[Enclosure]⁴

diffused through 77,000 parts of water gave with starch a light pink colour and the colour became blue when the proportion amounted to $\frac{1}{470}$ grain in 28,000 parts of distilled water—⁵ Arsenic—“D^r Traill⁶ has lately asserted that $\frac{1}{16000}$ part of a grain is precipitated by the silver test and that with ^1/10,000 grain the precipitate is visible to the eye. I have found that $\frac{1}{8000}$ grain dissolved in one drop of water gave a pale yellow film, but the result materially depended on the quantity of water present—   Thus $\frac{1}{4000}$ grain of arsenic in ten drops of water was not perceptibly affected by the test but $\frac{1}{2000}$ grain dissolved in four drops of water gave a decidedly yellow precipitate”—⁷

“This test (sulphuretted hydrogen) is extremely delicate in its reaction.—   It begins to give a yellow tinge when the liquid contains only $\frac{1}{4000}$ grain of arsenious acid in ten drops of water the arsenic therefore forming about ^1/40,000 part of the solution”—⁸

“There is no doubt that considerably less than the millionth part of a grain of arsenic may by Marsh’s test be rendered visible on a glass plate;⁹ it is possible to distinguish by the eye a piece of leaf gold which would weigh less than the ten-millionth part of a grain but the real question is whether the test will discover arsenic in a single drop of solution made by dissolving one grain of the poison in a million grains or sixteen gallons of water! If not the statement amounts to nothing for it is clear that if more than one drop of such an extremely diluted solution be taken the test is acting upon a larger quantity of arsenic than the above form of expression would indicate—   I have generally found that the fractional quantity stated to be detected referred rather to the degree of dilution than to the absolute quantity of poison present, whereas a test may fail to act as we have already seen either from the smallness of the quantity present or from the very large quantity of water in which it is diffused— The results of my own experiments are that where the arsenic is mixed with the acid liquid in a tube capable of holding two fluid ounces very faint and scarcely perceptible deposits begin to be formed on a glass plate with a quantity equal to the $\frac{1}{2160}$ of a grain the diffusion here being equal to two million times the weight of the poison—   With $\frac{1}{1080}$ grain in the same quantity of water the arsenic forming therefore one millionth part slight brown annular stains were procured. x x x x x x x x x— With $\frac{1}{720}$ grain the arsenic being in the proportion of about ^1/800,000 of the liquid the stains were much more decided but quite imponderable— With $\frac{1}{100}$ grain in one fluid oz of water, ^1/48,000 part, and $\frac{1}{67}$ grain in 2 fluid ounces, ^1/64,800, the deposits on glass were decided and characteristic—¹⁰

”Delicacy of Reinsch’s process—¹¹ This test failed to detect $\frac{1}{4000}$ grain in 30 drops of water the dilution being equal to 120,000 times the weight of arsenic— The deposit on copper commenced with a violet coloured film—when the quantity of arsenious acid was equal to $\frac{1}{3000}$ grain in 30 drops of water or under a dilution of 90,000 times its own weight    It was also very decided with $\frac{1}{2000}$ grain in the same quantity of water—x x x x x x— the copper was coated in a few seconds when boiled in a solution containing $\frac{1}{4000}$ grain in ten drops of water though the test had failed to detect the same weight of arsenic in three times that quantity of water—¹²

“Corrosive sublimate—   The protochloride of tin added to $\frac{1}{8640}$ grain of corrosive sublimate dissolved in one drop of water in a minute tube produced a dark grey discoloration from which in twenty four hours a black substance separated—   This however was in such small quantity that it was impossible to determine whether it did or did not contain any mercury”¹³

Prussic acid— “A standard solution was made in which each grain of liquid contained $\frac{1}{4400}$ grain of anhydrous prussic acid— One grain of this very diluted liquid placed on a plate of glass—gave a decided milkiness with nitrate of silver but when the same quantity was mixed with sixty grains of water nitrate of silver had no perceptible effect shewing the powerful influence of dilution on the action of a test $\frac{1}{4400}$ grain of prussic acid being in this instance diffused through 264000 parts of water— $\frac{1}{1460}$ grain in sixty grains of water, dilution 87,600, gave with nitrate of silver a milkiness but no precipitate— With $\frac{1}{440}$ grain, dilution 26400, the milkiness produced by the test increased—& the white cyanide became flocculent—& with $\frac{1}{220}$ grain, dilution 13200, the test gave a decided precipitate”—¹⁴