210 KNOTT AND TANAKADATE 



tiirbances in the Declination are very marked. Tliere is no evidence 

 whatever of a Declination less than 5° W. until we reach the stations 

 on the North East Coast. That is to say, the isogonic line of 5°, 

 wdiich general considerations would lead us to expect to run right 

 across the region from S. W. to N. E., seems to form an extra distinct 

 contour round the schistose granitic portion on the East. The declina- 

 tions at Hanamaki (Xo. 7) and Miyako (No. 10) are especially large. 

 So great in fact was the value of the declination at Miyako — ^^just ab(uit 

 a whole degree greater than was expected — that at first sight it looked 

 like a misreading of a whole degree on the azimuth circle. Three 

 distinct observations, however, were taken at this station — two by 

 means of the sun and one (three transits in all) by means of Polaris. 

 All the values agree to within 7', so that the large value of the decli- 

 nation is undoubted. This peculiarity may be purely local, or it may 

 be of wider extent. Our stations are not sufficiently mnnerous to 

 enable us to C(mie to any sure conclusion on this point. It should be 

 mentioned that a little to the south of Miyako considerable c[uantities 

 of irori ore are known to exist. 



If we take account only of the 50 selected Stations, the mean 

 probable error of a single observation is ± 7'. Three of the Stations 

 are conspicuous by reason of the largeness of the différences between 

 the observed and calculated values. These are Ishibashi (No. 1 ), 

 Shiogaina ( N"o. 5 ) and llachinohe ( No. 12 ). It will be noticed that 

 the declination at Shiogama is greater than is given l)y the formula, 

 and that the declination at Hachinohe is less. These Stations lie res- 

 pectively at the extreme South and at the extreme North of the region 

 of schists and granites that has just been discussed. Only a number 

 of observations at neiohbourino' localities could, however, determine 

 whether the peculiarities presented by Shiogama and llachinohe 

 belong to a genernl svstem or nre to l)e regarded as purely lornl. 



