280 On Irregularities olserved in 



ice saws, &c.) which were carried on the after part of the deck 

 for convenience in use, than it would have been, had it been raised 

 higher. This was proved by placing a compass on a plank ele- 

 vated for experiment in front of the companion, to the same height 

 as in the Isabella, namely, nine or ten feet. The points of no 

 error were found, in this position, to be about north and south, 

 and the amount of error at eight points, nearly twenty degrees ; 

 the same as in the Isabella : the greatest error at the same time, 

 by the Alexander's standard compass, viz. the one nearer the 

 decks, being 8° 20' at N.N.E. The dip was 84° 09'. 



The propriety of Captain Flinders's recommendation, to deter- 

 mine the points of no error in a fixed compass by actual observa^- 

 tion in every ship as soon as the distribution of iron is completed, 

 may therefore be considered as confirmed by the observations in 

 the Isabella and Alexander ; whilst his rule of proportion may 

 receive a verbal alteration to render it more suitable for general 

 application : so corrected, it would be as follows — the expres- 

 sions substituted being marked in italics, and the original words 

 entered in the margin. 



" The error produced in any direction of the 



East or west. ship's head, will be to the error at the point of' 



the greatest irregularity, as the sine of the 



Magnetic meridian, angle between the ship's head and the points of 



no error to the sine of eight points, or radius." 



Thirdly; Capt. Flinders's experience in the Investigator showed 

 that the maximum of error in the same compass would be dif- 

 ferent in different parts of the world, although the use of the com- 

 pass was confined to one particular spot in the ship, and every 

 precaution taken to avoid an interference with the distribution 

 of the ship's iron. 



It is worthv of remark, that by multiplving observations and 

 by comparing the series one with another, he was thus practically 

 led to trace a connection between the amount of the errors and 

 the dip of the needle ; a knowledge of the fact preceding, in his 

 mind, any theoretical suggestion that such might be the case. 



It does not appear indeed that the principal cause of this con- 

 nection was even subsequently known to him ; he perceived that 

 the influence of lo'*al attraction on the compass needle increased 

 as the dip became greater. He endeavoured to account for this 

 circumstance, on a supposition that all iron might receive an al- 

 solute increase in the intensity of its attractive power by ap- 

 proaching the magnetic pole. 



The increase, however, which was the subject of his observation, 

 was a relative one, being in comjjarison to the directive power 

 of magnetisili. A diminution in the latter would therefore pro- 

 duce the effect equally with an absolute augmentation in the for-' 



mer; 



