172 Scientific Proceedings, Royal Dublin Society. 
Tt will be seen that the earliest workers obtained an amount 
which was vastly too high : nor need any surprise be felt at this, 
as their method was eudiometric, and therefore not sufficiently 
delicate for the purpose. 
Dalton, in 1802, was the first to correct the extravagant 
figures of Humboldt, and by employing a different method to obtain 
results which were only about twice as large as the correct value. 
Thenard, in 1812, originated a different method, by which he 
obtained results which were surprisingly near the truth. 
De Saussure, whose researches extended over the period 1809- 
1830, and who made some hundreds of observations, did not 
materially reduce Thenard’s figure—but it is interesting to notice, 
as Blochmann has pointed out, that his figures almost steadily 
decreased as his work progressed. 
From this time until about the year 1870, the normal amount 
of atmospheric carbonic anhydride appears to have been taken as 
4 in 10,000 vols., and we fancy that many chemists are inclined t) 
quote this figure even at the present day. 
Later researches have, however, made it clear that a reduction 
on that amount of at least 25 per cent. must be made for the 
quantity in the “ fresh air” of most localities, with the exception of 
towns. . 
It is difficult to say exactly to what observers we are indebted 
for this revision. In 1848-1850 Marchand, at Halle, had found 
an average of 3°10 in 150 determinations; and, in 1865, Angus 
Smith an average of 3°36 in 92 determinations on the air of the 
country in Scotland. ‘Thorpe, in 77 experiments made in 1865- 
1866 on ocean air, had found 3:0, and I’. Schulze, a yearly average 
from 1868-1871, at Rostock, varying between 2°86 and 3-01. 
Marchand’s determinations do not appear to have been generally 
known, and the lowness of the other figures was at first attributed 
to the proximity of the places of observation to the sea; but the 
very careful work of Reiset partly at Paris, and of Mintz and 
Aubin, not only in France, but in various localities visited by the 
French expeditions sent out to observe the Transit of Venus, in 
1882, and subsequently, the results obtained by most other observers 
have shown that the true amount oscillates about an average of 
3 in 10,000, or according to Mintz and Aubin, 2°82 for the 
northern hemisphere, and 2°72 for the southern. 
