hoods. New pest invasions, and the exacerbation of exist­ing pest problems, are likely to increase with future climate change. Warmer winters will lead to an expansion of insect and pathogen over wintering ranges (Garrett et al., 2006); this process is already under way for some plant pathogens (Rosenzweig et al., 2001; Baker et al., 2004). Within exist­ing over winter ranges, elevation of pest damage following warm winters is expected to intensify with climate change (Gan, 2004; Gutierrez et al., 2006; Yamamura et al., 2006). Increased temperatures are also likely to facilitate range ex­pansion of highly damaging weeds, which are currently lim­ited by cool temperatures, such as species of Cyperus (Terry, 2001) and Striga (Vasey et al., 2005).      Several current AKST strategies for managing agricul­tural pests could become less effective in the face of climate change, thus potentially reducing the flexibility for future pest management in the areas of host genetic resistance, biological control, cultural practices, and pesticide use (Pat­terson,  1999; Strand, 2000; Stacey, 2003; Bailey, 2004; Ziska and George, 2004; Garrett et al., 2006). For example, loss of durable host resistance can be triggered by deacti-vation of resistance genes with high temperatures, and by host exposure to a greater number of infection cycles, such as would occur with longer growing seasons under climate change (Strand, 2000; Garrett et al., 2006). Recent evidence from CO2-enrichment studies indicates that weeds can be significantly more responsive to elevated CO2 than crops, and that weeds allocate more growth to root and rhizome than to shoot (Ziska et al., 2004). This shift in biomass al­location strategies could dilute the future effectiveness of post-emergence herbicides (Ziska and George, 2004; Ziska and Goins, 2006). Elevated CO2 is also projected to favor the activity of Striga and other parasitic plant species (Phoe­nix and Press, 2005), which currently cause high yield losses in African cereal systems.      In addition to range expansion from climate change, the future increase in the trans-global movement of people and traded goods is likely to accelerate the introduction of invasive alien species (IAS) into agroecosystems, forests, and aquatic bodies. The economic burden of IAS is US$300 bil­lion per year, including secondary environmental hazards associated with their control, and loss of ecosystem services resulting from displacement of endemic species (Pimentel et al., 2000; GISP, 2004; McNeely, 2006). The costs associated with invasive species damage, in terms of agricultural GDP, can be double or triple in low-income compared with high-income countries (Perrings, 2005).

6.1.2.1 Diversification for pest resistance
To enhance the effectiveness of agroecosystem genetic diver­sity for pest management, some options include shifting the focus of breeding towards the development of multi- rather than single-gene resistance mechanisms. Other options in­clude pyramiding of resistance genes where multiple minor or major genes are stacked, expanding the use of varietal mixtures, and reducing selection pressure through diversifi­cation of agroecosystems.
     Multigene resistance, achieved through the deployment of several minor genes with additive effects rather than a sin­gle major gene, could become an important strategy where highly virulent races of common plant diseases emerge, as in

the case of the Ug99 race of wheat stem rust for which major gene resistance has become ineffective (CIMMYT, 2005). Integration of genomic tools, such as marker-assisted selec­tion (MAS) to identify gene(s) of interest, will be an impor­tant element of future resistance breeding. Future breeding efforts will need to include greater farmer involvement for successful uptake and dissemination, e.g., farmer-assisted breeding  programs  where   farmers  work  with  research and extension to develop locally acceptable new varieties (Gyawali et al., 2007; Joshi et al., 2007). Better develop­ment of seed networks will be needed to improve local ac­cess to quality seed.       Gene pyramiding (or "stacking") has the potential to become a future strategy for broadening the range of pests controlled by single transgenic lines. For example, express­ing two different insect toxins simultaneously in a single plant may slow or halt the evolution of insects that are resis­tant, because resistance to two different toxins would have to evolve simultaneously (Gould, 1998; Bates et al., 2005). Though the probability of this is low, it still occurs in a small number of generations (Gould, 1998); the long-term effec­tiveness of this technology is presently not clear. The use of gene pyramiding also runs the risk of selecting for primary or secondary pest populations with resistance to multiple genes when pyramiding resistance genes to target a primary pest or pathogen (Manyangarirwa et al., 2006). Gene flow from stacked plants can accelerate any undesirable effects of gene flow from single trait transgenic plants. This could result in faster evolution of weeds or plants with negative effects on biodiversity or human health, depending on the traits (as reviewed by Heinemann, 2007). Finally, mixtures of transgenes increase the complexity of predicting unin­tended effects relevant to food safety and potential environ­mental effects (Kuiper et al., 2001; Heinemann, 2007).      Varietal mixtures, in which several varieties of the same species are grown together, is a well-established practice, particularly in small-scale risk-adverse production systems (Smithson and Lenne, 1996). While this practice generally does not maximize pest control, it can be more sustainable than many allopathic methods as it does not place high se­lection pressure on pests, and it provides yield stability in the face of both biotic and abiotic stresses. For example, varietal mixtures could play an important role in enhancing the durability of resistance for white-fly transmitted viruses on cassava (Thresh and Cooper, 2005). Research on vari­etal mixtures has been largely neglected; more research is needed to identify appropriate mixtures in terms of both pest resistance and agronomic characteristics, and to back-cross sources of pest and disease resistance into local and introduced germplasm (Smithson and Lenne, 1996).      In addition to varietal mixtures, future AKST could en­hance the use of cropping system diversification for pest con­trol through supporting and expanding, where appropriate and feasible, practices such as intercropping, mixed crop­ping, retention of beneficial noncrop plants, crop rotation, and improved fallow, and to understand the mechanisms of pest control achieved by these practices. The underlying principal of using biodiversity for pest control is to reduce the concentration of the primary host and to create con­ditions that increase natural enemy populations (Altieri, 2002). The process of designing systems to achieve multiple