Geological Society of London. 565 



and the term " migration " is suggested for the process in which 

 secondary hornblende is formed in plagioclase, as if bastite-substance 

 had been carried in solution along the cleavage-lines of the felspar 

 and by reacting upon its host had produced secondary hornblende. 

 Hornblende so formed is referred to as " emigrated hornblende." 

 Areas of uralitic hornblende generally extinguish under crossed 

 nicols in two portions. On a very favourable section it was possible 

 to determine that the two portions of hornblende are in definite 

 crystallographic orientation — namely, that of twinning on the 

 orthopinacoid, as is common in hornblende-crystals. 



A mineral resembling wohlerite was found to be relatively 

 abundant iu some of the more acid rocks of the Windy Lake eruptive. 



The nickeliferous rocks are cut by younger eruptives — stocks of 

 granite and dykes of olivine-diabase. 



A few pages are devoted to the subject of differentiation, in which 

 it is pointed out that no one of the generally accepted theories is 

 able to account for all the phenomena observed in the differentia- 

 tion of the nickel-bearing rocks of the Sudbury district. Stress 

 has been laid upon the part which gravitation undoubtedly 

 plays in producing heterogeneity in eruptive rocks. All the old 

 theories of differentiation are directed to explain the presence of 

 basic borders on more acid central portions, while they do not 

 account for those cases where the central portions of stocks are 

 basic and the margin acid. They fail also to give any explanation 

 of the commonest case, namely, the eruptive showing little or no 

 differentiation. These different cases are not only explained, but 

 predicted, by the application of the principle of gravitation to 

 slowly-cooling eruptive magmas. 



3. " On the Distribution in Space of the Accessory Shocks of the 

 Great Japanese Earthquake of 1891." By Charles Davison, Sc.D., 

 F.G-.S. 



The object of the author in this paper is to consider the geo- 

 graphical distribution of the numerous shocks which preceded and 

 followed the great earthquake of 1891. A brief summary of 

 Professor (Jmori's work on the distribution of the after-shocks of 

 this earthquake is given, and the difference between his method 

 of treatment and that adopted in the present paper pointed out ; the 

 author furthermore indicates possible sources of error in his maps, 

 and explains how these may be practically neglected. 



In a map of the coast within the Mino-Owari district, the 

 boundary of the area over which the principal shock was felt is 

 enclosed by a line which bifurcates towards the south. The longer 

 axis of this area coincides generally, as shown by Professor Koto, 

 with the direction of a fault-scarp which, however, is only prolonged 

 into the south-eastern fork of the disturbed area. This the author 

 speaks of as the main fault, and he inters the evidence of a secondary 

 iault running along the southern fork. 



In discussing the preparation for the great earthquake, reasons 

 are given lor believing that the distribution of earthquakes iii 



