_The Double Iodide Solution._--In a note appended to DR. MANSELL'S
communication on the calotype (Vol. ix., p. 134.), you state that having
lately prepared the double iodide solution according to the formula given
by DR. DIAMOND, in which it required 650 grains of iodide of potassium to
dissolve a 60-grain precipitate, you were inclined to believe, until you
made the experiment yourself, that DR. MANSELL must have made a wrong
calculation as to the quantity of iodide of potassium (680 grains) which he
stated was sufficient to dissolve a 100-grain precipitate, as the
difference appeared so small for a solution more than one-third stronger.
The small difference referred to with respect to the quantity of iodide of
potassium required, is owing to the amount of water used being in both
cases the same. A slight difference in the strength of a solution of iodide
of potassium makes a great difference with respect to the quantity of
iodide of silver it is capable of dissolving. Thus, if you remove a small
proportion of the water from a solution of the double iodide of silver by
evaporation, the slight increase of strength which the solution will
thereby acquire, will enable it to take up a much larger proportion of
iodide of silver than it already contains; and if, on the other hand, you
dilute it with a small proportion of water, its diminished strength (unless
the solution contains a great excess of iodide of potassium) will cause the
precipitation of a large proportion of the iodide of silver. And hence the
great variation in the amount of iodide of potassium which is found
requisite to form a solution of the double iodide of silver, under the same
apparent conditions with regard to the proportions of the other ingredients
employed, may be accounted for by the impossibility of _measuring_ off with
sufficient accuracy the proper proportion of water.
Whenever _exact_ quantities of liquids are required, recourse should always
be had to the balance, for no great accuracy can be depended upon by
measurement with our ordinary glass measures, even supposing them to be
correctly graduated, which is not always the case.