Abstract
Eight experiments, each having 15 treatments, were established. Sites with clayey, loamy and sandy soils were used. The treatments included herbicide spots of various sizes up to whole plots kept weed-free; polythene mats of various sizes up to whole plots (10 x 8 m) covered in polythene; and polythene mats of various sizes with the rest of the weeds within the plot treated with herbicide. In each experiment survival or growth or both increased the further weeds were kept from the trees. Growth was increased most on the sandy site. Survival was increased most in the experiment which used poor quality oak plants - from zero in the control treatment to 63 per cent in the whole-plot polythene treatment. Herbicide spots were generally better than polythene mats when the area treated was less than one square metre per tree, but not so good when larger areas were treated: this is attributed to weeds rooting under the mulches. Growth was increased when the weeds around the mulches were killed. Accidental herbicide damage, voles gnawing the trees from beneath the mulches, and soil anaerobism on poorly drained sites also influenced the results.