TRANSACTIONS OF THE SECTIONS. 45 



tage in its favour that the accuracy of the observation does not depend upon a 

 imifonn rate of variation of the resistance. The heat generated in the coils by the 

 passage of the electric currents employed, atfecting the two coils equally, is also 

 completely compensated in using eq^ual coUs. 



The late extensive conilagrations of hemp and other warehouses have suggested 

 to me the idea of applying this method for detecting in such stores any spontaneous 

 generation of heat. It also appears to me applicable in meteorological and other 

 scientific observations where uiaxunum and minimum thermometers are at present 

 used. 



On the Effect jproducecl on the Dtviation of the Compass hy the Length and 

 Arramjement of the Compass Needles; and on a New Mode of Correcting 

 the Quadrantal Deviation. By Aechibald Smith, M.A., F.B.S. ; and F. J. 

 Evans, B.N., Supenntetident of the Compass Department of Her Majesty's 

 Navy. 



This was the substance of a paper lately read to the Royal Society, and about to 

 appear in the forthcoming volume of the Philosophical Transactions. The follow- 

 ing is a summary of the results obtained. 



In correcting the de\-iations of a ship's compass in the usual way by magnets 

 and soft iron, if it is necessary to bring the correctors so ne^s the compass, and if 

 the needle is of such a length that its length bears a considerable proportion to 

 the distance of the correctors, an error is introduced which cannot be corrected in 

 the usual way. When caused by magnets, tliis eiTor is sextantal ; when by soft 

 ii'on, it is octantal. Mr. Evans, however, obseiTed that this error did not arise 

 when, instead of a single needle compass, an Acbniralty standard compass was used. 

 In this compass, instead of one needle there are four, arranged two and two at 

 angles of 15° and 45° on each side of the central line — an arrangement adopted 

 long ago in order to prevent the tcabhUng motion which a single needle card baa 

 when distm-bed. And. on submitting the matter to calculation, it appeared that the 

 error in question was wholly corrected when, instead of one needle in the central 

 line, there were two needles each at angles of 30° on each side of the central line, 

 and four needles placed as in the Admnalty compass ; the term involving the error 



having as a factor cos 3a in the first case, and cos — -^ — in the second, where a 



and /3 are the distances of the needles from the central line. It is therefore re- 

 commended strongly that all corrected compass cards should be constructed in 

 this way. 



The second part of the communication was a new mode of correcting the qua- 

 drantal deviation. In all ships, with a very few exceptional cases, this error is posi- 

 tive. It may be corrected by cjiiuders of soft iron placed on each side of the com- 

 pass; but when it is large, there are great practical difiiculties in making the 

 correction. It had been long ago observed by the late Capt. Johnson, R.N., that 

 when two compasses are arranged as in the double binacle compass, they produce 

 on each other a considerable deviation, being in fact a negative quadrantal deviation ; 

 and accordino-ly this arrangement had been prohibited in the Navy. It, however, 

 occurred to Mr. Evans to apply this arrangement to correct the usual positive 

 quadrantal deviation, and one pair of the compasses of H.M.S. Warrior are cor- 

 rected in this manner. In this case likewise the needles should be arranged in the 

 way above described. 



BemarTcs on H.M.S. Warrior's Compasses. By P. J. Evans, B.N. 



It may be considered interesting to the Meeting, as supplementary to the paper 

 read by Mr. Archibald Smith, to receive a brief notice on the magnetism of the 

 first of the great iron wai'-ships of the day, the ' Warrior,' and of the disposition 

 of her compasses. 



There is but little novelty in the arrangement of those on the upper deck, 

 excepting that it has been deemed desu-able to fiunish two standard compasses, 

 from the unavoidable proximity of the after one to a new featui-e which from the 



