172 KNOTT AND TANAKADATE 



Aluiost all the tr^nelling by water accoiDpliyhed by the Northern 

 Party occurred during six days of their stay in ^ ezo — otherwise, 

 with a few unimportant exceptions, theii- mute hiy wholly overhuid. 

 On tlie contrary, the Sourhei'U Party accomplished over all uearlv a 

 montli's truvelliny bv water. 



Section II. 



The Northern Party, as already mentioned, were provided with 

 the Portable Magnetometer (Elliott Protliers, No. (11) and Dipi)ing 

 Needle of the well-known Kew Pattern (Parrow A: Co. London, No. 

 24). In addition to these was the indispensable Chronometer (Negus 

 Sidereal, No. 1669). 



The chronometer was systematically <hecked by Sextant Obser- 

 vations, usually of the sun, sometiines of a [)lanet or star. When 

 circumstances were favourable, equal altitude obser\ations were taken. 

 Generally, however, this method was impracticable, since the plan of 

 survey decided upon prevented tlie Party sojourning more than 16 or 

 17 hours at most of the stations. The chief S(jurce of error in 

 determining the time by single altitude observations lay in the some- 

 what uncertain values of the latitudes and longitudes of many of the 

 stations. In-equently the only choice was to estimate these co- 

 ordinates from the authoritati\e maps pre]>ared by the Meteorological 

 and Military depîu'tments. For a few important stations (Hakodate, 

 Niigata, etc.) recent determinations of latitude and longitude were 

 a\ailablc; in other cases Inö's observations were utilised. The 

 clock-error was necessary for the determination of the declination, 

 and the rate for the determination of the horizontal intensity. Any 

 jjossiljlc eiTor in the latter can atfect the value of the horizontal 

 intensity by, at the most, uiie in the 5th signiticant tigurc. As the 

 declination was, in the great majority of cases, determined b\- 



