190 Experiments on Model Girders. 



design, characterised by the strictest adherence to the 

 indications of mathematical calculation compatible with 

 simplicity of construction. In general appearance, girder B 

 is hardly distinguishable from A ; a close inspection reveals 

 the following differences — (1) The replacing of the massive 

 ends, each of which consists of five separate pieces of metal 

 united by numerous rivets, by a single angle iron ; (2) The 

 omission of certain minor vertical members, upon which 

 there is no calculable load ;. (3) The introduction of angle 

 irons instead of the plain flat bars for the compression 

 diagonals. The time taken in construction was carefully 

 noted, also the weight of metal, number of rivets, and other 

 particulars, for which see subjoined tabular statement. The 

 great reduction in the amount of time expended in the case 

 of girder B was due mainly to the simplification in the con- 

 struction of the end vertical pieces. 



In testing, the girders were supported and loaded exactly 

 -as in the actual structure, the top members being braced to 

 prevent lateral deflection. The difference in strength, and 

 the mode and nature of the fracture in each case, was in 

 accordance with what was predicted on theoretical grounds. 



Girder A. Girder B. 



Weight of metal ... 1 lb. 9 oz. ... ... 1 lb. 5 oz. 



Time occupied in mak- 14 hours 10 minutes 5 hours 30 minutes. 



ing 



Actual breaking load 208 lbs 771 lbs. 



Calculated ditto 182 lbs 800 lbs. 



Eatio of breaking load 133 to 1 ... ... 587 to 1. 



to weight of girder 



Nature and position of Buckling of com- Terminal tension dia- 



fracture pression diagonals, gonal torn across 



leading to general through rivet hole, 

 collapse of structure 



