1891.] On tlie Plasticity of an Ice Crystal. 



FIG. 3. 



32i 



FIG. 4. 



jf the field and then turning it 7 till it looked as dark as possible, 

 this is done the axis lies in the principal plane of the instru- 

 icrit. It will be noticed that in each crystal the direction of the 

 jptic axis is almost uniform. I imagine that the two crystals existed 

 virtually in the bar, but that their optic axes were so nearly parallel 

 that in the polariscope they behaved as one crystal. The kind of 

 lear that must have taken place in the upper crystal is represented 

 fig. 4 by a number of layers of finite thickness slipped over one 

 mother. 



I cannot say definitely that the bending was either slower or faster 

 than in a bar all one crystal with the axis vertical. 



The part beyond the dotted line is perhaps due to the intrusion of 

 mother crystal completely overlapped by the main crystal, or perhaps 

 some alteration of the optical qualities due to elastic strain. 

 Exp. 7. Another bar cut from the same lump was a single crystal 

 rith the axis nearly longitudinal, inclined perhaps at 5 to the side of 

 bar. Breadth 10'7, depth 1O5, distance between supports 84 mm., 

 weight 1'29 kilos. After six hours, during which time the tempera- 

 ture had been between 1'7 and 0'6 C., the bar was found lying at 

 the bottom of the box broken into two pieces. It had bent so much 

 lat it must have slipped down between the supports and been broken 

 the fall. The two parts could be accurately pieced together. At 

 the dotted line there was a very rapid but not sudden change in the 

 lirection of the optic axes. The shape of the surfaces normal to the 

 )tic axes is shown in fig. 5 (p. 330). These sliding surfaces must 

 ive the geometrical property that the normal drawn at any point to 

 my point is also normal to all the surfaces it cuts within the bar. It 

 in fact parallel to the optic axis all along its course. 



z 2 



