netic engineering, to create genetically modified/engineered organisms (GMOs/GEOs) through transgenic technology by insertion or deletion of genes.        Combining plants with different and desirable traits can be slow because the genes for the traits are located in many different places in the genome and may segregate separately during breeding. Breeding augmented by molecular screening may yield rapid advances in existing varieties. This process, however, is limited by breeding barriers or viability in the case of cell fusion approaches, and there may be a limit to the range of traits available within species to existing commercial varieties and wild relatives. In any case, breeding is still the most promising approach to introducing quantitative trait loci (Wenzel, 2006). Emerging genomics approaches are showing promise for alleviating both limitations.

Genomics. Whole genome analysis coupled with molecular techniques can accelerate the breeding process. Further de­velopment of approaches such as using molecular markers through MAS will accelerate identification of individuals with the desired combinations of genes, because they can be rapidly identified among hundreds of progeny as well as improve backcross efficiencies (Baenziger et al., 2006; Re-ece and Haribabu, 2007). The range of contributions that MAS can make to plant breeding are being explored and are not exhausted (e.g., Kumar, 2006; Wenzel, 2006). It thus seems reasonable that MAS has the potential to contribute to development and sustainability goals in the long term, provided that researchers consistently benefit from funding and open access to markers. MAS is not expected to make a significant improvement to the rate of creating plants with new polygenic traits, but with future associated changes in genomics this expectation could change (Baenziger et al., 2006; Reece and Haribabu, 2007).        Regardless of how new varieties are created, care needs to be taken when they are released because they could be­come invasive or problem weeds, or the genes behind their desired agronomic traits may introgress into wild plants threatening local biodiversity (Campbell et al., 2006; Mer­cer et al., 2007).        MAS has other social implications because it favors centralized and large scale agricultural systems and thus may conflict with the needs and resources of poor farmers (Reece and Haribabu, 2007). However, breeding coupled to MAS for crop improvement is expected to be easily inte­grated into most regulatory frameworks and meet little or no market resistance, because it does not involve produc­ing transgenic plants (Reece and Haribabu, 2007). Varieties that are developed in this fashion can be covered by many existing IP rights instruments (e.g., Baenziger et al., 2006; Heinemann, 2007) and would be relatively easy for farm­ers to experiment with under "farmers' privilege" provided that suitable sui generis systems are in place (Sechley and Schroeder, 2002; Leidwein, 2006). The critical limitation of MAS is its ultimate dependence on plant breeding specialists to capture the value of new varieties; unfortunately, current and projected numbers of these specialists is inadequate (Re­ece and Haribabu, 2007).

Transgenic (GM) plants. Recombinant DNA techniques al­low rapid introduction of new traits determined by genes

that are either outside the normal gene pool of the species or for which the large number of genes and their controls would be very difficult to combine through breeding. An emphasis on extending tolerance to both biotic (e.g., pests) and abiotic (e.g., water stress) traits using transgenes is rel­evant to future needs.        Assessment of transgenic (GM) crops is heavily influ­enced by perspective. For example, the number of years that GM crops have been in commercial production (approxi­mately 10 years), amount of land under cultivation (esti­mated in 2007 at over 100 million ha) and the number of countries with some GM agriculture (estimated in 2007 at 22) (James, 2007) can be interpreted as evidence of their popularity. Another interpretation of this same data is that the highly concentrated cultivation of GM crops in a few countries (nearly three-fourths in only the US and Argen­tina, with 90% in the four countries including Brazil and Canada), the small number of tested traits (at this writing, mainly herbicide and pest tolerance) and the shorter-term experience with commercial GM cultivation outside of the US (as little as a year in Slovakia) (James, 2007), indicate limited uptake and confidence in the stability of transgenic traits (Nguyen and Jehle, 2007).         Whereas there is evidence of direct financial benefits for farmers in some agriculture systems, yield claims, adaptabil­ity to other ecosystems and other environmental benefits, such as reduced alternative forms of weed and pest control chemicals, are contested (Pretty, 2001; Villar et al., 2007), leaving large uncertainties as to whether this approach will make lasting productivity gains. The more we learn about what genes control important traits, the more genomics also teaches us about the influence of the environment and ge­netic context on controlling genes (Kroymann and Mitchell-Olds, 2005; MacMillan et al., 2006) and the complexity of achieving consistent, sustainable genetic improvements. Due to a combination of difficult to understand gene by en­vironment interactions and experience to date with creating transgenic plants, some plant scientists are indicating that the rate at which transgenic plants will contribute to a sus­tained increase in future global food yields is exaggerated (Sinclair et al., 2004).          Adapting any type  of plant  (whether transgenic  or conventionally bred) to new environments also has the po­tential to convert them into weeds or other threats to food and materials production (Lavergne and Molofsky, 2007; Heinemann, 2007). This problem is particularly relevant to transgenes because (1) they tend to be tightly linked packages in genomes, making for efficient transmission by breeding (unlike many traits that require combinations of chromosomes to be inherited simultaneously), and (2) the types of traits of most relevance to meeting development and sustainability goals in the future are based on genes that adapt plants to new environments (e.g., drought and salt tolerance). Through gene flow, wild relatives and other crops may become more tolerant to a broader climatic range and thus further threaten sustainable production (Mercer et al., 2007). An added complication is that these new weeds may further undermine conservation efforts.  The emer­gence of a new agricultural or environmental weed species can occur on a decade (or longer) scale. For example, it can take hundreds of years for long-lived tree species to achieve