."()() CONCKKNINC THE HI'KKCT ol' ;;KA\1TV OAVIS. 



Su|)('i"]MiSe(l on rliis is the ctfcet of i2,ravit v. whicli. as we have 

 seen, will, t'nr inm-ganic salts, be to increase the concentration 

 with the (k']ith. T havi- ivuiahly rciirc^scntcd the first effect by 

 the curve (1). Fig-. 4, and that of the second h_v cnr\e (2). They 

 will aive a n'sultant curve (3) of chanue in concentration with 

 dcjitli. an 1 if we suppose the first effect to l^e the lireater near 

 the surface, then the concentration will first decrease, not neces- 

 sarily uniformly, down to a certain point where the two elfect^s 

 just counterbalance, ami from thence will increase steadily fo 

 the hottoni. 



This would account for the actual distribution of the saline 

 contents of sea Avatei- as actually found from the summary of 

 the researches on the "Challenger" as i:,iveu by Dittnuu". (See 

 Enrj/clojKi'dia BvUannlca. !»th edit.. ''Sea Water"). 



"In ]daces where there is active dilution at the surface, the 

 salinity as a I'ule increases do-vvn to some 50 or 100 fathoms. 

 but thence downward it follows thi^ general rule, that is. it 

 decreases doivi} Ui SOO or 1 .000 fd/Jtoins. and thence i)H'reases 

 steadily to the bottom." 



Of the data available from the reports on the concentration 

 of the separate constituents, only that of the absorbed gases is 

 of value since the others wei'e determined i'(datively to the total 

 rlil'irino content. 



A)iart from the fact of the large increase of solnbility of 

 gases with teni|)eratnre. ihe increase of density with concentra- 

 tion, though jiositive, is extremely small auil difficult to measure. 

 (S e Ostwald, Solutions, p. 32). 

 It Avas found that : 



(1) The amount of nitrogen increases with the depth. 



(2) The amount of oxygen decreases with the depth. 



But as a matter of fact, the increase of density with con- 

 centration for oxya'cn is greater than for nitrogen, whicdi <»ught 

 therefore to show a greater increase with the depth. 



Dittmar explains the decrease i>{ oxygen as being due to 

 oxidation. 



(For the greater part of the mathematics in this ])aper 1 am 

 indebted to Dr. 11. I.. P.ronson. of the I)ei)artmcnt of Physics, 

 Palliousie rVdlege.) 



