256 ©. Davison—Creeping of Soil-cap by Frost. 
feet was brought to grade in the process...... The rock is 
gneiss and mica-schist with hornblendic and chloritic strata, 
inclined at a high angle, and decomposed, for the most part, the 
entire depth of the cut [not more than three to four feet in the 
portion sketched], presenting a banded section of variously-coloured 
earths. The most striking and novel peculiarity of this section is 
shown in the sketch (Fig. 1, ante, p. 255), viz. the gradual drawing 
out,—attenuation of these coloured bands, as the parts of them in 
succession were moved down the slope” (p. 358). 
After remarking on the great depth of frozen soil in Canada 
and Labrador during their present winters, and on the still greater 
depth to which it must have penetrated during the Glacial period, 
Mr. Kerr says: ‘The alternate freezing and thawing of the saturated 
mass of decayed rocks . . . . would of necessity produce just the 
movement and settling which are described above. That is, this 
freezing and thawing would give rise to precisely the same move- 
ments of the mass, and of the particles inter se, as are seen to occur 
in the true glacier, differing only in amount. In other words, these 
masses were earth glaciers, and these deposits may be denominated frost 
drift, as distinguished from proper glacial drift” (p. 352). The 
majority of these deposits Mr. Kerr refers to glacial age; but, con- 
sidering the small thickness of that illustrated in Fig. 1, he remarks 
that “the probability is strong that it is of recent (present) origin, 
the existing climate of Philadelphia being equal to the production 
of such effects” (p. 353). 
It is not clear from this description how Mr. Kerr supposes the 
freezing and thawing to have acted in producing this movement of 
the soilcap. My first impression after reading his paper was that 
the water in freezing would, by its expansion, force the mass of 
loose earth bodily down the slope. This may be the case in certain 
parts of the-mass, but that the explanation cannot be generally true 
will, I think, be evident after I have described the way in which 
I believe the movement must as a rule take place, together with 
some of the experiments made to test the theory. 
2. Imagine a layer of damp earth resting on an inclined surface, 
and exposed to the action of frost. The water in the interstitial 
pores will be frozen to a depth depending on the intensity and 
duration of the frost. If the particles of soil be not closely packed, 
the water will, in freezing, expand into the spaces between them, 
and the relative position of the particles may not be altered in 
consequence. But this cannot often be the case, for freshly-turned 
earth is soon rendered close and compact by a few showers of rain. 
The distances between the separate particles will thus as a rule be 
increased when the water between them is frozen. Now, the soil 
being compact and, except near the edges of the mass, continuous, 
the only direction in which expansion can readily take place is out- 
wards and perpendicular to the surface. Every particle of soil in 
the frozen layer will therefore be displaced from its original 
position along the line of the normal (or perpendicular) to the 
surface of the soil; and, if the water be equally diffused through- 
out, the amount of the displacement will be proportional to the 
