Pollination in the genus Arum appears to be in general a complex deceptive pollination process. The genus Arum is composed of 28 species, all belonging to the subgenus Arum, except A. pictum, the only species of the subgenus Gymnomesium, which is basal and sister to all other Arum species. The aim of this paper is to document the pollination ecology of the insular Arum pictum, the only Arum species to flower in autumn, on the island of Corsica (France). The anthesis cycle of A. pictum occurs during the day, attracting sphaerocerid flies and staphylinid beetles early in the morning and late in the afternoon of the first day. The pollen is released from the anthers early in the morning of the second day before the departure of the insects. Its thermogenic cycle matches the anthesis cycle with an original and unique, bimodal temperature pattern of the appendix (morning and afternoon), contrary to the unimodal pattern found in all other studied Arum species. Data from reproductive success and seed sets suggest that sexual reproduction in this species is limited by pollen (e.g. attracting lured insects) rather than by resources. The biology of this Western Mediterranean species appears to be different from other Western European Arum and close to some Eastern Mediterranean species. Further studies are needed to establish whether Arum pictum represents some ancestral stage or whether its peculiar biological traits are adapted to its insular distribution.

The 28 species of the genus Arum (Araceae) attract and temporarily trap insects (mainly flies, and beetles in a few cases) during a complex pollination process. At anthesis, the appendix of the inflorescence produces heat and emits a specific odour which attracts insects. The lured insects are trapped within the floral chamber when stigmas are receptive. They will be released about 24h later after pollen emission, ensuring pollen dissemination. Studies on the reproductive biology of the genus have shown some degree of variability in the pollination strategies: morphological variations, flowering and heating periods, odour types and the type of pollinating insects. Most of Arum species have never been studied in depth but data available from the literature indicate quite a high diversity of pollination strategy within this genus. Consequently, a general pollination model is not valid at the level of the whole genus. The origin of this diversity certainly results from the biogeographic history of the genus. The plants (i.e. species) have developed adaptations in response to different climatic, ecological and biotic (i.e. entomofauna) constraints (i.e. selective pressures) according to the various habitats occupied in the different regions of Europe and the Middle East. However, in the absence of phylogenetic data, it is actually impossible to determine how these different reproductive strategies have developed and evolved during the history of this genus.

Pollen-loaded insects are not obligatorily captured rapidly by odoriferous inflorescences after their escape from a ‘pollen-donor’ inflorescence, but may be caught two or three days later. In such a situation, can these insects be considered as pollinators (i.e.. pollen vectors) or just visitors? Our results confirm that pollen grains in both species Arum italicum and A. maculatum quickly lose their viability. In natural conditions, pollen must then be dispersed quickly between male phase and female phase inflorescences in order for the pollination to be efficient. In fact, it should happen during the first hours after female Psychoda are liberated by male phase inflorescences. This is because pollinators captured on subsequent days would most probably carry non-viable pollen and thus would not pollinate the inflorescence they visit. In natural conditions, pollen grains were viable for two days. By contrast, refrigerated pollen was viable for a longer time (4-5 days). Thus refrigeration at 8 or 15oC appears to be a good method to store pollen and prolong its viability.