produced by the mutual Reaction of Solid Substances. 295 



That the cold produced when ice and a salt are mixed is due 

 to rapid liquefaction of the ice is plain enough; but I have 

 seen no attempt made to explain the cause of the liquefaction, 

 until Ordway last year announced his theory of the " diffusion 

 of solids "in an address * before the American Association 

 for the Advancement of Science. 



We know that the molecules of a body are in a state of 

 constant oscillation, and that if a salt solution be placed in 

 contact with pure water, diffusion takes place until the mole- 

 cules of salt are equally distributed throughout the mass. 



So, too, when the solid is placed in water, solution follows, 

 or, in other words, diffusion. Now, when a salt and water, 

 both in the solid form, are in contact, there is probably the 

 same tendency to interpenetration. But a mixture of water 

 and salt molecules cannot remain in the solid form except at 

 a low temperature, and the rigidity of the solid state is over- 

 come, because oscillations of the water and the salt molecules 

 cooperate to produce a greater motion. 



Graham found that although sodium chloride is not at all 

 deliquescent, yet the saturated solution has a great affinity for 

 water. Therefore when the smallest quantity of the salt is 

 once in solution the first step is taken, and the melting of the 

 ice continues rapidly. If this is the true explanation of the 

 action of sodium chloride on ice, the problem is solved. 



When salts capable of metathesis are used, this physical 

 phenomenon is complicated by chemical reaction. Liquefac- 

 tion probably results when CaCl 2 . 6 H 2 and Ca(N0 3 ) 2 . 4H 2 

 are mixed, and in similar cases, because the crystallizing-point 

 of these two bodies together is lower than for each alone — just 

 as the freezing-point of salt water is lower than that of fresh 

 water, and as the fusing-point of an alloy is sometimes below 

 that of either of its constituents. 



Calorimeter. — For further experiments in which the reduc- 

 tion of temperature might be measured with some degree of 

 accuracy, it was desirable to secure a closed space in which 

 radiation and convection should be reduced to a minimum, 

 and the heat of the surroundings should be constant. A calo- 

 rimeter was therefore constructed somewhat like that used by 

 Berthelot in some of his investigations. 



It consists of a covered circular tank of fourteen-ounce 

 tinned copper, of about twelve gallons capacity, placed in a 

 much larger wooden case, the space between the walls of the 

 tank and case being filled with loose cotton. 



The upper surface of the tank has four wells, each to re- 

 ceive a cylindrical vessel of polished gennan silver resting on 

 * Proceedings Amer. Assoc. Adv. Science, vol. xxix. p. 293. 



