292 



THE CIVIL ENGINEER AND ARCHITECT'S JOURNAL. 



[September, 



mitted to the Society in 1S24, who after investigating tlie merits of 

 the plan — tryii^g some experiments upon models — and witnessing its 

 successful application inbuildings— finally testified to Mr. Ainger, their 

 sense of the merit of his contrivance fur strengthening beams, by 

 formally awarding to him the thanks of the Society. Mr. Ainger's 

 truss, as applied to a girder of 34 ft. span, consisted of two sets of iron 

 rods, (one on each side of a single beam of timber,) secured by screws 

 and nuts, to abutment plates, notched upon the upptr side at ihe ends ; 

 these rods descended so low as to admit of two supporting plates being 

 inserted between them and the beam, and forced up against its under- 

 side by the end screws, so as to give to the girder two intermediate 

 points of support; thus dividing the whole span into three bearing 

 spaces of equal extent ; the side view closely resembling fig. 1. This 

 girder is described as having answered very well, it is stated to have 

 carried a leaden flat for two years, without sensibly altering its form. 

 Mr. Ainger also recommended that notches should be cut in the 

 upper side of the beam extending to one-third of its depth, and filled 

 " with thin wedges of hard wood or metal, forcibly driven in," so as 

 to augment its strength and stiifness, by putting its upper side in a 

 state of compression before being loaded.* 



It was the article above referred to which first drew the attention of 

 the writer to this particular subject, and induced him, in company with 

 Mr. McClure, to undertake, in the year 1832, a few experiments upon 

 model girders. 



These models were accurately made under the direction of Mr. 

 McClure, and were _/? re IK number; each was composed of two parts, 

 or flitches, so united as to bear a strain together; each side, or flitch, 

 was 53 inches long, 1-75 inches deep, and 1-16 inches broad, having 

 together a sectional area of 4-06 superficial inches; and all the models 

 ■were made of clean, straight grained, and well seasoned white pine ;t 

 numbers 1, 2, and 3, were from the same plank, and Nos. 4 and 5 from 

 timber of the same lot and quality. 



Modd No. 1 : Was trussed in a mode similar to that prescribed by 

 Mr. Ainger; a single wrought iron rod exactly one-fourth of an inch 

 square, being secured upon the npper side, at both ends, by quarter 

 inch square pins, resting horizontally upon abutment plates one -twelfth 

 of an inch thick, which embraced each end of the girder; this rod de- 

 scended between the flitches so low as to admit the insertion of two 

 bearing plates, b and c, fig. 1, which were forced in between it and 

 the underside of the beams, at equal distances from the ends, and from 

 each other, so as to give to the girder a small camber. (See plan and 

 section, fig. 1.) And, in addition, 34 equidistant cuts were made in 

 this model, to a depth of one-third of the whole, and tightly tilled with 

 thin pieces of hard oak well driven in. 



Model No. 2 : Was similar to No. 1, (fig. 1,) in all respects except 

 that the hard oak wedges were omitted. 



Model A'o. 3: Was a plain stick, formed by nailing the two flitches 

 together, side by side, with four nails in the centre line, (see fig. 2, 

 side view.) 



Model J\'o. 4: Was a girder trussed exactly like Nos. 1 and 2, esi- 

 cept that a central bearing plate, one-sixteenth of an inch thick, was 

 inserted between the beam and truss rod at a, fig. 1. 



* This was conformable to the results of some experiments by M. Duhamel. 

 who show ed that when a bar of soft tvcod, such as willow, was cut one-third 

 through from the upper side, and this cut filled with a thin piece of hard 

 wood sluck in pretty tiyht. its ultimate transvasf strength was tlierehij iitcreastd 

 nhont 10 per cent. (Barlow's Essay on the Strength and Stress of Timber, 

 3rd. K<lit., 1826.) 



t This is the Pinua Strobus of Michaud's North American Sylva (vol. iii, 

 p. 159), the loftiest tree of the American forest ; it is employed here for an 

 immense variety of purposes, and is one of the most valuable timbers which 

 we pi'Ssess. It Is called Weymouth Pine, in Tredgold's Carpentry ; and ive 

 may here remark that the concurrence of three of our experiments indicates 

 that the ultimate strength of the white pine, used in our models— which was 

 of the best quality supplied Ly the liimuer market of Philadelphia — did not 

 exceed 868 lb. for sticks of the scantlirg, and length of bearing, used by us. 

 Wherefore the constant quantity c, of the table iu Tredgold's Carpentry^ 



(2nd Edit., London, 1828, p. 56,) calculated by the formula =--^ would b* 



509 instead of 658, as Tredgold has it, which last is certainly too high a num- 

 ber ; for the breaking weight of our plain stick, (model No. 3,) if calculated 

 by it, would have been 1122 lb., instead of 868 lb., (as we actually found it 

 to be,) or tliiriij per cetit more. 



Model No. 5 : Was a girder trussed in the common way — vtith one 

 king bolt, two iron abutments, two hard oak braces let into each flitch 

 nearly one-eighth of an inch — and keyed up to a slight camber. (See 

 fig. 3, which shows the girder with one flitch removed.) 



df. 



3 



All of these models were put under strain in the same manner, their 

 own weight being counterbalanced ; viz., by resting them against two 

 fixed supports, exactly 50 inches asunder, and applying the straining 

 force, by a vertical pull upwards, through an iron stirrnp with rounded 

 edges, slipped over the model — which was turned bottom up — and 

 attached to the shorter arm of an accurate steelyard, which weighed 

 up to 1,288 lbs.; and the successive strains were put on by moving 

 the weight of the steelyard, as in weighing, so as to add usually 56 lbs. 

 at a time ; the beam of the steelyard being adjusted to a level at each 

 addition of strain, and the deflections being carefully measured — from 

 a fine string joining the extremities of the models — by a diagonal scale 

 to hundredths of inches. (See fig. 4, which gives a general idea of 

 the experimental apparatus,/ being a girder under trial.) 



The subjoined tabular statement exhibits, at one view, the results 

 of the trials of the strength of the several model girders ; which it will 

 be recollected, were all of precisely the same scantling and bearing 

 length. 



It must be observed that our steelyard not having sufficient power 

 to break model No. 4, it was subsequently broken by a lever, in order 

 to ascertain the position of the neutral axis of fracture. 



The five following results were common to all the models when 

 broken, and though similar to what have been observed by other ex- 

 perimentalists, they may as well be stated. 



1. The neutral axis of fracture, occupied precisely the same place 

 in all, being at four-sevenths of the whole depth down from the top, 

 or one-fourteenth below the middle, (Barlow's experiments on fir 

 beams fixed it at five-eighths of the depth, which is nearly the same.) 



Giving, area of fibres crushed, 2-32 square inches. 



Ditto torn asunder 1'74 „ 



Total sectional area in superficial inches 4-06 



2. The first indications of fracture, always appeared on the upper 

 side, by the fibres crushing. 



3. The signs of crushing at the time of fracture could be traced 

 each way from the centre some five or six inches. 



