Diversity, structure and health of a cocoa based agroforest system in the Humid dense forest, East Cameroon

Cocoa is a major cash crop in Cameroon, where its production and export contributes significantly to the national economy and in poverty alleviation. Cocoa-based agroforestry systems (cAFS) have been recognised as a fair strategy for natural resource management, combining both the agricultural and conservation objectives. This study aims to (1) assess the diversity, (2) analyses the floristic and structural characteristics as well as (3) the dendrological features of the (Exploitation Agricole Betti) (EAB), a cAFS vast of 120 ha, located in the East region of Cameroon. Cocoa and associated tree species were counted in 21 sampling plots of 0.25 ha systematicaly settled all over the system (EAB) between 28 August and 22 October 2016. A total number of 3 147 stems was recorded and distributed in 2,599 cocoa trees and 548 associated trees. The overall diversity of the system is low. The density of the cocoa trees is 495.0 stems/ha, correponding to a success rate of 44.6%. The EAB is attacked by the black pod disease. The average Pod Rot Attacked Index (PRAI) is 0.35 ± 0.38; and this varies significantly according to the associated trees density and the season. Further studies should aim to (1) identify different cocoa varieties planted in the system, (2) identify correctly all the pests and diseases of the system, (3) assess the impact of associated tree thinning and cocoa tree Pruning on the pests or diseases attack and on the cocoa production, and (4) to explore the usage of associated trees in the system. This with the view to come out with a fair model cFAS to use in tropical humid forest zones.


INTRODUCTION
Cocoa or Theobroma cacao L., is a tropical tree, the most important genus of the Malvaceae family because of its commercial value. Cocoa was introduced in Cameroon since 1886 by the German colonial administration and is a major cash crop in Cameroon and many other countries of the tropical world, where its production and export contributes significantly to the national economy and in poverty alleviation. The cocoa sector is a source of employment for about four million individuals and it is Cameroon's major agricultural export crop. The revenue generated from cocoa exports accounted for about 14% of non-oil exports in 2012, particularly to Europe (Ngoe et al., 2016(Ngoe et al., , 2018. In the past years, Cocoa was cultivated mostly by smallholders who usually farm on 1 to 3 ha of land (Sonwa et al., 2007;ICCO, 2014). Today, many persons have versed in growing cocoa in large areas (10 ha and +). This has been made possible due to the huge work of sensitisation made by the Cameroon government, the Ministry of agriculture and rural development (MINADER) and the Society for Cocoa development (SODECAO) to be precised. The Cameroon's rural development strategy, adopted in 2005, and whose implementation was intensified within 2012-2020 aims to "Ensure food security, the sustainability of performance and achieve integration in exchanges". Cameroon government is set to increase cocoa and coffee production to an appreciable level through the rehabilitation and creation of new seed farms; production and dissemination of plant material; setting up of systematic and integral phytosanitary treatment of cocoa and coffee farms; emergency programs to save production as well as boosting financial resources (Achancho, 2013 ;République du Cameroun, 2006, 2009. The adult cocoa tree (Theobroma cacao L.) is a tree that can reach 12 to 15 m in height when growing in the wild. Its size and the importance and development of its foliage depend very much on the space available. Thus, when planting, the usual spacings allow the adult tree to reach an average height of 5 to 7 m. When it comes from the germination of a seed, the cocoa tree reaches its full development around the age of 10 years. However, it is productive well before this age since flowers and fruits appear in the third or fourth year, with full yield generally being obtained around six or seven years old. A wellmanaged plantation can remain profitable for at least 25 to 30 years. The cocoa tree fruit, called cherella while it is growing and then pods when it reaches its final size, reaches maturity after five to six months depending on the origins (Mossu, 1990). The pod, before maturity, can be either green, or more or less dark red-violet, or green, particularly pigmented with red-violet. The varieties cultivated in Cameroon are not homogeneous. In the space of sixty years, and under the effect of several administrations, the first introduced varieties hybridized. Most of the cocoa trees in place are of the Forastero variety made up of the forms Amelonado (with yellow pods) and Cudeamor (with red pods), the latter being the most numerous (Champaud, 1966).
There are 600,000 cocoa farmers across Cameroon, and it is a vital sector for rural communities. But cocoa is a fragile crop with yields that tend to decrease over time, putting farmers' livelihoods at risk. That's why the African Development Bank has committed to provide funding to IRAD, the Institute of Agriculture Research for Development, where research is focused on creating adapted seed varieties. The second-generation seed varieties developed by IRAD allow for an average yield of 2 tons per hectare, compared to the first generation developed in the 1970s and 1980s that produced around 1 ton per hectare (https://www.afdb.org/en/successstories/cameroon-new-seed-varieties-help-cocoa-cropsbloom-and-farmers-thrive-33940).
Forest stand structure refers to the stand structural attributes and stand structural complexity (McElhinny et al., 2005;Zenner, 2000cit. Sonwa et al., 2016. Stand structural attributes include measures such as abundance, diversity, basal area, richness. Such measures can thus help in having a quantitative idea on the habitat created by combination of many components on a forest stand. In the case of cocoa agroforest, the forest structure is altered by the opening of forest stand to grow cocoa trees. The main aim of the manager is to alter the forest structure in such a manner that it provides suitable conditions to the growing of cocoa. In the past, management was mainly constituted by the introduction of cocoa seedling and regular management to maintain certain amount of shade and understorey slashing to reduce competition with cocoa seedling/trees. With the recent cocoa crisis (Sonwa et al., 2005) characterized by the liberalization of the cocoa value chain, the constant management of associated plants include elimination of some trees and introduction of more socio-economically useful ones to provide shade but also timber and non timber forest products such as food, medicinal and service plants to household (FAO, 2002;Sonwa et al., 2007;Bobo et al., 2006;Zapfack et al., 2002). The result of this management is a structurally complex system with abiotic (e.g. microclimate, humidity, etc.) and biotic elements (e.g. trees, vines, etc.) which, depending on the age and plants species composition, define a habitat structure different from the one of mono-species system such as pure cocoa orchard or cocoa with one or two associated species cultivated in an intensive manner (Sonwa et al., 2016).
Agroforestry is a land use management system in which trees or shrubs are grown around or among crops or pastureland. Agroforestry has its roots in tropical food production systems. The diversification of the farming system initiates an agroecological succession, like that in natural ecosystems, and so starts a chain of events that enhance the functionality and sustainability of the farming system. Trees also produce a wide range of useful and marketable products from fruits/nuts, medicines, wood *Corresponding author. E-mail: lagardebetti@yahoo.fr/lagardeprunus@gmail.com. Tel: 00 (237) This intentional combination of agriculture  and forestry has multiple benefits, such as greatly  enhanced yields from staple food crops, enhanced farmer  livelihoods  from  income  generation,  increased  biodiversity, improved soil structure and health, reduced  erosion, and carbon sequestration (USDA National  Agroforestry Center: Agroforestry practices, https://www.fs.usda.gov/nac/practices/index.shtml).
Carbon sequestration is an important ecosystem service that agroforest systems can provide. Agroforestry practices can increase carbon stocks in soil and woody biomass. Trees in agroforestry systems, like in new forests, can recapture some of the carbon that was lost by cutting existing forests. They also provide additional food and products. The rotation age and the use of the resulting products are important factors controlling the amount of carbon sequestered. Agroforests can reduce pressure on primary forests by providing forest products (Montagnini et al., 2004). Agroforestry practices are highly beneficial in the tropics, especially in subsistence smallholdings in sub-Saharan Africa (Kuyah et al., 2016) and have been found to be beneficial in Europe and the United States (Schoeneberger, 2017).
Cocoa-based agroforestry systems (cAFS) have been recognised as a fair strategy for natural resource management, combining both the agricultural and conservation objectives. These systems dominate in Cameroon, and are different to the intensive systems of monoculture, due to their diversification and resilience, which ensure long term cocoa production (Sonwa et al., 2007;Jagoret et al., 2009). Cocoa agroforests generally result from the clearing of some large forest trees and the thinning of part of the understory in order to introduce young cocoa plants. Other large trees are left during the establishment of the agroforest. Crops such as banana and plantain are used to shade the cocoa seedlings. A cAFS is a multi strata artificial system which look alike the natural forest. The main characteristic being the mixture of many species which are in perpetual competition within or between different species. The farmer seeks to enhance the productivity and the resilience of the system through the mixture of cocoa and other trees or crops (Sonwa et al., 2016).
Plant in the forest can be easily influenced by light, water, air humidity, wind, nutrient, heat and other biotic components. Such variables are likely to be modifified by the structure of the forest or agroforest. A structure with high shade intensity is known to slow the cocoa development and favors black pod disease (Ruf, 2011;Kouadio et al., 2018). While with less shade, mirid attack can be a serious problem. Anyway, the main disease of cocoa in Cameroon is the black pod rot in cocoa causing 80-90% losses without chemical control with Phytophthora Megakarya as causal agent (Mfegue, 2012). Plant diversity can also be linked to the structure of cocoa plantation. It is generally admitted that complex cocoa agroforests are richer in biodiversity than cocoa orchards.
Studying plant diversity of cocoa agroforest has revealed that land intensifification, market access and population density was affecting agroforests composition (Sonwa et al., 2007).
The Exploitation Agricole Betti (EAB), is a cAFS vast of 120 ha, based in the East region of Cameroon. The first plots were settled between 2008 and 2015, with the Cocoa being the main culture. This crop is associated with bananas and many tree species which were left for diffefrent purposes including shade, biodiversity conservation and valorisation in terms of timber and nontimber forest products. Till now, no specific study has been conducted in that complex. The knowledge of the diversity of this complex in term of species composition, stand tree structure and cocoa health is essential as this is the first step for proposing fair management measures with the view to better sustain and ensure the resilience of that system. It is clearly admitted that one cannot manage what he does not know. The key assomptions formulated for the EAB cAFs are : (1) the EAB is diversified and this diversity may influence (2) the density and (3) the health of the cocoa. This paper aims therefore to (1) assess the diversity, (2) analyses the floristic and structural characteristics as well as (3) the dendrological features of the Exploitation Agricole Betti as key elements for sustaining that cAFS and make it a model for cAFS settled in the humid forest zone of Cameroon.

Study site
The Exploittion Agricole Betti or EAB is a cAFS, vast of 120 ha, located between the community forest of Mbeth II in the Diang subdivision (Lom and Djerem division) and the communal forest of Doumé in the Doume subdivision (Haut Nyong division), East Region of Cameroon. The EAB is situated between 4°22'-4°58' latitude North and 13°34'-13°61' longitude East, on the axis Bouam (on the national road n°1)-Dimako (on the national road n°10). The average altitude is 691 m.
The climate is an equatorial and Guinean type, characterized by four in-equal seasons: a great dry season which goes from December to mid-March; a small rainy season from mid-March to May; a great rainy season from mid-September to November; and a small dry season from June to mid-August. Climatic data ( Figure 1) considered are those found in the city of Bertoua, based at about 35 km of the EAB (https://fr.climatedata.org/afrique/cameroun/east/bertoua-1000032/). The average temperature is 23.7°C, with the maximum at 24.8°C in March and the minimum in July (22.5°C). The average annual rainfall varies between 1 000 and 1 600 mm. October is the most rainy month (280 mm), while January is the less rainy month (20 mm).   illustrates the climatic data obtained for the Bertoua city, located close to the EAB for the period 2011-2016 (https://fr.climatedata.org/afrique/cameroun/east/bertoua-1000032/). Soils are iron soils type. The EAB is located in the Guinean Congolese floristic region, in the low and medium altitude, in the domain of dense and rain semideciduous forest of Sterculiaceae and Ulmaceae (Letouzey, 1985). This zone has already been subjected to forest logging by the year 1970, which explains the general feature of old secondary forests observed in the field.
The EAB was chosen for this study both because it allows comparison of different farms or blocs with different ages and because its management and plantation structure are typical of the whole region, making it especially useful for a case study. The exploitation is composed of three farms (plantations). Each farm is composed of several blocs settled in different years between 2008 and 2015 as shown in Table 1.
The first plots of the EAB were settled in 2008. The main culture consists of the Cocoa, which is associated to the bananas (plantain and sweet bananas). The main objective of the promotor (the farmer) of the EAB is to ensure that the agriculture development is not detrimental to the conservation of the forest. The specific objective is to yield cocoa and bananas with a less disturbance of the natural milieu. For all plots, the technical itinerary is almost the same: preparation of the nursery during the dried season, December (year 1)-January (year 2) with seeds obtained from the Institute for research on agriculture and rural development (IRAD), clearing of the forest by removing herbs and lianas by February-March of the year 2, cutting of shrubs and sapplings in April-May, planting of the cocoa at 3 m × 3 m with seedlings of at least 6 months bred in the local nursery from mid August-October, and felling of medium and big trees with the chain saw on November year 2. The first clearing of the planted cocoa occurs in February-March of year 3, and this is done every four months. The felled trees are left on the ground, with the view to ensure the good return of the material (minerals). The promotor of the EAB does not use fires nor fertilizers.

Data collection
The method used to assess the abundance of cocoa and associated tree species in the EAB is the one call "method for forest management inventories". This method consists of counting the number of stems of the resource on a representative sampling area with sampling plots (counting units) settled systematicaly all over the farm (EAB) and to (if needed) estimate the stock at the level of the useful forest area. Cocoa and associated tree species were counted in sampling plots of 0.25 ha (100 m long × 25 m large) from 28th August to 22th October 2016, in blocs 1 (8 years; 18 ha) and 2 (7; 7) of farm 1, and in blocs 10 (5; 3) and 11 (3; 18) of farm 3. Table 2 presents the characteristics of the inventories. Twenty one sampling plots totalising 5.25 ha for a sampling rate (ratio of sampled area/useful area in %) of 4.77% were systematically settled and distributed as follow: Bloc 1 (10 plots; 2.5 ha; 13.9% as the sampling rate), Bloc 2 (3; 0.75; 10.7%), Bloc 3 (2; 0.5; 16.7%), Bloc 4 (6; 1.5; 8.3%).
In each plot, we identified the cocoa trees and the associated tree species trough their trade or common names, and we recorded dendrometrical and dendrological parameters. Dendrometrical parameters include circumference of the tree at 20 cm above the ground for the cocoa, and circumference at breast high (CBH) for associated tree species. Dendrological parameters were recorded with the view to assess both the productivity and the health of the system. Those parameters were recorded only on cocoa trees, in two phases : the first phase from August 28th to September 5th 2016, and the second phase from 17th to 22th October 2016 in the same plots. Information recorded included the number of healthy and sicked or rotten fruits (pods) per cocoa tree. For this first study, all pods with any sign of illness (or attack) including rotting and necrosed pods were classified as sick pods. Health pods were those which were not seen with any visible sign of attack. Plants were identified in the field with the aid of Mr DJENDJ MIASSE, the local botanical technician, responsible of the Community forest of Mbeth II. Specimens were collected and brought to the National Herbarium of Cameroon, Yaounde. Identification were made with the assistance of Dr Barthelemy TCHENGUE and Mr Eric NGANSOP. Databases on plants taxonomy including LEBRUN and STORK (https://www.villege.ch/musinfo/bd/cjb/africa/recherche.php?langue=fr), JSTOR (https://plants.jstor.org/compilation/Erythrophleum.ivorens e), PROTA (https://uses.plantnet-project.org/fr) and the Plant List (http://www.theplantlist.org/tpl1.1/search?q=) were used for eventual verifications.

Data analysis
Diversity was analyzed, the floristic and the structural characteristics, as well as the dendrological features of the two farms inventoried.

Diversity
Diversity indices include the Shannon Weaver index, the Simpson index and the regularity or the equitability index of Pielou. The Shannon Weaver index (H') allows to assess the diversity level of each group (farm) taking into account the proportion of each plant in the group. Shannon weaver index is calculated as shown in Equation 1.
Where, Pi = Ni/N, Ni = number of individuals (trees) of the species i, and N = total number of individuals for all plant species in the group. The Shannon index is sensitive to the variations of importance of scarce species (Peet, 1974). It is equal to zero when there is only one specie, and its maximal value is Log 2 (S) when all species have the same dominance (Dajoz, 2006). The Simpson index (D) measures the probability for two individuals withdrawn randomly from a given group, to belong to the same plant specie (Dajoz, 2006). Simpson index is calculated as shown in Equation 2.
The Simpson index is sensitive to the variations of importance of dominant or abundant species (Peet, 1974). The regularity or the equitability index of Pielou allowed to note the relative mess disorder of the population. It measures the diversity level reached by a group compared to its maximal level of diversity. It compares two groups which have different number of individuals (Grall and Coïc, 2005). The regularity index tends to zero when almost all stem or individuals are concentrated on one single plant species. It tends to 1 when all species have the same abundance. A weak regularity illustrates the importance of a few dominant plant species. The regularity is calculated as follow: E = H'/Log2S, with S being the total number of species (Dufrêne and Legendre, 1997). The concomitance usage of the three indices including the Shannon, Simpson, and Pielou allows to make a complete analysis of the structure of the communities of plants (Grall and Coïc 2005).

Floristic characteristics
Floristic characteristics used include the habitat, the phytogeographic range, and the type and mode of scaterring of seeds. The habitats of the plants refer to the habitat frequently used by the specie in the nature. These were identified as defined in Letouzey (1970Letouzey ( -1972 who distinguished five types based mostely to the degree of perturbation including : the primary forest (or the forest which is less perturbed), the secondary forest (perturbed), the culture (plantation), the forest edge and the swamp forest.
The phytogeography of the associated trees was evaluated based on the typology made in Central and West Africa by Lebrun (1947), modified by White (1979White ( , 1985 and later used by Sonke (1998) in the Dja Reserve, East Cameroon. This system recognises four levels of species distribution including : (1) species with large distribution in the world (pan-tropicals or species found in all tropical areas in the world), (2) species largely distributed in Africa or pluri-regionals (afro-tropical or species found in tropical africa and tropical oceanic islands such as Madagascar, Seychelles), (3) species with regional distribution (afro-malagasy or species found in Africa and Madagascar, western guinean or species which range extends from the West Africa to Cameroon and the Congo basin, Guinean congolese or species found in the guinean region), and (4) species with reduced distribution (central guinea-congo or species which distribution area ranges from Cameroon to the Democratic Republic of Congo).
The types and modes of scattering of seeds were defined according to the model used by Danserau and Lems (1957) and which includes: (1) plants or fruits scattered by animals or humans (hanging and adhesive fruits-HgAd, fleshy and indehiscent fruits-FlIn), (2) by the wind (fruits with aliform appendages-AlAp, fruits with feathery or silky appendages-FeSi, fruits not fleshy and relatively light-FlLi), (3) by the same tree or selfscaterring (dried or fleshy fruits scattered by the plant itself-DrFl; dried or fleshy fruits but heavy and indehiscent-HeIn).

Structural characteristics
Structural characteristics are distinguished in two groups including horizontal and vertical structures. Horizontal structure was expressed using the density, the success rate and the dominance of each specie. The density expresses the number of trees or stems per surface area as seen in Equation 3.
where Ni = the number of individuals for the species « i » in the group, Sa = the sampling area in hectare (ha) and Di = the density of the specie « i ». It is expressed in number of stems or trees/ha. The success rate (Sr) is the ratio in percent of the number of living cocoa trees (current density) to the number of planted trees (initial density= 1 111 trees/ha). As stated above (study site), cocoa was planted at 3 m × 3 m, which guives a planting density of 1 111 stems/ha. The success rate (Sr)

Plantation Bloc (Age) Richness Shanon (H) Pielou (€) Simpson (D)) Density
Farm 1 Where D is the diameter of the tree, π = 3.14, G is the stand basal area. The basal area is known to be a good indicator (of species dominance) in several silvicultural management and is gradually admitted as useful also in agroforestry management (Sun Hong-gang et al., Nissen and Midmore cit Sonwa et al., 2007). The relative dominance of the specie corresponds to the ratio of the basal area of the i-th specie over the total basal area of all the plants in the sampling area. Vertical structure was expressed using the Letouzey (1982) classification in forest, and later adapted in cAFS based in the humid forest zone of Southern Cameroon by Sonwa et al. (2016). According to this classification, plants with diameter below 20 cm can be classified as shrubs, those with diameter ranging from 20 to 50 cm are saplings, trees comprised within 50-100 cm of diameter can be considered as average trees, and plants with diameter above 100 cm are big trees.

Dendrological characteristics.
Dendrological characteristics in this study refers to the health of the cocoa pods. We evaluated the intensity of the diseases through the Pod Rot Attaked Index (PRAI) expressed as Equation 5. PRAI = ∑attacked fruits / ∑total fruits (5)

Statistical analysis
Data analysis was performed using the R version 3.5.1 (2018-07-02), Ri 386 computer packages. This bundleage served to make the one way ANOVA, for example to assess the variance of the PRAI in different farms and different periods of counting.

Structural characteristics
The density and the dominance (illustrated by the stand basal area) of each tree specie are presented in Table 5. Table 6 presents the synthesis for associated trees and cocoa. The overall density of the whole system is 599.43 trees/ha. The density of the associated tree species is 104.4 stems/ha for a stand basal area of 16.21 m²/ha, while the density of the coacoa is 495 stems/ha representing a success rate of 44.6% for a stand basal area of 9.74 m²/ha. In general, Farm 1 (average age 7.5   (1) type of scaterring of fruits : (1) plants or fruits scattered by animals or humans-An (hanging and adhesive fruits-HgAd, fleshy and indehiscent fruits-FlIn), (2) plants or fruits scattered by the wind-Wi (fruits withaliform appendages-AlAp, fruits withfeathery or silky appendages-FeSi, fruits not fleshy and relatively light-FlLi), (3) plants or fruits scattered by the same tree or self-scaterring-Ss (dried or fleshy fruits scattered by the plant itself-DrFl; dried or fleshy fruits but heavy and indehiscent-HeIn).
The six most important tree species according to their dominance or relative stand basal area are Musanga cecropioides, Ceiba pentandra, Terminalia superba, Sterculia rhinopetala, Cordia platytyrsa, Triplochyton scleroxylon. Figure 2 illustrates the distribution of stems in different diameter groups. Saplings (45% of stems; 47 stems/ha) and shrubs (35%; 36 stems/ha) are more represented. Average and big trees are les represented but totalise 12.18 m²/ha for stand basal area, representing 75.18% of the total basal area of the system.

Dendrological characteristics
A total of 23 904 and 29 692 cocoa fruits (pods) was recorded in phase a (dry season) and phase b (rainy season) respectively in the EAB as shown in Table 7. Number of trees and pods increase in phase b (+256 trees, +5788 pods) compared to phase a. The productivity expressed by the average number of pods per tree is 17.49 and does not vary significantly between the two phases of counting (ANOVA, df = 1, F = 2.152, P ˂ 0.143). But this productivity varies significantly between different blocs (ANOVA, df = 3, F = 342.8, P ˂ 2e-16), with bloc 1 having the highest productivity (23.61 pods/tree). The number of attacked fruits increases from phase a to phase b, while that of healthy fruits decreases. This is expressed by the Pod Rot Atacked Index (PRAI) which varies significantly from one counting phase to another (ANOVA, df = 1, F = 516.1, P ˂ 2e-16). We can note that the PRAI increased, the value obtained in the second counting phase (0.48) being times 2 of that of the first phase (0.19). The PRAI also varies significantly from one bloc to another, whatever be the phase of counting (ANOVA for Phase 1, df = 3, F = 27.45, P ˂ 2 e -16, and ANOVA for phase 2, df = 3, F = 27.45 P ˂ 2 e -16). Bloc 1 (0.19 in phase a and 0.58 in phase b) and bloc 2 (0.31; 0.54) having the highest PRAI compared to blocs 10 (0.12; 0.10) and 11 (0.05; 0.14). The PRAI varies significantly from one farm to another (ANOVA, df = 1, F = 342.5, P ˂ 2 e -16). The PRAI obtained in farm 1 (0.41) being time 4 high than the one obtained in farm 3.

Diversity
The age of the blocs varies from 3 to 8 years. This exploitation can be considered as very young, the cocoa being a perennial plant (Jagoret, 2011). The technical itinerary used in the EAB is different to the one proposed by the agricultural services and the Food and Agriculture Organisation for Cameroon (FAO, 2002). In fact, FAO (opcit.) suggests that the preparation of the field should be done using the following logistical steps : (1) clearing of the herbs, shrubs, and sapplings, (2) felling of the incompatible or antagonist trees, (3) planting the cocoa one or two years later, with seedlings coming from the nursery. The main difference resides on the fact that, the EAB plants cocoa before cutting shrubs, sapplings, and before felling trees.
Ulmaceae, and Sterculiaceae figure among te ten most cited families, which confirms the position of the surrounding forest in the domain of dense and rain semideciduous forest of Sterculiaceae and Ulmaceae (Letouzey, 1985). The overall diversity of the system characterized by the Shanon Weaver (H) is = 2.69 which   Ns: number of stems, Sba: stand basal area, RSba: relative stand basal area; Ds: density. is low (H < 3 bits) according to Frontier and Pichod-Viale (1995). It is even too low compared to the 3.06 obtained in the cAFS found in the Centre and South regions of Cameroon (Jagoret and Messie, 2008). Cocoa farmers maintain a high diversity in their farms with the view to have a permanent shade, and also to combat the quick invasion of weeds (Jagoret and Messie, 2008). The Pielou index is also low (E = 0.42 < 0.5), showing that a small number of tree species have the high number of individuals (Djego et al., 2012). Globally, H and E are low, indicating that the EAB is an homogenous and specialized milieu (Djego et al., 2012). Six tree species including Musanga cecropioides, Ceiba pentandra, Terminalia superba, Sterculia rhinopetala, Cordia platytyrsa, Triplochyton scleroxylon totalise more than 80% of the species dominance. A study tour was conducted in Cameroon, Côte d'Ivoire, Ghana and Nigeria. Preferred trees by farmers that cut across the four countries include Milicia excelsa, Terminalia superba, Triplochiton scleroxylon, Alstonia boonei, Recinodendron heudelotti. Those tree species are found in our system. (Asare, 2005).

Floristic characteristics
Species of the primary and secondary forest The average diameter of associated trees or the diameter of the medium associaed tree is 34.7 ± 28.10 cm, and this does not vary from one bloc to another (ANOVA, df = 3, F = 0.273, P ˂ 0.845). The average diameter of the cocoa trees is 6.78±2.03 cm and this diameter varies significantly in different blocs (ANOVA, df = 3, F = 174.5 P ˂ 2e-16) with Bloc 1 having the high diameter, 7.51 cm.  (Jagoret and Messie, 2008). Plants with regional distribution are more represented (65.2%), with the Guinean congolese being the most important phytogeographic type (52.2%). This also confirms the position of the surrounding and initial forest of the EAB in the Guineo-congolese forest domain (Letouzey, 1985). Plants which are scattered by animals are the most important; 74.1% of the trees having fleshy and indehiscent fruits, indicating the key role of wildlife (animals) in the forest regeneration (Kidikwadi et al., 2015;Beina, 2011). It also illustrates the vestiges or traces of primary forests in the area.

Structural characteristics
In theory, the cocoa density is supposed to be 1 111 stems/ha, since the seedlings were settled at 3 m × 3 m. The current cocoa density of the system is 495.0 stems/ha which is too low, at least times 2 low compared to the initial density. This density is also too low compared to the agronomic norms recommended by Braudeau (1969). It is even low compared to the 900 stems/ha suggested by FAO (2002) for Cameroon. Densities of 1 911 stems/ha were obtained for young cocoa farms in the Centre region of Cameroon (Jagoret, 2011), 1 168 stems/ha in the humid forest zone of the Centre and South regions of Cameoon (Sonwa et al., 2016) while 1 111 trees/ha were found in the South west region of Cameroon (Bobo et al., 2006). Our cocoa density is too low compared to the range of 1 028 and 1 212 of the shaded plantations around Dalo and Gagnoa in Côte d'Ivoire (N'goran;2003). According to Sonwa et al. (2016), three reasons explain the reduction of cocoa density including: (1) destruction of cocoa trees during the felling of big associated trees, (2) cocoa trees which died due to pest and disease or (3) non-replacement of dead trees with the intention of managing more associated plants. Some villages in the Centre region of Cameroon presented the same cocoa density obtained in our study, 495.0 stems/ha with the high management of exotic fruit trees being the main reason. In our case, the reduction of the cocoa density is mainly due to the destruction caused by the felling of trees, as the technical itinerary used consists of felling trees after plantation of cocoa. The age of the EAB varies from 3 to 8 years with a cocoa stand basal area of 9.74 m²/ha. This value is high compared to the 5 m²/ha found in humid forest zones of Centre and South regions of Cameroon. The stand basal area of associated trees of our system is 16.21 m²/ha, which is low compared to results obtained in other agroforest systems in Cameroon (Sonwa et al., 2016). The total stand basal area of our system (cocoa and trees) is 25.95 m²/ha. This value is low compared to the 30 -36 m²/ha, obtained in cFAS of the humid forest zones of Centre and South regions of Cameroon (Zapfack et al., 2002;Sonwa et al., 2016). This tends to show that the EAB system is less shaded compared to cFAS cited in other forest zones of Cameroon. In contrary, our system is too shaded, at least times 5, compared to the system of annual culture (4.9 m²/ha) obtained in the South West Region of Cameroon (Bobo et al., 2006). In natural forests, 48.7 m²/ha and 40.0 m²/ha were obtained in near-primary and secondary forest respectively in the South West region (Bobo et al., 2006), while a basal area of 35.68 m²/ha were found in a forest stand in South Cameroon (Guedje, 2002). These results confirm the fact that the cFAS basal area is somewhere between those of forests and the one of annual culture, and more specifically closer to a forest basal area value. Cocoa value for land restoration, enrichment of biodiversity and provision of sustainable incomes in less advanced regions has been appreciated (Dropdata, 2015). The total density of 599.3 stems/ha obtained in our case, is too low compared to the 1 489 and 1560 trees/ha obtained in other regions of Cameroon (Sonwa et al., 2016;Zapfack et al., 2002). The density of the associated tree species of the EAB is 104 trees/ha. This density is low compared to the 204 trees/ha and 321 trees/ha obtained in the forest humid zone of Cameroon (Jagoret, 2011 ;Sonwa et al., 2016). Saplings (45% of stems) and shrubs (35%) are the most represented groups. This corresponds to an inverse J shape structure and remains the same in the four blocs of the system. This structure is quite different to the same inverse J shape obtained in the humid forest zone of Cameroon (Sonwa et al., 2016), where shrubs were the most important group (56% of stems) followed by saplings (33%). The average diameter of the associated tree species found in our system is 34.7 cm, which is too high compared to the 26.7 obtained in the Centre and South regions (Sonwa et al., 2016). The density of big trees (4.6 trees/ha) is low compared to results obtained in the Centre, South and South west regions of Cameroon (Bobo et al., 2006;Sonwa et al., 2016).

Dendrological characteristics and the healthy of the EAB
Cocoa is affected by a range of pests and diseases with some estimates putting losses as high as 30 to 40% of the global production. Cocoa can be attacked by many pest species including fungal diseases, insects and rodents. Common diseases affecting cocoa production include witches broom, frostry pod rot, black pod disease, vascular streak Die back. Pod rot, also known as phytophthora pod rot is caused by the fungus Phytophthora spp. Three fungal species of the same genus are capable: P. palmivora, P. capsici, and P. megakarya. P. capsici and P. citrophthora cause pod rots in Central and South America, whereas P. megakarya causes significant pod rot and losses due to canker, and it is the most important pathogen in Central and West Africa, known as the aggressive of pod Rot pathogens. Visible symptoms for P. megakarya are the rotting or necrosed of pods. Pods can be attacked at any stage of development, and the initial symptoms are small, hard, dark spots on any part of the pod (ICCO, 2013;2014;2015;Guest, 2007;Ngoe et al., 2018). Although we did not yet identified the correct disease, we can say that the EAB cocoa pods were mainly attacked by the P. megakarya, according to the symptoms noted. The productivity of our system ranges from 2.99 pods/tree in bloc 11 of 3 years to 23.6 pods/tree in bloc 1 of 8 years old. It is clear that this productivity increases with the age of the plots. The productivity in healthy pods ranges for the oldest bloc (bloc1 of 8 years old) from 10 pods/tree in phase b to 18 pods/tree in phase a. The productivity of this specific bloc during heavy rainfall is within the range of 9-16 healthy pods/tree obtained in the cFAS of 15-25 years studied in the regions of Centre and South West Cameroon (Ndoumbè-Nkeng et al., 2009). The average Pod Rot Attacked Index (PRAI) obtained for the two phases of counting is 0.35 ± 0.38. This PRAI varies significantly from one phase of counting to another, from one bloc to another and from one farm to another. The PRAI increased significantly in one month interval, ranging from 0.19 in phase one (28th August-5th september) to 0.48 in phase 2 (17-22th October). This can be justified by the increase in humidity through the high rainfall. In fact the first counting phase occured during the small dry season (August), while the second counting phase occured during the big rainy season, in october considered as the most humid month of the area. The PRAI obtained in the rainy month is time 2.5 high than the PRAI obtained in dry month, meaning that higher rainfall may increase the PRAI. The expansion of the black pod disease (Phytophthora megakarya) during heavy rains is very common in the equatorial rainforest of Cameroon (Opoku et al., 2002;Atangana et al., 2013). A study conducted in the Centre and South West regions of Cameroon clearly showed that disease, pod rot to be precised, increased with increasing quantity of rainfall (Ndoumbè-Nkeng et al., 2009). Ndoumbè-Nkeng et al. (2009) found that the highest pod rot rate (PRR) incidence occurred in 2003 at Barombi-Kang (70.3%) located in the South West Region of Cameroon and Mbankomo (64.76%) in the Centre Region when the quantity of rainfall was very high (>2 200 mm). In contrast, the lowest losses were obtained in Goura (Centre region) in 2001 (1.15%) when rainfall was low (751 mm). However, in this case, (Goura), the production of healthy pods per tree was also low, probably meaning that the rainfall was not sufficient to induce good fructification. We obtained a maximum PRAI of 0.48, corresponding to the Pod Rot Rate (PRR) of 48% in rainy season in our system. This value is in concordance with the annual rainfall of our system which ranges from 1000-1600 mm ; classifying the EAB between the low and very high rainfall sites as indicated above (Ndoumbè-Nkeng et al., 2009). It is generally indicated that a minimum of 1000-1200 mm of rainfall is required in a cocoa plantation to get a good yield (Mossu 1990cit. Ndoumbè-Nkeng et al., 2009. The best cacao yield is obtained with an intermediate rainfall regime (1100-2000 mm), and our system is in this interval, which explains the good productivity (average number of pods/tree) observed (17.5 pods/tree). Similar results were observed in southwest of Ghana, where rainfall is higher and more regular than in any other cocoa region (Ruf, 2011). The cocoa tree grows well in combination with other tree species that give shade to the cacao trees and provide other benefits for the farmer, like food, fruit, timber and fuel wood. Shade trees reduce the stress of coffee (Coffea spp.) and cacao (Theobroma cacao) by ameliorating adverse climatic conditions and nutritional imbalances, but they may also compete for growth resources (Beer et al., 1988). In Ghana, farmers stressed the negative effect of competition for light. Under heavy shade, cocoa trees tend to grow tall in search of light, which makes harvesting more difficul (Ruf, 2011).
The PRAI varies significantly from one farm to another. The PRAI obtained in farm 1 is time 4 high compared to the one obtained in farm 3. Two reasons may explain the high value of the PRAI in Farm 1 including the productivity and the shade intensity. The high level of PRAI in farm 1 is firstly explained by the relative productivity of different blocs, and which is itself justified by the age of each bloc. It has been proved that the disease incidence increases with the production (Ndoumbè-Nkeng et al., 2009).
The risk of black pod is exacerbated by shade trees (Ruf, 2011). The variation of the PRAI in different farms may also be explained by the shade intensity, which is link to the density of the associated trees. The average density of the associated tree species of the EAB is 104.4 stems/ha, with Farm 1 (bloc 1 and bloc 2) having the high density (116.3 stems/ha) compared to Farm 3 (85 stems/ha). The level of shade increases progressively with the proportion (density) of associated forest trees. The high density of trees in the plantation, tends to create dense shading and subsequently permanent moisture, favorable to the development of the disease. An excessif shade creates a more humid microclimate which induces the proliferation of diseases such as the black pod diseases and reduces the cocoa yield (Mossu, 1990 ;Bouley 1998cit. Kouadio et al., 2018. This tends to confirm the assumption which states that agroforestry systems are traditionally seen as one of the causes of increased pest and disease incidence, in contrast with full-sun monocultures (Armengot et al., 2020). Studies have proved that shading reduces the final yields of the cocoa in term of healthy pods. The shade modifies the quantity of the light, temperature, the air movements, which have direct effects on the photosynthesis, the growth and the yield of the cocoa (de Almeida et Valle 2007cit. Kouadio 2018Braudeau 1969). The most relevant arguments why the farmers in Ghana would like to have (more) shade trees on their cocoa farm are "improvement of air and water quality" and "the increased lifetime of cocoa trees" (Hoogendijk, 2012).
Musanga cecropioides, Ceiba pentandra, and Triplochyton scleroxylon are listed by the Cocoa Research Institute of Ghana (CRIG) as un-desired tree species in a cAFS for different reasons including competition for water and other resources and also they harbour some diseases (Asare, 2005). The three species total 28 stems for a density of 5.33 trees/ha. Piptadeniastrum africanum and T. scleroxylon are undesire tree species in a cAFS since they use to compete for water and other resources with cocoa (Adou et al., 2016). Celtis sp, Klainedoxa gabunensis, Lannea welwitschii, Macaranga sp, Myrianthus arboreus, Piptadeniastrum africanum, Tretrapleura tetraptera are known as undesirable tree species in a cFAS (Lavabre, 1959). All these informations may also explain the high level of the PRAI in our system. Optimization of shade is an strategy to control pests and diseases (Staver et al., 2001;Atangana et al., 2013). Disease losses can be reduced through integrated management practices that include regular tree pruning and shade management, leaf mulching, regular and complete harvesting, sanitation and pod case disposal, frequent pod harvest, regular removal of infested pods, weed management, appropriate fertilizer application and targeted fungicide use (Guest, 2007;Armengot et al., 2020).

CONCLUSION AND RECOMMENDATIONS
This study aimed to assess the diversity, to analyse the floristic and structural characteristics, as well as the healthy of the EAB, a cFAS based in the forest zone of the East region of Cameroon. The current cocoa density of the system is 495.0 stems/ha, for a succes rate of 46.4%. Higher shading coupled to the high rainfall tend to increase the black pod disease attacks. Further studies should aim to (1) identify different cocoa varieries planted in the system, (2) identify correctly all the pests and diseases of the system, (3) assess the impact of associated tree thinning and cocoa tree prunung on the pests or diseases attack and on the cocoa production, and (4) to explore the usage of associated trees in the system. This with the view to come out with a fair model cFAS to use in humid forest zone in the East region ofi Cameroon.
This study aimed to assess the diversity and to analyse the floristic and structural characteristics of the Exploittion Agricole Betti (EAB), a Cocoa-based agroforestry systems (cAFS) vast of 120 ha, based in the forest zone of the East region of Cameroon. The current cocoa density of the system is 495.0 stems/ha, for a succes rate of 46.4%. The density of the associated trees is 104 trees/ha. The overall diversity is low (H < 3 bits and E < 0.5) indicating that the EAB is an homogenous and specalized system. Nineteen out of the 78 associated tree species identified are incompatible with the cocoa for different reasons The cocoa pods are mainly attacked by Phytophthora spp, the average Pod Rot Attacked Index (PRAI) being 0.35 ± 0.38. This PRAI varies significantly according to the associated trees density (degree of shade) and to the season. Higher shading coupled to the high rainfall tend to increase the PRAI.
Further studies should aim to (1) identify different cocoa varieries planted in the system, (2) identify correctly all the pests and diseases of the system, (3) assess the impact of associated tree thinning and cocoa tree prunung on the pests or diseases attack and on the cocoa production, and (4) to explore the usage of associated trees in the system. This with the view to come out with a fair model cFAS to use in tropical humid forest zones.