169 



ment is like a longitude determined by the mere difference of 

 time : both, executed with good instruments, and under fa- 

 vorable circumstances, are susceptible of great exactness ; 

 but when the meteorological variations, or the rate of the 

 chronometer, are not regular and uniform, it is impossible to 

 fix the limit of the errors, as we may do with success in dis- 

 cussing a geometrical operation, or the result of a series of 

 lunar distances. 



After having excluded the barometric measurement of 

 Borda, two others remain, which inspire great confidence, 

 but of which one appears to be somewhat too little, and the 

 other too great. We have already remarked, that their mean 

 result does not differ 0 003 from the geometrical measure- 

 ment -j and we shall not give a preference to the barometric 

 observations of Lamanon over those of Mr. Cordier, because 

 we think we have proved, that the result even of the trigo- 

 nometrical measurement may well be a few toises too 

 small, and Mr. Cordier made his excursion in very fine and 

 settled weather. This gentleman thinks, that his measure- 

 ment must have given a result near the truth, on account of 

 the numberless precautions which he took to avoid errors *. 

 The observation was made in the morning j and it is known, 

 that at this time of the day Laplace's formula makes the 

 heights too little, because his coefficient was deduced from 

 observations made at noon ; but on the other hand, Mr. Ra- 

 mond has rendered it probable, that the coefficient appro- 

 priate for our northern countries must undergo a slight 

 diminution to adapt it to the measurement of the heights 

 comprised between the tropics, or near the limits of the torrid 

 zone f. A compensation therefore took place : and this 

 compensation was not disturbed by the effects of the diurnal 

 variation of the barometer. I insist on this latter circum- 



* Ramond, p. 5 and 26. 

 t Ibid, p. 97. 



