• whether ecological effects may arise through the unrestricted release and use of genetically engineered soybean containing 'foreign' genes
• how the pattern of use of the herbicide glyphosate could change and the effects of such a change on human health and the environment
• potential for harm to human consumers as a result of the changes to the soybean

This determination must take into account the context of the soybean market and its trading arrangements. Soybean is an internationally traded commodity and Monsanto is a multinational corporation. The most reasonable expectation of any safety assessment is, therefore, that this will have considered the environment most at risk. This is one of the most striking shortcomings in Monsanto's environmental risk assessment for the US and European authorities which have relied only upon data from the United States, Canada, Costa Rica and Puerto Rico. Not only are there methodological problems with the experiments in these countries per se but they do not reflect the ecological situation in the Far East where ecological risks may be greatest.

This paper considers this aspect of Monsanto's risk assessment together with the rather optimistic assumptions they have made when coming to their conclusion of safety for human consumption and their neglect of the potential for harm arising out of changed patterns of use of glyphosate.

Monsanto has conducted field trials of RRS in Argentina, Canada, Costa Rica, Puerto Rico and the USA. In 1994 they also conducted semi-contained trials in Japan (Hino, 1994). Monsanto have made it clear that RRS will not be separated from other, conventionally bred soybean. From these data the inevitable conclusion is that RRS will be grown and traded worldwide including the Far East.

One of the main issues in the environmental risk assessment of genetically engineered crops is whether the introduced, foreign gene can be transferred through outcrossing to native, related species and cause genetic pollution. Although cultivated soybean is mainly self-pollinating, pollen can also be carried by bees to other soybeans and related wild or weed plants. Soybean can cross with other member of the genus Glycine which are found in Australasia including Japan. Natural hybridisation is known to occur between cultivated soybean and G. soya, a common weed in Japan (USDA/APHIS, 1994). There are no compatible relatives in the United States or Europe.

Thus the ecological dangers seem to be greatest in the Far East and the potential for RRS to cause ecological harm should be evaluated in those environments. In Monsanto's determination of this risk, when seeking permission to import RRS into Europe, they have only considered the possibility of gene transfer in Europe where there are no related species of plants. In the USA they have made a similar argument. Yet it is clear that it is not Monsanto's intention that the growth and trade of soybean be restricted to these countries.

Monsanto should evaluate the risk of gene transfer in Australasian countries and not only where RRS may be grown but also where it may be imported. As one ecologist has recently commented: "Given that crop seeds travel hundreds of kilometres between seed merchant, farmer and processing factory, spillage in transport is inevitable - and more worrying that pollen spread" (Crawley, 1996).

Another important issue in ecological (and agricultural) risk assessment is whether RRS could persist in the environment and displace natural flora and fauna or become a troublesome weed. Because soybean is susceptible to frost this has been used by Monsanto to claim environmental safety. However, the experimental trials described by Monsanto in their petition for deregulation in the USA and their market application in Europe have not properly investigated RRS's potential for persistence in the environment. Plots are routinely destroyed by 'disking' after experiments designed to tests agronomic traits not ecological impact.

In one trial volunteer RRS were detected in the following season, however, these plants were destroyed by being 'disked' into the ground and the observation seems never to have been investigated further. In addition, and of more concern, is that the potential for persistence in countries other than the USA or Europe has never been addressed (for example, in the Far East where frost will not regulate persistence). This has not been considered at all by Monsanto.

Furthermore, as with the potential for hybridisation, Monsanto have supplied no quantitative data on invasiveness or weediness at all (Kareiva and Parker, 1994). Monsanto have also used Baker's List to argue that RRS is unlikely to become a troublesome weed. This is a list of characteristics which are said to be commonly associated with weeds and which has been used to try and predict the weediness of plants (Keeler, 1989). The usefulness of this list has been questioned by scientists who have used it to see if it can predict the weediness of existing, known weeds and have shown that it is not reliable (Williamson, 1994).

Gene transfer and the persistence of RRS are important ecological questions. The gene coding for glyphosate tolerance may not be ecologically neutral. In the presence of glyphosate it would be expected to give a competitive advantage and thus the potential for the persistence and spread of RRS or related species with which it crosses must be treated seriously. Even in the absence of glyphosate the gene may persist and cause genetic pollution. Assumptions that herbicide resistance genes would not persist in the absence of herbicide because they would be 'genetic baggage' and thus confer a disadvantage have been undermined by experiments which showed that genetically engineered oilseed rape was no less able to survive than non-engineered rape (Crawley et al., 1993). Monsanto's ecological risk assessment for RRS is therefore deficient in several important respects because it:

• presents an unjustifiably optimistic scenario by assessing risks only in the U.S. and European context when RRS is likely to be grown and/or exported to environments of higher ecological risk in Australasia;
• has failed to use appropriate experiments to investigate ecological risk as trials have concentrated on investigating agronomic characteristics not invasiveness, weediness or potential for hybridisation;
• has relied upon unreliable predictive methods to argue that RRS is ecologically safe.

In the context of a global soybean market, Monsanto's determination of ecological harmlessness for RRS cannot be relied upon.

Claims are made that Roundup is a safe herbicide yet this is a contradiction in terms because a chemical used for its toxic effects on plants cannot be environmentally benign. There is evidence that Roundup can cause harm to the environment and human health even at current levels of use. Increased use of the herbicide Roundup may result in pollution of water and lead to a further decrease in wild plant diversity. It may also harm animals and beneficial soil micro-organisms. Glyphosate is believed to be immobile in soil as it readily binds to soil particles, however a new study indicates that glyphosate can be readily released from soil particles, and therefore may leach into water (Piccolo et al., 1994). The active ingredient of Roundup has already been detected in drinking water in Germany (Greenpeace, 1995).

Roundup can be toxic to fish to an extent which depends on several factors including the hardness of the water, the age of the fish and water temperature. In some situations concentrations as low as 10 parts per million of glyphosate can kill fish (WHO, UNEP & ILO, 1994).

Glyphosate also affects the growth and survival of earthworms. A study in New Zealand showed that glyphosate caused a reduction in growth and increase in mortality of the most commonly found earthworm in New Zealand (Springett & Gray, 1992). Glyphosate as is also toxic to many beneficial mycorrhiza fungi which help plants to take up nutrients from soils (Estok et al., 1989, Chakravarty & Chatarpaul, 1990, Sidhu & Chakravarty, 1990, Chakravarty & Sidhu, 1987).

Spray can drift up to 400 to 800 metres away from the site of application (Yates et al., 1978) and cause damage to wild plants and flowers. In turn the death of plants can have indirectly harmful effects on wild life such as insects, birds and mammals which depend on the vegetation for food or shelter (Cox, 1995 b).

Glyphosate residues have been found in strawberries (Cessna & Cain, 1992), lettuce, carrots, barley (U.S. EPA, 1993), and fish (Wang et al., 994, Folmar et al., 1979). These residues persisted long after the glyphosate was used. For example, lettuce, carrots, and barley contained glyphosate residues at harvest when planted a year after treatment (U.S. EPA, 1993). Because RRS are tolerant to glyphosate at twice the recommended levels of usage, farmers may be tempted to use extra or be less careful in their application rates.

The Northwest Coalition for Alternatives to Pesticides has undertaken a thorough review on glyphosate's toxicology, human exposure and ecological effects. They found that products containing glyphosate are acutely toxic to humans. Symptoms include eye and skin irritation, cardiac depression and vomiting (Cox, 1995 a). In California they found that glyphosate was the third most commonly-reported cause of pesticide related illness among agricultural workers (Cox, 1995 b, Pease et al., 1993). These toxicities seem to be a result of the so-called 'inerts' in some formulations of the product and are not fully understood. Findings such as these throw into question Monsanto's unsupported assertion for the benefits of Roundup such as that it is an 'environmentally acceptable herbicide'.

Other claims by Monsanto that the use of RRS and glyphosate may be cheaper than other options (which are never specified) go unsubstantiated and potential drawbacks such as the encouragement of the emergence of herbicide tolerant weeds through increased reliance on one herbicide go unmentioned. Recently there has been a report of glyphosate resistance in ryegrass in Australia (New Scientist, 6 July 1996, p.6). Resistance could be further encouraged by the use of herbicide resistant crops both through the increased use of a particular herbicide and/or the transfer of resistance genes to weeds.

Therefore, Monsanto's evaluation of the human and environmental safety of the change in use patterns of Roundup are inadequate because despite:

• evidence existing which suggests that formulations of glyphosate may be causing health problems even at existing levels of use;
• the potential for harm to non-target plants and other organisms through the increased use of a non-specific herbicide;
• the likelihood of increased residues in crops and groundwater.

Monsanto and the regulatory authorities have ignored the changing use of glyphosate in their evaluation of RRS. As a result no confidence can be placed in claims that the use of RRS together with Roundup will be safe.

RRS contains a gene from an Agrobacterium species which codes for an enzyme 5-enolpyruvylshikimate-3-phosphate (EPSPS). Glyphosate normally kills plants by inhibiting the EPSPS which is an important enzyme in one of the biochemical pathways of plants. The new form of EPSPS (CP4 EPSPS) which has been engineered into soybean is not sensitive to glyphosate and thus renders the soybean tolerant to it.

The biggest question mark hangs over the assessment of whether CP4 EPSPS could be allergenic in soybean products which contain the protein. There are no predictive assays of allergenicity. The latest findings that a gene from Brazil nut, which encodes for a methionine rich protein, which was engineered into soybean caused allergic reactions in sensitive humans despite earlier evaluations being negative (Nordlee et al., 1996) This has highlighted the potentially serious effects of transferring genes coding for novel proteins.

The discovery of the allergenic potential of Pioneer's genetically engineered soybean prior to its use as human food sources was thanks to a unique advantage: the donor organism for the gene, Brazil nut, was a known food allergen, and serum samples of persons known to be allergic to Brazil nuts were available for testing. Monsanto has a special difficulty in assessing the allergic potential of RRS because unlike the Brazil nut protein it does not come from a source recognised to be allergenic. Thus Monsanto have no specific tests they can conduct to gather evidence suggestive of safety. Monsanto have had to rely on more questionable methods such as comparison to other compounds rather than any empirical evidence. For example, Monsanto have emphasised that EPSPS does not have sequence homology with other known protein allergens. Yet one expert recently stated: "...it is impossible to identify most allergenic proteins prior to genetic manipulation. These proteins do not appear to have common amino acid sequences so comparison of their IgE-binding epitopes is probably fruitless" (Taylor,1994).

Monsanto also emphasise the small quantity of EPSPS that would be present in diets and that most allergens are major proteins in the diet. However, this is not inevitably the case (Taylor, 1994) and the CP4 EPSPS gene is being used in other crops too so exposure may not be restricted to soybean. CP4 EPSPS is also produced in higher quantities in engineered soybean than normal EPSPS in non-engineered soybean.

Taylor (1994) has also emphasised that assessments of digestibility should be carried out both with the isolated protein and in the form it will be ingested. It is not clear that Monsanto have conducted the tests with intact food under realistic conditions. Thus the Monsanto evaluation of the importance of the risk of allergenicity from RRS is deficient because: the risk of allergenicity may be low but is essentially unpredictable; food containing RRS protein will not be labelled, therefore the most important protective mechanism available to susceptible people, avoiding the harmful food, will not be open should problems emerge; the bacterial EPSPS gene may also be present in other, non-soy based products which will affect the risk.

The potentially serious risks associated with food allergies are unjustifiable when they are unpredictable and untestable and the normal ways of avoiding food allergens become unavailable. Monsanto's evaluation of the safety of RRS ignores these issues.

The whole issue of changed patterns of Roundup use has been ignored even though Monsanto's intention is to increase sales of glyphosate through sales of RRS. Impacts on the environment and human health are being left to large scale uncontrolled experiment in the environment. Whether the new form of the protein present in soybean will be safe is unpredictable. The nature of the potential harm is serious yet this feature and the impossibility of managing or avoiding harm should it arise has not been considered. Therefore, no confidence can be placed in assertions of the safety and benefits of RRS. They have excluded important factors from consideration.

Investing in RRS is to sanction an experiment with the environment and human health, the dangers of which may be unpredictable, irreversible and unmanageable. This experiment is being conducted in the interests of promoting sales and use of a toxic chemical. The package of chemical and resistant seed may be only of short-term usefulness to farmers as resistance problems emerge. Greenpeace believes these risks are not justifiable and food companies encouraging such risk taking by using RRS in their products are acting in Monsanto's not consumers' interests.

In addition RRS contravenes the principle of sustainability and the precautionary approach which is embedded in European law both through the EU's endorsement of the Rio Declaration and the precautionary basis of Directive 90/220 which covers the marketing of genetically engineered organisms. Risking irreversible damage to the environment which will be inherited by future generations should no longer be tolerated.(return)

Also see the related report "Not Ready for Roundup: Why Consumers and Farmers Should Avoid Monsanto's Genetically Engineered Soybeans".

Cessna A J and Cain N P (1992) Residues of glyphosate and its metabolite AMPA in strawberry fruit following spot and wiper applications. Can. J. Plant. Sci. 72:1359-1365.

Chakravarty P and Chatarpaul L (1990) Non-target effect of herbicides: I. Effect of glyphosate and hexazinone on soil microbial activity. Microbial population, and in-vitro growth of ectomycorrhizal fungi. Pestic. Sci. 28:233-241

Chakravarty P and Sidhu S S (1987) Effects of glyphosate, hexazinone and triclorpyr on in vitro growth of five species of ectomycorrhizal fungi. Eur. J. For. Path. 17:204-210

Cox C (1995 a) Glyphosate, Part 1: Toxicology, Herbicide Factsheet, Journal of Pesticide Reform, Fall 1995 Vol. 15, No 3, (from the Northwest Coalition for Alternatives to Pesticides)

Cox C (1995 b) Glyphosate, Part 2: Human Exposure and Ecological Effects, Herbicide Factsheet, Journal of Pesticide Reform, Winter 1995 Vol. 15, No 4, (from the Northwest Coalition for Alternatives to Pesticides)

Crawley M (1996) 'The Day of the Triffids'. New Scientist 6 July pp40-41

Crawley, M J, Hails, R S, Rees, M, Kohn, D & Buxton, J (1993) Ecology of transgenic oilseed rape in natural habitats. Nature 363: 620-623

Estok D, Freedman B and Boyle D (1989) Effects of the herbicides 2,4-D, glyphosate, hexazinone, and triclopyr on the growth of three species of ectomycorrhizal fungi. Bull. Environ. Contam. Toxicol. 42:835-839

Folmar L C, Sanders H O and Julin A M (1979) Toxicity of the herbicide glyphosate and several of its formulations to fish and aquatic invertebrates. Arch. Environ. Contam. Toxicol. 8:269-278

Greenpeace (1995) Pestizide im Grundwasser, Ergebnisse einer Befragung deutscher Gesundheitsaemter 1994, Greenpeace Bericht

Hino A (1994) Safety assessment system of field tests in Japan. Proceedings of the 3rd International Symposium on the Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms. November 13-16 1994, Monterey, California. University of California: Oakland. pp 177-182

Jewell L D (1988) Agricultural Statistics 1987 United States Department of Agriculture, Washington

Kareiva, P & Parker, I (1994) Environmental risks of genetically engineered organisms and key regulatory issues. Greenpeace International: Amsterdam

Keeler, K H (1989) Can genetically engineered plants become weeds? Bio/Technology 7: 1134-1139

Nordlee J A, Taylor S L, Townsend J A , Thomas L A and Bush R K (1996) Identification of a brazil-nut allergen in transgenic soybeans, The New England Journal of Medicine, Vol 334, No 11:688-692

Pease W S et al. (1993) Preventing pesticide-related illness in California agriculture: Strategies and priorities. Environmental Health Policy Program Report. Berkeley, CA: University of California. School of Public Health. California Policy Seminar.

Piccolo A et al. (1994) Adsorption and desorption of glyphosate in some European soils. J. Environ. Sci. Health B29(6):1105-1115

Sidhu S S and Chakravarty P (1990) Effect of selected forestry herbicides on ectomycorrhizal development and seedling growth of lodgepole pine and white spruce under controlled and field environment. Eur J. For. Path. 20:77-94

Springett J A and Gray R A J (1992) Effect of repeated low doses of biocides on the earthworm Aporrectodea caliginosa in laboratory culture. Soil Biol. Biochem. 24(12):1739-1744

Taylor, S L (1994) Evaluation of the allergenicities of foods developed through biotechnology. Proceedings of the 3rd International Symposium on the Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms. November 13-16 1994, Monterey, California. University of California: Oakland. pp 185-198

USDA/AHPIS (1994) Response to Monsanto petition P93-258-01 for determination of nonregulated status for glyphosate tolerant soybean line 40-3-2. United States Department of Agriculture, Animals and Plant Health Inspection Service, Biotechnology, Biologics and Environmental Protection.

U.S. EPA Office of Pesticide Programs. Special Review and Registration Division (1993) Reregistration eligibility decision (RED): Glyphosate. Washington, D.C. (September)

Wang Y, Jaw C and Chen Y (1994) Accumulation of 2,4-D and glyphosate in fish and water hyacinth. Water Air Soil Pollut. 74:397-403

Williamson, M (1994) Community response to transgenic plant release: predictions from British experience of invasive plants and feral crops. Molecular Ecology 3: 75-79

World Health Organization, United Nations Environment Programme, International Labour Organization (1994) Glyphosate. Environmental Health Criteria 159. Geneva, Switzerland

Yates W E , Akesson N B and Bayer D E (1978) Drift of glyphosate sprays applied with aerial and ground equipment. Weed Sci. 26(6):597-604