From 18 to 22 June this year, around 120 fans of floral biology met at Hyères-les-Palmiers on the French Côte d'Azur for the latest in a series of two-yearly workshops on 'Molecular Mechanisms Controlling Flower Development'. These international meetings usually take place somewhere on the Mediterranean coast, and Hyères provided an idyllic setting in which to meet up with colleagues and discuss science on the beach as well as in the lecture hall or poster room. One flowering plant that really did return to the sea, Posidonia oceanica, was to be found growing in the shallow coastal waters surrounding the conference venue, and many delegates could be seen splashing about and (no doubt) thoroughly investigating this marine monocot between scientific sessions. The workshop, of which details can be found at: http:// www.ens-lyon.fr/RDP/FlowerWorkshop2019/, was divided into seven oral sessions and formatted so that a maximum number of speakers could present their data in 20 min highlighted talks, 15 min standard talks, or 4 min 'flash posters'. In all, 74 oral presentations were made, involving nearly two-thirds of delegates, who came mostly from Europe, but also from as far afield as China, Japan, Vietnam, Mexico, and Washington State. It is impossible in a short review to cover all the subjects addressed at the meeting, so the following is a personal view of some of the highlights and novel themes to emerge. The workshop got underway with its traditional starting topic of 'Flowering', namely the complex and interconnecting networks used by flowering plants to initiate their reproductive phase in response to numerous environmental and endogenous cues. Much of what is known about flowering comes from the model plant Arabidopsis thaliana, which is a typical annual species that flowers once in its life cycle, promptly sets seed, and dies. It has been known for several years that multiple cues for flowering in Arabidopsis converge on the FLOWERING TIME (FT) protein, which moves from its site of synthesis in the leaves to the shoot apex where flower production is initiated. However, two recent research trends were evident from this year's workshop. First, several mobile signals other than FT are coming to the fore, including a potentially large list of metabolites that can promote flowering, as highlighted in a talk by Reyes Benlloch (IBMCP, Valencia, Spain). Secondly, several research groups are now focusing on alternative models with contrasting life cycles and flowering requirements. For example, the work of George Coupland's laboratory (MPI-Cologne, Germany) demonstrates that endogenous cues that depend on the age of the plant are of particular importance to flowering in perennial Brassicaceae (Hyun et al., 2019). Indeed, although the workshop included a devoted 'evo-devo' session, the prominent presence of new and emerging models (Fig. 1) in all seven scientific sessions meant that evolutionary considerations were discussed throughout the meeting. Many species require vernalization: cold treatment as a necessary prelude to flowering. However, the need for vernalization can cause unwanted delays in plant breeding programs, which is why Richard Immink's group (Wageningen University, The Netherlands) recently made a screen of some 9000 chemical structures to find a compound that could bypass the vernal-ization requirement (Fiers et al., 2017). Two compounds were found to have the desired effect in Arabidopsis, one of which might prove of practical use in crop species. Interestingly, this compound did not appear to act via any of the known dedicated components of the vernalization pathway, but instead induced FT by a novel mechanism, possibly involving the production of hydrogen peroxide. After deciding to flower, a plant must then determine how its flowers should be arranged, and so the workshop turned its attention to the subject of inflorescence architecture. The enlargement of model species from Arabidopsis was clearly apparent in this theme too, with several impressive talks featuring cereals, legumes, Solanaceae, and Asteraceae. One particularly interesting model for inflorescence architecture is cultivated rice, which has undergone the domestication process twice from distinct wild species in Asia and Africa. These domes-tication events appear to have resulted in parallel changes to the rice panicle, and a collaboration involving Stefan Jouannic and Helene Adam (IRD-Montpellier, France) and Ngang Giang Khong (Agricultural Genetic Institute, Hanoi, Vietnam) is taking a multidisciplinary approach to discover whether similar genetic targets and developmental processes were