development and their impacts will need to be evaluated in the future. The main challenge here will be to keep GM pharma and industrial crops separate from crops for food (Ellstrand, 2003; Ledford, 2007).

Environmental impacts of GM crops are inconclusive.

Both negative and benign impacts have been reported, depending on the studied system and the chosen comparator. Contradictory reports from laboratory and field studies with GM crops (Bt- and herbicide resistant) show a great diversity of impacts on non-target organisms, including arthropods and plants (Burke, 2003; O'Callaghan et al., 2005; Squire et al., 2005; Hilbeck and Schmidt, 2006; Sanvido et al., 2006; Torres and Ruberson, 2006). Some reports claim that GM crops do not adversely affect biodiversity of non-target organisms, or have only minor effects, while others report changes in the community composition of certain biocontrol taxa (Torres and Ruberson, 2006). Some reports find that the key experiments and fundamental issues related to environmental impacts are still missing (Wolfenbarger and Phifer, 2000; Snow et al., 2005). Another controversial topic surrounds claims that GM crops significantly reduce pesticide use and thus help to conserve biodiversity (Huang et al., 2002; Pray et al., 2002; Qaim and Zilberman, 2003; Bennett et al., 2004ab; Morse et al., 2004). Contradictory evidence has also been provided (e.g., Benbrook, 2003, 2004; Pemsl et al., 2004, 2005), which in part may be attributable to the dynamic condition of pest populations and their outbreaks over time. A further complication arises from the development of secondary pests which reduce the benefits of certain Bt crops (Qayum and Sakkhari, 2005; Wang et al., 2006). The effects of Bt crops on pesticide use and the conservation of biodiversity may depend on the degree of intensification already present in the agricultural system at the time of their introduction (Cattaneo et al., 2006; Marvier et al., 2007). A recent meta-analysis of 42 field studies (Marvier et al., 2007) in which scientists concluded that the benefits of Bt-crops are largely determined by the kind of farming system into which they are introduced, found that Bt-crops effectively target the main pest when introduced into chemical intensive industrial farming systems. This provides some support to the claim that Bt plants can reduce insecticide use. However, when Bt crops were introduced into less chemical intensive farming systems the benefits were lower. Furthermore when introduced into farming systems without the use of synthetic pesticides, (e.g., organic maize production systems), there were no benefits in terms of reduced insecticide use. In fact, in comparison with insecticide-free control fields, certain non-target taxa were significantly less abundant in Bt-crop fields. Most field studies were conducted in pesticide-intensive, large-scale monocultures like those in which 90% of all GM crops are currently grown (Cattaneo et al., 2006); consequently, these results have limited applicability to low-input, small-scale systems with high biodiversity and must be assessed separately. Introducing GM crops accompanied by an intensification strategy that would include access to external inputs

could enhance benefits for small-scale systems (Hofs et al., 2006; Witt et al., 2006).

Currently there is little, if any, information on ecosystem biochemical cycling and bioactivity of transgene products and their metabolites, in above and below ground ecosystems.

There are multiple potential routes for the entry of Bt-toxins into the ecosystem, but there is little information to confirm the expected spread of Bt-toxins through food chains in the field (Harwood et al., 2005; Zwahlen and Andow, 2005; Harwood and Obrycki, 2006; Obrist et al., 2006). One expected route would be embedded in living and decaying plant material, as toxins leach and exude from roots, pollen, feces from insects and other animals. There is confirmation of the presence of Bt toxin metabolites in feces of cows fed with Bt-maize feed (Lutz et al., 2005). Several experiments have studied the impacts of Bt-crop plant material on soil organisms with variable results ranging from some effects, only transient effects, to no effects (e.g., Zwahlen et al., 2003; Blackwood and Buyer, 2004). However, to date there has not been a study of the ecosystem cycling of Bt toxins and their metabolites, or their bioactivity.

Evidence is emerging of herbicide and insecticide resistance in crop weeds and pests associated with GM crops.

Since 1995 there have been reports of an increase from 0 to 12 weed species developing resistance to glyphosate, the main broad spectrum herbicide used in GM crops from countries where herbicide-resistant GM crops are grown (van Gessel, 2001; Owen and Zelaya, 2005; Heap, 2007). In addition, the use of glyphosate has greatly increased since the introduction of herbicide tolerant crops. With the exception of Australia (http://www.ogtr.gov.au/rtf/ir/dir059finalrarmp1. rtf: 2006, Australian Gene Technology Act 2000) no resistance management plans are required for the production of herbicide resistant crops; management strategies are required for insect-resistant Bt-crops, in most countries where they are grown. There has been only one report of an insect pest showing resistance to one of the commonly used Bt-toxins (Gunning et al., 2005). Strategies are needed for efficient resistance management and the monitoring of the spread and impacts of GM-resistance genes in weed and pest populations.

There are reported incidents of unintentional spread (via pollen and seed flow) of GM traits and crops.

The consequences from unintentional spread of GM traits and GM crops could be serious. GM traits and crops with varying levels of approval are spreading fast throughout the world; intentional spread occurs mainly through human