140 TENSION TKSTINC. 



When the. load has become large and the bar has 

 stretched to a considerable extent, it will be seen that the 

 scale of the metal that is, if the bar has not been turned 

 lupins to crack and peel off, and also the bar will be seen 

 to visibly diminish in diameter, this diminution being 

 at first uniformly distributed along the whole length of the 

 bar. 



When the maximum load is approached, however, the 

 reduction of diameter begins to assume a maximum 

 amount at one point of the bar. A distinct thinning takes 

 place at this point, and the specimen appears to have a 

 " waist." As this local contraction sets in, the load 

 lias probably reached its maximum amount, and it is 

 soon found that the bar will no longer support the load 

 which has caused this contraction. The explanation of 

 this is simple. When local contraction begins to manifest 

 itself, the area of the cross section of the bar begins to be 

 more rapidly reduced, and, although the load may not be 

 increased, the load per unit area rapidly becomes greater. 

 This increase of load per unit area, or stress, has the imme- 

 diate effect of still further increasing the local extension 

 and contraction, and therefore the stress also. When this 

 point is reached, the life of the bar is at an end, and 

 fracture soon takes place by the tearing apart of the bar. 



The local contraction only takes place in a marked 

 degree in the case of the more ductile metals, such as mild 

 steel, the better kinds of iron, copper, and metals of a similar 

 nature. By careful manipulation it is possible, in the case 

 of ductile metals, to ascertain both the maximum and the 

 breaking loads. In order to do this, the maximum load is 

 to be quickly noted, and then the jockey-weight rapidly 

 run back towards zero until the load just balances the stress 

 in the bar ; by carefully adjusting the load in this way 

 balance may be preserved until fracture takes place, and 

 the actual breaking load obtained. In ductile materials 

 this breaking load may be considerably below the maximum 

 load. 



Throughout the test, as the load is increased and the 

 bar in consequence stretches, the weighing beam will fall 

 from its horizontal position, and will have to be raised 

 again, and balance preserved by the hydraulic appliances 

 provided for taking up the strain of the specimen. In the 

 case of the more ductile metals there will have to be quite 

 a considerable interval of time between the increase ol the 

 load in each case and the measurement of the stretch, as a 



