358 Frederick Guthrie on Salt Solutions 



against supersaturation. The notion that in order to start the 

 crystallization of a supersaturated solution crystals must he 

 present in the air identical in chemical kind with those ready 

 to be formed in the solution, seems to be disproved by the fact 

 i hat supersaturation in regard to ice is of the very commonest 

 occurrence. In many cases a salt solution which should yield 

 ice at a temperature — 1° refuses to do so until — (t° + 5° or 6°) 

 is reached ; then ice-crystals are formed through the mass and 

 the temperature rises. Although we may indeed imagine ice- 

 crystals to be floating in the air about the surface of the cryo- 

 gen, we can scarcely conceive these microscopic crystals to 

 travel unmelted through the surrounding air, which may be at 

 10° or 15° C, so as to reach the liquid. And if it be argued 

 that the origin of the ice-crystals may be at or near the surface 

 of the salt solution in the experimental tube, I reply that crys- 

 tallization from a supersaturated solution occurs both when it is 

 covered with a layer of oil and when only the lower part of 

 such a solution is acted on by the cryogen. 



Method of observation. — Since the separation of ice strengthens 

 the solution, it is necessary to determine the temperature of 

 the very beginning of the ice-formation. The salt solution is 

 cooled in a test-tube until some ice is formed ; this is very 

 nearly completely remelted under constant stirring with the 

 thermometer, and then the tube is plunged momentarily into 

 the cryogen. The minute spicula of ice so formed are again 

 nearly remelted. The mean of four or five readings of the 

 thermometer when the minute quantity of ice begins to increase 

 is taken as the true temperature of ice-formation. In order to 

 save time and start the crystallization of ice in a solution which 

 one has reason to suppose is supersaturated with ice, the ther- 

 mometer-bulb may be plunged into the cryogen and rapidly 

 wiped dry, or into the mercury-cup to be described immedi- 

 ately. It soon becomes clothed with a film of ice-crystals from 

 the moisture of the air. These are indeed almost invisible, and 

 far too small in quantity to alter the strength of the salt solu- 

 tion when brought into contact with it, but amply sufficient to 

 determine the separation of ice if the solution be supersaturated 

 in regard to that body. 



When the tube containing the salt solution is plunged into 

 the cryogen till crystallization begins and then removed, por- 

 tions of the cryogen adhering to the tube may carry the ice- 

 formation too far, and by soiling obscure the tube. It is there- 

 fore found convenient to have standing in the freezing-mixture 

 a short wide-mouthed bottle of mercury, and for the final ob- 

 servations to plunge the tube into this and so avoid soiling. 



The solutions stronger than the cryohydrate were treated in 

 a similar manner. 



For the solubilities of the various salts at 0° I kept the solu- 



