MAGNETIC SUßVEi" OF JAPAN FOR THE EPOCH 189Ô.0. 31 



difficult to decide wliicli course the curves will take, and recourse 

 had to be made to the comiiion sense estimate or guess ivork. 

 To draw those curves with accuracy " over even the smoothest 

 hillside " to use Lord Kelvin's words with regard to the dis- 

 tribution of atmospheric electricity, " would infiniteh' transcend 

 human mathematical power." 



Mr. S. Nakamura and the writer starting separately on 

 different sheets obtained curves agreeing in general appearance, 

 but in particular details they differed widely in some places, the 

 curves in INIaps 1, 2, and 3 drawn on transparent sheets, are 

 a compromise made Ijy the writer. Actual values at each station, 

 corresponding to the nearest mark on the maps, are given 

 along with the curves, so that they can be reconstructed by any 

 one to suit his own view. The controversy which has risen in 

 this respect in the result of previous surveys is thus avoided. Those 

 curves were prepared on a larger scale of linear dimension five 

 times those given in this volume and were reduced by pantograph. 



§ 10. Mean Isomagnetics. 



The mean isomagnetics are re2)resented by empirical fornud^e 

 expressing magnetic elements in terms of longitude and latitude in 

 the usual way. Number of terms to be taken in such expressions 

 depends upon the character of distribution of these elements in the 

 country. By way of trial these elements were calculated for 12 

 points in Japan from the table of magnetic elements for the globe 

 corresponding to 1885.0 as given by Prof. Ad. Schmidt," in 

 which the expansion is carried to seventh harmonics. Table V 

 shows the distribution of these points. 



. * Aus dem Archiv der Deutscheu Seewarte XXI Jahrgang 1898 No. 2, p. til. 



