The Case Against GM

You hear a lot about “sound science” these days, mostly from the GM lobby and other people who are trying to convince you that you should put your faith in scientists - by which they mean themselves and their friends - and let them get on with the wonders they are going to bring us.

Curiously, “sound” isn’t a word scientists have traditionally used very much. The term first appeared in 1993 when Philip Morris and its PR firm founded a non-profit group called TASSC: The Advancement of Sound Science Coalition to fight against the regulation of cigarettes. Other people with similar aims have picked it up, and it now seems to be used mostly to mean arguments intended to convince the public that something that industry wants to do is safe.

Whatever word you prefer, we in the ISP are certainly interested in good science, in general and especially in decision making. But we don’t think the public should be quite so willing to take things on trust.

The public can and should demand that scientists justify their positions and they should participate in taking decisions in which science is involved. You don’t have to be a specialist to ask the questions, to judge whether the person you ask is addressing the question or avoiding it, or even to judge between two conflicting claims. As in a court of law, the people making the judgement don’t need specialist knowledge if they insist that the two sides each make their case as clearly as possible.

As an example, let me take you through a recent issue which may still be fresh in your mind, the UK government’s recent decision to allow commercial growing of maize that has been genetically modified so as to make it tolerant of certain herbicides.

Now there are lots of reasons why we shouldn’t be growing GM crops, certainly not until we are a lot more confident about the technology by which they are created, their effects on the environment or even our bodies by horizontal gene transfer, what happens in succeeding generations as the genome alters, allergies, and so on. We should also be asking whether the claimed gains are real or hypothetical, whether they could be achieved by improvements in conventional farming, and so on.

Ignoring all this, the government announced there was to be a series of Farm Scale Evaluations. If GM crops turned out to be less harmful to biodiversity, as measured by such indicators as the numbers of weeds that survived, that would prove they were beneficial and they would be licensed for commercial growing in the UK.

To fasten on this one point, important enough but only one among many important issues, seemed to us to be rigging the game in favour of GM. A bit like saying that the MOT should consist of testing only the tyres. No need to worry about the brakes or the steering.

Anyway the tests went ahead. The crops were grown on split fields, GM on one side and conventional on the other. Herbicides were applied to both, according, so we are assured, to a strict protocol.

It turned out that the biodiversity was less with GM oil seed rape and sugar beet than with conventionally managed crops. But the reverse was true for maize. Here there was more biodiversity on the GM plots than on the conventional ones. Accordingly, the Government announced that it would licence GM maize for commercial production, but not GM oil seed rape or sugar beet.

When you think about it, this was the best possible result for the GM lobby. On the one hand, it allowed the government to establish the principle that GM crops can be grown in the UK, which is surely more important to the industry than whatever benefits there may be from GM maize itself. At the same time, the government could claim to be being suitably precautionary, by not allowing the two other GM crops.

Just as they were about to announce this, however, there was an apparently unexpected development. The maize trials had involved 44 plots, and on 40 of these, the conventional maize had been managed with triazine based herbicides: atrazine, simazine and cyanazine. The EU announced that these were to be withdrawn from its list of approved chemicals. In future, farmers will not be allowed to use them.

(I said “apparently unexpected”, because I’m surprised that the researchers hadn’t known that the ban was likely to be imposed when they started the trials. The EU doesn’t move that quickly on regulation. Or perhaps they did know.)

In any case, the effect on biodiversity of the conventional management of maize had been measured using herbicides which were now to be banned. This threw the whole trial into doubt. As the researchers themselves later wrote:

Withdrawal of the triazine herbicides atrazine, simazine and cyanazine from approved lists of EU chemicals could therefore reduce or even reverse the reported benefits of GMHT [genetically modified herbicide tolerant] maize.

Now you would have thought the next step was obvious. If you accept - as I don’t but the government does - that the only legitimate objection you could have to GMHT maize is its effect on biodiversity as measured in these trials, and if the comparisons with conventional maize were carried out under conditions that no longer apply, then you should repeat the trials using the herbicides that farmers will use in the future. That’s going to involve some extra expense and a delay of at least one year, but if we’re trying to be precautionary, it must be worth it.

But the researchers found - or claimed they had found - a better way, and at the beginning of March a paper appeared in the on-line version of the journal Nature with the title:

Ban on triazine herbicides likely to reduce but not negate relative benefits of GHMT maize cropping.

There’s a lot of scientific words and statistics in the paper, but in fact it’s not hard to see what they did. So let’s go through it.

It happens that while triazine herbicides had been used on 40 of the 44 conventional plots, non-triazine herbicides had been used on the remaining 4. So there actually are some relevant data, which you can analyse.

When you do this, however, you find no significant difference between the measured biodiversity when GMHT maize was used and when conventional maize was used with non-triazine herbicide. So you can’t infer from the results of the trials that GMHT maize confers any benefits in this regard.

Nothing daunted, the researchers tried another approach. They pointed out that of the 40 plots on which triazine herbicides were used, it was applied pre-emergence (i.e. before any weeds emerged) on 16 and only post-emergence on the remaining 24. It was also applied post-emergence on the 4 non-triazine plots. As you’d expect, pre-emergence application tends to reduce the weed abundance.

They took the data from all the conventional plots on which herbicides were applied post-emergence, compared that with the results from the GMHT maize, found a significant difference, and concluded that the benefits of GMHT would be reduced but not negated by the banning of triazine herbicides.

Really? Surely what they had shown - if anything - is that the benefits of GMHT would be reduced if farmers were to stop using triazine herbicides pre-emergence. That’s interesting, if not surprising, but it’s not what the paper claims to be about.

On the one hand, there are 44 plots on which GMHT maize was grown. On the other, there are 28 conventional plots, of which only 4 were treated with non-triazine herbicides. How can you possibly claim that this tells us about the comparison between GMHT and the non-triazine herbicides which farmers are going to have to use?

Only if we suppose that there is no difference between the effects of triazine and non-triazine, surely. But that is what the paper is supposed to demonstrate, and now we find that it is actually an assumption. To quote from the paper:

“If this pooled category of herbicide regimes is indeed representative of weed control in post-triazine crops, and if the weed management in GMHT maize remains the same as observed within the FSE, then final weed numbers would still be larger than in conventional maize.”

Is it likely that what the authors call the pooled group is actually representative of what would happen with non-triazines? I don’t know, but such data as there are suggest the contrary. The authors themselves write:

“For the other conventional regimes, weed abundance for the few sites that were treated only with non-triazines was consistently slightly greater than those for the triazine.”

The question addressed by this paper was “Are the results of the Farm Scale Evaluations for maize still valid if triazine herbicides are replaced by non-triazine?” The answer is “Yes, providing we assume a priori that triazine and non-triazine herbicides are equivalent.”

That’s obviously true but it doesn’t actually tell us anything. Yet this paper passed peer review quickly and appeared on Nature’s web site just in time to be used to justify the decision to allow GM maize to be grown commercially in the UK. It was widely referred to by the media and by ministers, except that all they mentioned was the claim you find in the title and in the first paragraph (which in Nature serves as an abstract). Only if you read the paper do you find out what they really did. And you don’t have to be a scientist to understand that; you only have to be able to read.

There were other weaknesses in the trials too. For example, in any work on health, safety and the environment you have to allow for the fact that what happens in the real world may be quite different from what happens under idealised test conditions. Farmers are, understandably, interested in yields. The whole point of using herbicides is not to have lots of weeds competing with your crops. If you don’t care about the yield, you can use as little herbicide as you like, but that won’t tell you anything about what a real farmer will do. So how can we know how realistic the trials were if we don’t know how the yields on the GM and non-GM plots compared? Yet there are no data for yields.

The best I can do is to show you a few slides that were taken by people not connected with the experiment. The GM crops don’t seem to be doing very well. I don’t know if that’s because so many weeds have been left, but I certainly don’t see that the two half-plots can be considered equivalent.

A test that doesn’t touch most of the potential dangers of GM crops and doesn’t deal properly with the issues it does address. That’s the basis on which our government takes crucial decisions. It’s just not good enough.