French National InslUiUe. 357 



and make hci2;hts to be considered as too great : descend- 

 ing winds necessarily produce a contrary etiect. The i^io- 

 iTient wiien the equilibrium of the atmosphere is not dis- 

 turbed bv either of these causes n)ust therefore be chosen^ 

 and that moment is the middle of the day. But M. lia- 

 mond observed -also, that the descending winds prevail 

 oftene^• than others ; and he concludes that in general the 

 mean rcsidts of observations must give heights too small. 



To make a proper choice of the moment is not all ; no 

 less care and atttntion is rcquiftite in the choice of the sta- 

 tions : simultaneous observations, some made in \\\t place 

 the height of which is required, and others in a fixed place 

 the height of which above the level of the sea i« perfectly 

 known, arc necessary. Those also w"ho wish to verify a 

 formula must have the same knowledge of tlie h.eight of the 

 mountain to which the barometer is carried ; and that no 

 pbjection may be made to the conclusion, it is necessary that 

 the two stations should be sufficiently near, and that nothing 

 interrupts the communication, so that the atmospheric va- 

 riations which arise in the one may also take place in the 

 other. M. Rnmoixd found all these advantages united in the 

 Peak of Bigorrc, and the town of Tarbes, where M. Dangos, 

 a celebrated astronomer, was pleased ^ to take upon himsclt" 

 the corresponding observations. 



It was by these means, and with these attentions, that 

 M. Ramond found the correction of the co-efileient of La- 

 place ; after uhich he applied the formula thus corrected, 

 in conjunction with several other known formula^, to cal- 

 culate the aerostatic ascent of M. Gay-Lus53c, who rose to 

 the greatest height ever attained by man, snicc it surpasses 

 that of all the mountains on the earth. 



He applies also all these formulee to the observations 

 communicated to the class bv i\I. Humboldt, and which 

 were made on the highest mountains of Peru, and particu- 

 larly Chimborago, several hundreds of metres above the 

 point at v.hich Condaminc, the most intrepid of our acadc- 

 jnici^ns, was obliged to i^top. 



it results from all these formulae, that the formula of La- 

 place forms a pretty exact mean between alll the clher for- 

 mulae ; that it gives errors always very small, sometimes 

 more sometimes less ; and that the sum of thei-e errors di- 

 vided by the nu.mbcr of observations scarcelv indicates 

 l-]00dth as the ulterior correction of the co-efficient de- 

 termined by M. Ramond. 



This memoir is terminated bv an appendix, in which are 

 given models of all the calculation?, tables for shortening 



the 



