Background and Aims. The presence of novel structures in orchid flowers, including auricles, rostellum and bursicles on the gynostemium and a lobed labellum, has prompted long-standing homology disputes, fuelled by conflicting evidence from a wide range of sources. Re-assessment of this debate using an improved model is timely, following recent phylogenetic insights and on the cusp of a revolution in developmental genetics. Methods. We use new data from floral development and anatomy in the small-flowered terrestrial orchid Herminium monorchis as a model to explore organ homologies in orchid flowers within the context of a review of recent literature on developmental genetics. Key Results. The apex of the median carpel of Herminium is trilobed, and the bursicles develop from its lateral lobes, relatively late in flower ontogeny. The bursicles enclose the viscidia, which adhere to the tapetal remnants to form a caudicle linking the viscidium with the pollinium. The auricles are initiated earlier than the bursicles, but they also remain unvascularized. The deeply trilobed labellum possesses three vascular traces, in contrast with the lateral petals, each of which contains a single vascular trace. The two lateral labellum traces diverge from the traces supplying the two adjacent lateral sepals. Data from flower ontogeny and anatomy conflict with respect to organ homologies. Conclusions. Much progress has recently been made in understanding the exceptional differentiation shown by orchids among perianth segments, focusing on multiple copies of the DEF/AP3 subclass of B-class MADS-box genes. In contrast, untangling homologies of profound congenital union of multiple floral organs forming the orchid gynostemium is hampered by their profound congenital union, which we ascribe to overlap in gene expression between organs. Thus, the functional morphology of the orchid flower could ultimately reflect extreme synorganization and associated genetic integration. Analogizing the deeply lobed orchid labellum with a compound leaf, we speculate that KNOX genes could be implicated not only in their demonstrated role in spur development but also in the development of both the characteristic lobed morphology of the orchid labellum and the lobing of the median carpel that differentiates the bursicles and rostellum.