Lurts & Buake—The Carbonic Anhydride of the Atmosphere. 207 
the number whose trend for particular periods is either upward or 
downward, a seasonal effect may possibly be revealed. The results 
of such an analysis are as follows :— 
Number of Curves trending— 
Period. LEE Showing 
Upwards. | Downwards. 
January—February, .  . 10 9 p 
February—March, . ee 10 10 P 
March—April, : : P 9 12 Increase. 
April—May, . : : ¢ 9 12 Increase. 
May—June, . : : : 13 8 Decrease. 
June—July, . : , : 14 9 Decrease. 
July—August, é 5 ! 12 9 Decrease. 
August—September, F . 6 11 Increase. 
September—October, . : 12 7 Decrease. 
October—November, : : 10 ll Increase. 
November—December, . : 9 12 Increase. 
December—January, 6 14 Increase. 
Viewing the matter broadly, we think that the indications 
(masked as they no doubt frequently are by the interference of 
other factors) are in favour of seasonal variations, and that the 
increase caused by production and evolution of carbonic anhy- 
dride from the soil during the early spring receives a well-marked 
check in the summer months, to be followed in the latter period 
of the year by an increase when vegetable activity has ceased. 
The figures given by Petermann and Graftiau in their very 
careful work on the air at Gembloux in Belgium are in harmony 
with such variations, viz. :— 
Spring. Summer. Autumn. Winter. 
2°958 2919 2°927 2:958 
although the variations from the mean figure for the whole series 
of their analyses, viz. 2°944, are very small. 
