IS 2 
THE VOYAGE OF H.M.S. CHALLENGER. 
IV. — Interpretation of the Results. 
1 ;im '( >rrv to have to confess that I have not been as successful as I should have 
wish- -1 in drawing general conclusions from my numbers, and if I here reproduce my 
< i !• ivouis in this direction, I do so chiefly in the hope that some other person, having 
Hi"!- experience than I in dealing with statistics, may take up the problem after me, 
and perhaps be able to extract the latent propositions which are therein concealed. 
In tlu tables which I propose to give, he will find all the data arranged in the most 
convenient form, so that all he needs is at hand. 
T" I- jin with the surface-water gases, a glance at Table XI. shows that the volumes 
"f nitr-gen and oxygen brought out by analysis (Columns IV. and V.) differ more or less 
fr-un th-' numbers (Columns VI. and VII.) calculated on the assumption that the water, at 
its natural temperature ( 0 , had been shaken with constantly renewed air to complete 
- itur.ition under 700 mm. dry pressure. This is no more than one would expect. Even 
the surface water of the ocean cannot be expected, at any time and any place, to be in a 
• it-' -if absorptiometric equilibrium : and for a number of causes which I will proceed to 
- umenit-', beginning with what I conceive to be the less important of disturbing influ- 
• n s. (1) The pressure of the atmosphere is inconstant, though not in general far 
i - miiv. -1 from what my calculations suppose to prevail.* (2) The water is in a state of 
onstant progressive motion ; the sample collected at a certain place was only travelling 
through that place at the time, coming, in general, from a region of different tempera- 
t i (3) Supp"~ing even the water at a given place were in a state of stagnation, its 
t- nip i t u i - * would be subject to periodic variation; it would reach a maximum at a 
rtain hour during tin* day, and fall to a minimum at a certain hour after midnight. 
H i!- ■■ the qu antity (</) of air contained in a litre of water must be a periodic function of 
t ■ i r i - - (T) Ilk- wi-- ; and supposing we could assume that the absorptiometric condition of 
• i (in the ~ " n -e of perfect air-saturation) always adjusted itself instantaneously 
t > the pr-'vailing i--mp-i.it ure. the variation of 7 would follow a curve q=f ( T), the 
m i "f whieh would correspond precisely to the minima, and the minima to the 
in i, of the temperature curve f = </>(T). But absorptiometric exchange is a thing 
• u progress; hence the actual curve 7 „ = F (T) will, so to speak, lag behind the tlieo- 
■ 'ii-', and th- actual (j„ will never quite rise to the maximum nor quite fall to the 
in m --t q. At -ome hour in the early morning and at some hour after sunset, the 
‘ ■ iv. would I presume intersect each other, so that q 0 would become equal to q. 
But the Challenger samples were probably all collected at hours between these two points; 
; nee, on the basis of our assumption, we should presume the actual q 0 (and in a lesser 
■ ? ’ i - 1 ratio «, :n 2 of oxygen to nitrogen) to be somewhat greater than the 
■I i 'It! *, ii < - ii nt-lry-v procure of the nitrogen and oxygen, especially at high 
’ - ' 1 . fill- at in- ■ - j >ln- r -. und cow-ojuently on an average less than 760 min. 
