MAGNETIC SURVEY OF JAPAN. ^15 



half a uiile further .south. On the euuterii .shore of Chearegdow — as 

 the large island already mentioned is called — lies Oniizutare, looking 

 over against Kurosaki Cape. A Hrtle to the south is Shiinokijinia, 

 that is Shiinoki Island, helping to fill in the s[)ace between Cheareg- 

 dow and Kurosaki. Shiinokijinia is ahout half a inih', long ; and the 

 observations were taken at tlie extreme southwestern and iiorthe:istern 

 ends. Thus Wakwan, Pusan, Kurosaki, Kurosaki (^ij)e, Shiin(jkijima 

 Cape, Shiinokijinia and Oniizutare form a circle of points surrounding 

 Pusan Harbour ; and no two t)f them are further distinit from one 

 another than four (jr five miles. Within this very limited area very 

 strikino- irreo-ularities exist. The declination varies bv nearlv a whole 

 degree ; the dip by fidly the same amount ; the horizontal force by 

 1.5 per cent, of the whole value. Perhaps the most striking irregu- 

 larity of all is that displayed by the value of the dip at Shiinokijiina 

 Cape as compared with the values at Shiinokijinia proper (half a mile 

 U) the south-west) and at Kurosaki Cape (a mile and quarter to the 

 north-north-east). As Mr. Tanakadatehas pointed out, the difference 

 of fully a degree in the dips at the two Capes suggests a kind of horse- 

 shoe magnet witli its poles at these points. The whole series of 

 observations is a very good example of the powerful effect of l<3cal 

 disturbances. It should be said that the wliole district is quite hilly 

 and craggy. 



We shall now follow a very usual custom and grouj) our .stations 

 according to the geological character of the rocks in their neighbour- 

 hood. This can only be done in a very general way, since the geology 

 of Japan is still known but in the broid. There is, of course, no 

 doubt as to the geological characteristics of some stations ; but there 

 are many regarding which we can, by study of the geological map, 

 draw no very sure inferences. The following grouping of stations is, 

 therefore, only a tentative one. Three groups are distinguished. In 



