21 6 Observations on the Admiralty 



trasting the errors which they now had with those which 

 respectively belonged to them previous to their being sent 

 away, (the proportional rates for the interval having been applied 

 to them,) this gold chronometer which had kept a more uniform 

 rate than any whilst quiescent, was so much affected by the 

 journey, that a deviation from its rate amounting to more than 

 a second was the result, whilst the others may be considered as 

 having suftered no sensible inconvenience, for a tenth of a second 

 was the maximum of alteration observed in either. Finding this 

 occurrence where it was least likely to have been anticipated, I 

 suspected that perhaps the error might have been my own, 

 originating in inaccurate comparison ; to put this matter to the 

 test, all the same chronometers were transmitted, as also 

 was the silver one on the day following, and the same precau- 

 tions being taken prior to their dispatch, and subsequent to 

 their return, the same results relative to the gold one were ob- 

 tained ; whilst the silver one, when the proportional rate for the 

 intervening time had been applied, gave precisely the observed 

 error. This was no accidental circumstance, the experiments, 

 somewhat modified, were several times repeated, and uniformly 

 with similar results. Observing the valuable property pos- 

 sessed by the silver one, I returned it to its maker, requesting 

 him to lessen its rate, which was cumbersome ; this being ef- 

 fected, I frequently carried it about in my pocket, together with 

 the gold one, and whilst I could place no reliance upon the lat- 

 ter, the former scarcely ever led me into an error of the time ex- 

 ceeding one or two tenths of a second, notwithstanding it some- 

 times met with rather rough usage; indeed, this almost despised 

 chronometer was one of the best I ever saw. That it was differ- 

 ence of position which occasioned the error found in the gold 

 one, was inferred from a knowledge that it would bear very 

 well considerable alterations of temperature. A chronometer, 

 intended for the pocket, should be rigorously indeed examined, 

 as to the eflect produced upon it by alterations of position, 

 seeing that the mere reclining of its wearer on a couch, may 

 lead to considerable alterations of rate ; indeed, from many op- 

 portunities of judging, I am inclined to believe, that it is more 

 difficult to procure a good pocket, than a good box, chrono- 

 meter. 



I have here only alluded to some of the instances which have 

 occurred under my own observation, where chronometers, good 

 in one respect were bad in another. Others might have been 

 adduced, but those already mentioned, will, I hope, suffice. 

 They are assertions, it is true, the particulars of which are not 

 brought forward ; but the like are so frequently occurring, that 

 he, indeed, can have had no experience who has not found them 

 practically correct. To make a chronometer keep a good rate, 

 when exposed to no change of position except what it must ne- 



