1874.] Variation of the Magnetic Declination. 255 



world in modern times have given results so different that we must con- 

 clude either that the magnetic needle obeys different annual laws at each 

 place, or that the differences are due to instrumental errors. The con- 

 sequence has been that, after long, laborious, and expensive researches, it is 

 still a question whether the magnetic needle obeys an annual law or not. 



The results obtained at some observatories have made it very probable 

 that, if an annual law exist, the range of the oscillation must be very 

 small. It is therefore essential, in questioning any series of observa- 

 tions for this law, to be assured that the errors (instrumental or others) 

 are neither considerable nor systematic. 



I have concluded, from several series of observations made with sus- 

 pension-threads bearing unmagnetic or slightly magnetic weights, that 

 the systematic errors due to varying temperature or humidity are very 

 small when the suspension-threads are carefully constructed with fibres 

 from which the original torsion has been removed. Dr. Lloyd has con- 

 cluded that threads with fibres differently twisted may produce com- 

 paratively large annual variations in different directions, according to the 

 direction of the twist. There is little doubt, however, that the greatest 

 errors are due to the unequal stretching and rupture of the different 

 fibres which form the suspension-thread. 



When the instrumental errors may be so considerable compared with 

 the variations to be observed, it cannot be supposed extraordinary that 

 instruments in different places give different results; and it appears 

 essential so to eliminate the sources of error that two instruments in the 

 same place may tell the same story before we attempt to announce the 

 existence of any law. 



If at sea two or more chronometers are necessary in case one may be 

 affected by error, it seems not less necessary in scientific researches 

 requiring continuous observations for years, where errors are so difficult 

 of detection and elimination, that two or more instruments should be 

 observed. These considerations induced me to establish at Trevan- 

 drum two declination-magnetometers of different construction, placed 

 under considerably different atmospheric conditions; and it is to the 

 results of sixteen years' comparative observation from these two instru- 

 ments that I desire to draw the attention of physicists. 



Both instruments had suspension-threads made with the utmost care. 

 One, Dr. Lloyd's instrument, made by Mr. G-rubb, of Dublin, with a 

 magnet weighing nearly a pound, was placed in the large room of the 

 Trevandrum Magnetic Observatory, which was always more or less open 

 to the external air ; and, although covered by a cotton-wadded hood and 

 a series of boxes, it was much more liable to any errors due to atmo- 

 spheric actions than the other. Its chief source of error was, however, 

 connected with small movements of the telescope wire, although. that was 

 made to coincide, at varying intervals of time, with the transit-mark five 

 miles distant. 



