We investigated seed dormancy among species of Melastomataceae from Neotropical montane vegetation of Brazil. Four out of 50 studied species had dormant seeds:Miconia corallina(Miconieae), Tibouchina cardinalis(Melastomeae), Comolia sertularia(Melastomeae) andChaetostoma armatum(Microlicieae). For these four species, germinability of seeds collected in different years was always < 10% and the percentages of embryoless seeds and non-viable embryos were both insufficient to explain low or null germinability. This is the first unequivocal report of seed dormancy in tropical Melastomataceae. The production of seeds with permeable seed coats and fully developed, differentiated embryos indicates the occurrence of physiological dormancy. The reconstructed phylogenetic tree of the 50 species suggests that physiological dormancy evolved multiple times during the evolutionary history of Melastomataceae in this vegetation. Physiological dormancy evolved in species and populations associated with xeric microhabitats, where seeds are dispersed in unfavourable conditions for establishment. Therefore, drought-induced mortality may have been a strong selective pressure favouring the evolution of physiological dormancy in Melastomataceae. We argue that dormancy may have been independently selected in other lineages of Cerrado plants colonizing xeric microhabitats and dispersing seeds at the end of the rainy season. The contributions of our data to the understanding of seed dormancy in tropical montane vegetation are discussed.

A preocupação com o uso indiscriminado dos termos cipsela e aquênio para os frutos de Asteraceae tem sido freqüentemente destacada por especialistas na família. O presente trabalho foi desenvolvido com o intuito de verificar a existência de argumentos para justificar a adoção de um termo em detrimento do outro. Após análise histórica e anatômica, concluiu-se que há fundamento técnico para considerar cipsela e aquênio como diferentes tipos de frutos. Para Asteraceae, o correto é denominar cipsela, já que aquênios são formados somente a partir de ovários súperos, como em Plumbaginaceae.The worry about the indiscriminate use of the terms cypsela and achene for the fruits of Asteraceae has been frequently detached by specialists in this family. The present work was developed aiming to verify the existence of arguments to justify the adoption of a term against the other. After historical and anatomical analysis, we concluded that there is technical basis to consider cypsela and achene as different types of fruits. For Asteraceae, the correct is to call cypsela; achenes are only derived from superior ovaries, as in Plumbaginaceae

The ontogeny of cypselae and their accessory parts were examined using light and scanning electron microscopy for the species Campuloclinium macrocephalum, Chromolaena stachyophylla, Mikania micrantha, Praxelis pauciflora, Symphyopappus reticulatus, and Vittetia orbiculata, some of these being segregated from the genus Eupatorium. A layer of phytomelanin observed in the fruit appears to be secreted by the outer mesocarp into the schizogenous spaces between the outer and inner mesocarp; its thickness was observed to vary among the different species examined. The bristles of the pappus are vascularized, except in M micrantha, and have cells that are superficially projected and arranged acropetally; in S. reticulatus some of the projections are retrorse and a fracture line on the floral disk that is only seen in this species may indicate a double dispersal process. Numerous differences observed among the cypselae examined here reinforce earlier segregations of the genus Eupatorium sensu lato.