280 Proceedings of the Royal Society of Edinburgh. [Sess. 
(1) may be connected with the fact that between these two concentrations 
we probably have a\ 2 = const. 
Any attempt at explanation of the underlying causes of these very 
interesting phenomena is probably best postponed until further data 
become available. 
It remains to express my best thanks to Professor Clifton for his 
kindness in granting me facilities for carrying out the above research in 
the Clarendon Laboratory, Oxford. 
( Issued separately October 3 , 1907 .) 
