636 PROCEEDINGS OF SECTION H. 



It is true that some of the lacing members in the last three 

 columns were broken or crumpled in the testing machine, but this 

 may be taken to be the result, not the cause, of the collapse. We 

 may take it that the column bent as a whole, failing virtually, and 

 then, when the lateral deflection reached a certain amount, the 

 lacing yielded. This seems a reasonable inference, because these 

 three columns, 2A, 3A and 4A, failed in a manner quite different 

 from 2, 3 and 4. In the case of 3A, as mentioned, there probably 

 was some defect causing the premature yielding which there 

 occurred. 



The table of results (A) shows that the ultimate stress in the 

 longitudinal L bars in column 4A was 38,276 lb. per square inch, 

 which is approximately the yield point for the mild steel from which 

 the bars were rolled. It is not likely that any considerably higher 

 figure for the ultimate strength of a column can be obtained with 

 these longitudinal L bars, whatever system of dimensions of 

 lacing be adopted, for the yield point in compression tests of this 

 nature is virtually the ultimate strength. 



It may be that columns with longitudinal members, more 

 massive in proportion to the transverse dimension of the column 

 than those experimented with here, might have their stength 

 developed with lacing bars lighter in proportion than those in the 

 present case ; but it must be remembered that the question is often 

 one of stability. Many columns have channels instead of L bars. 

 The Quebec bridge had four channels as longitudinal members in 

 the strut which caused failure. The lacing bars there had a section 

 0'32 per cent, of that of the column itself, a vastly lower figure 

 than the 4|- or 5 per cent, which we found necessary here. The 

 Quebec bridge lacing, moreover, consisted of a double system 

 of braces, connected by transverse pieces. The existence of these 

 transverse pieces, which acted in tension, really weakened the 

 structure, because they compelled the diagonals to share largely 

 in the load on the strut which theyw^ere supposed merely to brace. 



The writer is of opinion that systematic tests, comparing 

 different sizes of lacing bars, and different systems of lacing, on 

 columns of various sections, would afford much better guidance 

 for the proportioning of struts than any formulae he has seen. In 

 any case, we should recognise that there is great uncertainty as to 

 eccentricity of a column's axis, or of loading, also that there are 

 unknown variations in the strength and elasticity of the material 

 of which a column is built, and, further, that accidental lateral 

 blows may be given, which may be cumulative with considerable 

 wind pressure ; besides all which, if the strut is not vertical, the 

 column's own weight tends to deflect it The attachment of a cross 

 beam in a bridge may deflect a strut. For all these reasons it is 

 urged that lacing bars should be used liberally and not cut down 

 for economy's sake. Extra dimensions in the parts forming 

 columns may be made to give great increase of strength, and the 

 cost is comparatively small. 



