388 Royal Society : — 



Now in this case it may be objected that a crystal of the sodic 

 sulphate hydrate was derived from one of the following sources : — 

 (1) from the air, (2) from the oil, (3) from the cork, or (4) from 

 the side of the flask. (1) It could not be derived from the air, 

 either of the laboratory or of the garden, because the solution 

 remained liquid long after the flask had been corked. (2) It could 

 not have been derived from the oil, because this was dispersed 

 through the solution in myriads of globules, without any nuclear 

 action, and the flask was left to repose for an hour after the oil had 

 been so dispersed. (3) Nor could any nucleus have been derived 

 from the cork, because the solution never touched it. Nor could 

 a minute speck of the sodic sulphate hydrate have fallen from the 

 cork; for the latter had been put into hot water, out of which it was 

 taken the moment it was put into the flask. It could not have been 

 derived from any of the hot water from the cork streaming down 

 the side of the flask to the solution, because sodic sulphate in 

 solution is in the non-nuclear anhydrous state, and also because 

 two hours had elapsed between the corking of the flask and the 

 solidification of the solution. (4) A crystal could not have been 

 derived from the walls of the flask, because the solution had been 

 briskly boiled in it, so that steam escaped with considerable force 

 from the neck after the small beaker had been put on. Hence 

 I am compelled to fall back on my former statement * — namely, 

 that the oil acted as a nucleus by the flattening of one or more of 

 the globules against the wall of the flask into the form of film. 



A globular flask, containing the same solution and the same oil, 

 was corked, and the oil dispersed in globules. It was left some 

 hours, with occasional shaking, so as to rattle the solution against 

 the side. The rattling motion was purposely less energetic than 

 in the former case, to avoid splashing against the cork. This 

 solution crystallized after it had been left a short time at rest ; and 

 it did so in large flat crystals, described in my second paper, very 

 different from the minute radial action noticed when crystallization 

 sets in from a point t. 



A similar flask, containing the same solution, was treated with 

 oil of bergamot, a drop of which was allowed to trickle down the 

 side of the flask. As soon as the oil touched the solution, crystal- 

 lization set in, the radiant-point coinciding with such point of 

 contact, and the whole surface was covered with fine lines diverging 

 from this point alone. 



In another case a drop of the oil of rosemary was deposited on 

 the centre of the solution. On shaking the flask, the solution 

 assumed the solid state. 



Two flasks, containing the same solution, crystallized under the 

 influence of oil of lavender. 



Five flasks, with the same solution, crystallized under the action 

 of sperm-oil. 



Experiment 2. Sodic sulphate — 2 salt, 1 water. In three flasks 



* Phil. Trans. 1871, p. 52 et se%. t Ibid. p. 54. 



