a global scale the quantity of active ingredient applied as herbicide and the energy required for manufacturing and field application is larger than all other pesticides combined (FAO, 2000a). In the developing country context, acute poisoning of agricultural workers from improper handling of herbicides also poses a significant public health risk that is linked to factors such as insufficient access to high-quality protective gear, poor product labeling, and low worker literacy rates (Repetto and Baliga, 1996). However, many of the newer classes of herbicide chemistry entering the market have much more favorable environmental profiles than commonly used insecticides and can be used at very low doses. Registration of new classes of herbicides has slowed (Appleby, 2005), which places a heightened imperative on maintaining the long-term efficacy of existing herbicides. There are also concerns for the sustainable use of compounds like glyphosate that are applied in conjunction with herbicide resistant crops (HRCs). Farmers using HRCs tend to extensively rely on a single herbicide at the expense of all other weed control measures, thereby decreasing long-term efficacy by increasing the odds of evolved herbicide resistance. However these worries are less of an issue in smallerscale systems where HRCs have not been previously used and seed systems make their widespread use less likely in the near future. Herbicides also have potential for reducing the cost of management of some important perennial and parasitic weed problems. Glyphosate is showing promise with farmers in Nigeria to reduce competition from the perennial grass Imperata cylindrica (Chikoye et al., 2002) and can reduce tillage inputs for management of other intractable perennial species, while in East Africa imazapyr herbicide tolerant maize has been introduced to combat Striga (Kanampiu et al., 2003).

Non-chemical control strategies can Iimit crop damage from weed competition.

Weed management attempts to reduce densities of emerging weeds, limit crop yield losses from established weeds, and promote the dominance of comparatively less damaging and difficult to control species. The first line of defense against weeds is a vigorous crop; basic crop management and cultural practices are important to maximize crop competitiveness and thereby reduce weed competition. Cultivars that are bred for competitive ability (Gibson et al., 2003), diverse crop rotations that provide a variety of selection and mortality factors (Westerman et al., 2005), and simple management changes such as higher seeding rates, spatiallyuniform crop establishment (Olsen et al., 2005), and banded fertilizer placement (Blackshaw et al., 2004) can reduce crop losses from uncontrolled weeds and, in some cases, reduce herbicide dependence. In conventional production settings, few of these options have been explicitly adopted by farmers. Cultural practice innovations for weed control work best if they are compatible and efficient complements to existing agronomic practices; hence, it is important to note the needs and constraints of farmers when developing new options for weed management (Norris, 1992). Hence participatory approaches are commonly used to ensure that

practices are appropriate to farmer needs (Riches et al., 2005; Franke et al., 2006).

Parasitic weeds are major constraints to several crops but a combination of host-plant resistance and management can control them.

Parasitic weeds such as Striga spp. and Alectra vogelii are major production constraints to several important crops, especially maize, sorghum and cowpea in SSA. Sources of resistance to S. gesneroides and A. vogellii were identified by traditional methods and the genes conferring resistance to and A. vogellii were subsequently identified using Amplified Fragment Length Polymorphism markers (Boukar et al., 2004) and successfully deployed in cowpea across W. Africa (Singh et al., 2006). Host-plant resistance to S. asiatica and S. hermonthica is now being deployed widely in new sorghum cultivars in East Africa but has been harder to find in maize. Inbred maize lines carrying tolerance to Striga have been developed and tolerance is quantitatively inherited (Gethi and Smith, 2004). However, the most successful strategy for controlling Striga in maize in West Africa is the use of tolerant cultivars used in rotation, and trap-cropping, using legumes, especially soybean, to germinate Striga seeds to reduce the seedbank (Franke et al., 2006). As Striga infestation is closely associated with low soil fertility, nutrient management, especially addition of nitrogen, can greatly increase yields of susceptible crops on infested fields. Farmers are now adopting green manures in legume/cereal rotations in Tanzania as a low-cost approach to reversing the yield decline of maize and upland rice (Riches et al., 2005). The interplanting of maize with Desmodium spp. within the "push-pull" system (Gatsby Charitable Foundation, 2005; Khan et al., 2006) is a promising approach to Striga suppression in East Africa. The broomrapes, Orobanche spp. are a major problem on sunflower, faba bean, pea, tomato and other vegetable crops in the Mediterranean basin, central and eastern Europe and the Middle East. Sources of resistance to broomrapes (Orobanche species) in a number of crops and the associated genes have been identified and mapped (Rubiales et al., 2006).

The increasing rate of naturalization and spread (i.e., invasions) of alien species introduced both deliberately and accidentally poses an increasing global threat to native biodiversity and to production.

Alien species are introduced deliberately either as new crops/livestock or as biocontrol agents; or by mistake as contamination of seed supplies or exported goods. Natural dispersal mechanisms account for only a small proportion of newly introduced species. This environmental problem has been ranked second only to habitat loss (Vitousek et al., 1996) and has totally changed the ecology of some areas (e.g., Hawaii). Negative economic and environmental