Estimation of thresholds for weed control in Australian cereals

A non-linear model relating crop yield to the density of weeds was fitted to nine Victorian weeds to evaluate their competitive abilities. The weeds were: Acroptilon repens (L.) DC. (creeping knapweed), Chondrilla juncea L. (Skeleton weed, Raphanus raphanistrum L. (Wild radish), lolium rigidum Gaud. (Annual ryegrass), Lithospermum arvense L. (White iron weed), Brassica tournefortii Gouan (Wild turnip), Lamium amplexicaule L. (Deadnettle), Fumaria parviflora Lam. (White fumitory) and Amsinckia hispida (Ruiz & Pav.) I.M. Johnston (Amsinckia). Where more than one experiment was available for a weed, the net return for a herbicide treatment over a range of weed densities was calculated to obtain the economic threshold density. Generally, the economic threshold densities within a weed species were the same order of magnitude, except for the perennial Chondrilla Juncea L., For this species data were collected in years of contrasting rainfall. The model used here is discussed in view of the threshold approach currently used in continental Europe.