Select Committee on European Communities Minutes of Evidence

Examination of witness (Questions 500 - 519)



  500.  As I understand it, the Bt bacterium is used as a spray in organic farming and no fuss is made about that, but a very great deal of fuss is made when it is introduced as pest resistance into crops. Does that not seem a bit illogical?

  A.  No, not illogical at all. The thing is, if you spray Bt it has a very short life in the field, it breaks down rapidly. Again when I first came into this ten years ago or so, when we first had Bt constructs to talk about, everybody said, "There is no evidence that any insect will develop resistance to Bt. It is a marvellous pesticide. All other pesticides they develop resistance to, but not Bt. We have been spraying it and no resistance has developed." If you put Bt into a plant and keep it there permanently so that it does not decay, you can, in fact, develop Bt resistance quite rapidly, within a few generations. So I think having Bt in the plant really is a very different scene and one which does need careful management and careful control.

Baroness Young of Old Scone

  501.  Could I ask a supplementary, Professor Williamson? You have described one environmental impact and assessed the risk of that, and that is what I would call the rampant species, the "thug" species such as rhododendron. Some commentators have suggested that there might be other impacts, particularly on beneficial insects and also perhaps with broader land use change. I wonder if I could ask for your comments on that?

  A.  Certainly on beneficial insects. One worry that has emerged about Bt in plants—and they actually tried it with the Novartis corn borer—is that if you then feed that corn to herbivorous insects and then feed those herbivorous insects to predatory insects, the predatory insects may show an effect. The same has been shown in Scotland, I think, with a different insecticide in potatoes, where you have ladybirds affected by the aphids. So I think knock-on effects through the food chain to beneficial insects are a possibility and it is one of the things that should be monitored for. Certainly we do not want to decrease the number of beneficial insects if we can possibly help it but modern agriculture does decrease beneficial insects anyhow. When you spray a broad spectrum of insecticides you lose your beneficial insects along with your undesirable insects.

  502.  The other question was on broader land use change.

  A.  I am sure that the development of genetic modification, if it becomes accepted in the European Union, will lead to land use changes, but I would say there all farming changes lead to land use changes. How long ago was it before we did not have yellow fields all over the United Kingdom? Now if you look out of your window you see yellow patches everywhere because oil-seed rape has become very common, and this is a land use change of sorts. Now we are getting more blue fields. I think agriculture is now using science to such an extent that it is going to change irrespective of genetic modification and there will be land use changes. I would not have said that there would be greater land use changes from genetic modification than otherwise, but I would say there will be land use changes from genetic modification. Again it is a question of coming back to the Government to say whether these land use changes are politically acceptable.

Lord Jopling

  503.  Professor, may I follow up an answer you gave to Lord Wade a few moments ago when you talked about a situation, during the process of genetic modification, where you used the phrase, or something like it, of "other genes creeping in".

  A.  That is right.

  504.  Could you explain to a group of non-scientists exactly what that means and how it can happen, with three particular questions in mind: is it always accidental or do scientists realise that other genes may creep into the process of genetic modification and say, "Well, so be it. It doesn't much matter,"; then when other genes creep in, to what extent is it easily detectable, or is it impossible to detect until the effect on the organism of those other genes emerge; and the third side of it is, is it avoidable to have other genes creeping in in the process of genetic modification?

  A.  They are quite deep questions. On the first one, whether it is accidental, the answer is no, it is done on purpose. They put these genes in in order to be able to make the construct and find out what they have made. Perhaps I could compare it with scaffolding on a building. It is difficult to build a building without scaffolding or a high crane or something, but when you have finished the building you take it away. In the genetic constructs that we have coming on the market at the moment, the scaffolding has not been taken away, it has been left in. So it is there deliberately and the people doing it were of the opinion that it was of no great consequence and I think they were surprised when European Committees said they did think it was of some appreciable consequence.

The second question you asked was whether it is detectable. I do not think I am the right man to ask this because it is microbiology but I would say it probably in all cases now is detectable, at a price. With DNA techniques you can detect practically anything in a nucleus if you are prepared to put enough money into developing appropriate probes and that sort of thing. It is not the sort of thing you are going to be able to do on the back of a tractor on a farm. It would be an expensive process as it stands at the moment but it could be done.

The third question you asked is whether it is avoidable, and I think the answer to that is, yes, it is, but you have at the moment to think about it when you start the process. When you first start to put a gene into a plant and you put it into the plasmid to be carried into the plant, you put in with the gene you are going to use—the Bt gene or whatever it is—an antibiotic gene or a herbicide gene or some other thing you are going to use, to allow you to select your right lines, you can flank that gene with what you might call weak links, things which are attackable later on. So you could make your original construct in such a way that you could effectively get out the scaffolding later on. But people are not doing that yet and getting it out now in the constructs that we have in the market would be very difficult for companies. I doubt it would be impossible but it would certainly delay them five years or more, I would imagine, while they went back and redid it from scratch. Having done it once they would know where they were going a bit faster but it would still take several years of breeding to succeed.

Chairman]  You will be able to see that, Lord Jopling, in detail and carry the discussion further when we visit the John Innes Centre in Norwich tomorrow.

Baroness Young of Old Scone

  505.  Could we cast our minds forward a bit. You have already touched on the issue of whether it was the wisest thing to start with a set of objectives that appear not to have as many public benefits as might be the case with other modifications. The concentration so far has been on herbicide tolerant and pest resistant crops. Are there other modifications that you think could be of less or different environmental concern, and the three groups I would like you to think about are the groups that allow crops to tolerate things that they otherwise do not: firstly, fungus resistance, salt and frost tolerance; secondly, the manipulation of nutritional property and quality standard, and thirdly, pharmaceutical production?

  A.  In your first category you put in fungus resistance, which I would regard as just another kind of pest resistance, but the other two you put in that category, the salt and frost tolerant, are, I think, developments which would need very careful ecological monitoring. At the moment, for instance, with maize, most maize grown in the United Kingdom at its northern limit needs a good summer to do well and that sort of thing and we are unlikely to get any escapes. I do not think you get many volunteers. There have been a couple of records of casual maize but not very many. If we got to a much more cold tolerant maize, I think we would get many more, so there will be a greater prospect of escape from modified maize crops if they were cold tolerant. Similarly with salt tolerant plants, you might well get some occurring in salty habitats, which include, of course, the verges of motorways, and you might get a nasty problem cropping up simply because of the salt tolerance. So I think salt and frost tolerance are likely to produce larger problems than herbicide tolerance certainly, problems of a different sort but maybe of the same problematical nature as pesticide tolerance.

In your second category on nutritional property, maybe you are thinking of Flavoursaver tomatoes and this sort of crop. I think most nutritional changes which occur in the edible parts of plants—leaves and roots and things—are not likely to lead to environmental problems, I think less likely so than other ones. But nutritional problems in seeds might. One of the earlier ones we had along was when they wanted to put a Brazil nut protein into oil-seed rape seeds. I do not think that actually developed anywhere but that could possibly be a problem. With oil-seed rape, you see it along roadsides, it produces masses of seeds into roadsides and yet hardly ever forms permanent populations. I think there are a few places where it might. This appears to be because mice and snails and I do not know what—fungi maybe—come and attack the seeds. If you change the chemical composition of the seed it might be that these things that eat it will no longer eat it, in which case you might then get a worse problem from oil-seed rape. So I do think we would need with those sorts of properties to try and develop experiments on seed survival, if possible, to see if we were likely to have problems with that.

The third one is a total black box, pharmaceutical products, but it is the same issue. A pharmaceutical product, as I said earlier, I think would be desirable but might nevertheless have ecological side-effects, and again because it might affect the population dynamics of the plant. The same would apply for plants producing plastics, which has been a suggestion. You would have a totally new chemical inside the plant which the herbivores of that plant had never met before and you do not know what effect that would have on the herbivores. The likely answer is it will have no effect at all but it might cause some indigestion so they stopped eating it or might even kill them. In which case you then get a really serious problem from feral crops, but again what I think you should do is experiment and find out.

Lord Rathcavan

  506.  Professor Williamson, you have made it known that you consider risk assessment based on hypothetical questions and small-scale trials to be effectively useless. What alternatives might there be?

  A.  If I could come back to your question, I was not actually referring to hypothetical questions. It will come in when we get to the end of the questions. It is the questions that DG XI ask us to ask which strike me as useless, because they are too qualitative. I think one should be asking quantitative questions. Small-scale trials, when I say they are useless, this is in relation to ecological problems. Small-scale trials are certainly useful for telling us whether the plant is behaving in the way it should behave as a physiological plant, whether it is growing properly and that sort of thing, but it does not usually tell us anything at all about the likely ecological behaviour. The question, I think, should be very much sharply focused on what possible ecological risks there are. I will be coming back to this, but I think one of the problems is that all the things which you might think would predict whether a plant will be invasive—the rate of reproduction, the number of seeds produced and that sort of thing—are on the whole rather bad predictors. So it is difficult to suggest to companies that they should put a lot of money into measuring things quantitatively, which may in the end be of no great use. Nevertheless, I think one should do some measurements of this sort. They should be quantitative measurements. We do want to know about the seed production of these modified plants; we do want to know about seed survival of these plants. It would be a better thing to have some idea about whether the dispersion patterns are different and that sort of thing. Primarily it is the immediate factors affecting the fitness of the feral plant you need to ask questions about and which the DG XI questions signally fail to ask about.

  507.  Would you regard invasive properties and the unknown outcomes in this field as being the most serious threat to identify?

  A.  From the ecological point of view it is the threat of invasion either by the plant itself or by the genes from the plant which have an immediate effect. You can only get secondary effects building up on the back of that, as we were talking about with predators. Certainly with virus resistant plants there are very considerable worries that they might, out in the wild, form new virus diseases. But all these are, in the way I use it, new invasions. So yes, I would say in general we are worried about biological material invading ecosystems in a novel way and are attempting to get a handle on how likely it is. I do think, however well you design your question and run your small-scale experiment, you are not really going to get very useful answers out of small-scale trials, however well tuned your questions are. We are going to have to go to larger-scale trials and monitoring. Perhaps I could use an analogy here with drugs in medicine. With medicine you start off with a small-scale trial to see if a drug does what you want it to do and then you go to larger-scale trials. It is usually in the larger-scale trials that you will find the nasty effects that you do not want to find, and the phase three trials is where you may lose your drive because of unacceptable effects, and I think this may happen with some of these modified plants. It is at the large scale where you are likely to have nasty effects which have to be cleaned up rapidly, but the probability of this is still small.


  508.  So large-scale trials means thousands of acres, does it?

  A.  Yes, commercial size.

  509.  It means commercial exploitation? You could not have such trials without commercial exploitation?

  A.  I think it would have to be done commercially and involve a lot of commercial farmers before you got to that scale. But it could still be that you might have a provisional licence to the farmers, that you could try growing this crop for such-and-such a period and see what effects there are. Then if you find no effects, you could say, "Okay, we will put it out for full-scale release." That will not, of course, necessarily protect you. A lot of these effects come on rather slowly and you may get a nasty surprise ten or 20 years down the line.

  510.  And these conditions would be covered under a monitoring system?

  A.  Yes, you would have to do the monitoring of the crops.

Baroness Young of Old Scone

  511.  Could I follow this up. You said as a result of large-scale commercial trialling, second-stage trialling, as it were, if adverse effects were detected there would be a need to clean up rapidly. How real do you think the possibility of a clean-up is? We have had evidence from other people who have said that once the genie is out of the bottle it is out and you are not able to get it back in again.

  A.  I think it is possible. It is difficult with plants. I could send you perhaps a manuscript I have on eradication of environmental problems. It is certainly true that if it is a large organism occurring in a well-defined area such as an island, it is much easier to eradicate it and most successful eradications have been done in this way, but not all of them. There is a famous case where African mosquitoes were eradicated in Brazil after they had covered a large portion of Brazil. That was an enormous exercise in the 1930s. So I think eradication of a feral crop plant would be possible but it would be very expensive.

Lord Grantchester

  512.  Could I expand this question of risk assessment. In your paper you ask a question yourself: "Is it possible to quantify the probability of a hazard happening; to quantify the risk to agricultural systems, the general environment and society a whole?" Could I ask you to answer your own question?

  A.  Yes. I think the paper singularly fails to answer the question at the end but papers often do that. I think at the moment one of the worries I have with the ACRE system is where we go through—I say "we"; I was with ACRE, I am no longer there—but ACRE goes through what it calls a risk assessment. Engineering friends who do real risk assessments of whether a chemical plant will go bang or a building fall down say these are not really risk assessments because they are not properly quantitative. At the moment it is very difficult to get enough good quantitative information of the right sort to do what you can really call a risk assessment, but it is possible to make some progress if you try a little harder than some people have. I would not go further than that. I think we can do better but I do not think we can do well.

Lord Willoughby de Broke

  513.  Professor Williamson, we touched on monitoring earlier on and I know you chaired the ACRE sub-committee on monitoring. Can you explain what monitoring could achieve, how it could be done, by whom and for what purpose?

  A.  ACRE produced a series of reports that look like this. I do not know if you have had the privilege of seeing them. I put "privilege" in inverted commas, I think; they are extraordinarily dull documents, but they do say, for example, "Guidance for experimental releases". They were concerned with small-scale monitoring, I think, in all those documents and we are now talking about both small-scale monitoring and large-scale monitoring. But I think broadly speaking we monitor for two reasons. One is compliance, that is to say if the product is let out under certain conditions, limited trials of some sort, and even a commercial trial can be under clearly defined conditions, then you monitor to see if the companies are obeying these conditions. The second reason we monitor is to see if there are unexpected effects. So you go out and say, "We don't expect any hard effect of this sort but there might be so we will go out in the areas surrounding the field and see what we can find", or "we will rely on an army of reporters who will be around the field and see what they can find." Again these are normal practices in medical trials. In medical trials and, as you have probably heard, in agricultural trials, on compliance, although companies tell us they are all splendid, some companies sometimes are not as splendid as they should be and we find that they do slip up and fail to adhere to the conditions. So monitoring even of small-scale trials does show up lack of compliance. On the small-scale trials you are not going to show up, as I said before, much of ecological interest; it would be of physiological interest. You need a larger scale to get to where you can monitor for ecological effects of interest. Does that help?

  514.  Yes. Could I follow that up by asking if you are aware of the monitoring of environmental impact in any other countries and whether it is effective?

  A.  I think the answer is I am not really in the field enough. I was involved with OECD about ten years ago and then I got to know what was happening in most OECD countries. I suspect where this technique is being used most rapidly is places like China, India and Israel, which are not even OECD countries. I anticipated your question and on Monday attempted to surf the web and looked for sites which are run by the United States Department of Agriculture to see what they said about monitoring and I failed to find anything at all. So I sent an e-mail but the problem with sending e-mails to America is it is rather late in our day and early in their day when you send them. I sent an e-mail on Monday to say could they tell me more about monitoring and by the time I left on Monday I did not have an answer. I was not in the Department on Tuesday, so I have to say except for monitoring for pesticide resistance round some of the crops which we do know is going on in America, I do not know of any monitoring. I would be surprised if in America there was much commercial size monitoring because most of the releases are done under a notification and not under a permit and that in itself implies that they do not think there is any need to monitor. So I think you will find that monitoring at the moment is likely to be rather a European phenomenon, but certainly not entirely so. I am going out to Australia maybe in the New Year and I am going to talk to people out there. I am told that genetic modification is a hot potato and I must be careful about what I say, so I may discover later on what the Australians are doing about monitoring but I am afraid I do not know at the moment.

  515.  Thank you very much for taking so much trouble.

  A.  It is no trouble to read it up. It is more trouble to find out if it is going to give you any useful information.


  516.  Nevertheless, in your published work you refer to the desirability, in your view, of having an international monitoring system. Could you say perhaps what you had in mind?

  A.  Yes. I came across two analogies with this one. I think the more helpful one is again in the medical field where the World Health Organisation does run an information network; they have a database in Sweden, I believe it is. I have something from the British Medical Journal of 1997 which describes it, which I can put into your papers if you wish. There you are looking for unusual adverse drug reactions and you get reports in. You can get many reports on adverse reactions having taken the drug, but, of course, many of them will not be adverse drug reactions; they will be adverse reactions of some other sort. So you have to sift out your problem. This is a major problem for the World Health Organisation and they have devised computer algorithms to sift through the data coming through. So that sort of system does exist at the United Nations level, and I suspect if one is going to trial internationally it would have to be a United Nations agency of some sort which did it. That is an information system and an international monitoring system. You used the word "separate" there. I did not have strong views on whether it should be separate or combined, but I do think if you set up a reporting network of some sort in a sense it is doing some of the monitoring for you. Again, if we do the large scale trials in this country, although we might require the companies to do some direct monitoring, I think we will be relying on the statutory agencies and on the agricultural community and on amateur naturalists and all sorts of people to form a reporting network if it is going to be effective. The thing about commercial trials is that it has to be big if you are going to have a probability of picking up the effect, and that means you will have to have an awful lot of people out there looking at the effect. They have to be told that people would like to know about these effects, they have to be told how to report them if they find them. I think this comes back to Lady Young's empire.

Baroness Young of Old Scone

  517.  I was wondering if we could get Professor Williamson to give us a flavour of the sorts of things he would see this monitoring network gathering data on?

  A.  If we are dealing with crop plants, the primary thing we want to know is whether the plant itself is forming self-sustaining populations outside agriculture. We also want to know for those which cross-breed with plants which are there already, whether they are getting interesting new hybrids. At the moment, the compliance monitoring is done by the Health and Safety Executive's inspectors who are certainly not expert botanists. They can see if the crop has been planted in the wrong place but I think they will be hard-pushed to distinguish oil-seed rape from wild turnip; it is not that easy. Who you need to do this are expert botanists, either professional botanists employed by English Nature, or amateur botanists in the Botanical Society of the British Isles. I suppose the analogy, and again you might be looking for effects this way, are with the bird surveys. I think the best surveys we have had in this country have been surveys of bird populations, which have certainly shown the massive environmental effects of agriculture, the massive decline in bird populations which you have probably all read about in the papers. These are done under contract from the JNCC, I believe, by the British Trust for Ornithology who are professionals who organise amateurs to count the birds. You have a whole army of amateurs out there counting birds, reporting back. Of course there is the recording and entering of it into data bases and analysing them, so it does cost an appreciable sum of money. You could say in this context it would be up to the companies to fund English Nature to fund the British Trust for Ornithology, though I have worries about that too with the corruption you get from contracts. But I do think it is going to have to be done by bringing in a large number of people. We might require companies to do something directly as well.

Lord Wade of Chorlton

  518.  You were referring earlier to therapeutic products and there is quite clearly a system laid down for them, with the various stages, number of patients to be treated and the number of volunteers to be found, all laid out in the legislation. The companies which produce the products are aware of the potential cost of the process and it is built into their costings and they can see the benefits coming out at the end of the day. Would you say that that basic principle is one which could be adopted for genetically modified crops and it could be laid down exactly what size of acreage, for how long and what issues should be monitored, so it is part of the company's overall cost to get that product into the market place? On the other hand, the pharmaceutical industry can carry high costs but possibly food companies could not, so in trying to give more confidence to consumers would it hamper it rather than support the product?

  A.  There are a lot of threads there. I think the answer is yes, the companies should allow for the cost. Whenever I talk to companies about this, they have always said that what they would like is firm and clear regulations so they can see what the cost is going to be and allow for it, or indeed, as you say, decide whether it is more expensive than they care for. I get the impression, reading the financial papers casually—I cannot say I read them at all intensively—that some of these companies are expecting to make a great deal of money out of this and therefore I would have thought they could take on the costs of some of these things at the same time. But it is a question of doing the calculation and seeing what the cost comes out to be. The problem is doing the regulation and saying what it is. You are talking about what the acreage should be and who should do the monitoring and the trouble is we are still very much on a case by case basis. With the Novartis insecticide resistant corn, the concern that came up was that we might get antibiotic resistance developing inside the guts of cows. I do not know if one is going to require the companies to go around and sample the gut contents of a large number of cows. It obviously can be done but it is not what they were geared up to do, they would have to bring in another section of their company or another company altogether. It would be very difficult, I think, to have a regulation which would indicate how many cows and for how long you would have to monitor for this. The company would have done better to have avoided the problem in the first place by excluding the antibiotic gene, as I was explaining to Lord Jopling, but the companies did not think of this when they were starting their development a decade or more ago. Yes, I think the companies should be expected to allow for it. I think it is not feasible to be able to give as firm an indication as the companies would like, and I would have thought the answer for the companies would be, as in many cases where you have unknown risks, that you would do it on an insurance basis. The Government of course insures itself and does not take out an insurance policy, smaller companies would inevitably take out an insurance policy for all sorts of risks, larger companies would no doubt have to decide whether they wish to have an insurance company for an unknown expected large thing or whether, being large enough, whatever comes they will be able to weather it.

Lord Grantchester

  519.  Following on from your point about antibiotics and cows, what I would like to ask you is, is the furthering of antibiotic resistance through genetically modified food a genuine cause for concern? How can the least safe practices be identified and phased out?

  A.  Is it a genuine cause for concern? It depends on the antibiotic resistance. If you have—and I am sorry if I am going to say things which you know very well already—a bacterium which is resistant to an antibiotic, it does it by having a gene that produces a product that fights the antibiotic in one of a number of ways. Then you can put that gene into a plant as a marker in the development of the plant. The gene will be under the control of a thing called a promoter and promoters are a branch of witchcraft in which I am afraid I am not trained. But I can say there is a difference between promoters which work in bacteria only, which are the sort which a bacterium normally has when it starts off, and promoters which will work in plants only. Most of the promoters which have been used for putting in antibiotic resistance genes in plants, are plant virus promoters, particularly the cauliflower mosaic virus promoter, and these on the whole will only work in plants. There are, of course, other promoters which will generally work in plants but which sometimes work in bacteria, or might even work in mammals, which is why I say it is all witchcraft. The only way to find out whether a promoter will work in a different organism at the moment is to try it. I have been critical of some people who say, "This product is safe because of such-and-such", before having done an experiment. I think with a bacterial antibiotic resistance gene in a plant under a plant-specific promoter, the concerns are very small indeed. It has got to get out of the plant into a bacterium, lose that promoter, gain another promoter, and then it will start acting. All that is remote, as the industry is likely to say, which means a very small probability. It is not impossible, practically nothing in this field is impossible, but it is very unlikely. The problem with the Novartis product is that they put in a bacterial promoter and that of course, when it gets into the cow's gut, could well get into a bacterium and start working straight away. That is very much more a cause for concern and that is why the British position on this was that there really was cause for concern and something should be done about it. So, again, it is a question of a particular case, because with antibiotic resistance there is always some slight cause for concern, usually very slight. In some cases using bacterial promoters is a considerable cause of concern. If I could take another case, where they are putting constructs into insect viruses. There they are putting in promoters which will drive the scorpion toxin gene, and the promoter is one which is new to that virus and new to the insect receiving it. There is a mild concern at the back of my mind that the product there actually could be toxic to some mammals, and that the promoter could conceivably work as a mammal promoter. It does not look like a mammal promoter but as far as I know people have not tried to see whether that new promoter does work in mammals. I think the answer to all these is that we need more focused experiments on the lines of, "Here is a possible risk, what experiment can we do to test it?" In that particular case I think the optimum test is possibly promoting activity in mammal cells, and I do not think they have done it. Did I answer the first part of your question adequately?

previous page contents next page

House of Lords home page Parliament home page House of Commons home page search page enquiries

© Parliamentary copyright 1999