495 Fifteenth Australian Weeds Conference Summary Potential social costs and externalities as- sociated with herbicide resistance have not generally been considered in economic analyses of the impact of herbicide resistance. The economics of managing herbicide resistance in weeds has focused on cost- effective responses by individual growers to the devel- opment of resistance at the individual field level, and the risk of resistance spread through weed mobility has been treated in economic analyses as if it were negligi- ble. In reality, weed mobility needs to be assessed on a case-by-case basis. Resistance has been shown in many cases to have resulted from multiple events, rather than by spread, however, there is also a likelihood of resist- ance mobility through the spread of pollen, seed and plants. Furthermore, growers may perceive that they will gain resistance via mobility rather than just their own actions. In cases where weed mobility is high, failure to recognise costs associated with externalities results in myopic behaviour by individual agents, and hence overuse of the herbicide resource. In this paper we discuss the evidence for weed mobility in various farming environments, and use a simulation model to consider the impact mobility may have on the onset of resistance, weed control costs and optimal herbicide resistance management strategies. Keywords Herbicide resistance, weed mobility, social costs, externalities, simulation model. INTRODUCTION Weed resistance to herbicides is an increasing problem world-wide, affecting the efficacy of major herbicides. Whereas externalities and social costs associated with resistance to pesticides and antibiotics have been con- sidered by economists (e.g. Laxminarayan 2003), this is not the case for herbicide resistance. Externalities exist when an activity undertaken by an individual has side-effects on others that are not taken into considera- tion by the first individual. Externalities can result in social costs: costs incurred off-site by individuals or society when externalities exist. If externalities are not taken into consideration then individual optimal use may be too high, as individuals may use products with insufficient concern about negative impacts of current use on future efficacy of the product for others. Weed and pollen mobility and social costs of herbicide resistance Sally P. Marsh 1 , David J. Pannell 1,2 and Rick S. Llewellyn 1,3 1 School of Agricultural and Resource Economics, University of Western Australia, Crawley, Western Australia 6009, Australia 2 CRC for Plant-based Solutions to Salinity, University of Western Australia, Crawley, Western Australia 6009, Australia 3 Present address: CSIRO Sustainable Ecosystems, Urrbrae, South Australia 5064, Australia Failure to consider externalities in analyses of optimal herbicide use is often justified by the higher mobility of many insects and diseases relative to weeds, and the consequent higher risk of off-site effects. The costs of herbicide resistance usually consid- ered when assessing optimal farmer use of herbicides are those associated with: i) the risk of poor future weed control and hence loss of crop yield, especially in situations where farmers do not realise they have resistant weeds; and ii) extra weed control costs, both in situations where resistant weeds are more expensive to treat or where management to prevent resistance is more expensive, as in the ‘Double-Knock’ treatment recommended to reduce the probability of the develop- ment of glyphosate resistance (Neve et al. 2003). The economics of managing weed herbicide resistance has focused on cost-effective responses by individual growers to the development of resistance at the individual field level, and the risk of resistance spread through weed mobility has been treated in eco- nomic analyses as if it were negligible (e.g. Gorrdard et al. 1996, Weersink et al. 2005). In reality, the risk needs to be assessed on a case- by-case basis. Some weeds are very mobile. Although in many cases herbicide resistance has been shown to result from multiple evolutionary events, rather than by spread, there is also a varying likelihood of resist- ance mobility through the spread of pollen, seed and weeds themselves, which will be dependent on a range of factors. The spread of herbicide resistance genes within a population is controlled in part by the extent to which a species cross-pollinates and also by the nature of the resistance gene. Rieger et al. (2002) have shown in canola the pollen movement of HR genes to be at least 2.6 km. Resistance has been shown to have spread from a single weed source in irrigation-based agriculture (Fischer et al. 2004), and research in Aus- tralia suggests that some separate glyphosate-resistant ryegrass populations in New South Wales are likely to have occurred through seed movement (Stanton et al. 2004). In the United States, most cases of glyphosate- resistant weed populations are reported in horseweed (Conyza canadensis L.), a weed whose parachute-type seeds are readily dispersed by wind, and some resistant