A semi-quantitative enzyme-linked immunosorbent assay (ELISA) for the detection of Rathayibacter toxicus in hay or pasture was used to estimate the degree of R. toxicus bacterial gall contamination in hay or pasture that was unsuitable for export but that may be suitable for feeding to domestic livestock. Based on experience of testing of fodder samples from pastures where livestock showed clinical signs, or outbreaks of annual ryegrass toxicity (ARGT), four relevant levels of bacterial gall contamination were selected. Several 1-kg samples of hay with no contamination by R. toxicus were spiked with these four different amounts of bacterial galls to provide five different risk categories considered relevant to the occurrence of ARGT. Extracts of the spiked samples were assayed to determine the range of ELISA results to be expected in each category. To validate these risk categories, all cases of ARGT diagnosed between 2000 and 2012 by the Animal Health Laboratories, Department of Agriculture and Food Western Australia, associated with the submission of suspected toxic fodder were allocated to the risk categories on the basis of the recorded results for fodder. In ∼15% and 79% of all cases, the fodder associated with outbreaks of ARGT fell into the ‘moderate risk’ and ‘high risk’ categories, respectively. The selected categories were considered to provide realistic estimations of the risk that fodder within them might cause ARGT if fed to livestock. This risk-level reporting has been adopted to enable informed decision making as to whether feeding a particular batch of hay or fodder to animals represents an acceptable risk.