Background and objectives
The head smut Sporisorium ophiuri is a candidate biocontrol agent for Rottboellia cochinchinensis (itchgrass), a serious grass weed in the humid tropics [1]. Seedlings are infected during germination and seed heads of infected plants contain only smut spores. The smut cannot be released until stringent host specificity tests have been conducted, so modelling is a useful means of investigating the potential of biocontrol in controlling the weed.

Materials and methods
We have adapted a simple model [3] with particular reference to development of integrated management packages for itchgrass in maize in the Guanacaste region of Costa Rica which combine herbicides, cover crops, and fallow control between crops [2] as well as the possibility of biocontrol. It is deterministic with seasonal transition probabilities. There are typically two cropping seasons per year and then a fallow dry season. Biocontrol was modelled as a constant proportion of itchgrass plants, seasonally, which set no seed.

To simulate the impact of head smut on an itchgrass population in a naturally developing epidemic, the pathogen population dynamics (highly simplified), were also incorporated into a model of itchgrass population dynamics. The proportion of itchgrass plants infected each year was calculated assuming that infection results from random contact between germinating weed seeds and the pathogen. The number of new smut propagules was dependent on the number of itchgrass plants infected.

Results and conclusions
The model suggested that the smut may be a useful control in combination with a cover crop. The smut will interact best with control measures which also decrease itchgrass germination, seedling mortality, seed production and proportion of the seed rain incorporated into the soil. According to the model, the smut caused a greater percentage reduction in the weed density when used in addition to the cover crop than if used on its own. However, cover crops themselves may have a yield penalty which further economic cost-benefit models of itchgrass control could be used to investigate.

Sensitivity analyses of the second model implied smut effectiveness is directly proportional to spore production and infectivity, and inversely proportional to spore mortality rate in the soil, suggesting potential trade-offs when selecting the best smut biotypes. However, equivalent trade-offs in these parameters did not result in similar behaviour patterns leading up to equilibrium. Decreases in pathogen mortality rate in the soil resulted in a much more stable pattern of behaviour than equivalent increases in spore production rate or infectivity.

1. Ellison CA, Evans HC, 1995. In Proceedings of the VIII International Symposium on Biological Control of Weeds, 2-7 February 1992, Lincoln University, Canterbury, New Zealand (eds. Delfosse, ES and RR Scott). DSIR/CSIRO: Melbourne, pp. 493-500.
2. Smith MC, Reeder R, Thomas M, 1997. Journal of Applied Ecology 34, 388-398.
3. de la Cruz R, Rojas E, Merayo A, 1994. Manejo Integrado de Plagas (Costa Rica) 31, 29-34.