184 REPORT — 1859. 



An altitude and azimuth instrument by Cary, on a tripod firmly placed, was 

 turned until the division on the glass scale of the magnet corresponding to 

 the magnetic axis coincided with the vertical wire of the telescope, and the 

 azimuth circle was then read. The altitude and azimuth instrument was then 

 turned towards the sun, the time at which his centre crossed the middle wire 

 was found by a chronometer, whose error and rate were known, and the read- 

 ing of the azimuth circle was noted. The latitude and longitude of the place 

 and the time of observation being supposed to be known, the astronomical 

 azimuth of the sun at the moment of observation is given by a well-known 

 formula. The difference between the readings of the azimuth circle for the 

 sun's centre and the magnetic axis enables us then at once to determine the 

 magnetic declination. The altitude and azimuth instrument was often allowed 

 to remain in position for some hours, during which time occasional readings 

 of the magnetic axis were taken, and at the same time the azimuth of some 

 fixed object was read occasionally, in order to see if the tripod-stand had 

 shifted. The silk thread by which the magnet was suspended was carefully 

 kept as free from torsion as possible, and the amount of torsion was moreover 

 examined and eliminated from time to time. The amounts occasionally found 

 to be present were always of such trifling consequence that no correction on 

 their account has been applied to the observations. 



The chronometer used was a pocket instrument by Arnold and Dent, No. 

 5155. Its rate appears to have been somewhat irregular, owing probably 

 to the motion it received in travelling. At almost every station, however, 

 altitudes of the sun were taken, by which the correct time, and consequently 

 also the error of the instrument.^were determined. 



For one or two stations no altitudes were taken, and consequently no chro- 

 nometer error determined. Here the following method was pursued. A 

 correction was applied to the last determined chronometer error, depending 

 upon the time that had elapsed since, and on the most probable chronometer 

 rate. The chronometer error so corrected was used in the azimuth observation 

 of the station whose altitude observation was wanting. 



During the years 1858 and 1859 self-recording magnetometers were con- 

 tinuously in operation at Kew Observatory, by means of which the magnetic 

 declination at any moment might be determined. The traces furnished by 

 the declination magnetometer have been reduced at General Sabine's office 

 at Woolwich, and the hourly means of the declination, free from disturbance, 

 for every month of both those years have been determined. This enables us 

 to say with great accuracy what correction ought to be applied to a declina- 

 tion observation taken at Kew at any hour of any month of any year near 

 this date, in order to reduce it to the 31st of December (mean of all the 

 hours) of the same year in which it was taken. 



Presuming that these corrections are also applicable to Scotland, and to 

 1857, they have been used in reducing the observations of declination taken 

 in that year to the epoch of January 1, 1858. In reducing those taken in 

 1858 a somewhat different method has been pursued. During that year the 

 Kew magnetometers, as already mentioned, were in operation. Suppose that 

 we take the mean of all the hours of January 1858, freed from disturbance, as 



fore the second year's survey its position had changed to 49 - (an inconsiderable difference). 

 At Thurso, on August 23, 1858, the magnet was observed erect and inverted, and 49 - was 

 still found to denote the axis. At Lerwick, the next station after Thurso, the glass scale 

 was wiped, after which the axis appears to have changed ; but, as in every observation after- 

 wards the magnet was viewed both erect and inverted, this shifting of the axis was of little 

 consequence. 



