50 Pbikfophkal Ufa of a ctmrmn Watch, 



. heavonl)' bodies taken at their pafl'age over the meredian, and the diAance of places from the 

 difference of the velocity of light and found. A pendulum to fwing feconds has ufually been 

 applied for thefe and fimilar purpofes, and in an obfervatory is found to be very convenient ; 

 but a watch by being more portable is calculated to be more general in its application, and will 

 ineivfure fmaller portions of time than any other inftrument that has been invented* ; befides, 

 it poffeffes this peculiar advantage, th^t in all fituatloips the btats thereof may be counted by 

 the ear, at the fame time that the obje£l of obfervation is viewed by the eye, fo that no lofs is 

 incurred, as mud inevitably happen, when the eye is ufed to view both the objeft and pen- 

 dulum or fecond-iiidex mfuecejfion, though it be ever fo quicic. — But it will be obje£ted here, 

 09 doubt^ that few watches meafure time accurately, and alfo that, from the diiFerent conftruc- 

 tions-of ' watches, the times correfponding to their beats vary in a very confiderable degree. 

 I allow thefe objecSions to be true, and conceive that the reafon may be attributed to them, 

 why the beat of a watch is not generally applied as the meafure of the lowed denomination of 

 •fubdivifions of time : I (hall therefore endeavour in this paper to obviate thefe objedtions, bjr 

 fhewing how any tolerably good watch, whatevetbe itsconftruftion, may be applied with ad- 

 vantage to many philofophical purpofes. 



We muft, in the firft place, confider that the portions of time which I propofe to have 

 mcafured by a watch are y/;ja// portions only, and thofc to be counted not by a fecond-hand, as 

 is the cuftom with medical men, but altogether by the beats ; in which cafe, if the vratch be 

 not liable to lofe or gain time confiderably in a day, the error in the rate of going will be 

 extremely minute in the time correfpondmg to any number of beats that the memory can re- 

 tain;, pr that the purpofes to which I propofe the application to be made will require ; and even 

 if tlie error in the rate of going te confiderable, fo as to amount to many minutes in a day, as 

 it is uniform, it may eafily be allowed for by a correftionf . Hence the firft objeftion, which re- 

 '!■ iates to the error oCcafioned by the rate of going of any Mratch, will conftitute no real obftacle 

 ■'''' to its application in the afcertaining of fmall portions of time, provided a fudden change of 

 'temperature be avoided at the time of ufmg it; for it will be neceffary that the rate of going 

 *- be eftimated when the temperature is the feme, or very nearly the fame, as when the watch is 

 '''^ 'ufed for philofophical purpofes ; fo that if it is ufually worn in the pocke^ it may be held in 

 ^"^ ^e hand to the ear, but if it be hanging in a room or in the open air where the rate of going 

 ' ' ' is afcertained, it muft be hung n?ar the ear, under fimilar circumftances, where any obferva- 

 "" tion is intended to be made by it. 



As to the other objeftion, which applies to the variation in the lengths of the beats of two 



different watches, owing to the difference of their conftru£tions, though they indicate hours 



^ ' imd minutes alike, it maybe removed very readily. All common watches have the fame num- 



♦ The beat of a watch is quicker than that of any other chronometer in general ufe ; but there have been 

 ioftruments made to divide the fecond into lOo partt. One of thefe, made ))y Whitehurft, and regulated by 

 : *fly, repeatedly meafured the time of fall of a leaden bullet (in fome txperimenti which I faw) with »o 

 greater variation than one hundredth part of the fecond. — N. 



t^ If the error were five roinuKi per day, the allowance would be left than one thrce-hunJredtb part. — N. 



ler 



