T. Terada; 



From this table, we may easily trace the cun'es of equal ex- 

 pectation corresponding to each given position of tlie centre, or the 

 locus of the positions of the centre bringing ecpial expectation for 

 each given district. For the sake of simplicity call tlie former 

 curves " the isohyets " for the given position of the centre, though 

 here, instead of the amount of the precipitation, the expectation is 

 meant ; the latter sets of curves may be called " the centre loci " 

 for the given district. To smootli down local irregularities, the 

 average was taken uf the expectations of four adjoining 2. 5° meshes 

 and the mean value was attril^uted to the centre of the four meshes. 

 Those sets of four meshes of which any one had the numl.)er of oc- 

 currences of depression, i.e. iYless than 5, were excluded. In this 

 way, the uncertainty due to the defect of data in tlie marginal re- 

 gions was avoided. From these averaged expectations we con- 

 structed the two sets of diagi'ams mentioned above. 



Figs. 1 to 20 show the isohyets for different positions of the 

 depression wliicli is marked with ¥: in each diagram. Figs. 21 to 

 08 gives the centre loci for each district, the middle point of which 

 is marked with •. 



