Pollinisation de l'artichaut (Cynara scolymus L.) par l'abeille domestique (Apis mellifera L.) en production de semences hybrides sous abris grillagés

Pollination of globe artichoke (Cynara scolymus L.) by honey bees (Apis mellifera L.) to produce hybrid seed under enclosure. Globe artichoke is currently multiplied mainly by vegetative means, but the planting of seeds could lead to easier cropping practices and improved yields. This is especially true with hybrid seeds which could be produced using the two genetic male-sterility systems available. For two years, we studied the effectiveness of honey bees as pollinators to produce hybrid seeds of globe artichoke under insect-proof enclosures. We used two adjacent $10 \times 6$ m tunnels covered with 1-mm mesh screen and planted with 2 male-fertile (MF) and 7 male-sterile (MS) lines (tab. I). Each tunnel was provided with a colony of 3,500 bees at the onset of MF flowering. Flowering of MS and MF lines was well synchronized (fig. 1), and the number of flower heads per plant was similar for all lines. Yet, there were over 10-fold differences in both years in the density of foragers that visited the various lines with extremes of 0.04 to 1.18 honey bees per head (fig. 2). Some foragers collected pollen, but these were rarely seen on MS heads. The ranking of MS lines in terms of forager density was similar over both seasons, which suggests that the differences in attractiveness among these lines were of genetic origin and probably resulted from differences in nectar availability or composition. Yield of achenes per plant also varied significantly among lines (fig. 3). The proportion of heads that were empty ranged from 23% to 100% among MS lines, and for each season was negatively correlated with the average forager density on each line (fig. 4). The number of achenes per head in the flower heads which were not empty did not vary significantly among MS lines (tab. II), which suggests that the pollination effectiveness of honey bee visits was similar among these different lines. This achene content was greater in MF heads, which is consistent with a greater pollination effectiveness of honey bees in this inflorescence due to the presence of pollen and some self-fertility in MF lines. The overall forager density was similar between the two seasons as was the proportion of empty MS heads, but the achene content of non-empty MS heads was significantly greater in the second year than in the first one (tab. II). This suggests that the pollination effectiveness of individual visits was greater in the second year. Pollen availability was similar in both seasons, but there were 5 times fewer foragers collecting pollen in the second year compared to the first one. The pool of pollen is limited when producing hybrid seed, especially in confined environments, and our results suggest that in these situations the level of pollen collection may negatively affect the pollination activity of honey bee colonies.