YG0 118 •32 j 10,880 ** ... Remarks. ( With tliis weight the top plate doubled up 1 foot C ) inches from the shackle. j Ult. deflection 1-31 After the top side had yielded to compression, the weights were removed, and the supports under the bottom having been lowered, the tube was supported on two cross bars passing through the tube, and the weights were again sus-pended upon it. These weights, when laid on, made no difference in changing the direction of the forces, as the top plate was ag - i i n98 EXPERIMENTAL INQUIRY ON THE forced upwards by compression, and that to a greater height than before, accompanied with increased distortion of the sides, which shortly became collapsed diagonally on each side of the shackle. In this experiment it is probable that some degree of tension was induced along the upper line of the top plate, as the extreme end of the tube was raised with some force as the weights were increased. This property of the weight raising the end over the fulcrum a a was strikingly apparent when the whole weight = 10,800 lbs. was laid on; but it did not appear to alter the conditions of the middle part, which was forced upwards by compression, and followed the same law as if it had been formed of a single beam. These appearances indicated a tendency of the two ends, then extended to half the distance between the supports, to act as a counterpoise, and not only to change the direction of the strain on the top side, but relieve the bottom, which in other respects must have borne the whole weight. From this it is inferred, that the tube in its full size would be greatly relieved by increasing its length on each side of the land-towers, as in the case of the Britannia Bridge, to the extent of half the span. As a further illustration of these views, the injured part of the tube was repaired by riveting an additional plate of the same thickness along the top side over that previously damaged. With this addition the experiment was again repeated. Experiment XVIII. repeated, August 10, 1845.—Rectangular tube same as before. Thickness of plates, top . . . = ’2850 inch. “ “ bottom . »*1425 “ “ “ sides . • = *1127 u Distance between the supports, 18 feet.STRENGTH, ETC., OF WROUGHT-IRON BEAMS. 99 1 vSi i i • m tion ip in lbs. . , I inches. . Ueflee- ; tion, load removed. • 1 Remarks. 7,520 j -68 r ■ 6,640 ”64 ■04 i i 9,760 ! "99 •10 10,880] 115 •16 ( With this Weight, 12,000 li the sides -were slightly! 12,00p 1-31 <*25 < puckered, indicating a tendency to force the topi 12,500 1*40 •32 ( side upwards. 13,120 1-6* ; •42 13*680 ... j f Paclrered as before by compression, the top platel \ doubling up 18 inches from the shackle. J /. Ult. deflection szr l-7l ■ 1 1 On consulting the two lajst experiments, it is obvious that no great increase of strength was obtained by doubling the thickness of the top plate. This may, however, be accounted for by the circumstance of the top plates being under instead of above the line of compression. In every description of girder composed of malleable iron, and probably of any other material, the upper side-should he elevated to a greater height than the line of ultimate deflection. It should always be above, but never below, the horizontal line of compression. Another cause of the failure of this tube, with a comparatively leg# weight than the increased thickness of the top would indicate, might be traced to the severe injuries which that part sustained in the previous tests. Hence followed the puckering of the top side, at a much earlier stage of the experiment than it should have done had the plates been sound and the line of the forces changed. The experiments on this form of tube are perhaps the more interesting* from the fact, that they exhibit certain defects of form which100 EXPERIMENTAL INQUIRY ON THE it may be desirable to avoid in girders of this description. If the parts suspended beyond the piers are intended to act as a counterpoise to the load, it will then become necessary to have the girder of uniform strength and texture, with a slight curvature of the top side about one-tenth the depth. With these precautions in the construction, the strength would be greatly improved, and, being subjected to severe strain, will follow the same law, as regards extension and compression, as those of a girder of the simple form. During the progress of Experiment XXII., when the elliptical tube, after being strengthened by an iron cellular fin, riveted along the top, was found defective in resisting the crushing force, it then occurred that a different construction might be introduced, so as to give strength and rigidity to that part. For this purpose I sketched out, and gave orders for the construction of, a tube with a corrugated top, forming two longitudinal cavities along the whole of its top side, as exhibited in the annexed sketch, Experiment XXIX. This tube was executed with considerable care; and having been submitted to the usual experimental test, the results were as follows: Experiment XXIX., October. 14, 1845^Rectangular tube, with a corrugated top, 19 feet 8 inches long, 15*4 inches deep, 7’75 inches wide, and 19 feet between the supports. 23 Thickness of plates, top . — = *115 inches each. “ bottom = T80 “ “ “ sides . = *070 “ Weight of tube = 500 lbs. Weight of shackle = 988 lbs. The tubes a a were 1*65 inch in diameter.STRENGTH, ETC., OF WROUGIIT-IRON BEAMS. 101 Weight in lbs. Deflection in inches. Deflection, load removed. 9S8 ’035 2,736 •110 4,468 T90 6,215 •270 7,924 •340 •020 9,636 •424 •050 11,334 •523 •062 13,041 •640 •095 14,7.51 •735 •125 16,490 •870 T86 IS,205 1-070 •276 19,065 1T55 19,918 1-270 •400 20,764 1-425 21,629 1-520 •590 22,469 ... ... . •. Uit deflection = 1’59 Remarks. Sr0®03 ( With the weight 18,205 lbs. tlie deflection increased ( -02 inch in three minutes. ( Broke by the side plate tearing from the top at two feet l from the shackle. A short time previous to the tearing of the sides from the top at the rivets, that part had begun to assume a slightly undulating appearance on one side, arising from the weakness of the side plate, which gave way near to the shackle. This was not, however, the only part that suffered under the strain, as the opposite side was tearing from the bottom plate, at the same time evidently showing a rapid approach to rupture on both the lower and upper sides of the tube. These parts exhibited important features in the due and perfect adjustment of the top and bottom, which in this case were calculated to resist, as nearly as possible, the forces acting upon them. Another property of considerable importance in this102 EXPERIMENTAL INQUIRY ON THE description of girder is its progressive tendency downwards to destruction. It is widely different from cast-iron and other crystalline substances in this respect, since, from its fibrous nature and greater ductility, it gives timely warning before rupture takes place. This property was noticed in several of the former experiments ; and in this experiment it became more apparent after the whole weight, 22,469 lbs., was laid on. With this weight more than three minutes elapsed before the experiment was completed, and the tube rendered unfit for use. Experiment XXX., October 10, 1845, on a malleable iron beam of the annexed sectional form, 11 feet 7 inches long, and 11 feet between the supports. Dimensions at a—1*000 inchesx2^ inches. „ ,, 6= *325 „ x7 ,, „ ,, c— *380 „ x4 ,, Weight of beam=:227 lbs. Weight of shackle ^885 lbs. Weight in lbs. Deflec-1 : tion in J . inches. J tion, load removed. j Remarks. S85 *04 2,581 •12 4 4,817 •20 1 H—.j, I 6,050 *26 If j 7,748 •35 1 %%%$.<-r- ] 9,498 ! ■ *46 11,258 •60 •09 fWith this weight the beam' became disto periments on a larger scale, in order to determine the form and proportions of the tubes. For these objects, an entifely new model tube, exactly one-sixth of the dimensions of the Britannia Bridge, was constructed; and having arranged the apparatus, the experiment was proceeded with as before. I have given the foregoing extracts from the experiments instituted to determine the form and strength of the Britannia and Conway Tubular Bridges. As these experiments were the first that had been made upon malleable iron beams, and as they arc of the highest importance, considered in relation to the extended application of wrought iron in the construction of buildings, their insertion in this work will not be without its use, if it aid in directing the attention of the younger branches of the profession to an effective as well as an economical distribution of a highly valuable material. Since the above experiments were made, others of equal interest, and bearing more directly upon the subject of beams for supporting floors, have been entered upon. These experiments were made on beams of the annexed form : Dimensions at «=3i x J-inch angle-iron. „ „ 6=-37xiG inches. ’ , "y*-* „ „ c=3£ x^-inch angle-iron. Weight of beam— 1380 lbs. Distance between the supports 24 feet, depth 1 a to beam, as indicated in formula, page 61. f This great national work was undertaken at tho suggestion and sole expense of my friend, Mr. William Dargan, the friend of Irelaud and the promoter of Irish industry.108 EXPERIMENTAL INQUIRY ON THE 0*0 X eo X 0*0 X • ■ * CO CO X X CO CO 0 1 o *- Tc a r- o O o P3 Weight of girder, 819 lbs.STRENGTH, ETC., OF WROUGIIT-IRON BEAMS. 109 Experiment to determine the Strength and Security of an open Trellis Girder, as shown in the foregoing Diagram (fig. 38). Oct. 12, 1852. W eight laid on in tons. Deflection in inches. Remarks. 1 2 •107 ( With three tons the diagonals a a were slightly buckled, 3 •187 and continued to lose their powers of resistance as the additional weights were laid on. 4 •250 *\ 5 •258 6 •266 7 •367 8 •500 9 •633 10 •638 The weights from three tons were all placed on the 11 "V 50 middle part of the girder, or within the two vertical 12 •875 „ struts cc. With 16 tons the diagonal bars a a had 13 •906 buckled 2f inches; with 20 tons, 3 inches; and with! 25 tons, 3£ inches. 14 1-000 15 1-034 16 1-070 17 1-088 18 1-130 1 19 1-162 20 1-185 J 21 1-188 09 1-195 23 1-203 24 1-218 > These weights were placed nearer the ends of the girder. j 25 1-224 26 1-240 At this point the diagonal struts or plates a a were bent 27 1-280 considerably, so as to render them totally inadequate Oft as supports to the upper flancli of the girder. 1 O i'Z 1 26 tons was left on the girder for three days without any 29 1-440 apparent increase in the deflection. 30 1-513 After laying on 32 tons, it was considered advisable to die- i •nar\ continue the experiment, as the girder was consider- ably crippled in the end diagonal stays a a, which were 32 1-624 puckered to the extent of four inches or upwards. 110 EXPERIMENTAL INQUIRY ON THE As some doubts were entertained as to the security of this form of girder, it was deemed expedient, on the part of Mr. Dargan and the Committee, to reassure the public of their safety; and for this object I was requested to visit Dublin and report thereon. To accomplish this object, 1 had two of the girders supported at the extremities at a distance of about four feet asunder ; and having covered the centre part of the top with a wooden platform, the girders were progressively loaded with iron in the foregoing manner, until a deflection of 1*62 inches was attained, when the experiment was discontinued. From the commencement the diagonal stays a a exhibited very defective powers of resistance, and, in fact, were of little value in supporting the upper flanch exposed to compression. To render these parts effective, they should have been constructed of T or angle iron, in order to giye the required rigidity in their resistance to a crushing force, which was pressing upon them in the direction of the abutments as the deflection of the girders increased. In every construction of this kind, it is desirable that the direction of the forces should be duly considered, in order to bring the bearing powers of all the parts simultaneously into action. After a weight of 32 tons was laid on the beam, the ultimate deflection was found to be 1 ’6 inches; and having removed the weights, the deflection was again taken, when there remained a permanent set of ’G5 parts of an inch. In the construction of girders of this description, there appear to be defects as well in respect to the form as to the distribution of the material. Throughout the experiments the diagonal beams b h were in a high state of tension, forming, with the bottom flanch, the chief element of strength, as a truss supporting the more rigid part ofSTRENGTH, ETC., OF WROUGIIT-IRON BEAMS. Ill the structure, or top flancli A A ; and this was accomplished without their receiving adequate support from the diagonals a a. On the contrary, the diagonals a a are not only weak and defective from their thin-plate form, but112 EXPERIMENTAL INQUIRY ON THE the cross diagonals e e are also inoperative, and become perfectly slack, from the effects of tension upon the diagonal bars b b and the bottom flanch, which, in fact, support the load. If the diagonals were made of T or angle iron, so as to act as struts, calculated to resist compression as well as tension, and the two centre diagonals entirely removed, leaving only a verticle T-iron stay in the middle, as at S, fig. 39, we should then have a girder of considerably increased strength, and that without any increase of material. This alteration, united to perfect workmanship in the connexion and union of the parts, would not only increase the strength, but greatly enhance the value of structures which are in such great demand for buildings of large dimensions, such as crystal palaces for industrial exhibitions, termini for railways, and others. The above diagram, fig. 39, shows the position of the load, 32 tons, laid upon the girders, and also the suggestion for dispensing with the middle diagonals e e, as shown in the original girder at fig. 38. Let us now proceed to determine a formula for calculating the strength of these trellis beams. We may regulate a trellis beam as an imperfect double-flanched beam, where the material of the section is collected at the top and bottom parts. We say imperfect double-flanched beam, because the connexion between the top and bottom parts is not so completely maintained as it would be by an unbroken plate or rib. However, with some allowance for this imperfection, we may calculate the strength of these beams on the same principle as the ordinary double-flanched beams. Thus we have -.T7- Gad , ^ Wl W — —j*"* and C —I —~~t~ < l 9 ad 4 where a is put for the area of the bottom part.STRENGTH, ETC., OF WROUGIIT-IRON BEAMS. 113 In order to determine the value of the constant for trellis beams, we have, from the Table of Experiments, W — 32}, l — 23£ x 12, a ~ 3 J, and d — 3 x 12; hence we have P_32!x23^ x!2_7^ 31X3X12 w 12 ad ••(0- lienee the value of the constant is pretty nearly the same as it is for tubular beams.PART III. ON THE CONSTRUCTION OF FIRE-PROOF WAREHOUSES. The following report was drawn up at the request of Samuel Holme, Esq., of Liverpool, in order to confirm the opinion of that gentleman as to the expediency of erecting all the new warehouses in Liverpool with fireproof materials. Believing the views contained in this report to be useful aud instructive, it is here reprinted for the information of the general reader. “ The serious nature of the late fires at Liverpool, Manchester, and other large towns, has induced an inquiry into the causes of these disasters, with a view to avert their progress, and to adopt measures for the better security of property, and the prevention of a calamity so injurious to the public as well as individual interest. In no other description of building have the effects of fire been so severely felt, nor have the provisions necessary for its s oppression been so disregarded, as in warehouses used for the stowage of commercial produce in maritime towns. In the manufacturing districts the same apathy has not prevailed, for in most places fire-proof buildings have been introduced; and considering their complete success, it is surprising that the same system has not been adopted in the construction of w arehouses and other buildings appropriated to the reception of merchandise. When we consider the extent and immense value of property contained in these edifices, it can scarcely be conceived that such a state of things should exist; and, moreCONSTRUCTION OF FIRE-PROOF WAREHOUSES. 115 particularly, amongst a body of men the most active and intelligent in Europe. Such, however, is the case ; and we have only to enumerate a few examples to show that a disregard of consequences, or a culpable ignorance of existing improvements, has pervaded the mercantile community for a number of years. This should not be, as the best description of buildings in which the manufactures of cotton, flax, silk, and wool arc carried on, arc, with few exceptions, almost entirely tire-proof; and upwards of thirty years have elapsed since iron beams, iron columns, and brick arches were first introduced in the construction of factories, as a security against fire. These facts ought not to have escaped the observation of the British merchant; and yet, in the face of so many examples, with one single exception,* it is only within the last few months that a non-combustihle material has been used in the construction of the immense magazines of Liverpool. In other parts of the empire the same laxity of application exists ; hut the effort so happily made at the port of Liverpool will, it is hoped, extend itself to the metropolis and every other sea-port in the kingdom. For these objects, and for the guidance of those who may feel disposed to adopt measures for saving a large rate of insurance, and for the further protection of their property, I would respectfully submit the following observations for consideration :— In the ages of antiquity we have only a few examples of fire-proof structures; and provided we except the monuments of the early Egyptians, and some of the public edifices of the Greeks and Romans, there are but few instances of buildings so erected as to afford any security against the ravages of fire. During the middle ages some of the Gothic * Messrs. Jevons constructed a fireproof warehouse ou the New Quay, Manchester, ten years ago.116 ON THE CONSTRUCTION OF churches and cathedrals were constructed almost entirely of stone ;* and, with these exceptions, there appears no evidence of an existing knowledge as to the benefits arising from the use of an entirely fire-proof structure. Probably, the want of cast iron, and the consequent ignorance of its use, was an insurmountable barrier to the development of the fire-proof system ; but in the present age these difficulties do not exist; and to neglect the means thus so libdrally supplied for the protection of life and property, would augur a want of discernment incompatible with the spirit and enterprise of the age. Latterly, the extension of commerce, and the great value of property which is daily consigned to the keeping of individuals and companies, have produced a different feeling; and, viewing the present engagements of merchants, with the amount of transfer from one hand to another, it is no longer matter. of surprise that measures calculated for the better security of property should be imperatively called for, and that in every instance where it is exposed to risk. The general character of warehouses has for ages been the same, the roofs and floors invariably being constructed of timber, with strong girders and wooden props; and these have, in most cases, been so injudiciously placed, as to cause considerable injury to the structure on every occasion when great weights have'to be supported. On referring to the greater number of these erections, it will be found that the props which support the floors have their ends placed immediately under the main beams; and these being successively supported upon each other, with the main beams intervening, the result is, that the fibres are thus completely crushed, particularly in the lower floors, by the superincumbent weight, * The cathedral of Milan is eoaairucted entirely of marble and glass.FIRE-PROOP WAREHOUSES. 117 and in many cases the beams are almost splintered, from the immense pressure to which they are subjected. Even in this imperfect construction the necessary precaution of wooden caps has not in all cases been adopted; and until the introduction of iron columns, with heads and bases covering a large portion of the beam, the timbers were in many instances seriously injured. The use of iron columns, although an improvement upon the old system of building, is nevertheless no security against lire; and it is obvious that no guarantee can he given so long as the structure is chiefly composed of limber, and the openings imperfectly closed by wooden doors and shutters. From this it is evident, that in order to give perfect security, warehouses should be constructed upon different principles, and these; may he enumerated as follow#, viz.: 1. The whole of the building to be composed of non-combustible material, such as iron, stone, or brick. 2. In order to prevent fire, whether arising from accident or spontaneous combustion, every opening or crevice communicating with the external atmosphere to be closed. 3. An isolated stone or iron staircase (well protected on every side by brick or stone walls) to be attached to every story ; and the staircase to he furnished with a line of water-pipes, communicating with the mains in the streets, and ascending to the top of the building. 1. In a range of stores, the different warehouses to be divided by strong partition-walls, in no case less than 18 inches thick, and no more openings to be made than are absolutely necessary for the admission of goods and light. 5. That the iron columns, beams, and brick arches be of strength sufficient not only to support a continuous dead pressure, but to resist the force of impact to which they118 ON THE CONSTRUCTION OF are subject by the falling of heavy goods upon the floors. Lastly, That in order to prevent accident from intense heat melting the columns, in the event of fire in any of the rooms, a current of cold air be introduced into the hollow of the columns from the arched tunnel under the floors. Adopting the foregoing divisions of the subject, it will he requisite to consider them separately. First, the whole of the building to be composed of non-combustible material, such as iron, stone, or brick. In the choice of material, much will depend upon locality, and the cheapness at which it can be obtained. In this country the best fire-proof buildings are generally composed of brick or stone, with iron beams and columns, properly framed and held together by rods built into the walls, and brick arches for the floors : which arches are supported by, and spring from, the,lower flanches of each beam, and are thus extended in succession on each floor from one end of the building to the other. These arches may be formed either in a longitudinal direction in the line of the building, or transversely, as circumstances may admit. The floors are generally laid with stone-flags or tiles upon the arches, after they are properly levelled, and filled up at the haunches with a concrete of lime, sand, and ashes. The flags or tiles, being well and solidly beddetj in mortar, form a durable and excellent floor. In buildings for particular objects, it id sometimes necessary to have wooden floors ; and, where found necessary, the boards are generally nailed, in the usual way, to sleepers imbedded in the lime-concrete, as before described, or, what is probably better, with a pavement of wooden blocks. This description of building, when properly constructed and surmounted by an iron roof, is perfectly imperviousFIRE-PROOF WAREHOUSES. 119 to the action of fire; and provided due regard be paid to the selection of a careful superintendent, both owners and occupants may rest satisfied as to the safety of the property. Secondly, In order to prevent fire, whether arising from accident or spontaneous combustion, every opening or crevice communicating with the external atmosphere to be closed. These are points which should never be neglected in fire-proof buildings. In warehouses, in particular, it is of vital importance; because in rooms or floors where combustible material is stored, nothing tends so much to the security of the building and its contents as a power to shut out and prevent the admission of air. For this purpose, an iron or stone staircase, surrounded by brick or stone walls, and communicating with the different floors by iron doors, should always be attached. This staircase should be easy of approach from without, with a covered opening at the top, and windows at each landing, in order to effect free ventilation, and a ready communication witly| every part of the building. Warehouses constructed upon this principle will effect almost perfect security; and, in the event of fire, will enable persons not only to approach the locality, but, in case of the casual admission of atmospheric air, the room might be .shut up, and the flames smothered, till an effectual remedy was at hand. For these objects, I would strongly recommend the iron doors, frames, and shutters, as constructed and used by Messrs. Samuel and James Holme, of Liverpool, to be fixed in every room. These doors arc made of double sheet-iron plates riveted to a skeleton frame, with a stratum of air between, which, acting as a non-conductor, is admirably adapted to the purpose for which they are intended.120 ON THE CONSTRUCTION OF Thirdly, An isolated stone or iron staircase, well protected on every side by brick or stone walls, to be attached to every story; and the staircase to be furnished with a line of water-pipes, communicating with the main in the street, and ascending to the top of the building. Under the second division we have already treated of the staircase, and the necessity which exists for having it perfectly distinct from other parts of the building: exclusive of this separation, it will be found still more secure by having a copious supply of water always at command. That supply should not only exist in the street-main, but should communicate with every landing by a brass cock and hose, till it terminates in a cistern, with a valve, on the top of the roof. This cistern should be of such capacity as would insure a sufficient supply of water in case of accident to the pipes in the street. The pipes, leather hose, and the requisite discharge of cocks, screw-keys, &c , should be kept in good repair, and the hose and screwkeys hung up at every landing, ready for use. These precautions will give additional security to parties bonding goods, as also to the owner of the property in which they are deposited. In addition to the above, it will be advisable that ail the cocks, hose, and screw-keys be made of one size, and the same as those used by the fire-brigade of the town. Before closing this part of the subject, I would observe, that an exceedingly simple and’ ingenious apparatus for extinguishing fire has been adopted by Joseph Jones, Esq., of Wallshaw, near Oldham. It cotSsists of a thin copper globe of nine* inches diameter, perforated full of small holes, and suspended from the ceiling of the different rooms, either in a mill or a warehouse. Each rose is (in case of need) supplied with water by lines of pipes commu-FIRE-PftOOF WAREHOUSES. 121 nicating with the mains in the street. In this form, Mr. Jones is not only in a position to discharge a flood of water into each separate room, but, from the peculiar shape of the rose, he is enabled (with a pressure of 200 feet acting upon the apertures) to disperse it to a distance of upwards of 40 feet in every direction. This is a certain and effectual method for extinguishing fire, and might easily be adopted in almost any important structure in large towns, where a supply of water and the necessary pressure can be obtained. Another important feature of this application is the facility and rapidity with which fires can be extinguished. The cocks are all on the outside of the building; and being carefully locked up and marked with numbers corresponding with the ^different, rooms, there is less risk of delay and confusion when an accident occurs. Fourthly, In a range of stores, the different warehouses to be divided by strong partition-walls, and no more openings to be made than are absolutely necessary for the admission of goods and light. Those" precautions become more apparent in every case where large piles of buildings areJ erected contiguous to each other, and where risk from fire is incurred in the communication of one part of the building with another. The Metropolitan Building Act has provided against accidents of this kind by the insertion of a clause wherein these precautions arc insisted upon; and by the introduction of partition-walls which divide the houses, the utmost security is afforded to that description of property. In contiguous buildings these partitions have their full value,; and it not unfrequently occurs that the property on each side has been saved from conflagration when a centre building haa been completely destroyed: hence the necessity for complete separation in every case where the buildings are contiguous.122 ON THE CONSTRUCTION OF In the construction of warehouses these precautions are the more important, from the increased value of the property therein deposited, and the greater risk to which, in some particular cases, they are subject. All warehouses should therefore be carefully separated from each other: and in forming the partition-walls, it might be a great improvement to have an open space of two inches up the middle, with proper binders* for the purpose of ventilation; as air, being a non-conductor, would, in case of fire, prevent the walls from being overheated, and afford a free communication with the atmosphere by the ascending current of air. They should also be built to some height above the roof, in order to prevent the possibility of communication with the adjoining stories, and to effect a complete separation of the different compartments into which they are divided.* To render the different flats or rooms of warehouses secure, it is a desideratum to have as few openings in them as possible. This is the plan adopted in the warehouses of Mr. Brancker, Dublin Street, Liverpool ; and they appear to be not only well calculated for the admission and transmission of goods on each side, but having no more windows than are absolutely necessary for the admission of sufficient light to effect the deposition and removal of merchandise, they are exceedingly well adapted for the double purpose of convenience and security. In every situation the iron doors and iron window-shutters already described should be used. It will be observed, that the security afforded by the iron doors and shutters will be of no rise unless they be closed and fastened every night before the warehouse is shut up. * The Liverpool Building Act has now rendered it compulsory that parapet walls shall be built up live feet above the gutters.FIRE-PROOF WAREHOUSES. 123 Fifthly, That the iron columns, beams, and brick arches be of strength sufficient not only to support a continuous dead pressure, but to resist the force of impact to which they may be subject by the falling of heavy goods upon the floors* This is one of the most important considerations connected with the security and construction of warehouses; and in order to remove .every doubt as to the stability of such a structure, I must refer to my highly talented and respected friend, Mr. llodgkinson, one of the first authorities in this or any other country on the strength of materials. To that gentleman the public are indebted for a series o theoretical and practical experiments on the Strength of beams and pillars, of the utmost value to architects, builders, and engineers. Any person choosing to make himself acquainted with the principles of Mr. Ilodgkineon’s experiments, and the results deduced therefrom, will find no difficulty in constructing beams and columns the strongest form, and at the same time insuring the proportional and requisite strength, accompanied with a great saving in material in all parts of the structure. On this part of the subject it will be ncce^ffifry to obi^ve, first, on the structure of beams, that until the publication of Mr. Ilodgkinson’s experiments, practical men were almost entirely without rule or any satisfactory theory on which to found their calculations on the form and distribution of the material. Now the subject is well understood, not only as regards the strength which is wanted, but also idle best and strongest form lor resisting the different .strains to which they are subjected. In warehouses containing goods th^^ strains are more varied than in factories# In the former, the] floors arc often loaded, to a great extent, with solid, dense material, at other times with light bales ; and the lower floors are frequently piled with casks containing mineral substances,124 ON THE CONSTRUCTION OF which produce not only a great amount of dead pressure upon the beams, but incur the risk of some of the heavier weights falling from some height upon the floor, and thus endangering the security of the structure by the fracture of the beam. These accidents are probably not frequent, but they should be guarded against; and the beams, arches, and columns should not only be calculated to resist the greatest load when operated upon by a dead weight, but the effects of impact produced by a body falling through a given space upon the floor. These calculations should apply to the first two floors of every warehouse, as the heavier descriptions of goods are almost invariably deposited in the lower stories. Mr. Hodgkinson, in searching experimentally for the strongest section, found that the old practice of making beams with equal ribs—such as recommended by former writers—was exceedingly defective : he proved a proportional between the top and bottom flanches; and the strain being less towards the ends of the flanches, it was reduced to the parabolic form, in order to give equal strengths throughout the whole length of the beams. This was an important discovery ; and as warehouse and factory beams are intended to be equally strong in every part, and sustain the load uniformly distributed, it is necessary to adopt the parabola in the form of the ribs, and to mark their relative properties with the body of the beams and with each other. In discussing these proportions, Mr. Hodgkinson demonstrates the curvature of the ribs as follows :—Suppose the bottom ribs to be formed of two equal parabolas, the vertex of one of them, AcB, being at c, d, C when, by the nature of the curve, any ordinate, d c, is as Ac x Be; the strength of the bottom rib, therefore, andFIRE-PROOP WAREHOUSES. 125 consequently that of the beam at that place, will be as this rectangle. It is shown, too, by writers on the strength of materials, that the rectangle Ac x Be is the proportion of strength which a beam ought to have to bear equally the same weight every where, or a weight laid uniformly over it. From this it would appear that the forms laid down by Mr. Ilodgkinson were rightly devised, and a great saving, not less than three-tenths, effected in the quantity of material used. Having pointed out the strongest form of beams, as applied to fire-proof buildings, it will be necessary in this place to refer to their strength, and to inquire into the nature of the strains to which they are subject. It has already been stated, that iron beams in warehouses have two distinct forces to contend against, that of direct pressure and the force of impact: with the former there is no difficulty, but the latter involves a proposition on which mathematicians arc not agreed. For practical purposes we may, however, suppose a case, such as a large cask of molasses, or box of heavy mineral substance, equal to one ton = 2240 lbs., Hilling from a height of six feet upon the floor. Now, according to the laws of gravity, a body falling from a state of rest acquires an increase of velocity, in a second of time, equal to 32 feet, and during that period falls through a space of 10 feet: this accelerated velocity is as the square roots of the distances ; and a falling body having acquired a velocity of 8 feet in the first foot of its descent, and 6 feet being the height from which a weight of one ton is supposed to fall, we have V (i x 8 = 2*449 x 8 = 19,502 for the velocity in a descent of 0 feet. Then, 10,592 x 2240 — 43,886 lbs., or nearly 20 tons, as the momentum with which the body impinges on the floor. In the present state of our knowledge, this momentum cannot probably be taken as the measure of the force of impact, but we may fairly estimate the latter as126 ON THE CONSTRUCTION OF exceeding that of momentum; and having these forces to resist, it will be necessary to guard against them, and to make the beams, columns, and arches, in the lower floors of such strength as will resist the blow, and neutralise its effect upon the floor. Although the iron beams and arches of a fire-proof floor may be sufficiently elastic to resist an impinging force* such as above described, it is still advisable to adopt other precautions, such as the bedding of timber along the top of the arches,* or to form the two lower floors entirely of wooden boards (three-inch plank) securely nailed to sleepers imbedded in concrete. This plan would give additional security, by the transmission of the impinging force over a larger surface; and, under these circumstances, the concussion w ould be made in the first instance on a soft elastic substance, before it could act upon the more rigid materials of iron beams and brick arches. In order, however, to remove all doubts as to security, it will be advisable to have stronger iron beams and columns in the tw^o low er floors; and having computed these strengths, they will probably be found nearly Correct in the ratio of 12 to 9. If on this data we take the breaking w eight of a beam, as suitable to the upper stories of a warehouse, at 22 tons, those of the lowTer storiea wrould require to be 29-32, or nearly 30 tons; and the^ columns, although less liable to fracture, w ill nevertheless be greatly improved by the introduction of a proportionate thickness of metal. Having, to ttye best of our ability, established the fact of perfect security in the use of iron beams and arches, the next point of inquiry will be as to the strength and * Since the above was written, I have been informed that the Act of Parliament for the regulation of fire-proof building's does not admit of any timber*whatever. In suoh case, I would advisj* the beams so to be made one-half stronger.FIRE-PROOF WAREHOUSES. 127 proportion of the columns. But before treating of this part of the subject, it may be proper to advert to the tie-rods, which are built into the walls and arches, and should unite the walls and girders as a species of net-work. These tie-rods are of great value, as they resist the strain of the arches, which, acting through their line of tension, not only secure the walls from being thrust out, but also retain the beams in the j>oraition best adapted to sustain the load. The usual practice in those districts is to have five lines of |-square rods in a width of 30 feet; two linos*'are imbedded in the wall, and the remaining three built into the arches: this is considered a perfectly secure building. But it must be borne in mind that cotton-mills are not subjected to heavy loads; and instead of live tie-rods of f-inch isquare, a warehouse should have seven lines of rods, each inch square. This will give a sectional area of about 11 inches iftSO feet, Avhich, taken at 25 tons to the square inch, will give a resisting tensile force of 275 tons. In factories, the resisting powers of the tie-rods seldom exceed 100 to 110 tons, which is und^r 4 tons to the foot, whereas the resisting forces in warehouse should not be less than from 9 to 10 tons to the square foot. In the construction of fire-proof buildings, it is not only necessary to secure the ends of the beams by extension-rods imbedded in the walls, but the arch-plates or “skewbacks” at each end should also be built into the wall; and this plate, as well as the ends of the beams, slightly raised above the level of the column, in order to allow for the settling of the walls, which invariably takes place as the weight increases in their ascent. For the strongest form and best position of columns supporting heavy weights, we must again refer to Mr. Ilodgkinson, as the very first authority. In his valuable treatise on the strength of pillars of cast iron and other materials, published in the Philosophical Transactions, Part128 ON THE CONSTRUCTION OF II., for 1840, and for which he received the gold medal of the Royal Society, will be found some of the most interesting and most useful experiments yet given to the public. From these researches it will be necessary to make some extracts, in order to ascertain the Jaws connecting the strength of cast-iron pillars with their dimensions, and to determine the best and strongest form adapted to the support of heavy weights. The first experiments were made upon solid uniform pillars, mostly cylindrical, with their ends rounded, in order that the force might pass through the axis; the next were of the same dimensions, with flat ends at right angles; and others, again, with one end rounded, and the other flat to the axis. They were broken at various lengths, from five feet to one inch (some with discs turned flat), and form a series of most interesting results. The pillars with discs give a small increase of strength above those with flat ends; but the approach to equality between the strength of pillars with discs, and those of the same diameter and half the length, with ends rounded, was nearly alike. The conclusion, as Mr. Hodgkinson observes, is, therefore, “ that a long uniform cast iron pillar, with its ends firmly fixed (whether by means of discs or otherwise), has the same power to resist breaking as a pillar of the same diameter and half the length, with the ends rounded or turned, so that the force would pass through the axis.” Mr. Hodgkinson, in the first experiment, gives the strength of cast-iron pillars with both their ends rounded and both flat; subsequently he experimented upon those with one end rounded and the Qtber flat, and in some cases with discs; and their results-being placed between those from the pillars with round and flat ends, gave the strength in a constant ratio, as under:—FIRE-PROOF WAREHOUSES. 129 Pillars. Breaking weight in lbs. Both ends rounded . . . 143 3,0 IV 7,009 7,009 16,493 One end rounded and one flat 6,278 13,499 13,565 13,557. Both ends flat 48V 9,007 20,310 These pillars, in each vertical column in this abstract, are of the same length and diameter; the strengths, therefore, in three different cases, reading downwards, are as 1, 2, 3, nearly, the middle term being an arithmetical mean between the other two terms. Mr. Ilodgkinson, therefore, found, by other experiments upon timber, wrought iron, steel, &c., that those, as well as every other sort and description of material, followed (as regards their strengths) the same laws; and that the strength of a pillar with one end round and the other flat is always an arithmetical mean between the strength of pillars of the same dimensions with both ends rounded and both flat. These are facts which should on no account be mistaken in the construction of fire-proof buildings; and it will be well to impress forcibly upon the public mind, that the principle is the same, however much they may vary in their ratio of strength. In treating of the strength of columns, I have endeavoured to establish principled which are not generally known, but which are proved to be fixed and determined laws, affecting the increase or diminution of strength, according as the ends are made round or flat. In order, therefore, to avoid error in the construction of buildings adapted for the support of heavy weights, it will be of some value to know, that the strength of pillars can be increased, according as their ends are shaped, in the numerical ratio of 1, 2, 3. Having investigated the subject at some length, it may be130 ON THE CONSTRUCTION OF necessary, before closing the report, to advert to a circumstance which appears to excite alarm, and increase the fears of individuals* respecting the safety of iron beams and brick arches as a perfectly fire-proof structure. It has been alleged, that in case of fire in any of the lower rooms in a warehouse, the intense heat generated by rapid combustion might melt the iron columns, and bring the whole edifice to the ground.* This is a possible, but a very improbable case, as an event of this .kind could never happen, provided the precautions enforced and inculcated in this inquiry be duly and properly observed. It is true, that negligence of construction on the one hand, and want of care in the management on the other, might entail risk and loss to an enormous extent; but it is no argument to say, that a warehouse built like a funnel, and provided with all the elements of conflagration, is attended wil^L risk, when it is w;ell known that a perfectly secures and perfectl^^tod fire-proof building can be erected, free from all the perils abov||fenumerated. In my own mind, there is not me shadow of a doubt as to the security of such a structure; and I do not hesitate to assert, that a well-built and properly-arranged fire-proof warehouse can not only be constructgflf, but may be to entail upon the commercial and manufacturing Mpmmunities of this country an important and lasting beMfit.” * There is only one instigate which has come to my knowledge of a fireproof building being injured by thj&tjufting of the columns, and that was at the works^of Messrs. Sharp./Jjipi3rts, & Manchester, where the pillars-were fixedBtt88p£of a steam-engine; and having a large quantity of wood them on the the boiler, for the purpose of drying, the heat became so intense as to c§SSfljl£em to bend, and ultimately to break. In this case therfront of^^^pOT|fe5S®®e was open, with a thorough draft direct across thmjsjjt^iug, which generated a most intense PJ^at, and caused the whole room to act as a reverberating furnace. Viewing the subject in this light, it cannot be c» „ 11 6 ■ „ 12 3 „ „ 13 0 „ 13 17 „ „ 14 14 „ ii 15 11 I „ 16 8 „ „ 17 5 „ „ 18 2 „ 18 19 „ » 19 16 „ „ 20 13 21 10 »> » 22 7 „ )> Deflection in inches. . -31 . -43 . -46 . -56 . -62 . *65 mh K • 75 . -81 . '84 . *90 . -92 . 1-06 . 1*12 . 1-18 . 1-25 . 1-31 . 1-43 . 1-50 . Broke with this <—4 in—> Fig. 57. sight.160 ON THE CONSTRUCTION OF Experimental Test of Beams 27 feet 4 inches between the Supports. tons. cwt. 18 15 on the centre 9 12 „ 10 9 „ „ 11 6 „ „ 12 3 „ n 13 0 ■ „ „ 13 ii i 14 14 | | 1511 „ n 16 8 „ 11 5 „ 18 2 I 18 19 „ „ 19 16 '** » 20 13 „ 216 n n 21 16 „ 29 6 „ Deflection in inches. •50 •56 •59 •65 •15 ‘84 *9$ m 1-06 1-12 1-18 1*31 1-43 1*56 1-71 1-84 1- 96 2- 12 On account of the danger of approaching too near the beam, the deflections, till 29£ tons were laid on, were act taken. The beam ultimately broke with that weight. About 200 of the loner beams havinor been tested with o o 12 tons in the centre, they deflected differently from f to £ of an inch. The same number of short beams, tested with 8 tons in the centre, deflected differently from rV to tb$inf deSqjibra at pp. 73 and 74, the timbers on each side would then have wen useful in preventing lateral flexure; but they would not have con* tribbted in any great degree to the vertical bearing powers of the beam. These defects afo the more apparent in the compound Construction of iron and wood, from the position of 4the iron plate and the difference in quality, as well, as the resisting powers of the material. When they are United in this form, they can never exert, at one and the same time, those duly proportionate Tjgwers of resistance .which inKJmogeneous material is sure to be fh!^ developed, and calculated to exhibit its full, powers of resistance, f THSyENp. ^:0/ aft i 0 RETURN ENVIRONMENTAL DESIGN LIBRARY TO—-► 210 Wurster Hall 642-4818 LOAN PERIOD T* 2 HOURS 2- 3 4 K 6 1 RESERV E Return books promptly. $5.00 processing fee may be levied. DUE AS STAMPED BE LOW UNIVERSITY OF CALljfgftlAt BERKELEY FORM NPJDDI 3A, ; BERKELEY, CA 94/20RARE CQ3MfifiMbDb