biological control of soil-borne diseases: some present problems and different approaches

Review

Biological control of soil-borne diseases: some present problems and different approaches Richard Campbell School of Biological Sciences, University of Bristol, Bristol BS8 1 UG, UK

Despite much recent interest in the biological control of soil-borne plant diseases it is difficult to find examples of commercial exploitation. This may be because inappropriate laboratory screening systems are used: screens must select for organisms adapted to the environment in which they will operate. Assessment of screening procedures is difficult on published information, but it seems that insufficient organisms are screened and the screens may sometimes not be as well targeted as might be hoped. An alternative approach, which has no commercial basis, is to develop traditional rotations, green manuring and organic amendment systems so that the microbiology is understood and can be manipulated, with or without added inocula, to achieve biocontrol. Development of added inocula is scientifically interesting and commercially fundable, but seems to have had only limited success at present. Traditional approaches are not ‘glamorous’ science, are not commercially viable because there is no product, and have received little funding, but are known to be effective in many instances although the microbiological reasons for this may not be understood.

Keywords: biological control; commercial inocula; screening systems; rotations; organic amendments; green manuring

There has recently been an increase in interest in biological control, because the public and many scientists believe that there are problems with the present levels and nature of pesticide use and, rightly or wrongly, perceive biocontrol as a safer method for the control of plant pests and diseases. It is difficult to find a reason a priori why the dispersal into the environment of a micro-organism (the biocontrol agent) producing a toxin (e.g. an antibiotic) should be intrinsically safer than using the toxin itself, provided that both are adequately tested for safety. One could make a case that the toxin delivered in a known amount at a given time with proven biodegradability could be safer than the micro-organism, which may or may not produce the toxin in unknown amounts at some time in the future depending on how well it survives or grows in the environment. A better case for improved safety could be made for those organisms operating by specific lysis or by siderophores. However, if the perception of greater safety exists, and the legislation on registration of biocontrol agents supports this perception by being less stringent than that for chemical pesticides (Lethbridge, 1989), then biocontrol agents will be researched and produced to meet the market.

The increased interest in biocontrol has been reflected in many reviews and books that have been published since Baker and Cook (1974) and Cook and Baker (1983) became the standard reference works. Some of the more recent reviews and books include Chet (1987), Campbell (1989), Baker and Dunn (1990), Hornby (1990), Baker (1991), Beemster et al. (1991),

Lewis and Papavizas (1991), Sivan and Chet (1992), and Chet (1993). Baker (1987) looked at research papers in this field and showed an exponential increase in the number of papers on biocontrol published since 1910, the main rise coming after 1960. The number of papers is rising faster than the number of plant pathologists, where the increase is more or less linear, indicating that an increasing number of research pathologists were turning to work in biocontrol. Even this is an underestimate, as the published papers are (it is hoped) matters of scientific interest, but any control that is really useful is likely to be sponsored by commercial concerns and published, if at all, only when development has been completed and patents obtained. The published papers are not, therefore, likely to report the work on the best organisms.

Lewis and Papavizas (1989) pointed out that most experimental biocontrol has been done by the application of selected micro-organisms to affected crops and they discussed the technology of fermentation, formulation and delivery and the hopes for the future development of genetic engineering in biocontrol. They, and others (Baker, 1987; Jutsum, 1988; Campbell, 1989, 1990; Powell and Faull, 1989) also consider the problems associated with biocontrol, especially the lack of consistent control from year to year or from crop to crop. This has many causes, including lack of survival in the environment, effects of environmental and edaphic factors on the live organisms (as opposed to the relatively inert chemical used for control), and interactions with other micro-organisms. Part of this lack of

0261-2194/94/01/0004-13 @ 1994 Butterworth-Heinemann Ltd

4 Crop Protection 1994 Volume 13 Number 1

Biological control of soil-borne diseases: R. Campbell

consistency results from the method(s) of selection of the biocontrol micro-organisms, and in their spectrum of activity in relation to the races of the pathogen; this is discussed further below. The second main theme of the present review is the use of cropping practices and other means to alter the resident microbial populations to favour antagonists and to inhibit the pathogen(s). This has been considered by various workers (Cook and Baker, 1983; Lumsden, Lewis and Papavizas, 1983; Baker, 1987; Cook, 1988; Hornby, 1990; Rovira, Elliott and Cook, 1990) and is often based on traditional agricultural techniques. It includes such things as the addition of green manures and other organic supple- ments and the manipulation of cropping systems to mix species or varieties in time (rotations) or space (mixed species cropping, polycropping, etc.; Kass, 1978; Francis, 1986). One of the problems with these latter systems of control is that there is usually no patentable or marketable product, and therefore no commercial interest and few research funds.

This review does not include the modes of action of antagonists, which have been discussed exhaustively by various authors (Campbell, 1989; Hornby, 1990; Andrews, 1992; Sivan and Chet, 1992; Whipps, 1992), or the considerations of antagonist inoculum and its formulation, commercial production and delivery (Lewis and Papavizas, 1989, 1991; Powell and Faul, 1989; Hornby, 1990).

Screening for biocontrol agents

General considerations

Screening for potentially useful crop protection agents, both chemical and biological, has been going on for many years. There are very well developed screening systems for chemical agents, which have been worked on by the agrochemical companies for the past 60 years at least. In recent years the number of chemical compounds screened in relation to useful products has increased enormously (Lethbridge, 1989), from about 2000 to 20 000 per useful product; this partly reflects the more stringent testing and registration procedures that eliminate potentially useful substances, which might previously have reached the market. It may also be that many of the useful chemical groupings have been found. Whatever the reason, it is now more difficult (although not impossible) to find novel chemicals that are both effective and safe, and this has increased research and development costs. In contrast, the research and development costs, including the screening, are usually considered to be less for biological agents (Andrews, 1992).

Screening for potentially useful micro-organisms has been done by the drug companies for many years, looking for antibiotic production and other characteristics useful to the pharmaceutical and fine chemical industries (Steele and Stowers, 1991), but in these cases the main object of the search is to produce organisms for fermentations, whereas with biocontrol agents the

quest is for organisms to put back into the field. This leads (or should lead) to different search strategies which, in the case of the pharmaceutically important organisms, will be based on cultural characteristics and selection of laboratory strains,- but for biocontrol the organisms need to be based on controlling the disease on the living plant, preferably under field environmental conditions. This is not necessarily ‘a bad thing’ because it gives another facet of the organism which can be selected for in the screen, and means that the screens can be carefully directed to this end (Cook, 1990). Weller and Cook (1983) showed that antibiotic- producing fluorescent pseudomonads from take-all suppressive soils, which grew on wheat roots, were involved in the control of take-all disease (caused by Gaeumannomyces graminis var. tritici). They then used this information to limit their search for potential biocontrol agents, and developed a rapid screening test based on these principles (Weller, Zhang and Cook, 1985). There may be other criteria that are considered important and which limit the screen: for example, if the aim of the screen is to produce industrially useful organisms, then considerations of later fermentation may be included so that the organism grows quickly on easily available, cheap nutrients such as might be used in a fermenter (Campbell, 1986). It is pointless to spend time and money on developing a good control agent in the specialist laboratory that cannot economically be grown under industrial conditions.

Size of screen

This targeting of the screens implies that it may be possible to screen fewer organisms per useful product than is the case for chemicals and this could again reduce the costs (Lethbridge, 1989). Nevertheless, the screen should still be of a reasonable size: in the worst cases, biocontrol agents may be selected from chance laboratory contaminants that produce impressive inhibition zones on Petri plates; however, it is also true that chance observations on chemicals may lead to their inclusion in screens, together with those selected in a more systematic manner. There is little information in the literature on the size of screens used in biological control studies: usually, it seems rather small, such as the initial studies in Weller and Cook’s system which looked at a few hundred isolates (Weller and Cook, 1981). This is also true of the study by Hebbar, Davey and Dart (1992) looking for the control of Fusarium on maize: they tested a total of 502 isolates from four soils. Campbell, Renwick and Coe (1986) and Renwick, Campbell and Coe (1991) looked at about 2000 isolates to find antagonists to take-all and other diseases. If the discovery rate is anything like that for chemicals, even allowing for the directed search in biological control, the screen should certainly look at thousands of organisms to have a reasonable chance of success. Perhaps this is one of the reasons for the lack of success of many biological control agents: the research and

Crop Protection 1994 Volume 13 Number 1 5


development are less costly because fewer organisms are screened than chemicals and the chances of finding really good organisms are thereby diminished. Even after all the selection and laboratory work, Powell and Faull (1989) estimate that only 5% of biocontrol agents (of all types) actually work in the field trials.

Choice of host, pathogen and potential control agent

The theory of screening biological agents has been looked at by many workers (Andrews, 1985, 1992; Weller et al., 1985; Campbell, 1986; Lumsden and Lewis, 1989; Merriman and Russell, 1990; Renwick et al., 1991; Campbell and Sands, 1992). There is general agreement that factors such as suppressive soils (see below) and genera containing known antagonists are a useful starting point in the absence of other information, allowing the targeted approach discussed above. How- ever, some care is necessary because retrospective analysis of data from screens suggests that the targeting may not always be effective. In the screen carried out by Renwick et al. (1991) for take-all control, sites with take-all decline were selected as a likely source of antagonists, to the extent that >70% of samples came from these sites; however, these yielded only 20% of the isolates finally selected on their ability to control disease on the live plant.

Similarly, selection of organisms by known modes of action (siderophore and antibiotic production by inhibition on plate tests, for example) can miss potentially useful organisms; when the selection was based on tests in viva regardless of the mode of action, more than one-half of the organisms did not show such inhibition of the pathogen in culture when subsequently tested, and would not therefore have been selected by tests in vitro (Renwick et al., 1991). There is general agreement that laboratory tests in general show poor agreement with the ability of the potential control agent to operate in the presence of the host plant and disease under field conditions (Deacon, 1991; Andrews, 1992). Despite this, the plate tests for inhibition are often the first part of the screen, simply because of the ease with which they can be done and the impression (albeit false) that they are selecting useful organisms; the organism so obviously inhibits the pathogen on the agar plate that it is hard to believe that it will not continue to do so in the field. It has, however, been shown many times that the degree of inhibition, and even the presence of any inhibition, varies with the medium used (Hebbar et al., 1992; Renwick et al., 1991), let alone environmental conditions outside the laboratory. Andrews (1992) presents a compromise scheme that utilizes tests both in vitro and in vivo, but then only seriously considers those organisms passing the plant tests for continuing study as field agents, although organisms passing the plate tests may contain useful characteristics (useful genes), such as antibiotic production, for later study. The main problem with plant tests in vivo is that they

are more time consuming, but at least the time is more likely to be usefully spent, and even these tests can be made speedy and reliable (Weller et al., 1985). It is important that the screens are conducted at realistic temperatures, soil moisture levels, etc., that would be found in the field and that organisms are selected that will operate in such conditions (Deacon, 1991; Powell, 1992).

Screens that do use the host plant and disease, generally use only one host and one strain of the pathogen, at least in the initial tests. It seems that some biocontrol agents may not be very specific in their action. In the screen for take-all control of wheat (Renwick et al., 1991) it was thought that the organisms should come mostly from wheat roots, for colonization of the host is, rightly, considered important in biocontrol (Weller, 1988) and there is good evidence of species and even cultivar specificity among plant growth- promoting rhizobacteria (Astrom and Gerhardson, 1988). However, of the potentially useful organisms tested for the control of take-all on wheat, nearly one- third came from the roots of non-cereals, one-third from barley and only one-third actually came from wheat; in subsequent tests most of the isolates colonized wheat roots (Renwick et al., 1991). One of the best organisms from this screen is being studied further for possible use against Pythium, Rhizoctonia and Fusarium on a variety of plants, and is not, therefore, limited to wheat take-all control, although that was the original target and screening pathogen. This type of result is true of other biocontrol agents: Trichoderma is one of the few commercially available biocontrol agents, and various species and strains work on many different pathogens on an assortment of hosts, from rots of woody plants to damping-off of herbaceous seedlings (Papavizas, 1985). Biocontrol agents that are host and/ or pathogen specific are advantageous from the registration and ecological points of view: they are unlikely to kill anything but the pathogen they are designed for. However, they may not be commercially viable; what is needed in the market place is a broad spectrum of control so that the agent may be used on many diseases and many hosts under different environmental conditions, giving it a large and financially viable market (Powell, 1992). These are two irreconcilable alternative aims of a screening system.

Notwithstanding this broad spectrum of action, there are, of course, many examples of pathogen species or even pathogen race specificity known, the control agent affecting only one race of the pathogen or affecting one race more than another. This has been shown for take- all and for Rhizoctonia (Campbell, 1990) and is to be expected in those biocontrol agents that work by producing antibiotics and other toxins; resistance to antibiotics (and fungicides) is known to develop. There is the well-known case of the bacteriocin, agrocin 84, produced by Agrobacterium strain K84, which is effective against most, but not all, nopaline-utilizing strains of the pathogen Agrobacterium tumefaciens, because it uses the same membrane transport system to



enter the target pathogen (Thomson, 1987). Resistance to agrocin 84 is known to occur.

This information suggests that screens of any sort should use several strains of the pathogen, and of the host if they are whole-plant screens, which would further complicate and lengthen the process. In practice it is likely that the initial screens will be done against one race of the pathogen and then the specificity of potentially useful isolates will be checked later, but some organisms could be lost if they did not react favourably to the initial race.

Another related possibility is that several strains, species or genera of biocontrol agent could be used as a mixed application in the hope that if one does not work the other will. The effect of adding back a complete soil mix is often better than that of any individual organism (Alabouvette, Rouxel and Louvet, 1979) and effective combinations of specific antagonists are known (Baker, 1991). The mix may be designed so that one strain works in dry conditions and another in waterlogged soils, or in different soil types, or against different races of the pathogen, etc. The first problem is that mixtures, although they can be effective (Weller and Cook, 1983; Baker, 1991), are not necessarily the sum of all the individual components. Antagonists can antagonize each other, or compete for limited nutrients or space. Added to this is the problem of testing combinations: with more than two or three strains, the number of possible combinations of strains, formulation products, etc. becomes impracticable and there are no realistic screening procedures to handle effects of multiple strains. Strains selected as single antagonists are usually combined later, in the hope of better results. Some workers advocate general changes in soil microflora by changing conditions in the soil (Baker, 1987), rather than expecting to find and maintain one all-powerful superorganism. Such changes in agricultural practice are discussed below.

In considering the small amount of information available it is difficult to draw general conclusions about screening procedures. No doubt, commercial companies and other laboratories working on the problem have developed their own solutions, but until more of these are analysed and published no generaliza- tions can be made; the screen developed for each project is probably unique, being some compromise between the reduced number of organisms that can be screened with the best whole-plant/pathogen techniques and the doubtful validity of many assays in vitro. In view of the lack of knowledge of which are the ‘best’ antagonists and how they work, it seems advisable to keep the initial screens as general as possible to isolate many different types of organism, from different environments apart from the obvious ones, and with different modes of action against the pathogen. This will lose the advantage of some targeted screening which in theory should produce the most efficient yield of useful organisms, but the quality of the information used to choose the targets seems not to be very good at present.

Biocontrol by the manipulation of existing soil populations

Definition

The foregoing discussion has considered some commercial aspects of the search for biocontrol agents, but there are many forms of biocontrol that do not involve the direct application of commercially produced inocula but rely on the manipulation of the existing soil populations. In their definition of biological control, Cook and Baker (1983) included the use of ‘. . . cultural practices . . that create an environment favourable to antagonists, the host plant resistance or both . . .‘. This includes such things as the use of rotations to reduce the inoculum potential of pathogens, the application of manures and other organic amendments to the soil and the ploughing-in of green manure crops. There is also the possibility of using intercropping (Francis, 1986) or undersowing crops with other plants that may later be used for green manures or for other purposes. One of the problems with these methods of control is that there is no specific organism(s) involved that may be commercially produced and sold. Another factor is that many of the methods and techniques are traditional, so there is no possibility of patenting a novel development. For both these reasons there is no commercial interest and little commercial money available for research, no matter how effective such systems might be, and yet the research is needed to understand and possibly to improve or adapt the systems to modern agricultural conditions. In some cases, convincing scientific evidence is needed before a method, based on tradition or anecdote, can be advocated for general use. Unfortunately, the present climate (in the UK at least) for research to be ‘relevant’, and to have clear commercial links and the prospect of commercial exploitation, has militated against continued research in this area.

Rotations

Rotations have been used for many years to reduce the effect of pathogens (Garrett, 1944). One of the aims is to deprive the pathogen of its host, so that it has to survive for long periods in the soil during which it might die of starvation or be lysed by natural soil organisms. There are exceptions to this, where pathogens can switch hosts or can survive saprophytically (Coley- Smith, 1979). There are, of course, plant nutritional and other agricultural reasons for rotations.

Crops in short rotations, or at the extreme in monoculture, sometimes show reduced yields that are not caused by known nutrient deficiencies, although the effect can be masked or alleviated to some extent by the use of additional fertilizer (Kupers, 1979). Such crops are without obvious disease, but they just do not grow well. There are also replant problems on a longer time-scale with trees and other perennial crops. It has been shown that in apple, for example, there is a replant problem that is correlated with an increase in



deleterious bacteria and other micro-organisms in the soil (Catska et al., 1982). There is also evidence of both growth-promoting and growth-inhibiting Pseudomonas strains around the roots of citrus (Gardner, Chandler and Feldman, 1984), another crop with a replant problem. The most completely understood example of this is the reduction in yield in crops of potatoes grown in the Dutch polder soils in short rotations, which has been investigated by Schippers and co-workers (Schippers, Bakker and Bakker, 1987). These workers developed bacterial strains that restored the yield reduction by competing with the deleterious bacteria producing cyanide, which inhibited potato growth (Geels and Schippers, 1983). The competition was for available iron, which is necessary for cyanide production by the deleterious bacteria; the beneficial pseudomonads use siderophores with a higher affinity constant for iron. Extensive work has been done by genetic manipulation of the ability to produce the siderophores, which showed that this was the most likely mechanism of action. Siderophores are also implicated in many other plant diseases and in biological methods of control (Swinburne, 1986).

Yield reduction in wheat monoculture can be restored by the use of fumigation (Rovira, 1990), but the mechanism of this action is not known. It may be a simple case of pathogen control, for Cook, Sitton and Haglund (1987) have shown that one of the problems is the increase in Pythium, which is eliminated by the fumigation. However, the effects of fumigation may be by way of a biological mechanism, for it has long been known that Trichoderma especially is resistant to some soil fumigants and increases dramatically after fumigation. It is the Trichoderma that is then responsible for the control of the disease that the fumigation was designed to treat (Bliss, 1951; Munnecke et al., 1981; Strashnor et al., 1985). There are many other examples of these effects, where fumigants and other pesticides change the soil microflora to the disadvantage of the pathogen, in various programmes of integrated control of diseases where both chemical and biological means are used (reviewed by Papavizas and Lewis, 1988).

There is, therefore, scope for the investigation of microbiology of other monoculture, short rotation and replant problems, as they could be controlled either by cultural methods that alter the microbiology, by fumigation to alter the soil microflora, or by the direct introduction of bacteria or fungi for the control of deleterious organisms after creating an available niche.

Naturally occurring suppressive soils

The effects of rotations also impinge upon the phenomenon of suppressive soils (Schneider, 1982; Rouxel, 1991). There are some soils in which the lack of disease has a physical basis, such as the clay content (Toussoun, 1975), and here the pathogen is usually prevented from establishing in the soil; this is not linked to biological control and is not further discussed here. There are, however, situations where plants grown in some soils

do not develop disease symptoms even though they are susceptible to the pathogen, and virulent races of the pathogen are known to be present or are introduced; micro-organisms are generally associated with the lack of symptom development in such soils. This may be linked with some abiotic factors, such as the levels of organic and inorganic nutrients or the pH, but there are microbiological components as well, through which the abiotic factors are expressed. For example, Lumsden et al. (1987) reported on the suppressive soils of traditional agriculture of the Mexican chinampa system, which involves the addition of high levels of organic material as manures and crop wastes and also mineral nutrients from aquatic sediments; such soils have very little Pythium damping-off, and furthermore are suppressive to introduced Pythium. This is associated with high levels of microbial activity, especially of fluorescent pseudomonads and saprophytic Fusarium spp. Kao and Ko (1986a) showed that soils suppressive to damping- off caused by Pythium splendens were associated with high calcium levels and high numbers of soil micro- organisms. They also showed (Kao and Ko, 1986b) that the addition of calcium and alfalfa meal (to increase microbial numbers) reduced damping-off in conducive soils. This type of suppressiveness may be part of the more general phenomenon of the addition of organic amendments, which is discussed hereafter.

Other reports of suppressive soils, such as in the Chateaurenard region of France, are due to competition between the natural saprophytic Fusarium species and the pathogen Fusarium oxysporum f.sp. melonis (Alabouvette, Couteaudier and Lemanqeau, 1986), although in this case there is no obvious interaction with the addition of organic material as in the Mexican systems described above. Suppressiveness is lost when the soil is heated, and suppressiveness can be transferred between soils by small soil inocula: both of these are characteristic of suppression caused by micro-organisms. In other regions the suppressiveness for F. oxysporum f.sp. melonis has been induced in a naturally conducive soil by repeated crops of melons; again, this involved the development of populations of saprophytic Fusarium strains that competed with the pathogen (Sneh, Pozniak and Salomon, 1987).

Suppressiveness may develop in the soil during the course of particular crop sequences or during prolonged monoculture where again there is repeat cropping with one host species. The classicexample of this phenomenon is take-all decline, which has been studied for many decades and has been written about by numerous authors, but is still not fully understood (Gerlagh, 1968; Hornby, 1979; Rouxel, 1991). It is, however, likely that the reduced incidence of the disease after repeated crops of wheat is caused in part by a change in the microflora, in the soil and on the wheat roots, to one which is antagonistic or at least unfavourable to the take-all fungus, Gaeumannomyces graminis var. tritici (Hornby, 1979; Simon and Sivasithamparam, 1990). Soil from fields that have grown wheat continuously can be suppressive even to introduced inoculum of the



fungus, apart from the reduced inoculum in the natural soil (Cook, 1981). However, the situation is more complicated than this might imply: some break crops in the monoculture (alfalfa, oats or potatoes) reduced the inoculum of take-all, as expected from the rotation effect, but resulted in the soil losing its antagonism to the take-all fungus, i.e. it had become conducive (Cook, 1981). A break crop of beans allowed the take- all inoculum to persist and, in addition, destroyed the suppressiveness of the soil and so resulted in serious disease. A break crop of maize has also been shown to destroy suppression of the take-all and to allow the survival of the inoculum (Lucas et al., 1989). However, where wheat was maintained in monoculture, or there was a grass break crop, there was some remaining inoculum of take-all, but the soil remained suppressive to that inoculum and to introduced inoculum of Gaeumannomyces (Cook, 1981). This effect of grass in maintaining or creating suppressiveness to take-all is known from other situations: a grass ley encourages the growth of G. graminis var. graminis and of Phialophora species on the grass roots and these compete with G. graminis var. tritici on the following wheat crop (Deacon, 1973).

Green manures

The grass ley is, of course, ploughed in before sowing the wheat and this can give rise to other effects. Ploughing-in a green crop (i.e. using it as a green manure) encourages the general microbial activity in the soil as the organic material is decomposed, and this releases nutrients into the soil. There are, therefore, at least two effects of the addition of organic matter - the nutrient effect on the pathogen growth and disease expression and the increase in either general or particular microbial activity.

An example of where these combinations of factors have been studied in some detail is again given by take- all, which is arguably the disease most studied in intensive agriculture for control by biological and cultural means (in desperation, one suspects, as all other control measures are ineffective). There have long been systems for growing wheat, and especially barley, undersown with legumes and grasses (Garrett, 1944). Early work on the Chamberlain system had shown that take-all was not serious in barley when it was undersown with a mixture of rye grass (Lolium perenne) and medic (also called trefoil, Medicago lupulina) which was subsequently ploughed-in as a green manure (Yarham, 1979). At the time, the success of the system was thought to be due to the locking-up of nitrogen in winter by the undersown crop and the later release of this, after decomposition, to the new growing crop (Garrett and Buddin, 1947). We could add to this, in the light of more recent knowledge, that the rye grass may have contributed by encouraging Phialophora (see above) and the increased microbial activity during decomposition could have encouraged microbial antagonism to the take-all fungus. A similar system,

involving the undersowing of wheat with Medicago, has been investigated further with regard to the microbiology. There is no doubt that take-all was reduced by nearly half, and there seem to be two separate stages in which control is obtained - during the growth of the mixture of the crop and the undersown plants (a mixed cropping effect) and then the green manure effect (Lennartsson, 1987, 1988). These two components of the system can be separated so that wheat is first grown in mixed cropping, and then the effects of green manuring with the Medicago are studied in separate experiments. The mixed cropping significantly reduced take-all but had little effect on the ‘total’ bacterial numbers in the soil (measured by dilution plates). It is possible that the microflora was changed in composition, even though in total the numbers were the same, but this was not detected. Such changes in microflora, by one plant influencing another growing next to it, are known to occur in other situations (Keswani, Kibani and Chowdhury, 1977; Newman et al., 1979). The separate green manuring with the Medicago halved the take-all infection and significantly increased the ‘total’ numbers of bacteria, particularly the fluorescent pseudomonads (Lennartsson, 1988); thus, both the mixed cropping and the green manuring decrease the take-all. The green manuring effect is complicated, in that there is a nutritional factor, especially an increase in available nitrogen, which reduces take-all anyway; this effect was also shown by Garrett and Mann (1948) for barley. However, if this nutritional effect was eliminated by the addition of fertilizer to equalize the available nitrogen, the control still occurred, and the reduction in take-all was greater than would be expected from the amount of nitrogen added in the green manure (Lennartsson, 1988). Thus, with green manuring the nutritional effect occurs, but there is something else in addition, which is thought to be the effect of the increased and changed microflora. The green manuring is therefore biological control because it acts, in part at least, through changes in the soil microflora, especially the fluorescent pseudomonads. Other groups of micro-organisms may also be involved: actinomycetes antagonistic to take-all increase in response to various green manure treatments that also control take-all (Ehle, 1966), and Phymatotrichum omnivorum is controlled by the increase in Trichoderma when green manures are ploughed in (Streets and Bless, 1973).

A word of caution should be added at this point, in that some crop residues and green manuring systems can make disease worse, apparently by encouraging the growth of the pathogen. Thus the use of Sesbania (a tropical woody legume) green manure exacerbated damping-off diseases caused by Pythium and Rhizoctonia, despite increasing the numbers of pseudomonads (R. Campbell and M. Begum, unpublished). However, Sesbania green manure has also been shown to reduce disease (Phymatotrichum; Streets and Bless, 1973); thus, it is not Sesbania itself that is the problem, but the interaction between it and the particular fungi. Other work with damping-off diseases, using Medicago



as a green manure, suggests that the time between the incorporation of the green manure and the sowing of the crop can be important and that the initial increased levels of disease are reduced as the decomposition proceeds and antagonist populations increase (R. Campbell, unpublished). Other deleterious effects of crop residues have long been known (Cook, Boosalis and Doupnik, 1978; Lumsden et al., 1983; Baker, 1991); these include the increase in pathogens when infected residues are incorporated, the ability of the pathogen to colonize and grow on the fresh residue, and the production of phytotoxic compounds during the decomposition.

Organic amendments

The general addition of organic material, as well as crop residues, can reduce disease. Traditionally it has been known that adding farmyard manure in sufficient quantities can reduce disease (Heitefuss, 1989). Thielaviopsis basicola causes root rots of many plants and can be controlled by the addition of organic matter; the nutrients in the organic amendments stimulate the pathogen spores to germinate and they then lyse, though whether this is because of antagonistic micro- organisms has not been determined (Papavizas, 1968). Similarly, Fusarium oxysporum f.sp. elaeidis has been stimulated to germinate by farmyard manure and then subsequently lysed (Oritsejafor and Adeniji, 1990). Lumsden et al. (1983), Baker and Cook (1974) and Cook et al. (1978) cite many other similar examples, but usually the microbiological information is lacking.

If specific organic compounds are used, then the mode of action may be more obvious. Thus, the addition of chitin to soil increases the general soil population and especially the numbers of actinomycetes, which are well known for the breakdown of such recalcitrant material; correlated with the rise in the actinomycetes is a fall in the inoculum of Fusarium oxysporum f.sp. pisi and a reduction in pea wilt (Khalifa, 1965). Chitin addition, because it stimulates chitin-degrading micro-organisms, may reduce the numbers of nematodes and other organisms that contain significant quantities of chitin in their walls and exoskeletons.

Another aspect of this same phenomenon is the recent interest in the use of various novel growing media for plants. The use of peat lands for forestry and agriculture, and to a lesser extent the extraction of peat for these products, has serious effects on wetlands and other ecologically sensitive areas. Many different factors have therefore combined to enforce a reappraisal of peat-based growing media, soil conditioners and similar products. There are now more restrictions on the disposal of sewage sludge by dumping at sea or on land; on the other hand there is an excess of various materials with a high C:N ratio, such as straw from cereals, and this has been exacerbated by the restrictions on, or prohibition of, the burning of straw and stubble. There are also waste products of forestry and other industries,

such as bark and coconut fibre, that may be used as substrates for horticulture. Domestic waste is now becoming expensive to dispose of by landfill or burning. There are, therefore, many attempts to combine sewage sludge or animal slurry with waste plant products and refuse in various composting systems to produce peat-free growing media and soil conditioners. These have been found to vary greatly in their ability to suppress disease. The relevant information has been summarized by Whipps (1992) and Hoitink and Fahy (1986). The latter authors also discuss similar suppressive properties of some types of peat, such as Finnish light sphagnum peats. Some of the suppressive composts and peats are thought to reduce disease because they encourage the growth of Tricho- derma and other antagonists to pathogens such as those causing damping-off and, indeed, Trichoderma or bacteria may be added to the substrates, which have a greatly modified microflora after heating either by the composting process (Kwok et al., 1987) or by artificial steam treatment to remove the pathogens. In many cases, however, although the composts are suppressive, the microbiological information is not available.

Conclusions

The main question to consider is whether biocontrol will ever become a realistic alternative to chemical control of plant disease. Taken in its broadest sense, the answer is affirmative, because biological control by rotations, etc., was the only way of disease control before chemical control was developed. However, the use of chemicals, as one component of intensive Western agriculture, has allowed the massive increase in food production since the Second World War, although unfortunately this has been paralleled or even exceeded by the rise in world population (Powell, 1992). We now therefore have a situation in which we are dependent on the high yields possible with modern crop varieties grown with chemical inputs of many sorts, and there is general agreement that the same yields cannot be obtained by present crop varieties without the use of chemical fertilizers and pesticides. This may change if varieties better suited to low- nutrient inputs are used, but this is a possibility for the future, not an immediate prospect. There is, therefore, no hope of a complete or rapid switch to biological methods of disease control (assuming that they existed), or organic farming systems in general, without un- acceptable results for the food supply of the world. We need, therefore, to regard biological control as one of several possible ways of limiting disease, and to accept that for the foreseeable future it will be used in competition, or at best integrated, with chemical methods of pest and disease control. We need to develop biocontrol strategies with this in mind, and screening systems should ensure that biological agents tolerate the common agrochemicals with which they will inevitably come into contact.



One of the bridges between biological and chemical control is that biological control agents as introduced inocula are being developed largely by the chemical companies, which are aiming to produce products using technology, application methods and marketing similar to those used for chemicals, although obviously with different formulations, etc. Biologicals at present and in the future will therefore be designed to be at least compatible with chemical control agents, and should not need to compete with them.

For the successful (i.e. commercially useful) production of biocontrol agents it is essential that realistic screens are used, with environmental parameters such as temperature, soil microbial populations, and pathogen and host plants present. Industry urgently needs some marketable products if it is to continue to support this approach to disease control. Until recently, much of the experimentation that was called biological control was more to do with the interactions of micro- organisms under laboratory conditions; although this was, perhaps, good science, it was not an attempt at realistic field control of disease. Biological control by the use of added inocula is one line of approach and it may well yet be successful, but it is proving to be more difficult and complicated than was initially supposed (Chet, 1993). The problems with introduced inocula, and the possibility of re-examining traditional cultural means of control, are demonstrated by the extensive study of take-all control. As is often the case, with hindsight, the initial hopes for artificial inocula look exceedingly optimistic, not to say naive.

The use of genetically engineered micro-organisms may eventually become important in biological control and at least one, Agrobacterium, is available now in some countries. There is the temporary problem of restrictions or difficulties with getting clearance for the release of such micro-organisms, but suitable protocols have been, or will be, developed and it is unlikely that this will be a serious impediment to commercial exploitation. Much more of a problem is that at present we do not know what determines survival or colonization by an organism in the agricultural environment, and in many cases we do not have a clear understanding of the mode of action. It is therefore not yet possible to ‘design’ an organism for biological control, because we often do not know which characteristics to add or remove. Even if more commercially useful genetically engineered organisms are developed in the near future, we still have the problems involved in screening to select the most suitable strains. Thus, extensive use of genetically engineered organisms is a prospect for the future, but much more information is needed on microbial ecology and on what characteristics make a good control agent before they become a reality.

Manipulating existing soil populations with green manures, etc., is another way of controlling disease biologically, or possibly of enhancing the effect of the added agents, and has environmental benefits compared with the possible problems of using genetically engineered control agents, for example. Some of the

approaches using large quantities of organic materials may be too costly or labour intensive for Western agriculture, but may be better suited to a low- technology subsistence agriculture. Rotations, manipulation of crop mixtures and green manuring have been known about for many years, and although the science behind them has been neglected in the shadow of the undoubted success of chemical control measures, they may merit further investigation as a means of reducing the use and environmental impact of pesticides. Some organic systems can be adapted for use in intensive agriculture and are amenable to mechaniza- tion. We need to understand much more about these systems, especially in terms of the microbiology of the soil and the decomposition of organic matter, before we can apply them to the best advantage. They are known to work in traditional agriculture but have been sidelined by the dramatic developments in chemical control and by the search for commercially profitable inocula of biocontrol agents.

References

Alahouvette, C., Rouxel, F. and Louvet, J. (1979) Characteristics of Fusarium wilt-suppressive soils and prospects for their utilization for biocontrol. In: Soil-borne Plant Pathogens (Ed. by B. Schippers and W. Cams) pp. 165-182, Academic Press, London

Alahouvette, C., Couteaudier, Y. and Lemanqeau, P. (1986) Nature of the intrageneric competition between pathogenic and non- pathogenic Fusarium in wilt-suppressive soil. In: Iron, Siderophores and Plant Disease, NATO ASI Series A, Life Sciences (Ed. by T. R. Swinburne) pp. 165-178, Plenum Press, New York

Andrews, J. H. (1985) Strategies for selecting antagonistic organisms from the phylloplane. In: Biological Control on the Phylloplane (Ed. by C. E. Windels and S. Lindow) pp. 3144, American Phytopatho- logical Society, St Paul, Minnesota

Andrews, J. H. (1992) Biological control in the phyllosphere. A. Rev. Phytopathol. 30, 60%635

AstrBm, B. and Gerhardson, B. (1988) Differential reactions of wheat and pea genotypes to inoculation with growth-affecting rhizosphere bacteria. Plant Soil 109, 263-269

Baker, K. F. (1987) Evolving concepts of biological control of plant pathogens. A. Rev. Phytopathol. 25,67-85

Baker, K. F. and Cook, R. J. (1974) Biological Control of Plant Pathogens, Freeman, San Francisco, California, 433 pp

Baker, R. (1991) Diversity in biological control. Crop Prot. 10, 85-94

Baker, R. and Dunn, P. E. (1990) New Directions in Biological Control: Alternatives for Suppressing Agricultural Pests and Diseases, Alan R. Liss, New York, 837 pp

Beemster, A. B. R., Bollen, G. J., Gerlagh, M., Ruissen, M. A., Schippers, B. and Tempel, A. (1991) Biotic Interactions and Soil Borne Disease. Developments in Agriculture and Managed Forest Ecology, Vol. 23, Elsevier, Amsterdam, 428 pp

Bliss, D. E. (1951) The destruction of Armillaria mellea in citrus soils. Phytopathology 41, 665-683

Campbell, C. L. and Sands, D. C. (1992) Testing the effects of microbial agents on plants. In: Microbial Ecology; Principles, Methods and Applications (Ed. by M. A. Levin, R. J. Seidler and M. Rogul) pp. 689-705, McGraw-Hill, New York

Campbell, R. (1986) The search for biological control agents against plant pathogens: a pragmatic approach. Biol. Agric. Hort. 3,317-327

Campbell, R. (1989) Biological Control of Microbial Plant Pathogens, Cambridge University Press, Cambridge, 218 pp



Campbell, R. (1990) Biological control of soil-borne diseases. Proc. 1990 Brighton Crop Prot. Conf. Pests Diseases 2, 607615

Hoitink, H. A. J. and Faby, P. C. (1986) Basis for the control of soil- borne plant pathogens with composts. A. Rev. Phytopnthol. 24, 93- 114

Campbell, R., Renwick, A. and Coe, S. (1986) Antagonism and siderophore production by biocontrol agents, plant growth promoting organisms and the general rhizosphere population. In: Iron, Sidero- phores and Plant Disease, NATO ASI Series A, Life Sciences (Ed. by T. R. Swinburne) pp. 179-189, Plenum Press, New York

Catska, V., Vancura, V., Hudska, G. and Prikryl, Z. (1982) Rhizosphere micro-organisms in relation to the apple replant problem. Plant Soil 69, 187-197

Hornby, D. (1979) Take-all decline: a theorist’s paradise. In: Soil- borne Plant Pathogens (Ed. by B. Schippers and W. Gams) pp. 133- 156, Academic Press, London

Hornby, D. (1990) Biological Control of Soil-borne Plant Pathogens, CAB International, Wallingford, Oxford, 479 pp

Jutsum, A. R. (1988) Commercial application of biological control: status and prospects. Phil. Trans. R. Sot. Ser. B 318, 357-373

Cbet, I. (1987) Innovative Approaches to Plant Disease Control, John Wiley and Sons, New York, 372 pp

Cbet, I., Ed. (1993) Biotechnology in Plant Disease Control, Wiley- Liss, New York, 373 pp

Kao, C. W. and Ko, W. H. (1986a) Suppression of Pythium splendens in a Hawaiian soil by calcium and micro-organisms. Phytopathology 76, 215-220

Coley-Smith, J. R. (1979) Survival of plant-pathogenic fungi in soil in the absence of host plants. In: Soil-borne Plant Pathogens (Ed. by B. Schippers and W. Gams) pp. 39-57, Academic Press, London

Kao, C. W. and Ko, W. H. (1986b) The role of calcium and micro- organisms in suppression of cucumber damping-off caused by Pythium splendens in a Hawaiian soil. Phytopathology 76, 221-225

Cook, R. J. (1981) The influence of rotation crops on the take-all decline phenomenon. Phytopathology 71, 189-192

Kass, D. C. L. (1978) Polyculture cropping systems: review and analysis. Cornell Int. Agric. Bull. 32, 169

Cook, R. J. (1988) Management of the environment for the control of pathogens. Phil. Trans. R. Sot. Land. Ser. B 318, 171-182

Keswani, C. L., Kibani, T. H. M. and Cbowdhury, M. S. (1977) Effect of intercropping on rhizosphere population in maize (Zea mays L.) and soybean (Glycine max Merrill). Agric. Environ. 3, 363- 368

Cook, R. J. (1990) Twenty five years of progress towards biological control. In: Biological Control of Soil-borne Plant Pathogens (Ed. by D. Hornby) pp. 1-14, CAB International, Wallingford, Oxford

Khalifa, 0. (1965) Biological control of Fusarium wilt of peas by organic soil amendments. Ann. Appl. Biol. 56, 129-137

Cook, R. J. and Baker, K. F. (1983) The Nature and Practice of Kupers, L. J. P. (1979) Agronomic aspects of soil-borne pathogens. Biological Control of Plant Pathogens, American Phytopathological In: Soil-borne Plant Pathogens (Ed. by B. Schippers and W. Gams)

Society, St Paul, Minnesota, 539 pp pp. 357-370, Academic Press, London

Cook, R. J., Bnosalis, M. G. and Doupnik, B. (1978) Influence of crop residues on plant disease. In: Crop Residue Management Systems (Ed. by W. R. Oschwald) pp. 147-163, American Society of Agriculture, Madison, Wisconsin

Kwok, 0. C. A., Fahy, P. C., Hoitink, H. A. J. and Kuter, G. A. (1987) Interaction between bacteria and Trichoderma hamatum in the suppression of Rhizoctonia damping-off in bark compost media. Phytopathology 77, 12061212

Cook, R. J., Sitton, J. W. and Haglund, W. A. (1987) Influence of soil treatments on growth and yield of wheat and implications for the control of Pythium root rot. Phytopathology 77, 1192-l 198

Lennartsson, M. (1987) Cultural Control of Take-all, PhD thesis, Department of Botany, University of Bristol, UK, 337 pp

Deacon, J. W. (1973) Phialophora radicicola and Gaeumannomyces graminis on the roots of grasses and cereals. Trans. Br. Mycol. Sot. 61, 471478

Deacon, J. W. (1991) Significance of ecology in the development of biocontrol agents against soil-borne plant pathogens. Riocontrol Sci. Technol. 1, 5-20

Lennartsson, M. (1988) Effects of organic soil amendments and mixed species cropping on take-all disease of wheat. In: Global Perspectives on Agroecology and Sustainable Agricultural Systems, Vol. 2 (Ed. by P. Allen and D. Van Drusen), pp. 575-580a, University of California, Santa Cruz, California

Ehle, H. (1966) EinfluB der Grlndtingung auf die Actinomyceten- population des Bodens unter besonderer Beriicksichtigung der gegen Ophiobolus graminis Sacc. Wirksamen Antagonisten. 2. PflKrankh. PflSchutz 73, 326-334

Lethbridge, G. (1989) An industrial view of microbial inoculants for crop plants. In: Microbial Inoculation of Crop Plants (Ed. by R. Campbell and R. M. Macdonald) pp. 1 l-28, IRL Press, Oxford

Francis, C. A. (1986) Multiple Cropping Systems, Macmillan, New York, 383 pp

Lewis, J. A. and Papavizas, G. C. (1989) Selection, production, formulation and commercial use of plant disease biocontrol fungi: problems and progress. In: Biotechnology of Fungi for Improving Plant Growth (Ed. by J. M. Whipps and R. D. Lumsden) pp. l71- 190, Cambridge University Press, Cambridge

Gardner, J. M., Chandler, J. L. and Feldman, A. W. (1984) Growth promotion and inhibition by antibiotic producing fluorescent pseudomonads on citrus roots. Plant Soil 77, 103-113

Lewis, J. A. and Papavizas, G. C. (1991) Biocontrol of plant diseases: the approach for tomorrow. Crop Prof. 10, 95-105

Garrett, S. D. (1944) Root Disease Fungi, Chronica Botanica Co., Waltham, Massachusetts, 177 pp

Garrett, S. D. and Buddin, W. (1947) Control of take-all under the Chamberlain system of intensive barley growing. Agriculture 54,425- 426

Lucas, P., Sarniguet, A., Collet, J. M. and Lucas, M. (1989) RCceptivitC des sols au pietin Cchaudage (Gaeumannomyces graminis var. tritici): influence de certaines techniques culturales. Soil Biol. Biochem. 21, 1073-1078

Garrett, S. D. and Mann, H. H. (1948) Soil conditions and the take- all disease of wheat: X. Control of the disease under continuous cultivation of spring-sown cereal. Ann. Appl. Biol. 35, 43-42

Geels, F. P. and Schippers, B. (1983) Reduction in yield depression in high frequency potato cropping soil after seed tuber treatment with antagonistic Pseudomonas species. Phytopath. Z. 108, 207-214

Gerlagh, M. (1968) The introduction of Ophiobolus graminis into new polders and its decline. Neth. J. Plant Pathol. 74, l-97

Lumsden, R. D. and Lewis, J. A. (1989) Selection, production, formulation and commercial use of plant disease biocontrol fungi: problems and progress. In: Biotechnology of Fungi for Improving Plant Growth (Ed. by J. M. Whipps and R. D. Lumsden) pp. 171- 190, Cambridge University Press, Cambridge

Lumsden, R. D., Lewis, J. A. and Papavizas, G. C. (1983) Effect of organic amendments on soilborne plant diseases and pathogen antagonists. In: Environmentally Sound Agriculture (Ed. by W. Lockeretz) pp. 51-69, Praeger Publishers, New York

Hebbar, K. P., Davey, A. G. and Dart, P. J. (1992) Rhizobacteria of maize antagonistic to Fusarium moniliforme, a ‘soil-borne fungal pathogen: isolation and identification. Soil Biol. Biochem. 24. 979- 987 -

Lumsden, R. D., Garcia-E., R., Lewis, J. A. and Frias-T., G. A. (1987) Suppression of damping-off caused by Pythium spp. in soil from the indigenous Mexican Chinampa agricultural system. Soil Riot. Biochem. 19, 501-508

Merriman, P. and Russell, K. (1990) Screening strategies for biological control. In: Biological ControlofSoil-borne Plant Pathogens (Ed. by D. Hornby) pp. 427-435, CAB International, Wallingford, Oxford

Heitefuss, R. (1989) Crop and Plant Protection, English edn, Ellis Horwood, Chichester, 261 pp



Munnecke, D. E., Kolhezen, M. J., Wilbur, M. J. and Ohr, H. D. (1981) Interactions involved in controlling Armillaria. Plant Dis. 65, 3X4-389

Newman, E. I., Campbell, R., Christie, P., Heap, A. J. and Lawley, R. A. (1979) Root micro-organisms in mixtures and monocultures of grassland plants. In: The Soil-Root Interface (Ed. by J. R. Harley and R. Scott Russell) pp. 161-173, Academic Press, London

Oritsejafor, J. J. and Adeniji, M. 0. (1990) Influence of host and non-host rhizosphere and organic amendments on survival of I:usarium oxysporum f.sp. elaeidis. Mycol. Res. 94, 57-63

Papavizas, G. C. (1968) Survival of root-infecting fungi in soil. VI. Effect of amendment on bean root rot caused by Thielaviopsis hasicola and inoculum density of the causal organism. Phytopathology 58, 421-428

Papavizas, G. C. (1985) Trichoderma and Gliocladium: biology, ecology and potential for biocontrol. A. Rev. Phytopathol. 23, 23-54

Papavizas, G. C. and Lewis, J. A. (1988) The use of fungi in integrated control of plant diseases. In: Fungi in Biological Control Systems (Ed. by M. N. Burge) pp. 235-253, Manchester University Press, Manchester

Powell, K. (1992) Is biological control the answer for sustainable agriculture? Chem. Ind. March 1992, 168-170

Powell, K. A. and Faull, J. L. (1989) Commercial approaches to the use of biological control agents. In: Biotechnology of Fungi for Improving Plant Growth (Ed. by J. M. Whipps and R. D. Lumsden) pp. 25s-275, Cambridge University Press, Cambridge

Renwick, A., Campbell, R. and Cue, S. (1991) Assessment of in vivo screening systems for potential biocontrol agents of Gaeumannomyces graminis. Plant Pathol. 40, 524-532

Rouxel, F. (1991) Natural suppressiveness of soils to plant diseases. In: Biotic Interactions and Soil-borne Diseases (Ed. by A. B. R. Beemstcr, G. L. Bollen, M. Gerlagh, M. A. Ruissen, B. Schippers and A. Tempel) pp. 287-296, Elsevier, Amsterdam

Rovira, A. D. (1990) The impact of soils and crop management systems on soil-borne diseases and wheat yields. Soil Use Mgmt 6, 195-200

Rovira, A. D., Elliott, L. F. and Cook, R. J. (1990) The impact of cropping systems on rhizosphere organisms affecting plant health. In: The Rhizosphere (Ed. by J. M. Lynch) pp 389436, John Wiley and Sons, Chichester

Schippers, B., Bakker, A. W. and Bakker, P. A. H. M. (1987) lnteractionsofdeleteriousand beneficial rhizosphere micro-organisms and the cffcct of cropping practices. A. Rev. Phytopathol. 25, 339- 3%

Schneider, R. W., Ed. (1982) Suppressive Soils, American Phyto- pathological Society, St Paul, Minnesota, 88 pp

Simon, A. and Sivasithamparam, K. (1990) Effect of crop rotation, nitrogenous fertilizer and lime on biological suppression of the take- all fungus. In: Biological Control of Soil-borne Plant Pathogens (Ed. by D. Hornby) pp. 215-226, CAB International, Wallingford, Oxford

Sivan, A. and Chet, I. (1992) Microbial control of plant diseases. In: Environmentul Microbiology (Ed. by R. Mitchell) pp. 335-354, Wiley-Liss, New York

Sneh, B., Pozniak, D. and Salomon, D. (1987) Soil supprcssiveness to Fusarium wilt of melon, induced by repeated croppings of resistant varieties of melon. J. Phytopathol. 120, 347-354

Steele, D. B. and Stowers, M. D. (1991) Techniques for the selection of industrially important micro-organisms. A. Rev. Microbial. 45,89- 106

Streets, R. B. and Bless, H. E. (1973) Phymatotrichum Root ROI, Monogr. 8, American Phytopathological Society, St Paul, Minnesota

Strashnor, Y., Elad, Y., Sivan, A. and Chet, 1. (1985) Integrated control of Rhizoctonia solani Kiihn by methyl bromide and Tricho- derma harzianum Rifai Agar. Plunt Pathol. 34, 146-151

Swinburne, T. R., Ed. (1986) Iron Siderophores and Plant Disease, NATO AS1 Series A, Life Sciences, Plenum Press, New York, 351 pp

Thomson, J. A. (1987) The use of agrocin-producing bacteria in the biological control of crown gall. In: Innovative Approaches to Plunt Disease Control (Ed. by 1. Chet) pp. 213-228. John Wiley and Sons, New York

Tousooun, T. A. (197.5) Fusarium-suppressive soils. In: Biology and Control of Soil-borne Plant Pathogens (Ed. by G. W. Bruehl) pp. 145-151, American Phytopathological Society. St Paul, Minnesota

Weller, D. M. (1988) Biological control of soilborne plant pathogens in the rhizosphere with bacteria. A. Rev. Phytopathol. 26. 379-407

Weller, D. M. and Cook, R. J. (1981) Pseudomonads from take-all conducive and suppressive soils. Phytopathology 71, 264 (abstr.)

Weller, D. M. and Cook, R. J. (1983) Suppression of take-all of wheat by seed treatments with fluorescent pseudomonads. Phyto- pathology 73, 463469

Weller, D. M., Zhang, B.-X. and Cook, R. J. (1985) Application of a rapid screening test for selection of bacteria suppressive of take-ail of wheat. Phytopathology 69, 710-713

Whipps, J. M. (1992) Status of biological disease control in horticulture. Riocontrol Sci. Technol. 2, 3-24

Yarham, D. J. (1979) The effect on soil-borne diseases of changes in crop and soil management. In: Soil-borne Plant Pathogens (Ed. by B. Schippers and W. Cams) pp. 371-383, Academic Press. London

Received 22 April 1993 Revised 7 May 1993 Accepted 7 May 1993


biological control of soil-borne diseases: some present problems and different approaches

Documents