Prof. J. Thomson on Crystallization and Liquefaction. 399 



free to melt and cause an equal quantity of water to freeze to the 

 other*, it will follow according to the axiom, now supposed to be 

 established, that the application of the stress will cause this action 

 to occur. 



The case of crystals in their solutions might be stated almost in the 

 same words as the case of ice in ice-cold water : but it is to be observed 

 that, in their case, the necessity for the translation of one chemical 

 substance through another (the salt through the dissolving liquid), 

 and not of heat or cold alone, causes a great slowness of the process, 

 as compared with that of the yielding of the ice, in ice-cold water, 

 to applied stresses. 



At an early stage of the considerations which led to the opinions 

 on the influence of stresses on crystallization and liquefaction described 

 in the present paper, the question arose to me : — Is a spiculum or 

 single crystal of ice, which has solidified itself in the interior of water, 

 and is therefore not colder than the water, plastic ? Or would it, 

 when in the water, and attached by one end, as for instance to a crust 

 of ice lining the containing vessel, gradually bend upwards by its own 

 bouyancy in the heavier water ? My idea is that it is not plastic- I 

 cannot conceive of the growth of a crystal proceeding with one con- 

 tinuous or uninterrupted structural arrangement, if during its growth 

 the part already formed undergoes permanent change of form, such 

 as would be due to any plastic or ductile yielding. I think we must 

 suppose the molecules in the interior of one crystal to be so locked 

 into one another, by the forces of crystalline cohesion, that any 

 one of them, or set of them, would experience a difficulty in making 

 a beginning of the change of state from solid to liquid. I have not 

 succeeded even in forming any clear conception of continuous cry- 

 stalline structure admitting of what may be called ductile or malleable 

 bending (that is, bending beyond limits of elasticity such as occurs 

 in lead, copper, tin, and many other metals), and still remaining of 

 the nature of one continuous crystal. What in soft or malleable 

 crystals of copper or other metals, deposited in the electrotype process, 

 may be the nature of the change of molecular arrangement induced 



* The supposition here assumed, however, of there being perfect freedom for 

 either of two pieces of ice, which are immersed in the same water, and are 

 alike free from stresses, to melt, and, by giving out its cold, to cause an equal 

 quantity of water to freeze to the other, will probably not meet with assent at 

 present from all, as it appears to be a prevailing opinion that water and ice in 

 contact are not in a state of perfect indifference as to retaining or interchanging 

 their conditions. It is supposed that ice has a property of tending to solidify 

 water in contact with it, and the more so if there be ice on both sides of the 

 water than if on only one side. Again, it is supposed that ice is essentially colder 

 than water in contact with it, and that the water must continually be giving off 

 heat to the ice. Both these opinions are inconsistent with the supposition here 

 assumed. I conceive, however, that that supposition is amply confirmed by 

 the fact that it was involved essentially throughout the reasoning, by which I 

 was led to conclude that the freezing-point is lowered by increase of pressure, 

 and to calculate the amount of the lowering. That reasoning led to true results, 

 and I believe it could not have done so unless the supposition were true, that 

 when water and ice are present together their freedom to change their state on 

 the slightest addition or abstraction of heat, or on the slightest application of 

 mechanical work tending to the change, is perfect. 



