of Terrestrial Magnetism. 443 



teorological causes, and the magnetic effect of the sun. The 

 former do not reach their extremes for a considerable time after 

 the corresponding astronomical phases, while the latter have an 

 almost exact coincidence with them. 



We do not pretend, however, that there are not considerable 

 difficulties in the way of this hypothesis ; and although it ex- 

 plains very well certain very singular facts, — as, for example, the 

 interval of six hours between the diurnal maxima and minima, 

 a fact the explanation of which has never, as far as I am aware, 

 been even attempted on any other hypothesis, and which yet is 

 so marked in all the magnetic variations in the mean latitudes; 

 also the singular exception which it suffers at the equator, be- 

 coming simple for the horizontal and for the vertical components, 

 and various other points, — yet we must confess that there are 

 some irregularities which our formula? do not explain. Of this 

 nature is the fact, that at St. Helena, and generally under the 

 equator, the period for the declination of the needle appears to 

 be rather eight hours than twelve, so that it presents sometimes 

 three maxima. Without repeating here what we have said else- 

 where in general terms, viz. that these periods may find their 

 explanation in those terms of the formulae which we have neg- 

 lected*, we may say that this fact may simply depend on the 

 configuration and nature of the ground near to the places of ob- 

 servation. Thus, for example, at St. Helena, an island situated 

 in the midst of the Atlantic, and entirely volcanic, the distribu- 

 tion of magnetism must be very different from what it is at a 

 place in the interior of a continent ; and we know, in fact, that 

 the isogonal lines rapidly change their direction in passing from 

 seas to continents, and this explanation may also apply to equa- 

 torial stations near the coast. We know too how much the 

 vicinity of magnetic bodies may influence the diurnal variation 

 of the needle f. This explanation seems to be confirmed by 

 the fact that the diurnal curves derived from the declination at 

 St. Helena resemble much more closely those of other countries 

 than the annual diurnal mean. In fact, the latter is exclusively 

 due to the horary angle, and therefore more strictly dependent 

 on the distribution of the earth's magnetism round the place of 



* We have in every ease worked out these formula: completely, and there 

 have resulted terms which, developed in series, give sines and cosines of 

 triple arcs ; from this we may deduce the explanation of a period of the third 

 of a day. But as this development may be made in every periodical func- 

 tion, it does not give a legitimate proof of the truth of the explanation, 

 until we have demonstrated theoretically that these terms may have co- 

 efficients of considerable magnitude, a point which I have not yet inves- 

 tigated. 



f See the volume of Arago already cited, where he relates the experi- 

 ments of Barlow on this subject, p. i'J-. 



