EXPERIMENTS WITH BUILT-UP COLUMNS. 635 



the lacing points in the first six columns were not close enough to 

 each other to prevent secondary flexure. The ratio of length to 

 least radius of gyration was smaller for each entire column than 

 for each section of it between lacing points. 



Columns Nos 3 to 6 all failed by secondary flexure at about 

 19,000 lb., nothing being gained by having the lacing in 4, 5 and 6 

 heavier than in No. 3. In other words, the lacing bars in No. 3, 

 which had an area equivalent to 18"4 per cent, of that in the longi- 

 tudinals, were sufficiently strong to develop the strength of the 

 longitudinals when these had lacing points about 9 J in. apart. 



Having got these results, which, under the circumstances, 

 were as satisfactory and consistent as could have been expected, 

 the next step taken was to try the effect of placing extra pieces 

 of lacing in columns resembling Nos. 2, 3 and 4, so as to reduce the 

 distance between lacing points. The new columns were numbered 

 2A, 3A, and 4A. Their lacing bars had about the same weight 

 as those in 2, 3, and 4, but extra pieces were placed transversely, 

 extending from each apex in the former system of lacing to midway 

 between the lacing points on the opposite longitudinal bar. The 

 efifect of introducing these extra pieces was to prevent secondary 

 flexure in the longitudinals. 



Column 2A was 17T per cent, stronger than 2. 

 Column 3 A was less than 1 per cent, stronger than 3. 

 Column 4A was 21-3 per cent, stronger than 4. 



The added metal, in each case, was 14-7 per cent, of that pre- 

 viously there. The result in the case of column 3A may evidently 

 be discarded, because some defect, which observation failed to 

 discover, may have existed ; or, as subsequent investigation 

 made to appear probable, the lacing bars in 3A may possibly have 

 been 6 per cent, lighter than those in 3. The writer had not time 

 to confirm this. 



The result, then, of the three last tests is to show that the 

 system of lacing adopted in them is sufficient for developing the 

 strength of the longitudinal portions of the columns, for each of 

 the columns failed by primary flexure — i.e., they bent sideways, 

 each as a whole, without secondary flexure. So that the experi- 

 ments, few in number and small in scale as they are, serve to 

 demonstrate that, with the particular longitudinals here made use 

 of and with these dimensions for the columns, L lacing bars, whose 

 weight per yard is about 18 or 20 per cent, of that of the longitudinal 

 bars, and arranged as in these columns, are able to prevent 

 secondary flexure and are strong enough to hold the parts of the 

 column together. 



Further experiments, on a different scale, would be needed to 

 show whether or not similar proportions would hold in the case 

 of square built-up columns, with other sizes of longitudinal angle 

 bars, or with different proportions of length to width. | 



