IMPACT OF PEST CONTROL MEASURES ON THE ENVIRONMENT

Paul Jepson (Integrated Plant Protection Centre, Oregon State University, USA)

There is much pressure and controversy at present, particularly in the US, between groups that maintain intensive agriculture is not very damaging to the environment and environmentalists who maintain it is hugely and universally damaging. Neither view is correct. A much more educated approach is needed to the development and use of insecticides.

There are many ways in which pesticides can enter the environmental system. Pesticide contamination of freshwater and groundwater is quite common. Contamination patterns are complex- there can be noticeable pulses of arrival of pesticides into water through the season that correlate with spraying and climatic events, especially rainfall. In Oregon there are concerns about the effects of these pesticide pulses on behaviour of salmon, and normal development of frogs and amphibians.

Laboratory testing of pesticide pollutant effects tend to be carried out using Daphnia, a species that doesn't occur very widely in freshwater ecosystems like those in Oregon. This means there is a question about whether or not such tests using individual standardised organisms are ecologically relevant. Testing with multiple, locally collected species enables results to be ecologically relevant, and for the estimation of hazardous concentrations.

With terrestrial invertebrates, it is also possible to carry out field measurements to detect the direct exposure to sprays of beneficial insects. For example, this can be done by adding a fluorescent marker to the spray and collecting insects to determine if they have been dyed. In estimating the impacts of spray drift, for example on Lepidoptera in field margins, if the potential dose of exposure of foliage with a single spray is known, it is possible to use laboratory-collected dose-response data to estimate impacts. Thus, good predictions can be made into the toxicity on non-target organisms, but unfortunately these techniques are not often used.

Long-term effects in the field are not closely correlated with laboratory obtained measures of toxicological susceptibility. Carabid beetles can be badly affected by only a few sprays but this is not seen in the short-term, only in large scale, multifield experiments. However, even with a supertoxic material that kills 100% of beetles each time it was applied, simple models tell us that it would be necessary to spray 50% of a farm before exceeding a 50% likelihood of carabid beetle extinction, due to the dispersal rate and mobility of the organisms.

A great deal is known about the impacts of pesticides on non-target organisms but these techniques are not widely applied in simple tests of how toxic different practices are and more of these types of data are needed.

Large spatial and temporal scales

In the Netherlands, some carabid ground beetles have been studied for 120 years and there has not been any apparent effect of modern agriculture. However local extinctions can occur at the farm scale with over-reliance on pesticides.

It is not possible successfully to extrapolate from the small scale upwards or from large scale downwards. Equally, long-term data on ecological impacts will never be available for more than a few organisms in more than a few habitats, but there are large datasets available from regulatory packages that show the impacts of pesticides on non-target organisms. Some of these data sets can be used to predict effects on other similar organisms in similar systems.

Pesticide decay profiles

In ecotoxicological studies, the aim is to establish if the Predicted Environmental Concentration (PEC) of a pesticide is greater than the concentration predicted to have no effect. One factor to take into consideration is that pesticides are transient because they will decay over time.

The Ecotoxicological Recovery Time (ERT) for any particular pesticide is the time following application that the residual dose subsides to the "95% protection level" for the community (e.g. soil invertebrates). This is the point at which, theoretically, 95% of exposed insects will be unaffected by the pesticide dose, and it is argued that general population recovery can take place after this point. . It is possible to examine the decay profile for a number of pesticides and calculate the ERT . For Dimethoate the ERT is 0.15 years, for Benomyl it is 3.8 years i.e. general population recovery for soil invertebrates will only take place following the ERT . Pesticides can be ranked according to a joint assessment of their toxicity and their persistence, and this improves our ability to predict responses in the field. .

The focus of attention tends to be on the effects of acute toxicity, but there is much less understanding of the profile of decay of chemicals and long-term effects. The use of organophosphates was never optimised as there was no understanding of how to use them properly. This was very unfortunate because of the large scope of properties of these chemicals, such that one well-timed spray in a season could be extremely effective.

Modelling of stream insects and their response to pesticides

Aquatic insects are a ubiquitous part of the aquatic environment and have an important role in nutrient cycling. Species have been selected for research on the basis of their abundance and life history. Stoneflies (predators), mayflies (herbivores) and caddis flies (detritivore species) have been tested. Profiles of toxicological impacts have been generated using bioassays to determine hazardous concentrations to the community.

Most bioassays are based on super-exposure and effects are therefore much higher than would occur at the actual rate of exposure that would normally be encountered. There is a need to vary the exposure time and consider pulsed exposure. There is a clear trade-off between concentration, exposure time of the organism and toxicity. There is also a vital need to understand the impact of direct and indirect effects at the community level.

Paul Jepson has worked with Zeneca Agrochemicals, Pond Action at Oxford Brookes University and other University researchers, to develop a web-based approach to link pesticide fate in ponds with toxicology and invertebrate life history. It is possible to generate a recovery model for sample populations and vary factors such as population size, growth rate and immigration rate of new organisms.

Donald Baird at the Institute of Aquaculture in Stirling, and others, have recently examined data concerning the effects of chlorpyriphos contamination of drainage ditches on invertebrate communities in the Netherlands. It was a community-based approach examining functional feeding groups: primary producers, herbivores, predators and detritivores. Levels of nutrients and pesticides were manipulated in the system. Nutrient manipulation boosted primary production and increased the numbers of herbivores and carnivores. Addition of pesticides killed certain herbivores, which led to higher standing crops of some of the primary producers and caused both direct and indirect effects on the predators. The data were examined in terms of impacts on the whole community. This is an exciting piece of work and shows that the future direction of research should now be to pursue a more community-based approach to understanding pesticide impacts in agroecosystems.

Transgenic agriculture

The exposure pathways of insects to plant incorporated protectants (PIPs) is very different to the exposure pathways to pesticides in conventional crops. Some taxa of beneficial organisms are clearly not exposed to PIP's and therefore will not be affected- this is a difficult message to convey to the sceptical public. In contrast, some beneficial taxa will be exposed to the PIP toxin for the whole season because a pathway of exposure exists. Use of PIP's is arguably a very inefficient way of administrating pesticides. Impacts of large scale adoption of GM crops may be extreme for exposed and affected taxa, and the challenge is to identify how numerous these organisms are, and to determine the significance of impacts relative to conventional pesticides and alternative methods of pest control.

Indirect effects may be more important than direct effects, for example specialist assemblages of parasitoids can be reduced by indirect effects following reduction of target pest populations. Selection of suitable taxa for monitoring is a far more complex problem than for conventional pesticides. In the US there has been a recent extension by the EPA of the registration of Bt corn and Bt cotton, and the need for more field-derived census data and toxicological testing has been recognised. There will now be long-term monarch butterfly and avian studies.

Unpublished data by Graham Head et al. showed that abundance of predatory bugs is higher in Bt cotton than in conventional cotton, because they escape pesticide sprays. Spiders also have higher survival rates for the same reason. However, more ladybirds are found in conventional sprayed cotton than in Bt cotton because aphid attacks are more common in the former.

In a study of Australian pest-tolerant cotton in 1998-1999, Bt cotton received 10 sprays per year, as opposed to 18 sprays on conventional cotton i.e. the Bt cotton was still receiving a considerable amount of spray treatment. There may be ways of growing cotton that are far superior in helping beneficial insects.

Additional modelling approaches

Chris Topping (formerly of The Glasshouse Crops Research Institute, Sussex, but now based in Denmark) uses a Geographical information System (GIS) to compare survival of non-target organisms in different agricultural landscapes, using a model called ALMaSS. He models distribution of nest sites for skylarks in imaginary and real agricultural landscapes with different amounts of insecticide use. This is a unique approach and is gaining significant interest from the pesticide industry and regulatory agencies.

In Oregon, researchers are examining future changes in the agro-environment using new ways of modelling agricultural landscapes and watersheds. One group at the University of Oregon is modelling land-use planning over the next 20 years. It is addressing questions such as "what will the future landscape look like if there is a great emphasis on conservation, or conversely, on urbanisation?" Predicted landscape changes if urbanisation is given priority would be a loss of old-growth timber in the area as well as a reduction in water quality.

Paul Jepson and John Bolte are examining what happens when the priorities of people who live in a watershed are also considered. Such stakeholder objectives may include water quality, water quantity, habitat quality and social and aesthetic issues. The approach in this study is to work with stakeholders to assess their objectives and to measure the impacts of different practices by investigating factors including hydrology, stream temperature, sediment transport, habitat quality, biodiversity, landscape costs, and social goals. These factors are then modelled in a spatially-explicit landscape generator. The stakeholder can vary the priority they assign to different factors and the model predicts how the landscape will change according to the different order of these priorities.

The two new concepts of producing models with a biological approach to predicting pesticide fate into ground and surface waters, as well as models that provide tools for stakeholders will be powerful future research tools.

Conclusions

Pesticides have impacts that need to be mitigated. Some compounds clearly have intrinsically higher impacts than others. However, the influence of other anthropogenic factors is also vitally important.

There is a need to move beyond merely describing effects (i.e. explaining how toxic pesticides are) and to move towards management and mitigation of these effects. More comprehensive and dynamic landscape models will be required to achieve this.

It is necessary to return to an era where farm biodiversity is monitored and experimental regimes developed that examine the functional role of agrobiodiversity.

Scientists are often appointed into biocide development groups and agroecosystem management without the benefit of ecological training- this should change.

Discussion

• In the UK in 2003 there may be risks increased impacts on some beneficial insects when certain types of pesticide become restricted and more toxic or less efficient alternatives take their place. The impact will depend very much on the pesticide group involved. If growers are given less choice, an unintended consequence is that they may use more toxic chemicals on a larger scale. Mildly toxic effects on a large scale can be just as harmful as acute toxicity at a small scale- this is clear to ecologists but has not been acknowledged by the pesticide regulatory sector.

• It may be possible to develop more benign pesticides, but there is concern that many of these new products (e.g. Bt sprays to control Lepidoptera) will be used more frequently because spray application and delivery techniques have not been well developed. Diversity in the market place is important as it allows people to make educated decisions about which spray to use most effectively.

• Regulatory decisions are based on short-term and small-scale studies. In the US in particular there has been a regression to single-species testing that only provides information about the intrinsic toxicity of a compound - this has meant a backward step of around 50 years in terms of the quality of the data considered in the regulatory decision. Much is known about the toxicity of a compound but very little about its effects in the field.

• It is important, however that regulators and growers should have separate domains. Regulators shouldn't interfere once the decision to pass pesticides has been made as this can cause many problems.

• In the UK and Europe a major pollution concern is not the spraying process itself but the cleaning of spray tanks and other sources of contamination. There are new regulations that direct the farmer to use clean water to flush out the spray tank during the final spraying. This avoids the tank being washed out in the farmyard and the chemical washing into drains. Research in Germany indicated that 80% of the stream contamination in some orchard areas was as a result of washing the tanks and not due to spray drift. The most seriously contaminated watersheds in America are in urban areas. This is a reminder that pesticide use is not just by growers; the worst examples of contamination are in areas they are not responsible for.

• Modelling at a landscape scale is an appealing idea, although there is a great danger of the compounding of errors due to the complexity of the parameters included. An additional difficulty is that most land is privately owned so private decision-making processes occur at microscales when the aim is to model at macroscales. This may be mitigated by programmes that encourage growers to collaborate such as the ESA (Environmentally Sensitive Area) schemes.

HUMAN HEALTH ISSUES AND REGULATORY IMPLICATIONS

Terry Tooby (JSC International Ltd, UK)

Human health issues are responsible for some of the pressures driving the decision-making process about pesticide use. In trying to establish risks of pesticides to human health it is necessary to use animal testing and to extrapolate to humans. There are accepted practices but this whole system of testing is based on many assumptions. In order to determine the impact of pesticides on the nation's health, there is the need for a more holistic view. This will be achieved by bringing together people from different sectors and discussing what the primary objectives for the future ought to be.

The concept of "Human health" includes: mental and social well-being, lifestyle factors, environmental factors and the absence of disease or infirmity. The importance of agriculture in human health will depend on issues such as whether UK residents obtain most of their food from the UK or from elsewhere and also the relative proportion of fresh or processed food eaten. There is a considerable problem of communication, with many members of the public viewing the countryside as a "theme park" and supermarkets as their primary source of food. This is partly because consumers don't buy directly from farmers.

A white paper was recently published on food safety by the EU (final draft in 2000). This paper was prompted by the BSE and E. coli scares, plus the problems of food residues. As a result of this paper, the EU has established a new body, the European Food Standards Agency (EFSA), which will be responsible for pesticides and food residues.

It is clear that consumer issues are now driving the controls of pesticides. In the UK, the report of the policy commission on the future of farming and food says the public are becoming increasingly concerned about the food they eat. People are concerned about:

• Use of pesticides and fertilisers
• Animal welfare
• Feeding practices
• Use of antibiotics

This report says that England's food and farming industry today is unsustainable in every sense of the term. This thinking will dictate and overlie many of the future decisions on pesticide usage and decisions.

The report said about pesticides:

• The risk from pesticide use must be reduced.
• There was a need to disseminate information on integrated farm management and organic farming.
• There is concern about the damage that the past use of pesticides has had on the environment (especially organochlorines and organophosphates) and that members of the public are concerned about the harmful effect these may still have.

• The report conceded that the majority of our food will continue to be grown with the help of pesticides for the foreseeable future and that the regulatory system in the UK is good.
• It is important to try to minimise the negative consequences arising from the use of pesticides.

There were 823 active substances available on the market in 1993; now after a European review this has been reduced to around 240. Many of these chemicals were not supported by residue or environmental data. No reason is given for withdrawing an active substance- it may not be unsafe, just uneconomic. This can be misconstrued, adding to the concern the public already has.

Human health issues

There are potential risks to both the operator applying the pesticide to a crop and to the end consumer of the crop. For the operator, the majority of contamination occurs through handling the concentrated product, during mixing and loading procedures. There is also exposure during spraying. In the UK it is possible to ask the operator to wear protective clothing- this is not true for all areas of Europe.

Regulators use the most extreme case on which to base the rules e.g. they assume that an operator will be spraying without using a tractor cab or without protective clothing. This then becomes the model on which regulation is based. Workers may re-enter a treated crop and bystanders do not wear protective clothing. Exposure may be a critical problem, especially in enclosed areas like glasshouses.

For the consumer, it is important to remember the different categories of adults, children, toddlers and infants, all of who may potentially suffer different risks, either acute or long-term.

EU regulation, establishing the EFSA, includes more than just pesticides; it covers animal feed, novel food and GMOs as well as issues of contamination and residues. There are many added complications, such as the fact that many insecticides are used on stored products. There are also the impacts on water. Europe has determined that no pesticide should be found to exceed >0.1μgl^-1 in drinking water, regardless of toxicity of the particular pesticide. There has been a global review of the data supporting existing chemicals but it is a very large problem to do this. Certain countries started national review programmes and this prompted the EU to also begin to implement a review.

Global Crop Protection Market

It is important to remember that as well as food standard agencies and consumer issues, an additional factor is the decline in the number of major crop protection companies through amalgamation. Re-registration procedures are likely to severely limit the number of available pesticides. There is a concern that with the rush to review and withdraw, driven by numbers only there may be a loss of significant biological activity in the "armoury".

Of the 190 insecticides currently being reviewed in the EU:

93 are under evaluation (and this evaluation may take years)

94 have already been withdrawn (mostly due to economic, rather than safety problems)

2 have been listed in Annexe 1

These are worrying figures.

Assessment of risk

Potential human health risks from pesticides are established using rats and mice and the data are extrapolated to humans. Risk is calculated as the hazard multiplied by the exposure. This procedure needs refining and there is a need to characterise what "risk" really is. The same chemical can have very different risks depending on how and when it is applied. For example, modifying methods of application could mean that a chemical that would otherwise be withdrawn due to operator risk could be rendered safe to use.

Risk management uses a nationally-based system and worst-case scenarios. There is significant room to improve this. Schemes such as LERAPs have begun to address this type of problem but there is a need for substantial local assessment. All aspects of risk need to be communicated to consumers. A typical dilemma might be whether to ban an organochlorine pesticide or to tell consumers to change their food preparation method to make the crop safe to eat.

Risk characterisation

For the operator, the measurement of risk is the Acceptable Operator Exposure Level (AOEL). For the consumer, the measurement of risk is the Acceptable Daily Intake (ADI). It has been recognised that some insecticides including organophosphates and pyrethroids have the potential for acute effects, so another measure of risk is used - the Acute Reference Dose (ARD).

Information is currently obtained from acute toxicity data by an LD50 test on rats. There is hence a need for good quality sub-chronic studies to examine short-term impacts. Tests on potential developmental or teratology impacts on humans are also based on animal studies. If effects are found in more than one species, the safety margin for human exposure would be increased. It is also important that tests should also be carried out using several generations of laboratory animal. For tests on the risk of carcinogenicity, rodents are used as models but are known to suffer spontaneous tumours, which makes it difficult to assess the likelihood of effect on humans. If tumours were found to occur in many different species as a result of exposure to a pesticide it would indicate a much more likely problem for humans.

There is a need for a better understanding of metabolic processes and also to obtain data by studying humans who have suffered pesticide poisoning. There is also the need for better information on pesticide degradation routes and rates. It is vital to understand how the pesticide residues behave in the crop and whether its metabolites will be significant to health.

Routes of uptake

The operator risk (AOEL) is based on inhalation or dermal or oral exposure, although dermal exposure is the most likely to be a problem. In contrast, the consumer risk is oral, through food consumption. In order to assess the risk of pesticide exposure by consumption, the Theoretical Maximum Daily Intake (TMDI) is calculated. This figure will be an overestimate, assuming maximum rate of consumption of food with a maximum residue level and assuming no pesticide degradation. The TMDI needs to be estimated for the four key consumer groups: adults, children, toddlers and infants. Two components of diet are used- the national estimated daily intake and national estimates of short-term intake (to assess acute impacts).

In order to control the residues in food crops it is necessary to establish Maximum Residue Levels (MRLs). These are used as the legal requirement for food to be acceptable, and is therefore effectively the trading standard for both UK and imported food. However, most consumers assume these levels are toxicologically significant, when in fact they are not.

Another issue is the different ways in which food is eaten. Processed foods and stored products typically combine a large number of plant products from within one crop (or even from different crops from different countries). This gives a large safety margin over potential residues because of the likely mixture of the original sources of the food. A much greater problem is with single vegetables and fruits, where there is potentially huge variation between residue levels in individuals items and they are eaten individually. A single carrot or apple could hence exceed the Acute Reference Dose or Acceptable Daily Intake. For this reason, pressures from consumers also tend to concentrate here.

Hence there are issues of variability that must be taken into consideration. It is difficult to characterise the risk from such exposure. One of the ways forward is to examine risk from the point of view of probability. Monte Carlo assessment can be used to determine the chances of any individual taking two consecutive high residue components in their diet.

Isomers

Some pesticide compounds may include isomers and unless the metabolism and rates of degradation of each isomer is known, the regulators will only take the worst case scenario into consideration. If it is possible to separate out the biologically significant isomer it may reduce the rate of pesticide application.

Unfortunately, for some compounds isomers cannot be cost-effectively separated. Additionally, some isomers are differentially degraded between their target (insects) and mammals. Or they may be differentially degraded in both rodents and humans so a rodent model can't be used to assess likely impacts on humans. The regulatory implications mean that an ADI and ARD are based on the worst case scenario.

To conclude, there is a need to bring a much greater understanding of the exposure of both the operator and consumer to pesticides into risk assessment. Unless this is done, the default regulatory decision about any pesticide may be to withdraw it because there will be no data to support safe use.

Discussion

• Ordinary food contains many potentially dangerous chemicals. There are 10,000 carcinogenic natural pesticides in plants and each person consumes several grams worth every day, but residue from a synthetic pesticide provides only 1/20,000 of this total. Current concern about pesticide residues is about a relatively insignificant issue. Food is regarded as though it is pristine to begin with, then farmers add health harming chemicals to it. Food itself may have intrinsically unpleasant properties. It is possible that genetic modification can remove some of these unwanted characteristics.

• High doses of pesticides need to be given to experimental animals to demonstrate toxic effect in order to estimate a high margin of safety in humans. The population does not share this understanding- they see a dangerous problem, whatever the dose concerned. The communication of the concept of risk is an important issue. Consumers need the information to make their own choices and they need to know the levels of uncertainty involved when information is presented to them. At the level of the individual there is always the risk that a particular person will be prone to effects from a pesticide residue.

• Figures from the Food Standards Agency put the risks from pesticide residues into context. It lists the number of premature deaths in per year in the UK due to diet-related factors: 35, 000 cancers are due to poor diet, 80, 000 circulation deaths are due to poor diet, 50 deaths by food poisoning are caused by diet, there have been 15 cases of CJD but there are no known deaths from pesticide residues. Poor diet is the major problem. If there is a risk from pesticide residues, it is well hidden. The problem is that people perceive that they have no power over issues such as pesticide residues and CJD and this is why it concerns them.

• There is an issue about the potential problems of cocktails of chemicals on human health. A recent report from the Food Standards Agency implies that there is not a significant cocktail effect between different pesticides given the current levels of residues. However there is the potential for residues to react with the 10⁸ natural products in existence. This produces a problem of enormous magnitude. Regulatory systems in different countries seem to working effectively despite the limitations, because of the stringent built-in precautions. Thousands of compounds are rejected before they even reach the point of mammalian screening because they have already revealed toxic properties. The vast majority of screening therefore occurs within industry.

• Science and technology deliver superior analytical tools to which new foods are then subject. New products like GM foods go through more stringent testing than other foods were ever subjected to. There are problems with the regulatory process keeping up because of this. It is important to try to estimate if all the money spent on pesticide regulation is appropriate.

• Another issue is the impact of imported overseas food. Figures such as MRL levels are unobtainable in areas like Africa and DFID has to aid small producers in countries like Kenya because of EU imposed standards for food. It is recognised in Europe that as compounds are withdrawn from use, it effectively restricts their use worldwide because the EU country will not accept imported food containing that compound. An import tolerance level needs to be established for imported crops, but it is not certain who will set this.

• When the EU expands so that its 15 member states become 22, new members need to show that they are adopting the EU legal framework for pesticide regulation "in principle". Countries such as Bulgaria do not have the staff to take on such review programmes. Therefore such countries may question existing procedures and thus introduce more pragmatism into the regulation process. At present there is a considerable Northern European bias, and Southern European members are not as effective at putting their views across. It is desirable that this should change.

• Many of the discussions in this meeting have been based around the use of fast-acting chemical insecticides and their safe use. Apart from the environmental considerations, chemical approaches have a built in obsolescence and resistance will occur. It is important also to think about imaginative alternatives and a diversity of approaches.

• New approaches, such as the use of mycoinsecticides can be hampered by the regulatory system. However, to move any type of potential pesticide from the laboratory into practice is expensive. There needs to be equivalent investment in alternative methods of control to that invested in chemical control.

• It was suggested that it might be helpful to find a term to replace Integrated Pest Management (IPM). The new term should convey the sense of a better use of pesticides, with more specific agents that are better applied. It would communicate the idea of promoting the better use of pesticides under appropriate circumstances. This could be used to address the gap in funding and support for this type of research. Suggested terms included:

• All growers use multiple pest control practices- rotations, timing, cultivation, choice of variety, choice of fields to grow crop in, i.e. there are lots of pest control practices that do not involve pesticides. Some of these have been forgotten during the pesticide generation. The use of chemicals is not incompatible with biodiversity, but they need to be used correctly. It is not sufficient to replace the paradigm of pesticides with a surrogate- such as a viral insecticide or GM crop. The challenge now is to exploit functional biodiversity more effectively and make it economically viable.

FUTURE TRENDS

Tony Trewavas (Edinburgh University)

Overall summary

There needs to be a conceptual change in the way the problem of pest control is considered. A systems approach is needed to better understand both the structure and sustainability of farms and how to control pest populations. The farmer is the key control for constructing sustainable agriculture.

Some of the important drivers and future issues in agriculture include:

• The global price of wheat and rice is currently very poor.

• Global population is estimated to increase from 6 billion to 8.3 billion by 2025 and to 10 billion by 2050. This will have a considerable impact on UK agriculture exports and trade.

• Environmental issues must be taken into account and incorporated into future agricultural policy.

• The taxpayer supports farming and is becoming concerned and aware about food production methods, animal welfare, GMOs and the environment.

• There may be slow removal of tariffs on imported food. If world agricultural trade barriers were all removed, there would potentially be no UK agriculture because most UK crops can be grown more cheaply elsewhere.

Immediate Future Directions.

Farms that produce basic commodities such as wheat will become larger - this is already occurring as companies take over existing small farms. Small farms will also survive, but only if they provide complimentary niche products such as organic food or neutraceuticals. A theoretical justification for the continued increase in numbers of large farms is already available and rests on the demonstration that the distribution of wealth in population is stochastic and uneven. Simply put, in bad times, as at present for farmers, the wealthier, larger farm can afford to invest more money in new crops, new technologies and new machinery. Investment returns are often irregular, but are proportional to the amount invested. The rich simply make more profit because they have more money to invest. Smaller farms will be taken over unless they can find a niche market which is complimentary and not competitive with big farms.

Large farms do raise certain environmental concerns. The "Boarded Barns" study at Ongar, Essex compared conventional, integrated and organic farming over a period of 10 years. The study concluded that Integrated Farm Management was the best solution to the requirements of environmental and landscape demands, animal welfare, and the need to maintain farmer income and land use efficiency. This very detailed and crucial study reported that 85% of bio-diversity was found in hedgerows and field margins, with the type of crop grown in the field having a bigger impact than the mode of farming. Despite many measurements consistent differences in numbers of predatory arthropods on organic or conventional fields were not observed contradicting many assumptions about organic farming. Legislation may be needed to ensure field margin and hedgerow maintenance on large, efficient farms.

New methods in farming.

Newer technologies to provide both economic and environmental benefit are emerging.

• Precision agriculture uses metre-by-metre knowledge of previous yields, organic content, soil chemistry, structure and pest history to match sowing rate and eventual mineral application precisely with crop canopy expansion minimising waste. Continuous assessment of soil nutrients should soon be possible on the same basis using GM sentinels, plants that report water and nutrient status.

• Conservation agriculture commonly called Integrated No-Till or Minimal Till agriculture starts with the recognition that ploughing is the most damaging soil treatment. Instead crop residues are left on the soil surface substantially reducing erosion and direct drill is used to sow seed. Compared to ploughed organic or conventional fields nitrate run off is halved and integrated nutrient management reduces it further, improving water quality and reducing the expense of purification. The soil now becomes a carbon sink, beneficial to climate change policies. Earthworm and farmland bird numbers soar; bigger field margins boost bio-diversity enormously. Crucially, fossil fuel use is one third that of other farming methods on a per yield basis. New approaches to disease control are however necessary but the shared experience of individual farmers through the inter-net could create an enormous compendium of valuable knowledge and developing experience on disease control for No-till and stabilise its introduction. The use of herbicide resistant GM crops is the simplest way of introducing No-Till to the UK. But ill-informed knowledge about agriculture and lack of understanding amongst some members of the public leading to agitation against GM crops, may slow the introduction of these remarkable environmental benefits. Landscape, animal welfare and margin and hedgerow maintenance optimise wild life and bio-diversity.

• Organic agriculture is popular with some members of the public but few farmers because the market is less certain. It has strict rules on animal welfare and field margins and an inspector system to police the use of organic regulations. The primary difference with conventional agriculture is the rejection of soluble minerals and synthetic pesticides. However the downside is that lower yields necessitate higher prices and land is thus used inefficiently. Although manures of animal origin and green leguminous manures are used for fertiliser, these cause pollution through nitrate run-off that is similar in amount to a well-managed conventional farm. Analyses of animal manures indicates order of magnitude variability in the major mineral constituents of N P K and Ca no doubt accounting in part for lower yields. These elements wash directly into the soil and are inherently no different from applying soluble minerals. However the organic materials decay more slowly, releasing further minerals for plant growth over a longer time period. A major problem with organic agriculture is the presence of higher populations of weeds that necessitate the use of the plough for control thus damaging the soil and releasing carbon dioxide to the atmosphere. Excess phosphate in manure also induces breakdown of organic material. Although the desire by organic supporters to improve the quality of UK agriculture is commendable, it must be asked whether the assumptions on which organic agriculture is based are credible. The Soil Association emphasises the necessity for maintaining the quality of the soil but ploughing is required because herbicides are banned. For 158 years winter wheat has been grown continuously here at Rothamsted Experimental Station UK, on soil fertilised only with minerals or only with farmyard manure. Identical yields have been obtained throughout that time. While no one would recommend continuous mono-culture cropping as a form of farming, the measurements indicate that greater numbers of micro-organisms in the soil or organic material have no direct relation to wheat yield or sustainability. Further measurements at Rothamsted have also indicated farmyard manure may cause the greatest nitrate pollution of waterways. In compensation organic food offers the consumer a lower level of pesticide residues, that is considered irrelevant to human health by toxicologists because the background of natural food pesticides is so much higher.

If there are current problems with UK agriculture many can be identified as arising from poor management. Pointing out to economically-minded farmers that nitrate and pesticide pollution is money down the drain as it were might change attitudes. Should farmers be licensed to farm after demonstration of managerial competence or qualification much as in other professions?

Systems approaches to farms and farm management.

All living systems are constructed from highly inter-connected networks. Cells are constructed from numerous interconnected chemicals, tissues from interconnected cells and individuals from interconnected tissues. Populations and ecosystems are constructed from numerous interconnected species. The recognition of this systems concept as underpinning all biology and in turn human political and economic systems has had profound impact on understanding their behaviour although it has had little impact currently on farms and farming. Two complimentary approaches are used to investigate the structure and properties of systems. The reductionist assumption is that the operation of the parts determine overall systems behaviour; the systems assumption states that the structure of the whole orchestrates the behaviour of the parts. Systems understanding can often appear counter-intuitive but farmers can recognise that their farm structure determines how they choose their particular farming methods and crops.

Farms are complex but purposeful dynamic systems that take various inputs and generate the output of food and farmer income. Strictly speaking only modelling can investigate the behaviour of such systems but numerous stochastic factors such as price volatility, catastrophic events, poor weather, sudden pest infestation, new technology, storage capacity make it difficult to model at the single farm level. However good farmers are intuitive systems thinkers seeing the essential connections between all the biological occupants of their farm. A good farmer appreciates that care in animal welfare leads to greater milk or protein yields. Good farmers use adaptive management, if necessary buying or selling land or even taking additional occupations as need be to supplement income in poor times. Farm experience and knowledge is hard won. Part of the farmers intuitive systems thinking encompasses seeing the value of maintaining margins, hedgerows that are sites for pest predators and some bird species to help control pests. It gives rise to the construction of beetle banks and other devices on integrated and organic farms to help control crop pests.

Systems thinking emphasises hierarchy, connectance, coherence, synergy and most particularly context. The importance of context cannot be over emphasised. All farms are in part unique structures with variable soil quality on a field-by-field or even metre-by-metre basis, variable but very local climate, variable weed and pest population that is inconsistent throughout the farm. Furthermore the local landscape, trees and other farming and local economic activities interact with the farm itself. It is possible to use a sensitivity analysis to detect the parts of any network that are the most vulnerable to manipulation; good farmers intuitively know the most sensitive constraints that limit income and yield and inevitably vary from year to year. Or if they don't the good farmer experiments and learns that way, a process that needs encouragement.

Stability, sustainability or resilience are system properties and are determined by the particular structure. Most investigations that have taken place have examined what causes the farm system to collapse and in most cases it seems to be poor financial management rather than poor farming that is responsible although the two might be inter-related. However there are two simple rule of thumb considerations that can be introduced.

• Diversity commonly makes for stability. Thus a farmer growing many crops instead of one is more likely to have a stable income; in finance diverse investments again ensure income stability. On this basis crop rotation helps stabilise output, use of multi-lines of resistant and susceptible genotypes in fields should help stabilise against disease and reduce fungicide use. Although it is often thought that greater diversity in ecosystems could make for stability this is contentious. Instead particular kinds of systems structures where organisms are complimentary to each other and where many organism only weakly interact with others but in which certain pivotal organisms are present that interact densely with many others seem to provide substantive stability. Such models may provide a basis for understanding stable farming systems countrywide.

• A second suggestion derives in part from that above. Attempts are being made to construct a minimal bacterial cell that can survive under discrete laboratory conditions. Estimates indicate that 3-400 genes should be sufficient. Yet most bacteria have 3000 genes and the suggestion is that the majority of bacterial genes provide for control systems that enable the bacteria to grow under a multiplicity of different environmental conditions. In this case, stability is provided by numerous control systems rather than one or two.

Applied to farming it can be seen that greater stability will be attained by farming systems that have access to the greatest number of control elements to deal with a variable economic environment, and fluctuating weather and pests. Out of the present farming systems, integrated no-till has the greatest flexibility available since the farmer can choose not only to farm no-till but can change to straight integrated, or conventional or even organic although the latter without appropriate approval from organic associations. The no-till farmer can use a range of pesticides as necessary or even those used by organic farmers. In other words managerial flexibility, adaptive management is the solution that comes from this brief analysis. The straight organic farmer has the least flexibility constrained by regulations and a supposed desire for self-sufficiency that isolates the farm from the stabilising influences of other farming systems and production. Primary problems with organic farming are the inability to control adequately pests and weeds. Even when pesticides are used by the organic farmer, it needs to be pointed out that this represents a reductionist approach in a form of farming that claims to be just holistic. Screening for organic pesticide residues is not currently undertaken and should be. It is perhaps ironic that the two extremes of farming, the unthinking conventional farmer that merely reads instructions on bottles and the unthinking organic farmer that merely lives by the regulations are currently the least sustainable forms of agriculture.

In trying to meld these two recommendations together certain suggestions emerge that need further research. Farmers would be better off continually varying their herbicide and pesticide regime, continually varying the types of herbicide or pesticide as well; change crops frequently in fields, longer rotations, diversify ploughing methods and changing them between different fields. Use the traditional ley (wheat undersown with legumes) variably. Use different genotypes in fields although there are economic problems with acceptance of mixtures of genotypes for human food but not for animal feed. The aim is variety of controls not one monotonous solution to a multifaceted problem.

Systems approaches to integrated pest management.

Integrated pest management represents the perfect example of systems approaches applied to pests. Pests, pest predators (insects, viruses, fungi, some vertebrates), disease organisms, crops, pest food recognition and pheromone sensing form a highly-interconnected system that is used as the knowledge base for integrated pest management. How such systems are persuaded to sustainably reduce pest populations is still uncertain. But impacts at many places on the system are more likely to succeed than just one; as critics of pesticide-use only frequently point out.

It is not difficult to see why pesticides used on their own do have limitations. A network is a robust but sometimes fragile structure (a single mutation can, for example, destroy a bacterial cell). But the robustness comes from the coherence and integrity of the network structure that underpins it. Impact at one point has numerous influences but constraint from the rest of the system resists change. Usually the pesticide kills the pest predators that could help keep the pest in check. Or pesticides destabilise the whole pest/pest predator network which simply collapses leading to flushes of uncontrolled pest reproduction. The pest in the longer term can also acquire resistance quickly.

In a pest/pest predator network what is needed is to strengthen the connections on the pest predator side and thus inevitably weaken those those on the pest side. In integrated pest management various various possibilities are available in terms of increasing pest predator numbers by deliberate inundative release, provision of environments for predator reproduction (such as beetle banks or crop refuges) and the provision of food for pest predators when pest numbers are small. Different rows of crops in the same fields can confuse predators but have current problems in easy harvesting. The introduction of specific pest viruses and pest fungal disease should also help. Genetic manipulation should also provide a variety of possible openings particularly in the improvement of viruses.

Here again a lesson from Nature suggests strategy. Some plants use a moving target model in dealing with pests. As soon as herbivory is detected, resistance mechanisms are randomly increased throughout the plant. The pest is then placed in the uncertain position of not knowing whether the next leaf will be edible or will kill. The response of many pests is to move away from such fields and to try elsewhere. It is now possible to raise the expression of the insecticidal protein, Bt, in the chloroplasts of plants to such high levels that 100% kill is detected. (Chloroplast location has the added advantage of not being expressed in pollen). However there is often a yield penalty when producing resistant plants. The moving target model could be mimicked by mixing susceptible with resistant crops in the same field. The susceptible plants could then act like a refuge reducing the progress towards resistance and would compensate for some loss of yield. The more likely result would be that certain mixtures would be very resistant, the proportion determined by experiment.

Conclusion.

If the farm is viewed as a system it is quite clear that the farmer is the pinnacle of the hierarchy of subsystems that control farm activity. Pesticide and organic regulations undercut the experience the farmer has of his own farm, knowledge which is often hard won and necessary for a robust farm. However in all the discussions about new forms of farming at the end of the day it is the farmer who is the lynchpin determining methods and crops. And the necessity for making a profit is the usually the highest objective because without profit the farm does not survive. Any form of farming can survive if the public buy the produce and pay the price.

All farmers should be encouraged to think in much greater detail about the nature of their farm; good farmers already do but poor farmers probably do not. As in so many areas of human activity poor management and thus poor farming leads to failure, disease, pollution and soil degradation. How agriculture develops depends in the main whether farmers will go along with what is suggested and in the usual terms is whether it will increase profit. Resilient systems depend on growth in communication, knowledge, organisational capacity and wealth. While governments cannot easily legislate about the last one, increased farming activity should generate greater levels of wealth in which more farmers benefit. However governments can accelerate the growth of communication and knowledge and encourage growth in organisational capacity. Those four important system factors have been mentioned in one way or another in this script and are a recipe for creating a resilient system in the face of an uncertain climactic and world economic climate.

Discussion

• Historically, the development of organic agriculture has been a response to reductionist demands. At present the main aims of research and development are to try to increase the robustness of the system as much as possible. There is anecdotal evidence about methods to try to increase robustness, for example multiple cropping systems which suffer few pest problems.

• In conceptual terms, integrated farm management doesn't differ very greatly from organic management. The difference is that one system continues to use pesticides if necessary. If a diversity of crops are grown simultaneously it will stabilise the field, but there will be reduced yield compared to just growing a single crop. Diversity in the cropping can lead to diversity of production for the market, which is a good stabiliser.

• It is hard to consider the farm as a single system in the context of the global food market but there is the need to put artificial boundaries around any system because ultimately the whole world is a system. The farmer reacts adaptively to the system as it changes. Organic farmers try to ring-fence their farm so nothing that affects their organic method of farming is allowed as an input. This is unrealistic, and the bigger picture must be acknowledged.

• The farmer's understanding of management and farmer education is critical. There have been problems in the past with a patriarchal attitude to farmers, where they have been issued with "global" instructions e.g. for pesticides application. This took away a sense of value of the individual knowledge the farmers had and there is a need to return to this principle.

• In organic farming it is clear what procedures are permissible and forbidden through the rules from the Soil Association. In contrast, it means nothing to the consumer if a group of farmers are using ICM- as consumers don't understand the principles involved. Guidelines are available to ICM farmers through LEAF (Leaf.org) and includes prescriptions on issues such as hedgerows boundaries and animal welfare.

GENERAL DISCUSSION

Application technology

In the next 5-10 years it is clear that insecticides will still be the major form of pest control. There is the need to examine appropriate technology for application i.e. finding ways of better applying the product, minimizing harm to the environment and reducing residues. It is unlikely that companies will carry out this type of research as their money tends to be invested into areas of regulation and formulation.

In improving application there should be a greater focus on the crops that the public are most concerned about. Multidisciplinary approaches are important and the merging of research institutes may help with this. There should also be better opportunities for reciprocal transfer of information between the UK and overseas interests.

Regulation

It is a mistake to blame the regulatory system for the relative lack of development of new technologies like biocontrols and the use of botanical pesticides. The drawbacks of commercialisation of such products are an issue that should be addressed.

It is possible for the regulation process to move quickly; in 1995 the "Escort" meeting in the Netherlands devised a new regulatory system for the protection of beneficial insects in Europe. Within four years this was revised into a better system. Short timescales can be effective in reaching solutions.

Loss of the arsenal against pests

A principal concern is that in the event of a real health or food emergency, the chemicals to address this emergency will no longer be available. In the future, the public and politicians may ask why scientists knew that tools were going to be lost but did not speak out, and had no new tools ready to address future emergencies.

There is a need to educate farmers to accept slower acting chemicals and for new active ingredients and new modes of action. Only four to five companies will be able to afford the background science in-house to make major moves forward in novel insecticide discovery. There is a case for significant public sector funding into alternative pest control options.

When considering the evolution of plant-insect interactions, insects are often treated as if they are a uniform group that attack the plants in the same way. This is clearly untrue, for example the distinction between sucking and chewing insects.. The reason why plants exist is because they are extremely well defended against insect attack. There is great scope to further harness the natural arsenal of defences that plants have and to exploit their available biodiversity. It is easier economically to find a magic bullet that will kill all insects in all circumstances. However,there is a need to acknowledge the differences between different types of insects and find appropriate control measures.

It is essential to consider the pesticides that are currently available as a finite resource because there is a real potential for resistance to develop quickly. Chemical companies should not sell a single active new ingredient and these should always be combined with others to protect long-term efficacy.

The need for applied ecology

There is a need for more research in applied ecology and a shift in funding to favour this approach. There is also a need to intensify the use of new molecular tools to understand better individual pests and the economic damage they cause in different systems of production.

Questions that merit further study include: why some pests are more prone to developing resistance than others and the dynamics of colonisation of natural pests in a crop.

The functional value of biodiversity must be remembered. In establishing priorities for conservation it is important to understand the compensatory effect on natural enemies as a consequence of reducing pesticide use.

Appropriate/new technology

This meeting has not substantially discussed the use of conventional plant breeding for resistance against insect pests and sadly there has been a loss of public sector funding in this area.

The sophisticated use made of pesticides in the UK is not transferable across the world. If pesticides are used without the rest of the support system of technology, they will not work successfully.

It is not possible to solve tomorrow's agricultural problems with today's technology. This is very true for insect control. New and appropriate tools are needed that are integrated into an ecosystem in a holistic fashion. In some cases it is worth revisiting existing technologies to assess their potential future benefit.

There needs to be a ratchet mechanism in British agriculture to facilitate the adoption of new technology. Future progress will be based on the evolution of current thinking rather than a revolution of something completely new.

Consumer issues

There is a strong need to "decriminalise" pesticides in the public perspective. The fact that there is a toolkit available incorporating a number of options for pest control should be a reassuring message for consumers. It is important to involve consumers and farmers in decision-making processes in order to build back trust.

Consumers should feel reassured, particularly on the subject of rigorous regulation. It is important to find a way to tell consumers that there is a need for some pest control. Communication of issues of risk and uncertainty are also important, particularly in the new technologies such as GM crops.

Farmers

Farmers should be encouraged to move away from a uniform approach to their farms (e.g. cutting hedges around the whole farm at the same time) and to treat each part in an appropriate way to maximise diversity.

There should be more cross-referencing between the interests of different farming systems, whether organic or conventional. There should also be cross-referencing between intellectual interests and economic interests. There is a strong imperative for change and the fragility of world food security means that effective pest control cannot be forgotten.

A summary of the four main points from the meeting was suggested:

Implementation: of existing and emergent technologies

Integration: of pest management options

Interaction: between all stakeholders

Innovation: to provide new technologies