35 



udført en Række Forsøg til Bestemmelse af de organiske 

 Stoffes Mængde, om at meddele mig de fornødne Data. 

 [følge ham ere de organiske Stoffes Mængde kun under- 

 kastet meget smaa Variationer, og affarver 1 -Litre Sø- 

 vand gjennemsnitlig 3 Mgr. Kamæleon, som under Forud- 

 sætning af Reduction til < (xydulsalt kun kan afgive hen- 

 imod 0.8 Mgr. Surstof, hvoraf det lettelig vil kunne indsees, 

 ut ilen dannede Mængde Kulsyre ikke kan være meget 

 stor, naar Surstofforbruget selv ved en saa energisk Oxyda- 

 timi som ved Anvendelse af Kamæleon er saa lidet. Og 

 hvad der især tyder paa, at den Oxydation, som kan 

 foregaa ved Vandprøvens Henstand under ordinære For- 

 holde, fuldstændig maa kunne negiigeres, er den < >m- 

 stændighed, at Schmelck har fundet Søvandéts Evne til 

 at • affarve Kamæleon ligestor'. hvadenten Vandprøvén imder- 

 søges frisk eller forst efter aarelang Henstand.! 



At noget afKulsyren ved den lange Henstand skulde 

 , være fordampet, er der beiler ingen Rimelighed .for. naar 

 man erindrer, at Sovandet er en alkalisk Vædske, som 

 ikke indeholder det mindste Spor af fri Kulsyre. og som 

 med saadan Kraft holder paa sin surt bundne Kulsyre. at 

 den i timevis kan koges i det af Jacobsen angivne Luft- 

 udkogningsapparat med det Resultat, at kun en meget 

 liden Brøkdel uddrjves. Der er aabenbart større Fare for 

 det Modsatte. nemlig at der skulde kunne absorber.es noget 

 af Atmospluerens Kulsyre. men for Dndgaaelsen af denne 

 Feilkilde er der sørget ved omhyggelig Opbevaring paa 

 velproppede Flasker. 



Resultaterne synes heller ikke at tyde paa. at Hen- 

 standen skulde have været til Skade for Vandprøvernes 

 ' Brugbarhed. idet' de alle uden Hensyn til den -meget for- 

 skjellige Varighed af det Tidsrum, hvori de have henstaaet, 

 vise.nogenlunde det samme Resultat, og jeg tror saaledes. 

 ikke at tåge meget Feil, naar jeg anser de af mig udforte 

 ( (bservationer som i alt Væsentligt ligesaa gode. som om 

 de havde været udforte ombord paa friskt optagne Vahd- 

 prøver. 



Hvad angaar de af. selve Methoden og Experiménta- 

 tionen Hydende Observationsfeil. da vil de hyppigt udførte 

 Controlbestemmelser kunne give de fornødne Bidrag til 

 Bedømmelsen af deres Størrelse, saaledes' er ved 13 Con- 

 trolbestemmelser for den neutralt bundne Kolsyres Ved- 

 kommende den gjennemsnitlige halve Differeiits mellem 2 

 paa samme Yandprove udforte Observationer bestemt til 

 0.11 Mgr. per Litre . (Maximum 0.6) og for den 'surt 

 bundne Kolsyres Vedkommende til 0.59 Mgr. per Litre 

 (Maximum 1.35). Foruden den variable Feil vil der imidlertid 

 i sidste Tilfælde ogsaa være en constant. idet de,t ikke 

 lader sig undgaa. at man ved Arbeide i en kulsyreholdig 

 Atmosphære vil erholde noget om end meget lidet for hoie 

 Resultater. Naar Feilene ved Bestemmelsen af den surt 

 bundne Kulsyre er lunden at være størst, da er Grunden 

 dertil for en stor' Del at soge deri. at man ved Retitration 

 med Natronlud. Især naar Rosolsyre anvendes, som Index, 

 erholder en meget skarp Endereaction. medens dette paa 

 lang Vei ikke i samme Grad er Tilladde, hvor Barvt re- 



ine witb tlie necessarj data, Ac6ording to his obserations, 

 the aniount of organic matter varies bul very slightly, 

 1 litre of sea-water discolouring on an average '■'< "'" per- 

 manganate of potash; and this quantity, assuming extteme 

 reduction. cannot give t off more than aboul 0.8 m «" ofoxygen. 



Heiice it is obvioUS that the amiiunt of Carbonk aeid can- 

 not be very large, considering the limited consumption of 

 oxygen even witb the use of permanganate of pbtasb. But 

 that the oxidation wbich can ordmarily result' from allow- 

 ing the water to stand over may be altogetber ignored, 

 is mure particularly indicated by the fact of Schmelck hav- 

 iilg foiind the property in sea-water of discolouiing perman- 

 ganate of potash to be invariably the same. whetber the 

 samples are freshly drawn or have been preserved for year,s 

 together. 



Nor is there reason to assumé, that any portion of 

 the carbonic acid shouhl have escaped by evaporation dur- 

 ing the Long ihtervalj if we bear in mind that sea-water is 

 an alkaline fluid, wbich does not contain the smallest trace 

 of free carbonic acid. and wbich retains that present in 

 bicarbonates with such vigour, that it may be boiled for 

 hburs together in the apparatus devised by Jacobsen for 

 boiling out air and not part with more than a mere frac- 

 tiou. Nav. there is obviously danger of the reverse. viz. 

 that some of the carbonic acid present in the atmosphere 

 will be absorbed; Imt that source of error was effectually 

 avoided by kéeping the water in bottles provided with 

 tight-titting glass stoppers. . 



Judging. too. from the results. wbich were very nearly 

 the same for all the samples, irrespective of the difference 

 in the length of the period during wbich they bad been 

 preserved. the interval that bad elapsed previous to examin- 

 ation ditl not appear to have bad any injurious effect on 

 the water for experimental purposes : and hence I ■ feel 

 tolerably convinced that my observations in all essential 

 particulars are as reliable as if they had been conducted 

 on board with freshly drawn ■ water. 



With respect to the errors of observation involved in 

 the method itself. numerous test-determinations will serve 

 as an approximate standard by wbich to compute their 

 magnitude: thus. for instance. half the difference between- 

 two observations 'witb the same sample of water was found. 

 for the carbonic acid forming carbonates. to be. 0.1 1'"'"' per 

 litre (maximum 0„6), and 'for the carbonic acid forming bi- 

 carbonates to be 0.59'"'"' per litre (maximum 1.35). Exclusive 

 of the variable error there will also. in the latter case. be 

 a constant one. inasmuch as the results of experiments 

 performed in an atmosphere containing carbonic acid must 

 necessarily be somewhat, if but a very little. too high. 

 W hen the error in the determination of the carbonic acid 

 forming bicarbonates is found to be greatest, the reason 

 will be chietiy this. that retitration with soda-lye. more espe- 

 cially if rosolic acid has been selected as the index. gives 

 rise to a very decided terminal reaction. wbich does not 

 result on the baryta water beins titrated with oxalic acid: the 

 reaction is then much less obvious. It must be likewise 



