6l2 



MOTONORI MATSUYAMA 



These results seem to show some decrease of rigidity with rising 

 temperature. Near the melting-point of ice, its rigidity will 

 decrease more rapidly than at lower temperatures, so that the 

 rigidity-temperature curve will be concave toward the temperature 

 axis. If we consider rigidity to depend upon the temperature t up 

 to the second power and calculate the coefficients from the fore- 

 going data, we obtain 



«= (0.18—0 .095 /— 0.0020 /^)X 10^ c.g.s. 

 This is represented by the curve in the following figure (Fig. 2). 

 the observed points being denoted by crosses. 



1-5 



K ^1 



__ ._^___-,p 



,^^^^^t^'^ 



t-^" — 1! 't 



I Mil ~F I ~T I I I 



Temperature in centigrade 

 Fig. 2. — Modulus of rigidity of ice composed of parallel crystals at various tem- 

 peratures. Crosses denote the values when shear is parallel to the optic axes of the 

 crystals and dots the values when shear is parallel to their basal sections. 



Similar observations were made for ice bars with the optic axes 

 of the constituent crystals parallel to the axis of the bar. The 

 test piece was 20.95 cm. long and 1.93 cm. in diameter. The 

 crystals were for the most part several millimeters in diameter 

 and never so long as to reach from one end of the bar to the other. 

 The mean deviation was 10' .by 44.0 gm. at — 6°.o C. Using the 

 known values of the constants, the rigidity was calculated to be 



w=i .8X10^ c.g.s. 



The following table gives the value of n determined from differ- 

 ent experiments: 



Temp. (C.) «Xio~^ 



— 9.2 2.1 



-8.6 1.8 



-6.0 1.8 



-5-6 1-7 



