Hydrochloric acid also rendered the toxic agent of the ash soluble in water, but proved unsuitable for our work, as it was found impossible to obtain neutral residues by mere evaporation on the bath. At first one of the heavy metals or members of the H_{2}S group[161] was suspected, but on passing H_{2}S into the slightly acid extract of the ash no active precipitate resulted, but the filtrate remained active.[162] A special Marsh test was, however, made for arsenic and antimony with negative results. A test for tungsten with zinc and hydrochloric acid proved negative. Members of the ammonium sulphid group were then suspected, but while ammonium hydrate alone gave a heavy white precipitate, this precipitate, as also the black one with ammonium sulphid, proved inactive save when not thoroughly freed from acid (used for solution). The action of this ammonium sulphid precipitate on rabbits was watched for sixteen days, but without result. Nevertheless, the writer still suspected some of the rare earths.[163] Sestini[164] had found that if certain plants were nourished with a solution of a beryllium salt, in the ash of these plants could be shown the presence of beryllium. Two grams of beryllium chlorid were fed in aqueous solution to a rabbit weighing 1,800.2 grams. In four days this animal lost 241 grams and died. The stomach showed the same general pallor seen in chronic locoed rabbits, but no ulcers. The tests for beryllium by Sestini's method, however, failed to show beryllium in the active loco plants examined. Thorium chlorid, cerium chlorid, and lanthanum chlorid in 2-gram doses and zirconium chlorid in 3-gram doses produced no chronic symptoms in rabbits or, in fact, any disturbance. Titanium chlorid, 2.5 grams, evaporated in the air and then fed in an emulsion to a rabbit, also proved inactive, but this inactivity may have been due to its insolubility. Thallium nitrate c. p., in aqueous solution, in 2-gram doses, killed a rabbit weighing 2,154.6 grams in two hours and fifteen minutes. The stomach in this case, while pink, was not hemorrhagic. Zirconium chlorid has an astringent taste, and if fed repeatedly will cause the metallic astringent action. On boiling an acetic acid solution of the ash with sodium acetate a precipitate formed.[165] The presence of zirconium was thus suspected and Dr. E. C. Sullivan, of the United States Geological Survey, estimated it to be present in the ash of a sample of _Aragallus lamberti_ in about 0.01 per cent zirconium oxid, with also 0.1 per cent titanium dioxid.[166] Zirconium chlorid, 3 grams, was fed in aqueous solution to a rabbit weighing 850.5 grams. This rabbit lost 96 grams in seven days, and was then fed 3 grams more of the same solution and the following day 2 grams more. It died eight days later, weighing 656 grams. The stomach and intestines were contracted, but showed no ulcers. However, 4 grams killed a rabbit in two hours and thirty-two minutes.