To obtain supplies of the alkaloid, platydesmine, which were required for further study of its conversion into dictamnine, the leaves of Geijera salicifolia Schott., the best recorded source of the former alkaloid, were extracted. The results of Johns and Lamberton were duplicated but in addition other fractions of the extract were examined. The "acidic" fraction yielded umbelliferone but the "phenolic" fraction, which presumably contained zanthoxylin which has been recorded as a constituent of the essential oil, was accidently lost. The "neutral" fraction gave, in addition to β-sitosterol, two other products. The first of these was identified as the complex coumarin, geiparvarin, whose isolation from the leaves of G. parviflora Lindl, and structure determination have been described by Lahey and co-workers. The second product, which was beautifully crystalline, melted fairly sharply at 120-121° and with several solvent systems on thin-layer chromatography plates behaved as a single substance. However, gas chromatography demonstrated clearly that it was a mixture of two components, and n.m.r. spectroscopy allowed them to be identified as the coumarins, geijerin and dehydrogeijerin (approximately 1 : 2). Geijerin (m.p. 121°) has been isolated from the bark of G. salicifolia by Lahey and Wluka, and dehydrogeijerin (m.p. 130-131°) accompanies geiparvarin in the leaves of G. parviflora. Since the chemistry of these substances has been thoroughly explored, no further work was done on the mixture.