64 
was 3 '00 volumes. From Saussure and Boussingault’s 
observations a decided difference may be traced between 
the amounts of carbonic acid in the air of day and night on 
the land, the air of the night containing more carbonic acid 
than that of the day in the proportion of 100 to 92 ; but, on 
comparing the experiments made at night with those of the 
day, it appears from the author’s observations that no such 
differences are to be discerned in sea air. The means of the 
two series are — 
For the Day. For the Night. 
(24 < observations) (20 observations) 
Carbonic acid in 10,000 vols. of air... 3*011 2*993 
The non-accordance of these results with those obtained 
by Lewy is undoubtedly mainly due to the difference in the 
methods of analysis employed. Lewy’s determinations were 
made in Regnault’s eudiometer in which, as is well known, 
the volume of gas is maintained constant throughout the 
analysis, the relative changes being estimated by the varia- 
tions in the pressure of that constant volume over or under 
that of the atmosphere, measured in millimetres of mercury ; 
and thus, as Dr. Frankland as already pointed out, a varia- 
tion in volume, which in the older method of Bunsen would 
appear considerable, is in that of Renault only represented 
by a small numerical expression, and consequently a very 
slight error of observation (say to the extent of one-twentieth 
of a millimetre) would exercise a very considerable influence 
on the result. It deserves also to be noted that Lewy’s 
analyses were not made until eighteen or twenty months 
after the air had been collected, and although Lewy satisfied 
himself by experiment that air contained in glass tubes for 
such a length of time experiences no alteration in composi- 
tion, yet Regnault has subsequently shown that the exact 
determination of the amount of atmospheric carbonic acid 
cannot be made in air thus preserved, since the glass absorbs 
a portion of this gas. 
