Moniliophthora perniciosa (witches' broom disease of cacao)

Cocoa pods infected with M. perniciosa can contain diseased seeds (Cronshaw and Evans, 1978; Frison and Feliu, 1989).

Almost all seeds from diseased pods with extensive lesions are killed or are not viable (Frison and Feliu, 1989).

The pathogen can be transmitted in the hypocotyl of an ungerminated seed and it will remain viable as long as the seed is alive. Disease symptoms only become apparent once the seed has germinated (Cronshaw and Evans, 1978). Discharged basidiospores germinated and were infectious to pre-germinated cocoa seeds causing hypocotyl hypertrophy plus root and shoot reduction (Almeida et al., 1997).

The methods of controlling M. perniciosa fall into three broad groups; phytosanitation (or sanitation), spraying with chemicals, and the use of resistance, but in operation they often entail modifications in the way cocoa is grown. Each method or combination of methods needs to be appraised in terms of cost and benefit, and against a background of fluctuating cocoa prices this presents other problems. If witches' broom is controlled, then other pod diseases in the plantation must also be taken into consideration or their incidence may increase, thereby reducing the economic benefit of successful witches' broom control (Rudgard et al., 1993).

Sanitation is based on the concept that by removing infected plant parts the production of inoculum will cease, or at least be reduced to low levels, depending on the efficiency of the removal. This approach has remained the basis of witches' broom control since it was first proposed at the beginning of the twentieth century. In practice, sanitation entails the pruning of brooms as diseased pods generally contribute little inoculum and can be removed during harvesting. The frequency of pruning is related to broom formation, basidioma production and cropping period in the particular area and takes advantage of any dry season when flushing is low and dry brooms can be seen more readily. Essentially it depends on the long (induction) period for basidiomata to develop on brooms which means that the latent period can be several months. There is generally only one disease cycle per rainy season; that is, within the span of one year the disease is monocyclic.In cases where removal of individual brooms is feasible, vegetative brooms should be cut off at least 15-20 cm below the point of infection. Diseased material on cushions should be carefully removed by cutting it off as close as possible to the bark. Diseased pods together with their peduncles should be removed whenever healthy pods are harvested.Diseased tissues are distributed through the branches and canopy of the trees, with the bulk of brooms (both cushion and vegetative) and diseased pods found generally in the high canopy, with only relatively few on the trunk and lower branches. The removal must be as thorough as possible to have any chance of success.It is important to ensure that all removed brooms and diseased pods reach the ground and do not remain suspended on branches within the canopy. Elimination of the diseased tissues from the plantation after pruning is optional in some circumstances but necessary in others. The practice is recommended if annual sanitation has lapsed for more than one year, because older brooms on the ground could produce basidiomata within a few days of rainy conditions if not removed. Where sanitation is regular, pruned tissues can be left in the plantation, provided that removed diseased material should not be heaped or left uncovered on the ground in plantations, and leaf-litter should be allowed to accumulate naturally, or should be placed over the removed diseased material. If phytosanitary removal is combined with structural pruning, it is likely that branches with brooms will be cut off the tree, in which case such material should be cut into small pieces. The material can then be spread on the ground so that it is all in contact with the leaf litter, as this speeds up decomposition.Phytosanitation must be correctly timed, and the choice of times can be determined using epidemiological studies which show the duration and the timing of the climatic and disease cycles. In many cases, additional (secondary) removals would be required to supplement a main (primary) sanitation pruning to ensure effective disease management. Phytosanitation should not be attempted where it is not possible to complete the primary removal at the recommended time. No secondary removal should be applied without a primary removal. Sanitation is most effective where it is routinely practised every year, so that numbers of second-year brooms are kept to a minimum. For the primary removal, a number of criteria for selection of the correct time can be specified. Phenological data show that the periods of greatest incidence of brooms and the periods of minimum production of basidiomata coincide during the annual drier season, where one occurs. At this time, most of the diseased tissues are necrotic and easier to distinguish. The objective of the primary removal is to remove diseased material before the onset of the rains (and thus the onset of basidioma production), particularly as the limited amount of sporophore production in the early part of the rainy season can be significant. In areas with no distinguishable drier season, the primary removal is best timed 1-2 months after the end of the main period of broom formation. A delayed dry season or early onset of rains requires, respectively, a late or an early start in the application of the schedules. A secondary removal is required in some areas to cope with the late-emerging brooms and those missed in the primary removal, and in most cases, to further reduce the number of productive brooms before the main period of pod set and development.The rate at which witches' broom builds up in a new planting will depend on the amount of background inoculum arriving and the efficiency of any phytosanitation implemented. In isolated plantings, especially in low-risk areas, it would be possible to maintain disease at insignificant levels provided sanitation is started early, and practised conscientiously every year. In high-risk areas, background inoculum alone may cause appreciable pod losses however thoroughly the planting itself is cleared of brooms.In established plantings the most accessible point at which to interfere with the disease cycle remains the production of inoculum, and the key management method is the removal of potential sources of inoculum to sanitize the planting. In plantations with canopies less than 3.5-4 m tall phytosanitation is generally practicable provided that methods of sanitation are practised thoughtfully. However, reductions in pod loss may not be significant if neighbouring areas of cocoa are not sanitized. Phytosanitation methods need to be qualified where the amount of pod loss to witches' broom does not justify action to control the disease. However, accumulated cushion infection and death from witches' broom can also be a threat to cocoa production. There is no direct evidence that reduced cushion activity decreases pod set, but the risk should be avoided where the potential for cushion disease incidence is high, and so phytosanitation is recommended even in areas of low incidence of cushion disease.Research has shown that careful broom removal reduces disease incidence dramatically in some, but not all, situations. Successful phytosanitation is jeopardized when inoculum from adjacent or nearby farms reaches the sanitized area in sufficient quantities to cause appreciable pod loss and abundant infections on shoots and flower cushions. It is not possible to predict precisely the outcome of sanitation for a particular situation, but some indications can be given. The distribution of cocoa varies greatly across regions and countries, from almost continuous to sporadic. In all plantings of cocoa where phytosanitation is practised in the surrounding cocoa, there is a high probability that pod losses can be reduced substantially. In discontinuous plantings (farms separated by at least several hundred metres) where phytosanitation is not practised in the neighbouring cocoa, the chances of achieving a successful reduction in pod losses are moderately strong. Prediction of any reductions in pod losses is more difficult for a sanitized field of cocoa surrounded by continuous plantations where brooms are not removed. The probability of significant reductions will be smaller in areas where disease incidence is high, but the possibilities for yield increases will be greater than in low-incidence areas.The rehabilitation of old or heavily diseased plantings where cocoa has exceeded 5 m in height involves the application of treatments to increase the cocoa production of low-yielding trees. Where witches' broom is present, application in single plantations may not be successful due to the risks from background disease pressure. Rehabilitation of a witches' broom-infected plantation also needs to be associated with a commitment to regular maintenance and regular sanitation pruning to maintain any improvement in yield. High disease incidence can force farmers to change from intensive management, because cumulative cycles of broom formation cause heavier pod losses.Cocoa plantations fall into two categories for the recommendation of treatments. The first category has cocoa with a branch structure originating below 2 m. Such trees should receive a structural pruning to reduce height and improve access to the canopy, followed by a sanitation pruning.The second category consists of cocoa that lacks any branch structure below 3-4 m and trees should be pollarded at between 0.5 and 1.5 m from the ground to stimulate chupons for the regeneration of the tree. Bud grafting onto basal chupons is an additional possibility. Plantations of cocoa with the seedling habit may contain individuals that are exceptionally heavily attacked by witches' broom, which may also be unproductive. Such trees could be eliminated and replaced by better material. In cases where the potential production of the existing cocoa is generally low, interplanting with improved seedlings followed by gradual removal of the old trees (Turrialba Method) might be preferable to rehabilitation.The success or failure of these treatments is influenced by the time of year in which they are carried out. The most important consideration is the phenology of the recuperating trees and the microclimate to which the regenerating trees will be exposed. In shaded plantings where variations of microclimate are less violent, the selected pruning treatment should be applied towards the end of the rainy season. This allows the canopy to start reforming during the dry season and continue into the early rainy season when there is no risk of infection. The recommendation only needs to be adapted for unshaded plantings where daytime temperatures are excessive (over 32°C) in the dry season (Brazilian Amazon). In such cases, the selected pruning treatment should be applied at the middle/end of the dry season.All treatments generate considerable quantities of infected material pruned onto the ground within the plantation, and much of this is capable of producing basidiomata. All the branches, foliage and brooms should be cut up, making a compact layer of organic matter that will be quickly covered with fallen leaves and decompose rapidly. Mineral oil may then be sprayed onto the cut material on the ground to inhibit sporulation on the brooms.These treatments are intended to increase pod production in the medium term, but will immediately reduce pod production as they involve the removal of either part or most of the fruit-bearing wood and/or the leaf area of the canopy. The beneficial effects of such practices on yield and the time taken for recovery have only been partially investigated. Early results are encouraging, and indicate that about 1-3 years are required for the two recommended rehabilitation pruning methods. The Turrialba Method involves no appreciable yield loss if carefully managed.

The use of chemicals for the control witches' broom disease has shown theoretical promise which does not bring success in practice. For instance, it has been shown in some cases that certain fungicides can significantly reduce pod disease incidence, although in other cases no effect has been found using the same main ingredient (Laker and Rudgard, 1989). The dose, interval and number of sprays have all been found to affect activity. Commercial application of fungicides for the control of witches' broom has not been adopted in any cocoa-producing country, because increases in production have not given sufficient economic return to motivate farmers into applying the treatments. Two strategies for developing economically effective fungicide control may be viable. The pods may be sprayed with a broad-spectrum residual fungicide or the canopy may be sprayed with a systemic fungicide both to inhibit broom formation and to eradicate mycelium from shoots and flowers. Cocoa has a very different spatial arrangement of spray targets to the majority of crops, and spray equipment typically used on cocoa has been developed for general purposes and does not suit cocoa. This has been a significant inhibitory factor to the development of fungicide usage. Pods are relatively small targets which are dispersed throughout the lower architecture and canopy, up to a height of several metres. The greater part of any chemicals sprayed in cocoa plantations lands on the litter layer. Chemicals are also lost when pod surfaces are saturated with spray and run-off occurs. Residues are eroded by rain and canopy drip, although some redistribution by splash can occur between pods in the canopy. Pods are most susceptible in the period of greatest growth, which means that fungicide residues are diluted by surface expansion. These factors lead to loss or dilution of any active ingredient and result in the breakdown of disease control.Alternative methods of treatment have been found where, with only one application, surface concentrations of protectant fungicide were maintained above lethal doses for sufficient periods, and it is possible that such techniques may be made commercially viable. Correct timing is essential if treatments are to be effective. Epidemiological data can provide information on the period of highest risk to pods in terms of the presence of basidiomata, and also the periods with the greatest numbers of susceptible pods. The comparative epidemiology experiment enables researchers to identify the most effective times that protective chemicals might be applied for control of pod disease from witches' broom.Systemic compounds for protection of vegetative or flower cushion activity might be sprayed on to the canopy, and eradicant compounds could be used to kill developing mycelium in green brooms. Knowledge of the activity of each infection court would be fundamental to the correct use of fungicides. Compounds have been found that translocate sufficiently well with cocoa plants to allow their use in soil applications in young trees. Such compounds might cause problems of residues in the cocoa beans. If those reservations were dealt with satisfactorily, applications would have to be timed slightly earlier than foliar sprays, to allow time for the chemical to be translocated to growing apices. Compounds have also been proposed for use as antisporulants to be sprayed on to brooms, whether on the tree or the ground (e.g. mineral oil). Research has identified the danger that certain active ingredients can stimulate greater basidioma production, and the development of antisporulants is unlikely to receive a high priority.

Within cocoa plantations there is a natural biological control of M. perniciosa in fallen brooms by other micro-organisms, some of which might be exploited either by manipulating the environment of the brooms or by using the antibiotic effect of substances they produce. The discovery of a hyperparasitic fungus, Cladobotryum amazonense, on basidiomata of M. perniciosa has also prompted research on its potential in biological control. However, no organism has been found with sufficiently aggressive antagonistic activity under field conditions to merit further development.Culture extracts of certain antagonistic fungi have been shown to have inhibitory activity against basidiospore germination and saprophytic growth of M. perniciosa, but such extracts seem to be merely fungistatic active ingredients that would have to be bulk manufactured and formulated for wide-scale use. Similar difficulties in reaching brooms with sprays of such extracts would exist to those of reaching canopy pods with fungicide sprays.

It is widely recognized that improvements in long-term control of witches' broom can best be achieved by the use of resistant cocoa planting material, and the development of such material is urgently required. The possibility of selecting for resistance was considered in the earliest reports of witches' broom. Many cocoa breeders have noted that not all trees in germplasm collections from wild cocoa populations are equally susceptible, and in earlier work advised the selection of the best bearing trees rather than those which appeared less susceptible, as the disease was considered controllable. This approach was also used in Trinidad after extensive screening within the Trinitario population failed to detect any highly resistant material. However, in this further programme using yielding ability as the main criterion in a situation of high pathogen pressure, there was also a bias towards witches' broom resistance and some clones, such as ICS 95, showed considerable resistance expressed as fewer developed brooms.Meanwhile, a search for immune or resistant trees was made in South America in several expeditions, notably those of FJ Pound in the 1930s. Three clones from Pound's collection, designated Scavina 6, Scavina 12 and IMC 67, contributed a high degree of witches' broom resistance to progeny bred from them to improve bean size. These Trinidad Selected Hybrids have been widely planted in Trinidad since the late 1950s and appear to have been a major factor in reducing witches' broom there to its present relatively low levels. However, when apparently resistant selections of Scavina 6 were sent to Ecuador they became severely infected, probably because they encountered a more aggressive strain of M. perniciosa.Breeding for witches' broom resistance was subsequently started in other countries, often following further searches for germoplasm, for example, the programme at the Tropical Experiment Station, Pichilingue, with material collected by Desrosiers and von Buchwald. While there are selections with some resistance to isolates of M. perniciosa in Brazil, Trinidad and Venezuela, the level of resistance in commercial cocoa is still less than satisfactory. In other countries such as Bolivia, Colombia and Ecuador which have more virulent isolates the situation is even less satisfactory.Some resistance to M. perniciosa clearly exists in cultivated cocoa but little is known about its mechanism. In cultivated cocoa, there appears to be a complete range in severity of symptoms depending on pathotype and cocoa clone. Some accessions with relatively little witches' broom and high yields are of obvious interest but, equally, so are those which yield well despite high rates of infection.