413 



Respecting the percentages of oxygen found by the analyses 

 I have very little to say. They exhibit the usual variations: 

 21.015 to 20.92. An isolated observation of 20.84 is probably 

 due to some very local cause or perhaps to an analytical error. 

 Taking no notice of this one the average of the determinations 

 comes out as 20.960, or slightly higher than that accepted for 

 Europe, 20.93. Regnadlts analyses indicate that the percentage 

 in the Tropics is a little lower than in Europe. 



The results of the COg -determinations are very remarkable 

 and unexpected. Whereas everywhere else, in temperate Europe, 

 in the Tropics and in the Southern Hemisphere, the percentage 

 of carbonic acid is about 0.03 and varies from 0.02 to 0.04 at 

 the utmost, I have found percentages up to 0.07 and variations 

 from 0.025 to this figure. 



I have, of course, tested these surprising results in every 

 possible way and 1 must confess that I have tried again and again 

 to explain them away as errors. But there is no such possibility. 



The samples of air were sometimes taken directly into the 

 analysis-apparatus, but in most cases they were taken into short 

 test-tubes and analysed about half an hour later. The corks 

 of the test-tubes were soaked in paraffine-wax. No detectable 



