39^ NOTES OF A NATURALIST. 



My own impression, as the result of such study as I have been 

 able to give to the subject, is that, in the present state of our 

 knowledge, the reduction of barometric observations for the 

 height of mountains made by day, and in summer, in temperate 

 latitudes, may best be effected by the formula proposed by 

 M. de St. Robert ; while for observations made at other seasons, 

 and in the tropics, I should prefer the formula proposed by 

 Mr. Riihlmann. 



Before closing these remarks, I may refer to an ingenious 

 suggestion made by M. de St. Robert in a paper published in the 

 journal Les Mo fides in Paris, in 1864, the substance of which is 

 to be found in the Atti delF Academia delle Scienze di Torino 

 for 1866, p. 193. Impressed with the difficulty of approximating 

 in practice to a correct knowledge of the distribution of tempera- 

 ture in the air between the summit of a mountain and a lower 

 station, the author sought to escape from it by seeking a phe- 

 nomenon, susceptible of observation, which should give a direct 

 measure of the mean density of the air in the space between the 

 two stations. He pointed out that the velocity of sound supplies 

 such a measure, and that, given the barometric pressures at the 

 higher and lower stations, the angle of elevation of the former, 

 measured by a theodolite and corrected for refraction, and the 

 exact time required for sound to traverse the interval between 

 them, the height is given with a near approximation to accuracy 

 by a simple formula. The error arising from air currents, which 

 increase or diminish the velocity of transmission, would be 

 readily eliminated by discharging a fire-arm simultaneously at 

 both stations, observing the interval between the light reaching 

 the eye and the report becoming audible, and taking the mean 

 of the intervals observed at both stations. 



M. de St. Robert does not disguise the practical difficulty of 

 measuring the time interval with the requisite accuracy, but 

 he thinks that it may be obtained within a fifth of a second. 

 The error in the result is inversely proportionate to the time 

 required to traverse the distance, and where the stations are as 

 distant as is compatible with the sound being audible, its amount 

 for an error of a fifth of a second is inconsiderable. 



This suggestion has not received the attention which it seems 

 to deserve. It possesses the advantage that the observations 

 may readily be repeated with little trouble or cost, and that the 



