DISCUSSION. 149 



obtaining them were described by him as erroneous and faulty. 

 From an inspection of the stresses and graphic stress diagrams 

 prepared by Mr. Haycroft it is apparent that he has ignored the 

 important action of the wedges altogether ; and were brace Y 3 

 omitted, as he advocated, the design would be so materially 

 altered that the stresses then obtained by him, even if correct 

 for bis design, could not possibly be correct for the original 

 design. Although Professor Warren has given us much useful 

 and practical information with regard to timber and its appli- 

 cation to bridgework, I consider that had the paper contained 

 more information on the strength of long columns of the timber 

 referred to, and also the determination of u f" and " a " in 

 Gordon's Formula, its value would have been greatly enhanced. 



H. H. Dare, b.c.e. — Mr. Yicars has demonstrated very clearly 

 the peculiarity of the stresses in the truss under discussion ; and 

 taken in conjunction with Professor Warren's paper his remarks 

 should prove of great interest, since, so far as I am aware, there 

 is little or no previous literature on this subject. Mr. Haycroft 

 has apparently gone astray in neglecting the fact that no arrange- 

 ment of loads can affect the diagonals so far as tensile stress is 

 concerned, and that for that reason, only the loads producing the 

 maximum compression in each bar need be considered. Take Y \ 

 for example. The only loads which can affect it are those at (1), 

 (2) and (3), and it is immaterial whether the other loads be on 

 the truss or not, since they cause no direct or counteracting stress 

 in the bar. As given on page 140, the maximum stress is the 

 summary of the effects of these three loads. The principle of 

 determining the maxima stresses for live and dead loads by 

 multiplying the stresses obtained for the dead load by the pro- 

 portion between the two classes of loading, has been condemned 

 by Mr. Haycroft as " certainly incorrect for the bracing and ver- 

 ticals " ; but without doubt this is by far the simplest and most 

 correct method of calculation under existing conditions. The 

 action of the wedges and the necessity for the bars Y 3 and Y 8 have 

 already been explained by Mr. Yicars and there is no necessity 

 for further reference to those features in the design, but I should 

 like to bring before the meeting the following simple investigation 

 into the relation of stresses in a rectangular ironbark beam. 

 Assuming modulus of rupture at 18000 lbs. per sq. in., and ulti- 

 mate shearing stress at 2000 lbs. per sq. in., in order that the beam 

 may be equally as strong for shearing as for direct stresses, the 

 following should obtain : — Supposing the beam to be loaded with 

 a uniformly distributed load, 



Maximum B. Mt. = W H ~ ; Mt. R. = \fb d 2 



• ' 8 ~ 6 / ° tl > VY ~ 6 x I ~ :i I V/' 



