at LemlngtQn Priors. 35c 



acid, which was ufed in this experiment, was added fome 

 diftilled water, and the liquor was filtrated: it was then 

 faturated with an alkali ; a fmall precipitate fell, which was 

 proved (by the ufual methods) to contain both manganefe 

 and iron. 



6. I have noticed -a variation {%.) in the effecT: of muriatic 

 acid when added to the water. I have obferved a fimilar va- 

 riety in the refiduum itfelf ; which is, that fometimes it has 

 been found foluble in the muriatic acid : when this happens, 

 the addition of an alkali precipitates the refiduum in its ori- 

 ginal cryftalline form ; and this it does before the acid is fa- 

 turated. But moft commonly the muriatic acid does not 

 diffolve it at all. Further, it has been faid above (vn. i.) 

 that the acids precipitate a cryftalline fubftance from oxyge- 

 nated muriat of iron. This fubftance I have found to con- 

 tain manganefe. But what belongs to this place to obferve 

 is, that it is not always, indeed it is but rarely, that this ef- 

 fect can be produced in any great degree : in a very fmall 

 degree it may always be obferved ; but, when I firft remarked 

 it, the precipitate was very copious, fo that enough was 

 readily collected for examination by the blow-pipe. The 

 lalt of manganefe does not fhew this appearance in the 

 (mailed degree. It depends, therefore, on fome peculiarity 

 of the iron ruft, but precifely on what I cannot take upon 

 me to determine. 



As the oxygenated falts unite with all the other falts of 

 the water, and confequently cannot be feparated by fpirit of 

 wine, I have found it impoffible to determine the quantity 

 of them. 



I had concluded from the experiment (i.) that this water 

 contained muriat of lime; and the following remark con- 

 firmed me in my error: I mention it, as I think it probable 

 that others have been led into miftakes from the fame caufe. 

 I reduced fome of the water, by evaporation, to about two 

 (iunce-meafures ; taking for granted that by this proce/s 

 A a 2 nearly 



