Abstract: Swingle^[14-15] divided Aurantioideac into two tribes, one of which, Clausereae was further divided into three subtribes, namely, Micromelinae, Clauseninae and Merrilliinae. Mic romelinae and Merrilliinae each have one genus, whereas Clauseninae has three genera. Morphologically, the Clauseneae is a natural tribe, the five genera are related in a linear sequence, starting with Micromelum as the most primitive and progressing in sequence to Glycosmis, Clausena, Murraya and Merrillia. Chemical studies also support this linear relationship, as revealed by the degree of oxygenation and complexity of the 3-methyl carbazole alkaloids, from CH₃ and C1₃ in Glycosmis to CHO and C₁₈ in Clausena and COOH and C₂₃ in Murraya^[20-21]. Distribution of flavonoids also indicates the progression from Clauseninae to Merrilliinae^[18-19]. Extensive work has been conducted on the chemistry and taxonomy of the genus Murraya ^{[1,5,7-13,16]}, and the data from these studies clearly indicate the presence of two distinct groups. Based on a combination of morphological and chemical differences, we agreed with Tanaka^[16-17] in dividing Murraya into two sections, i.e. section Murraya and section Bergera^[1]. However, our previous study^[1] has not touched on the relationship between the two sections. Tanaka^[16-17] placed section Bergera before section Murraya, and indicated that the former is close to Micromelum and the latter to Merrillia. Swingle^[14-15], on the other hand, put taxa of section Murraya ahead of those of section Bergera, presumably suggesting that plants of section Murraya are more primitive than those of section Bergera, this arrangement was followed by Huang^[2-3]. The two conflicting viewpoints would have direct bearings on the interpretation of the trends of biogenesis of prenylated indole and carbazole alkaloids, as well as on the weighing of the relative advancement of the morphological characters within the genus, such as in the assignment of indices of divergence and in the construction of Wagner Divergence Diagrams. Without more objective criteria, we find it difficult to select one of the two systems. In order to determine the relationship between the two sections of Murraya, we decided to study plants of related genera, with the hope that the chemical data may shed light on the problem. A plant that attracted our attention is Merrillia caloxylon (Ridley) Swingle. So far, only eupatorin and a few other flavonoids have been reported from the fruit of this species^[4,6]. Although Me. caloxylon belongs to Merrilliinae, a subtribe next to Clauseninae, Tanaka^[16-17] believes that it is close to Murraya section Murraya. Swingle^[15], also suggested that this species might have developed from the same stock that gave rise to Mu. paniculata. If their interpretations were accurate, we would expect that Me. caloxylon would also contain yuehchukene and 8-prenylated coumarins. The presence of the antiimplantation agent would not only open up a new source of the compound but also help us judge the relationship between the two sections of Murraya. It is in this context that we studied the chemical composition of Me. caloxylon. Indeed, root and stem bark of Me. caloxylon were found to contain the antiimplantation indole alkaloid yuehchukene (1), and the 8-prenylated coumarins sibiricin (II) and phebalosin (III), as well as 3-(3-methy1-buta-1,3-diene) indole (IV) and eupatorin (V.) Details on the chemical profiles are reported in another paper. Through this exercise, we have confirmed the close relationship between Merrilliinae and Murraya section Murraya, plants of both taxa contain yuehchukene and 8-prenylated coumarins, but no carbazole alkaloid. Root and stem bark of Me. caloxylon, like those of plants of section Murraya, are strawcolored to pale whitish. Its leaves also bear wings along the rachis an in Mu. alata, and the seeds are also villous. However, Me. caloxylon has long trumpetshaped flowers 55-60 mm long, much larger than those found in other rutaceous plants. Its fruit is ob long, up to 11 cm long and 8 cm across, bearing a thick and warty pericarp, exuding a very stick mucilage when cut, and containing numerous seeds (>30). The plant was known to exist in the Malay Peninsula and north Sumatra^[15], but, according to David Jones (per. comm.) of the University of Malaya, is now only available in cultivation in Malaysia and Singapore, a limitation to any further exploitation as an additional source of yuehchukene. Besides confirming the close relationship between Merrillia and section Murraya, we may also conclude that section Bergera is close to Glycosmis and Clausena, since they are known to contain carbazole alkaloids but no yuehchukene. Accordingly, we find Tanaka’s arrangement more acceptable: plants of section Bergera are more primitive than those of section Murraya, the former is close to Clausena whereas the latter (notably Mu. alata) to Merrillia. The relationship among the gonera with in Clauseneae may be illustrated as follow: Micromelum→ Glycosmis→Clausena→Murraya sect. Bergera→Murraya sect. Murraya→Marrillia. Accordingly, we may decide that the following character states are more primitive among plants of Murraya and Merrillia: root and stem bark dark brown, leaf rachis wingless, flower small, fruit purple-black with few seeds, and seed coat glabrous. In contrast, strawcolored or pale whitish bark, winged leaf rachis, large flower, red or yellow fruit with many seeds and villous seed coat can be regarded as more advanced characters. Acknowledgments Partial support was received from the World Health Organization Special Programme on Human Reproduction and the Kevin Hsu Research Fund (to YCK) and Commonwealth Science Council (to PPHB). The staff of the Singapore Botanic Gardens and the Forest Research Institute of Malaysia are thanked for their assistance in collecting plantmaterial.

Key words: Chemotaxonomy, yuehchukene, 8-prenylated coumarin, Clauseneae, Murraya