352 Prof. G. F. Armstrong. Diurnal Variation [Apr. 29, 



Yols. of CO. in 

 10,000 vols, of air, 



1868 (October to December) inclusive 2*8943 



1869 (January to December) inclusive 2*8668 



1870 (January to December) inclusive 2-9052 . 



1871 (January to July) inclusive 3*0126 



11*6789 



These give a final average on the four years of =2*92 vols. 



C0 as against 2*96 vols. C0 2 in 10,000 of air obtained in the present 

 instance ; a difference, that is, of only 004 vol. C0 2 in 10,000 of air. 

 This close correspondence between the results is important on 

 account of the method of determination pursued being essentially the 

 same in each case, with the exception that Schulze employed jars of 

 only 4 litres capacity as compared with 14 litres in the present case. 



But what is chiefly interesting in this comparison arises from the 

 fact of these low averages obtained by Schulze having been attributed 

 to the contiguity of Rostock, where the experiments were made, to the 

 sea. They approximated so closely to Dr. Thorpe's Atlantic results* 

 for the day, namely, 3*011 vols. C0 2 in 10,000 of air, that their 

 lowness was generally regarded as being due to the influence of sea- 

 air, and this solution has been the more forcibly urged also owing to 

 the fact that, in using a modern method of determination, an average 

 of 347 similar experiments made at Duhmef by Fittbogen and 

 Hasselbarth, has been found to give as much as 3*34 vols. C0 2 , and 

 17 other determinations made by Henneberg at Weend, near 

 Gottingen,J 3*2 vols. C0 2 in 10,000 vols, of air, while Truchot has 

 obtained considerably higher figures still. 



A similar explanation might with equal plausibility be applied to 

 the low results more recently obtained by Reiset§ in the open country 

 near Dieppe. For, although he employed the more perfectly absorbing 

 method of the aspirator and Woulfe's bottles, and for the remainder 

 followed Pettenkofer's plan of determination, yet, as a mean of 92 

 experiments made between September, 1872, and August, 1873, his 

 average is only 2*942 vols. C0 2 in 10,000 of air. This result, it will be 

 seen, agrees in a still more remarkable manner with that arrived at in the 

 Grasmere experiments, the difference between the two averages being 

 only 0*018 vol. C0 2 in 10,000 of air. But the sea-air explanation can 

 scarcely be so confidently relied upon in the latter instance as in some 

 of the others. Rostock and the place at which the air was collected 

 near Dieppe, are both close to the sea, the one on the embouchure of the 

 Nebel, and the other not more than 5 miles from the coast. Grasmere, 



* " Journ. Cliem. Soc," [2], v, 189. 



f " Watts' Chem. Diet.," Suppl. Ill, 132. 



X Ibid. 



§ " Compt. Eend.," 88, 1007-1011. 



