436 JOSEPH BARRELL 



could rise to indefinite heights. This is because the depth of 

 maximum stress-difference would lie at about one-sixth of the 

 wave-length below the mean level of the surface. With increasing 

 wave-length the height of the waves could accordingly be greater 

 without increasing the stress-difference at the trough-line of the 

 waves. The gradient would, however, have to become more 

 gentle; in other words, the amplitude would have to increase at a 

 lesser rate than the wave-length. If the strength of ice were meas- 

 ured by its rigidity it could stand permanently in masses one-third 

 or one-fourth as steep and high as these theoretic limits for granite 

 mountains, without failure by plastic flow. Yet, on the contrary, 

 the great ice fields spread out by flowage of their bases, although 

 their surfaces possess very gentle gradients. The distinction 

 between strength and rigidity in the movement of glaciers is thus 

 clear. The strength of glaciers is limited by the amount of the 

 stress-differences needed to produce slow movement by recrystalli- 

 zation. 



Johnston and L. H. Adams have applied this theory of yielding, 

 well known as an explanation of glacial motion, to all plastic flow, 

 and argue that even for those substances, such as the metals and 

 rocks, in which cubic compressibility raises the melting-point, 

 shear greatly lowers it for the parts under stress.^ They argue from 

 a physico-chemical basis that the most plausible explanation for 

 flow in metals is that the shearing strain is great enough on individ- 

 ual points to produce a change of phase of individual molecules 

 from solid to liquid, even at ordinary temperatures. 



Apart from theory as to its explanation, the phenomenon of 

 welding of iron shows for high temperatures a low elastic limit and 

 ready passage beyond into plastic flow. For iron and steel, 

 furthermore, the influence of temperature upon the rigidity has 

 been investigated. Pisati gives the following equations in which 

 n is the value of the modulus of rigidity for temperature t.^ For 

 iron — 



Wi= 8ii X io^(i — . 000,206/— . 000,000, 1 9^^-}- . 000,000,001,1/^), 



'"On the Effect of High Pressures on the Physical and Chemical Behavior of 

 Solids," Am. Jour. ScL, XXXV (1914), 205-53. 

 'Smithsonian Physical Tables (1904), p. 76. 



