148 
THE VOYAGE OF H.M.S. CHALLENGER. 
No. 193. ] “ Air collected deck, 2 p.m., November G, 1875.” Lat. 37° 55' S., long. 
93° 56' W. 
\ >1. of sample = 309 '64 ; minus CO., = 309'41 ; contraction = 0 '23. Residue divided. 
Portion I. — VoL - 162*25 ; plus hydrogen, 286*45 ; after explosion, 148*28; c = 102*17; oxygen = 34 - 057 
m 20 ’99 per cenL 
/’ Hi II. VoL 147*11 ; /'/'/*• hydrogen = 231 *85 ; after explosion = 139 *40 ; c = 92*45 ; oxygen = 
30 8 17 =« 20*93 per cent 
Summary. 
Percentage of Oxygen. 
[No.] 
I. 
IT. 
Mean. 
92 
20*99 
20*89 
20*94 
194 
21*05 
21*00 
21*02 
192 
20*99 
21*01 
21*00 
193 
20*99 
20*95 
20*97 
By way of comparison, let us state that Rrgnault found for the percentage of oxygen 
in air freed from carbonic acid in — 
5 samples from the Atlantic Ocean, .... from 20*918 to 20*965 
1 „ Ecuador, ..... 20*960 
„ Switzerland, ..... „ 20*909 to 20*993 
lb Bunsen found, in 28 samples collected at Heidelberg, from 20‘84 to 20‘9G3 per 
cent.; mean of all the 28 results, 20 924. 
Th< composition of unpolluted air, in short, is the same everywhere; hence, the 
proper mode of interpreting our 8 analyses is to Bay that they brought out results whose 
greatest deviation from their mean 20‘984 was 0*066, and this for an apparatus of the 
mod r#t pretensions of mine is a very high degree of precision. 
Unfortunately, the same degree of exactitude could not be attained in the analyses of 
*m]'l< - <*f ;iir from -< a-water ; the deviations from each other of numerous duplicate 
nslyses were in general considerablj greater. I can explain this only by the unavoid- 
able aheorptiometi action of the relatively large amount of caustic potash used on the 
'"ijb ■* introduc <1 into the pipette. In the case of atmospheric air analyses, the 
r !_ , ‘*nt, by the action of tin* air of the laboratory, has already come into the 
j * in- f ri* condition which it tends to assume in the analysis. In the case of 
M ];. C of ga - from - a-wat< r of variable composition, the elimination of the 
inavoida : • ly arcompnnied by addition of nitrogen or oxygen from the 
onj l* above it. If I had to do the work again, I should certainly use 
m< r- ly mm.*. I, cau tit potash for the absorption of the carbonic 
arid. 
