52 EIGHTH REPORT— 1838. 



tions with a needle so circumstanced can furnish even an ap- 

 proximation to the absohite dip." 



I may add to the preceding remarks, that the discordances 

 thus noticed far exceeded the limit of either diurnal or irregular 

 fluctuations of the dip in England, as far at least as these phae- 

 nomena have hitherto been the subject of observation. 



An attentive consideration of the various sources of error to 

 which dip observations might be liable, — of those which were 

 already guarded against, and of those which still remained un- 

 provided for, — induced the belief, that a considerable part at least 

 of the discrepancies in question, and of similar discordances ex- 

 perienced elsewhere, were occasioned by the axle, on which the 

 needle rests on the agate planes, not being perfectly cylindrical. 

 Careful observers on the continent had already noticed' defects of 

 workmanship in this respect ; and had been led thereby to have 

 needles made, in which the axle, instead of being permanently 

 fixed to the needle, was secured in its place merely by strong 

 friction, and could be taken out, turned a portion of a circle on 

 its own centre of rotation, and replaced; thus enabling the 

 points of the circumference of the axle in contact with the sup- 

 porting planes to be varied in successive trials. At Captain 

 Ross's desire, Mr. Robinson undertook to have four needles of 

 this description made, for one of which Mr. Frodsham, whose 

 chronometers are so well known for their excellence, undertook 

 to make the axle. On these needles being completed, they were 

 tried each in four different positions of the axle, — that is to say, 

 the axle being secured, an observation of the dip was made in 

 the usual manner, and with the usual reversals : — the axle was 

 then removed, turned on its own centre a portion of a circle, 

 replaced, and the dip again observed : — in like manner, a third 

 and fourth change was made in the position of the axle, and the 

 dip observed at each. The process thus described was twice 

 repeated with each needle. Of the four, Mr. Frodsham's axle 

 proved the best ; but the trial clearly manifested in all the im- 

 perfection which had been apprehended. The results with the 

 needle furnished with Mr. Frodsham's axle are given in the 

 subjoined table, where that needle is designated as No. 1. 



With this experience Mr. Robinson undertook to replace the 

 axles of the other three needles with three which should be the 

 workmanship of his own hands. On these being tried, the dis- 

 crepancies of each in the four positions were less than of any of 

 the four axles in the former trial, but still amounted to several 

 minutes. The results of the best of Mr. Robinson's axles have 

 been selected for illustration, and are those of No. 2. in the sub- 

 joined table, 



i 



