deposition of solid particles. upon them from the surrounding fluid, that would not have been so deposited without the presence of the previous solid portions; a fact sufficiently proved by the growth of fine crystals of ice in ice-cold water. This effect was admirably shown in Mr. Harrison's freezing apparatus, where beautiful thin crystals of ice, six, eight, and ten inches long, would form in the surrounding fluid; and these crystals, which could not be colder than the surrounding fluid, exhibited the phenomena of reg elation when purposely brought in contact with each other.
The next point may be considered as an assumption: it is that many particles in a given state exert a greater sum of their peculiar cohesive force upon a given particle of the like substance in another state than few can do; and that as a consequence a water particle with ice on one side and water on the other, is not so apt to become solid as with ice on both sides; also that a particle of ice at the surface of a mass in water is not so apt to remain ice as when, being within the mass, there is ice on all sides, temperature remaining the same. If that be admitted, then regelation is sufficiently accounted for. Difference of temperature above or below that of the changing points of water is not alone sufficient to cause change of state, the change being independent of temperature throughout a large range. At such times the particles appear to be governed by cohesion. Cohesion resolves itself into the force exerted on one particle by its neighbours, and this force seems to me to be sufficient, under the circumstances, to account for reg elation.
Supposing this to be the true view of the state of things, then a particle of ice within ice can exist at a temperature higher than a like particle of ice on its surface in contact with water; and though it does not appear at present how a higher temperature could be communicated to the interior of a mass of freezing ice than that existing over its surface, still there may be principles of action in radiation, and even in conduction and liquefaction, producing that effect. Assuming, however, that a piece of freezing ice is in such a state, then, if it were to be pulverized, it ought to produce a mixed mass of ice and water colder than the ice was before. Such seems to be the result in one of Prof. Forbes's experiments, in which ice rapidly pounded showed a temperature of 0°.3 Fahr. below the tem-
