Observations upon the Dip of the Magnetic Needle. 85 



The needle originally furnished is cylindrical, with a steel axis, 

 and with this a kw observations were attempted in 1838, at Nor- 

 folk and vicinity, and Madeira and Bahia, but without the means 

 of reversing its poles ; and as a decision of the Navy Commission- 

 ers, in May, 1838, deprived me in common with other officers of 

 the same class, of the quarters which we had previously occupied, 

 and which Congress have since restored, and thereby exposed my 

 instruments to the hazards of a steerage or midshipman's mess, 

 and consequent damage, I did not deem the previous results com- 

 parable with observations made after I had procured a magnet, 

 and therefore rejected them altogether. At Rio Janeiro I procu- 

 red two additional needles, with brass axes and steel pivots, and 

 adjusted them by the method described in this Journal, Vol. xxxvir, 

 p. 277, and subsequently added the adjusting screws there sug- 

 gested. 



No. 1, on the original needle, was distrusted after a few of the 

 first trials, and latterly has been rejected altogether. Nos. 2 and 

 3 are flat ; No. 3 being broadest in the plane of motion, and No. 2 

 in that perpendicular to it. The indications of No. 2 are found 

 in practice to accord best with each other, although No. 3 ap- 

 peared to have been best made ; a circumstance probably due to 

 the difficulty of so reversing the poles of a magnet, in the ordi- 

 nary way, that the resultant of the forces, after each operation, 

 may continue constantly parallel to the plane perpendicular to 

 the vertical, in which it was situated previously. Otherwise a 

 new source of error is introduced, which it is not easy to elimi- 

 nate, except by multiplying observations, and one which it is to 

 be feared is not always sufficiently regarded. It seems therefore, 

 that in a needle for reversion the mass of magnetic matter should 

 be, as nearly as may be convenient, in a plane passing through 

 its axis and extremities. By this means, though we cannot de- 

 stroy them entirely, we may reduce within narrow limits the 

 errors resulting from differences in the direction of the planes of 

 the magnetic resultants, such as are likely to be the consequence 

 of a want of homogeneity in the metal, or any diversity or care^ 

 lessness in the mode of magnetizing. 



Montevideo, 2d to 4th March, 1839, mean of series with Nos. 

 1 and 2, upon the top of Mrs. Jane Martin's house, near the Ca- 

 thedral and public square or plaza, . . . 33° 36'-5 

 Without the walls, near Mr. Fides Baracca, about one mile 

 northwest of the Cathedral — ^mean of Nos. 1, 2 and 3. 



