In order to follow a systematic method in researches of this nature, it is advisable to divide the materials under examination into three parts: one portion is preserved, in order to ascertain its physiological effects on animals, the chemical analysis having failed to give positive results. The other portions are submitted to analysis, but with slightly different objects in view; one is subjected to a series of tests which are adapted, under all circumstances, to place the chemist on the track of the poison present, and which, in some cases, may even give conclusive and definite results. Should these tests furnish only _indications_ of the nature of the poison, the remaining portion serves, with the assistance of this information, to establish beyond doubt the identity of the substance. INDICATIVE TESTS. Two cases may present themselves: the materials to be examined possess either an alkaline (or neutral) or an acid reaction. As the methods to be pursued in either of these cases differ somewhat, they will be treated separately. THE SUBSTANCE POSSESSES AN ACID REACTION. The materials are mixed with water, placed in a retort provided with a delivery-tube which dips in a solution of nitrate of silver, and heated over a water-bath: if a _cyanide_ be present, hydrocyanic acid will be disengaged, and a white precipitate of cyanide of silver formed: this is examined as previously directed (_vide_ p. 50). In case a precipitate is not produced by the above treatment, more water is added to the retort, and the mixture boiled for about an hour, care being taken to collect the evolved vapors in a well-cooled receiver. The portion remaining in the retort is thrown on a filter and the filtrate obtained united with the distillate. The residue remaining on the filter is next washed with boiling absolute alcohol, the washings being added to the aqueous solution. In this way, the suspected substances are divided into soluble and insoluble portions, which are examined separately, as directed below. _a._ LIQUID PORTION. If the addition of alcohol caused a precipitation of animal matters, these are separated by filtering the solution. The filtrate is then placed under a bell-jar over concentrated sulphuric acid until its volume is considerably reduced. The solution may contain organic and inorganic bases and acids. In order to detect all bodies that are present, the following course is pursued: