METEOROLOGY AND DEEP TEMPERATURES 459 



meteorological station, and showed that de Saussure had a very 

 sound knowledge of what should be done at such an observatory. 

 In those days there were no self-recording instruments, and de 

 Saussure did his best to secure readings at uniform intervals 

 throughout the whole period, arranging for simultaneous observa- 

 tions to be made at Chamonix (3445 feet) and Geneva (1230 feet). 

 A full transcript of the whole of the observations from de Saussure 's 

 original copy was published by his grandson in 1891 in the Memoir es 

 of the Geneva Physical and Natural History Society on the 

 occasion of its centenary. Only a summary of the results appears 

 in the Voyages dans les Alpes. Observations of a very accurate 

 mercurial barometer which required much care were made as far 

 as possible every two hours from 4 A.M. to midnight, with the 

 definite object of ascertaining the daily range in order to find the 

 best hour of the day for the barometric determination of height. 

 From various causes, more often affecting the low-level observa- 

 tions than those on the mountain, a detailed comparison could 

 only be made for the period of twelve hours from 8 A.M. to 8 P.M., 

 but these gave results of great interest and value. It is certainly 

 remarkable that observations extending over only half the day and 

 continued for a single fortnight, should have revealed the order of 

 diurnal range which has been confirmed by years of continuous 

 observation at high and low level observatories a hundred years 

 after his time. 



He showed clearly that the hour of the two daily maxima and 

 minima of barometric pressure differed with altitude. Hence 

 he deduced the practical conclusion that observations of the 

 barometer for measuring heights are best made about noon, when 

 both high and low levels have nearly their mean daily pressure. 

 In the morning and evening the range at high and low levels is at 

 opposite extremes, hence at those times the risk of error in height 

 determinations is at its maximum. 



Temperature observations were made also at two -hourly 

 intervals from 4 A.M. to midnight and compared with those at 

 lower stations. These showed a diminution of temperature on 

 the average of 1 Reaumur for 100 toises, or 1 Fahrenheit for 

 282 feet, an extremely close approximation to the value now 

 generally accepted. 



Recognising that there was a greater range of temperature 

 between day and night and between summer and winter near sea- 

 level than at higher altitudes, he calculated that at heights of 

 between 7000 and 8000 toises (say 40,000 feet) the temperature of 

 the air would be uniform at all hours and at all seasons in other 

 words, that the air at such a height would neither be warmed nor 



