Notwithstanding the confidence with which _Dr. Bostock_ has supported the pretensions of the _Sulphate of Copper_, as an infallible test for arsenic, its validity has been lately called in question, and it has been stated that a _decoction of onions_ has the property of imparting to the copper precipitate, produced by a fixed alkali, a green colour and appearance completely analogous to that which is occasioned by the presence of arsenic. This opinion was boldly advanced, and supported, on the trial of _Donnall_, before alluded to, and of which we have given a very ample report in the _Appendix_. Since this event an opportunity occurred which enabled the author to examine this alleged fact, by a fair and appropriate series of experiments,[256] the result of which has satisfactorily proved that the opinion was grounded on an optical fallacy, arising from the _blue_ precipitate assuming a _green_ colour, in consequence of having been viewed through a yellow medium.[257] The phosphoric salts may also, under similar circumstances, be mistaken for arsenic; for the intense blue colour of the _phosphate of copper_ will, when viewed through a yellow medium, necessarily appear green. Such instances of optical fallacy are by no means uncommon in the history of chemical reagents; thus _corrosive sublimate_ has been said to possess alkaline characters, in consequence of appearing to turn the syrup of violets green, whereas this apparent change is to be solely attributed to the optical combination of the yellow hue of the sublimate with the blue colour of the violet. Whenever therefore such a source of fallacy can be suspected, the operator would do well to repeat his experiment on white paper, in the manner we have already pointed out, when treating of the silver test; and let it be remembered that the results, when obtained in glasses, should always be examined by day light, and viewed by reflected, and not by transmitted light. _Dr. Bostock_ observes, that a weak solution of the sulphate of copper, without any addition, when held between the eye and the window, frequently presents a greenish tinge. It should be also known that the usual reaction of the _ammoniuret of copper_, upon a diluted solution of arsenic, is prevented by the presence of _tannin_; strong tea may therefore render the test inefficient.