The architectural spectrum of the rain-forest appears to be qualitatively the same in New Guinea as previously reported in tropical Africa or America. However, the study of Asiatic and Melanesian trees leads to a broader understanding of the architectural range in pantropical families of flowering plants, and suggests an approach to the understanding of many problems in taxonomy, phylogeny, biogeography, and forest ecology. THE ARCHITECTURE OF TREES in the tropics, a new subject for botanical research, has already been applied to taxonomy, as, for example, in the Euphorbiaceae (Halle 1971). In the present paper, I wish to introduce another aspect of tree architecture, which is biogeographical rather than taxonomical. The introduction of the topic of tree architecture by Halle and Oldeman (1970) was mainly founded upon African and American records, and was comparatively poor in original data about Asiatic trees. Taking into account the particular richness of the Asiatic flora, and its well-known importance for tropical botany, filling such a gap was a preliminary essential for the satisfactory formulation of the problem of architectural evolution in trees. A visit to Australian New Guinea in 1972 provided an opportunity to make a first study of the architecture of Asiatic and Melanesian trees, the main results of which are summarized below.' All the architectural models previously discovered in Africa and America were found easily in the lowland rain forest around Lae, and in the submontane rain forest around Bulolo, Morobe District. A significant, if negative, result is that-contrary to a priori views-no new architectural models were found during my visit to Asia. Thus, as far as I know, the architectural spectrum of the rain forest appears to be qualitatively the same in all parts of the tropics. Of course, the possibility of quantitative variation within the spectrum cannot be denied, some models being favored in some countries either for historical reasons or by environmental conditions, but this supposition is speculative, as a study which would reveal this still remains to be done. Even if the total number of known models has not been augmented, this study of Asian trees leads to a better understanding of the architectural range in several pantropical families of flowering plants. From the 76 New Guinea species from 45 families whose form was studied, I extract 11 which were observed in particularly good circumstances. Schuurmansia henizgsii K. Schum. OCHNACEAE Cf. NGF 40163 from Red Hill, near Lae, Morobe District. Leeuwenberg's Model. This very common small tree, usually less than 10 meters high, is one of the finest plants around Lae at flowering time. Figure 1 a and b represents a well-grown flowering specimen. All the serial axes are orthotropic and equivalent to one another, the only variation being a quantitative one such that, when the branching increases, the size of the successive axes, and the length of their leaves, progressively decreases, as advanced by Corner (1949) in his Durian Theory. Apical inflorescences appear simultaneously on all the upper axes, the lower ones remaining sterile from then on. Owing to these architectural features, Schuurmansia heningsii belongs to Leeuwenberg's model, as defined by Halle and Oldeman (1970). Apart from the characteristics of the model itself, Schuurmansia heningsii offers some individual morphogenetic features, such as the stilt-roots supporting the seedling axis, and the periodic growth which is indicated by the series of bud-scale scars at each level of temporarily interrupted growth. This is the first record of Leeuwenberg's model in Ochnaceae. However, in the same family, the African Lophi'ra alata Banks possesses an architecture related to the present one, although its aerial axes, after the first growth, are no longer equivalent to one another, since some of them grow upright to form a sympodial trunk (Kwan Koriba's model). 1 I wish to express my gratitude to John S. Womersley, head of the Division of Botany at Lae, to the members of his staff, particularly Mark J. E. Coode, for their help in my field-work, and to Anna Cook and Bokinane, for their assistance. BIOTROPICA 6(1): 43-5