Latitude and Temperature in Barometric Hypsomeiry. 149 



Again, for St. Bernard as the lower and Mont Blanc as the upper 

 station, 



B' 0-56803 A' 7'6 H x 2463, 



b' 0-42429 a -9'1 h x 4787-7; 



which has heen calculated as Ex. 2 above. 



But taking the data from the Annuaire da B. des L., 1865, 

 p. 324, we have for Geneva as the lower and Mont Blanc as the 

 upper station, 



B 729-65 M' 18-6 A' 19*3 H x 408, 

 b 424-05 m -4-2 a —7-6 h r 4815'9. 



That is, the height of Mont Blanc above the sea, when calculated 

 from observations at Geneva, St. Bernard, and the summit, is deter- 

 mined as 4787*7 metres, but when calculated from observations at 

 Geneva and the summit only, is determined as 4815*9 metres, or 28'2 

 metres more. This is striking enough, but it is by no means clear that 

 even the smaller amount may not be too large*. 



Mr. Glaisher's balloon ascents offer a very convenient series of 

 examples on account of the comparative closeness of his observations. 

 I have therefore calculated two, Tables III. and IV., p. 156, which 

 are important from their height or remarkable changes of tempera- 

 ture, first, by determining the height of each station from the lowest 

 (which I call the total method) ; and secondly, by calculating the 

 height of each station from the height of the next lower station 

 (which I call the gradual method). I have added the differences 

 of level between the stations as determined from both methods 

 and the differences between them, which are important for dis- 

 covering how the discrepancies between the two results are produced 

 by temperature. Each station is lettered. Two letters against a 

 number, as a h 5720, show that the height of the station h above 

 the sea is found as 5720 feet, when station a is taken as the lower 

 station with the height assigned to it in the same column. The 

 distance a A is termed an interval. A careful examination of these 

 results will show that the gradual method is probably the most 

 trustworthy. 



* In the Ann. Met. de F. (1. c.) M. Plantamour calculates the height of St. 

 Bernard by Bessel's formula (taking account of the humidity of the atmosphere 

 according to his hypothesis, which is, however, not in accordance with Mr. 

 Glaisher's observations) as 2473 metres. In the Annuaire de la Societe Meieoro- 

 logique de France, 1853, p. 249, M. Plantamour gives the heigbt of the basin 

 of the barometer at the hospice of St. Bernard as 2493 metres, but does 

 not there state how this result was obtained. These heights being respectively 

 10 and 30 metres greater than that calculated by Laplace's formula, would, 

 if adopted as the height of the lower station in the second calculation, give 

 results more nearly in accordance with those in the third calculation. The ob- 

 ject here, however, is to examine the action of Laplace's formula only, and hence 

 the height assumed for St. Bernard must be that due to that formula. But 

 different data give different results for this height. Geneva and St. Bernard 

 are too widely separated horizontally, and have generally too great a differ- 

 ence of temperature, to enable us to calculate the whole height in one section 

 with any degree of confidence, as there are probably many abnormal intermediate 

 changes of temperature which, as will be seen, tend to vitiate the result. Nor 

 can any reliance be placed on adopting the mean barometric pressures and 

 temperatures. If any mean be taken, it must be the mean of many heights 

 separately calculated from their individual data. 



