82 cosmos. 



of temperature in those parts of the atmosphere which are 

 accessible to us. Neither the principal epochs of diurnal and 

 annual alterations of declination at the different hours of the 

 day and night, nor the periods of the mean intensity of the 

 terrestrial force* coincide with the periods of the maxima 

 and minima of the temperature of the atmosphere, or of the 

 upper crust of the earth. We may remark that the annual 

 alterations were first accurately represented by Sabine from 

 a very large number of observations. The turning points in 

 the most important magnetic phenomena are the solstices 

 and the equinoxes. The epoch at which the intensity of the 

 terrestrial force is the greatest, and that at which the dip- 

 ping-needle most nearly assumes the vertical position in 



* " So far as these four stations (Toronto, Hobarton, St. Helena, 

 and the Cape), so widely separated from each other and so diversely 

 situated, justify a generalization, we may arrive at the conclusion that 

 at the hour of 7 to 8 A.M. the magnetic declination is every where 

 subject to a variation of which the period is a year, and which is every 

 where similar in character and amount, consisting of a movement of 

 the north end of the magnet from east to west between the northern 

 and the southern solstice, and a return from west to east between the 

 southern and the northern solstice, the amplitude being about 5 min- 

 utes of arc. The turning periods of the year are not, as many might 

 be disposed to anticipate, those months in which the temperature at the 

 surface of our planet, or of the subsoil, or of the atmosphere (as far as 

 we possess the means of judging of the temperature of the atmosphere) 

 attains its maximum and minimum. Stations so diversely situated would, 

 indeed, present in these respects thermic conditions of great variety ; 

 whereas uniformity in the epoch of the turning periods is a not less 

 conspicuous feature in the annual variation than similarity of char- 

 acter and numerical value. At all the stations the solstices are the 

 turning periods of the annual variation at the hour of which we are 

 treating. The only periods of the year in which the diurnal or horary 

 variation at that hour does actually disappear are at the equinoxes, 

 when the sun is passing from the one hemisphere to the other, and 

 when the magnetic direction, in the course of its annual variation 

 from east to west, or vice versa, coincides with the direction which is 

 the mean declination of all the months and of all the hours. The 

 annual variation is obviously connected with, and dependent on, the 

 earth's position in its orbit relatively to the sun around which it re- 

 volves ; as the dmmal variation is connected with, and dependent on, 

 the rotation of the earth on its axis, by which each meridian success- 

 ively passes through every angle of inclination to the sun in the round 

 of 24 hours." Sabine, On the Annual and Diurnal Variations, in the 

 second volume of Observations made at the Magnetic and Meteorological 

 Observatory at Toronto, p. xvii.-xx. See also his memoir, On the An- 

 nual Variation of the Magnetic Declination at different periods of the 

 Day, in the Philos. Transact, for 1851, pt. ii., p. 635, and the Intro- 

 duction of his Observations made at the Observatory at Hobarton, vol. i., 

 p. xxxiv.-xxxvi. 



