17=
days and repeatedly examined in ultraviolet light (360 mp). Simultaneously an ordinary silica gel plate and an alumina plate were treated in the same way. On the two latter plates, after a relatively short time (I-Z h), the spots darkened and the fluores-. cence weakened (anthracene and pyrene) or almost disappeared (3: 4-benzopyrene, perylene, I : Iz-benzoperylene and chrysene) , while on the caffeine-impregnated plate they were completely unchanged after four days. The fluorescent spots developed from a coal tar sample on each of the three plates behaved in a similar manner to the corresponding spots of the pure compounds. Examination of coal tar. The tar was an untreated sample. A solution was made up of 59 mg of the tar in I ml of benzene and a total of So t_cl was spotted on the plate in ten spots. On another plate a single spot of the tar was run together with reference substances. From the positions and fluorescence colours of the spots after five runs, and from absorption spectra of the extracted spots the following hydrocarbons were identified (figures in brackets state amounts of the hydrocarbons found, as per' cent of weight of tar) : anthracene (0.21), pyrene (o.#),, 3: &benzopyrene (o.Ig), perylene (0.03) and I : rz-benzoperylene (0.12).