[Illustration: THE DETECTIVE--NEW STYLE--IN THE LABORATORY.] Probably, in most cases, the ordinary medical attendant is able to tell whether a person is dying a natural death or is being carried off by some deadly drug. His position, however, is not a pleasant one. It is impossible to be certain; and, in order to make a full investigation, he must suggest either that the victim is committing suicide, or that someone else, perhaps his wife or son, is committing murder. And, after all, the signs in the living are very obscure. Of course, if a person is foolish enough (as many are) to drink sulphuric or nitric acid, his mouth and throat are burned as if he swallowed coals of fire, the former leaving black and the latter yellow stains; but when the poison is arsenic, or opium, or strychnine, the symptoms are very like those of certain diseases. When the cholera was last in London, a father, mother, son, and daughter dined together. Immediately after dinner, all, except the son, became suddenly ill, and died in a few hours, with the symptoms of arsenic poisoning. The son, who was always quarrelling with the rest of the family, was arrested on the doctor's report and charged with murder. But a _post-mortem_ examination showed that cholera was the real cause of death. Apoplexy, in the same way, is very like opium poisoning; and hydrophobia, lock-jaw, and even some cases of hysteria, closely resemble poisoning by strychnine. Still, when a healthy man grows suddenly ill soon after a meal, the doctor keeps his eyes open, and if death follows he has a pretty shrewd idea of what caused it. At all events, he feels perfectly justified in assuming that the case is not a normal one. He therefore hands over to the analyst the jars and other receptacles containing the portions of the subject's body likely to bear traces of the poison, knowing full well that if any poison is there the analyst will infallibly detect it. The analyst begins by making a series of what may be called "brews," mincing, pounding, boiling, cooling, filtering, decanting, and distilling, over and over again. In these operations various solvents are used in succession, plain water separating out one class of poisons, alcohol dissolving out another group, benzol taking up a third, naphtha a fourth, ammonia a fifth, and so on. This preliminary work takes, not hours, but days to perform. At an early stage in it the operator discovers such volatile poisons as prussic acid, chloroform, carbolic acid, and phosphorus, if any of them be present. Later on he comes across the alkaloids, such as strychnine, digitalin, cantharidin, and other terrible poisons of that class. Finally, the residue of the animal matter with which we have supposed the medical detective to be experimenting is mixed with hydrochloric acid, and distilled once again, after which it can contain no poison except one of the metals.