87 . . .. . . .. .. .. .. * . 92 .. .. .. . . * . * . .. . . 95 Current situation . . . . Significance of residues in soil . . .. .. .. .. .. 101 Future possibilities 102 Conclusions Bibliography Pollution of Soils I. J. Graham-Bryce, M.A., B.Sc., D.Phil., and G. G. Briggs, B.Sc., Ph.D. * . Rothamsted Experimental Station, Harpenden, Herts * . Introduction Routes of arrival in soil . . .. .. .. .. Interaction with soil Breakdown .. .. . . * . .. .. .. . . .. .. 0 . .. .. 103 . . .. 104 .. Persistence . . .. .. . . .. . . .. .. 88 . . .. .. .. .. .. .. 89 .. . * . . .. .. .. 97 .. 0 . . . . . .. .. .. .. INTRODUCTION Materials intended to improve crops, and waste products from other activities, have always become incorporated in soils.Some of these materials are persistent, for example even in Roman times the use of a persistent pesticide- arsenic-was recommended. Until recently, however, deleterious effects were localized and sporadic. Awareness of radiochemical fallout and the occur- rence of persistent synthetic pesticides in soils has led to much public concern, and pollution of soils is now widely regarded as an important problem. The concept of pollution is imprecise and judgement as to what constitutes a pollutant is very much a matter of opinion. The materials involved are not all unequivocally injurious: some may be harmless or beneficial when present at small concentrations or in other situations. Others are ambivalent; for example organochlorine insecticides may control undesirable soil insects while also killing beneficial insects and causing other unwanted effects.Harmful concentrations of biologically active chemicals may arise naturally. For example, soils may contain concentrations of elements such as copper, zinc and molybdenum large enough to be toxic to plants or stock, although smaller concentrations are essential for proper growth. In this article, we define a polluted soil loosely as one which contains enough of a given material to cause undesirable effects in the soil, the atmo- sphere above it or water draining from it. The community must decide what it regards as an undesirable effect. Acceptable residue levels in crops have usually been set by dividing the concentration likely to harm human beings by a substantial safety factor.From this point of view a polluted soil is one that gives crops containing more than this arbitrary level. However, using the crop alone as an indicator gives no information on the amount of material either in water running off the soil, or in the air above the soil, or the effect Graham-Bryce and Briggs 87 on organisms in the soil other than the plant. Detection of the specified amounts requires adequate methods of analysis, and the prominence of per- sistent pesticides and radioactive fallout, which evoke rather special popular alarm, is to some extent an accident resulting from the existence of extremely sensitive and rapid analytical techniques. Fractions of a picogram of organo- chlorine insecticides can be detected by gas chromatography with electron- capture detection, and if analytical methods in general were equally sensitive, contamination by other pollutants might cause increased concern.Analysis of natural materials by gas chromatography at the greatest sensitivities is difficult and some reports of very small amounts can be regarded sceptically. There are reports of DDT residues being identified in soil samples which were sealed and stored before it was used as an insecticide. Nevertheless, pollution of soils is a real problem and its implications are particularly serious for materials such as 90Sr and organochlorine insecticides, which have a long life and can become transferred through stages in food chains.The same basic principles govern the behaviour and ultimate fate of any material reaching the soil. The climate, and the physical and chemical properties of the soil and the applied material, are the principal factors influencing the distribution between the soil, the atmosphere above it and water draining through or running off it. Given a suitable combination of these factors, usual agricultural methods could give rise to polluted soil, atmosphere or water. In practice there are remarkably few materials giving serious problems. With the benefit of hindsight, it seems probable that these could have been predicted, and we shall attempt to discuss the principles involved, using examples mainly from the behaviour of pesticides. We have confined our attention strictly to soil.However, in considering processes that lessen concentrations in soil and therefore decrease pollution, it must be emphasized that removal by volatilization and leaching merely pass on the contaminant to another part of the environment where it may cause even greater problems. Failure to recognize that dangers to the community do not cease when persistent materials are removed from the immediate area being considered has contributed to the present difficulties. Biologically active chemicals only cease to be potential dangers when they are degraded into harmless decomposition products. ROUTES OF ARRIVAL IN SOIL Application of agricultural chemicals adds extraneous materials ranging from solid fertilizers to gaseous fumigants directly to soil. Fertilizers, sys- temic insecticides and many herbicides, such as the substituted ureas and triazines, are applied to be taken up by plant roots.Soil insecticides and fumi- gants are applied to control soil-borne pests such as wheat bulb fly, wire- worms and nematodes. R.I.C. Reviews For optimum performance, agricultural chemicals are formulated in many different ways including liquids, granules and dusts. They may be applied to the soil surface or incorporated either uniformly or placed in various ways, or may be in the form of seed dressings. As a result, there is a wide range of initial distributions of concentration in the soil, from a uniform value several centimetres deep to bands of large concentration on the surface.88 In addition, much of the pesticide applied to shoots and leaves reaches the soil by being washed off by rain, by leaf fall, or simply because application inevitably cannot be restricted to the desired target. Similarly, other polfu- tants reach the soil unintentionally. Much of the material discharged into the atmosphere eventually settles on the soil or is brought down in rain. Radio- active fallout is a well publicized example on a worldwide scale. Dust, smoke and SO2 from fuel are also widespread and their effects on soil acidity around cities are well established. The quantities of material involved are often large compared with amounts of 1-2 kg ha-1 (0.1-0.2 g m-2) commonly applied for pesticides, and in heavily industrialized areas as much as 1 kg of dust and grit may be deposited per square metre each year, although this material may have small biological activity.On a more local scale specific materials may be potentially dangerous. For example, smoke from brickworks contains considerable quantities of fluorine and gives rise to concern, because large amounts of fluorine can be retained by growing plants which are eaten by stock. A notorious example of localized pollution is the Lower Swansea Valley, where severe contamination of the soil by Cu and Zn, from metal smelting during the 18th and 19th centuries, devastated the traditional plant life. INTERACTION WITH SOIL Materials in soil become distributed between the soil solution, soil air and soil solids, and the nature of this distribution greatly influences subsequent behaviour.Partition between the solution and air at equilibrium can be predicted from a knowledge of the solubility and vapour pressure. Although some chemicals have such small volatility that the air phase may be neglected, many pesticides have significant vapour pressures and vapour movement often makes a considerable contribution to their transport to sites of action. This applies particularly to fumigants such as methyl bromide and D-D mixture (dichloropropene plus dichloropropane) applied to control nematode pests inhabiting the soil. Partition onto the soil solids occurs by sorption on the highly reactive surfaces, and also insoluble compounds may be formed with soil constituents.Soils contain very small particles, especially in the clay fraction, where particles are of colloidal size and have an enormous specific surface. Typical values for the readily accessible surfaces of whole soils measured by nitrogen adsorption range from 25-100 m2 g-l, so there is a very large interfacial area at which non-specific surface adsorption can occur. The nature of these surfaces ranges widely-the soil colloids include layer lattice silicates, various oxides and hydroxides and organic materials. Many of the surfaces carry permanent or pH-dependent charges ; for example, the surfaces of the principal clay minerals consist of planes of oxygen atoms or hydroxyl groups associated with localized permanent negative charges caused by isomorphous replace- ment of cations within the lattice. The edges of clay minerals, and carboxyl and phenolic groups associated with the organic matter give rise to pH- dependent charges.From the nature of these surfaces, therefore, it will be clear that there are also many ways in which more specific adsorption can take place and that a large proportion of many pollutants will become Graham-Bryce and Briggs 89 associated with the surfaces. Even in an ‘air-dry’ soil the relative humidity is enough for most clay surfaces to attract at least a monolayer of water so that, except in the very top soil during hot dry conditions, this sorption takes place in the presence of water which competes for the surfaces. The extent of adsorption can be described by adsorption isotherms deter- mined from slurry experiments with large solution/soil ratios.For pesticides, such isotherms have usually been found to fit the empirical Freundlich equation, x/m = KCn (where x/m is the amount adsorbed per unit weight of soil, K and n are constants and C is the equilibrium concentration in solution) with values of n in the range 0.7 to 0.95, although usually only a small fraction of the avail- able surface is occupied by the adsorbed molecules. Values of I(, reflecting the affinity of the pesticides for the surfaces, range widely. In the slurry experiments most of the sorption occurs within minutes and equilibrium is usually reached within a few hours. In relation to subsequent behaviour, the extent to which adsorption is reversible when solution concentrations decrease is very significant.The possibility of irreversible sorption has some- times been suggested, although with no clear theoretical reasons, but where tested by properly designed experiments the process has always been shown to be largely reversible. However, desorption may be considerably slower than adsorption, particularly on soils with much organic matter. The bipyridylium herbicides, paraquat and diquat, provide examples of extreme sorption which is responsible for their outstanding property of being inactivated in contact with soil. The strong sorption brings the solution concentration to immeasurably small amounts at normal rates of use so that the herbicide becomes unavailable to plant roots.These herbicides exist as cations with a flat configuration and the strong binding results from both electrostatic forces and forces resulting from the close approach of a highly polarizable flat ion to the flat charged clay surfaces. Most other pesticides are less strongly sorbed, although they may still be present predominantly in the adsorbed state. Many studies have shown that, for un-ionized materials not involved in ion-exchange reactions with the charged sites, the extent of sorption in different soils is related to the content of organic matter rather than to the mineral fraction, as it is with the bipyridylium herbicides. Many of these pesticides are hydrophobic and presumably they have affinity for the hydrophobic surfaces in the organic matter.Correlations between partition coefficients in oil/water systems and sorption by soil have been obtained for such compounds. Some pesticides, such as sodium chlorate, TCA (trichlor- acetic acid) and some of the organophosphorus insecticides such as dimeth- Paraquat cation I,l’-Dimethyl-4,4’-bipyridylium ion CH,- / -7 D iq uat cation 9,IO- Di h yd roSa, I Oa- d iazon iap hen an t h rene ion 90 R . I. C. Re views Ci&. COOH TCA Trichloroacetic acid Dimethoate S /I (CH30)2PSCH2CON HCH3 Di methyl S-(N-met hyl- carbamoyl methyl)- phosphorothiolothionate oate, are only very slightly sorbed and may be regarded as the opposite end of the scale from the bipyridylium herbicides.Adsorption, by decreasing the concentration available in the soil solution or soil air, has a pronounced effect on almost all aspects of later behaviour. In this review we are interested in its influence on processes leading to loss of the contaminant from the soil. There may be considerable losses during application of a calculated dose of pesticide to soil; for example, spray may be lost by drift, or dust may be blown away, and some pesticide may also be retained by the machinery used for application. Apart from these losses during application, disappearance may be considered under two broad headings: transfer and decomposition, both of which may be affected by adsorption. Over a long period, significant mixing may occur through the activities of soil fauna such as earthworms and mites, but there has been no systematic study of these processes. Except for this rather uncertain biological transport, mobility within the soil depends solely on the physical processes of molecular diffusion in soil solution and, for volatile materials, in the soil air space; and mass flow, that is, transport by flowing water.Diffusion in soils is retarded by the tortuous and restricted pathways through the pores and by adsorption, which immobilizes the portion of the diffusing material associated with the solid. Diffusion in soil water is slow and because the average displacements are proportional to the square root of time, it decreases in effectiveness with time. (Average displacements over a period of one month are likely to be of the order of 1 cm.) It is therefore negligible over long distances.Vapour diffusion is approximately 10000 times faster than diffusion in solution at the same volume concentration, so that even for materials with large waterlair partition coefficients it may be a more significant mechanism. Near the soil surface, air movement and changes in barometric pressure may assist vapour diffusion and increase evaporation. Evaporation may be very important in removing pesticides from soil. It has been calculated that a pesticide with a molecular weight of 200 and vapour pressure of 10-4 mmHg (0.013 Pa) could lose as much as 20 kg ha-1 month-1 from an inert surface during a temperate summer, whereas usual amounts applied are in the region of 1-2 kg ha-l.Although this loss would be diminished by sorption and solution in soil water, these factors would be partially offset by the upward capillary movement of water to replace that evaporating from the soil, which would cause the pesticide to concentrate at the surface. Also, many pesticides have vapour pressures considerably greater than 1 0-4 mmHg. At first sight an equally obvious route for removal from soil would be leaching in drainage water, but its importance is probably frequently over- estimated. Water is lost from soil by evaporation and transpiration by plants, Craham-Bryce and Briggs 91 2-Chloro-4,6-bisethylamino- I ,3,5- C,H,NH Si mazine triazine BREAKDOWN NHC,H, as well as by drainage.In summer, particularly in the drier parts of Britain, evapotranspiration usually exceeds precipitation and the soil is drying out rather than transmitting water to drainage. Over the year as a whole, for an average rainfall of 75 cm in south-east England, drainage to water tables is likely to amount to only about 25 cm although this depends considerably on local topography. The effectiveness of this percolating water in leaching a pollutant depends on the extent of sorption; the soil may be regarded as acting like a crude chromatographic column. Thus, moderately sorbed materials such as simazine may be used selectively to control shallow rooting weeds in deeper rooting crops, such as field beans, because very little leaching takes place provided the soil contains enough organic matter.In contrast, weakly adsorbed materials like sodium chlorate are readily distributed throughout the root zone. Sometimes there may be channelling (for example during thunderstorms) down the large cracks caused by shrinkage in clay soils during summer, but the importance of this has not been assessed. Except for materials of slight solubility, such as DDT, leaching is unlikely to be limited by solubility. Two cm of rain provides 200 000 kg of water per hectare, enough to dissolve 1 kg ha-l of a pesticide with a solubility of 5 ppm. Adsorp- tion allows the soil to accommodate far more molecularly dispersed pesticide than could be dissolved in the soil water alone. It has sometimes been sug- gested that solubility and ease of leaching are related.From what has been said it will be clear that adsorption rather than solubility governs leaching, and these two properties depend on quite different factors. Apart from application losses and volatilization from the soil surface, materials are lost from soil principally by microbial or chemical decomposi- tion, or a combination of both. Usually plant remains initially decompose rapidly in soil, but a small fraction remains as very stable polymeric materials. Radiocarbon dating of soil organic carbon gives an average age of about 1000 years for the carbon in surface soils, clear evidence of the great stability of some natural materials in soil. For pesticides, many well known microbial conversion steps have been observed in soil, such as oxidation and reduction, N- and 0-dealkylation, nitrile, amide and ester hydrolysis, dehalogenation, aromatic ring hydroxyla- tion and epoxide formation from alkenes.For materials that are readily metabolized there is usually a lag phase followed by a rapid decay. Organisms have been isolated that degrade various pesticides in pure culture and the breakdown pathways determined. But the situation in soil is much more complex because there is a mixed population of organisms and sorption or 92 R.I.C. Reviews chemical reaction can remove materials, that are easily metabolized, from solution. For example, there are at least three pathways known for the metabolism of 2,4-dichlorophenoxyacetic acid (2,4-D) in solution but the pathway in soil is not known.Where decay is slow, chemical and microbial breakdown are difficult to distinguish but many stable materials are probably decomposed incidentally during the decomposition of soil organic matter by the action of extracellular enzymes. For example, this has been suggested for picloram (4-amino-3,5,6-trichlorpicolinic acid), which decomposes at a rate proportional to the amount of organic matter in the soil, and so proportional to the microbial activity. Agricultural soils are usually moist and well aerated, although they can be anaerobic at times, especially in less accessible and finer pores. Conditions are therefore suitable for hydrolysis, and oxidation-reduction reactions and chemical transformations of added chemicals by these reactions can be expected.In vitro, many of the potential hydrolysis reactions proceed much faster at the extremes of pH so that in agricultural soils which are mostly near neutral, slow rates might be expected. However, because of electrical double-layer effects in solutions near charged surfaces, the pH in the environ- ment of adsorbed pesticides may differ considerably from that measured in dilute salt suspension, leading to faster reactions. Chemical hydrolysis of nitriles, esters and activated aryl and alkyl halides has been demonstrated. For compounds such as the chlorotriazines, reaction to form the inactive hydroxytriazines takes place on the surface of the soil colloids. Dichlobenil (2,6-dichlorobenzonitrile) is apparently hydrolysed in solution and sorption onto organic matter slows this hydrolysis.Malathion and some other organophosphorus insecticides are hydrolysed in radiation- sterilized soils by a heat-labile factor extractable with sodium hydroxide. Many fungicides rely for their activity on reaction with thiol, amino and other nucleophilic groups in organisms and so can react with suitable groups on soil F' 2,4-D 2,4-Di c h I o ro p h en ox yaceti c aci d Picloram CI 4-Amino-3,5,6-trichloropicolinic acid Dichlobenii 2,6-Dichlorobenzonitrile Graham-Bryce and Briggs 93 Malathion !I I Trifluralin S COOCzHj 5-[ I ,2-Di(ethoxycarbonyl)- ethyl]dimethylphosphorothiolothionate (CH30)zPSCHCHzCOOCzH5 NO, q- c1 Cl 2,6-Di n itro-N,N-di propy l-4- trifluoromethylaniline Dieldrin endo-5,8-dimethanonaphthalene I ,2,3,4,10, I O-Hexachloro-6,7-epoxy- I ,4,4a,5,6,7,8,8a-octahydro-exo- I ,4- organic matter.Another example of reaction with soil organic matter is pro- vided by the herbicide trifluralin which is converted in soil to a substituted o-phenylenediamine that condenses with organic matter. Reductive cleavage of the alkali-extracted organic matter regenerates the diamine. Many agricultural chemicals have been observed to undergo photo- chemical reactions in solution, but these do not seem to be very important in practice, although small amounts of photoproducts of dieldrin and aldrin have been detected in soils treated with large amounts and there is some evidence for photochemical breakdown of paraquat.Overall kinetics of disappearance from soils are usually close to first order with amounts customarily applied to soil. Where there is chemical break- down by reaction with active sites in soil the distribution of the material becomes important. When active sites are few, the concentration of the chemical may be large in relation to the number of sites. Breakdown will then tend to be independent of concentration when the chemical is not distributed through a large volume of soil, either initially or by subsequent leaching. This may explain the increased persistence of the triazine herbicides during dry summers. Similarly, the rate at which picloram decomposes is proportional to concentration at small concentrations but independent at large ones.First-order kinetics have been successfully used to calculate long term residues from dieldrin treatments, using a half-life of about four years. Two examples illustrate the interaction of some of these factors in decom- position. Anhydrous ammonia applied below the soil surface will be lost partly as vapour unless the surface is sealed by compaction behind the injector. There will be a zone of large concentration around the injection point where pH is high and condensation reactions with soil organic matter can occur. Ammonium ion is held by ion exchange and some can be fixed in the interlayer spaces of expandable-lattice clay minerals. As the ammonia diffuses outwards, pH falls and microbial oxidation to nitrate occurs. Micro- bial activity is temperature-dependent and conversion to nitrate is much 94 R.I. C. Re views CI \ Diuron 3-(3,4-DichlorophenyI)- I , I -dimethyl- urea faster during summer than spring. Nitrate is not sorbed by soils and is readily leached. Diuron (N’-3,4-dichlorophenyl-N,N-dimethylurea), a pre-emergence herbi- cide, is fairly strongly sorbed by organic matter. Micro-organisms convert it slowly to the N-methyl derivative, which is more water-soluble but more strongly sorbed than diuron and is therefore much less herbicidally active in soil although about equally active in nutrient solution. A further microbial N-demethylation produces the still more strongly sorbed and herbicidally inactive 3,4-dichlorophenylurea, which then breaks down to 3,4-dichloro- aniline.The amine is considerably less strongly sorbed and has a moderate vapour pressure. When concentrated, two molecules can couple to give a mixture of products including 3,3’,4,4’-tetrachloroazobenzene, a strongly sorbed and very stable compound which probably has some biological activity. However the decomposition of diuron in the field is so slow (about a year for complete disappearance) that the amine is decomposed almost as fast as it is produced and concentrations do not build up sufficiently for bimolecular steps to be observed and only unimolecular decomposition steps are important. Finally, contaminants may be removed from soil by growing crops. Al- though this decreases contamination in soil, it is undesirable because it brings residues directly to stock or people.Extensive precautions are taken to limit residues of pesticides in crops and such uptake is a very minor mecha- nism for removing unwanted materials from soils. It may be calculated that even if a crop contained the largest concentrations of residues permitted, only a small fraction of the applied material would be removed from the soil. PERSISTENCE On the basis of this discussion of the factors affecting the behaviour of chemicals in soil, the properties of a potentially dangerous pollutant can be described. To prevent losses by leaching or evaporation, it would be fairly strongly sorbed and have small volatility, but not to such an extent that it would become deactivated and unavailable to exert any biological activity in the soil.It would be stable chemically even under the extreme conditions of the adsorbed state and it would not form a suitable substrate for micro- organisms. As would be expected, materials reaching the soil fit these proper- ties to different degrees and have a wide range of persistence. Of course, persistence is a variable property and is affected by factors such as climate and soil type, but approximate times for at least 75 per cent of the applied dose to disappear are given in Table 1 for some representative compounds. These times relate only to persistence in soil; it must be emphasized again that volatilization and leaching may contribute to removal of the more stable materials so that their life in the environment as a whole may be considerably 95 Graham-Bryce and Briggs Table 1. Persistence of materials in soils: approximate times for at least 75 per cent of added dose to disappear longer.In most cases it is very difficult to assess the relative importance of the different processes contributing to disappearance. In spite of the complexity of the processes involved and the large variations with climate and soil type, the kinetics of disappearance often approximate to those of a first-order reaction with the rate of disappearance proportional to the amount present. This is a very broad generalization and the last residues often disappear more slowly than would be predicted, but it does allow potential accumulation of repeatedly applied chemicals to be calculated approximately.There is an initial build-up until the rate of loss between applications has increased to equal the amounts applied. For half-lives of up to one year, residues will not be more than twice the annual application, whereas a half-life of four years leads to a maximum accumulation of approximately six times the annual dose. Not all pesticides show these kinetics. Potentially more dangerous situations arise when the original pesticide decomposes to a further toxic residue which is not lost, but accumu- lates in proportion to the amount of pesticide originally applied. Such be- haviour has been suggested for heavy metal and arsenic compounds; for 1-5 Disulfoton Diethyl S-[2-(ethylthio)ethyI] phosphorothiolothionate Chiorfenvinphos 2-C h loro- I -(2,4-d ichloropheny1)-vinyl diethyl phosphate p,p’-DDT I , I, 1 -Trichloro-2,2-di-(4- chloropheny1)ethane Compound Methyl bromide Ammonia 2,4-D Disulfoton C hlorfenvinphos Diuron S i mazi n e Dieldrin DDT 96 Normal rates of use per application (kg ha-1) 1000 up t o 100 I 1-2 1-4 0.5-2 0.5-1 -5 1-5 Use Fumigant for soil sterilization Fertilizer Foliar herbicide Systemic organophosphorus insecticide 0 rgano p hos p hor us soi I insecticide Soil-applied substituted-urea herbicide Soil-applied triazine herbicide Soil and crop organochlorine insecticide Soil and crop organochlorine insecticide CCI, Persistence <I week 3 weeks 1 month 6 weeks 6 months 8 months I year > 3 years > 5 years 0 CHCI CI I e C I CI !H - - R.I.C. Review example the herbicide methane-arsonic acid (MSMA) is metabolized by soil micro-organisms to arsenate, which is not dissipated and can accumulate steadily in soils. For a pesticide to cause major pollution problems it must be used widely at large doses, as well as have the appropriate properties. This implies that it will not be very specific in its action and must be cheap. The present situation with regard to pollution of soil may be related to these requirements, together with the properties of different materials reflected by their persistence times such as those shown in Table 1.CURRENT SITUATION In relation to the criteria for a potential pollutant, most herbicides are too short lived to cause problems. Persistence is obviously an undesirable property because of the danger of harming subsequent crops, and excessively persistent compounds would therefore never be developed commercially. Herbicides that are stable are either not toxic enough to organisms other than weeds or not sufficiently available in soil to produce changes in flora and fauna significant in comparison with those brought about by growing a crop. As herbicides are often required to act selectively between weed and crop, blanket treatments of large areas with a single compound are unlikely and the appropriate chemical is usually chosen for each specific purpose.The different sensitivity of crops in a rotation usually requires a different material each season. Dosages are small and repeated treatment during a season unlikely. The possibility of crop failure is a powerful constraint to over- dosing and insidious unknown accumulation is unlikely because presence of the herbicide would be revealed by plant damage. For all these reasons, herbicides do not fit the requirements of an ‘ideal’ pollutant well and it is not surprising to find that pollution by herbicides is not widespread. However several herbicides can persist for more than a year and individual ‘carry-over’ problems have been reported, particularly with triazine and substituted urea herbicides which are widely used.Such problems are most likely in a rotation where a sensitive crop succeeds a resistant one. The increasing trend towards growing the same crop, especially cereals, for several successive years may lead to more cases where a single compound is used repeatedly, but fortu- nately the materials most commonly used in this way are not very persistent. 2,3,6-Trichlorobenzoic acid (2,3,6-TBA) and 3,5,6-trichlorpicolinic acid (picloram) can persist for long periods and may leave residues in straw that are toxic to plants grown on straw bales in greenhouses, but fortunately, as with most herbicides, these compounds are relatively harmless to mammals. Fumigants are usually even more transitory than herbicides, but insecticides show a very wide range of persistence and the longer-lasting ones are potential causes of trouble, especially as some are toxic to organisms other than pests.Some of the organochlorine insecticides such as DDT and the cyclodienes seem almost tailor-made to fit the requirements for a pollutant discussed earlier. They are sparingly soluble and not very volatile. They are strongly sorbed, are inert chemically and are not readily decomposed by micro- organisms. It should perhaps be emphasized that the persistence of organo- chlorine insecticides was originally regarded as a very desirable feature Graham-Bryce and Briggs 97 because it offered prolonged protection against soil-inhabiting insect pests from a single application. However, this property, combined with toxicity to many different insect pests and small cost, led to very widespread use, so that their prominence as pollutants can be readily understood.Most reports of pollution of soils by pesticides relate to these organochlorine insecticides. The problems are increased because these materials are lipophilic and can be stored in fatty tissues and concentrated from soil or water through food chains. Widespread concern has led to extensive if somewhat inconsistent monitoring for persistent pesticides in many countries and to a voluminous literature on the subject. Most of the investigations with soils are for agricul- tural land, although the presence of organochlorine insecticides in air and rain indicates that residues probably occur very widely in untreated land also.Table 2 shows some typical residue figures from various sources ; a much more extensive summary is given in the review by Edwards, cited at the end of this article, from which these are taken. The most common contaminants in agricultural soils are DDT and its relatives, which were detected in almost all samples analysed. Average values for all the surveys are of the order of a few parts per million. Orchard soils contain the largest amounts, reflecting the intensive treatment given to fruit trees. Repeated spraying during the season of 1-2 kg ha-1 per application to control insect pests on the foliage has given rates of up to 40 kg ha-1 a-1 for many years in some orchards. Much of this eventually reaches the soil and residues exceeding 100 ppm are not uncommon, especially in the US.98 Persistence of residual herbicides when applied at excessive rates. This barren patch in a potato crop marks the spot where a tractor stood with the herbicide spray left on five years before. (Effects are still visible after nine years.) R.I.C. Reviews No. of related compounds y-BHC sites Heptachlor and epoxide Chlordane Dieldrin Crop ref. Country Aldrin - - A B C 0.20 7. I 0.3 9.7 21 5 10 2 0.09 0.15 - - 0.4 0.5 0.02 - 0.15 - 0.04 0.02 0.06 4 5 6 9 4 0.2 0.3 0. I 0. I 0.3 - 0.8 I .2 1.4 1.5. 3.2 9.5 61.8 5 I1 8 Table 2. Mean residues of organochlorine insecticides in soils (ppm) DDT and Great Britain potatoes orchards arable orchards Canada sugar beet pasture corn cereals glasshouse tobacco vegetables orchards us turf and cu It ivated cotton desert and prairie arable 0.04 - 0.33 - - t D Davis, Ann.appl. Biol., 1968, 61, 29 Pest. Mon. J., 1968, 2, 93. 2.9 0.03 - - 2.4 I .6 227 3 5 41 Butcher and R. T. Murphy, J. econ. Ent., 1965, 58, 1026 0.06 B C. A. Edwards, N.A.A.S.q. Rev., 1969, no. 86, 47 C B. N. K. E J. E. Fahey, J. W. G W. L. Trautmann, 0.7 t = trace refs: A G. A. Wheatley, J. A. Hardman and A. H. Strickland, PI. path., 1962,81, I I D C. R. Harris, W. W. Sans and J. R. W. Miles, J. ogric. Fd Chem., 1966, 14, 398 F C. Lahser and H. G. Applegate, Texas J.Sci. 1966, 18, 12 CI+ CI Aldrin 1,2,3,4,10, I O-Hexachloro- I ,4,4a,5,8,8a- hexahydro-exo- I ,4-endo-5,8- dimethanonaphthalene CI CI Chlordane I ,2,4,5,6,7,8,8-0ctachloro-3a,4,7,7a- tetrahyd ro-4.7-methanoindane CI CI CI CI Heptachlor I ,4,5,6,7,8,8-Heptac h loro-3a,4,7,7a- tetrahydro-4,7-methanoindane CI CI Many soils used extensively for growing vegetables also contain large amounts of DDT. DDT is only slightly more persistent than dieldrin and the quantities of aldrin/dieldrin and DDT used in agriculture are comparable, so that it is not surprising to find that this is the next most important residue, although amounts are usually considerably less than of DDT. Compared with DDT and aIdrin/dieldrin, quantities of other chlorinated insecticides used in Great Britain are extremely small and very few residues have been found.However, in the United States, relatively large amounts of chlordane and heptachlor are common. The large residues of organochlorine insecticides are often the result of prac- tices unlikely to be repeated with present knowledge. Addition of persistent pesticides to fertilizers, and ‘insurance’ treatments given without knowledge of whether pests were present or not, have often contributed unnecessarily to residue problems, and the practice of frequent blanket spraying, for ex- ample in orchards, with a single broad-spectrum pesticide is now regarded as highly undesirable because of its effects on predators and beneficial insects, and the danger of selecting resistant populations of pests, Although residue problems with synthetic organic pesticides probably receive most publicity now, earlier inorganic materials were potentially at least as dangerous.Lead and arsenic residues from lead and calcium arsenate insecticides often reached alarming levels, again mainly in orchard soils where figures of several hundred ppm were not unusual. The use of these insecticides has declined since the war, but some old orchard soils remain unproductive because the residues are phytotoxic to subsequent crops. Al- though the danger with the organic arsenicals now used as herbicides and defoliants is less, their prolonged use should be carefully watched. 100 R. I. C. Re views Other agricultural chemicals cause less concern.Some of the older copper fungicides and mercury seed dressings leave residues of heavy metals which are persistent, and although occasional examples of copper killing soil fauna in orchards have been reported, these are not widespread in this country. Quantities used in seed dressings are small so that mercury residues are not large compared with natural amounts in soil. More recent synthetic organic fungicides have tended to be reactive compounds that are soon decomposed. The major plant nutrients added in fertilizers are nitrogen, phosphorus and potassium. Phosphorus is immobilized by forming various sparingly soluble phosphates and potassium is firmly held in the soil by ion exchange. Am- monium ions are also held by ion exchange but are converted by micro- organisms to nitrate and in this form nitrogen is very mobile and easily leached.In grassland, the crop intercepts most of the nitrate but with arable crops large losses can occur in spring during wet seasons. Clearly a large supply of nutrients is desirable to grow large yields but nitrate leached from the soil may be one of the nutrients stimulating the growth of micro-organisms in streams, lakes and reservoirs. The possibilities of nitrogen and phosphorus being moved from agricultural soils to water supplies have been discussed by Cooke and Williams. On a smaller scale, sewage cake from town sewage works is used as manure in some areas. This often contains large amounts of heavy metals from industrial eWuents, and can therefore cause pollution by heavy metals in soil.These are very persistent and can be taken up by plants so they may be a considerable danger in localized areas. SIGNIFICANCE OF RESIDUES IN SOIL The significance of the pesticidgresidues found in soils is difficult to assess. Except in ‘carry-over’ problems of herbicides or phytotoxic effects of arsenic residues in old orchard soils, there is no evidence that pesticide residues generally affect soil fertility. Even in the badly contaminated orchard soils, established deep-rooted trees continue to grow well although soil structure may have deteriorated and a surface mat of organic matter accumulated because the soil fauna have been killed. There seems to be little direct hazard to the consumer because few residues are transmitted to the fruit, although use of such soils for root crops, which absorb residues more readily, might contaminate food supplies. However, organochlorine residues are common in soil fauna, particularly earthworms inhabiting treated soils, and there is evidence that soil invertebrates may accumulate the insecticides to larger concentrations than in the soil.Birds that feed on these animals are at risk. The long-term biological and ecological significance of these residues is even less clear. There is no doubt that the composition of the soil fauna and soil micro-organism population is disturbed by pesticides, but whether this is permanent and what effects it has are largely unknown.Without good experi- mental information, it is wisest to proceed with caution, as is now generally recognized. The existence of large residues in soil is also undesirable as it acts as a reservoir for contamination of water supplies and other parts of the environment. Graham-Bryce and Briggs 101 We suggested earlier that with present knowledge current soil pollution problems might have been predicted. The dangers from long-lived materials soluble in lipids are now apparent and, with the present emphasis on less persistent chemicals, if DDT and some other organochlorine insecticides were newly introduced, they would probably not now pass the approval schemes operated by many countries. For many established uses, DDT and other chlorinated hydrocarbons are the best materials available and their use will continue until there are effective alternatives.The kinds of danger caused by these compounds are appreciated and perhaps more attention should be given to predicting other types of potential danger which have not so far been considered. Some processes which could be important will be discussed, but it must be emphasized that these examples are used only to illustrate possible types of mechanism and there is no evidence that these dangers exist in practice, Two old established materials that have been little investigated are penta- chloronitrobenzene (PCNB) and the fumigant mixtures containing dichloro- propene and dichloropropane (D-D mixture). PCNB has been used on a small scale in the US at about 110 kg ha-l to control potato scab. It is reported to persist in soil and its breakdown products are not known. Reduction of nitro groups to amines does occur in soils and, with such large doses, con- densation of the amines to highly chlorinated azo compounds is a possibility.These azo compounds are stable and soluble in fat and further work seems Deterioration in the structure of an orchard soil caused by repeated application of copper fungicides t o the foliage. (Left) Orchard soil showing weakly developed structure and accumulated surface mat of organic matter resulting from absence of earthworms. (Right) Neighbouring soil, uncontaminated by copper fungicides. (Courtesy Dr J. M. Hirst) FUTURE POSSIBILITIES R.I.C.Reviews 102 desirable if PCNB is likely to be used widely. It should be pointed out that the necessity for applying PCNB is dictated by the consumer. There is nothing wrong with potatoes whose skin is affected by scab, but they sell badly in the prepacked market. The consumers’ choice gives the farmer no alternative but to try control measures. Soil fumigation has long been a practice for some crops that repay cost of treatment. Recent work with dichloropropene/dichloropropane mixtures makes fumigant treatment in arable crops likely. There is little published work on the fate of these materials in soil and even at the smallest rates used, about 70 1 ha-l, a small percentage conversion to polymeric products could leave a significant residue of a chlorinated, fat-soluble material of unknown biological activity.Paraquat and diquat are relatively stable and can be detected chemically for considerable periods after application. But they are bound so strongly to mineral particles that they are not available to exert biological activity once they reach the soil. Their continued use in large quantities is certain and repeated applications could lead to much potentially active material being held in the deactivated adsorbed form. As highly-polar doubly-charged cations, they lack the lipid solubility that allows concentration of chlorinated hydrocarbons in food chains or by aquatic organisms but some questions about their eventual fate remain to be answered. CONCLUSIONS Residues of organochlorine insecticides and some other pesticides already present in soils will take several years to disappear, even without further additions.Although applying materials such as activated carbon to contami- nated soils lessens their effects, such treatments are unlikely to be practicable on a large scale and there is little that can be done to accelerate the removal of persistent pesticides. Where possible, changing to non-persistent pesticides such as the organophosphorus and carbamate insecticides, together with more careful use of persistent materials, seem to be the only ways to lessen residues. However organophosphorus and carbamate insecticides are generally much more toxic to mammals than organochlorine compounds and are therefore more dangerous to handle and cause greater short-term risks. Also persistence is a desirable feature for controlling some pests such as soil insects and the degree to which replacements for the organochlorine insecticides are success- ful is often largely a measure of the degree to which they persist. There seems little possibility that pesticides can be abandoned with the increasing world demand for high quality food, so that it is essential that the potential dangers are fully appreciated and taken into account in their use. Because residue chemists exist to find residues, it is no surprise that ana- lytical methods are becoming more and more sensitive. A needle weighing one gram in a haystack weighing 20 tons is at a concentration of 0.05 ppm and detection at these levels would be an easy analytical feat using modern methods. One cannot be complacent about pollution arising from agricultural sources, but certainly in Britain it is relatively minor in comparison with that from people and industry. It is simple to ban the use of a pesticide; cutting sewage output requires more drastic action. Graham-Bryce and Briggs 103 BIBLIOGRAPHY G. W. Cooke and R. J. Williams, ‘Losses of nitrogen and phosphorus’ rom agricultural land.’ Symposium on eutrophication, Society of Water Treatment and Examination, 1970. C. A. Edwards, ‘Insecticide residues in soil,’ Residue Rev., 1966, 13, 83. C. A. Edwards, ‘Persistent pesticides in the environment.’ Chemical Rubber Co. Critical Reviews in Environmental Control, February 1970. C. G. L. Furtnidge and J. M. Osgerby, ‘Persistence of herbicides in soil,’ J. Sci. Fd Agric., 1967, 18, 269. C. A. I. Goring, ‘Physical aspects of soil in relation to the action of soil fungicides,’ Ann. Rev. Phytopath., 1967, 5, 285. H. Martin (ed.), Insecticides and.fungicides handbook, 3rd edn. Oxford : Blackwell Scientific, 1969. L. J. Andus (ed.), The physiology and biochemistry of’ herbicides. London and New York: Academic, 1964. J. D. Fryer and S. A. Evans (eds), Weed control handbook, 5th edn, vol. 1. Oxford: Blackwell Scientific, 1968. R.I.C. Reviews 1 04