Evaluating the farm-scale trials

The findings of the farm-scale evaluation (FSE) of spring-sown genetically modified crops were recently published in a special issue of the Philosophical Transactions of the Royal Society. The FSE was designed to evaluate farmland biodiversity, comparing genetically modified herbicide-tolerant (GMHT) crops to conventional crops.

That the evaluation involved GM crops is not relevant: herbicide-tolerant crops can also be, and indeed have been, developed by conventional methods. GM is a process not a product – and, as demonstrated by this study, each new product (whether it is GM, conventional or organic) needs to be evaluated on a case-by-case basis using rational evidence-based science.

Indeed, the FSE was actually a herbicide experiment, comparing the direct effects of different herbicides and treatment regimes. The robustness of the data surprised the research teams in being remarkably consistent across a wide variety of soil types and weeds in different parts of the country. It was also the first large-scale ecological study of its type.

Over three growing seasons, GMHT maize, sugar and fodder beet, and oil-seed rape were compared in about 65 split fields to their adjacent non-GM equivalents. Similar studies on autumn-sown oil-seed rape should be published in spring 2004.

There is no early or pre-emergence spraying for the GMHT crops, so it is not surprising that the FSE found higher weed numbers and biomass early in the growing season. But for beet and rape the regime for GMHT was found to be more effective at weed control than for the conventional crops, giving lower late-season weed biomass and seeds. Indirect effects further down the food chain on the animals that feed on weeds – such as bees and butterflies – showed a net reduction in numbers.

In maize, however, the opposite result was found: GMHT maize herbicide regimes support more weeds and feeding invertebrates than conventional maize. For all three crops there was an increase in detrivores (animals that feed on the increased amount of weed detritus arising from later spraying). Importantly, differences in biodiversity between crops – for example, maize versus rape – are greater than differences between GMHT and conventional rape, or beet or maize. There were also lower inputs of herbicides for the GMHT crops.

The FSE did not evaluate differences in yields of GMHT compared to conventional crops. But since farmers elsewhere choose GMHT crops we must assume that they outperform conventional varieties. And since GMHT gives better weed control it makes farming more efficient, which allows: more food to be grown, cheaper food, less land under cultivation, increased profits for farmers, reduced farm subsidies, or any combination of the above, according to the rules set by the government. Nonetheless the FSEs were of course presented to us as bad news for the environment by most of the national press and the anti-GM pressure groups.

But let us look at these FSEs in the context of agriculture in the UK as a whole. Any increase in agricultural efficiency must be at the expense of the wild ecosystems with which they compete. Conventional and organic farming both ‘damage the environment’ and ‘reduce biodiversity’. In the UK, 75 per cent of the land is given over to farming, with the remaining 25 per cent shared with cities and towns, roads and managed forestry, leaving little in a virgin wild state of what was once mainly forest. The destruction of hedgerows, the use of fertilisers and silage, the introduction of winter-sowing and of ‘foreign’ crops have all added to the destruction of the natural environment and the reduction in biodiversity.

Thus we are faced with choices in how to provide the balance we desire between wild ecosystems and the agriculture and industry with which they conflict. And clearly we should use the most economical and least damaging methods to achieve this desired balance.

For example, beet and rape support more biodiversity than maize (which is much taller, inhibiting the growth of weeds beneath). Crops in the FSE, in ranging between one and four per cent cultivated land area, are dwarfed in their coverage of the national cropped area by the cereals – for example, winter wheat is about 44 per cent. Yet the cereals, which were not the subject of this FSE, also support less biodiversity than beet and rape.

Surely it is more rational to make large-scale strategic decisions about which crops to grow for biodiversity? Dramatic changes in biodiversity could be effected by changing the crops grown – such as growing less wheat or maize and using beet and rape as break crops more extensively. Relatively small changes here might dwarf the effects on biodiversity of GMHT beet and rape in relation to conventional varieties.

But inefficient agriculture requires more land. If GM technology produces crop varieties with increased yields, another strategic decision would be to maintain current production quotas, so leaving more land for wildlife.

The FSEs set a new precedent in being the first studies of their kind conducted before the implementation of a new technology, allowing a decision as to whether to shut the stable door while the horse is still inside. The studies took three years and cost almost £6million; they have given radical new insights into the ecology of agricultural systems that may allow future predictive mathematical modelling to be performed that may make future evaluation cheaper and more rapid.

However, the studies were based on very specific herbicide regimes. For example, the maize studies indicated that GMHT crops supported greater biodiversity than conventional crops – but atrazine, the herbicide used by 75 per cent of the farmers in the trial, has now been banned. So presumably, any decision as to whether to approve the use of GMHT maize will require its evaluation against the new conventional herbicide regime that replaces atrazine.

Indeed, the logical and most honest extrapolation from this FSE is surely that any future change in a herbicide or spraying regime on any crop – be it GM, conventional or organic – will also require a farm-scale evaluation. If not, what was the point of this recent FSE? But will newly evaluated herbicides remain still-born within an FSE, to become superseded coincident with their approval by the continual need to re-evaluate them against the latest new technology? The FSEs were funded by the public sector. So if this Kafkaesque nightmare is really unleashed upon us, and indeed this new precedent is set, who will pay for these future evaluations?

Meanwhile, back on the organic farm, the ‘organic horse’ bolted long before ‘stable doors’ had been invented: organic farmers use old-fashioned herbicides that, because they are less specific in their toxicity, are rather poisonous beyond their intended victims, the weeds. Yet they are in use, because their adoption preceded the regulations now required for approval of modern herbicides. As we know, organic farmers and consumers are concerned for the environment – surely they should therefore also be willing to submit their herbicides to the same farm-scale evaluations?

Organic farmers should also accept that the process of GM technology, which indeed is by definition a very organic technology, is simply another tool for plant breeding and can also provide future new products. In other words, new plant varieties that can be more readily grown by the ‘organic’ process. For example, GM crops that themselves produce safe pesticides against target herbivorous insects would pre-empt the need to spray with pesticides. It would also make organic food safer by preventing the damage inflicted by feeding insects that allows invasion by fungal pathogens producing carcinogenic toxins. The fact that GM can be used to produce these kinds of crop varieties, suggests that there is no contradiction between organic and GM.

Conrad Lichtenstein is professor of molecular biology at Queen Mary, University of London. He is committed to the public understanding of science within the GM debate and is a member of the CropGen panel.

Read on:

spiked-issue: Genetics