A CONTINUOUS RECORD OF ATMOSPHERIC NUCLEATION. 

 Second Series. — 110 = 327000. White (Electric) Light. 



85 



' Alternative, 22200, 2800, o. 



Fourth Series.' — 110 = 546000. White Light. 



2 Cycles nearly absent. Third series omitted. 

 ' Alternative, 25300, 10400, 4900, o. 



The green coronas reproduce the orders of nucleation ah-eady attributed 

 to them, showing ^ = 90,000, 110,000, 90,000, 90,000, a favorable result in view 

 of the large differences of subsidence encountered. 



27. Smaller and larger pressure differences. — To further elucidate the 

 effect of the subsidence error and to exhibit the sequence of coronas more 

 closely, experiments were now made with lower and higher exhaustions than 

 the above. In view of the smaller steps of pressure, it was also thought that a 

 more definite location of the cusps of the cycles would be possible. This, how- 

 ever, was not the case, for if the interval between observations is too small, the 

 coronas are never free from distortion. Table 14 in particular, in which the 

 interval is but one minute, gives evidence of this, but it is even present in 

 the succeeding tables where the intervals are two minutes or more. 



To reduce the data, the values of m must first be computed. If the latent 



