402 C. Barus — Possibility of a Colloidal State of Gases. 



permanently with the same nucleus. One may note that the 

 smallest nuclei occur in liquids of greatest, the largest nuclei 

 for liquids of least specific inductive capacity, an inference 

 already drawn for ionizing solvents in a different connection 

 by others (cf. Nernst's Theoretische Chemie, p. 365). 



Finally, the order of condensation here implied should be 

 noticed : the electron by its mere presence condenses the 

 nucleus or molecular cluster, the latter being always so small 

 an aggregation of molecules as to remain optically quite inap- 

 preciable. The nucleus (for Kelvin's thermodynamic reasons) 

 condenses the visible water globules of the coronas seen on 

 exhaustion. The uniformity of nuclei obtained by shaking 

 liquids may also be thus accounted for, supposing that the 

 available electric charges are produced by friction. Their per- 

 sistence in the presence of electrolytes (HC1, etc.) follows 

 more naturally than by the concentration hypothesis adopted 

 provisionally, elsewhere. An interesting feature is the side 

 light thrown on the nature of cohesion. In the nucleus the 

 electron is saturated wholly or partially by a definite average 

 number of molecules, small in electrolyzing and large in non- 

 electrolyzing liquids. However the nucleus may be produced, 

 whether from phosphorus, the flame, etc., from a charged 

 point, by the X-rays, etc., or from a comminuted electrolyte as 

 in shaken nuclei,* it is conceived to be a molecular cluster held 

 together by one or more electrons. The cluster owes its dif- 

 fusion velocity to molecular bombardment in the usual way. 

 This velocity will, therefore, decrease as the conditionsf favor- 

 able to unilateral bombardment decrease, or as the nucleus 

 increases in size from molecular dimensions, indefinitely. 



It seems to me that with these experiments one is approach- 

 ing the true relations of the nucleus and the ion. Whether 

 these relations will continue to hold for liquid electrolytes I 

 hesitate to affirm : but if a definite ion must travel in a defi- 

 nite solvent within a definite nucleus, then the present theory 

 predicts Kohlrausch's law and other allied facts which it is 

 better to assert in connection w T ith quantitative results. 



Brown University, Providence, E. I. 



* Science, xv, p. 426, 1902. 



f See my "Experiments with Ionized Air," Smithsonian Contributions, 

 Washington, 1901. 



