M. C. Lea on a new test for Hyposulphites. 223 
ra containing one twelve-thousandth gave a well marked 
ink liquid. 
nm containing one twenty-five-thousandth gave a salmon 
or. 
The experiment was not carried further because the salmon 
color in the last-mentioned trial showed that the test had then 
reached its practical limit. Ido not doubt that even with one 
hundred thousandth a coloration could be obtained, but it would 
not have the specific distinctness given by the earmine and rose 
shade previously described. 
_ A few words remain to be said as to the best mode of apply- 
ing this test. 
Thave recognized in solutions of sesquioxyd of ruthenium a 
strong tendency to decompose by dilution: dilute solutions have 
4 strong tendency to gradually deposit their ruthenium as oxyd. 
And even before the slightest sign of a precipitate appears, in 
immediately upon dilution, solutions show a tendency to 
change their reactions. So that I find it invariably better on 
diluting the ruthenium solution for use in testing, to boil it (as 
Thave elsewhere pointed out in speaking of the dilution of ru- 
thenium) with a few drops of hydrochloric acid, and this even 
although the solution is to be immediately afterward rendered 
alkaline by ammonia. To ascertain with certainty that this im- 
proved the delicacy of the reaction, I made comparative experi- 
ie on two ‘tions of the same ruthenium —s ina ; 
ound that the coloration by h ulphite was at least three, 
umes Stronger in the case of de yore that had been boiled 
with HCl than with that that had not. tae 
‘S ammonia was thereafter immediately added, it might ap- 
that the function of the hydrochloric aci form hy- 
In using this reaction for the detection of small quantiti 4 
. 
Very little ruthenium is present. 
—— Eyposulphite, it is useful to remark that i ane a 
