138 EXPLORATION OF THE UPPER ATMOSPHERE 



Comparing these results with those obtained at other points under the 

 auspices of the Weather Bureau in the summer of 1898 we find a close agree- 

 ment. Practically all of the Nassau experiments were conducted in the after- 

 noon. Taking from Bulletin F, of the Weather Bureau, the results of the 

 Washington, D. C, afternoon observations and the mean of the afternoon 

 observations at 17 localities in the United States, and placing them alongside 

 the Nassau observations, the rate of decrease is seen to be almost identical. 



COMPARATIVE RATE OF DECREASE IN TEMPERATURE PER 1000 FEET. 



The rate of decrease in temperature computed from averages for each 

 100 feet is also shown in Fig. G (a). 



. DECREASE IN PRESSURE. 



As the altitude of the meteorograph is determined trigonometrically we 



may compute the average rate of decrease in pressure during four flights as 



shown in the following table, which also shows the number of observations 



employed and the average temperature of the air in each column of 1000 feet : 



RATE OF DECREASE IN PRESSURE. 



Altitude Number of Average Fall of 



observations. temperature. 0. 1 inch. 



Feet. o Per feet. 



0-1000 16 80.4 95 



1000-2000 25 78.3 103 



2000-3000 26 69.8 109 



3000-4000 7 64.6 105 



Means. 73.3 103 



In Fig. 6 (&) the rate of decrease in pressure is shown graphically. The 

 plotted observations for which the curve is drawn are averages for each suc- 

 cessive 100 feet. The total number of observations employed in the construc- 

 tion of the curve is about 75. 



RELATIVE HUMIDITY. 



Observations of relative humidity at the surface were obtained by means 

 of a Richard hygrograph, checked by direct determinations of humidity by 

 means of a sling psychrometer. The readings derived from the kite hygro- 



