Testa and seed appendage morphology in Aeschynanthus (Gesneriaceae): phytogeographical patterns and taxonomic implications

The current sectional classification of the genus Aeschynanthus Jack, essentially based on seed morphology, presents some problems of species placement. A comparative SEM survey of seed and seed appendages was undertaken in order to assess the value of this classification. Seeds of 99 taxa (that is about two-thirds of the estimated total) were examined and found to fall into two types, A and B. Type A has spiral testa cell orientation, papillae formed from a single cell and short smooth appendages. Type B is recognized by the straight orientation of the testa cells, combined with the presence of papillae formed from the raised ends of two adjacent cells on the long hair-like appendages and usually on the testa. Only six of the investigated species did not fall into either category. Three have straight testa cell orientation combined with single-cell papillae and short smooth appendages; the papillae and appendage characters place them in type A. Three have spiral testa cell orientation and short smooth appendages but the testa cells have slightly raised ends; these are also placed in Type A. The three subtypes in Type A are equivalent to the sections Haplotrichium s.s., Microtrichium and Aeschynanthus, but the divisions are less clear than those within Type B. However, other morphological characters support sectional separation. Type B subdivides into three: two subtypes equivalent to sections Polytrichium and Diplotrichium, and a third encompassing section Xanthanthos together with part of the current sect. Haplotrichium, and here referred to as sect. X. There is sufficient morphological correlation with seed type to make the sectional position of many species clear without recourse to seed, particularly in sects Polytrichium, Diplotrichium, Haplotrichium S.S. and Aeschynanthus. There is strong correlation between seed type and geographical distribution. Sects. Microtrichium and Aeschynanthus, with Type A seed, are essentially Malesian. Groups with Type B seed are largely confined to mainland south and south-east Asia, except for sect. Polytrichium which is more widespread, possibly due to the greater effectiveness of a coma of hairs in wind dispersal. It is suggested that Type A seed, probably sect. Microtrichium, is the least determined and Type B sect. Polytrichium the most derived seed type. Based on these findings a revised key to the sections is provided.