Do widely-spaced trees in agroforestry plantations grow satisfactorily ?

Would wide-spaced trees in agroforestry plantations grow correctly ? The question is crucial to the agroforestry issue, and very often questioned. In forestry plantations, high initial stocking rates compensate for the loss of trees. In agroforestry plantations, almost every single planted tree has to be a winner. We present the growth results of most experimental French agroforestry plantations aged 5 or more, located on public or private properties, and in contrasted climates : Mediterranean, dry central temperate plains, wet central mountains and cold oceanic areas. In most of these plantations, trees have been protected by a tree shelter (TubexTM), to ensure protection against cattle or sheep in pastures or to facilitate its maintenance in tree-crops associations. Diameter and height growths of the trees have been measured for 5 to 8 years. When available, the growth of the trees in the agroforest is compared with a forestry control treatment. A quality assessment of the stem is necessary to complete the diagnosis. In most comparative plots, trees in agroforests grow satisfactorily, and often faster than control forestry trees. The early impact of the treeshelter may biased the comparison, when only height increments are measured. After a fast height growth inside the shelter, the trees need often 2 years to recover a more stable shape and resume a normal height growth. During this emergence crisis, the trees are very sensitive to water or nutrients competition from the sward or the crop. Some accidents were observed, mainly due to cattle. The stability of sheltered trees is still a concern in some agroforests. In some experimental plots, fast growth patterns are observed, with annual height increments higher than 1 m, and diameter increments of 2 cm or more. In such places, the final harvest could take place in about 40 years. Most basal boles are however not yet fully formed. Some agroforests were not successful, and the reasons behind the failure are discussed. From these early results we can deduce some rules for designing new agroforests, regarding tree stocking, beating up policy, the use of tree shelters and the intensive care required by trees in an agroforest.; Cultiver des arbres forestiers à large espacement avec des cultures intercalaires présente certaines difficultés. La croissance de ces arbres risque d'être perturbée par la compétition des cultures associées, par l'absence d'ambiance forestière, par les dégâts dus aux animaux ou aux machines agricoles. De plus, les densités utilisées ne permettent pas de tolérer des taux d'échec élevés, si l'on veut parvenir à un peuplement adulte de densité suffisante. Les parcelles agroforestières installées à partir de 1988 en France commencent à permettre d'évaluer la réussite de telles plantations. Elles donnent des résultats de croissance individuelle des arbres très satisfaisants, et en général meilleurs que dans les parcelles forestières témoins limitrophes. On n'observe pas de différence de croissance pour des densités de 100 à 400 tiges par hectare. Compte tenu des difficultés de formation de la bille de pied par taille, ces résultats permettent de recommander des densités de plantation en agroforesterie d'environ deux fois la densité finale du peuplement que l'on souhaite constituer.