The absence of a terminal flower in inflorescences ('open inflorescences') is currently explained by the maintenance of putative stem-cells in the central zone (CZ) of the inflorescence meristem (IM) governed by the CLAVATA-WUSCHEL regulatory loop. Disruption of this regulatory pathway, as in Arabidopsis TERMINAL FLOWER LOCUS 1 mutants, leads to terminal flower production. However, recent studies in other taxa reveal novel mechanisms of inflorescence termination; for example, the SEPALLATA-like MADS-box floral identity gene GERBERA REGULATOR OF CAPITULUM DEVELOPMENT 2 in Gerbera excludes the retention of a CZ as an ontogenetic cause for the openness of these inflorescences. Moreover, comparative histological studies show that the retention of a CZ in the IM, mostly a feature of the 'typical open families', is absent in open inflorescences of other families. Concerning these groups, new evidence suggests that spatial constraints at the IM could play a role at the time when terminal flower production (or not) is determined. This indicates that the multiple loss and re-gain of the terminal flower in angiosperms is necessarily based on more than one ontogenetic pathway.