482 Royal Society .•-— 



always in nearly the same direction. The dip of this resolved por- 

 tion will be about 1 7i°* 



It is also found that a declination peak corresponds to a peak of 

 either force, except in the case of the great disturbance of August to 

 September 1859, during the most violent portion of which a peak of 

 the declination corresponded to a hollow in either force. The length, 

 however, of a declination peak does not bear an invariable ratio to 

 that of a force peak — this ratio varying much from one disturbance 

 to another, but not much from one part to another of the same dis- 

 turbance. In this last case, however, the variation of the ratio, 

 although not great, is yet greater than that of the ratio between the 

 two force peaks ; so that it is somewhat difficult to obtain similar 

 peaks when comparing the decimation curve with that of either 

 force. 



It thus appears that the force which acts upon the magnets does 

 not vary much from one part to another of the same disturbance, 

 and it therefore becomes possible to give the elements of this force, 

 which will thus characterize the disturbance. 



The author then attempts, by means of comparing similar appear- 

 ances, to represent the force at work for each disturbance between the 

 beginning of 1858 to the end of 1860. The great disturbance of 

 August to September 1859 is here remarkable as one in which two 

 independent disturbing forces seem to have acted at once, — one of 

 these being of the normal type, in which all the elements were raised 

 or depressed together, while in the other the declination was raised 

 when both elements of the force were depressed. 



It will be observed that this method of analysis does not completely 

 determine the disturbing force, but merely fixes the line of its 

 resultant action, along which the force itself may be either positive 

 or negative ; or, again, there may be two nearly opposite forces acting 

 against one another, the visible disturbance denoting merely the 

 difference in strength between the two ; and there is some reason to 

 think that this last supposition represents the true state of the 

 case. 



For while the definite relation which exists between the peaks of 

 the two force-components shows that all disturbing forces affect these 

 in nearly the same way, yet sometimes, though very rarely, in the 

 general progress of the curve one of the elements will be above the 

 normal while the other is below it. Now, this may be accounted for 

 in the following manner. Suppose we have a disturbance producing 

 an elevation in the horizontal force represented by +40, and one in 

 the vertical force represented by +20. This will be of the normal 

 type. Suppose now that at the same time we have another force 

 nearly similar, whose action on the two force-elements is represented 

 by —39 — 21. This is also sufficiently near the normal type. The 

 result of these two disturbances superimposed will be + 1 and — 1, 

 showing that the one element is raised above its normal position, 

 while the other is depressed below it. This idea of two opposite 

 forces acting simultaneously in disturbances is that entertained by 

 General Sabine from other considerations. 



