Persistence of Some Varieties of Pasture Grass and Their Susceptibility to Invasion by Other Species

Many of the reseeded pastures of the better known grasses when planted at Frankenwald Field Research Station have proved unsuitable. This lack of success was in certain cases due to poor germination and establishment, e.g. Panicum coloratum. In other cases, e.g. Chloris gayana, good stands were established in the first year, but these deteriorated rapidly in succeeding years, being soon replaced by other volunteer grasses, chiefly Eragrostis curvula,* E. plana, Cynodon dactylon, Sporoboluspyramidalis and Hyparrhenica hirta. Eragrostis occurs in the second grass stage of secondary succession as is found on the sandy soils of the Transvaal Highveld, being later replaced by Hyparrhenia hirta (RoseInnes 1939). The increase of Eragrostis in veld where heavy applications of nitrogenous fertilizer have been applied has clearly been shown by Meredith (1947). Further, the effect of fertilizer in maintaining succession at the Eragrostis stage has been demonstrated by Davidson (1955) and Leigh (1958). It is desirable to maintain a pure stand of the selected variety of grass planted. Volunteer grasses, besides often lowering production per area, frequently give rise to difficulties associated with management. With this in view, the following experiment was carried out to determine whether certain types of E. curvula were more persistent and less susceptible to invasion by Hyparrhenia hirta and dicotyledonous weeds than other types. Certain varieties of Eragrostis curvula, which is a perennial bunch grass, are widely used as ley grasses in certain areas of South Africa. PROCEDURE The experiment was carried out at Frankenwald Field Research Station which is situated on latitude 260 41' S. and long. 280 06' E. at 4950 ft (1509 m) above sea level. The soils, derived from pre-Witwatersrand old granites, are acid, sandy, laterized and shallow. The annual rainfall is 30 in. (76 cm). The experimental layout consisted of a randomized block with three replicates. In each block there were twenty varieties of E. curvula and one variety each of E. plana, Digitaria smutsii and Chloris gayana. The experiment was planted in broadcast stands in February 1956. After a careful study of the varieties known collectively under the name of Eragrostis curvula, it was found possible to group the varieties under five main types to which the names 'Curvula', 'Chloromelas', 'Robusta blue', 'Robusta intermediate' and 'Robusta green' were assigned. This classification was based primarily on such morphological characters as leaf colour, leaf structure, inflorescence size and general growth habit. It was, however, realized that such characters are influenced by growth and fertilizer treatment. Nevertheless the results of further studies of such characters as anatomy