In plant communities, the roots of different species are linked by an extraradical network of hyphae of arbuscular mycorrhizal (AM) fungi. The existence of such network has major ecological consequences since it has been assumed that minerals and carbon (C) can be redistributed within the network and hypothetically between plants. The possible C-transfer between autotrophic plants via the AM fungal network has been abundantly investigated but the results still remain equivocal. In the present study, we investigated how these AM fungal networks are constructed and studied the C-transfer through these networks, from an autotrophic plant (1) to another autotrophic plant, growing under decreased light conditions, or (2) to a seedling. We demonstrated that strains belonging to the Glomeraceae and the Gigasporaceae markedly differed in their capacity to interconnect AM fungal networks from the same isolate by means of anastomoses. While the Glomeraceae strains were able to do so, such process was never observed with the Gigasporaceae. The adequate culture conditions for the C-transfer study were obtained by associating an autotrophic plant to an AM fungus under in vitro conditions. This system, first developed for a single plant with Solanum tuberosum, was extended to two plants of Medicago truncatula, physically separated in a bi-compartmented Petri plate, but with their root systems interconnected by a network of Glomus intraradices. Following the labeling of the donor plants with 13CO2, we observed in both experiments the C-transfer from the donor to the receiver plants via the AM fungal network. However, the transferred C remained in the receiver's roots. Fatty acid analysis in the receiver's and control's roots of the seedlings further indicated that 13C enrichment of the root cells did not result from a direct transfer of C from the AM fungus to the root cells, but more probably from indirect pathways. From these experiments, it is tempting to conclude that C cannot be transferred from a M. truncatula donor plant into the cells of a M. truncatula receiver plant. However, our observations do not preclude the possibility that such C-transfer via AM fungal networks exists with other plant-fungus combinations. Research on this issue should therefore be focused on stronger C-sinks, such as myco-heterotrophic plants.