Planted or naturally regenerated young beech trees exibit form defects. Two causes are studied in this paper : light as an environmental factor and genetic variability. A few definitions first (fig. 1). Short shoots have 3 to 5 internodes, they never branch, but carry 77 p. 10 of the total foliar area. They are called exploitation shoots. Long shoots have 6 to 10 internodes, they may branch. They represent 63 p. 100 of shoot dry weight. They are called exploration shoots. Only exploration shoots may be polycyclic, i.e. have several growth waves in one growing season. If apical dominance is strong and located on the same meristem, growth is monopodic-straight (fig. 2, a and b). Growth becomes sympodic-straight when apical dominance is transferred to a lateral meristem (fig. 2, c). When apical dominance weakens, growth is monopodic or sympodicforked according to the number of meristems functioning together (fig. 2, d to f). The effect of light was studied in natural regeneration areas at 1200 meters above sea level, in Aigoual Mountains, Southern France. 45 areas, each containing 30 seedlings. were sampled under 3 light levels corresponding to 3 levels of canopy cover : full-light, half-light and shade. Characteristics measured concerned vigour : total height, types of shoots (exploration or exploitation) : and form (mono or sympodic, straight or forked). No polycyclism was present at this elevation. A direct analysis of these characteristics (tabl. 2 and 3) shows that full-light increases the number of exploration shoots and the total height. Half-light or shade favour exploitation shoots which never branch. The form is best in the latter conditions, but seedlings do not grow. A second analysis (tabl. 4) eliminating seedlings without exploitation shoots was applied. It shows that when a seedling has produced exploration shoots its forking risk is weakest in full-light. These results are very important for forestry management but they need confirmation in two directions : extension to low elevation regeneration areas and the effect on seedling form of a variation in light intensity due to part or total removal of the pre-existing stand.The genetic variability of vigour and form was studied in four 8-year-old provenance test (tabl. 1). Characteristics observed were the same as previously, but, since polycyclism was present in a number of seedlings, its role related to height-growth and form was also analysed. Height-growth is variable in 3 of the four provenance tests (tabi. 5). In the fourth test, named Ecouves, a poor and heterogeneous site, provenance variability is masked by a strong site effect. The choice of the best provenance produces a gain in height which roughly corresponds to one growing season shoot increment and which can easily be interpreted in economic terms (tabl. 6). Variability of form is absent in the Orléans trials (tabi. 5), but a slight one is detected at Ecouves and Chaud, where there is no provenance-environment interaction on forking. Improvement in form (tabl. 7) is high but very difficult to express in economic terms. Because of the high provenance environment-interaction, no provenance shows enough flexibility to be the best at sevcral sites both for vigour and form. Provenance tests will have to be established in thc main regions where beech is planted and compromise provenances will have to be looked for. Such compromises can already be detected in our trials. The positive effect of polycyclism on height growth is evident, but it becomes a troublesome characteristic for form. When limited to the terminal shoot it is favorable, but when appearing on laterals it may end as forks in the following growing season. Polycyclism is connected to abundance : abundance of light, of nutrients, of space. Planting techniques including spacing, fertilization and early thinnings should be closely adapted to soil fertility, water supply and canopy consistancy if planted under a pre-existing shade stand.