507 Vol.46, n. 4 : pp. 507-514, December 003 ISSN 1516-8913 Prined in Brazil BRAZILIAN ARCHIVES OF BIOLOGY AND TECHNOLOGY AN INTERNATIONAL JOURNAL Variaion and Is Disribuion in Wild Cacao Populaions from he Brazilian Amazon Luiz Anônio dos Sanos Dias 1*, Júlio Pones Barriga, Paulo Yoshio Kageyama 3 and Caio Márcio Vasconcellos Cordeiro de Almeida 4 1 Deparameno de Biologia Geral/BIOAGRO; Universidade Federal de Viçosa; 36571-000; Viçosa - MG - Brazil. Delegacia Federal de Agriculura do Maranhão (DFA/MA); 6500-500; São Luís - MA - Brazil. 3 Deparameno de Ciências Floresais; Escola Superior de Agriculura Luiz de Queiroz ; Universidade de São Paulo; C. P. 09; 13418-900; Piracicaba - SP - Brazil. 4 CEPLAC/SUPOC/SERPE; Av. Gov. Jorge Teixeira, 86; Nova Poro Velho; 78904-300; Poro Velho - RO - Brazil ABSTRACT A sample of 64 progenies (30 cacao rees as a whole) from four Brazilian Amazon basins was colleced and evaluaed on he basis of 15 frui and seed rais. Nesed univariae analyses of variance showed significan variaion across progenies and basins. However, mos of he variabiliy appeared o be due o among rees and basins differences. The mulivariae analysis showed ha he differeniaion in cacao populaions occurred among basins. Since cacao diversiy was predominanly found in rees wihin basins and among basins, one should opimize he collecing process by aking as many rees as possible saring from few progenies and many river basins. These findings seemed o validae gene conservaion effors made o dae o preserve he cacao geneic resources and provide insigh ino he cacao geneic srucure aiming cacao collecion, managemen and improvemen. Key words: Theobroma cacao L., cacao rees, geneic srucure, collecing wild germplasm, Amazon river basin INTRODUCTION The cacao ree, Theobroma cacao L., is an insecpollinaed mixed perennial species, domesicaed by he Mayas, housands of years ago (Dias, 001a). Originally culivaed rees belonged o he Criollo racial group (Theobroma cacao var. cacao). Non-Criollo groups of cacao rees exis in heir wild sae in Souh America in he Amazon basin, and hey consiue he Amazon Forasero racial group (Theobroma cacao var. sphaerocarpum). Amazon Foraseros have been culivaed only during he las wo or hree cenuries (Cheesman, 1944; Cuarecasas, 1964). Presenly, his racial group accouns for over 80% of he world s producion of cacao. Boh Criollos and Foraseros are disinguished basically by frui and seed characerisics. Triniario is a hybrid group beween he wo, developed from he naural hybridizaion ha ook place in Trinidad, when he Criollo planaions were ravaged by disease in he mid-18h cenury and he los rees were replaced by Forasero, inerplaned wih he remaining Criollo (Cheesman, 1944). The puaive cacao cenre of origin is believed o be in he upper waers of he Amazon ribuaries Napo, Puumayo (Iça River in Brazil) and Caqueá (Japurá River in Brazil) a he foo of he Andes * Auhor for correspondence
508 Dias, L. A. S. e al. (Cheesman, 1944), currenly border of Colombia and Ecuador. From his region, cacao dispersed hroughou he Norhern and Souhern regions having he Andean Cordillera working as he barrier ha allowed he differeniaion beween Criollos and Foraseros. One can find Criollo populaions from Cenral America o Norhern Souh America. Amazon Forasero populaions are from he Lower and Upper Amazon. In oher words, Criollos are confined o he pacific wesern slope while Foraseros are found in he Amazon basin in he Guyanas, Surinam and Venezuela. Afer revisiing in deph all he hypoheses exisens, Dias (001b) proposed a new scenery for he origin and dispersal of cacao based on paleogeological, anhropological and hisorical evidences and, sudies of populaion geneics made wih molecular markers. The Brazilian Amazon region consiues one of he mos exensive reserves of cacao geneic variabiliy. In his region, expressive variabiliy regarding plan vigor, yield and is componens, resisance o wiches broom disease, size and leaf coloraion, and frui and seed characerisics has been found (Barriga e al., 1985; Almeida e al., 1987; Almeida and Almeida, 1987; Barley e al., 1988; Almeida and Dias, 001). In an effor o preserve cacao geneic resources and o broaden he crop geneic basis, he Execuive Commission for Planning Cacao Culure (CEPLAC) he insiuion ha is responsible for he cacao Brazilian research has colleced and incorporaed ino heir germplasm banks more han,000 accessions since 1965 (Barriga e al., 1985; Almeida e al., 1987; Almeida and Almeida, 1987; Almeida e al., 1995). Despie he grea imporance of his work, a he momen, he ransomed variabiliy represens only a small sample of he overall variabiliy of he species. On he oher hand, wild populaions are under eminen risk of exincion due o deforesaion, developmenal aciviies such as wood and mineral exploiaions, large hydroelecric projecs and by he expansion of he agriculural fronier. For insance, in he las 0 years, approximaely 550 housand km (14%) of he Amazon fores were cleared. Therefore, i is imperaive ha he cacao genepool be colleced, conserved and sudied for he presen and fuure. In fac, cacao has been considered by he Inernaional Board for Plan Geneic Resources (IBPGR, 1981) as a prioriy crop for conservaion. However, here are few sudies abou he levels of geneic variaion mainained wihin wild populaions and abou he way his variaion is geographically disribued. To manage he cacao geneic resources efficienly, we mus know how geneic variaion is disribued and wha facors influence is disribuion. There is some evidence ha he cacao disribuion is disconinuous, wih presence of small populaions confined in he viciniy of he riverbanks along he alluvial plains (Barley, unpublished work). Knowledge abou he level and he spaial disribuion of he variabiliy will provide a scienific basis for he collecion and conservaion of cacao germplasm. In he case of he diversiy o be concenraed in river basins his informaion will be used o opimize he collecion of germplasm for gene conservaion programs. The role of rivers in he diversificaion of Amazonian bioa was already recognized by Alfred Russel Wallace in he 1840s (Wallace, 1889). Pound (1938) and Almeida e al. (1987) described differeniaion of cacao populaions by river basins; however, his hypohesis has never been invesigaed. The goals of his sudy, using morphological daa available from four river basins (see Table 1 and Fig. 1), were: i) o assess he diversiy and is disribuion paern wihin and among cacao populaions from he Brazilian Amazonia; ii) o evaluae he hypohesis of differeniaion among hem by basins and; iii) o examine he cacao collecion sraegy carried ou by CEPLAC. Table 1 - Deails of basins sudied Basins Range of laiude Range of longiude Mean Range of annual aliude (m) rainfall (mm) Japurá 01 0 5-0 0 4 S 64 0 8-69 0 56 W 80 650-3650 1983 Amazon 0 0 38-03 0 08 S 56 0 44-58 0 18 W 50 > 000 1987 Purus 07 0 48-09 0 03 S 67 0 06-68 0 41 W 15 1750-50 1981 Ji-Paraná 11 0 0-11 0 11 S 61 0 5-6 0 39 W 300 1500-300 1984 Year of collecing
Variaion and Is Disribuion in Wild Cacao Populaions from he Brazilian Amazon 509 MATERIALS AND METHODS Plan maerial: The presen sudy was carried ou in he Cacao Geneic Resources Saion José Haroldo germplasm bank (ERJOH). The ERJOH is locaed in Mariuba, Pará Sae, Brasil (la 10 0 1 S, long 48 0 13 W, aliude 1 m), 18 km away from Belém, he capial of he Sae. For he presen work, a oal of 30 cacao rees from four river basins (Table 1 and Fig. 1), represening a considerable par of he river basin diversiy of he Brazilian Amazon region, were sampled and measured for frui and seed rais. I was assumed ha four basins, which were randomly chosen, were a represenaive sample ou of he river basins exising in he Amazon region. Wihin each basin, 16 open-pollinaed progenies were sampled, wih five rees/progeny being evaluaed. Figure 1 - Map of Brazilian Amazonian region, showing he four basins cacao ree collecion (1. Japurá,. Amazon, 3. Purus and 4. Ji-Paraná). For frui descripors, random samples of five fruis/ree were colleced and he following rais were measured: shape (SP), basal consricion (BC), apex form (AF), lengh (LG), widh (WD), lengh/widh raio (LG/WD), weigh (WG), frui surface rugosiy (FS), ridge pair appearance (RP), primary furrow deph (PF), wall hickness (WT), and seed number/frui (NS). To analyze seed descripors, he seeds from he same five fruis were exraced, peeled, and couned and heir we weigh (WW), dry weigh (DW), and form (SF) measured. The characerizaion occurred during January 1993. The procedures followed for he characerizaion were as deailed by Engels e al. (1980). Descripion of basins: Aiming a he collecion of cacao germplasm, Almeida e al. (1987) divided he Brazilian Amazon region ino basins (186), classified as follows: 4 of firs order, 60 of second order and 1 of hird order. Four basins (see Fig. 1 and Table 1) randomly sampled were parially sudied: one of firs order (middle Amazon River basin), wo of second order (Purus
510 Dias, L. A. S. e al. and Japurá River basins), and one of hird order (Ji-Paraná River basin). Deails such as where, when and by whom a paricular progeny from each river basin was colleced could be found elsewhere (Almeida e al., 1995). The samples aken represened radiional cacao crop regions (Amazon) and regions where cacao grew sponaneously (Purus and Japurá); regions subjec o flooding (Purús, Japurá and Amazon) and non-flooded regions (Ji-Paraná); regions from differen geographical posiions as cenral (Amazon), norh (Japurá) and souh (Purús and Ji- Paraná); he region comprising he puaive cacao cenre of origin (Japurá); he region which appears o indicae some human inervenion in he cacao spread (Ji-Paraná); and he region in which cacao rees have low or no sympoms of wiches broom disease. Saisical mehods: Two mehods were used o deermine he differeniaion in morphological rais: nesed univariae analyses of variance and mulivariae analysis. The univariae approach calculaes variance componens a each of hree levels of populaion subdivision: among basins [B i ], among progenies wihin basins [P j(i) ], and among rees wihin progenies wihin basins [T k(ij) ]. The random effecs B i, P j(i), T k(ij) were assumed independen, wih means equal o 0 and variances equal o σ b, σ p, σ, respecively. Then, he oal variance esimae σ T was pariioned as follows: σ T b = σ + σ p + σ The nesed analyses of variance were performed using he following model: Y ijk = µ + B i + P j(i) + T k(ij) where Y ijk = average value of he rai Y measured on five fruis in he k h ree of he j h progeny in he i h basin; µ = overall mean; B i = effec of he i h basin; i = 1,,..., 4; P j(i) = effec of he j h progeny in he i h basin; j = 1,,..., 16; T k(ij) = random error defined by he effec of he k h ree of he j h progeny in he i h basin; k = 1,,..., 5. The forma of he nesed univariae analyses of variance and he expecaion mean squares are shown in Table. The mulivariae approach o describe differeniaion beween basins was based on he D -saisic of disance of Mahalanobis (1936). I is he mos robus measuremen of disance, wih he propery of being scale invarian and, herefore, i is dimensionless (Dias and Kageyama, 1998). The D disance beween wo basins i and i on p characers is defined as: D = d ~ ' W 1 d ~ where ~ d is a vecor of differences beween he basins ' averages i and i for all p characers and d is is ~ ranspose. W is he p x p wihin-basin covariance marix. In order o achieve D values significance levels, F saisics was applied. All he saisical analyses were performed using he SAS package (SAS Insiue Inc., 1989). The univariae analyses of variance and he D Mahalanobis disance were performed by using he NESTED and CANDISC procedures, respecively, from SAS. Table - Forma of nesed analysis of variance and he expecaion mean squares Sources of variaion d. f. Expeced mean squares Enry Tes agains enry Basins (B) (a -1) σ + σ + σ 1 Progenies (P) wihin (B) a(b - 1) σ + nσ 3 Trees (T) wihin (P) ab(c - 1) σ 3 Toal abc 1 n p p np b RESULTS AND DISCUSSION For all morphological rais sudied, highly significan differences across progenies and basins were revealed (Table 3). In mos cases, he progeny componen of variance was raher low for all rais. For seven of he frui rais, he ree variance componen was higher. Conversely, for seven oher rais concerning he seeds and he frui shape such as LG/WD, FS, WT and NS, he
Variaion and Is Disribuion in Wild Cacao Populaions from he Brazilian Amazon 511 basin variance componen was higher. In general, he pariioning of variaion wihin and among populaions (river basins) revealed ha mos variaion was deeced wihin populaions, as in oher sudies wih cacao ree, involving isozymaic (Ronning and Schnell, 1994; Sounigo e al., 1997) and RAPD (Russel e al., 1993) markers. Many oher ree species have revealed he same paern of variaion disribuion. Tropical rees showed high levels of geneic variaion, conained mosly wihin he populaions (Dias and Kageyama, 1991). D values compued for differeniaion beween basins (Table 4) ranged from 4.9 o 6.6. Japurá and Ji-Paraná basins showed o be he mos divergence, while Purus and Amazon were he mos similar basins. Considering Cheesman s (1944) hypohesis, he Japurá River (Caqueá River in Colombia) basin comprised he puaive cacao cenre of origin which could explain is relaive greaes divergence. When submied o F saisic, all D values were highly significan agreeing wih resuls found by he previous univariae es. Consequenly, each basin was geneically disinc from each oher. Table 3 - Summary of he nesed analysis of variance and he variance componens (in %) for each source of variaion Sources of variaion Basins (B) Progenies (P) wihin (B) Trees (T) wihin (P) Trais Mean Squares Variance Componen (%) Mean Squares Variance Componen (%) Mean Squares Variance Componen (%) Frui SP 1 0.334** 9.38 0.063** 1.37 0.05 69.5 BC 1.348** 5.41 0.078** 1.75 0.039 61.84 AF 1.644** 4.44 0.170** 37.81 0.08 37.75 LG 1.441** 5.07 0.15** 9.0 0.030 45.90 WD 0.059**.97 0.030** 36.76 0.007 60.6 LG/WD a 1.144** 64.78 0.01** 5.86 0.006 9.36 WG 165.154** 38.76 6.133** 14.59.391 46.64 FS 35.44** 6.58 0.504** 8.71 0.00 8.70 RP 0.007 0.00 0.110** 38.67 0.06 61.33 PF 553.13** 17.70 45.708** 11.3 5.431 70.98 WT.341** 6.15 0.036** 10.36 0.01 7.48 NS 0.757** 43.95 0.07** 18.90 0.007 37.15 Seed WW 0.650** 47.10 0.016** 10.55 0.007 4.35 DW 89.3** 64.53 1.45** 1.09 0.397 3.37 SF 44.09** 67.35 0.63** 11.15 0.173 1.50 d.f. 3 60 56 ** P < 0.01 a (x 10 6 ); 1 acronyms are defined in he ex; % of oal for each source of variaion in each rai. Table 4. Mahalanobis disance (above diagonal) o river basins and P values by F saisic (bellow diagonal) River basins Japurá Amazon Purus Ji-Paraná Japurá 0 8.658 1.6837 6.663 Amazon 0.0001 0 4.9548 14.0855 Purus 0.0001 0.0001 0 8.6841 Ji-Paraná 0.0001 0.0001 0.0001 0 This resul appeared o confirm ha he differeniaion of cacao wild populaions from he Brazilian Amazon region evolved according o he each river basin sysem. Alhough we would no be able o disinguish he several hypoheses for his cacao populaions spaial differeniaion paern based on river basins, i is possible o discuss hem. In fac, i may be correlaed wih he large disance among
51 Dias, L. A. S. e al. he sudied populaions or may reflec he acion of pas evoluionary processes. In he firs case, he isolaion among he four populaions, all more han 800 km apar, prevened he gene flow among hem. In he second case, such paern of differeniaion could be achieved by he combinaion of he Riverine barriers (Paon e al., 1994), Refuge (Haffer, 1969) and he Amazon Lagoon (Klammer, 1984) hypoheses. According o he Riverine barriers hypohesis, he developmen of he Amazon River sysem, caused by Cenozoic uplif of he Andes, promoed a series of vicariance evens. As a consequence, he fragmenaion of ranges of widespread ancesral species occupying ha region ook place. Demes a opposie margins or a basins have he gene flow reduced, wih naural selecion and geneic drif as main driving forces of divergence beween populaions. Hence, he inrabasin differeniaion among populaions is less pronounced han he inerbasin differeniaion. In addiion, he gene flow is expeced o be more resriced across large Rivers and across heir mouhs han across small ones and heir headwaers. Unforunaely, he proof for his hypohesis would be he sudy of differeniaion among cacao populaions on adjacen rivers, which was no done here. However, he fac of each basin o be geneically disinc from each oher, as revealed by applying he D saisic, proved ha exchanges of fruis beween far upsream and far downsream from he Japurá, Purus and Ji-Paraná basins for middle Amazon basin did no exis or i was inexpressive. The Refuge hypohesis is based on cycles of Pleisocene climaic changes, which resuled in corresponding cycles of fragmenaion and coalescence of ropical fores species. Populaions or enire communiies, isolaed in refuges, evolved independenly under naural selecion and geneic drif. Boh hypoheses described above accoun for he classical alloparic speciaion model. In some exension, he river basins sudied (see Fig. 1) correspond wih he refuge areas proposed by Prance (1973) for four families of rees, as follows: Japurá, Amazon, Purús, and Ji-Paraná basins wih Imerí, Manaus, Tefé and Aripuanã refuges, respecively. Refuges areas have played an imporan role in he evoluion of wild cacao populaions from French Guyana, as argued by Lachenaud (1997). Alhough he Amazon Lagoon hypohesis has been resriced o geology, i also could play an imporan role in he Neoropical diversificaion (Marroig and Cerqueira, 1997). During he las four million years (4 Myr Before Presen), in he Plio-Pleisocene, he rising level of sea became a barrier for he discharge of he Amazon waer sysem, leading o formaion of he Amazon Lagoon. The recurren oscillaion in he sea level for as long as Myr B.P. resuled in a displacemen of fores areas o he bordering of he presen Amazon highlands. Thus, he differeniaion occurred a highlands of basin borders increased for he isolaion effec of enlarged Rivers. In urn, he exen of geneic differeniaion beween and wihin river basins a a given spaial scale depends on he disance and magniude of gene flow, he scale of naural selecion and, he geneic drif. Cacao rees presen some specializaion in heir pollinaion mode. In erms of managemen, he knowledge of pollinaor populaion becomes as imporan as he knowledge abou he ree species serviced by pollinaors. Unforunaely, we know virually nohing abou his subjec. Pollinaion is mainly achieved by a number of species of midges of he genus Forcipomya and, accidenally, for oher insecs (Chapman and Soria, 1983). Ineresingly, hese midges are so small ha hey are difficul o see and o sudy. Moreover, hey fly from one ree o anoher foraging a shor disance, frequenly no more han 60 m (Posnee, 1950). Hence, pollen flow occurs among neighboring rees causing he formaion of family clusers. In sudying he relaionship beween he disribuion of geneic variaion wihin and beween ropical ree species and he gene flow via pollen, Loveless and Hamrick (1987) argued ha he smaller fligh disance of pollinaors he larger geneic variaion beween populaion. This saemen can be direcly applicable o cacao germplasm differeniaion by basin. Resriced gene flow and wihin-basin differeniaion due o microsie selecion pressure increases he formaion of neighbourhoods. Resriced gene flow also reduces effecive populaion size, causing increase of geneic drif and of inbreeding due o maing among relaives. Furhermore, even having complee and hermaphrodie flower, an incompaibiliy sysem operaes in he cacao species (Cope, 196), favouring oucrossing, wihou excluding selfpollinaion. Wihin he same populaion some cacao rees may be enirely self-incompaible, while ohers may display self-compaibiliy o varying degrees. Thus, cacao rees wih differen
Variaion and Is Disribuion in Wild Cacao Populaions from he Brazilian Amazon 513 oucrossing behaviours compee wihin each populaion. CONCLUSIONS The presen sudy provides insigh ino cacao geneic srucure aiming a cacao collecion, managemen and improvemen, i.e., he river basin appears o be an imporan facor in variaion srucuring of cacao populaions from he Amazonia. Neverheless, i is necessary o be cauious since characerizaion has been done wihin one year only. Thus, his insigh could be influenced by he effec of genoype-year ineracion. An adequae collecing sraegy mus ake ino accoun he spaial disribuion of he geneic variaion. Because cacao diversiy is predominanly concenraed in rees wihin basins and among basins, aking as many as possible rees for few progenies from many basins one should opimize he collecing process. Moreover, he lile undersanding of he organizing geneic srucure of he cacao wild populaions requires he collecing of larger samples. RESUMO O nível e a disribuição da variação em populações silvesres de cacau é pouco conhecido. Para a sua avaliação colearam-se, em quaro bacias hidrográficas, 64 progênies (30 cacaueiros), as quais foram avaliadas para 15 caracerísicas de fruo e semene. Análises univariadas de variância evidenciaram significane variação enre progênies e bacias. A análise mulivariada revelou que a diferenciação enre as populações esá esruurada por bacias. Uma vez que a maior pare da variação concenrou-se enre cacaueiros denro de bacias e enre bacias, para se oimizar o processo de colea de germoplasma deve-se colear muios cacaueiros procedenes de poucas progênies e muias bacias. Esses resulados validam os esforços de colea de germoplasma feios aé o presene para preservar os recursos genéicos de cacau e fornecem subsídios para o enendimeno da esruura genéica visando a colea, o manejo e o melhorameno do cacaueiro. REFERENCES Almeida, C. M. V. C. and Almeida, C. F. G. (1987), Colea de cacau silvesre no Esado de Rondônia. Rev. Theobroma, 17, 65-9. Almeida, C. M. V. C.; Barriga, J. P.; Machado, P. F. R. and Barley, B. G. D. (1987), Evolução do programa de conservação dos recursos genéicos de cacau na Amazônia brasileira. Boleim Técnico, 5 (CEPLAC/DEPEA, Belém). Almeida, C. M. V. C.; Machado, P. F. R.; Barriga, J. P. and Silva F. C. O. (1995), Colea de cacau (Theobroma cacao L.) da Amazônia brasileira: uma abordagem hisórica e analíica. Poro Velho : CEPLAC/PLANAFLORO. Almeida, C. M. V. C. and Dias, L. A. S. (001), Recursos genéicos. In: Dias, L. A. S. (Ed.). Melhorameno genéico do cacaueiro. Viçosa : FUNAPE. pp. 163-16. Barriga, J. P.; Machado, P. F. R.; Almeida, C. M. V. C. and Almeida, C. F. G. (1985), A preservação e uilização dos recursos genéicos de cacau na Amazônia Brasileira. In: 9 h Inl. Cocoa Res. Conf. London : Cocoa Producers' Alliance. pp. 73-79. Barley, B. G. D.; Machado, P. F. R.; Ahner, D. and Barriga, J. P. and Almeida, C. M. V. C. (1988), Descrição de populações de cacau da Amazônia brasileira. I observações preliminares sobre populações de Alenquer, Pará. In: 10 h Inl. Cocoa Res. Conf. London : Cocoa Producers' Alliance. pp. 665-67. Chapman, R. K. and Soria, S. J. (1983), Comparaive Forcipomyia (Dipera, Ceraopogonidae) pollinaion of cacao in Cenral America and Souhern Mexico. Rev. Theobroma, 13, 19-139. Cheesman, E. E. (1944), Noes on he nomenclaure, classificaion and possible relaionships of cacao populaions. Trop. Agric., 1, 144-159 Cope, F. W. (196), The mechanism of pollen incompaibiliy in Theobroma cacao L. Herediy, 17, 157-18. Cuarecasas, J. (1964), Cacao and is allies: a axonomic revision of he genus Theobroma. Conrib. US Na. Herb., 35, 377-605. Dias, L. A. S. (001a), Melhorameno genéico do cacaueiro. Viçosa : FUNAPE. Dias, L. A. S. (001b), Origem e dispersão de Theobroma cacao L.: novo cenário. In: Melhorameno genéico do cacaueiro. Viçosa : FUNAPE. pp. 81-17. Dias, L. A. S. and Kageyama, P. Y. (1991), Variação genéica em espécies arbóreas e conseqüências para o melhorameno floresal. Agrorópica, 3, 119-17. Dias, L. A. S. and Kageyama, P. Y. (1998), Comparison beween mulivariae mehods applied for he evaluaion of geneic divergence in cacao (Theobroma cacao L.). Braz. Arch. Biol. Technol., 41, 199-06.
514 Dias, L. A. S. e al. Engels, J. M. M.; Barley, B. G. D.; Enríquez, G. A. (1980), Cacao descripors, heir saes and modus operandi. Turrialba, 30, 09-18. Haffer, J. (1969), Speciaion in Amazonian fores birds. Science, 165, 131-137. IBPGR (1981), Geneic resources of cocoa. Rome : IBPGR. Klammer, G. (1984), The relief of he exra-andean Amazon basin. In: Sioli, H. and Junk, W. (Eds.). The Amazon: Limnology and landscape ecology of a mighy ropical river and is basin. pp. 47-83. Lachenaud, P. (1997), Geneic/axonomic srucuring of he Theobroma cacao L. species fresh hypoheses. INGENIC Newsleer, 3, 10-11. Loveless, M. D. and Hamrich, J. L. (1987), Disribución de la variación en espécies de arboles ropicales. Rev. Biol. Tropical, 35, 165-175. Mahalanobis, P. C. (1936), On he generalized disance in saisics. Proc. Na. Ins. Sci. India, 1, 49-55. Marroig, G. and Cerqueira, R. (1997), Plio-Pleisocene Souh America hisory and he Amazonas Lagoon hypohesis: a piece in he puzzle of Amazonian diversificaion. J. Comp. Biol.,, 103-119. Paon, J. L.; Silva, M. N. F. and Malcolm, J. R. (1994), Gene genealogy and differeniaion among arboreal spiny ras (Rodonia: Echimydae) of he Amazon basin: a es of he riverine barrier hypohesis. Evoluion, 48, 1314-133. Posnee, A. F. (1950), The pollinaion of cacao in he Gold Coas. J. Hor. Sci., 5, 155-163. Pound, F. J. (1938), Cacao and wicbroom disease (Marasmius perniciosus) of Souh America wih noes on oher species of Theobroma: repor on a visi o Ecuador, he Amazon valley and Colombia April 1937-April 1938. Yuille s Prinerie, Trinidad and Tobago. Prance, G. T. (1973), Phyogeographic suppor for he heory of Pleisocene fores refuges in he Amazon basin, based on evidence from disribuion paerns in Caryocaraceae, Chrysobalanaceae, ichapealaceae and Lecyhidaceae. Aca Amazonica, 3, 5-6. Ronning, C. M. and Schnell, R. J. (1994), Allozyme diversiy in a germplasm collecion of Theobroma cacao L.. The J. of Herediy, 85, 91-95. Russel, J. R.; Hosein, F.; Johnson, E.; Waugh, R. and Powell, W. (1993), Geneic differeniaion of cocoa (Theobroma cacao L.). Mol. Ecol.,, 89-97. SAS Insiue Inc. (1989), SAS/STAT user s guide, version 6. 4. ed. Cary : SAS Insiue. vol.. Sounigo, O.; Umaharan, R. and Sankar, A. (1997), Assessing geneic diversiy in he Inernaional Cocoa Genebank of Trinidad (ICG,T) using isozyme elecroforesis. INGENIC Newsleer, 3, 11-16. Wallace, A. R. (1889), A narraive of ravels on he Amazon and rio Negro. New York : Dover Publicaions. Received: November 19, 001; Revised: July 18, 00; Acceped: November 01, 00.