Field studies of the effects of pre-sowing weed control and time of sowing on tropical perennial grass establishment, North-West Slopes, New South Wales

Abstract

Two separate experiments on the North-West Slopes of New South Wales investigated: (1) the effects of differing lengths of pre-sowing weed control (2 years, 1 year, and 3–4 months in the spring immediately before sowing) on the soil seed density of liverseed grass (Urochloa panicoides) from November 2005 to November 2007 and its seedling density at establishment (December 2007), as well as on subsequent herbage mass and plant frequency of this species and Panicum coloratum var. makarikariense cv. Bambatsi (Expt 1); and (2) the effects of cultivar, time of sowing (November 2006, January 2007, and March 2007), and defoliation in the autumn after sowing, on the over-wintering ability of 5 tropical perennial grasses (Expt 2). The cultivars sown in Expt 2 were Bambatsi, Dichanthium aristatum cv. Floren, Chloris gayana cv. Katambora, Digitaria eriantha var. eriantha cv. Premier, and Bothriochloa bladhii var. glabra cv. Swann. At sowing (November 2007), liverseed grass seed and seedling densities in Expt 1 were lowest (P < 0.05) in the 2-year pre-sowing weed control treatment (15 and 13/m2, respectively), and the subsequent herbage mass of Bambatsi (January and October 2008) was highest (P < 0.05) in this treatment. In Expt 2, 59 frosts occurred in 2007, and Floren, Bambatsi, and Swann sown in March of that year over-wintered poorly and had lower (P < 0.05) herbage mass and plant frequency in the following spring compared with the other sowing times. These differences were also evident in autumn and spring 2008, up to 23 months after sowing.