30 On the Distribution of Cyclonic Precipitation in Japan. 



where r and 6' are the polar coordinates, with O as tlie origin and 

 OA as the prime vector. It wih be seen, that whatever may be 

 the form of F^rd'), the effect of the second term will be to shift the 

 centre of area of the loci loci toward the direction d = - - e. 



From the above, it will be seen that the first and the second 

 factors are essentially not very different, and there would have been 

 no special need at all to distinguish them, if Ave are everywhere 

 concerned with humid land only. In the latter case, the essential 

 factor is only tlie distribution of the isotherms, and the results of 

 the al)Ove considerations in the paragrai)hs under (1) and (2), may 

 be summarized as follows : The maximum precipitation will be 

 expected on the right front quadrant when we look out from the 

 centre of depression toward a direction having the warmer half plane 

 on the right side, provided the isotherms are maintained nearly 

 stationary regardless of the cyclone. The latter condition will be 

 nearly fulfilled when the general temperature distril)ution is de- 

 termined by an extensive " centre of action." 



When we consider the three factors as separate, it will be seen 

 that the second and third generally conspire in winter, but tend to 

 cancel each other in summer, provided the land and water are of 

 sufficient extent. For a narrow strip of land such as Japan, the 

 second factor will play no important part when a cyclone of con- 

 siderable dimensions is concerned, since in such a case the assump- 

 tion does not apply that the land behaves as a reservoir of heat de- 

 termining the stationary isotherms, regardless of the disturbance 

 due to the cyclone itself. Meanwhile the third influence remains, 

 since it will be independent of the latter disturbance. 



Next the first and the third influences conspire when a coast 

 line running more or less in west-easterly direction, face to the 

 water on its southern side, but tend to cancel each other when the 

 water is on the northern side. The resultant effect is well illu- 

 strated by comparing the centre loci diagrams for P and J districts. 

 That the first influence predominates in these cases, is shown by 

 the fact that the district in question lies generally on the eastern 

 side of the point of maximum expectation, though yet the eccentri- 

 city is more decided in the case of the Pacific coast than in the 



