260 C. Davison—Creeping of Soil-cap by Frost. 
further down the slope. One of the red grains had crept # mm., and 
the other 1 mm. down the slope. The soil particles must therefore 
have descended along a line inclined at an angle of about 17° to the 
normal to the surface of the soil. The normal rise of the surface 
particles was about #5 of the depth of the frozen soil, and their creep 
about +45 of the same depth. 
8. Now, if the creeping of the soileap through the action of frost 
were due to the expansive force of the freezing water urging the 
mass down the slope, it is clear that the downward movement should 
take place during the frost and not during the thaw ; but the experi- 
ments just described show that this is not the case, unless it be along 
the edges of the mass of frozen earth. Again, if this explanation 
were the correct one, the whole frozen layer at any point would be 
‘displaced equally throughout its depth, and there would be a rupture 
of continuity between the frozen and unfrozen parts of the soileap. 
Hence, any line inclined to the surface would, after a succession of 
frosts of different degrees of intensity and duration, exhibit a series 
of breaks forming miniature faults parallel to the surface of the 
ground. And, if the principal frosts of every season penetrated to 
approximately the same depth, then, at that depth, the fault-displace- 
ment would be greatest. It need hardly be said that this inference 
is not supported by Mr. Kerr’s section given in Fig. 1. Both reason- 
ing and experiment are in favour of the view that creeping takes 
place by a normal rise during the frost and a more or less vertical 
descent during the thaw. 
9. It would be interesting to determine, if we could, the length 
of time required for the production of the effects illustrated in Fig. 1, 
but the data are too imperfect to enable us to form more than a very 
rough estimate. Judging from the figure, the creep near the surface 
is on an average about three times the depth of the bends. If the 
amount of creeping of the surface particles be the mean of the 
values given by the experiments recorded above, namely, +7 of the 
depth of frozen soil, and if there be only one important frost and 
thaw eyery winter, then the time required may have been about 531 
years. But the assumptions involved in making this estimate are 
too great to allow us to place any reliance whatever upon its 
accuracy. It is just possible, however, that it may indicate the 
order of magnitude of the real interval of time. 
10. Lastly, the conditions which are most effective in producing 
the creeping here considered are neither those of a temperate climate, 
in which there may be many alternations of frost and thaw in one 
winter, nor those of an arctic or glacial climate; though probably 
nearer those of the latter than of the former. Other conditions being 
the same, the amount of creeping (measured by the product of the 
volume of soil moved into the average displacement down the slope) 
varies as the square of the depth to which the frost penetrates. 
Thus, if in one place, the soil be frozen to a depth of one foot, and, 
in another, to a depth of three feet, the creeping of the surface- 
particles in the latter will be three times as great as in the former, 
and the volume of soil affected also three times as great: hence the 
