Narrative of the surveying voyages of his majesty's ships Adventure and Beagle (vol.1): between the years 1826 and 1836 : describing their examination of the southern shores of South America, and the Beagles's circumnavigation of the globe

518 OBSEKVATIONS OF INTENSITY. At Rio tie Janeiro, which was the first station observed at in South America, the cylinder was vibrated in August 1826, Sep- tember 1827, and December 1828 ; in the intervals between these dates are comprised the greater part of the observations on the east side of South America. There is no direct observation at Rio subsequently to December 1828 ; but we are enabled to supply the time of vibration, which would have been observed had the cylin- der been employed at Rio on June 1, 1830, in the following man- ner. We have seen that on the 15th September 1827 the time of vibration was observed at Rio ; on the 18th December following it was observed at Monte Video. These observations give the inten- sity at Monte Video relatively to that at Rio, subject to whatever change of magnetism the cylinder may have undergone in the interval of three months. This comparison was repeated in the following year, on Captain King's return from Monte Video to Rio, the interval being nearly of the same duration, and the order of the experiment reversed, the passage being in this instance from Monte Video to Rio, it having been before from Rio to Monte Video. On the supposition of an uniform, or nearly uniform rate of change in the cylinder, the errors arising therefrom during the two passages would be of opposite kinds, and should compensate each other in a mean of the two comparisons. Calling the force at Rio unity, these comparisons give its value at Monte Video respec- tively as follows, namely, September and December 1827 ... M97-1 October and December 1828 ... 1-207 P^^^" ^'^"^• On the 1st of June 1830, being then on his return from the west coast of South America, and on the eve of sailing for England, Captain King again observed the time of vibration of the cylinder at Monte Video ; whence, through the preceding comparison, we obtain the time of vibration at Rio, which should belong to the same date. We have thus a fourth date at Rio, which, added to those enumerated above, will include the whole of the South Ame- rican stations ; and we have only to distribute in each interval the loss of magnetism which the observations shew to have taken place from one date to the next, in the manner which may appear most suitable. There is no very obvious indication that the loss was other than gradual ; and by considering it uniform in each separate