TIIR STIlUC'l'intR OF TTIK T^IIff^LETrs. 1;;^7 



wholly evaporate their water, shoiiUl be more stable than other saline nuclei. The 

 results show emphatically that this is not the case. 



Nuclei generated from hygroscopic bodies in the above tables, and their rate 

 of evanescence, i.e., their absorption velocity, h, is not exceptional and is no greater 

 in the first case or less in the second than that of saline bodies in general. Hence, 

 as the nucleus is necessarily a solution in case of the former solutes (CaClg, etc.) it 

 is reasonable to suppose it always to be. It has been sliowu that the pressure 

 decrement dp < 2 cm. of mercury (my apparatus in its present form does not allow 

 me to go below this, though the limit is much lower) is more than sufficient to 

 precipitate the nuclei produced by shaking. For such sinall decrements the 

 equation .29 sp/p = tf0/(273 + 0), where 6 is the temperatui'e iu degi-ees C, may be 

 assumed. Now the decrement of vapor pressure, (5;r, coi-responding to 66, is at 208 

 about d'7t = .llSd ; whence Sti = .25 cm. for the observed excessive 6 p. In other 

 words, the vapor depression of a few millimeters is certainly much more than is 

 required to stop the evaporation of the nucleus, so that if this depression is to be 

 due to the solute, the solution need not even be very concentrated. In case of 

 HgSO^ at 20° the nucleus would hold more than 75 % of water, and at lower 

 temperatures much greater dilution would suffice. 



49. Temperature and harometric pressure. Conclusion. — It is obvious from 

 the last paragraph that the temperature and pressure of the medium must play a 

 very important role iu relation to the constants of the nuclei obtained. In the 

 present research I have made no attempt to determine this, not only because the 

 vapor pressure relations of concentrated solutions must be more fully investigated, 

 but because the method adopted above of producing the nuclei in question, is 

 too crude for the purpose. It will in the first place be necessary to find a limiting 

 value of the nuclei producible by agitation per cubic centim., or at least to make 

 this nucleation a definite function of controllable conditions. The method which 

 I have had under way (though the results cannot be included in the pi-esent re- 

 port) consists in producing comminution by the impact on each other of two 

 independent very fine jets of the same solution under lai-ge hydraulic pressure. 

 In this way not only is the nucleating mechanism increased iu efficiency but the 

 availability of apparatus wholly of glass is at hand. It follows, then, that the 

 opportunities for retaining the solvents absolutely pure are much enhanced, }.ar- 

 ticularly in case of the hydrocarbon solvents. Finally, the coirelative electrical 

 behavior of nuclei is, under these conditions, more definitely open to investigation. 



In the present volume, however, direct evidence for the diffusion of nuclei is 

 the final issue. 



