APPENDIX. 



It has already been faid that the watch was ufed only to 

 prevent an error of whole minutes, in eftimating the time 

 gained by the pendulum in twenty-four hours ; the exacSt 

 period of twenty- four hours being determined by the revo- 

 lution of the fun. 



In order to obtain the acceleration of the pendulum, 

 the original obfervations are transferred from the foregoinor 

 table, to that which follows, for the convenience of ar- 

 ranging them according to the length of the intervals, 

 beginning with thofe of the fhortell duration :. fo that the 

 conclufion from each period becomes a check upon thofe 

 that follow. 



In this \.2^Q the jirfi column refers to the original obfer- 

 vations, from which a conclufion is here to be drawn; 

 thus, in the firft line, we find 27 — 30, by which is meant 

 that a conclufion is to be drawn in this line from obfer- 

 vations 27 and 30, that is, from the acceleration of the 

 pendulum from four hours, forty-fix minutes, ten feconds 

 and a half, to five hours, twenty-four minutes nine feconds 

 in the afternoon, July 17. 



T'hefecond column exprefles the interval of time by the 

 watch, between each pair of obfervations referred to in 

 the firft. 



"The third column fhews how much the pendulum 

 gained on the watch, in each period exprefled in the 

 fecond. 



The fourth column fhews the mean height of the Ther- 

 mometer for each period. 



rhe 



163 



