270 THE ROYAL SOCIETY OF CANADA 



is far below the pressure of the crystal itself, while a scratched crystal 

 effloresced at once. The affinity between the water molecules and 

 the other atoms must be of considerable magnitude, and doubtless 

 increases as the size of the particle decreases. 



My own experiments on the size of particles necessary to induce 

 crystallization may be briefly stated, since better results may be 

 obtained in the future. The determinations were with sodium sul- 

 phate and salol. 



Sodium sulphate. 35 grams of the anhydrous salts were dis- 

 solved in 100 grams of water. The solubility diagram shows that at 

 33° anhydrous salt separates. Below 18° the stable modification is 

 the heptahydrate, and this separates out at 6° or 7°. When a solution 

 containing crystals of the heptahydrate is inoculated with a crystal of 

 the decahydrate, the clear heptahydrate crystals almost at once 

 appear to effloresce; they turn, apparently, to the decahydrate and to 

 the anhydrous salt. After a few months the change to the deca is 

 complete, except for a few points of anhydrous salt. 



These solutions may have air drawn through them for days 

 without separation taking place if the air be sterilized. In general, 

 the air in my laboratory contained one crystal germ in about 30 or 

 40 cc. If the air be drawn quickly through a tube bent at right angles, 

 the small particles are usually retained at the bend, and crystallization 

 only takes place when the liquid is forced up the tube through which 

 the air passes. With a straight leading in tube, crystallization occurs 

 quicklv. 



Size of Crystals. 



Glass tubes, an inch and a half in diam.eter, and from 30 to 40 

 centimeters long were carefully cleaned, and in them to the depth of 

 5 centimeters the supersaturated sodium sulphate solution was placed. 

 These were plugged with cotton wool and carefully "sterilized." 

 After standing for two weeks, a large, wide- mouth bottle from which 

 the bottom had been cut, was fastened over the tube, covered with a 

 sheet of paper, and the whole fixed as nearly as possible perpendicular 

 to the earth's surface. The cotton wool was then withdrawn by 

 means of an attached wire, and the apparatus left until crystallization 

 was noticed. From the length of tube, the velocity of the falling 

 particle was obtained; and then by the aid of Stokes' Law, the size 

 of the particle was calculated. Millikan's and Cunningham's results 

 show the failure of Stokes' Law for particle small compared with the 

 mean free path of gaseous molecules; and my results, for several 

 reasons, I regard as little better than guesses. I have, however, 

 (obtained by this method numbers running from 10'^^ to 10"^* gram. 



