346 On Mr. Bonvy castle's Disserlut'/on on (he 



sort of micrometer. But it must be observed, that in attempting 

 such a construction, great nicety is requisite, not only in cuttinjij 

 the wedges so that the refraction in each shall take place at right 

 angles to the axis, but also in cementing them together, so that 

 the axes of the two wedges shall be at right angles to each other. 

 Audit may further be remarked, that even then, unless the pen- 

 cil of light pass truly in the common plane of refraction of the 

 wedges, four images will be formed, so as to destroy the effect 

 of the combination. 



LIV. On iV/r. Bonnycastke's Dissertal'ton on the Influence of 

 Masses of' Iron on the Mariners' Compass published in our 

 55th Fbhune. 



OoME time ago we received from a Correspondent, N., the 

 first of the subjoined communications, with an intimation tliat 

 the Editor might either communicate the article to Mr. Bonny- 

 castle, or publish it in the Philosophical Magazine, as he n)ight 

 think would prove most agreeable to that gentleman, to whom 

 no disrespect was intended by the author. The Editor in con- 

 sequence sent Mr. N.'s letter to Mr. Bonnycnstle, who has since 

 returned it with the letter which we have subjoined to that from 

 Mr. N.* 



To Mr. Tilloch. 

 Sir, — In reading Mr. Bonnycastlc's Dissertation on the " In- 

 fluence of Masses of Iro'.i on the Mariners' Compass," 1 have 

 met with a part which carries with it considerable difficulty. At 

 page 453 of the 55th volume of your Philosophical Magazine 

 he says, 



" Were the needle acted on by no other attraction than that 

 of the sphere, its position would be such that 



tan. S' = 2 cos. <p ; 

 where I' in this case is the dip. This formula follows imme- 

 diately from the equations (1) and (2) ; for from them we have 

 2. cos. 9 : sin <{: : ; 1 : cos 8' 

 cos S' = | tan. tp 

 tan 8' = 2 cot.cp." 

 Now I do not mean to dwell upon the enunciation tan 8'= 

 2 cos $ differing from the result tari o' = 2 cot <J., for that is the 

 consequence evidently of a press error; but I own that I neither 



* This communication should have appeared in our last Number, Inrt was 

 somehow mislaid by the Printer. 



see. 



