214 Prof. Millikan : New Modification of the Cloud Method 



a sudden movement of 2 mm., and within ten seconds had 

 crept back to its original zero. One degree of change in 

 temperature was found by a separate experiment to produce 

 a deflexion of 8*5 mm. Hence the maximum fall in tem- 

 perature in the expansion as indicated by this couple was 

 about i° C. 



This couple was then replaced by another which was made 

 of 1 mil copper wire and 1 mil iron wire, and had therefore 

 but about one-seventh the thermal capacity of the first. An 

 expansion of 20 cm. of mercury produced in this case a 

 sudden deflexion which corresponded to an apparent instan- 

 taneous fall in temperature of o, 8 C. As before, within 

 9 or 10 seconds the original zero had been altogether 

 regained, and within 6 or 7 seconds it had been regained to 

 within less than half a degree. Both before and after this 

 test the couple was placed inside a two-litre bottle containing 

 water, and an expansion of 20 cm. of mercury produced by 

 turning a stop-cock of about 4 sq. cm. opening, which 

 ^connected this bottle with a 100-litre carboy within which 

 the pressure was maintained at 55 cm. of mercury- In both 

 ■cases the deflexion was about ten times as much as when the 

 couple was between the plates p and n of the apparatus 

 shown in the figure. Within a period of 15 seconds the 

 initial temperature had been regained. 



A more complete study of the relation between the volume 

 and shape of the vessel and the temperature existing at given 

 time intervals after the sudden expansion of a saturated 

 vapour is now in progress. The above experiments, however, 

 are sufficient to demonstrate, at least, that with the sort of fog 

 •chamber here used the temperature existing* midway between 

 the plates six seconds or more after expansion does not differ 

 appreciably from that of the room. Since all measurements 

 made by the method herein described are taken at times 

 which begin between 6 and 15 seconds after an expansion, it 

 is evident that the coefficient of viscosity of the medium at 

 these times must be considered to be that corresponding to 

 the temperature of the room. Furthermore, even when the 

 observations are made by Wilson's method and are begun 

 within two or three seconds of the time of an expansion, there 

 can be little doubt that the mean temperature during the ob- 

 served time of fall is, approximately at least, the temperature 

 of the room. 



Still a third reason for confidence in this conclusion is 

 found in the following observations. In order to control the 

 temperature of the chamber, water-jackets J l5 J 3 were placed 

 about its two branches in the manner indicated in the figure. 



