Plastic Crystals of Ammonium Nitrate. 9 



the experiments. The part that is shown with a larger scale 

 along the time axis, shows the portion of the bend which 

 may be called initial. In this case, '05 cm. depression 

 resulted at once on loading, and *10 cm. within two minutes. 

 After thirteen minutes this had increased to '12 cm., which 

 represents the initial bend. In obtaining the measurements 

 plotted in figs. 7 and 8, the successive loads were put on at 

 •as nearly equal intervals as possible, and thus fair values 

 proportional to the initial bend were obtained. It will be 

 seen, however, that, though there is a large initial bend, it 

 is not possible accurately to discriminate between it and the 

 subsequent bending, which tends to become slower and 

 slower as time goes on. 



It will be seen that the bending curves for a crystal 

 supported near its ends are similar to those obtained when 

 one end was clasped and the load applied at the other. 



An attempt was made to observe the amount of stretching 

 of a crystal clamped vertically and supporting a con- 

 siderable weight from its lower end. Many difficulties, 

 however, were met with. Thick crystals could only be 

 obtained of an inch or two in length. It was necessary to 

 observe marks on their surface with a cathetometer. On 

 each occasion, before any definite increase in length could be 

 observed, fracture occurred. In one case of breakage under 

 considerable load, the stress was calculated and the same 

 crystal arranged to have the same maximum stress in a 

 bending experiment. The rate of deflexion was exceedingly 

 slow ("01 cm. in one and a half months). Thus the stress 

 that resulted in fracture in the direct tension experiments 

 was hardly large enough to give observable deflexions by 

 the bending method. It seems, at first, that this means the 

 failure of the method of experiment most likely to lead to 

 results the meaning of which can be understood. And, no 

 doubt, if this method were successful, important and easily 

 interpreted results would ensue. However, it must be 

 remembered that we have no direct evidence that a simple 

 tension will result in a plastic flow. This point will be 

 discussed more fully later in the paper. 



The change of form is not due to a simple viscous flow, 

 for the rate of bending for a given crystal and load is not 

 constant, but generally decreases from the time of loading 

 onwards; this is the characteristic of a plastic change. 

 Thus, there is not a unique value for the viscosity of the 

 substance. It may, however, be of interest to proceed as 

 if the flow were viscous and to calculate from the slope of 

 the portion of the curve after the initial rapid bending, a 



