Résumé
Triclonal seed gardens of Coffea canephora were planted in the Cote d'Ivoire from 1972 to 1976, in order to produce hybrid varieties, as a complement of clonal varieties used by extension services.
Four trials, planted between 1975 and 1983, show that the half-sister progenies produced in those seed gardens are very inferior to the selected clones used as controls: the production of most of them was less than 50 % of the clones'. However, when re-analysing the data from former hybrid trials (progenies of controlled crosses), one can see that, in the nineteen combinations produced in the seed gardens, eight have a productivity similar to the clones.
Thanks to the possibility of using enzymatic markers, the genetic origin of plants from half-sister progenies issued from one seed garden was analysed.
The results show that the low value of those progenies is due, in those triclonal seed gardens, to the production of a high ratio of seeds from less productive hybrid combinations.
Differences between flowering dates of the mother clones, that led to the choice of the triclonal design, are in a large part responsible for this.
This led to the use of a biclonal design producing only one selected hybrid, even though there may be a risk of under pollination.