Disease and Sustainability in the Cocoa Agroecosystem


Harry C. Evans
CABI BIOSCIENCE
UK Centre (Ascot)
Silwood Park
Ascot, Berks. SL5 7TA
United Kingdom

Introduction

Theobroma cacao and its relatives are common understorey trees in the forests of the Amazon Basin (Cuatrecasas, 1964). In this centre of diversity, the genus Theobroma appears to be relatively widespread and locally abundant and, undoubtedly, constitutes an important component of the forest ecosystem. In the forests of Par?- State (Brazil), for example, up to 14 Theobroma trees have been recorded per hectare, with similar densities in eastern Ecuador and Peru (author, pers. obs.); whilst Allen (1982) recorded five trees of T. cacao per hectare in the Rio Napo region of Ecuador. The closely related genus Herrania occupies a comparable niche in the densely-forested, eastern slopes of the Andes, but is also common in the forests of the western Andes (Schultes, 1958). In such habitats, the amount of light reaching the lower forest canopy is probably the single, most important factor affecting the phenology of the understorey trees. As Cuatrecasas (1964) noted, these understorey trees demand constant high humidity and protection against direct radiation and evaporation; whilst Johnston (1912) considered that, "Theobroma does not resist even short dry seasons without the protection of dense shade and local humidity". Thus, growth is controlled and relatively slow, as are both the flushing and flowering patterns, and the populations of pests and diseases attacking the plant are similarly regulated by shade density. In accordance with its forest origins, cocoa was traditionally grown under thinned forest or with planted, permanent tree shade. However, in the 1950's work in Trinidad and Ghana showed that cocoa actually grows more vigorously and produces more flowers and, therefore, produces more crop in full sunlight compared with shade (Adomako, 1983). Nevertheless, there is a price to pay and it was subsequently discovered that this high production phase is of short-term duration compared with the life-span of a cocoa tree, and that after ten years or so, there was a marked decline in tree vigour due to a combination of factors, including increased water stress, due to increased soil moisture evaporation, excessive leaf transpiration and damage by insect pests. In other words, cocoa cultivation without shade was found to be non-sustainable. Although diseases have not been directly implicated in this decline (Adomako, 1983), pathogens do constitute the most serious production constraint for cocoa (Gotsch, 1997), and thus, if these are not successfully managed, then production will aways remain low. The most problematic of these pathogens are now analysed in relation to cocoa agronomy and sustainability.

 

Witches' Broom Disease

Of the major cocoa diseases, it is probable that only witches' broom disease, caused by the basidiomycete fungus, Crinipellis perniciosa (Stahel) Singer, is truly coevolved with cocoa, indeed, it was this disease, which was noted as long ago as the Eighteenth century in Amazonia (Silva, 1987), that, until recently, prevented expansion of cocoa as a plantation crop in its centre of origin. This pathogen has had a dramatic impact on commercial plantations geared for maximum production at an early age, typically involving vigorously growing, heavily flowering cocoa hybrids, intermixed with or based on "resistant" clones, planted in unshaded blocks in open situations with a high soil fertility: a recipe for disaster when a pathogen evolved to invade and colonise meristemmatic tissues is present (Evans, 1981a).

C. perniciosa is just such a pathogen, since it is intimately linked with the physiology of its host. It is favoured by increased tree vigour: the more susceptible tissues or invasion courts, such as flushes, flower buds, cherelles, the more infection. Evans (1981a) undertook a case study of the disease, based on a commercial plantation in Ecuador, but similar scenarios have been played out since in other South American countries. An analysis of the possible inter-relationships between cocoa agronomy and disease incidence is given in Figure 1. It is concluded that absence of shade, or growing cocoa as an orchard crop, favours disease epiphytotics. As long ago as the 1930's, it was noted by Briton-Jones & Cheesman (1931) that, with or without the activity of insect pests, any reduction in shade results in defoliation of cocoa followed by a flush of growth often coinciding with sporulation of the pathogen and thus increased infection. In addition, as well as eroding the "resistance" or increasing the susceptibility of the host, absence of shade also increases the fitness of the pathogen, both directly and indirectly. Evans & Sol?>rzano (1982) showed from a large scale dispersal experiment in Ecuador that the sporulation potential of C. perniciosa in unshaded plantations is significantly higher than in shaded cocoa because the witches' brooms are subject to direct sunlight and drying winds, ideal conditions for maintaining the viability of the colonising mycelium and reducing the impact of secondary, invasive organisms. Thus, exposed brooms produced more than three times the amount of inoculum (basidiocarps) than brooms in shaded conditions moreover, at the end of the season, the latter showed extensive decomposition by opportunistic invaders, whilst the former remained intact, without evidence of secondary invasion or colonisation, thereby increasing not only the amount of basidiocarps produced but also the duration of sporulation.

It was also noted that growing cocoa in unshaded blocks without wind breaks results in increased air movement or turbulence (Evans, 1981a) and that air currents especially favour inoculum movement at the plantation edges, with the air flowing close to ground level from the open area into the trees, up under the canopy in an updraught, and then back into the open where it meets the downdraught created by the open situation and thus back into the plantation (van Arsdel, 1967). Faced with this situation, it is not surprising that both pod and flower infections were extremely high (Evans, 1981a). Indeed, the self-incompatible hybrids appeared to be systemically infected since almost every flower cushion became broomed. The situation may be further exacerbated by low populations of pollinating insects in unshaded, "sterile" cocoa plantations which may lead to continuous flowering and, of course, to a constant source of susceptible tissues. Such massive cushion infection has been observed regularly in wild Theobroma trees in disturbed Amazonian rain forest, when widespread understorey species such as T. speciosum, are suddenly exposed to full sunlight and inherent physiological stability is lost. In these situations, where the upsurge in flowering activity coincides with environmental conditions which favour basidiocarp production, complete trunk infection can occur (Author, pers. obs.). Ironically, Pound (1938) considered that this species has resistance to the disease, since he encountered only limited broom formation on trees during his surveys in the Amazonian forest.

Once this cycle of infection is underway in the unshaded plantation system, there is little hope of slowing its progress and, indeed, many of the panic measures to contain the disease, such as ill-timed or injudicious pruning, fertiliser and pesticide application, may only serve to fuel it. Insect damage, as well as colonisation by opportunistic, necrotrophic fungi and parasitic algae (Cephaleuros mycoides, for example), may lead to extensive die-back and gradual debilitation of the stressed, unshaded trees, particularly if high solar insolation and low water status coincide. As the canopy fails to close or breaks-up, the plantation rapidly becomes uneconomic and, in effect, sustainability has been lost. Probably such a scenario is being acted out in an old illustration of a heavily pruned plantation in Surinam at the turn of the century (reproduced in Evans & Prior, 1987), which depicts a heavily pruned, probably lightly-or non shaded plantation. As it turned out, the strenuous efforts to manage the disease by the Dutch planters failed to save the previously buoyant cocoa industry of Surinam from collapse.

 

Monilia or Frosty Pod Rot

This pod disease, caused by the mitotic fungus, Moniliophthora roreri Evans et al., the anamorph or asexual stage of an unknown basidiomycete, is also thought to be a coevolved pathogen of the genus Theobroma or Herrania, originating not in Amazonia but in the rainforests of the Pacific coast. Rorer (1918) reported the disease on wild species of "cacao"such as T. bicolor (cacao blanco) and Herrania balaoensis (cacao de monte) in the Andean foothills of western Ecuador. This coastal, rather than Amazonian origin was further supported by the discovery of a new species of Theobroma (T. gileri) in the isolated, dense rain forests of north-west Ecuador (Cuatrecasas, 1964), which was reported as having Herrania-like fruits commonly infected internally by a fungal disease which matches the description of frosty pod rot. Such internal infection, with only minimal external symptoms, has been noted previously in collections of Herrania spp. in Ecuador (Evans, 1981b), and it has been speculated that this represents a survival strategy of the fungus in rain-forests where infection courts (young pods) are severely limited, due to the scattered, low density nature of the host and the rare occurrence of flowering and fruiting in the understorey. Thus, gradual decay of the diseased pods over time will result in a slow, long-term release of inoculum as the pod walls erode and the powdery, internal, "resting spores" are released.

The impact of shade or no shade on disease development is much more problematic than with witches' broom disease, and hence more difficult to interpret (Fig. 2). However, because M. roreri infects only the pods, then the disease cycle is intimately linked with that of pod production and, therefore, microclimatic conditions at the pod surface, or in the trunk region in general, coupled with air movement, probably constitute the most critical factors in driving the disease cycle. Potentially, the more favourable microclimate (i.e. high humidity) associated with pods in shaded cocoa will favour spore germination and hence the initial infection process. Conversely, the increased air movement in unshaded cocoa, especially when grown in unprotected blocks will strongly favour rapid drying out of the spore masses and, therefore, efficient release and dispersal within and between trees. Because the sporulation phase is fuelled by the massive internal colonisation of the pod by the fungal hyphae, the pathogen can sporulate abundantly using the host's water reserves and is independent of external climatic factors. Degradation or natural removal of the infected pod as an inoculum source over time will be slower in unshaded cocoa since secondary microorganism activity will be lower compared with the more humid, shaded plantation, thereby increasing its potential infectivity span, as the spores can survive for prolonged periods (Evans, 1981b).

Thus, there are pros and cons for and against shade and, on the basis of current knowledge, it is difficult to correlate pod disease losses with cocoa agronomy. The analysis (Fig. 2) suggests that, in unshaded cocoa, there are more factors favouring increased disease levels compared with shaded. However, the model needs to be tested experimentally before firm conclusions can be reached. Whatever the situation, control measures based on good cocoa agronomy are required if this dangerous pathogen is to be successfully managed within its current area of distribution and before it spreads into other "exotic" cocoa-growing regions. This disease alone could make cocoa cultivation uneconomic and, therefore, unsustainable.

 

Black Pod Disease

As with M. roreri, the Phytophthora species which cause black pod disease are primarily pod pathogens, although both stems and leaves can also be attacked, depending on the species complex involved and the prevailing climatic conditions. Thus, the pod microclimate is a critical factor in the infection cycle, particularly when motile spore stages are involved. Logically, therefore, any reduction in humidity beneath the canopy, in particular by removal of shade or shelter belts, should create more unfavourable microclimatic conditions for infection and, thereby, directly result in a decrease in black pod disease incidence. Sporulation will also be reduced under such conditions, especially as the air movement increases. However, balanced against this will be the changing pattern of rainfall without a buffered shade canopy which simulates the forest upperstorey, potentially leading to its greater impact on spore dispersal. For example, direct rain hitting the pod surface and soil will serve to increase the intensity and velocity of rainsplash, which, if combined with high winds, could alter the dispersal gradient in favour of increased long-distance spore movement (Maddison & Griffin, 1981).

Unlike the aforementioned diseases, black pod disease has been shown to be closely linked with invertebrates, which may act directly or indirectly as vectors (Evans, 1973; Gregory & Maddison , 1981). Of particular importance are tent-building ant species (Evans, 1971), which can initiate or expand infection foci. Leston (1973) has discussed how the overall ant composition of a cocoa farm can impact on its ecology and affect pest and disease levels, and, how this so-called ant mosaic is determined by the agronomic features of the farm or plantation, and, in particular, the quality and quantity of shade. Undoubtedly, the ant mosaic in shaded cocoa will be considerably more complex than in unshaded farms and the ant genus or species composition can be altered agronomically, perhaps to favour an association which involves a non-vector or, which actively discourages tent-builders. For example, Crematogaster striatula, an important vector in Ghana since it uses black pod tissues in tent construction, is favoured by heavy shade (Evans, 1973); whilst Pheidole megacephala tends towards more open, unshaded conditions and has been found to be of particular importance as a vector in Nigeria (Taylor & Griffin, 1981). Their relative abundance, as well as that of other genera, may also be influenced by aggressive ant genera, such as Oecophylla, which, although tending mealybugs, do not build protective tents. Thus, these genera could impact, either directly or indirectly, on the disease cycle and critically affect black pod disease levels. As Fig.3 implies, however, this is still a highly speculative area and considerable basic epidemiological as well as ecological data remain to be collected and analysed before manipulation of the ant mosaic can be considered as part of an integrated management strategy.

 

Cocoa Swollen Shoot Virus (CSSV)

As discussed by Posnette (1981), since cocoa is an understorey forest tree, it has been grown traditionally under the shade of taller trees, which typically in West African cocoa farms, comprises "survivors" of the indigenous forest complex. Unfortunately, because a relatively high proportion of these shade trees which were left belonged to related Sterculiaceae genera, often being conserved for their useful by-products, these acted as sources of coevolved pests and diseases.Thus, plant-sucking homopterans readily migrated to the new host and, as a consequence, vectored the indigenous virus (CSSV) from the infected but predominantly symptomless shade trees. Therefore, as shown in Fig. 4, shade can have an indirect, detrimental impact on cocoa by harbouring virus inoculum. In addition, heavy shade also favours the main mealybug vector, Pseudococcoides njalensiis. Nevertheless, removal of shade, and high forest in general, can also be linked to a greater disease incidence. Bigger (1981 a & b) associated the deforestation of southern Ghana with subtle changes in the insect fauna of "exposed" cocoa farms, which resulted in a decline in P. njalensis populations and a corresponding increase in more mobile vector species, culminating in an increased rate of spread of CSSV. Once again, as for black pod disease, the composition of the ant mosaic, and, in particular, the mealybug-tending genera, will be critical in the disease cycle and this, in turn, will be influenced by the agronomy or shade regime of the cocoa farm. Thus, similar arguments, concerning the importance of the ant mosaic in the disease cycle and its posible manipulation, can be put forward for CSSV as those proposed for black pod disease. However, for the time being at least, these remain speculative.

Since CSSV also disrupts the cocoa root system, as well as affecting the shoots, virus-infected trees are more prone to water stress. Thus, unshaded cocoa is more predisposed to stress due to increased evapotranspiration directly from the plant, as well as from the soil, particularly during periods of drought and high solar insolation. It has been proven that such stressed cocoa trees with low bark turgor pressure are more susceptible to Phytophthora cankers, probably because the resistance mechanisms in the bark are more easily overcome by the pathogen, resulting in large spreading or unchecked cankers which may prove to be lethal to the tree (Evans, 1981c). Increased invasion of weakened cocoa tissues may also be due to opportunistic fungi (Calonectria), insect pests, parasitic algae and higher plants, which may ultimately result in the gradual but irreversible decline in tree health and the abandonment of cocoa as a viable crop.

 

Discussion :

An attempt has been made in this paper to assess the impact of one agronomic variable (shade or no shade) on the major diseases of cocoa. Shade is probably the single most important factor controlling the physiology, and hence the growth of cocoa, since this species, and indeed the genus Theobroma, evolved as understorey forest trees. It is both difficult and dangerous to deal in isolation with a single variable or a single disease, without taking into account all the other variables which impinge on the crop. Thus, of necessity, the assessment has been somewhat simplistic. Nevertheless, some firm conclusions can be reached, assuredly with witches' broom disease and probably with CSSV.

Johnson (1912) wrote that, "The cultivation of certain inter-crop and catch-crops with cocoa might tend to check disease diffusion; and belts of trees, planted at suitable distances apart throughout the plantation would serve both to protect the cocoa trees from wind and deter the spread of disease." Certainly, the Brazilian (Rondonia) and Ecuadorian experiences with witches' broom disease lend support to this statement. Unprotected, unshaded, and hence unbuffered cocoa, based on vigorous hybrids, favours disease epiphytotics when witches' broom disease is endemic. The physiologically unstable trees, combined with increased air movement within and above the canopy, which results in rapid drying-out of the brooms, and, potentially to increased sporulation in time and space, as well as in efficient spore dispersal within and between cocoa plantings, are the factors which drive the infection cycle. Frosty pod rot may also increase in this situation as the dry spores are released in moving air, which, if increased due to absence of shade or windbreaks, would disperse the spores more readily and efficiently. However, there may also be a trade-off between the conditions affecting spore dispersal and those conducive to infection, making it difficult to assess the relative importance of shade versus no shade on the incidence of M. roreri. This also applies to black pod disease, which is probably even more dependent on high humidity within the cocoa canopy, and to free moisture on the pod surface. The situation here, however, is further complicated by the still poorly understood interrelationships between disease-invertebrate vectors-ant mosaic and shade. A similar complex chain of associations could be constructed for CSSV, making it difficult to interpret the importance of any one link within the chain. An added variable to further complicate the equation, is the effect of shade on the entomopathogenic fungi which attack the vectors and the components of the ant mosaic, as well as cocoa arthropods in general. Evans (1974, 1982, 1989), following studies of tropical forest and tree crop ecosystems in Africa and South America, concluded that these display the richest and most diverse interactions between ants (and arthropods in general) and entomopathogenic fungi. It was further concluded that cocoa farms tend to have less pest problems when the conditions reflect those of the natural habitat, the multi-storeyed tropical high forest, and that, in such situations, the pressures from entomopathogenic fungi are significantly higher than in degraded forest and poorly shaded cocoa (Samson et al., 1988). Evans (1974) considered, therefore, that the nearer the conditions for cocoa cultivation approach those of the natural forest, the less will be the danger of a breakdown of natural regulation. However, dense shade is associated with low pod production, as well as with high humidity which favours pathogens such as Phytophthora. Measures to control the latter, particularly the use of fungicides, will reduce the effectiveness of the entomopathogenic fungi, which may, ultimately, result in pest outbreaks. As well as impacting directly on arthropod pests and potential vector species, entomopathogenic fungi may act as important regulators of those arthropods which indirectly influence the disease cycles through a myriad of interactions, most, if not all, of which are still not understood. These need to be addressed, as well as the vexed question of shade (both density and composition) so that an integrated management strategy can be developed which will enable cocoa to be grown sustainably whilst maintaining a commercially acceptable level of production.

Fig.1. Comparison of shade and no shade on witches' broom disease

Shade No shade
Favours controlled flushing and flowering = fewer infection points Favours excessive flushing and flowering = increased infection points
Favours pollinators and synchronous flowering = fewer infection points Discourages pollinators, increases flowering, often asynchronous in self-incompatible trees = increased infection points
Favours humid microclimate & rapid degradation of brooms = reduced sporulation & low inoculum potential Favours rapid drying-out of brooms & more sporulation over longer periods = increased sporulation and high inoculum potential.
Buffered air movement

= inefficient spore dispersal

= reduced infection

?Y

PHYSIOLOGICAL STABILITY

?Y

CONTROLLABLE ENPHYTOTICS

?Y

SUSTAINABLE

Increased horizontal air movement & convection currents

= increased infection

?Y

PHYSIOLOGICAL IMBALANCE

?Y

UNCONTROLLABLE EPIPHYTOTICS

?Y

NON-SUSTAINABLE

 

Fig. 2. Comparison of shade and no shade on frosty pod rot

Shade No shade
Buffered air movement

= inefficient spore dispersal

= reduced infection

Slow drying of spore masses

= reduced spore release

Increased horizontal air movement & convection currents

= increased infection

Rapid drying of spore chains

= favouring rapid spore release & efficient dispersal

Favours rapid degradtion & colonisation of pod by secondary invaders

= decreased survival

Favours rapid drying & mummification of pod

= less secondary invasion

= increased survival

Favourable microclimate for spore germination

= increased infection

Potentially low humidity

= reduced spore germination

= reduced infection

 

Fig. 3. Comparison of shade and no shade on black pod disease

Shade   No shade
Humid microclimate

= increased sporulation & dispersal

= increased infection

  Drier conditions

= decreased sporulation

= reduced infection

Canopy disrupts/ reduces impact of rain

= reduced dispersal efficiency

  Impact of rain increases

= increased dispersal efficiency

Complex ant mosaic ?z

 

?Y

Less chance of a vector genus becoming dominant

= potentially reduced disease levels

STILL

SPECULATIVE

unknown impact on

disease levels

?o Simplified ant

mosaic

?Y

Possible increase, if conditions

favour a vector ant genus

Less water stress

= high bark turgor pressure

= reduced stem cankers

  Increased water stress

= variable bark turgor pressure

= increased stem cankers

 

Fig. 4. Comparison of shade and no shade on CSSV

Shade   No shade
Increased density of secondary forest hosts

= Increased inoculum sources

  Decreased density of secondary forest hosts

= decreased inoculum sources

Favours mealybug spp. With low mobility

= less dispersal of CSSV

= less infection

  Favours highly mobile mealybug spp.

= increased dispersal

= increased infection

Complex ant mosaic ?z

?Y

Less chance of a vector-tending ant genus becoming dominant

= potential reduced disease

STILL

SPECULATIVE

Unknown impact on disease levels

?o Simplified ant mosaic

?Y

Possible increase in disease if conditions favour mealybug tending ant genus.

Less water stress; infected trees maintain canopy

= less cankering

= less die-back (Capsid/ Calonectria)

= less mistletoe

?Y

SUSTAINABLE

  Increased water stress

= increased cankering

= increased die-back

= increased mistletoe

 

 

?Y

 

NON-SUSTAINABLE

 

References: