umn so as to stiffen the latter and check the transverse vibration from passing trains.

XIII. PROPORTIONING COLUMNS FOR DIRECT LIVE AND
DEAD LOADS AND IGNORING THE EFFECTS OF BENDING
CAUSED BY THRUST OF TRAINS AND LATERAL VIBRATION.

The practical effects of this fault can be seen to best advantage by standing on one of the high platforms of one of the elevated railroads of New York City. The vibration, by no means small, from an approaching train can be felt when it is yet at a great distance. Some may claim that this vibration is not injurious; but they are certainly wrong, for what does it matter, so far as the stress in the column is concerned, whether the deflection be caused by vibration or by a statically applied transverse load, so long as the amount of the deflection is the same in both cases? It takes metal, and considerable of it, to make columns strong enough to resist bending properly; and a sufficient amount should be used to attain this end.

XIV. OMISSION OF DIAPHRAGM WEBS IN COLUMNS SUBJECTED
TO BENDING.

If the diaphragm web be omitted in such a column, reliance must he placed on the lacing to carry the horizontal thrust from top to bottom. But even if the lacing figure strong enough to carry it, which is unusual, it is wrong to assume it so, for the reason that one loose rivet connecting the lacing-bars will prevent the whole system from acting, as will also a lacing-bar that is bent out of line. Decidedly every column that acts as a beam also should have solid webs at right angles to each other.

XV. INEFFECTIVE ANCHORAGES.

On account of both rigidity and strength, every column ought to be anchored so firmly to the pedestal that failure by overturning or rupture would not occur in the neighborhood of the foot, if the bent were tested to destruction. The flimsiness of the ordinary column-foot connection is beyond description.