Influence of Environmental Factors on Cacao Biometric Attributes

Agrotrópica 28(2): 123 - 140. 2016. Centro de Pesquisas do Cacau, Ilhéus, Bahia, Brasil INFLUENCE OF ENVIRONMENTAL FACTORS ON CACAO BIOMETRIC ATTRIBUTES Guilherme Amorim Homem de Abreu Loureiro1,2, Quintino Reis de Araujo1,2, Raul René Valle1,2, George Andrade Sodré1,2, Lindolfo P. Santos Filho1, Sérgio J. R. Oliveira2, Paulo A. S. Dantas1, Lícia R. Couto1, Matheus R. Lopes1 1 Cocoa Research Center (CEPEC/CEPLAC), 45600-970 - Ilhéus, Bahia, Brazil. [email protected]; 2 State University of Santa Cruz (UESC), 45662-000 - Ilhéus, Bahia, Brazil. This study aimed to investigate the influence of different sites of cacao (Theobroma cacao L.) cultivartion, represented by different soils and cropping systems, on biometrics attributes of pods and beans. The 12 study sites are in the cacao region of Bahia, Brazil, in the humid zone. The variation and the statistical differences between the means of these attributes clearly indicate the influence of genotype and environment interaction. Attributes pod wet biomass, beans with mucilage wet biomass and number of beans were positively correlated with the cultivation sites represented by Argisol Red-Yellow Eutrophic cambisolic and Argisol Red-Yellow Dystrophic Cohesive abrupt soils. The highest values of length and dry weight of dry cacao beans correspond to Argisol Red-Yellow Eutrophic cambisolic. The dystrophic soils are related to lower dry weight values of cacao beans. The selection of cacao cropping sites by biometric attributes also highlighted the Argisol Red-Yellow Eutrophic cambisolic and Argisol Red-Yellow Dystrophic Cohesive abrupt soils. Understanding the variability of cacao biometric attributes emphasizes the importance of using technologies for achieving sustainable production of cacao with quality. Key words: Theobroma cacao L., plant biometrics, post-harvest, cacao quality. Influência de fatores ambientais sobre atributos biométricos do cacau. Este estudo teve como objetivo investigar a influência de diferentes locais de cultivo de cacau (Theobroma cacao L.), representados por solos e diferentes sistemas de cultivo, sobre atributos biométricos de frutos e amêndoas. Os 12 locais de estudo estão na região cacaueira da Bahia, Brasil, na zona úmida. A variação e as diferenças estatísticas entre as médias destes atributos indicam claramente a influência da interação genótipo e ambiente. Os atributos biomassa úmida de frutos, biomassa úmida de amêndoas com mucilagem e número de amêndoas foram positivamente correlacionados com os locais de cultivo representados pelos solos Argissolo Vermelho-Amarelo Eutrófico cambissólico e Argissolo Vermelho-Amarelo Distrófico coeso abrúptico. Os maiores valores de comprimento e peso seco de amêndoas de cacau correspondem ao Argissolo Vermelho-Amarelo Eutrófico cambissólico. Os solos distróficos estão relacionados com menores valores de peso seco de amêndoa de cacau. A seleção de locais de cultivo de cacau por atributos biométricos também destacou o Argissolo Vermelho-Amarelo Eutrófico cambissólico e Argissolo Vermelho-Amarelo Distrófico coeso abrúptico. A compreensão sobre a variabilidade de atributos biométricos do cacau enfatiza a importância do uso de tecnologias para uma produção sustentável do cacau de qualidade. Palavras-chave: Theobroma cacao L., biometria de plantas, pós-colheita, qualidade do cacau. Recebido para publicação em 16 de fevereiro de 2016. Aceito em 28 de julho de 2016. DOI: 10.21757/0103-3816.2016v28n2p123-140 123 124 Loureiro et al. Introduction Cacao (Theobroma cacao L.) is a tropical species whose agricultural production directly involves approximately six million people (FAO, 2003). Cacao beans are the main raw material for the manufacture of chocolate (Beckett, 2009). However, cacao is still produced by small farmers with a daily income equivalent to US $ 1.25, below the absolute poverty line (Cocoa Barometer, 2015). To reverse this panorama is primarily needed to invest in agricultural education and finance projects that enable producers to add value to the crop. In this actions it is important the joint engagement of cacao science and governmental institutions. For example, among current technological and environmental challenges facing the world cacao production (FAO, 2003; WCF, 2014) can be cited: increased production and productivity by deployment of higher yielding and tolerant cultivars of the main diseases and pests (Pereira, 2001; Ayestas et al., 2013; Dias, 2001; Lopes et al., 2011; Monteiro & Ahnert, 2012; Muniz et al., 2013), phytotechnical use of technology and agricultural mechanization in the stages of planting, pruning, harvesting and postharvest processing (Adzimah; Asiam, 2010; Bentley et al., 2004; Caires et al., 2014; Icco, 2009; Somarriba, 2004; Wood & Lass, 1985), irrigation (Carr & Lockwood, 2011), the use of crop residues (Sodré et al., 2012), intercropping (Icco, 2010; Müller & GamaRodrigues, 2012; Ruf, 2015), and quality certifications related to environmental services and the nutritional quality of the produced beans (Afoakwa, 2010; Amores et al., 2009; Badrie et al., 2015; Jacobi et al., 2015; Obeng & Aguilar, 2015). In this technological context, biometric characterization of pods and cacao beans is important for studies of production and productivity of the crop and may contribute to the selection of genetic materials with better agricultural performance (Alexandre et al., 2015; Dias & Resende, 2001; Kobayashi et al., 2001). The high genetic variability in cacao requires these agronomic components in assessment strategies that should be considered by breeding programs (Lopes et al., 2011). Cacao genotypes tolerant to witches’ broom, disease caused by the fungus Moniliophthora perniciosa (Stahel) Aime & Phillips-Mora, have been propagated as the main strategy to control the disease Agrotrópica 28(2) 2016 (Pereira, 2001; Lopes et al., 2011; Monteiro & Ahnert, 2012). This research is based on Porto Hibrido 16 (PH16), a clonal variety selected from a population of interclonal crosses between cacao trees of the Forastero (Amazon) and Trinitario groups (whose parents are unknown), located at Porto Hibrido farm in São José da Vitória, Bahia, Brazil (Moreau Cruz et al., 2013). The objective of this study was to investigate the influence of different sites, represented by different soils and cropping systems on biometrics attributes of pods and cacao beans. Materials and Methods Sites The 12 study sites are located in the cacao region of Bahia (Table 1) in the humid zone, according to the climatic classification of Thornthwaite: B4r A’, B3r A’, B2r A’, B2r B’, B1r A’, B1r’ A’, B1w A’ (SEI, 2014). These sites (Table 1), are cultivated with PH-16 under different cropping systems, different averages density of shade trees per hectare in different soils according to the Brazilian System of Soil Classification (SiBCS) (2006) and its correspondence in Soil Survey Staff (2006) (Figure 1; Table 1). Soil and cacao pods sampling Soil is the factor that stratifies the environment of the sample source and the experimental units are the cacao trees grafted with PH-16 from wich pods were collected. Each study site with approximately one hectare was divided into three collection areas, characterized by the same soil and same cropping system (Table 2). Pods were collected at a distance of approximately 100 m from the soil identifying points for SiBCS (EMBRAPA, 2006) in the three collection areas (Table 2). Thus, the origin of each sample of pods and beans corresponds to a properly identified and classified soil in each study site, as shown in the example of Table 2. The beans and beans mucilage were obtained from a composed sample of 50 mature pods (Table 2). Cacao sampling occurred on November of 2008, during the second harvest period (August to January). 125 Environmental factors on cacao biometric attributes Table 1 - Summary information about the study sites (soils) cultivated with PH-16 cacao genotype in the cacao region of Bahia, Brazil Average Geographic Soil Cropping City Acronym of Site Soil density of Coordinates Classification Systems the SiBCS1 Taxonomy shade trees/ha 1 13º 40’ 30" S, 39º 14’ 27" W Nilo Peçanha 2 13º 44’ 38" S, 39º 30’ 10" W Gandú PVAd 3 13º 45’ 21" S, 39º 20’ 25" W Piraí do Norte 4 Latosol Yellow Dystrophic cambisolic Hapludox Cacao2 x Rubber Tree3 150 Argisol Red-Yellow Dystrophic tipic Hapludult Cacao2 x Erythrina4 60 PVAd Argisol Red-Yellow Dystrophic abrupt Hapludult Cabruca5 60 13º 46' 07.0" S, Ituberá 39º 17' 52.0"W LAd Latosol Yellow Dystrophic tipic 5 13º 51’ 08" S, 39º 17’ 54" W Ituberá LVAd 6 14º 31’ 14" S, 39º 15’ 45" W Uruçuca 7 14º 51’ 36" S, 39º 14’ 42" W Itabuna 8 14º 51’ 47" S, 39º 06’ 47" W Ilhéus 9 15º 17’ 04" S, 39º 28’ 43" W 10 15º 23’ 08" S, 39º 26’ 04" W 11 15º 23’ 15.1" S, Camacan 39º 25’ 48.6" W 12 16º 29’ 02" S, 39º 23’ 56" W LAd cam Typic Hapludox Cacao2 x Rubber Tree3 350 Latosol Red-Yellow Typic Hapludox Cacao2 x Rubber Tree3 Dystrophic tipic 400 PVAe cam Argisol Red-Yellow Eutrophic cambisolic Hapludalf Cabruca5 50 CXd Cambisol Haplic Dystrophic tipic Dystropept Cabruca5 35 LVAd arg Latosol Red-Yellow Dystrophic argisolic Hapludox Cabruca5 70 Arataca PAd lat Argisol Yellow Dystrophic latosolic Hapludult Cabruca5 35 Camacan PVAd Argisol Red-Yellow Dystrophic tipic Hapludult Cabruca5 35 PVA ali Argisol Red-Yellow Alitic tipic Hapludult Cabruca5 35 Argisol Red-Yellow PVAd coe Dystrophic Cohesive abrupt Hapludult Cacao2 x Rubber Tree3 400 Porto Seguro SiBCS - Brazilian System of Soil Classification (EMBRAPA, 2006). 2Theobroma cacao L. 3Hevea brasiliensis (Willd. Ex Adr de Juss) Muell. Arg. 4Erythrina fusca Lour. 5Cabruca is an ecological system of agroforestry cultivation where cacao trees are grown under native trees of the Atlantic Forest of South of Bahia (Lobão et al., 2007). 1 Table 2 - Summary information of pods and cacao beans sampling Site1 SiBCS2 Cropping Systems Collection Area (100 m radius) Simple Sample Pod (Post-Harvest processing) 3 Pod4 (biometrics) Bean (biometrics) 1 15 90 50 2 15 90 50 3 15 90 50 ... ... ... ... ... ... ... 50 15 90 1 50 15 90 2 12 PVAd coe Cacao x Rubber Tree 50 15 90 3 1 Site: area of approximately 1 hectare. 2SiBCS: Brazilian System of Soil Classification (EMBRAPA, 2006). 3Total number of pods used for cacao post-harvest processing. 4Number of pods used in biometric evaluation. 1 LAd cam Cacao x Rubber Tree Agrotrópica 28(2) 2016 126 Loureiro et al. LEGEND 01 - LAd cam 02 - PVAd 03 - PVAd 04 - LAd 05 - LVAd 06 - PVAe cam 07 - CXd 08 - LVAd arg 09 - PAd lat 10 - PVAd 11 - PVA ali 12 - PVAd coe Cartographic Projecyion UTM System Daturn Sirgas 2000 24S Preparation: Antonio Fontes Faria Filho CEPLAC - CEPEC - SENUP 2015 Figure 1 - Map with the geographic scope of the study sites, represented by 12 soils cultivated with PH-16 cacao genotype. Agrotrópica 28(2) 2016 Environmental factors on cacao biometric attributes Biometric characterization of pods 15 pods chosen of the 50 pods collected were used to represent the three repetitions of each study site to make biometric characterization (Table 2). The attributes: number of beans per pod, pod wet biomass, husk wet biomass, content of pod (beans with mucilage and placenta) wet biomass were evaluated. The wet biomass was determined with a semi-analytical balance. Biometric characterization of beans The physical dimensions of length, width and thickness were determined with a digital pachymeter of 200 mm in the cacao beans. The dry cacao beans weight from moisture between 6-7% was determined by semi-analytical balance. The moisture content of cacao beans was determined gravimetrically, using an oven with air circulation at a temperature of 105°C to constant weight according to the AOAC method 977.10 (2005). Statistical Analysis Statistical procedures used in this study were performed in the R Core Team program (2013). Package ‘stats’: Shapiro-Wilks normality test, Bartlett homoscedasticity test (R Development Core Team, 2013). Package ‘nortest’: Kolmogorov-Smirnov normality test (Lilliefors correction) (Gross & Ligges, 2012). Package ‘MASS’: Box-Cox transformation (Ripley et al., 2015). Package ‘ExpDes’: Analysis of Variance (ANOVA) and Scott-Knott test (Ferreira, Cavalcanti, & Nogueira, 2013). Package ‘Lattice’: Graphics (Sarkar, 2015). Package ‘bpca’: Biplot applied to Principal Component Analysis (Faria et al., 2013). Selection of sites by cacao biometric attributes The scores and sums of scores to select the sites related to the best cacao production characteristics were obtained from the groups generated by multiple mean comparisons by the Scott-Knott test. Results and Discussion Pods biometrics The mean pod wet biomass of 653 g found in this study (Table 3) was similar to the mean of 655.6 g observed in another study with the same cacao genotype (Moreau Cruz et al., 2013). The cacao pod 127 has many morphologically difference between genetic materials (Kobayashi et al., 2001). It is subjected to environmental influences and physiological changes during the various stages of growth and development, when changes occur in their biometric attributes (ALmeida & Dias, 2001; Britto and Silva, 1983; Lopes, 2000; Machado & Almeida, 1989). In this evaluation, it is hypothesized that the higher the wet biomass from ripe pod, the higher the increase in production associated with the weight of beans (Table 3). In a study with the same clone, it was observed a mean cacao husk wet biomass value of 475.8 g (Moreau Cruz et al. 2013), lower than the mean value of 527 g found in this work (Table 3). Biometric attributes of the cacao husk are heavily influenced by physiological maturation processes and it is an important component of pod wet biomass (Britto & Silva, 1983; Machado &Almeida, 1989). Our work highlights the cacao bean wet biomass as an important production component in the biomass partitioning. This study highlights the group of lowest means of husk wet biomass (Table 3), because husk wet biomass is not considered a positive characteristic for post-harvest processing. Pod wet biomass content has an overall mean of 121g (Table 3). In a study conducted with the same genotype (Moreau Cruz et al., 2013) there is no information about the pod wet biomass content of this component, however, adding the means of beans with mucilage (pulp) and placenta wet biomasses gives an approximate value of 175.4 g, nearly 30% higher than the value observed in this study (Table 3). It is considered that the highest cacao pod wet biomass content (Table 3) is associated with an increase in final weight of the dry beans. Cacao pod wet biomass is also related also with environmental conditions like rain, humidity and temperature (Britto & Silva, 1983). In Argisol Red-Yellow Dystrophic Cohesive abrupt production of fresh beans was approximately twice the amount produced in LAd - Yellow Latosol dystrophic tipic (Table 3). This result suggests that the soil eutrophic character was responsible for the difference in production, but environmental factors are also important to define the productivity and quality of cacao beans (Britto & Silva, 1983; Dantas, 2011; Pinto, 2013). It is observed that both ANOVA coefficients of variation for wet biomass attributes, pod, husk and content of pod (Table 3) are approximate values, which Agrotrópica 28(2) 2016 128 Loureiro et al. Table 3 - Summary of analysis of variance, Scott-Knott test and descriptive analysis of the wet biomass of pod, husk and content of pod of PH-16 cacao clone Source DF Pod Husk ---- g ---- Content of Pod1 Mean Square Soil2 Error Total 11 168 179 CV (%) 83255** 19292 74593** 13704 4399** 678 21,2 22,2 21,5 Soil 01 LAd cam 02 PVAd 03 PVAd 04 LAd 05 LVAd 06 PVAe cam 07 CXd 08 LVAd arg 09 PAd lat 10 PVAd 11 PVA ali 12 PVAd coe Mean ± Standard Deviation (n = 15) 681 ± 117 a 720 ± 119 a 573 ± 159 b 611 ± 102 b 757 ± 185 a 711 ± 152 a 674 ± 132 a 535 ± 139 b 599 ± 147 b 742 ± 152 a 673 ± 145 a 564 ± 92 b 541 ± 105 a 591 ± 116 a 459 ± 136 b 520 ± 90 a 621 ± 161 a 555 ± 138 a 547 ± 11 a 420 ± 127 b 467 ± 119 b 631 ± 130 a 545 ± 128 a 432 ± 77 b 134 ± 24 a 126 ± 17 b 108 ± 33 b 83 ± 16 c 127 ± 32 a 151 ± 19 a 123 ± 31 a 109 ± 26 b 130 ± 34 a 108± 26 b 124 ± 27 a 128 ± 20 a Overall mean (n = 180) Mininum Mean ± Standard Deviation Maximum 310 653 ± 152 1060 240 527 ± 135 950 53 121 ± 30 192 Soil: 01 LAd cam - Latosol Yellow Dystrophic cambisolic, 02 PVAd - Argisol Red-Yellow Dystrophic tipic, 03 PVAd - Argisol Red-Yellow Dystrophic abrupt, 04 LAd - Latosol Yellow Dystrophic tipic, 05 LVAd - Latosol Red-Yellow Dystrophic tipic, 06 PVAe cam - Argisol Red-Yellow Eutrophic cambisolic, 07 Cxd - Cambisol Haplic Dystrophic tipic, 08 LVAd arg - Latosol Red-Yellow Dystrophic argisolic, 09 PAd lat - Argisol Dystrophic latosolic, 10 PVAd - Argisol Red-Yellow Dystrophic tipic, 11 PVA ali - Argisol Red-Yellow Alitic tipic, 12 PVAd coe - Argisol Red-Yellow Dystrophic Cohesive abrupt. DF - Degrees of Freedom. CV - Coefficient of Variation. Significance levels by test F: (**) = 1% of error. 1 can display a standard for the variability of biometric attributes of PH-16. The overall mean of the beans with mucilage wet biomass corresponds to 92 g (Table 4), lower than the mean of 148,4 g observed in another study with PH-16 (Moreau Cruz et al., 2013). The variation of the beans with mucilage wet biomass (CV = 24.6%) (Table 4) can be related to biometrics and physiological differences of the pod at harvest, particularly related to the hydration and beans mucilage. The final weight of the dry cacao beans is related to wet biomass of in natura beans, being an important production component (Beckett, 2009; Britto and Silva, 1983; Engels et al., 1980; Lopes, 2000; Sánchez et al., 1996). Agrotrópica 28(2) 2016 Placenta wet biomass content observed in this study corresponds to 29 g (Table 3) was higher than the mean value of 27 g for the same genotype (Moreau Cruz et al., 2013). For processing, the placenta is removed so that the beans are fermented with mucilage (Wood, 2001). This study highlights the group of lowest means of placenta wet biomass (Table 3), considering that the higher weight of placenta in natura suggests that the pods are not in the ripening ideal point to postharvest process. In the same cacao clone (Moreau Cruz et al., 2013) was observed an mean of 43 beans per pod, approximate value to the mean of 40 beans observed in this study (Table 4). This study highlights the group 129 Table 4 - Summary of analysis of variance, Scott-Knott test and descriptive analysis of the beans with mucilage and placenta wet biomass, and number of beans of PH-16 cacao clone Beans with mucilage Placenta ------- g -------- Number of beans Mean Squared Soil Error Total 1 11 168 179 CV (%) Soil 4045** 515 165** 37 179** 58 24,6 21,2 18,8 Mean ± Standard Deviation (n = 15) 1 01 LAd cam 02 PVAd 03 PVAd 04 LAd 05 LVAd 06 PVAe cam 07 CXd 08 LVAd arg 09 PAd lat 10 PVAd 11 PVA ali 12 PVAd coe 105 ± 23 b 95 ± 18 b 82 ± 28 c 56 ± 15 d 93 ± 27 b 120 ± 15 a 95 ± 28 b 86 ± 23 c 103 ± 28 b 75 ± 21 c 96 ± 24 b 104 ± 15 b 30 ± 4 a 31 ± 5 a 25 ± 7 b 27 ± 4 b 33 ± 8 a 31 ± 7 a 29 ± 6 a 24 ± 7 b 26 ± 6 b 33 ± 7 a 30 ± 6 a 25 ± 5 b 42 ± 7 a 39 ± 8 b 38 ± 9 b 33 ± 8 b 40 ± 7 b 46 ± 4 a 40 ± 10 b 41 ± 10 a 38 ± 8 b 39 ± 9 b 40 ± 6 b 47 ± 4 a Overall mean (n = 180) Mininum Mean ± Standard Deviation Maximum 30 92 ± 27 153 13 29 ± 7 45 18 40 ± 8 54 1 Soil: 01 LAd cam - Latosol Yellow Dystrophic cambisolic, 02 PVAd - Argisol Red-Yellow Dystrophic tipic, 03 PVAd - Argisol Red-Yellow Dystrophic abrupt, 04 LAd - Latosol Yellow Dystrophic tipic, 05 LVAd - Latosol Red-Yellow Dystrophic tipic, 06 PVAe cam - Argisol Red-Yellow Eutrophic cambisolic, 07 Cxd - Cambisol Haplic Dystrophic tipic, 08 LVAd arg - Latosol Red-Yellow Dystrophic argisolic, 09 PAd lat - Argisol Dystrophic latosolic, 10 PVAd - Argisol Red-Yellow Dystrophic tipic, 11 PVA ali - Argisol Red-Yellow Alitic tipic, 12 PVAd coe - Argisol Red-Yellow Dystrophic Cohesive abrupt. DF - Degrees of Freedom. CV - Coefficient of Variation. Significance levels by test F: (**) = 1% of error. of higher means of number of beans, because the number of beans is an important component of production used in cacao breeding (Beckett, 2009; Britto & Silva, 1983; Engels et al., 1980; Sánchez et al., 1996). The statistical differences of the ANOVA F test showed a variation of around 20% between biometric variables of the PH-16 cacao clone as a function of cultivation sites represented by soil (Table 4). However, these statistical differences (Table 4) can be related to several factors that were not considered in this study, for example, pod yield per plant (nutrient partitioning and biomass), competition of light, water and nutrients between plants, plant age, fertility and soil management, disease and pest attacks, and other genetic, environmental and phytotechnical factors that interfere with the physiological and cacao dendrometric characteristics (Almeida & Valle, 2007; Britto and Silva, 1983; Engels et al., 1980; Lopez Baez, 1995; Machado and Almeida, 1989; Sánchez et al., 1996). Positive correlations between pod wet biomass and husk wet biomass (r = 0.98) (Figure 2A) and with the placenta wet biomass (r = 0.98) (Figure 2C) were found. A positive correlation between the pod wet biomass and content of pod wet biomass (r = 0.58) (Figure 2B) was also observed. There is biological evidence that the husk wet biomass of PH-16 correspond to approximately 80% of the total pod wet biomass (Table 3). The high correlation between these attributes (Figure 2A) may indicate that both are subject to the same influence of Agrotrópica 28(2) 2016 130 Loureiro et al. Figure 2 - Correlation between cacao pods biometrics attributes of PH-16 clone (r = Pearson correlation coefficient; **Significance at 1% level; n = 180). environmental factors related to the physiology of growth. This information is important for breeding and nutritional management of cacao trees, because they relate to the partition of nutrients and biomass. These biometric attributes may be decisive for the choice of genotypes with different agricultural potential, particularly regarding cacao quality. The husk wet biomass also showed a positive correlation with placenta wet biomass (r = 0.95) (Figure 2D). With regard to partitioning of nutrients the husk has largely of the nutrients directed by cacao trees is growing of the pod, showing the large export of nutrients for this plant organ (drain) (Pinto, 2013; Sodré et al., 2012). The placenta is the conductive part of Agrotrópica 28(2) 2016 pod nutrients to the beans (Muniz et al., 2013), and a high correlation with the husk (Figure 2D) may indicate that both are important in the partition of nutrients and biomass in cacao trees. Primarily, the content of pod wet biomass is positively correlated with beans with mucilage wet biomass (r = 0.97) (Figure 3A), followed by positive correlation with number of beans (Figure 3C) (r = 0.67) and, finally, with placenta wet biomass (r = 0.59) (Figure 3B). It was also observed that beans with mucilage wet biomass was positively correlated with the number of beans (r = 0.70) (Figure 3D). According to this correlation (Figure 3D), pods with more beans tend to have higher beans with mucilage wet biomass 131 Environmental factors on cacao biometric attributes Figure 3 - Correlation between cacao pods biometrics attributes of PH-16 clone (r = Pearson correlation coefficient; **Significance at 1% level; n = 180). as verified in Tables 3 and 4, following an opposite relationship with respect to the groups of means of wet biomass of the pod and husk. However, because they are partial values of the pod wet biomass total value; these interpretations were also based on the graphical analysis of the correlations (Table 3). Among some attributes, there was wide graphic dispersion of sample observations and lower correlation coefficients (Figures 3B; 3C). This can be explained by the great variability of biometric in cacao pods (Lopez et al., 2011). In higher plants, both dry matter as well as nutrient partitioning are phenomena of growth and development closely related to genetic factors (Taiz & Zeiger, 2013). The husk of the Forastero group is thicker than of the Criollo group, and the Trinitarian cacao husk shows intermediate thickness (Bartley, 2005). PH-16 is a hybrid of Forastero with Trinitarian groups (Moreau Cruz et al., 2013). The value of the ratio pod wet biomass and number of beans in Cacao Common (Forastero group) approaches 15 (Loureiro, 2012), whereas in this study, the value observed for the ratio in PH-16 was approximately 17 (Tables 3 and 4). The ratio between pod wet biomass and number of seeds may mean plant nutrients invested in biomass conversion is particular directed to the seeds (beans) (Marenco & Lopes, 2009; Agrotrópica 28(2) 2016 132 Loureiro et al. Taiz & Zeiger, 2013). However, not only the number of beans must be considered an important component of plant production, but also the bean weight, especially after processing, when they will be marketed (Garcia, 1985; Garcia, 1973; Santana, 1981). Principal components analysis (PCA) biplots of cacao pods biometric attributes of PH-16 clone are shown in Figures 4 and 5. It is observed a correlation structure similar to the structure of bivariate correlations in Figures 2 and 3. PC2 (41.0,8%) A: Sites identified by soils (SiBCS) PC1 (54,74%) PC2 (41.0,8%) B: Sites identified by cropping systems PC1 (54,74%) Figure 4 - Principal Component Analysis Biplots. Factors indicate the relative weight of the variables on the axes. Cacao pods biometric attributes of PH-16 clone: pod wet biomass (POD), husk wet biomass (HUS), content of pod (beans with mucilage and placenta) wet biomass (CON), beans with mucilage wet biomass (BWM), placenta wet biomass (PLA), number of beans (NOB). Places represented by soils (Brazilian System of Soil Classification - SiBCS): Latosol Yellow Dystrophic cambisolic (1_LAd cam), Argisol Red-Yellow Dystrophic tipic (2_PVAd), Argisol Red-Yellow Dystrophic abrupt (3_PVAd), Latosol Yellow Dystrophic tipic (4_LAd), Latosol Red-Yellow Dystrophic tipic (5_LVAd), Argisol Red-Yellow Eutrophic cambisolic (6_PVAe cam), Cambisol Haplic Dystrophic tipic (7_Cxd), Latosol Red-Yellow Dystrophic argisolic (8_LVAd arg), Argisol Dystrophic latosolic (9_PAd lat), Argisol Red-Yellow Dystrophic tipic (10_PVAd), Argisol Red-Yellow Alitic tipic (11_PVA ali), Argisol Red-Yellow Dystrophic Cohesive abrupt (12_PVAd coe). Numbered soils according to the longitudinal direction North-South. Agrotrópica 28(2) 2016 Environmental factors on cacao biometric attributes 133 PC2 (28,88%) A: Observations identified by soil classes PC1 (64,68%) PC2 (28,88%) B: Observations identified by cropping systems PC1 (64,68%) Figure 5 - Principal Component Analysis Biplots. Factors indicate the relative weight of the variables on the axes. Cacao pods biometric attributes of PH-16 clone: pod wet biomass (POD), husk wet biomass (HUS), content of pod (beans with mucilage and placenta) wet biomass (CON), beans with mucilage wet biomass (BWM), placenta wet biomass (PLA), number of beans (NOB). Observations identified by soil classes (A) and cropping systems (B). Agrotrópica 28(2) 2016 134 Loureiro et al. Table 5 is a summary of the PCA of cacao pods biometric attributes of PH-16 clone explained by Biplot graph. The Biplots represent the cacao pods biometric attributes of PH-16 clone, which vary according to the study sites represented by soils and their classes, different cropping systems, the average density of shade trees per hectare and the geographic coordinates (Figures 4 and 5). Graphs A and B in Figure 4, have the same structure between variables and objects because they are the same ACP (Table 5). Graphs A and B in Figures 5 also have the same structure between variables and objects (Table 5). The objects shown in A and B were renamed for interpretation purposes (Figures 4 and 5). The main components represented in Biplot graphs A and B of Figure 4 have eigenvalue higher than 1, and retains 96% of the total variance of the data for interpretation based on the means (Table 5). The CP represented by the graphs A and B of Figure 5 also have eigenvalue higher than 1 (one), retaining about 94% of the total variance of the data (Table 5). The variables number of beans (NOB), content of pod wet biomass (CON) and beans with mucilage wet biomass (BWM) are positively correlated (Figure 3). Graphically, the site 06 - PVAe cam (Argisol RedYellow Eutrophic cambisolic) in Cacao Cabruca system (Cab), with an average density of 50 shade trees per hectare, and the site 12 - PVAd coe (Argisol RedYellow Dystrophic Cohesive abrupt) in Cacao x Rubber Tree intercropping system, with an average density of 400 shade trees, are positively correlated with these variables (Figure 3). Besides the natural fertility, soil management conditions, such as correction of acidity and fertilization, are related to the eutrophic character of the soil; therefore, the correlation of these agronomic characteristics of cacao pods (NOB, CON and BWM) may be associated with good soil management. Site 04 - Lad (Latosol Yellow Dystrophic tipic) showed a negative correlation with these same variables (NOB, CON and BWM) (Figure 3). Sites 10 - PVAd (Argisol Red-Yellow Dystrophic tipic), 05 - LVAd (Latosol Red-Yellow Dystrophic tipic) and 02 - PVAd (Argisol Red-Yellow Dystrophic tipic) were positively correlated with the attributes pod (POD), husk (HUS) and placenta (PLA) wet biomasses. These soils have in common the "Dystrophic tipic" character, showing that low soil fertility favors biometric attributes as HUS and PLA, which are not as important as the attributes related to the production of beans (NOB, CON and BWM). All observations from biometric variables were also explored in relation to the three soil classes Argisols (Arg), Cambisols (Cam) and Latosols (Lat), and also in relation to the three types of cropping systems Cacao Cabruca (Cab), Cacao x Erythrina Tree (CxE), and Cacao x Rubber Tree (CxR) (Figure 5). Clusters between soils related classes (SiBCS) with cropping systems or longitudinal arrangement of the geographic coordinates (Figures 4 and 5) were not observed. But Table 5 - Summary of Principal Component Analysis of cacao pods biometric attributes of PH-16 clone Summary POD HUS CON BWM PLA NOB Eigenvalue Retained Variance Accumulated Variance 12 soils and cropping systems with average density of shade trees All observations by soil types and cropping systems PC1 PC2 PC1 PC2 -0.49 -0.42 -0.41 -0.34 -0.50 -0.25 0.30 0.41 -0.41 -0.49 0.28 -0.51 -0.46 -0.41 -0.44 -0.38 -0.46 -0.28 -0.32 -0.43 0.34 0.47 -0.31 0.52 6,01 0.55 0.55 5.21 0.41 0.96 26.36 0.65 0.65 17.61 0.29 0.94 POD – pod wet biomass (g), HUS – huks wet biomass (g), CON – content of pod (beans with mucilage and placenta) wet biomass (g), BWM - beans with mucilage wet biomass (g), PLA – placenta wet biomass (g), NOB – number of beans. PC – Principal Component. Agrotrópica 28(2) 2016 135 Environmental factors on cacao biometric attributes even if the level of detail of this study (Figures 4 and 5) were not sufficient to explain these differences directly, indirectly they indicate that biometric attributes of cacao pods are under the influence of genotype x environment interaction (Almeida & Valle, 2007; Dias & Kageyama, 1985; Engels et al., 1980; Garcia, 1973; Icco, 2008; Machado and Almeida, 1989; Monteiro et al., 2011; Sánchez et al., 1996). Beans biometrics The highest mean of dry cacao beans length of 18 mm correspond to site 6 represented by Argisol RedYellow Eutrophic cambisolic (Table 6); this site is also related with the highest mean of cacao beans dry weight of 1.4 g (Table 6). As already discussed, the eutrophic character of the soil associated with good fertility conditions is directly correlated with the production of components related to cacao beans, both in wet biomass (Tables 3 and 4; Figure 4) and dry biomass (Table 6). It was observed that the lowest means of dry cacao beans weight are related to soils with dystrophic character (Table 6), showing the relationship between soil fertility and production of cacao beans. The variation in values of dry cacao beans length shown in this study (Table 6) is similar to the variation observed in another study employing the same genotype (Moreau Cruz et al., 2013). The mean of dry cacao beans width corresponds to 11 mm, this value was lower the mean of 14.3 mm observed in another study with PH-16 clone (Moreau Cruz et al., 2013). Our study analyzed dry cacao beans and another study analyzed in natura cacao beans, the absence of post-harvest Table 6 - Summary of analysis of variance, Scott-Knott test and descriptive analysis of the biometrics attributes of dry cacao beans of PH-16 clone Length Source DF Width ------- mm -------- Thickness Dry Weight g Mean Squared Soil Error Total 1 11 1068 1079 CV (%) 14,9** 4,8 10** 2,6 UR UR 4,8** 1,2 12,7 14,8 UR 17,1 Soil1 01 LAd cam 02 PVAd 03 PVAd 04 LAd 05 LVAd 06 PVAe cam 07 CXd 08 LVAd arg 09 PAd lat 10 PVAd 11 PVA ali 12 PVAd coe Mean ± Standard Deviation (n = 15) 17.6 ± 2.1 a 17.0 ± 2.2 b 16.6 ± 2.1 b 17.2 ± 2.1 b 16.9 ± 2.3 b 18.0 ± 2.2 a 17.6 ± 2.1 a 17.1 ± 2.4 b 17.4 ± 2.1 a 17.8 ± 2.4 a 17.5 ± 2.3 a 17.6 ± 2.1 a 11.0 ± 1.4 b 10.9 ± 1.9 b 10.3 ± 1.4 c 10.9 ± 1.4 b 10.6 ± 1.5 c 11.4 ± 1.8 a 11.1 ± 1.5 b 11.6 ± 1.8 a 11.0 ± 1.6 b 11.2 ± 1.8 a 10.9 ± 1.7 b 11.0 ± 1.7 b 6.5 ± 1.2 6.1 ± 1.0 6.1 ± 0.9 6.1 ± 1.1 6.1 ± 1.0 6.7 ± 1.0 5.9 ± 1.3 6.0 ± 1.2 6.1 ± 1.1 6.2 ± 1.0 6.3 ± 1.1 6.1 ± 0.9 1.3 ± 0.2 b 1.2 ± 0.2 c 1.2 ± 0.2 c 1.2 ± 0.2 c 1.2 ± 0.2 c 1.4 ± 0.2 a 1.2 ± 0.2 c 1.2 ± 0.2 c 1.2 ± 0.2 c 1.3 ± 0.2 b 1.3 ± 0.3 b 1.2 ± 0.2 c Overall mean (n = 180) Minimum Mean ± Standard Deviation Maximum 10.8 17.4 ± 2.2 23.8 7.3 11.0 ± 1.7 15.4 3.3 6.2 ± 1.1 9.6 0.67 1.2 ± 0.2 1.93 Soil: 01 LAd cam - Latosol Yellow Dystrophic cambisolic, 02 PVAd - Argisol Red-Yellow Dystrophic tipic, 03 PVAd - Argisol Red-Yellow Dystrophic abrupt, 04 LAd - Latosol Yellow Dystrophic tipic, 05 LVAd - Latosol Red-Yellow Dystrophic tipic, 06 PVAe cam - Argisol Red-Yellow Eutrophic cambisolic, 07 Cxd - Cambisol Haplic Dystrophic tipic, 08 LVAd arg - Latosol Red-Yellow Dystrophic argisolic, 09 PAd lat - Argisol Dystrophic latosolic, 10 PVAd - Argisol Red-Yellow Dystrophic tipic, 11 PVA ali - Argisol Red-Yellow Alitic tipic, 12 PVAd coe - Argisol Red-Yellow Dystrophic Cohesive abrupt. DF - Degrees of Freedom. CV - Coefficient of Variation. UN – Unrealized. Significance levels by test F: (**) = 1% of error. 1 Agrotrópica 28(2) 2016 136 Loureiro et al. and all biochemical transformations related were the probable causes of the differences of these values (Table 6). The overall mean of dry cacao beans weight with 67% moisture corresponds to 1.24 g (Table 6). This values are highest than the value of 1.0 g (for samples of 100 cacao beans) recommended by industry (CCCA, 1984), also is recommended that not more than 12% of the cacao beans should have a variance larger or smaller than 1/3 of mean dry weight. However, these and other requirements of the chocolate industry were determined without considering the genetic diversity and biometric variation of cacao genotypes in different growing conditions. Hybrid cacao beans can vary widely regarding their biometric dimensions (Beckett, 2009; Engels et al. 1980; Almeida, 2001; Dias and Rezende, 2001; Sánchez et al., 1996). This result indicates that the dry cacao beans weight of clone PH-16 was slightly r = 0.63** A influenced by the soil in which it was being cultivated. The length, width and dry weight and are the seed of morphological and agronomic characteristics as descriptors used in breeding programs to differentiate the cacao genotypes, and are also subject to the degree of development and ripening of pods (Bekele & Butler, 2000; Engels et al., 1980; Lopes, 2000; Bartley, 2005; Mattietto, 2001; Sánchez et al., 1996). The cacao beans length showed positive correlation with cacao beans width (r = 0.63) (Figure 6A) and bean dry weight (r = 0.58) (Figure 6B). Besides the length, cacao beans thickness also showeds positive correlation with cacao bean dry weight (r = 0.54) (Figure 6C). The dry weight of dry cacao beans is an important attribute for the final marketing of cacao (Beckett, 2009; WCF, 2014), and the correlations shown Figure 6, the attributes that represent the cacao beans dimensions are directly related to dry weight. r = 0.58** B Figure 6 - Correlation between dry cacao beans biometrics attributes of PH-16 clone (r = Pearson correlation coefficient; **Significance at 1% level; n = 1080). Agrotrópica 28(2) 2016 137 Environmental factors on cacao biometric attributes necessary to emphasize that this evaluation does not provide the best performance between means that belong to the same group generated by Scott-Knott test because regardless of higher or lower values within the group, the averages were classified with the values 0 (zero) or 1 (one) (Table 7). Selection of cropping sites by cacao biometric attributes The evaluation criteria used in the selection of cacao cropping sites were (Table 7): means of attributes interpreted as good production characteristics received a score of 1 (one), and the means unselected received a score of 0 (zero). For example, values from groups with the highest means of beans with mucilage wet biomass and number of beans were converted to 1, and the other values to 0 (zero). Another example of interpretation, values from groups with the lower means of husk and placenta wet biomass were converted to 1, and the other values to 0 (zero) (Table 7). According to the sum of the scores of cacao biometrics attributes of PH-16 clone selected from the Scott-Knott test, sites 06 - PVAe cam and 12 - PVAd coe showed the best performance with 7 and 5 the best means of the studied biometric attributes, respectively (Table 7). Sites 02 - PVAd and 04 - LAd have the lowest performances, both with a score of 1 point (Table 7). This method of evaluation (selection by scores) does not replace the multivariate statistical methods, but, it proved to be able to differentiate levels of study of factor (source of variation) showing similar results of nivariate (Tables 3, 4 and 6) and multivariate analyzes (Figures 4 and 5) objectively. However, it is Conclusions Cacao biometric attributes of pods and beans were influenced by different cropping sites. Attributes content of pod (beans with mucilage and placenta) wet biomass, beans with mucilage wet biomass and number of beans were positively correlated with the cultivation sites represented by Argisol Red-Yellow Eutrophic cambisolic and Argisol Red-Yellow Dystrophic Cohesive abrupt soils. The highest values of length and dry weight of dry cacao beans correspond to Argisol Red-Yellow Eutrophic cambisolic. The dystrophic soils are related to lower dry weight values of cacao beans. The selection of cacao cropping sites by biometric attributes also highlighted the Argisol Red-Yellow Eutrophic cambisolic and Argisol Red-Yellow Dystrophic Cohesive abrupt soils. Table 7 - Selection of cacao cropping sites by mean groups of biometrics attributes generated by the Scott-Knott test Soil1 01 LAd cam 02 PVAd 03 PVAd 04 LAd 05 LVAd 06 PVAe cam 07 CXd 08 LVAd arg 09 PAd lat 10 PVAd 11 PVA ali 12 PVAd coe Pod2 Dry Bean3 POD HUS CON BWM PLA NOB LEN 1 1 0 0 1 1 1 0 0 1 1 0 0 0 1 0 0 0 0 1 1 0 0 1 1 0 0 0 1 1 1 0 1 0 1 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 1 1 0 0 0 1 1 0 0 1 1 0 0 0 0 1 0 1 0 0 0 1 1 0 0 0 0 1 1 0 1 1 1 1 WID 0 0 0 0 0 1 0 1 0 1 0 0 Sum WEI 0 0 0 0 0 1 0 0 0 0 0 0 4 1 2 1 2 7 3 4 4 3 3 5 1 Soil: 01 LAd cam - Latosol Yellow Dystrophic cambisolic, 02 PVAd - Argisol Red-Yellow Dystrophic tipic, 03 PVAd - Argisol RedYellow Dystrophic abrupt, 04 LAd - Latosol Yellow Dystrophic tipic, 05 LVAd - Latosol Red-Yellow Dystrophic tipic, 06 PVAe cam Argisol Red-Yellow Eutrophic cambisolic, 07 Cxd - Cambisol Haplic Dystrophic tipic, 08 LVAd arg - Latosol Red-Yellow Dystrophic argisolic, 09 PAd lat - Argisol Dystrophic latosolic, 10 PVAd - Argisol Red-Yellow Dystrophic tipic, 11 PVA ali - Argisol Red-Yellow Alitic tipic, 12 PVAd coe - Argisol Red-Yellow Dystrophic Cohesive abrupt. 2Pod: POD – pod wet biomass (g), HUS – husk wet biomass (g), CON – content of pod (beans with mucilage and placenta) wet biomass (g), BWM – beans with mucilage wet biomass (g), PLA – placenta wet biomass (g), NOB – number of beans per pod. 3Dry Beans: LEN – Bean length (mm), WID – Bean width (mm), WEI – Bean dry weight (g). Agrotrópica 28(2) 2016 138 Loureiro et al. Understanding the variability of cacao biometric attributes emphasizes the importance of using technologies for achieving sustainable production of cacao quality. Acknowledgments This paper is part of the project “Linking soil quality and cacao quality in Bahia, Brazil”. To run fundamental steps of this project, the Researcher Quintino R. Araujo (coordinator) was supported by the Brazilian National Council for Scientific and Technological Development (CNPq) with a Postdoctoral fellowship. Literature Cited ADZIMAH, S. K.; ASIAM, E. K. 2010. Design of a cocoa pod splitting machine. Research Journal of Applied Sciences 2:622-634. AFOAKWA, E. O. 2010. Chocolate science and technology. Oxford, Wiley-Blackwell. 296p. ALEXANDRE, R. S. et al. 2015. Caracterização de frutos de clones de cacaueiros na região litorânea de São Mateus, ES. Revista Brasileira de Engenharia Agrícola e Ambiental 19:785-790. ALMEIDA, C. M. V. C. 2001. Ecologia de populações naturais. In: Dias, L. A. S. D. ed. Melhoramento genético do cacaueiro. Viçosa, MG, FUNAPE - UFG. pp.129-162. ALMEIDA, A-A. F.; VALLE, R. R. 2007. Ecophysiology of the cacao tree. Brazilian Journal of Plant Physiology 19:425-448. ALMEIDA, C. M. V. C.; DIAS, L. A. S. D. 2001. Recursos genéticos. In: Dias, L. A. S. D. ed. Melhoramento genético do cacaueiro. Viçosa, MG, FUNAPE - UFG. pp.163-216. AMORES, F. et al. 2009. Entorno ambiental, genética, atributos de calidad y singularización del cacao en el nor oriente de la provincia de esmeraldas. Quevedo, Los Ríos, Equador, INIAP. Boletín Técnico nº 135. 120p. ASSOCIATION OF OFFICIAL ANALYTICAL CHEMISTS (AOAC). 2005. In: Horwitz, W.; Latimer J. R. G. W. eds. Official Methods of Analysis of the Association of Official Analytical Chemists. Gaithersburg, Maryland, AOAC. Agrotrópica 28(2) 2016 AYESTAS, E. et al. 2013. Caracterización de árboles promisorios de cacao en fincas orgánicas de Waslala, Nicaragua. Agroforestería en las Américas (Costa Rica) 49:18-25. BADRIE, N. et al. 2015. Cocoa agronomy, quality, nutritional, and health aspects. Critical Reviews in Food Science and Nutrition 55(5):620-659. BAROMETER, C. 2015. Cocoa barometer 2015. Available. in: http://www.cocoabarometer.org/ International_files/Cocoa Barometer 2015 USA.pdf>. Acessed in: January 08, 2016. BARTLEY, B. G. D. 2005. The genetic diversity of cacao and its utilization. Wallingford, UK, CABI Publishing. 341p. BECKETT, S. T. 2009. Industrial chocolate manufacture and use. London, Chapman and Hall. 720p. BEKELE, F. L.; BUTLER, D. R. 2000. Proposed short list of cocoa descriptors for characterization. In: Working procedures for cocoa germplasm evaluation and selection. Rome, CFC/CCO/IPGRI Project Workshop. pp.41-48. BENTLEY, J. W. et al. 2004. Neighbor trees: shade, intercropping, and cacao in Ecuador. Human Ecology 32 (2) :241-269. BRITTO, I. C.; SILVA, C. P. 1983. Medidas biométricas do fruto do cacaueiro durante seu desenvolvimento. Sitientibus (Brasil) 2(3):59-66. CAIRES, S. A. et al. 2014. Spatial analysis for management zone delineation in a humid tropic cocoa plantation. Precision Agriculture 16(2):129147. CARR, M. K. V.; LOCKWOOD, G. 2011. The water relations and irrigation requirements of cocoa (Theobroma cacao L.): A review. Experimental Agriculture 47(4):653-676. COCOA CHOCOLATE AND CONFECTIONERY ALLIANCE (CCCA). 1984. Cocoa beans: chocolate manufactures' quality requirements. London, CCCA. 19p. DANTAS, P. A. S. 2011. Relação entre fertilidade do solo e nutrição do cacaueiro no Sul da Bahia. Dissertação Mestrado. Ilhéus, BA, UESC. 85p. DIAS, L. A. S. 2001. Novos rumos do melhoramento. In: Dias, L. A. S. ed. Melhoramento genético do cacaueiro. Viçosa, MG, FUNAPE - UFG. pp. 531578. Environmental factors on cacao biometric attributes DIAS, L. A. S.; KAGEYAMA, P. Y. 1985. Combiningability for cacao (Theobroma cacao L.) yield components under southern Bahia conditions. Theoretical and Applied Genetics 90:534-541. DIAS, L. A. S.; RESENDE, M. D. V. 2001. Estratégias e métodos de seleção. In: Dias, L. A. S. ed. Melhoramento genético do cacaueiro. Viçosa, MG, FUNAPE - UFG. pp.217-287. EMPRESA BRASILEIRA DE PESQUISA AGROPECUÁRIA (EMBRAPA). 2006. Sistema Brasileiro de Classificação de Solos. 2. ed. Rio de Janeiro, RJ, EMBRAPA. 306p. ENGELS, M. M. et al. 1980. Cacao descriptors, their states and modus operandi. Turrialba (Costa Rica) 30 (2):209-218. FOOD AND AGRICULTURE ORGANIZATION OF THE UNITED NATIONS (FAO) 2003. Agricultural commodities: profiles and relevant WTO negotiating issues. Rome, Italia, Economic and Social Development Department. 89p. FARIA, J. C. et al. 2013. BPCA: Biplot of multivariate data based on principal components analysis. Available in: <http://cran.r-project.org/web/ packages/bpca/bpca.pdf>. Accessed in: March 13, 2014. FERREIRA, E. B. et al. 2013. ExpDes: Experimental Designs package. Available In: <http://cran.rproject.org/web/packages/ExpDes/ExpDes.pdf>. Accessed in: January 20, 2014. GARCIA, J. J. S. 1985. Beneficiamento, armazenamento e padronização do cacau. In: COMISSÃO EXECUTIVA DO PLANO DA LAVOURA CACAUEIRA (CEPLAC). ed. Sistema de produção do cacaueiro na Amazônia Brasileira. Belém, Pará, CEPLAC /SUPOR. pp.86-104. GARCIA, J. R. 1973. Estudo de alguns índices de crescimento e produtividade para seleção juvenil em híbridos de cacau. Dissertação Mestrado. Turrialba, Costa Rica. IICA. 89p. GROSS, J.; LIGGES, U. 2012. Nortest: Tests for normality. Available in: <http://cran.r-project.org/ package=nortest>. Accessed in: January 11, 2014. INTERNATIONAL COCOA ORGANIZATION (ICCO). 2008. Manual of best known practices in cocoa production. Moscow, Consultative Board on the World Cocoa Economy. CB/16/2Rev.1. 9p. INTERNATIONAL COCOA ORGANIZATION (ICCO). 139 2009. Guidelines on best known practices in the cocoa value chain. Moscow, Consultative Board on the World Cocoa Economy. CB/16/2Rev.1. 10p. INTERNATIONAL COCOA ORGANIZATION (ICCO). 2010. Inventory of diversification practices on cocoa farms. London, United Kingdom, ICCO. 9p. JACOBI, J. et al. 2015. Farm resilience in organic and nonorganic cocoa farming systems in Alto Beni, Bolivia. Agroecology and Sustainable Food Systems 39(7):798-823. KOBAYASHI, R. S. et al. 2001. Caracterização morfológica de frutos e sementes de clones de cacaueiros (Theobroma cacao L.) silvestres da Amazônia Brasileira. Belém, Pará, CEPLAC/ SUPOR. Technical Bulletin nº19. 58p. LOBÃO, D. E. et al. 2007. Cacau cabruca - sistema agrossilvicultural tropical. In: Valle, R.R. ed. Ciência, tecnologia e manejo do cacaueiro. Itabuna, Gráfica e Editora Vital. pp.290-323. LOPES, A. S. 2000. Estudo químico e nutricional de amêndoas de cacau (Theobroma cacao L.) e cupuaçu (Theobroma grandiflorum Schum) em função do processamento. Dissertação Mestrado. Campinas, SP, Unicamp. 130p. LOPES, U. V. et al. 2011. Cacao breeding in Bahia, Brazil - strategies and results. Crop Breeding and Applied Biotechnology (S1):73-81. LOPEZ BAEZ, O. 1995. Caracteristicas del fruto del germoplasma de cacao, Theobroma cacao L. seleccionado en Rosario Izapa, Chiapas. Agricultura Técnica En México 21 (2):127-137. LOUREIRO, G. A. H. A. 2012. Atributos qualitativos de solo e amêndoas de cacau comum: revisão, análises e interpretação de relações. Monografia em Engenharia Agronômica. Ilhéus, BA, UESC. 87p. MACHADO, R. C. R.; ALMEIDA, H. A. 1989. Estimativa do volume do fruto do cacau. Revista Brasileira de Fisiologia Vegetal 1: 115-117. MARENCO, R. A.; LOPES, N. F. 2009. Fisiologia vegetal: fotossíntese, respiração, relações hídricas e nutrição mineral. Viçosa, MG, UFV. 451p. MATTIETTO, R. A. 2001. Estudo das transformações estruturais e fisco-químicas durante o processo fermentativo em amêndoas de cacau (Theobroma cacao L.) e cupuaçu (Theobroma grandiflorium Schum). Dissertação Mestrado. Campinas, SP, Unicamp. 164p. Agrotrópica 28(2) 2016 140 Loureiro et al. MONTEIRO, W. R.; AHNERT, D. 2012. Melhoramento genético do cacaueiro. In: Valle, R.R. ed. Ciência, tecnologia e manejo do cacaueiro. Brasília, DF, MAPA. pp.1-29. MONTEIRO, W. R. et al. 2011. Manejo de cacauais seminais pela poda e correção do sombreamento permanente. Ilhéus, BA, MAPA/CEPLAC/CEPEC. Technical Bulletin nº 201. 30p. MOREAU CRUZ, J. F. et al. 2013. Assessment of the fermentative process from different cocoa cultivars produced in Southern Bahia, Brazil. African Journal of Biotechnology 12(33):5218-5225. MÜLLER, M. W.; GAMA-RODRIGUES, A. C. 2012. Sistemas agroflorestais com cacaueiro. In: Valle, R.R. ed. Ciência, tecnologia e manejo do cacaueiro. Brasília, DF, MAPA. pp.407-435. MUNIZ, M. R. A. et al. 2013. Exportação de nutrientes pelos frutos de cacau de diferentes clones cultivados nas Fazendas Reunidas Vale do Juliana. Addubare RR Agroflorestal, 24: 5-9. Available in: <http://www.rragroflorestal.com.br/downloads/ informativos/addubare24.pdf>. Accessed in: Febrero 14, 2015. OBENG, E. A.; AGUILAR, F. X. 2015. Marginal effects on biodiversity, carbon sequestration and nutrient cycling of transitions from tropical forests to cacao farming systems. Agroforestry Systems 89 (1):19-35. PEREIRA, A. B. 2001. Melhoramento clonal. Dias, L. A. S. ed. In: Melhoramento genético do cacaueiro. Viçosa, MG, FUNAPE - UFG. pp. 261-384. PINTO, F. C. 2013. Fertilidade do solo e partição de nutrientes em cacaueiros. Dissertação Mestrado. Ilhéus, BA, UESC. 108p. R DEVELOPMENT CORE TEAM. 2013. R: A language and environment for statistical computing. Viena, Austria, R Foundation for Statistical Computing. Available in: <from http://www.r-project.org>. RIPLEY, B. et al. 2015. Support functions and datasets for venables and ripley's MASS. Available in: <https:/ /cran.r-project.org/web/packages/MASS/ MASS.pdf>. Accessed in: December 20, 2015. RUF, F. 2015. Diversification of cocoa farms in Côte d'Ivoire: complementarity of and competition from rubber rent. In: Ruf, F.; Schroth, G. eds. Agrotrópica 28(2) 2016 Economics and ecology of diversification. Netherlands, Springer. pp.41-86 SÁNCHEZ, P. et al. 1996. Caracterización y establecimiento de un banco de germoplasma de cacao criollo en el litoral aragueño. Caucagua, Edo. Miranda, Venezuela, FONAIAP- Estación Experimental Miranda. 57p. SANTANA, C. A. M. 1981. Beneficiamento e padronização do cacau. Uruçuca, BA, CEPLAC/ DEPED/EMARC-UR/NAGRI. 46p. SARKAR, D. 2015. Lattice: Multivariate Data Visualization with R. Available in: <http://cran.at.rproject.org/web/packages/lattice/lattice.pdf>. Accessed in: January 12, 2015. SODRÉ, G. A. et al. 2012. Extrato da casca do fruto do cacaueiro como fertilizante potássico no crescimento de mudas de cacaueiro. Revista Brasileira de Fruticultura 34 (3) :881-887. SOIL SURVEY STAFF. 2006. Keys to soil taxonomy. Washington, USDA Natural Resources Conservation Service. 338p. SOMARRIBA, E. 2004. ¿Cómo evaluar y mejorar el dosel de sombra en cacaotales? Agroforestería En Las Américas 41-42 :120-128. SUPERINTENDÊNCIA DE ESTUDOS ECONÔMICOS E SOCIAIS DA BAHIA (SEI). 2014. Tipologia Climática - Segundo Thornthwaite - Pluviometria 1943-1983 / Temperatura 1961 - 1990 - Estado da Bahia. Available in: <http://www.sei.ba.gov.br/site/ geoambientais/cartogramas/pdf/ carto_tip_clim.pdf> Accessed in: February 2, 2014. TAIZ, L.; ZEIGER, E. 2013. Plant physiology. Sounderland, Massachusetts. Sinauer Associates Inc. 954p. WORLD COCOA FOUNDATION (WCF). 2014. Cocoa market update. Available in: < http:// www.worldcocoafoundation.org/wp-content/ uploads/Cocoa-Market-Update-as-of-4-12014.pdf>. Accessed in: May 5, 2014. WOOD, G. A. R. 2001. Quality and inspection. In: Wood, G. A. R.; Lass, R. A. eds. Cocoa. Oxford, Blackwell Science. pp.119-126. WOOD, G. A. R.; LASS, R. A. 1985. Cocoa. London, Longman. 620p. l