124 cosmos. 



needle during the years 1794 and 1795 in Fort Marlborough, 

 on the southwestern coast of Sumatra, as well as at St. 

 Helena.* The results which were then obtained drew the 

 attention of physicists to the great decrease in the quantity 

 of the daily alterations of variation in the lower latitudes. 

 The elongation scarcely amounted to three or four minutes. 

 A more comprehensive and a deeper insight into this phe- 

 nomenon was obtained through the scientific expeditions of 

 Freycinet and Duperrey, but the erection of magnetic sta- 

 tions at three important points of the southern magnetic 

 hemisphere — at Hobarton in Van Diemen's Land, at St. 

 Helena, and at the Cape of Good Hope (where for the last 

 ten years horary observations have been carried on for the 

 registration of the alterations of the three elements of terres- 

 trial magnetism in accordance with one uniform method) — 

 afforded us the first general and systematic results. In 

 the middle latitudes of the southern magnetic hemisphere the 

 needle moves in a totally opposite direction from that which 

 it follows in the northern ; for while in the south the needle 

 that is pointed southward turns from east to west between 

 morning and noon, the northern point of the needle exhibits 

 a direction from west to east. 



Sabine, to whom we are indebted for an elaborate revision 

 of all these variations, has arranged the horary observations 

 that were earned on for five years at Hobarton (42° 53' S. 

 lat., variation 9° 57' east) and Toronto (43° 39' N. lat., va- 

 riation 1° 33' west), so that we can draw a distinction be- 

 tween the periods from October to February, and from April 



* Phil Transact, for 1795, p. 340-349, for 1798, p. 397. The re- 

 sult which Macdonald himself draws from his observations at Fort 

 Marlborough (situated above the town of Bencoolen, in Sumatra, 3° 

 47' S. lat.), and according to which the eastern elongation was on the 

 increase from 7 A.M. to 5 P.M., does not appear to me to be entirely 

 justified. No regular observation was made between noon and 3, 4, 

 or 5 P.M. ; and it seems probable, from some scattered observations 

 made at different times from the normal hours, that the turning hours 

 between the eastern and western elongation fall as early as 2 P.M., 

 precisely the same as at Hobarton. We are in possession of declina- 

 tion observations made by Macdonald during 23 months (from June, 

 1794, to June, 1796), and from these I perceive that the eastern vari- 

 ation increases at all times of the year between 7h. 30m. A.M. till 

 noon, the needle moving steadily from west to east during that period. 

 There is here no trace of the type of the northern hemisphere (Toronto), 

 which was observable at Singapore from May till September ; and yet 

 Port Marlborough lies in almost the same meridian, although to the 

 south of the geographical equator, and only 5° 4' distant from Singa- 

 pore. 



