LIEUT.-COLONEL SABINE ON TERRESTRIAL MAGNETISM. 15 



No. 8. is the only needle which presents a difference from the other needles exceed- 

 ing in value a five-hundredth part of the time of vibration. If therefore the magnet- 

 ism of Nos. 5. 7. 9. 1 1. 12. and 13. suffered any change in the interval comprehended 

 by the comparison, the alteration must have taken place to an equivalent amount in 

 each of the needles : a coincidence the less probable, because they had all been pre- 

 viously exposed to greater extremes of natural temperature than were experienced at 

 the stations visited between Panama and San Bias. 



The difference which No. 8. presents from the mean of the other needles is equivalent 

 to P-7 in 480 seconds, its time of vibration, or to a proportional loss of magnetism. 

 It does happen occasionally that results with the same needle corrected for tempera- 

 ture, will differ from each other to this amount, in cases when the subsequent return 

 to the original time of vibration manifests that the magnetism of the needle has un- 

 dergone no change, or at least no permanent change : but as the difference exceeds 

 the probable error of observation, as will be presently shown, and as moreover nearly 

 the same difference appears at all the subsequent stations, when the results with No. 8. 

 are compared with those of the other needles, I have regarded it as an indication of 

 an actual loss of magnetism sustained by No. 8., at some time between the observa- 

 tions at Panama in March 1839, and those at Mazatlan in November of the same year, 

 rendering that needle less fit than the others for intermediate deductions : and I have 

 allowed for this loss at all the stations subsequently to San Bias, by deducting -00312 

 from the logarithm of the square of its time of vibration. 



The observations at Mazatlan with No. 7? foi" the comparison of its time of vibra- 

 tion with that of the other needles, were made on the 29th of November. On the 

 30th of November and 2nd of December, this needle was employed in experiments to 

 ascertain the effect on its time of vibration of differences of temperature, by vibrating 

 it in air of the natural temperature, and in air heated by means of boiling water. No 

 memorandum has accompanied the observations of any accident having occurred, 

 either in putting the needle away after the conclusion of the observations of the 29th, 

 or before the commencement of those of the 30th, but a comparison of the results on 

 the three days manifests that the magnetism of the needle sustained an alteration in 

 that interval : 



November 29. Corrected time of vibration 550'9 seconds 

 November 30. Corrected time of vibration 556-7 seconds 

 December 2. Corrected time of vibration 555-9 seconds. 



The observations at the next station, San Bias, confirm this direct evidence of a 

 change, as is seen in the following statement, which shows the quotients of No. 7. at 

 Mazatlan and San Bias compared with the mean quotients of the other needles. 



1839. 



Mazatlan < «,^ * \ U * 



I Nov. 30 and Dec. 2 



' San Bias Dec. 6 and Dec. 19 



