14 



sig fra *-m st;illi - I ' Vil ' nans ^ ua b' ser - da han nemlig. til 

 Udfældning af Svovlsyren har benyttet sig af den salpeter- 

 . sure Baryt. hvorved man — som senere er bevist — faar 

 et stærkt forurenset Bundfald. 



Af den sidste og dé foregaaende Tabeller fremgaar 

 det. at fra hvilken Side man end ser alle disse Bestem- 

 melser af Kalk. Magnesia og Svovlsyre, kan de ikke med 

 Bestemthed bevise nogen Foranderlighed i Søvandets Sam- 

 mensætning: men det bor tillige bemærkes, at de Afvigel- 

 ser. der er vanskeligst at forklare, isærdeleshed fremkommer 

 ved Svovlsyrebestemmelserne. 



Den Antagelse, at Havet i hele sin Dybde er en ens- 

 artet Blanding, hvori den nøiagtigste kemiske Analyse neppe 

 kan paayise nogen Forskjellighed, bekræftes ved de her 

 foreliggende Undersøgelser i endnu høiere Grad end ved 

 de tidligere. deg har nemlig ikke nogensinde fundet saa 

 store Uoverensstemmelser sum de. der paa enkelte Steder 

 er fremkomne ved Forchhammers og andre Kemikeres 

 Analyser. • 



Retter man nu < )pinærksomheden paa de fortyndede 



Vandprover. hvis Egenvægt er under 1.0260, vil man hel- 

 ler ikke her tinde nogen Forandring i det konstante For- 

 hold mellem Bestanddelene. Station- 300, 339 og 350, 

 hvor Egenvægt og Klorgehalt er formindskede ved nærlig- 

 gende Ismassers Smeltning. viser en nøiagtig tilsvarende 

 Forskjelliglied for. de ovrige Bestanddeles Vedkommende. 

 I Station 143. der er beliggende i Nærheden af den norske 

 Kyst, og hvis Overfladevand som folge heraf er noget for- 

 tyndét, har jeg ved to Bestemmelser fundet en temmelig 

 hoi Kalkgehalt. medens derimod Magnesia og Svovlsyre 

 fuldkommen retter sig efter Klormængden. Dette fortjener 

 maaske en Smule ( )pmærksomhed, da der jo er Omstæn- 

 digheder, som taler for en Tiltagen af Kalkmængden ved 

 Kysterne. 



I Station 256, hvor Overfladevandets Saltstyrke er 

 reduceret til 2.02 pCt., viser dog Forholdet mellem Be- 

 standdelene saa smaa Afvigelser fra det almindelige. at 

 disse fuldst.endig kau forklares som Feil i Bestemmekserne. 

 der ved Våndets Fortynding naturligvis taber i Xoiag- 

 tiglied. 



Kjender man Søvandets Klorgehalt, vil man af de i 

 Tabelleme angivne Korlioldstal kunne beregne de ovrige 

 Bestanddele med saa stor Nøiagtighed. at de fundne Tal 

 for Kalk. Magnesia og Svovlsyre kmi undtagelsesvis vil 

 differere mere end 0.002, 0.005 og 0.003 pCt. fra dé direkte 

 Bestemmelser i Tabel i. 



Porat lette Sammenligningen med de tidligere Ana- 

 lyser har jeg i etterfølgende Tabel paa samme Vis som 

 Forchhammer forbundet Klor og Svovlsyre med Kalk og 

 Magnesia til! de Salte, som i Almindelighed antages at 

 forekomme i Søvandet. Al Kalk er regnet til Svovlsyre, 



tain a. larger quantity of sulplmric acid. since he found the 

 proportion of that constituent between the 60th and 62nd 

 parallels of latitude to be 0.230 per cent. But, in all 

 probability. this incongruous result must be ascribed to a 

 constant error in his analyses, precipitatmg as he did the 

 sulplmric acid with nitrate of baryta, which — as will be 

 afterwards shovm - - gives a niuch contaminated deposit. 

 From the last Table, it is sufticiently clear. that how- 

 soever these determinations be regarded. they cannot be 

 assumed to furuish absolute proof of a variable relation 

 subsisting between the constituent parts of sea-water; and 

 we must bear in mind, that of such differeuces as have 

 bitherto defied the penetration of Observers, the most strik- 

 ing refer to sulphuric acid determinations. 



The hypothesis which assumes the Ocean to consist 

 throughout its entire depth of one homogeneons fluid, in which 

 the most aecurate of chemical analyses shall fail to detect dis- 

 similarity of composition, has veceived from the experiments' 

 here described probably stronger contirmation tlian from 

 any that have gone before them. Indeed none of my own i 

 results exhibit a want of agreepient so considerable as that 

 met with in some of Forchhammers and other chemists' 

 analyses. 



Again. passing to the diluted samples (with a spe- 

 citic gravity under 1.0260), no disturbance will be apparent 

 in the constant relation subsisting between the component 

 parts of the water. At Stations 300. 339. and 35o. where 

 the specitic gravity and the proportion of chlorine are 

 reduced by the melting, in the immediate viciuity. of large 

 masses of ice. the other constituents exhibit a correspond- 

 ing difference in amount. In two samples drawn at Sta- 

 tion 143. in close ' proximity to the Norwegian coast. where 

 the surface-water is accordingly somewhat diluted. I determin- 

 ed a comparatively large amount of lime. whereas the 

 proportion of magnesia and sulphuric acid was in strict 

 proportion to that of the chlorine. This result is,. perhaps, 

 deserving of notice. inasmuch as there are grounds for 

 assuming the amount of lime to increase near the coast. 



At Station 256, where the proportion of salt in the 

 surface.-water is reduced to 2.02 per cent. the disturbance 

 in the normal relation subsisting between the several con- 

 stituents was so slight, that it could be wholly accounted 

 for in each case as au error of analysis, the determinations 

 beihg by reason of the dilution of the water proportiou- 

 ately less aecurate. 



Griveh the amount of chlorine in sea-water, the other 

 constituents may be calculated from the proportional valiies ' 

 given in the Tables witli such accuracy, that the hvures 

 found for lime. magnesia, and sulphuric acid will rarely 

 differmore than 0.002, 0.005, and 0.003 per cent from the. 

 direct determinations set fortb in Table I. 



With a view to facilitate comparison with the eaflier 

 analyses. 1 have in the ann exed .Table, following the exam- 

 ple of Forchhammer. combined chlorine and sulphuric acid 

 with lime and magnesia, to form the salts which are. gener- 

 ally assumed to occur in sea-water. Thus. all lime is 



