579 



second, so lliat no place can be assigned to them on the axis of 

 tiie abscissae. When, however, the minimnni-, and the maximum- 



Fig. 1. 



dosages are known, a place maj' be fonnd for them on the axis 

 of the ordinates, and when assnming the law to hold good also for 

 these substances, their hypothetical place on the axis of tlie abscissae 

 may be determined by erecting a perpendicular. We have plotted 

 the graph accordingly and thus given a value for rubidium as well 

 as for caesium. We know then the presumable nnmber of ions that 

 will be emitted under the given premisses per gram and per second. 

 For our pure preparation of caesium it appears to be 55 per gram 

 and per second. With such a small number of ions we can expect 

 a photogra[)hic effect only after 9 years. It is easy to understand, 

 therefore, that up to the present endeavours to produce any effect 

 of caesium upon a sensitive plate, have not been successful. 



The 55 ions per gram and per second that, according to the 

 hypothesis of the corpuscular equivalence, should belong to pure 

 caesium, caimot really belong to the caesium as such, but must be 

 due to the impurity of the commercial preparation, which had been 

 removed from the caesium in the following way: 



Addition of copper sulphate; perfusion of sulphureted hydrogen 

 for \/', hour; after 24 hours removal of the precipitate of copper 



38* 



