T. Terada : 29 



side. OG shows the direction of the barometric gradient which 

 makes an angle 6' wâth OA. Here the essential difference of 

 land and sea is assumed to consist in the difference of the 

 coefficients of friction, n and n' respectively; the other thermal 

 behaviours are supposed to be equal on both sides. It has been 

 shown in the above cited paper that the intensity of the ascend- 

 ing current due to the discontinuity of the horizontal flux across 

 AB is given by 



o = D cos(^' + £) 



where D cos £=-|(cos5^-cos5^0. tg^=7' 



D sill £ = |(sinÀ?^ — sin^^^''), tgç^' = ~^- 



If il'=Q,V, ^' = 42', then D= -0.3255, ^--Vè\ 



Since D is usually negative, the ascending current at will be 



maximum when 6'=7T-e, and zero when ^'=— ( ^ + ^) ^^ 2"""^' 

 For ö'= -('.^ + e) to -^)-^» the vertical current will be directed 



downwards; hence the expectation of precipitation at will be 

 none for these directions if there is no general ascending current 

 proper to the depression. In the case wlien O lies within the 

 inner region of a cyclone, the ascending current at this point 

 proper to the cyclone will be enhanced when the centre lies on the 

 B side of PQ, but weakened wdien it hes on tlie opposite side. If 

 there exists no such hydrodynamical influence, the trace of the 

 cyclonic centre bringing equal expectation to O, will evidently be 

 a circle with O as the centre. In the special case, when the 

 gradient is proportional to the distance from the centre, or G=br, 

 R will be independent of r, as may be seen from (3) since here the 

 temperature is assumed constant. In general cases, however 

 R^R' will be a function of r and 6' as may be seen from (2). 

 Denote this by 



Taking the hydrodynamical influence in account, the centre loci 

 will be given by 



F{rd') + Dco^id' + = const 



