126 COSMOS. 



At St. Helena the movement of the end of the needle, which 

 points to the nortli, is entirely opposite, in the months from 

 May to September, from the direction which it follows in the 

 analogous hours from October to February, it has been 

 found after five years' horary observations, that during the 

 winter of the southern hemisphere, in the above-named peri- 

 ods of the year, while the sun is in the northern signs, the 

 northern point of the needle has the greatest eastern varia- 

 tion at 7 A.M., from which hour, as in the middle latitudes 

 of Europe and North America, it moves westward till 10 

 A.M., and remains very nearly stationary until 2 P.M. At 

 other parts of the year, on the other hand, namely, from Oc- 

 tober till February (which constitutes the summer of the 

 southern hemisphere, and when the sun is in the southern 

 signs, and therefore nearest to the earth), the greatest west- 

 ern elongation of the needle falls about 8 A.M., showing a 

 movement from west to east until noon, precisely in accord- 

 ance with the type of Hobarton (42^ 53^ S. lat.), and of oth- 

 er districts of the middle parts of th% southern hemisphere. 

 At the time of the equinoxes, or soon afterward, as, for in- 

 stance, in March and April, as well as in September and Oc- 

 tober, the course of the needle fluctuates on individual days, 

 showing periods of transition from one type to another, from 

 that of the northern to that of the southern hemisphere.* 



Singapore lies a little to the north of the geographical 

 equator, between the latter and the magnetic equator, which, 

 according to Elliot, coincides almost exactly with the curve 

 of lowest intensity. According to the observations which 



* Sabine, Observations mads at the Magn. and Meteo7\ Observatory at 

 St. Helena in 1840-1845, vol. i., p. 30 ; and in the Phil. Transact, for 

 1847, pt. i., p. 51-56, pi. iii. The regularity of this opposition in the 

 two divisions of the year, the first occurring between May and Septem- 

 ber (type of the middle latitudes in the northern hemisphere), and the 

 next between October and February (type of the middle latitudes in 

 the southern hemisphere), is graphically and strikingly manifested 

 when we separately compare the form and inflections of the curve of 

 horary variation in the portions of the day intervening between 2 P.M. 

 and 10 A.M., between 10 A.M. and 4 P.M., and between 4 P.M. and 

 2 A.M. Every curve above the line which indicates the mean decli- 

 nation has an almost similar one corresponding to it below it (vol. i., 

 pi. iv,, the curves A A and BB). This opposition is perceptible even 

 in the rvDCturnal periods, and it is still more remarkable that, while the 

 type of St. Helena and of the Cape of Good Hope is found to be that 

 belonging to the noi-thern hemisphere, the same earlier occurrence of 

 the turning hours which is observed in Canada (Toronto) is noticed in 

 the same months at these two southern points. Sabine, Obscj'v. at 

 Hobarton, vol. i., p. xxxvi. 



