Extraction of the leaves of Eremophila glabra R.Br. collected from Rottnest I. consistently afforded a colourless crystalline compound, C₂₀H₂₆O₈, which formed a dibromide, C₂₀H₂₄O₈Br₂. Only a trace of the same substance was obtained from the plant collected on the sand dunes of Rockingham, south of Perth. Six of the oxygen atoms were identified as methoxyl groups and the compound showed no infrared absorption corresponding to the presence of hydroxyl or carbonyl groups. Hence, it was considered likely that the remaining two oxygen atoms were attached by cyclic ether linkages, probably in a lignan of the 3,7-dioxabicyclo[3.3.0]octane type. This has been confirmed by the well-established degradation technique used on this group of compounds, namely, the action of nitric acid on the dibromides (Erdtman and Gripenberg 1947; Hearon and McGregor 1955). The product in this case, 4-bromo-5,6-dinitropyrogallol trimethyl ether, indicated the structure (I) for the compound C₂₀H₂₆O₈. The presence of benzyl ether linkages was demonstrated by hydrogenolysis, with sodium in ethanol, which afforded the diol (II). The loss of the central methoxyl groups would be expected for a pyrogallol trimethyl ether under these conditions as illustrated by the behaviour of elemicin (Semmler 1908), pyrogallol trimethyl ether (Thoms and Siebeling 1911) and 5-propylpyrogallol trimethyl ether (Sonn and Scheffler 1924). Permanganate oxidation of (II) yielded the expected 3,5-dimethoxybenzoic acid. At the time the structure of the lignan became clear, Dickey (1958) described the isolation of a diglucoside, liriodendrin, from Liriodendron tulipifera L. Dependent on the methods used, three isomeric aglycones lirioresinol A, B, and C were obtained by hydrolysis of the glucoside. Dickey concluded on the basis of infrared comparison with synthetic (±)-syringaresinol (Freudenberg and Dietrich 1953) that the aglycones were stereoisomeric with the synthetic compound. Further degradation was shown to be analogous to that of syringaresinol although comparison with authentic samples was not made. Since then Pearl, Beyer, and Dickey (1958) have isolated from the spent sulphite liquor of aspen, three compounds of which one appears to be a further isomer of lirioresinol. The identity of our material with lirioresinol-B dimethyl ether has been established by comparison with an authentic sample provided by Dickey. The cis-fusion of the tetrahydrofuran rings is established by the optical activity of the diol (II). The alternative trans-fusion would be expected to lead to a meso-diol. Equilibration under the hydrogenolysis conditions would lead to racemization.