I No. 1 fandtes 0.0595 og 0.0(302 pCt. OaO. 



I ,. 11 — 0.0580 ,. 0.059(1 .. 



I ..' 26 — 0.0590 _,. 0.0596 ,. 



I ,. 33 0.0585 ,. 0.0595 ,. 



I n 34 _ 0.0590 ,. 0.0593 ,.' 



I n 35 _ 0.0585 .. 0.0588 „ 



I „ '36 — 0.0577 ,. 0.0579 ,. 



I .. 49 0.0578 .. 0.0582 ,. ,. 



I h so _ 0.0583 ,. 0.0588 ,. 



Den største Differents mellem Bestemmelserne i et 

 og samme Vand er altsaa 0.001 pCt. 



Til Bestemmelsen af Magnesia inddampedes de to 

 Filtrater fra Kalkiældingerue i en Platinskaal til omtrent 

 150 CC. og bundfældtes derpaa med fosforsurt Natron og 

 en san stor Mængde koncentreret Ammoniakoplosning. at 

 denne kom til at udgjøre omtrent 1 / 3 af Vædskens Volum. 

 Iagttager man ikke den Forsigtighedsregel at tilsætte et 

 saadant Overskud af Ammoniak, vil man let faa uoverens- 

 stemmende Resultater. Ved at anvende forskjellige Ammo- 

 niakmængder fandt jeg i et og samme Vand følgende Tal 

 for Magnesia. 



Med 9 Dele Vædske og 1 Del kone. Amm. 0.2044 pCt. MgO. 

 .. r. .. „ 1 „ • 0.2061 .. ,. 



.. 2 ,. ,.1 ,. 0.2140 .. 



I et andet Vand fandtes paa samme Maade: 



Med 9 Dele Vædske og 1 Del kone. Amm. 0.2054 pCt. MgO. 

 ,. 2 .. • .. 1 ,. — 0.2112 ,. 



Nødvendigheden af at tilsætte et saadant Overskud af 

 Ammoniak for at faa en nøiagtig Bestemmelse af Magnesia 

 efter denne Methode er altfor lidet fremkævet i de analy- 

 tiske Læreboger. Rimeligvis er denne mindre paaågtede 

 Feilkilde Grunden til. at flere Kemikere bar fundet meget 

 uoverénsstemmende og lave Tal for Magnesia, i Søvandet. 

 Forcbbammer. hvis Resultater med Hensyn til Kalk- og 

 Magnesiamængden stemmer godt overens med mine Ana- 

 lyser, inddamper Filtratet fra den fosforsure Ammoniak- 

 magnesia til Tørhed og faar ved Opløsning af Residuet i 

 Ammoniak altid en liden Rest af det fosforsure Salt. 

 En saadau [nddampning er imidlertid ikke nødvendig, 

 naar man blot før Fældningen koncentrerer Vædsken til 

 det mindst mulige Volum og tilsætter det nødvendige Over- 

 skud af Ammoniak. Naar Bundfaldet derpaa efter 24 

 Timers Eenstand frafiltreres, giver Filtratet ved Inddamp- 

 ning ingen' Hest af fosforsur Ammoniakmagnesia. 



Men uagtel disse Forholdsregler kan Magnesia ikke 

 bestemmes med samme Sikkerhed i Havvandet, som Kalk 

 og Svovlsyre. Nedenstaaende Tal viser Differentserne mel- 

 lem Kontrolbestemmelserne : 



I No. 33 fandtes 0.2160 og 0.2200 pCt.' MgO. 



I .. 35 — 0.2180 .. 0.2207 ,. 



I .. 36 - 0.217;} ,,.0.2180 „ 



Største Differents = 0.004 pCt. 



In Sample 1 the proportion was 0.0595 and 0.0602 p.ct. CaO. 



In — 11 ,. 



In — 26 ,. 



In — 33 ,. ,, 



In — . 34 ,. 



In — 35 ,. „ 



In — 36 ,. ,. 



In — 49 ,. 



In — 50 ,. 



Hence the greatest difference between a ny two de- 

 terminations with the same sample of water is 0.001 per cent. 

 For determining the magnesia, the two solutions from 

 which the lime-precipitates bad been filtered off were evap- 

 orated in a platinum dish to about 150 M , and tben treated. 

 for precipitation, with phosphate of soda. and with a eon- 

 centrated solution of ammonia measuring one-third of the 

 volume of the fluid. Unless ammonia be added in such 

 excess. there is a risk of obtaining incongruous results. 

 On treating different quantities of the fluid with the same 

 proportion of ammonia. I found the following figures for 

 magnesia. 



With 9 parts of fluid and 1 pårtof conc.amm. 0.2044 p.ct. Mg< ). 

 „ 5 „ „ ,. .. 1 „ „ ■ 0.2061 ,. 



„ 2 .. .. .. .. 1 .. ,. 0.2141» .. .. 



In another sample of water, the percentage tbus found 

 was — 



With 9 parts of fluid and 1 part of conc. amm. 0.2054 p.ct. MgO. 

 „ 2 ,. ,. ,. .. 1 .. ,. 0.2112 ,. ,. 



The necessity. witb this method. of adding ammonia 

 in so great excess, to ensure a satisfactory determination 

 of magnesia, bas [not been sufticiently empbasized in ana- 

 lytical class-books'; and to this comparatiyely disregarded 

 source of error. may no doubt be attributed the fact of 

 divers chemists having obtained for magnesia, in their ana- 

 lyses of sea-water. results alike incongruous and low. 

 Forcbbammer. whose lime and magnesia determinations 

 exhibit close agreement with my own. evaporates to dryness 

 the solution from which the double phosphate of ammonia 

 and magnesia has been filtered. and can. on dissolving the 

 i-esidue in ammonia. invariably detect traces of the phos- 

 phate. Meanwhile, evaporation does not constitute a necessary 

 phase of the process, provided the fluid, previous to precip- 

 itation. be concentrated to the least possible volume. and 

 tben treated with ammonia in due excess; for. on the pre- 

 cipitate, after 24 hours, being filtered off. 

 evaporated to dryness. will not leave a 

 phosphate of ammonia and magnesia. 



But. even assuming every precaution, 

 sea-water «loes not admit of being determined with the same 

 degree of accuracy as du lime and sulphuric acid. The 

 annexed figures show the extent to which the test-deter- 

 minations were found to dilter. 



ln Sample 33 the proportion was 0.2160 and 0.2200p.ct.MgO. 

 In — 35 .. .. ,. 0.2 ISO .. 0.2207 ,. .. 



In — 36 .. „ .. 0.2173 .. 0.21 su .. .. 



Greatest difference: 0.004 per cent. 



the solution. if 

 trace of double 



the magnesia in 



