6 Mr. W. N. Bond on the Properties of 



local stresses, experiments were next made by supporting a 

 crystal near its two ends on small steel pegs, and loading it 

 in the middle by means of a small loose wire loop (see 

 PI. I. fig. 1). This loop would produce much less local dis- 

 turbance than the wedge-form of end-clamp. This fresh 

 method of support is discussed more fully later in the paper. 



The crystal used for the test shown in fig. 5 was broken 

 near one end as a result of the test, and was but little bent 

 in the remaining portion. It was, therefore, used agaiu for 

 this new method of experiment. It was supported on points 

 2 25 cm. apart, and loads of 50, 100, and 200 grams were 

 applied seriatim. The deflexion was exceedingly slow, and 

 a load of 400 grams was, therefore, applied. This experi- 

 ment gave results shown in fig. 6. On removing the load, 

 the elastic return could not be detected with certainty. On 

 applying a load of 200 grams in the opposite sense (i. e. by- 

 inverting the crystal) no deflexion could be detected. On 

 again reversing and applying 200 grams still no effect was 

 observed. When 600 grams load was applied, rapid bending 

 took place and fracture followed. 



Thus again we find that small loads produce a very small 

 deflexions and large loads considerable change. The 400 



Fig. 6. 



gram load in fig. 6 ought to give the same maximum stress 

 as about 75 grams in fig. 5. And, if the bending took 



