EXPERIMENTS WITH BUILT-UP COLUMNS. 639* 



The ratio in the case of the lacing bars was, in general, greater 

 still. That for column 2A, for instance, was 78 to 1 ; and yet 

 this lacing proved sufficient to develop the strength of the column 

 as a whole — i.e., the column failed as a unit, without appreciable 

 distortion such as would result from the yielding of parts. 



The common rule, and a good one to follow, is to make the 

 ratio of length to least radius of gyration, for each portion between 

 lacing points of a longitudinal component of a column, the same 

 as for the whole column. The present experiments support the 

 rule — e.g., secondary flexure was observed when the ratio was 

 about 55 to 1, and the lacing bars strong enough, while the columns 

 proved to be satisfactorily braced against secondary flexure when 

 the lacing points were so spaced as to make the ratio 28 to 1. 



On the other hand, it must be borne in mind that experiments, 

 such as those carried out by Talbot and Moore, show that any 

 lacing bar may act in compression. Now, if the same stress per 

 unit of area be allowed in lacing bars as in the principal com- 

 ponents of a column, it would follow that the ratio of length to 

 least radius of gyration should be the same for lacing bars as for 

 the other portions of the column. The practice is to make this 

 ratio far greater for the lacing than for the principal bars, and the 

 present experiments seem to show that the practice is to some 

 extent warranted. This leads to — 



(3) The object of lacing. In the writer's opinion the object 

 is to hold the longitudinal members together and so prevent them 

 from yielding singly, by lateral deflection, as long columns. This 

 is not the same thing as providing against deflection of the column 

 as a whole. It seems incorrect to design a column as a beam 

 subjected to bending, because the column, under working loads, 

 should not deflect as a whole, nor, with proper workmanship, 

 should there be any perceptible initial deflection. Moreover, it is 

 only in exceptional cases that loads are applied eccentrically. Nor 

 is it sufficient to consider each longitudinal component separately, 

 designing the lacing to resist bending in two planes, for it must 

 be remembered that the lacing bars fixed to one longitudinal are 

 merely attached to other longitudinals, which may tend to bend 

 in the same direction as the first. This is not the same thing as 

 staying an upright post from an immovable wall. 



There appears to be no method of deducing a rational formula 

 for determining the dimensions of lacing-bars, connecting longi- 

 tudinal column components with each other. Their raison d'etre 

 is (1) to prevent flexure of individual longitudinal components, 

 and (2) to transmit any equahty of loading from one such com- 

 ponent to the other. What appears to be needed is a series of 

 tests, covering built-up columns of all good designs. Nothing but 

 actual tests can enable us to make satisfactory allowance for the 

 accidental imperfections in workmanship, the variation in elasticity 

 of the materials used in construction, and unintentional eccentricity 

 of loading. 



