46 REPORT — 1860. 



sustain at least 4000 impacts, bending them through one-third of their ulti- 

 mate statical deflection. They ascertained also, that when the load was 

 placed upon the bars without impact, if the deflection did not exceed one- 

 third of the ultimate deflection, the bar was not weakened ; but that if the 

 deflection amounted to one-half the ultimate deflection, the bars were broken 

 with not more than 900 changes of load. With wrought iron bars they 

 found no perceptible effect from 10,000 changes of load, when the deflections 

 were produced by a weight equal to half the statical breaking weight. 



These experiments are interesting so far as they go, but they are very in- 

 complete as regards wrought iron. For wrought-iron bars they were not 

 continued long enough, nor do they apply to those larger constructions in 

 which the homogeneous bar is replaced by riveted plates. The influence 

 of change of load on riveted constructions possesses a special importance, 

 from its bearing on the question of the proper proportion of strength in 

 plate and tubular bridges. Do these constructions gradually become weak- 

 ened from the continual passage of trains? and is it requisite to make allow- 

 ance for such a deterioration by increased sectional area of material in their 

 original construction? These questions I have sought to solve by the fol- 

 lowing experiments. 



As the load is brought upon bridges in a gradual manner, the apparatus 

 is designed to imitate as far as possible this condition. A riveted beam is 

 fixed on brickwork supports, 20 feet apart. Beneath this is placed a lever 

 grasping the lower web of the beam, and fastened upon a pivot at the ful- 

 crum. At the other extremity it carries the scale and weights. This lever 

 is lifted clear of the beam, and again lowered upon it by means of a connect- 

 ing rod attached to one of the arms of a spur-wheel placed at a considerable 

 distance overhead. In this way any required part of the breaking weight 

 can be lifted off and replaced upon the beam alternately by the revolution of 

 the spur-wheel. The apparatus is worked night and day by a water-wheel, 

 and the number of changes of load is registered by a counter. 



The girder subjected to vibration in these experiments is a plate girder of 

 20 feet clear span, and of the following dimensions : — 



Sq. in. 



Area of top : 1 plate, 4 in. x £ in 2 - 00 



„ 2 angle-irons, 2 X 2 X ^ 2*30 



4-30 



Area of bottom : 1 plate, 4 in. x \ in 1*00 



„ 2 angle-irons, 2 X 2 X ^ 1*4 



2-40 



Web, 1 plate 15i X | 1'90 



Total sectional area 8*60 



Depth 16 in. 



Weight 7 cwt. 3 qrs. 



Breaking weight (calculated) 12 tons. 



This beam having been loaded with 6643 lbs., equivalent to one-fourth of 

 the ultimate breaking weight, the experiment commenced. 



