D.L. Trudgill and M.S. Phillips
Yield losses are influenced by the pathogenicity of the species of nematode involved, by the nematode population density at planting, by the susceptibility and tolerance of the host and by a range of environmental factors. Because of this, available models only estimate yield losses as proportions of the nematode-free yield. Estimating threshold levels further involves various economic calculations. Consequently, predicting yield losses and calculating economic thresholds for most nematode/crop problems is not yet possible. What is needed is more field-based information on the relationship between nematode population densities and crop performance, and various approaches to obtaining such data are described. Measuring the population density, especially of Meloidogyne species, is a major problem which needs addressing.
Nematode population dynamics are also density dependent and are influenced by host growth, the reproductive potential of the species and by various environmental factors. Consequently, modelling nematode population dynamics is an equally impressive science. Again, good field data are required but the complicating effects of biological control agents, host susceptibility differences and environmental factors, and errors associated with measuring initial population densities, may mean it is practically impossible to predict reliably the multiplication rates of most nematodes, especially those with several generations per season.
The nematode, the host and the environment are the three interacting variables influencing the extent of yield loss in infested soils. An understanding of the mechanisms and principles involved in these interacting relationships is basic to being able to predict yield reductions from estimates of pre-planting nematode population densities (Pi).
When modelling the damage caused to plants by root-feeding nematodes certain basic principles apply. These are:
· Damage is proportional to the nematode population density.
· The degree of damage is influenced by environmental factors.
· The yield harvested is determined by the amount of light intercepted by the crop, by how efficiently the intercepted light is converted into dry matter, and finally by how that dry matter is partitioned into non-harvested and harvested yield. For some crops significant variations in moisture content will also affect final yield.
The above principles can be simply stated but are more complex in practice. Damage may be proportional to the nematode population density, but there are several qualifications of this statement. The relationship is usually curvilinear, increasing numbers of nematodes having proportionally diminishing effects. There is some evidence that at low densities the host plant can repair the damage and that growth may even be slightly stimulated. Seinhorst (1965) termed the population density (Pi) at which damage first became apparent as the tolerance limit (T).
Equally, at very high values of Pi, increasing numbers of nematodes may not further reduce dry matter productivity. Seinhorst termed this the minimum yield (m). There are various reasons why m may occur; there may be some growth before attack starts or after it finishes, and a significant biomass may be planted (e.g. potato tubers). However, m applies to total dry matter and because of effects on partitioning, the harvest value of m may be greater or less than that for total dry matter.
The third parameter in the Seinhorst equation is z, a constant slightly less than one. The equation is:
y=1 where Pi£ T
where y is the yield.
An important qualification is that y is expressed as a proportion of the nematode-free yield. Hence, according to Seinhorst, the greater the yield potential the greater the loss in tonnes per hectare for any value of Pi.
The Seinhorst equation is usually plotted with Pi on a logarithmic scale, producing a sigmoidal curve (Fig. 1). In practice T is usually small and the Pi value at which m is reached is so large that it is only the central part of the curve that is of practical use. Oostenbrink (1966) suggested that this approximated to a straight line. The equation for such a line is:
y = y(max) - slope constant × log Pi
Even the simplified Oostenbrink relationship is not very helpful. Yield is still expressed in proportional rather than real (tonnes per hectare) terms. Also, there is no way of applying the relationship without considerable experimentation to determine the slope of the regression.
The slope of the regression varies for several reasons. These include differences in pathogenicity (capacity to cause damage) between species, e.g. Meloidogyne spp. may be inherently more damaging than Tylenchus but we have no measure of their relative pathogenicities. Different plant species and varieties within species differ in their tolerance (capacity to withstand nematode damage). Also, there are large environmental influences on the damage suffered and particularly how that damage is translated into effects on final yield.
An important consideration, often overlooked, is the basis of measuring Pi. Usually it is given as numbers per gram of soil. A more appropriate measure is per unit volume of soil as this allows for bulk density differences. Numbers per gram of root is probably the most appropriate, but is difficult to measure because it is always changing. This latter aspect becomes important when trying to relate results from experiments where root densities are very different, e.g. pot and field trials.
A further problem is encountered when considering damage by nematodes that have two or more generations in the lifetime of a crop. Usually the Pi is measured at planting, but on a good host population of, for example, Meloidogyne spp., they can increase from below the value of T to a level in mid-season where they cause significant damage. Even so, it is a race between increasing Pi and increasing plant size that brings with it increasing tolerance (in Seinhorst terms, increasing m). In such situations suitability as a host (susceptibility) and tolerance can have a marked effect on the degree of damage.
In summary, both the Seinhorst and Oostenbrink equations are, without the addition of a substantial amount of additional information, purely descriptive and cannot be used to predict actual yield losses.
Mechanisms of damage and environmental effects on damage
Damage is proportional to the intensity of attack; this is often proportionally greater in sandy soils where nematodes can move more freely, than in heavier soils where movement is impeded. Adequate soil moisture is essential for free movement so attack is often limited as soils dry out later in the season. Temperature also influences the rate of nematode movement, but plant growth is usually equally affected.
FIGURE 1: The relationship between proportional yield loss and initial population density as modelled by Seinhorst (1965)
Primary damage to the attacked roots can be attributed to mechanical damage associated with feeding or invasion, to withdrawal of nutrients, and/or to more subtle physiological effects. Generally damage reduces the rate of root extension. This reduces the rate of uptake of nutrients and water and, if any become limiting (and they usually do, even for crops without nematode damage), top growth rates are reduced. This reduces the rate of increase in light interception and carbohydrate synthesis and hence the capacity of the plant to generate more roots to overcome the limitations imposed by nematode damage. Such appears to be the main mechanism of damage by potato-cyst nematodes (Globodera spp.) whose effect is further increased by reductions in root efficiency, revealed in a decrease in root: shoot ratio. Further damage is associated with withdrawal of nutrients by the developing females (resistant cultivars of potato are often less damaged than susceptible cultivars) and by secondary pathogens such as Verticillium dahliae. The central role of nutrient uptake is revealed, however, by the substantial ameliorating effect on damage of additional fertilizer.
With Meloidogyne spp., impaired water relations appear to contribute substantially to reduced rates of top growth. This is probably because the developing giant cell systems interfere with and disrupt the developing xylem. Clearly, with such damage, effects on growth and yield are likely to be greater where the plants are on the threshold of becoming moisture stressed. Other effects include reduced photosynthetic efficiency and these are reviewed in Trudgill (1992).
Effects on light interception and utilization
There is a good correlation in many crops between percent ground cover (i.e. the percentage of ground occupied by a plant or a crop, when viewed from above, that is covered by green leaves) and percent light interception. Most annual crops start as individual, separate plants and a reduction in growth rate is directly reflected in ground cover and hence light interception. As they grow the leaves of neighbouring plants merge to form a continuous canopy. Nematode damage that only delays the production of a continuous canopy, and hence 100 percent light interception, will have a smaller effect on final yield than damage which prevents the crop achieving such full cover. Premature crop death will also proportionally reduce yield.
Several environmental interactions have already been mentioned. Soil type clearly has an effect because it influences nematode movement as well as being a nutrient and water supply to the host. It can also influence nematode survival during periods of stress and will certainly influence the species composition of nematode communities. The effect of fertilizer practice and of water availability has also been mentioned, but these in turn will interact with host genotype and husbandry factors such as spacing and time of planting. Recent studies of potato-cyst nematodes illustrate some of the interactions and are briefly summarized below:
· The interaction between two potato cultivars of different tolerance and rates of compound fertilizer and the nematicide aldicarb was studied at a site with a sandy soil (Trudgill, 1987). In this trial, the site was uniformly heavily infested with Globodera pallida and the tolerant cv. Cara produced tops that were generally twice the size of intolerant cv. Pentland Dell. Consequently, Cara tended to produce many more leaves than were required to give 100 percent ground cover. The yield of the Cara was increased equally by a half and a full rate of aldicarb whereas that of the Pentland Dell was increased more by the full rate. Similarly, increasing rates of fertilizer proportionally increased the yield of Pentland Dell untreated with nematicide more than it did that of treated Pentland Dell or untreated Cara. This trial and several others showed that initially the G. pallida proportionally decreased the top growth of both cultivars to the same degree, supporting the basic proportional model proposed by Seinhorst.
· A series of five trials on different soil types tested the same five potato genotypes in plots with a wide range of initial populations (Pi) of G. pallida. Excellent regressions between Pi and tuber yields were produced (Fig. 2) revealing differences in tolerance between genotypes and in overall rates of yield reduction at the different sites. Further analysis showed that variations from a basic model similar to a simplified Seinhorst curve (without T or m) could be partitioned into genotype and site effects. The former were common across sites and the latter across genotypes.
FIGURE 2: The relationship between initial population density of Globodera pallida and tuber yield for tolerant, moderately tolerant and intolerant genotypes at two sites with contrasting yield potential
This information provides the basis for predicting the effects of G. pallida on the tuber yields of different cultivars classified on their degree of tolerance and of sites classified by their soil type. However, the losses are still predicted as a proportion of the nematode-free yield. To have a prediction of the actual loss in tonnes per hectare requires an estimate of the yield potential of the cultivar and site, which requires yet further modelling.
Only with this information can yield losses be accurately quantified in financial terms and the tolerance limit identified. The alternative is to extrapolate from the available trial data and make allowances on the basis of experience for the obvious possible environmental influences. Methods of estimating yield loss are therefore of central importance and are considered below.
Methods of estimating yield losses
Pot studies can be used to determine some of the basic information on yield-loss relationships, but because of environmental differences and interactions, field studies are also needed. There are two approaches: one is to use nematicides at relatively uniformly infested sites; the other is to work at sites with a range of population densities but which are uniform in other respects. A combination of both approaches is often a happy compromise. The former gives practical information on the effectiveness and potential value of a particular treatment but tells little about the nature of the relationship. It also suffers from the criticism that nematicides have a range of side-effects. The latter has the benefit of producing information on the relationship between Pi and yield, but it requires experimental errors to be minimized. Because Pi estimates have large errors, accuracy is improved by reducing plot size and by taking and processing multiple samples from each plot. However, plot size must be large enough to obtain a realistic yield and adequate guard plants are essential.
Another option is to establish many small plots in large but otherwise uniform fields. These can be at random, in a grid pattern or along known trends in Pi. The plots can be split and a nematicide applied to one half. For each plot the Pi and yield are determined. The results will produce a scatter of points, hopefully with yield decreasing as Pi increases. Much of the scatter is due to errors in estimating Pi and yield, and it can be minimized by taking the average of all the results within each error band. Such an approach needs:
i) a wide range of initial populations;
ii) a uniform field;
iii) a large number of plots (100 or more); and
iv) the plots to be part of an otherwise uniform crop.
Control measures aim to protect the treated crop from damage, and to prevent nematode multiplication and so reduce the threat to the next susceptible crop in the rotation. Most cost-effective and successful is the growing of resistant varieties. However, while these will prevent nematode multiplication, they are often as vulnerable to damage as a susceptible variety. In yield-loss studies resistant varieties can be a very useful tool for preparing plots with reduced populations without the side-effects associated with other treatments. Between vulnerable crops, rotations involving non-hosts are almost essential. Nematicides, whether natural or artificial, are a last resort and should not be used as a crutch to compensate for poor management. They are always costly and frequently toxic and environmentally damaging. However, their side-effects can make them attractive in some situations; the oximecarbamates control a broad range of pests, until they develop resistance, while the fumigant nematicides release nitrogen, further increasing yields.
Nematodes have various reproductive strategies. Some grow large and have long life cycles with low rates of population increase (K strategists), others are relatively small, have short life cycles and potentially higher reproductive rates (r strategists). An endoparasitic habit with induction of giant cells or other rich and continuously available food sources, reduces exposure to predation and other stresses and further increases reproductive potential. A reduction in the number of active juvenile stages further decreases development time, thereby reducing generation time and increasing the potential for multiple generations in a season. A wide host range completes the adaptation of pathogens such as some Meloidogyne spp., which can be regarded as the ultimate plant-parasitic nematode r strategists. Many Longidorus spp. are examples of K strategists.
It is a characteristic of K strategists that they do best in stable environments where populations are usually close to the equilibrium density (the population density that can be sustained). In contrast r strategists increase rapidly where the environment is favourable, often overshooting the equilibrium density. Severe damage to the host occurs and the population crashes. This can occur with repeated cropping of hosts as Pi increases, environmental influences on sex determination reduce multiplication, parasites of the nematode increase in number, and increasing damage to the host and competition for feeding sites progressively reduces multiplication.
Consequently, nematode multiplication rates are strongly density dependent. Again, the question of how density is defined arises. Usually it is expressed as the numbers of nematodes per gram or ml of soil, but the units that directly affect the nematode are those that are root related, e.g. number of root tips and/or length or weight. Hence, a cultivar with twice the root mass of another will, except at low densities where the multiplication rate is the maximum, support a higher multiplication rate. Similarly, tolerant cultivars that maintain a greater root mass as Pi increases than intolerant cultivars, will have a greater equilibrium density and maintain a greater multiplication rate at high pre-planting population densities.
Overall multiplication rates are determined by the intrinsic maximum rate of multiplication, which is influenced by nematode species, the susceptibility (defined as all those qualities favouring the nematode) of the host, and the various environmental factors that influence both the nematode and the host.
Nematode multiplication can be modelled in different ways. For migratory nematode species that multiply continuously, Seinhorst (1966) proposed the following formula derived from a logistic equation:
Pf = aEPi/(a - 1)Pi + E
where a is the maximum rate of increase and E is the equilibrium density at which Pf = Pi. For sedentary nematodes with one generation at a time, e.g. potato-cyst nematodes, Seinhorst (1967) proposed an alternative model based on the competition model of Nicholson (1933):
where a is again the maximum rate of multiplication, and 1 - q is the proportion of the available space which is exploited for food at a density of Pi= 1.
Jones and Perry (1978) also proposed a model for sedentary nematodes with a logistic basis derived from the observation that sex determination is density dependent. Their model includes parameters that reflect fecundity and the proportion of the population that does not hatch.
All three models, in their most basic form, show maximum rates of multiplication at low initial densities. As Pi increases the rate of multiplication is reduced as an upper asymptote is reached (Fig. 3). In reality, the shape of this curve is modified as Pi increases due to the increasing damage inflicted and the loss of roots. With the Jones and Perry model this is exacerbated as space lost as a result of root damage increases the competition between invading nematodes, resulting in an even greater shift in the sex ratio towards male production than would otherwise be the case. Thus the approach to the asymptote is slower and indeed the asymptote is reduced below the theoretical level. Further increases in Pi can inflict so much root damage that the population increase becomes negative and the population size is ultimately reduced.
All the equations mentioned require modifying by including a damage function such as that of Seinhorst, which also allows the differences in tolerance between cultivars to be taken into account. The damage func- tions used model proportional differences, and further modification may be required to account for absolute differences in plant size.
Another plant characteristic that affects population increase is the host status of the plant. Differences can be modelled in terms of the maximum multiplication rate or the space required for successful multiplication (Seinhorst models), or in terms of fecundity or effects on the sex ratio (Jones and Perry model).
An important method of expressing and comparing the effects of different cultivars or cropping regimes is to consider the equilibrium density, i.e. the point at which Pf = Pi. This density is usually observed at a Pi which is larger than that which gives the largest Pf (Fig. 4). In practice this equilibrium density is reached after a period of oscillation about the equilibrium density. The size of the oscillations will be determined by the tolerance and resistance of the host. Tolerance and resistance will produce small oscillations, while susceptibility and intolerance can result in large oscillations. Indeed, these two factors can interact to the extent that a tolerant but partially resistant cultivar can produce a higher equilibrium density than an intolerant susceptible cultivar (Fig. 5).
FIGURE 3: The theoretical logistic relationship between initial population density and final population density and the relationship when roots are damaged
FIGURE 4: The relationship between Pf and Pi when a tolerant and an intolerant host are grown
FIGURE 5: The relationship between Pf and Pi contrasting the response when an intolerant susceptible host is grown to that which occurs when a tolerant but partially resistant host is grown
Care needs to be taken in devising management strategies for the control of nematodes to balance the benefits of tolerance against the benefits of resistance, to ensure that while yields are maximized, nema-tode populations are not raised to levels that are damaging to other cultivars.
Models can be used to examine and explore nematode management strategies, but need to take into account the effective population if this is less than the actual population, and the decline in the numbers of nematodes in the absence of a host crop.
Jones, F.G.W. & Perry, J.N. 1978. Modelling populations of cyst nematodes (Nematoda: Heteroderidae). J. Appl. Ecology, 15: 349-371.
Nicholson, A.J. 1933. The balance of animal populations. J. Animal Ecology, 2: 132-178.
Oostenbrink, M. 1966. Major characteristics of the relation between nematodes and plants. Meded. Landbou. Wageningen, 66: 1-46.
Seinhorst, J.W. 1965. The relation between nematode density and damage to plants. Nematologica, 11: 137-154.
Seinhorst, J.W. 1966. The relationships between population increase and population density in plant-parasitic nematodes. I. Introduction and migratory nematodes. Nematologica, 12: 157-169.
Seinhorst, J.W. 1967. The relationships between population increase and population density in plant-parasitic nematodes. II. Sedentary nematodes. Nematologica, 13: 157-171.
Trudgill, D.L. 1987. Effects of rates of nematicide and of fertiliser on growth and yield of cultivars of potato which differ in their tolerance of damage by potato-cyst nematodes (G. rostochiensis and G. pallida). Plant & Soil, 104:185-193.
Trudgill, D.L. 1992. Resistance to and tolerance of plant-parasitic nematodes in plants. Ann. Rev. Phytopathol., 29: 167-192.