406 REPORT — 1856. 



scribed. These defects are invariably present when the plates are not sound ; 

 but in other respects, where the bars which form the pile are clear and free 

 from rust or scoria, and are well-welded in the rolling process, the wide dif- 

 ference between the temperature of one side and that of the other produces, 

 apparently, no injurious effect on the strength of the plate. It is, however, 

 widely different when the whole of the plates are exposed to the same de- 

 gree of temperature, as in this position the strengths are increased or dimi- 

 nished according as the temperature approaches or recedes from the point 

 where the strength is a maximum. 



In order to show how the results were obtained, it will be necessary to 

 describe the apparatus and the mode of conducting the experiments. 



The apparatus consisted of a powerful wrought-iron lever, Plate IV. A, 

 figs. 2 and 3, capable of imparting a force of more than 100,000 lbs,, or 45 tons 

 per square inch to the specimen to be broken. The lever is supported in a 

 cast-iron standard or frame B, arranged for the reception of specimens of the 

 material to be subjected to a crushing force or tensile strain. On the short 

 arm of the lever the plates and bars (one of which is seen at a) were 

 suspended by a shackle c, and held down to the bottom of the cast-iron 

 standard by the rod and screw e; on this rod the box, b, was fixed, and pre- 

 pared to hold a bath of oil or water, in which the iron to be broken Avas 

 immersed. Below this box was a fire-grate, d, for heating the liquid in the 

 bath to the required temperature, and this grate could be drawn backwards 

 from the box b, when the required temperature was attained or when it be- 

 came too high. The fulcrum of the lever is shown atf, and the scale in 

 which the weights were placed at g. The cast-iron standard was firmly 

 bolted to the heavy balks of timber upon which it stands, and the pressure on 

 the specimen was adjusted by placing weights in the scale. 



The plates experimented upon were of the form shown in fig. 4, reduced 

 at a, to 2\ inches wide, and at 6 to 2 inches wide, in order to secure frac- 

 ture at the part of the plate immersed in the liquid in the bath. At each 

 end two holes are drilled 



to receive the bolt which '°' 



fixed them in the shackles. 

 The wrought-iron bars 

 were formed in a similar 

 manner. They were f inch 

 in diameter, reduced to f 

 of an inch at a, and to ^^ 

 inch, or ^ inch at b. The 

 shackles were made to clasp 

 the bars below the shoul- 

 ders so as to apply the strain ^^^- ^• 

 requisite to cause fracture. t^ i ^^^ 



It is evident that the weak- . 1 h~"^ „ h „ 'h-^^ 



est part of the bars being ,^^2- "^ ^..-^.p<,.~^-^i--^ 



within the bath, breakage •*■ -J-"'* "^ 



was sure to occur at that 



point where the temperature was raised or lowered to the required degree. 



With these preparations, the experiments proceeded as follows : — the bar 

 to be broken was fixed between the shackles of the lever; and, if necessary, 

 the bath was filled, and the fire drawn close under it; as soon as the intended 

 temperature was attained, the lever was let down by the crab, and weights 

 carefully added to the scale until the bar broke. During the process the 

 temperature was observed from time to time, and the fire adjusted accord- 



"V_ 



O - H» o 



r" 



-4-— ><r G-k- — ->k— 3-'-^-— 4-— ->j 



