Recent advances in our knowledge of grass phylogeny suggest the need for critical reevaluation of certain morphological structures and their homologies within the family. Particularly in the woody bamboos, inflorescence morphology has been regarded as complex and difficult to compare with that of other grasses. Application of the unified concepts of Troll, Weberling and others provides a firm basis for the assessment of inflorescence homology among bamboos, and ultimately between bamboos and other grasses. The spikelet, with its series of overlapping, distichous bracts some of which bear flowers in their axils, is regarded as a simple, bracteate, polytelic, spicate inflorescence. The stalk of the spikelet is therefore a peduncle, and the aggregation of spikelets is termed a synflorescence. Thus there are no true panicles, racemes, or spikes in the traditional grass ‘inflorescence.’ The terminal spikelet of the main axis of the synflorescence is the main florescence, and lateral axes comprise paraclades and coflorescences. Pseudospikelets, with one or more gemmiparous bracts present at the base of the spikelet proper, are correlated with the presence of a subtending bract and associated prophyll at each axis within the synflorescence. The presence of pseudospikelets and associated bracts on the one hand, and true spikelets with no associated bracts on the other hand, are viewed as representing two extremes of a spectrum of morphological variation, with intermediate states represented by certain taxa. Phylogenies derived from molecular data suggest that either spikelets or pseudospikelets, or possibly both, had multiple origins within the woody bamboos.

Key words: bamboo, inflorescence, phylogeny, Poaceae