There is great diversity in inflorescence architecture within grasses, yet patterns of inflorescence evolution are difficult to establish by morphological analysis alone. A more comprehensive approach is to couple morphological description with comparative studies of inflorescence development and the optimization of morphological and developmental characters onto robust phylogenies. We have used the GPWG (2001) phylogeny of the grasses, together with finer-scale phylogenetic analyses in the panicoid grasses to investigate inflorescence evolution. We have found that differences in the timing and position of initiation of primordia at the apical meristem can produce distinct inflorescence morphologies amongst clades. For example, members of the Triticeae (including wheat) have distichously arranged inflorescence branches, while maize and other panicoid grasses have polystichously arranged inflorescence branches. On a finer phylogenetic scale, differences between closely related genera and species often involve developmental shifts in 1) the numbers of axes initiated, 2) the numbers of orders of axes initiated, and 3) the extent to which axes elongate. In the panicoid taxa Setaria, Pennisetum and Cenchrus, we are investigating the genetic control of inflorescence branching using quantitative trait loci analysis of a cross between two species of Setaria. The QTL analysis shows that there are several QTL involved in controlling differences in inflorescence architecture, and that some of these correlate to inflorescence development genes in maize.

Key words: Cenchrus, grasses, inflorescence development and evolution, Pennisetum, QTL, Setaria