The biggest environmental-impact study of genetically modified crops conducted anywhere in the world has produced largely mixed results.
Scientists tested three biotech crops and found the cultivation of two - an oilseed rape and a beet crop - to be more harmful to many groups of wildlife than their conventional equivalents.
A GM spring rape was tested in the three-year trials
The production of a third biotech plant - a maize - was shown to be kinder to other plants and animals than the normal crop.
The results of the trials will be used by the UK Government, along with other information, to make a decision on whether or not to allow the engineered plants to be commercialised in the country.
The outcome of the £6m three-year study conducted at some 60 sites across Britain was reported on Thursday in eight lengthy papers in the journal Philosophical Transactions Of the Royal Society: Biological Sciences.
The head of the research team, Dr Les Firbank, said: "The results are clearly important to the debate about the possible commercialisation of GM crops.
"But, they also give us new insights that will help us conserve biodiversity within productive farming systems."
The so-called farm-scale evaluations (FSEs) set out to look at a narrow set of issues related to the impact on the environment of herbicide-tolerant GM crops.
These plants can be sprayed with a particular weedkiller and still prosper while other "pest" plants in the field are killed.
The FSEs tested the idea that the alternative management practices involved in the production of these crops would make no difference to biodiversity in the field.
The scientists' work rejects this.
They grew the GM plants and their conventional equivalents side by side, and then observed the wildlife in among the crops and at the field margins.
Birds and bees
The FSEs showed that some insect groups, such as bees (in beet crops) and butterflies (in beet and rape), were recorded more frequently in and around the conventional crops because there were more weeds to provide food and cover.
There were also more weed seeds in conventional beet and rape crops than in their GM counterparts.
Such seeds are important in the diets of some animals, particularly some birds. However some groups of soil insects were found in greater numbers in the GM beet and rape crops.
In contrast, growing GM maize was better for many groups of wildlife than conventional maize. There were more weeds in and around the biotech maize crops, more butterflies and bees around at certain times of the year, and more weed seeds.
"The results of these Farm-Scale Evaluations reveal significant differences in the effect on biodiversity when managing genetically herbicide-tolerant crops as compared to conventional varieties," Dr Firbank said.
"One of the key points to remember is that the results are only applicable to the three crops studied, and only under the regimes of herbicide usage which were employed."
The trials, which tested GM oilseed rape and maize produced by Bayer CropScience, the UK arm of German biotech giant Bayer BAYG.DE AG, and sugar beet made by US agrochemicals producer Monsanto, did not investigate the plants' impact on human health.
Neither have the trials looked at how GM traits might flow into the wider environment through pollen spread - although another team will report on this at a later date.
Dr Firbank: A valuable contribution to the debate
There were protests against the FSEs, and crops in some trial fields were pulled up.
The trial results will now be assessed by Acre (Advisory Committee on Releases to the Environment), the agency that will advise the government on their implications.
A decision by ministers on whether to commercialise the crops could come later this year, or early in 2004.
The GM rape tested was a spring variety. A winter rape is also being investigated and the results of its evaluation will be published next year.
The environmental group Greenpeace criticised the scope of the trials and dismissed them as a political fudge.
It said their outcome did prove, however, that GM corporations were wrong when they claimed herbicide-tolerant plants could have biodiversity benefits through a reduction in the use of agrochemicals.
And executive director Stephen Tindale added: "The real comparison should be between GM and organic agriculture. But organic is so obviously better for the environment that the GM industry refused point blank to have this included in the trials.
"The trials are simply comparing one highly damaging form of agriculture with one that's even worse."
Dr Paul Rylott, of the Agricultural Biotechnology Council (ABC), countered:
"These results confirm what industry has long argued. The flexibility of GM crops allows them to be grown in a way that benefits the environment."
He added: "Activist groups claimed that GM crops were in effect 'green concrete' and would 'wipe out' wildlife.
"These studies show that this sort of scaremongering is not supported by the evidence. On the contrary - this evidence shows that GM crops are more flexible and can enhance biodiversity."
Dr Mark Tester, a senior lecturer in plant sciences, at Cambridge University, commented: "To generalise and say 'all GM is bad' or 'all GM is good' is a crude over-simplification, and these new results provide classic evidence of the complexity of the real issues.
"It's excellent to see such thorough work addressing the effects of GM crops, and this will hopefully inject some rationality into the debate."