Beschreibung:
<jats:p>Groups of mature, non‐lactating sheep and cattle grazed a <jats:italic>Nardus stricta</jats:italic> community during the growing season for six consecutive years from 1984 to 1989. Three unreplicated treatments were applied by continuous variable stocking to maintain between‐tussock sward height at (a) 4·5 cm by cattle grazing and (b) 4·5 cm or (c) 3·5 cm by sheep grazing. Diet composition and herbage intake were measured from 1984 to 1987 on three occasions in the growing season, and live weight and stock density were recorded from 1984 to 1988. The diet of cattle usually contained more dead herbage, <jats:italic>Nardus</jats:italic>, sedges and rushes but less forbs and other fine‐leaved grasses than the diet of sheep. Principal coordinate analysis showed different trends across years in diet composition, especially between the sheep and cattle treatments at 4·5 cm. Diet digestibility was usually higher for sheep treatments than for the cattle treatment. Diet digestibility and herbage intake increased between 1984 and 1985, and 1986; they also declined from spring to late summer. Regression of diet digestibility on independent principal components — which were derived from diet composition measurements — showed that the two most important principal components accounted for 72% of the variation in digestibility. Stock‐carrying capacity (kg LW × d ha<jats:sup>−1</jats:sup>, calculated from live weight, grazing days and stock density) was greater on the cattle treatment than on either sheep treatment. Stock‐carrying capacity also increased more in successive years on the cattle than on either sheep treatment, and it was greater on the sward maintained at 3·5 cm than at 4·5 cm by sheep. These increases in stock‐carrying capacity were generally positively associated with the increase in the percentage specific frequency both of live material and of the more productive grasses in the swards. These data indicate that sheep‐only stocking tends to <jats:italic>Nardus</jats:italic> dominance and suggest that further study using productive cattle — either alone or mixed with sheep — is needed, preferably in association with measurements of floristic change both within and between tussocks.</jats:p>