To avoid major negative impacts of the widespread adoption of biofuel species, whether they are exotic species, natives, or novel constructs, we need a system for screening their weed potential. Australia is an important global center of biodiversity and also has major cropping industries to protect. Prevention of the entry of further weeds is therefore a major national priority. The Weed Risk Assessment (WRA) system was developed and implemented for importation decisions in 1997; it has since been introduced into other countries and is probably as good as any system currently in operation. However, we need to be aware of the limitations of any system, to address these, and to work toward improved or alternative systems. WRA is a very simple spreadsheet requiring answers to questions about a species' life-history traits, dispersal, habitat suitability, impacts on other species, and history overseas, which are then added together and compared with numerical decision criteria. Its predictive powers are limited by this simplicity and by the complexity of human attitudes toward risk and impact. Alternative risk-management methods are available but, even so, the capacity for improvement is limited. It is quite possible, therefore, that in using any trait-based system to assess the negative risks of importation or interstate translocation of biofuel species, we will wrongly reject a valuable species or approve a species that turns out to be a major weed. It is suggested that, rather than attempting to improve a single-tiered decision-support system (the quarantine “sieve”), a multitiered system (nested sieves) would lead to a more effective system and greater cost-effectiveness. The key to this would be a postentry screening process for those species that have successfully passed through the WRA system.