 |
Agriculture and Environment Biotechnology Commission (AEBC) | |||
 |
Meetings |
 |
 |
|
|
FOURTEENTH COMMISSION MEETING ROYAL
SOCIETY OF EDINBURGH, 22-26 GEORGE STREET (NB: AUDIBILITY DIFFICULT WHEN THE MICROPHONE WAS NOT USED) CHAIRMAN Welcome
to the Thursday morning meeting of the AEBC.
I am sorry this is distorting a bit, but the reason I am speaking
into this is that we will be taping the session that is about to follow. We
have this morning the pleasure to welcome to the meeting of the Commission
Professor Joe Perry, who is at Rothamsted Research, and Dr Mike Wilkinson,
who is the Reader in Plant Genetics in the University of Reading.
They are going to be each of them providing us with a short
presentation and then there will be an opportunity for us to pose some
questions to them. This
session is part of the evidence-gathering by the Consumer Choice
sub-group, and members of the Commission will have amongst their papers a
paper each by Mike Wilkinson and Joe Perry setting out some of the
background to their research. I
am going to invite Joe to talk to us first.
PROFESSOR
JOE PERRY Thanks
very much. There are two
things I wanted to talk about today, both to do with separation distances.
The first thing, which is not in the papers that you have had, is
the possible effect of insects and how that might affect our ability, by
increasing separation distances, to push down the degree of
cross-pollination. This graph
shows the degree of cross-pollination on the y-axis, plotted against
distance downwind here on the x-axis, and describes the results from a
simple model that I have built. Now
I think it is fair to say that the relationship between cross-pollination
and distance downwind is known about through trial and error, but it is
not an area of science which has a great deal of evidence associated with
it in the literature. There
is some material in the literature, but not a huge amount.
Really, we are kind of catching up since about the mid-1990s when
people became very interested in this area for reasons concerned with GM
and there is more recent research that has been done.
There is also an important paper recently published by Rieger from
Australia, work on oilseed rape, that I will try and come back to. Do
remind me if I don’t. But
for the moment let’s say that we understand what goes on maybe up to 200
metres and maybe down to about 1% cross-pollination quite well, again
mainly through trial and error, mainly through commercial activities, the
need to retain seed purity. But
we don’t, I think, understand what goes on beyond that distance, and in
particular there have been various reports in the media, and in the
scientific literature, about the possible effects of insects, in taking
viable pollen relatively long distances downwind.
It is not just insects, there are reports of relatively unusual
weather patterns such as convection currents, weather fronts coming
through, etc. So that
is really the background for this model. The
straight line here is what we might expect with exponential decline of
cross-pollination as we go downwind, and so (to make sure that I quote
this absolutely accurately, in front of such an audience I don’t want to
get this wrong), for the straight line I have assumed normal windborne
dispersion, and there we have got an average dispersal distance downwind
of about 37.5 metres and that would lead to 10% cross-pollination in the
first 86 metres, 1% at 173 metres, and 0.1% at 260 metres.
So under that model, by having a separation distance of about 260
metres, we could push down the degree of cross-pollination of conventional
or organic crops by GM down to 0.1%.
That is all very well, but what about the possible effect of these
long distance movements of insects, such as bees, taking viable pollen? So
what I am going to do is to, on top of that wind dispersed exponential
distribution , I am going to postulate another distribution for
insect-dispersed pollen and that has got a much longer tail, so the
average distance that I am assuming now is about 500 metres, but there
will not be many events which are mediated by insects or by unusual
weather. So let’s assume
that that is just 1% of all the cross-pollination events are mediated by
insects or some other long distance dispersal mechanism.
This makes very little difference to the 1% cross-pollination up
until we get 200 metres downwind. Look
at these two graphs: the red is just the wind-dispersed alone, and the
blue is the mixture where I have got the wind-dispersed plus just 1% of
this extra component which is long distance movement.
The point to note here is that the curves at relatively short
distances, up to say 200 metres in this model, coincide, so there is no
way of knowing, if you are in this area, whether we have wind dispersal
alone or whether we have wind with a bit of insect also.
But
the point is, looking back at the original graph, these are the results
from wind alone and the dotted line is the results from the mixture
distribution. So the two
curves coincide up to 200 metres, but beyond 200 metres they begin to
diverge, not only do they diverge, but we can see that the insect one with
long distance transport reaches a sort of plateau, and so we can see that
in order to push the degree of cross-pollination down to 0.1% we would
need to go a lot further than the 260 metres that we had before under the
wind-alone model, right out to over a kilometre. So
what is the message here? The
message is that at the moment I don’t think we have got the evidence to
say what situation we are in. The
evidence is not good for whether we are in windborne alone or wind plus
insects. Previously we were
not concerned, so that is why I think there hadn’t been these studies
done. If you want to study cross-pollination at very low levels you
are going to need a lot of effort. It
is not the sort of experiment you would go into lightly because to detect
such small percentages you really need a lot of man or woman power out in
the field collecting data, it is a very expensive thing to do. QUESTION: I
am just wondering why does the insect borne population only go downwind?
I can’t see that, insects are fairly happy to fly in all
directions. PROFESSOR
PERRY No,
you are absolutely right. QUESTION: And
therefore if you looked upwind you would be able to distinguish the
difference between windborne and insect borne pollination. PROFESSOR
PERRY Yes,
but to actually find insect borne pollination upwind still requires a
great deal of effort because then you are looking for very small
percentages. But you are
absolutely right, insects don’t just travel downwind. So
the message there then I think is: but before I go back to the message I
should say that studies have been commissioned by DEFRA and are going on,
are starting fairly soon, I think next year, into looking at long distance
pollination in oilseed rape and in maize, and hopefully we will have a
much better idea after those, but of course they will take about 3 or 4
years. The
second thing I should say is that there is recently a very extensive
paper, published by Mary Rieger, looking at the percentage
cross-pollination of oilseed rape in Australia and that had a message
which is part-optimistic and part-pessimistic I think concerning this
debate. The plus was that actually the percentages that she detected
were relatively quite low, often below 0.1%, 0.03% typically, but there
wasn’t much evidence of a sharp decline as you increased the separation
distances. So
the overall message here I think is that the decline in cross-pollination
at distances up to 200 metres may or may not apply at greater distances,
you might get these relatively long distance cross-pollination events and
they may induce strong non-linearities into this relationship between the
degree of cross-pollination and the separation distance required to
achieve very low levels. And
then if that were the case it would be only with very substantial
increases in separation distances that you could achieve the very low
levels of contamination of 0.1% that the government have recently
mentioned. And
finally on that particular subject there is also the question of cost and
I alluded to it before in saying the reason that people haven’t done
this sort of work in the past is a mixture of need and cost. Now generally, to detect a particular percentage
contamination you would want to take a sample, and quite a large sample,
to estimate low contamination rates, and in general, I won’t go through
this in any detail, all I will say is that in general there is a
relationship between the degree of purity that you wish to attain and the
sample size that you need to detect contamination, and in general that
relationship is that if you want to increase your purity level by tenfold,
say from 1% to 0.1%, you would need to increase your sampling effort by
about the same degree. This
analysis does not take account of the type 2 error.
There is a type 1 error, you don’t want to allow through a batch
which is contaminated and deem it as being pure, so you want to take
sufficient samples. On the
other hand there is the type 2 error, you also don’t want to condemn a
batch as impure where the actual parent population is OK. So, in general
then, if you want a tenfold increase in purity you need a tenfold increase
in sample size, and of course there are cost implications to that. That is the first bit of work that I wanted to describe.
And
the second part is the part that you I think got from the paper, and this
really asks the question: Well, if separation distances were increased,
what would be the implications for the government’s declared policy of
if GM were commercialised, of co-existence between the various forms of
farming? So this is a very
simple sort of exercise in which we say: well, let’s assume we have got
a number of farms. Now you
could either look at this from the point of view of existing GM farms
needing some distance around them within which you would not wish to grow
organic, or you could look at it from the other point of view.
Since GM is not yet commercialised I will look at it from the point
of view of organic farms and a distance around them in which you would not
be allowed to grow GM, defined by the separation distance.
So let’s assume that these black squares are organic fields and
that we have a certain separation distance, so we would not be allowed to
grow GM in these grey areas which surround the black organic fields. Now in this particular example we have got a large white
area, so we have got a high proportion of land which is still available to
grow GM if we wanted to. But
obviously that will decrease with two factors:
first of all it would decrease with the number of organic farms, so
the more organic farms we have got, proportionately the less space there
is to grow GM; and of course
if the separation distances are increased that again would lead to a
decrease in the area in which you would be able to grow GM.
So in this exercise I just took a random distribution of organic
farms and looked at the proportion of area, effectively which is white, on
a relatively large scale, and we need to do this spatially and with
simulations because of course there is often overlap of the areas bounded
by the separation distances. So
as we would expect, there is a relationship between the proportion of land
that is available for GM cropping and the separation distance, and as the
separation distance increases then the amount of land available generally
decreases, but the surprising thing, I suppose the interesting thing for
me scientifically, is that it is rather sensitively dependent on how much,
in this case organic, agriculture you have got to start with.
So
here we have got a number of scenarios.
This one, A, is roughly equivalent to what you would expect for
maize where you have got about 3% of land which is organically farmed and
maybe 1% of organic farmers will grow organic sweetcorn, and this shows
really that not only at current separation distances, but even if we went
out to the 3 kilometres proposed by the Soil Association you would
effectively have co-existence. As
we move round, however, the scenarios change.
Scenario B represents what would happen if 10% of those farmers
that were organic decided to grow sweetcorn, and there of course we can
see that an increase in separation distances is beginning to have a bit of
an effect, maybe not a great effect, and then as we come round to scenario
D here we have got 70% of organic farmers growing sweetcorn.
But of course the 3% of farms which are organic at the moment is
predicted to increase. The
target of the organic movement is 30%, but there is debate about whether
that will be achieved through economic pressures.
Let’s say it goes up to 20%, that would be scenario C, you have
got 10% of those 20% of farmers that farm organically growing sweetcorn,
and here, in scenario E, 70% grow sweetcorn.
And when we are around these two scenarios C and D, we can see that
with the separation distance as it is at the moment, we have got about
half of the land available for GM, which is reasonable if the government
wanted to commercialise and ensure co-existence, but with an increase in
separation distance, the proportion comes right down, and of course these
are on logit scales, so here we have got about 0.05 and then you would
really question whether co-existence was possible. So
just to summarise, there may be, if we increase separation distances,
there may be limited space to grow both organic and GMHT crops
simultaneously, and that may conflict with the government’s stated
policy of co-existence. Also,
that a decline in cross-pollination at short range may not apply at great
distances and a small amount of weather or insect pollination may have a
relatively big effect, and then we might require separation distances of
the order of kilometres to force thresholds to very low levels, and
finally that an increase in purity by whatever factor increases the costs
of testing by about the same factor. DR
MIKE WILKINSON What
I thought I would do is just very quickly go through the ethos of the kind
of work that we are doing and to put it into some sort of context and then
I will describe some of the work which follows on from the paper that you
have been given. OK,
these are just some of the major issues which have been raised concerned
with geneflow. We are
actually only interested in most of our activity in movement of geneflow
from the GM crop into wild relatives.
We are particularly interested in relatives in natural habitats and
semi-natural habitats, as opposed to weeds, although we are looking at
weeds as well. The reason for
that in the long term is that natural habitats are generally more species
rich, than are (those) in farm agricultural environment and therefore
there is greater sensitivity to environmental change;
and secondly, natural habitats are not normally subjected to the
same sort of rigorous management that occurs actually on farm, so again
the plants are more important. So
the simple question, how can we look at geneflows of course for unwanted
ecological change in a natural environment, how do we go about doing that?
Well rather than give you very complicated models which we are
making, I thought I would just give a hypothetical very simple black box
example. So here we have a GM
crop and all this with big long arrows, somehow by geneflow it causes the
herbivore of a natural wild relative to become depressed, the numbers
decline to such an extent that a parasitoid (phon) which only feeds on it
becomes extinct. OK, that is
a fairly big hazard and very much a hypothetical one.
The way that you go about doing the quantification of this kind of
risk is that you break the process into a series of intervening steps, so
you build a pathway of risk. And
in this particular case, as in most cases, in fact all cases involving
what I call vertical geneflow, geneflow mediated by pollen movement or by
seed, the first step is the formation of F1 hybrid in the region, then
usually the transgene has to stabilise within the genetic background of
the wild relative and then it spreads, and so on. In
the ideal world it would be nice if we could assign probabilities over a
particular time frame for each of these steps.
Now I have just given hypothetical numbers there, but the thing to
really note is that these probabilities are accumulative, so at this
start, which is where we are at, the probability of getting all the way
down to depressed herbivore numbers requires each of those steps to be
fulfilled, therefore the chances of me doing that, that, and that, relies
on the first step being completed, then the second step, then the third
step and fourth step and so on, and each step has a certain probability so
these probabilities multiply as you go along.
The basic probability theory, meaning that as you go down a
pathway, as long as you can put numbers to each of these steps, the
probability diminishes as you go through from the start point.
However,
the key thing to note is that as you go at the start your error functions
also accumulate as you go down the pathway, so every time you make an
estimate you have an error function, so your estimate for F1 hybrid
geneflow might be, I don’t know, 80% plus or minus 5%, your next gene
stabilisation might be whatever plus or minus such and such an error
function, and these error functions on the first stage passes through to
the second phase, so your level of uncertainty accumulates as you go
through the pathway. Therefore
the advantage of this type of approach is that if at any stage you get to
a point where effectively you have got negligible or zero, effectively
zero probability over a time frame which you predefined, then later steps
in the pathway effectively become not high priority to examine. The disadvantage is that we have got to watch very carefully
on these error terms. It
follows that everything relies on the error terms accumulating, it means
that the first step, our very first, has got to be as accurately measured
as possible and for all geneflow to wild relatives pathways, and there are
many that you can imagine, and many that you can’t imagine, that they
all have the same first step; the formation of F1 hybrids.
We have got to get that as accurate as possible, and that is what
we are trying to do. We are
trying to estimate F1 hybrid formation between a model obviously, we are
using brassica napus or oilseed rape as the model, and its closest wild
relative in the UK, bargeman’s cabbage, turnip rape, call it what you
will. Now this species grows
naturally only on the riverbanks of rivers, mostly in England.
It also grows very occasionally as a weed but the two are actually
genetically distinct. I am
just focusing now on gene movement to this wild relative.
Now
as Joe has just mentioned, pollen movement by wind is actually quite well
understood over 200 metres or so, which is about a field.
The key difference between oilseed rape to oilseed rape gene
movement, and oilseed rape to a wild relative movement, is that most of
the pollen does naff all, it doesn’t do anything, it is only a small
proportion of pollen that actually succeeds in causing seed set, so there
is a dampening of the effect of the airborne pollen, whether it is by wind
pollinated or by insect, it doesn’t matter.
So much so that we actually know, even when you have a commercial
field of oilseed rape right next door to, literally one metre apart from a
natural population of this bargeman’s cabbage, we have found that their
hybridisation rates are of the order of 0.5%.
Given that these populations are typically 75 individuals to 100
individuals, 0.5% is not many. So for the vast majority of hybrids then
are going to occur when an oilseed rape field comes next to one of these
wild plant populations. Right,
well we know that the wild populations naturally grow by rivers, so
essentially we have got three elements:
oilseed rape, river, wild population.
When do the three come together?
Essentially we have to do it as several different points.
First of all we have to find how often the oilseed rape is next to
the river, then we have to find out when the wild relatives are next to
the river, and then when they all come together.
It is not as easy as it sounds. This
is oilseed rape, this is where it is across the UK. These are DEFRA figures, but we have been using satellite
imagery. This, surprisingly
enough, is not a map of the UK, this is a map of the rivers in the UK.
There are a lot of them, surprisingly enough.
If you overlaid the two, that gives you the frequency and
proportion of oilseed rape fields next to rivers.
That is not of very much use.
Until we started, that was the best data that was available in the
UK for where the wild relative was. The
big problem with that is that these spots represent 10 kilometre squares
and it could be one plant represented there, or it could be a million
plants, we have no idea at all. So
the way round that is we have done extensive river surveys across the
whole of the UK and we have made use of a massive survey of all the major
herb area across the UK, we have requested all their brassica wild
specimens and we have also been in contact with local botanists.
So we now know, and believe me, my students hate me, for 200
kilometres I have sent them walking along rivers, and they not only walk
along them, they measure the number of plants and exactly where they are,
we are using GPS positioning, so we now know not only which rivers carried
that crop, but the density of them and how big these populations are. QUESTION: You
should say … DR
WILKINSON Oh
yes, well one of the perks of the job is that somebody has got to go in
the boat and it is a position of responsibility, you might crash it or
anything, so I took that responsibility, understandably. Just
to give you an example, this is a series of dots, but in real terms this
end is for the posh people, that is Eton, and also for the equally posh
people, that is Oxford, so we started from one posh place to another posh
place, and inbetween we found 1,118 populations of Brassica rapa (phon).
So we know a lot about the density, and that is just on one river
system, we have looked at lots, in fact 17, that is just the population
sizes. Now
we are in a position to say, ‘right which river catchments contain
brassica rapa?’. These are them, you will notice there is a big red bit in the
middle, that big red bit in the middle is actually not just one river
catchment, but because England has all these canals, these river
catchments are all connected to one another, so it is a continuum, so we
treat that as a oner. Right,
we have to add in other factors which limit the distribution of brassica
rapa, altitude, absence of flooding which I can’t go into because it
could take an hour on that, and the absence of brackish water. If we do that we can predict the distribution of riverbank
rapa across the UK, or across mainland Britain.
Apologies to our friends from Northern Ireland, I am afraid I
haven’t got enough data from there yet, I couldn’t get the boat hire.
So here we are, this is where the brassica rapa is in density, and
when we overlap, as I said this is the river bank and when you pop the two
together that is where the three elements come together. We
are now, I have to say we are in the process of converting these into
actual numbers per annum, these are very preliminary so they are not
published data and they are not the final data, but just so that you can
have some sort of idea how this translates, this big blob up here, which
represents a reasonable amount of it, before we put in a calculation to
adjust for long range geneflow, that I think it was about 1500 hybrids per
annum in that area there, so what we are trying to do now is compile an
estimate of the total number of hybrids that we are expecting between
oilseed rape and this wild relative per annum across the whole of the UK.
The importance of that is that will allow us then to as a start
point for the estimates as we go down this pathway, but equally we can put
in correcting factors for if we want to put in approaches which will limit
the number of hybrid formations, for instance if we were to put the
transgene on the chloroplast , or if we were going to put the transgene on
to a chromosome which does not normally transmit into the wild relative. CHAIRMAN Thank
you very much … DR
ROGER TURNER Again,
repeating thanks, that was jolly good, I am not sure that even I
understood all the fancy bits about sympatrics and things like that,
perhaps we can draw those out. The
general thing I think is that people are focusing in on separation
distance as the panacea, but there are lots of other ways you can prevent
pollen flow. So I don’t think we should get too concerned about whether
it is x, or 2x or 10x. I
think the other point, I am always fascinated by in these sorts of studies
is if you look and think that pollen flow and geneflow is as we see it in
some of these models, plant breeders say to me:
“you know it ain’t that easy”. If it was that easy we could
breed millions of new varieties every week, so it is not that easy to get
the genes to move to where you want them.
I think that is another important point that we need to think
about, and then of course you get into all the bits that both Joe and Mike
pointed out. As Mike was saying at the end, it depends whereabouts in the
genome the particular gene is and whether it is “mobile” or not.
So there are all those sort of background factors that come into
play in this sort of exercise. And
the other thing I find quite fascinating about it all is what I call the
“so-what factor”. It does all of this, so what? If you look at oilseed rape, it is actually herbicide
tolerant, even if it is a wild population it will tolerate certain
herbicides, the mechanisms whereby that tolerance mechanism operates
varies depending on the herbicide. And
was it my friend Darwin who said “Nature red in tooth and claw…” and
you know natural populations are pretty robust, it is the crop that
deserves our sympathy, and our care, and our love. The wild species can
probably look after themselves? Anyway,
I have rambled on enough. I
throw it open to general questions. QUESTION: I
have two questions to Joe. The
first one was that in your combined graph that tailed off at about 0.1%,
in the picture you showed us of it, it really went almost totally
horizontal at 0.1%, and in order to go below 0.1% on the scale you were
looking at one would have gone for tens of kilometres or hundreds of
kilometres. I just wondered
does your model allow you to look further out and see what it looks like
further along? PROFESSOR
JOE PERRY It
appears to go totally horizontal, but it doesn’t go completely
horizontal. To go down to
0.01% you would then go down to about I think just over 3 kilometres.
But I don’t think that the actual values are at all important
here, what is important is the principle.
I think in neither of the models that I have showed would I think
that they were the sort of models that you would use for prediction, or
for accurate statements, they are more models to demonstrate a point using
fairly realistic parameters that we know are close to the mark but I
wouldn’t want to give you the idea that those values that I put up
there, in either of the graphs that I showed, were likely to be precise in
practice. QUESTION: My
second question was based around a visit that some of us did to a seed
production farm a month or so ago where they have to have fairly high
levels of purity in order to be certified for seed production, and yet
they seem to be using separation distances of a couple of metres between
one crop of weeds for one strain and another strain.
So that seems to be where at the moment when one has to produce
reasonably genetically pure populations for seed production, they seem to
be living with separation distances of literally as far as I can step.
I was just wondering how that ties in with both the sort of
empirical data in your modelling that we are talking about 200 metres to
3,000 metres. PROFESSOR
PERRY I
think the answer there is that each crop, and sometimes different
varieties of different crops, will require different separation distances.
So, for example, for oilseed rape, the separation distance required
for seed purity varies according to what type of crop it is, in the sense
of is it completely male fertile, is it partially male sterile, is it a
varietal association? And the
distances can indeed vary quite greatly according to the crop and the
species. The model parameters
weretaken from the separation distances that are used in the Farm-Scale
Evaluations where they are designed to limit the degree of
cross-pollination down to 1%. But
I think it is a good point, and there are other points here, so I think
the answer is one needs to, as is often the case with GM, one needs to
take everything on a case by case basis.
But it also raises the question, if you want to set a threshold,
should you set one which is an average for several batches of seed,
several seed lots, or do you want to set a threshold which would be more
stringent so that none of those
seed batches attains the threshold, so are you looking to set a threshold
for average or for complete safety? And
that kind of question, and there is a related question of, well OK over an
entire field the average may be 1%, but we certainly know that averages
are misleading here and that spatial clustering, patchiness, of
cross-pollination will almost certainly occur, and so there will be parts
of a recipient field, if it is cross-pollinated, that are much greater in
percentage terms than other parts. When
that crop is harvested, that might lead to produce being differently
affected, so although the average may be 1% it may be that certain parts
of the field may be as high as 5, or 10, or greater.
And then of course there is the related question of are you talking
about seed purity or food purity, and I think it is important to at least
to raise the issue that there is a difference between these, that if I
take GM wheat and mill it and produce flour, that just because 1% of the
ears are GM doesn’t mean that 1% of the produce has GM content. QUESTION: A
question for Joe and a question for Mike, if that is all right.
The first one is one of these issues around principles, what are
the issues that we are going to have to take into account.
One of the things that you didn’t talk about was the size of the
pollen sought, which seems to have been important elsewhere in Canada, and
I wonder if you could shed any light on that and how you took that into
account and think we could do. Because
it was one of the things I asked Mary Rieger about was her work in
Australia and in fact they didn’t know, because they didn’t know how
much of that commercial variety had been sold, but she thought it was very
small, so in fact although the levels were quite low that they found, is
there an issue if you have got an increasing pollen source, that you are
going to have a greater load of pollen. So I would just be interested in your comments on that one. PROFESSOR
PERRY Could
you just explain what she said was low? QUESTION: The
proportion of oilseed rape which was grown that was the herbicide tolerant
variety, she thought was low but she didn’t know because the data
wasn’t available? PROFESSOR
PERRY I
completely agree that the degree of cross-pollination is going to depend
on the size of the donor fields and of course, as you imply, the greater
the proportion of, if it were GM, which it wasn’t in her case, but if it
were in this case, the greater the proportion of GM that was grown, the
total aggregate amount of pollen there would be and therefore, yes, the
greater the cross-pollination. And
I think it is actually often a complex relationship, the relationship
between, spatially, a block of donor plants here giving pollen and the
degree of cross-pollination by a recipient can depend upon the distance
apart, as we heard, but also in what is in between, what barriers there
are. And in particular there
are ways, I think Roger mentioned, there are ways of mitigating this by
growing barrier crops around either the donor or the recipient.
And you are absolutely right, it is a very complex situation.
There are models developed I think by the French, the so-called
MAPOD model, but I searched for it in the literature and could find no
good description of it, it is based on an earlier model called Genesys.
And I think the problem with very large simulation models of that
sort, which are multi-parameter, and which try and take many, many,
factors into account, is that they are often not very accurate for
prediction purposes. QUESTION: I
was just wondering if Mike might be able to help us, just focusing on
recent information I think we will have seen reported but won’t know a
lot about, we talked about oilseed rape almost exclusively this morning,
about the sugar beet studies which have been done around geneflow and
whether you know about those, because that is irrelevant in the short term
as well? DR
WILKINSON We
have been looking mostly at oilseed rape.
The reason we have looked at oilseed rape is because in the UK it
was, at the time we started, by far the most likely species to be
commercialised which had wild relatives available.
Now there are plenty of studies on lots of other crops, sugar beet
in particular, there have been some studies.
The key thing is their floral biology is completely different from
oilseed rape, oilseed rape is predominantly, while the brassicas are
predominantly insect pollinated, but in an agricultural sense it is insect
and wind, whereas sugar beet and wild sea beet is wind pollinated and
therefore it tends to have a greater dispersal distance.
Sea beet, as you can imagine, is limited to the sea coast and
therefore can equally be studied in this way, I know Rikke Jorgensen is
actually working at that now, isn’t she?
Yes. You have to look
at every crop separately. There is no way round it.
One of the things that I didn’t mention though is that when I
went through that pathway is I sort of didn’t make the distinction
between information which is generic for the crop, and information which
is specific to the transgene crop combination, and as you go through the
pathway you increasingly become transgene specific.
So where the effort is at the moment, and where the effort should
be at the moment, is on the top generic level so that we can actually
quantify geneflow rates and recruitment and stabilisations, which are by
and large transgene independent, certainly the F1 hybrid formation for
most transgenes, but as we go down the pathway then increasingly as
selection comes in then it is the transgene that becomes increasingly
important. So I wasn’t
really sure what he was trying to get at as far as the Sea Beet examples.
The key thing to remember with beet though is beets are both
diploid, triploid and tetraploid, whereas the wild relatives are all
diploid. QUESTION: Well
I suppose I was interested in getting some insight to what the situation
would be for where we have a crop and a wild related species in this
country. Do you see what I
mean? Because like oilseed
rape has been less discussed than sugar beet, which what I understand from
the literature is also, and increasingly at that first F1 hybrid stage,
that people are increasingly looking at that and making estimations of
that probability, which seem to be going up rather than down as people
look more carefully, as I understood it, and I haven’t read the most
recent stuff, and that was why I wondered whether you could enlighten us
on it really. DR
WILKINSON I
think what you will find, as you go through the literature, and this
applies to all crops, you have to separate the sorts of studies that we
are talking about. Hybridisation
rates in field trials are hazard identification more than anything, they
are no more than estimates of hybridisation rates as a weed.
Hybridisation rates in the field, there are very few of those,
there are an increasing number of them but they are more relevant in this
particular context. And you
are right, some of them are going up and some of them are going down, but
context is everything in this and it is not just it can be high under
these circumstances, you have to say well how often do these circumstances
arise over the region that you are interested in. QUESTION: So
are there any data? East
Anglia for us, are there any data you can think of which is relevant for
us? DR
WILKINSON No.
There are, yes, but in the case of Sea Beets, of course you have to
remember that sugar beet doesn’t normally flower, only a small
proportion of the bolters actually flower, (so) it is not the whole crop,
unlike oilseed rape. So a priori we would guess that the rates should be a
lot lower, but having said that the species barrier is weaker.
To my knowledge nobody has actually got those data yet, but it
needs to be done. QUESTION: I
want to ask a very practical question really, picking up on something that
Joe said earlier about variability. I
am just trying to talk in a little bit more detail, that if we decide to
go, say, for a 1% threshold in cross-pollination, do we actually have
enough information about the variants in the tail really to be able to set
separation distances so that people who wanted to grow GM-free crops would
be 99% certain that they would not exceed that threshold?
And if we haven’t got enough data, what exactly do you think
would be required to get that data? And
finally, how much extra might be required if we moved from 1% to 0.5%,
which currently seems to be what the debate is in the EU at the moment? PROFESSOR
PERRY I
think that is a very difficult question.
I think that I would say that if we reflect commercial practice,
which I believe that 1% does broadly, I think we have enough information,
but only because commercial practice has maintained that degree of purity
over the years, rather than the fact that there have been detailed data
published in the literature. I
am not aware of much published data that relates to the tail of the
distribution, and I would like to see much more data from NIAB – the
National Institute of Agricultural Botany – and I would think that maybe
your group could well benefit from asking them about what other procedures
for maintaining purity of conventional seed lots.
I have asked them, and I haven’t actually had a very clear reply.
So this is probably my lack of knowledge, but I am not aware of what just
the straightforward basic run of the mill tests that are done, I would
like to know more about those, I would particularly like to know about
what sample sizes are used before we could even answer the question of
whether commercial seed purity was being maintained to the degree that we
I think all imagine it is. I
don’t think there is enough data in the literature if we were to go down
to 0.5% to give a very good estimate of what separation distances would
have to be. And in framing
that question, you would need to tell me in what percent of cases you
wanted that to be so, i.e., you wanted to achieve 99.5% purity in what
percent of cases? because quite clearly by the nature of biological
variation that we are talking about, you won’t achieve them in all
cases. But we could put
confidence limits on those. So
I don’t believe there is much good evidence in the literature to enable
us to make a very decisive stab at what happens to maintain 0.5%. DR
WILKINSON Can
I just add something to that? When
you are talking about detecting 0.5% contamination, you have got to
consider the technique that you are actually using to do that. Now as I understand it, we are talking about quantitative PCR
methods. Quantitative PCR
can get down to 0.5%, but you have got to be very careful who is
doing it and whether the system is up and running to be able to do that
reliably, so there is an extra added variant caused in by the technique
itself for detection, quite apart from the sampling considerations. QUESTION: I
would interject, Joe was making the comment about purity of seed.
It is relatively easy to do at the moment because you are using
morphological characters in your measurement, so you can pick out within a
variety what the impurities are coming from other varieties in the field.
But as I read it, the problem with the GM is that you can’t
easily measure that, so you have to then come down, as Mike was saying, on
some form of PCR. It is not easy to distinguish by the naked eye. PROFESSOR
PERRY If
you wanted to detect contamination in oilseed rape, presumably you could
take (say) 10,000 seeds, sow them with applied glyphosate and see then how
many survived, and that would give you a quantitative – a pretty good
quantitative – estimate of the percentage. PANEL
MEMBER The
problem there is you have got your crop, you have harvested it, you have
to do it and then make a judgment, rather than actually do it there and
then in the field, and if the crop is impure in the field you can actually
rogue out the ones you don’t want. QUESTION: Hi.
Dozens of questions really. Trying
to put some of what we have heard this morning in the context of the work
we have actually got to try and do, and some of these may be issues I may
have misheard or not understood. But
Joe, in your first paper, the model that you have created there, I presume
that is a crop specific model and what was the crop in this instance?
Was it oilseed rape? PROFESSOR
PERRY It
was a sort of generic model, it wasn’t crop specific. It certainly could apply to oilseed rape, it could also apply
to maize, although maize isn’t insect pollinated, the pollen would still
be affected by things like turbulent conditions and weather fronts. QUESTION: But
you end up with absolute distances, which surely will vary from crop to
crop. PROFESSOR
PERRY Yes,
they do. QUESTION: And
so that is a model which is about units rather than metres as it were in a
real situation, one would have to do that model, I assume, for every crop
in a more specific way. PROFESSOR
PERRY Absolutely
right. QUESTION: You
can’t just roll that out, so we shouldn’t be taking from that actual
distances because the crop is not specific in that one, you are
demonstrating how these things work. PROFESSOR
PERRY That
is exactly what I tried to answer in response to Mathew’s question. QUESTION: Sorry,
… The second one, because
obviously there is a huge amount of work because you want every crop, and
probably every variety of crop that you are looking at listing, which is I
think one of the concerns we have, how will we able to maintain this
momentum of energy going into this kind of thing all the time.
The second issue, just to make sure that I am clear about this, you
were saying that your assumption was that perhaps only 1% of
cross-pollination might be mediated by insect or unusual weather
conditions, but surely that is going to end up with, when you get an
unusual weather pattern in an area, a very lumpy problem, you might
suddenly have quite a massive issue in one region at one time because of
unusual weather conditions. Is that correct? PROFESSOR
PERRY Yes. QUESTION: So
it is a low likelihood but then you might suddenly get a spike. PROFESSOR
PERRY That
is a very good point. It is a
bit like when they used to measure radiation, the government said that
there were radiation emissions of below such and such a percent, and that
may have been the case, but there were hotspots and spikes, as you call
them. So yes, a lumpy
distribution. QUESTION: A
third point. DR
WILKINSON Can
I just interject with that? Yes,
that is true, but you have got to temper that with the behaviour of the
insects, because although you do get long range dispersal of insects, not
all insects will then carry on happily pollinating where they happen to
find themselves. QUESTION: I
was thinking more of the weather pattern issue than insects, I would think
it was more likely to happen with weather than insects perhaps. DR
WILKINSON Right,
OK. QUESTION: The
third point, which is a more general one but relates to your first paper,
one of the things that concerns me when we are looking at seed purity in
particular, is that a lot of farmers – organic and conventional – are
saving their own seed for cost reasons and what happens when you have,
have we modelled when you have your 0.1% contamination, for want of a
better word, how does that multiply up in the field over, say, 5 or 6
generations of saving your home grown seed, because that is what is
happening in reality on a lot of farms.
Have we got any data on that? PROFESSOR
PERRY I
wasn’t aware that farmers were saving their seed, because seed like
an F1 hybrid tends to lose vigour. QUESTION: But
farmers save all their own seed, we pay royalties on the seed that we
save, so we have to pay a royalty to the seed house, but some varieties we
can’t buy commercially now, and actually it is much cheaper to clean
your own seed on the farm and to replant it.
So this is something which is happening everywhere now, especially
as margins get tighter, they are buying in new varieties, they are just
paying the royalty on the seed, and quite often saving their own. ROGER
TURNER Yes,
Helen is right, there is a very high level of farm saved seeding used by
farmers in the UK. I won’t
get into whether it is economically better or not, we can talk about that
later…, but the law says on F1 hybrids that farmers can only save those
with the permission of the breeder. QUESTION: Inaudible. DR
WILKINSON In
this situation, if GM crops went ahead, I am sure one of the conditions
would be no farm saving. QUESTION: …
this is not about those who are growing the GM seed, but those who
aren’t, and who have a low level of contamination under the threshold
which would be acceptable for food use, but want to save that seed on the
farm and might do that – as we do – for 5, or 6, or 7 generations, or
more. And what happens in the
field, when you have that low level of contamination, how does that
multiply up over time, is that something that is going to be very
relevant? PROFESSOR
PERRY I
think the answer to that is that if you have, say, 1% contamination per
year, the next year it might go up to 1.1%, and the year after 1.11%. QUESTION: But
this is when you have actually … with no separation distances, because
you will then have your 1% within the field, so it is next to the plant
rather than with your 200 metres separation distance, or whatever it might
be. PROFESSOR
PERRYI am trying to give you an answer in general terms, and I think the
answer remains that if in one year your seeds were 1% contaminated, and
you saved them all, then the next year you would sow 1% contamination and
receive – what would you receive – that is right, you would receive an
extra 1%. So yes you are
probably right, yes, you would go up to say about 2% in year 2 and so on,
yes, OK, good. DR
WILKINSON Can
I point out that with the F1 hybrids, the reason that companies make F1
hybrids is because F1s are by definition uniform, and as soon as you self
(phon) them that uniformity goes, it is genetic, the trait which from the
two parents segregate out at the next generation, so generally it is not
in the farmers interest at all to self an F1.
That is why they are so popular for breeders. ROGER
TURNER There is the Hardy-Weinberg Law about genetics.
It varies from crop to crop etc, etc, and all the general caveats,
but in fact a gene will decrease over time, it doesn’t increase, and it
goes down very rapidly and the numbers are very, very, very small. QUESTION: Inaudible. PANEL
MEMBER We
did on a European Union Working Group look at this question, I think
Helen’s is a really important question.
The first thing is that there are actually very few hybrid
varieties in UK agriculture, so you can almost put that to one side as
essentially seeds where it is practical genetically to save seed.
And the issue is very much about, certainly from the different
scenarios we have looked at, whether there are volunteers in the field,
whether you grow it on the same field and if it is oilseed rape, whether
you come back to that field the next year, or the following year, and so
on. The contribution of volunteers from a previous GM crop.
And so it is quite a difficult calculation to do.
And we looked at, well I can say more about that to Helen privately
I think. QUESTION: OK.
A final one. I just wanted to make the point that in a separation distance
model that obviously we are not talking just about organic farmers here, I
think it is very important we are starting to look at co-existence, but at
the moment the market is saying we are not very sure we want to buy this
stuff, and I think a lot of farmers will say we want to keep our crops
GM-free, conventional farmers will want to say that, and I think we all
accept that as a Commission. So
how does that model change when you start to put in the obviously lower
separation differences, because one assumes there will be a higher
threshold to discuss, when you are looking at all conventional farmers
growing these crops as well, how much room does that leave us? PROFESSOR
PERRY Could
you explain why the threshold should be different for conventional
farmers? QUESTION: Well
I was just assuming that it might well be.
If it is not, that is fine, let’s just work on the same
thresholds then for both. So
if we are saying that it is not just an issue of organic farmers saying we
don’t want to have contamination, but obviously all farmers who don’t
want to grow GM crops want to be able to maintain the purity for a non-GM
market, which is the likely market scenario we are going to be facing over
the next, say, 10 years. How
does that model then look? PROFESSOR
PERRY I
think in general if, as you imply, the black squares, if they were
representing not just organic but organic and
conventional, the model would say well there is going to be obviously more
of them, and therefore it would move us into the situation that there will
be much less land even than the graph that I showed available to grow GM,
if thresholds were the same and if the separation distances were the same.
But of course they are not -within the Farm Scale Evaluations, the
separation distances are not the same; there is a difference between
organic and conventional. QUESTION: I
have a question, but if I could just make a couple of comments about some
of the other things that have been talked about.
I think the point about learning from the NIAB experience is an
important one. I understand
John McLeod is attending the meeting at CSL in York.
He is a former director of NIAB and I think it would be important
to pin him down on that kind of question about the practicalities and the
experience we have so far on achieving particular levels of seed purity.
On the question that Matt raised about crops, different distances
and so on, the biological reasons why cereals only need a few metres is
that pollination happens without the flower opening, and so all the
pollination happens and you get a bit of pollen shed but it is very, very
little, and usually pollination has already happened then. Whereas with
the oilseed rape, that is insect pollinated with obvious consequences, and
wind pollinated like grasses, they sort of push their anther out and
expose them into the wind and there are some very elegant practical
experiments where they looked at the plume of pollen, the plume of
pollination downwind and so on. That
is quite old work but it is quite relevant.
My
question is really just to get a reaction.
If one says, or argues, that this pollination is OK in the risk
assessment, ACRE and everything, one comes to the view, or ACRE comes to
the view that pollination is not really a hazard in terms of environment
and so on, and you could argue that if there is a perceived, or evidence
that a particular transgenic will cause environmental harm, then one
should question whether they should ever go out.
So if one concludes from the regulatory process that geneflow is
more about defining a product, it is the identity of a product, and that
is what we are trying to fathom and trying to work out in terms of
thresholds, it is about taking this view, it is about defining what an
organic crop, what standard it should be at in terms of genetic purity,
what a non-GM crop and so on. Now
the seed multiplication people have had enormous amount of experience,
decades of experience of achieving 99.7% purity. Now there are some questions that we need to ask about that,
you know to what extent can they be absolutely certain of achieving those
levels of purity, and with seed multiplication and certified seed
production there is the possibility of long distance flyers and there are
stories of bees loading up with honey, getting trapped in the train and
going 200 miles and so on and doing some pollination.
There is no way you can control that but it is likely, well it is
pretty well certain, it depends very much on the recipient population and
on the size of it, but it is pretty certain that that pollination event
will lead to an extremely low pollination frequency.
As I say, if it lands on three plants and does quite a lot of
pollination then the proportion of seeds in that three plant block could
be quite high, but if the recipient crop is a large crop, and that
generally will be what will happen in agriculture, then the frequency
would be very, very small. Now
we know all that happens, we know it happens with certified seed
production, I am just wondering to what extent we need to worry about the
long distance flying bee, or the long distance rare pollination, because
that is always going to be very, very small, and if we are to accept
thresholds and quality standards, defining products, then we are going to
have to tolerate a low level of genetic mixing, it is a fact of life.
I am just interested in both of your perspectives on that because I
think Mike also has views on that. PROFESSOR
PERRY I
think this is partly a scientific question, and it is quite a long
question as well, I think I have got it, it is partly a scientific
question and partly a non-scientific question.
For the non-scientific aspect I think my view would most closely
coincide with the Christian bio-ethicists, Bruce and Horrocks et al, in
their book Modifying Creation,
which was published last year, and there they question the rights of
different groups to impose their own agendas on society at large, and they
ask how far is society obliged to bow to one group who claims that the
growing of GM crops represents a commercial threat, primarily because of
the way they have chosen to define their own business;
and on the other hand, how far is another group justified in
claiming it as their commercial right to have the opportunity to grow GM
crops, and they conclude that neither can presume an exclusive claim, and
I think I would go along with that ethically.
Scientifically, I think you pointed out that a threshold that can
be realistically aimed at can’t ever be zero, if you have got both GM
and non-GM, or GM and organic. The
threshold should be considered on a case by case basis, depending on the
crop, depending on the use. And
you mentioned ACRE, I am prepared to trust ACRE and agree with them that
if there is an identified risk to human health or to the environment,
there is no suitable threshold and it would be better not to grant consent
at all. But if there is no
risk then the question is more a political one than a scientific one and
it should be agreed between growers and government, the food industry,
consumers and bio-ethicists. So
I would go along with the ACRE view that if it is safe then it is not up
to scientists to apply, to determine, a threshold, not up to scientists
alone to determine a threshold. DR
WILKINSON I
think it is horses for courses. I
think when we are talking about contamination, if you want to call it
that, from one field to the next, in an agricultural sense that is not, in
my view, so much a big environmental issue, I think it is an agricultural
issue and it is a political issue. Because
certainly for the crops that I work on, the genie can go back in the
bottle for the vast majority of cases.
With long range geneflow into the natural environment, into natural
populations, I think it is important that you know about long range
pollinations, but only to adjust figures so that you can actually have a
realistic feel for how large overall, because a lot of rare events can add
up to a reasonable number of events.
So therefore it is important from my point of view that we have an
estimate of long range geneflow so that we can have a more accurate
overall estimate of the number of hybrids that we get, but as far as field
to field contamination goes, frankly I think I personally would be more
worried about farm practice and seed contamination I think than pollinated
… QUESTION: The
group of us looking at co-existence I think see it as our job to try and
get a handle on the practical prospects for co-existence in the light of
some of the thresholds that are being proposed, so whether that is the
current zero aimed for by the organic movement, or the 0.5%, or the 1% EU
level. And I think what has
been very interesting this morning is learning that there are all sorts of
complexities to what a 1% may mean, it sounds straightforward but it can
be about what it is 1% of, how much it samples to average a 1%, the point
about a particular field averaging at 1% but there may be part of it that
ends up as more of that in foodstuff.
All of this I think is very much what we are trying to get a sense
of. And my question was
whether either of you in assessing what might happen had looked at, for
oilseed rape, the impact of volunteer population, the potential impact of
volunteer population. Because
one of the things that I think has been a continuing problem, even in ACRE
assessments, is things like simple spillage from trucks, so volunteer
populations of rape springing up in places you didn’t expect them, so
while your models may say well actually this organic field is separated by
a kilometre from the nearest commercial rape, there is a road down the
middle with a huge volunteer population because there has been shed seed
down it. And similarly, Mike, in your paper you say maybe the way of
limiting geneflow to wild relatives is to ban the growing of the GM crop
in a field near a river, again there may be opportunities for shed seed or
volunteer populations, or in a rotation, there was an example recently I
think of stubble regrowing and flowering in November to yield a pollen
source that was entirely unexpected.
So have either of you been able to factor in those things or could
you just comment on the way in which we should try and factor them in in
terms of assessing the realistic prospect? DR
WILKINSON The
paper that I gave you in actual fact represents the last bit of work which
preceded our current work, and I have to say if you looked at the area
that we looked at, and you looked at the distribution of brassica rapa,
there is hardly any there, but we weren’t to know that at the time. So I don’t think it would be a realistic option actually to
ban the growing of fields next to rivers now that we know exactly how many
there are. There are other
ways of limiting geneflow. As
far as volunteers and feral populations, I spent a lot of my early
research working on longevity, as has Rosie, of feral populations, the
turnover of feral populations, and also looking at volunteers, although
Pete Lutman’s work is more relevant to that.
Certainly seed spillage is massive, of oilseed rape, it is all over
the place and a lot of the work by Rosie and Mick Crawley has shown that
the turnover of these populations, and certainly our data backs that up,
the turnover of these populations is fairly great inasmuch as these
populations tend towards extinction, except you have to measure it in
terms of the seed in the soil rather than in terms of the plants above
ground. If you look at the
plants above ground it is a misleading figure, and one of the big problems
with it is you see what seems to be oilseed rape all the time by roads and
on construction sites, on construction sites it tends to be agricultural
soil being moved actually, on the side of roads it tends to be spillage
during transport, but it is a dynamic because these are just being
respilled all the time, and the same with volunteer populations, it is
being spilled all the time during their harvest process usually.
And this is one of the reasons I said seed is probably more
important than pollinated contamination, because quite apart from the fact
that genetically you have got two copies of the transgene, also it happens
on such a large scale that I think you do have to take that into account.
We have looked into wild populations, and with the wild brassica
rapa growing by the river, I think we have surveyed a few thousand
populations now and so far I think I have found two where there is a mix,
so that gives you an idea of the scale, it is not really a great problem
in the air with the truly wild, simply because where the farmers don’t
tend to take their seed there is a little bit of seed spillage
occasionally, but very rarely, into those populations they very rapidly
get competed out. PROFESSOR
PERRY Could
I just begin by saying that you said something that I didn’t say.
I did say that parts of the field may be more contaminated than
others, I certainly didn’t imply that the food content would be above
the contamination levels necessarily from that, because in general I was
trying to make the point earlier that food content is often much lower,
much, much, much lower than the degree of percentage contamination of
seed, so I didn’t mean to imply that and I don’t actually think it is
in general terrifically
likely that 1% contamination is going to lead to 1% food content, it would
be much, much, much less by orders of
magnitude. I have little to add to what Mike has said, except to
emphasise that the difference between, in all the discussions we have had
this morning there has occasionally been a confusion between geneflow into
wild relatives, and that is an environmental problem, and contamination of
crop produce which is kind of a food-safety, consumer type of problem, an
agronomical problem, and I think the two do need to be clearly separated. As far as the geneflow to wild relatives go, yes I think that
Mick Crawley’s work showed that with herbicide tolerant crops this is
not going to be a huge problem. And
Rosie’s work, again I think that has been very influential, particularly
her review in Trends in Ecology and
Evolution, in which she pointed out that the key thing, and ACRE have
pointed this out too, is whether the wild relative gets a selective
advantage and might thereby increase or whether there is no reason to
suppose a selective advantage in which case we expect local extinction of
the invader or weedy near-relative hybrid. QUESTION: I
too want to go back to the question the group is exploring on behalf of
the Commission and the whole issue about both consumer choice and
co-existence, because it seems to me that the logic of what you were
suggesting to us, Joe, there were some pictures you were showing us,
combined with what Helen was saying about the need not just for organic
but for non-GM farmers in this new world scenario to be able to show that
they have actually GM-free, whatever the tolerance level is, produce.
What that seems to suggest to me, this is kind of common sensical,
but you have illustrated it, that the pattern of land use for each of the
three farming techniques will have an impact on the availability of land
for each of the other farming approaches.
Whether or not you can achieve co-existence and guarantee consumer
choice, let’s assume we can sort the consumer choice issue out and that
all of the very many complex questions you have been talking about in
terms of separation distances in order to support consumer choice can be
resolved, so let’s take that for granted.
But whether you can then have co-existence, it seems to me,
doesn’t boil down, the key ethical question there is not the one that
you re-ran from that particular publication, it is not about whether
farmers have a right to choose their mode of farming, but rather where
they can choose to do it. And
the logic of the model is that there must be a pattern of land use which
will provide us with more opportunity to assure co-existence.
Now one extreme is if all the GM farms were in East Anglia there
would be a lot of land left in other parts of the country for organic
production. That tends to
suggest to me that it may be that we want to go down the path of
co-existence, or effectively licensing land for certain sorts of farming,
rather than allowing an individual farmer in place (a) to choose which
farming method they adopt, irrespective of the other types of farming
around them, so it is an optimum land use pattern.
Is that right, do you think, is that logic right that there is a
better pattern of land use which would allow more co-existence and that
that might need to be managed? PROFESSOR
PERRY I
don’t think I have got the experience in agricultural economics to be
able to answer that question. I
guess you are talking about ideas of zoning where you would grow a certain
crop in a particular part of the country and then the question would be
well how big would those areas be. I
think it is important to realise that the American model of having their
environment well separated from their intensive agriculture is not the
case in our country where we have got a heterogeneous spatially patchy
mosaic of both farming and semi-natural habitat and wilderness areas all
mixed together. Actually for
me a much more important question is if you did go to commercialisation,
under what circumstances you would allow it in terms of the indirect,
ecological effects, not the effects of co-existence.
I don’t think I can comment further. DR
WILKINSON Can
I just add one little thing. If
it is a major problem and it is unacceptable to have a mix in the way that
you suggested, or it has been suggested, not all GM varieties and
associations are actually equal in this regard, for instance you can have
a male sterile GM cultivar with a pollinator which is non-GM, in which
case the geneflow dynamics are zero effectively unless the transgene
breaks down, and there are other technical ways of limiting geneflow that
way. QUESTION: I
think what I am suggesting, one of the pictures that Joe shows us, I think
it was a scenario where organic land use had grown to 20% and we had
substantially bigger separation distances than we have now, so it is an
extreme case. But let’s
take Helen’s point and let’s assume that conventional farming is in
this picture too, so it is probably more realistic than the pattern you
showed. There are pockets,
because of the positioning of organic and conventional farms, the pockets
of land that remain in that model for GM production might simply not be
sufficient to allow for co-existence as GM produce.
Now in reverse, if you turn your model on its head and map GM farms
and conventional farms, they may not be sufficient for organic farms, so
it is the sort of rather random nature of the land use which we currently
allow because we allow people to elect individually what they will farm,
and where, which has a very real impact in this world of co-existence that
we might want to allow, on whether or not we can achieve our aim, in other
words it is having three types of farming because the pattern of organic
farming might rule it out, or the same for GM or conventional farming,
which suggests some kind of management of who can grow what where.
I am not suggesting that we have all our wheat in one area, it is
less about crops, it is more about organic, GM and conventional farming. PROFESSOR
PERRY I
think that the message from my second paper is very much that if
separation distances are increased greatly, I can’t see how it would be
possible to have any reasonable chance of co-existence. And given what we
have heard about the decision by ACRE, that of whether or not it is safe
for human health and the environment, I suppose one would ask the
question: Well, if it is safe at 1% it is not much safer at 0.5%, it is
either safe or it isn’t safe. And
therefore I suppose I am saying I would have thought that the government
would need to think very carefully before increasing separation distances
greatly. CHAIRMAN Thank
you. We are I think nearing
the end of the session, but Mathew has I think a further question. QUESTION: …
It is not only brief but it is way outside my own area of expertise, it is
a legal one, which is putting myself out on a limb slightly.
But this is from the Liability Group and I think it is relevant to
this co-existence question that this group is considering, I think it is
just worth mentioning something that wasn’t entirely obvious to me and I
learnt from one of the people that came to talk to us, which is that –
correct me, lawyers, if I am wrong – but I think there is a legal
principle that the first use of land is not considered to give you a prior
right, such that in the context of either increasing organic production or
the commercialisation of GM crops in either way, being there first and
doing your kind of farming first doesn’t currently, I understand, give
you any right to stop someone coming along and doing something that
clashes with that, and that seems to me to be a real confounding, or a
really important point to bear in mind when one is thinking about how to
set up co-existence regimes. CHAIRMAN Thank
you very much. I would, on
behalf of the Commission, like to thank very much Joe Perry and Mike
Wilkinson for talking to us this morning and to commend them for the
clarity of both their presentations and also the way in which they
responded to questions, which I think has been hugely helpful both for the
scientists on the Commission but also for those of us who aren’t
scientists and have assisted our understanding of the issues and of their
great complexity in a way which probably hadn’t been clear to us before.
So thank you very much for joining us this morning, we are greatly
appreciative of it. We
are now reaching the end of the (off-microphone …) (END OF TRANSCRIPT) |
 |
|||||||||||||||||||||||||||||||||||||||||
|
 |
||||
| Back to top |  |
|
 |
