TERRESTRIAL MAGNETIC INTENSITY. 29 



have generally found not to differ by much more than one minute from the mean, 

 when the observations have been made in favourable circumstances, the instru- 

 ment being in perfect adjustment. 



58. The method of making the dip observations has always been by a com- 

 plete series of Eight Observations ; four with the magnetism in each direction. 

 The reversal of the poles has never been omitted. One of the needles (marked 

 A. 1) gave a difference of about a degree when the magnetism was changed, in- 

 dicating a displacement of the centre of gravity by far two great, and consequently 

 introducing an error depending on the intensity of magnetization. Since 1838 

 this error has been reduced to less than a half. 



59. With the needle A. 2, on which I place most confidence, I have generally 

 found the difference of readings after successive displacements of the needle, and 

 allowing it to come to rest, so insignificant, that I have very generally omitted 

 this process, unless some discrepancy has led me to suspect an error, when I 

 have repeated it over and over until the true result was clearly apparent. I am 

 aware that this abbreviation of a tedious process will appear to many persons 

 exceptionable. I have made repeated. experiments in both ways, and with the 

 assistance of another observer, our readings being separately recorded, and I am 

 persuaded that whilst the condition of the axis remains nearly perfect, this mos^t 

 harassing operation may be greatly abridged. 



60. The INTENSITY NEEDLES employed in 1837, were the same as those which 

 I used in 1832, and the methods of reduction employed are identical with those 

 described in my former paper, art. 11, &c.* I need not here repeat them. Find- 

 ing that Cylinder No. 1. still retains very nearly its magnetic constancy, so as 

 to render any correction for epoch almost unnecessary, I have confined my deduc- 

 tions to observations made with it ; and since Paris did not enter into the circuit 

 of stations in 1837, I adopted, as fundamental, the relative horizontal intensity 

 at Edinburgh, determined in 1833 and 1835, and since fully confirmed, viz. 0.840, 

 that at Paris being 1.000. Consequently its time of vibration at Paris is con- 

 stantly reckoned 247 S .70 as before (art. 21).f 



61. The journey I performed in 1837, was not undertaken for the purpose of 

 making magnetic observations. They are, therefore, neither numerous nor regu- 



* Ed. Trans, vol. xiv. p. 5. 



t By art. 18, we found for Cylinder No. 1, at Edinburgh, 



Log. Time Log. Ratio 



h 300 Vibrations. Annual Change. 



1829, July 9. 11 2.90765\ .00043 



1832, June 2. 11 2.90890<^ 99956 



1833, May 7. 5 2.90849? 

 1835, May 4. 1 2.90915</ 



To these we may now add, 1837, Apr. 27. 1 2.90900/ 



1838, May 10. 1 2.90970/ 



