91 
and alcohol, the solution is mostly converted into a crystalline 
magma; and if it is desired to separate a small proportion of 
calcium, which we will suppose to be present, the magma 
must be filtered, dissolved, and subjected to the treatment 
again and again to separate the sulphate of calcium, when, if 
the quantity of that salt present be very minute, it must be 
wholly lost. Of course these remarks do not apply to solu- 
tions of calcium and magnesium salts containing much of the 
former, except in a modified degree. What is true of the 
sulphuric acid and alcohol process, is true in a more extended 
sense of the oxalate of ammonium process. I have precipitated 
within a trace the whole of the magnesium present in a 
considerable quantity of solution of chloride of magnesium, 
simply by successive additions of oxalate of ammonium, — the 
solution being concentrated to its original bulk after the last 
addition of the re-agent. Yet, in working with this re-agent 
the rule is, that enough of it must always be added to trans- 
form all the magnesium salt into oxalate, since oxalate of 
calcium is soluble in solution of chloride of magnesium. That 
some magnesium salt must precipitate with the lime salt 
under such conditions is obvious ; and that it does so is well 
known, and is, though incompletely, provided for by the 
process being directed to be repeated upon the precipitate first 
obtained. This process, therefore, is also one of fractional 
precipitation, and for it to approach success, the operator 
must know pretty nearly beforehand how much calcium, in 
proportion to the magnesium present, he has to deal with. 
Nevertheless, it is unquestionable that in skilled hands, 
either of the two last processes is capable of giving close 
approximations to the truth, when the quantity of calcium 
present amounts to a few per cent of the mixed salts. When 
the quantity of calcium is less than 1 per cent, I do not think 
it is possible to estimate it accurately by any of these pro- 
cesses ; and when the proportion is larger, the processes are 
at least more troublesome, have a wider limit of experimental 
