56 ~ NOTE ON REGELATION. 
cause of its buoyancy assumed a position of stability. The threads 
were about 1} inch long, so that a piece of ice, when depressed 
sideways and then left to itself, rose in the water as far as it could, 
and into its stable position, with considerable force. When, also, a 
piece was turned round on its loop asa vertical axis, the torsion force 
tended to make it return in the reverse direction. 
Two similar blocks of ice were placed in the water with their 
‘opposed faces about two inches apart; they could be moved into 
any desired position by the use of slender rods of wood, without 
any change of temperature in the water. If brought near to each 
other and then left unrestrained, they separated, returning to their 
first position with considerable force. If brought into the slightest . 
contact, regelation ensued, the blocks adhered, and remained ad- 
herent notwithstanding the force tending to pull them apart. They 
would continue thus, even for twenty-four hours or more, until they 
were purposely separated, and would appear (by many trials) to 
have the adhesion increased at the points where they first touched, 
though at other parts of the contiguous surfaces a feeble thawing 
and dissecting action went on. In this case, except for the first 
moment and in a very minute degree, there was no pressure either 
from capillary action or any other cause. On the contrary, a tensile 
force of considerable amount was tending all the time to separate the 
pieces of ice at their points of adhesion ; where still, I believe, the 
adhesion went on increasing—a belief that will be fully confirmed 
hereafter. 
Being desirous of knowing whether anything like soft adhesion 
occurred, such as would allow slow change of position without sepa- 
ration during the action of the tensile force, I made the following 
arrangements. The blocks of ice being moored by the threads 
fastened to the lowest corners, stood in the water with one of the 
diagonals of the large surfaces vertical ; before the faces were brought 
into contact, each block was rotated 45° about a horizontal axis, in 
opposite directions, so that when put together, they made a compound 
block, with horizontal upper edges, each half of which tended to be 
twisted upon, and torn from the other. Yet by placing indicators in 
holes previously made in the edges of the ice, I could not find that 
there was the slightest motion of the blocks in relation to each other 
in the thirty-six hours during which the experiment was continued. 
This result, as far as it goes, is against the necessity of pressure to 
