C. Davison—Creeping of Soil-cap by Frost. 209 
of the slab. It was then left some days, and afterwards placed 
outside on the window-sill of an unused room, on the evening of 
January 3, 1889; and the distances of the needle-point from the 
lower edge of the bar and one of the lines perpendicular to it were 
carefully measured with the aid of a pair of bow-compasses.* 
The minimum temperature during the following night was 27° F. 
and the soil was evidently frozen through by the next morning. 
The needle-point had risen 14 mm. in a line almost exactly per- 
pendicular to the surface of the soil; and it remained in this position 
until the night of Jan. 7-8, when a thaw set in, by which the ice in 
the soil was entirely melted in about twelve hours. During this 
interval the position of the index was recorded several times ; and 
the displacements measured, though very small, showed that the 
-needle-point descended in practically a straight line, so as to end 
at the same distance from the lower edge of the bar as at starting, 
but 2 mm. further down the slope. The direction of descent was 
therefore not quite vertical, but approximately bisected the angle 
between the vertical and the normal to the surface of the soil. Also, 
the normal rise of the surface-particles was about 33; of the depth 
of the frozen soil, and their creep downwards about z$z of the 
same depth. 
At first sight, it seems possible that this creep may have been 
only apparent, and due to the weight of the index, which is of course 
much greater than that of an equal volume of soil. But I noticed 
carefully that, after the thaw, there was no gap between the lead 
slab and the surrounding soil; and this could hardly have been the 
case if the movement of the surface-particles had been very different 
from that of the index. The following experiment, however, appears 
to me to remove this objection, if it be one. 
7. Experiment 2.—(Box B). An index, similar to that used in 
the preceding experiment, was placed on the surface of the soil, in 
order to measure its normal rise. Two fine grains, half a milli- 
metre in diameter, of very hard red drawing-chalk were also 
inserted into very small holes made in the soil to receive them. 
One of these was exactly in a line with the edge of a long pin 
(a lady’s hat-pin) fixed at right angles to the bar, and penetrating 
the soil. The other was just underneath the point of a similar pin 
constrained to slide in a straight groove perpendicular to the lower 
edge of the bar. Before the soil was frozen, the point of the pin 
was drawn up a short distance, and brought close down to the soil at 
the end of the frost and again after the thaw. . 
The box was put outside on February 9th, and by the 12th inst. 
the soil was frozen through. The needle point of the index had 
risen normally 32 mm., and the two red grains had also risen 
normally, and probably the same distance. Later on Feb. 12th the 
thaw began, and the interstitial ice was quite melted the next day. 
After the thaw, the needle-point of the index was found at ¢ mm. 
above its original distance from the lower edge of the bar, and 1 mm. 
1 The measurements thus made are probably correct to one-eighth of a millimetre. 
