Theobroma cacao L.

Authority

Cuatrecasas, José. 1964. Cacao and its allies, a taxonomic revision of the genus Theobroma. Contr. U. S. Natl. Herb. 35: 379-614. pl. 1-12.

Family

Malvaceae

Scientific Name

Theobroma cacao L.

Type

Type.-Sloane Herbarium vol. 5, p. 59 (BM, lectotype); Paratype: Tournefort pl. 444, fruit-lectotype. Types of Theobroma cacao.-The only specimen of T. cacao in the Linnaean herbarium is the number 934-1, but this specimen is posterior to the publication of Species Plantarum (1753), since it bears the annotation "PI. Surin. 1775. p. 13." The type has to be sought among the bibliographic references of Linnaeus. In Species Plantarum, Linnaeus refers to Hortus Cliffortianus. At my request, Mr. Sandwith was kind enough to examine the Hortus Cliffortianus herbarium at the British Museum and found no specimens of Theobroma in it. However, Hortus Cliffortianus page 379 seems to give a key to the matter, where Linnaeus writes "Flores a nullo bene depicti, multo minus descripti sunt, . . . Sloane mihi inspiciendi copiam fecit, . . ." and gives the quotation "Hist. Jam. pl. 160." Linnaeus was especially preoccupied with the flowers; he wanted to know exactly the structure in order to be able to place Theobroma in the right place in his sexual system; the previous literature did not give him the answer, and neither did the Sloane plate. However, from the above paragraph we may infer that Sloane sent him flowers at his request, surely very few, which may be the reason why there are none of them in the Linnaean herbarium; from Sloane's flowers Linnaeus found the floral structure of Theobroma by himself and drafted his diagnosis. The original flowers (surely dissected) having disappeared, the corresponding specimens in the Sloane Herbarium have to be considered the isotype. Messrs. Dandy and Sandwith found the specimens in the Sloane Herbarium at the British Museum. Mr. Sandwith writes: "We found the corresponding specimen in the Sloane Herbarium and it is obviously the source of (in fact part of) the plate, though not identical; there are leaves, detached flowers, fragments of fruit and one seed; it is the vol. 5 p. 59." There is the possibility that Sloane lent Linnaeus these specimens, but even if this were not the case, we may assume that Linnaeus used flowers taken from this specimen which would thus be an isotype; therefore, the sheet page 59, volume 5, of the Sloane Herbarium is to be chosen as lectotype for the flowers and leaves. According to Sandwith and Dandy "there is also what appears to be a duplicate, leaves only, in the Plukenet Herbarium volume of Herb. Sloane. The leaves of both specimens are reddish brown and glabrous beneath with reticulate tertiary veins." Sloane's plate 160 is a paratype, but another even more important paratype is Tournefort's plate 444. Tournefort is the only reference given by Linnaeus in his original description of the genus in Genera Plantarum (1754), and its citation precedes the reference to Sloane in Hortus Cliffortianus. I propose it as the lectotype for the fruit, because the Tournefort drawing is perfectly defined, leaving no doubt that the characters are of the Criollo type. On the other hand, the fruits shown on Sloane's plate 160 are not unquestionable, even though they are very pointed; they are among the variations found in Criollo populations. But the reasons which compel me to consider the flower specimens in the Sloane Herbarium as the lectotype do not apply to the fruit, the origin of which may have been different from the origin of the flowers and leaves, for Sloane collected in several places and countries. A fruit typification by the Tournefort plate fits well the definition of T. cacao L. sensu stricto given by Pittier and later authors. It is obvious that all these authors described cultivars and that the cacao described by the earlier authors was of the type Criollo, as can be inferred from the illustrations, and also from the descriptions of Hernandez, Pison, Plukenet, Merian, Weinmann, Tournefort, Catesby, and Gaertner.

Description

Description - Tree usually 4-8 m. high, rarely taller (up to 20m.), with the growth of the sympodial stem by subterminal, lateral, upright shoots (chupons); primary branching by successive whorls of normally spreading branches; young branchlets terete, grayish green or brownish, densely or sparsely pubescent, with slender, patulous, acute, simple or furcate hairs 0.1-0.3 mm. long, later glabrate, more or less striate, rugulose and sparsely lenticellate; stipules subulate, very acute, 5-14 mm. long, 0.5-1.5 mm. wide at base, pubescent or puberulous, deciduous. Leaves coriaceous or chartaceous, more or less rigid, alternate, distichous on the normal branches, green; petioles pubescent or tomentose, with simple, acute, slender, rather dense, spreading hairs, thickened pulvinate at the ends, 1.5-2 (1-3) cm. long, on orthotropus stems 3-10 cm. long; blades subobovate-oblong or elliptic-oblong, slightly asymmetrical, rounded or obtuse at base, attenuate and cuspidate at apex, acute or subacute, usually entire or slightly and irregularly sinuate, green above, pale when dry, glabrous except for the pubescent or puberulous midrib, the midrib linear, prominent, the secondary nerves filiform, the fine veins reticulate, often slightly prominent, lighter green beneath, glabrous or with very sparse, minute, simple, furcate or stellate hairs, rarely puberulous, the midrib thick and prominent, the secondary nerves 9-12 each side, prominent, subpatulous, then ascending, near the margin curving, slendering, anastomosing, the tertiary nerves prominent, the minor veins reticulate and prominulous, 15-50 cm. long, 4-15 cm. broad, the acumen 1-2.5 cm. long. Inflorescences on the trunk and branches, usually borne on small tubercles, the cymose branchlets short, knotty, persistent, the cymose peduncles 1-3 mm. long, stellate-pubescent, hirtellous and with scattered glandular trichomes; bracts ovate or ovate-oblong, amplectant, pubescent; bracteoles ovate-oblong, acute or subacute, 0.5-1.2 mm. long, pubescent, deciduous; pedicels capillary, rigid, pale green, whitish or reddish, 5-15 mm. long, hirtellous with rather dense, thin, acute, patulous, stellate or furcate hairs and sparse pluricellular, glandular, capitate trichomes; buds white, whitish green, lilac, or reddish, ovoid or oblong-ovoid, acute, 5-7 mm. long, subglabrous or sparsely puberulous with slender stellate hairs and thicker, glandular, stipitate trichomes; sepals thick-membranaceous, lanceolate or oblonglanceolate, acute, white, greenish white, pale violaceous or reddish, slightly 3-nerved, shortly (0.5-1 mm.) united at base, 5-8 mm. long, 1.5-2 mm. wide, minutely tomentose at margin, outside sparsely pubescent with stellate and glandulose hairs, or glabrous, inside glabrous or with rare glandular trichomes; petals contorted in aestivation, thick-membranaceous, the hood 3-4 mm. long, 5-2 mm. wide, obovate, rounded at apex, white, 3-nerved, the nerves papillose inside, the lateral ones very thick, usually purplish or red, the medial prominent only upwardly; petal-lamina pale yellowish, 1.5-2.5 mm. long, 1.5-2 mm. wide, obovate or rhombic-obovate, attenuate at apex, acuminate or subtruncate and shortly mucronate, rarely blunt, entire or sinuate at margin, pedicellate at base, the pedicel linear, 3-nerved and sometimes 4- or 5-nerved at apex, 2-4 mm. long, 0.2-0.5 mm. wide; androecium tube rather thick, 1.2-1.5 mm. long, glabrous; staminodes 4-6 mm. long, 0.6-0.7 mm. broad at base, narrowly subulate, very acute, erect in bud, erect or somewhat flexuous in anthesis, the middle vein thick, angular, red or purplish, minutely papillose-pilose, the thin margin whitish, more or less revolute, ciliate, with slender, flexuose simple hairs; stamens diantheriferous, the filaments glabrous, patulous or recurved, 1.5-2 mm. long, the anthers about 0.4 mm., with rounded cells; ovary oblong-ovoid, obtusely pentagonal or 5-ridged, about 1.5 mm. high, 0.8 mm. thick, glabrous or usually glandular, covered with more or less copious white or reddish, pluricellular, stipitate glands; styles 5, glabrous, adherent, 1.5-2.5 mm. long. Fruit subbaccate, variable in shape, from globose to fusiform and acute, and with a very smooth to a strongly ridged and rugose or verrucose surface; pericarp consistently fleshy and thick (5-15 mm.), and usually made of two, more or less conspicuously different, carnose layers (epicarp and endocarp) separated by a thin, ligneous membrane (mesocarp), the endocarp limited by a firm epidermis inside, the inner wall of the shell, the epicarp sometimes differing in color and firmness, rich in mucilaginous cells, limited outside by the hard epidermis of the fruit, the mesocarpic membrane sometimes reduced to isolated bundles of fibers or lacking, the endocarp also sometimes lacking. Seeds (20-40) usually arranged in 5 rows, but sometimes when large arranged in 4 or 3 rows, the five radial walls initially separating the 5 cavites in the earlier stages reabsorbed long before the maturation of the fruit; seeds ovoid, ellipsoid, amygdaloid, more or less complanate through mutual pressure or almost round in cross section, variable in size (20-40 mm. long, 12-20 mm. broad), the integuments brown, subcoriaceous, the surrounding pulp white, sweet, the cotyledons white, purplish, violet, or intermediate in color.

Common Names - Cacao, cocoa, cacao dulce, cacao criollo, cacao calabacillo, cacao forastero, cacao amelonado, cacao trinitario, cacao lagarto, and many other adjectives according cultivar varieties and regional or local cultivar forms. Also: Bizoya, yagabizoya (Reko), deghy (Otomi), caocauatzaua (Zoque), kako (Mixe), cahequa (Tarascán), Chudechu (Otomi) in Mexico after Standley; cacahuatl, cacahoatl, cacahoaquahuitl, quauhcacahoatl, mecacahoatl, xochicacahoatl, tlacacahoatl, cacahoacuahuitl, cacaotlquahuitl (Nauhatl), kicou, kicob, cuculat, pacxoc, cucuh, caco in Costa Rica and Guatemala; cacao chuncho in Peru. For Costa Rica and Chiriquí, Pittier (1902) gives the following Indian names: ko (Térreba), koó (Brunka), kajo (Guatuso), kuá (Guaimí), kno (Penonomé), doló (Doraske), tsirú (Bribrí), (Cabecara); according to him the Bribrí Indians use the following names for some varieties of cacao: murú-uak, tsipá-uak, xi-uak, and betsún-uak; Standley (1937) mentions, also for Costa Rica: kuk (Rama), tsirú-kurú (Cabécara), kao-krá (Brunka) [Brunca] and kau (Tiribí). Cacao silvestre, cacao de monte, wild cacao, are names often used in different countries whenever a cacao tree is found apparently spontaneously growing. It is a fact that although the other species have native Indian names, for Theobroma cacao only the name "cacao" is recorded from the whole of South America, whereas this species has many indigenous names in Central America.

Discussion

Theobroma cacao is variable in its characters, especially with regard to the color, size, and shape of the parts of the flower and the fruits. These are variations to be expected of an ancient crop spread throughout a very wide area. Based on some of these more or less consistent variations, Bernoulli described three species, T. -pentagonum, T. leiocarpum, and T. saltzmannianum, and Chevalier T. sphaerocarpum. The shape of the fruit is the main defining character, except for T. saltzmannianum, which was based on petal characters, probably from an abnormal specimen. The few floral characters given for the other three species are irrelevant or inconstant. Theobroma pentagonum is defined by having elongate, gradually and acutely narrowed, warty pods which are strongly pentagonal and 5-ridged; it has white seeds. It was described from the Atlantic coast of Guatemala and is called "cacao lagarto." Theobroma leiocarpum was characterized by ovoid pods, almost smooth, with five very shallow furrows and a glabrous ovary; the color of the seeds was not stated; it was found in plantations on the Atlantic coast of Guatemala, and stated to be rather rare. Theobroma sphaerocarpum was described from cultivated specimens on São Tomé island, western Africa, characterized by its nearly spherical, slightly 10-furrowed, almost smooth or slightly rough pods, and violet cotyledons. Schumann considered the three Bernoulli species as mere local variations of T. cacao and therefore unworthy of taxonomic consideration. Some authors have followed Schumann in this view, but there is a tendency to accept T. pentagonum as a different species, because of its characteristic fruit form and thinner pericarp. For many years there has evidently been confusion in the taxonomy of cultivated cacaos, the main problem being insufficient knowledge of the wild populations of T. cacao. There are many citations of places in Central and South America where T. cacao is said to have been found wild, which may have been true in some cases, but not in others where we may be dealing with the remains of abandoned plantations. The discovery of wild cacao by Stahel in the rain forests of Mamaboen Creek in Surinam (confirmed later by Myers) and a few other places very distant from populations, is very significant; the cacao is of the Amelonado type. Ogilvie also found it in abundance on Black Creek, a branch of the Essequebo River, and along the Berbice River in British Guiana. He found it along the Rupunumi River at Rewa and Quitaro Creeks, at Kuduwini Kassikedju or Dewar Wow, etc., on the upper Essequebo River, and also on the Quitari River (according to Myers). Robert Schomburgk also cited wild cacao in several places in British Guiana on the Cutari, at the head of the Correntyne River, and also at Quitaro, and Richard Schomburgk found it on the Upper Pomeroon River. Wild cacaos also have been found, according to Stahel, in the Upper Oyapok in French Guiana (Myers), all of the Amelonado type, implying that this population of the Guiana highland area westwards to the Amazon valley may have been the home of this cultivated variety. In Brazil there has been found wild cacao near the Guianas and, according to Ducke, at the upper Rio Branco, northeast of Obidos, and at Francez on the middle Tapaj6z; he adds that the area of spontaneous T. cacao includes a greater part of the Hylaea and that it is of the form "leiocarpum." Preuss (1901, p. 247) found wild cacao in eastern Ecuador, also of the Amelonado type. Huber found spontaneous cacao in the upper Purus river, and along the Río Ucayali, Peru. Pound also claimed to have found spontaneous cacaos in the upper Amazon basin. The spontaneous cacao trees found by the Anglo-Colombian Expedition in the forests of Caguan (Caquetá) and Río San Miguel offer some doubt about their indigenousness. I found myself, in 1939, in the rain forests of Río Guaviare wild trees of T. cacao, tall and well developed, but we have to be cautious in accepting them as indigenous, because the Guaviare River is said to have had more extensive plantations on its banks in earlier times. An interesting find was made in British Honduras by Sampson, who encountered wild trees of the Criollo type in the Southern forests. In Mexico, Miranda recently found a completely indigenous cacao tree in a forested region of Chiapas (Selva Lacandona), very distant from any population; the natives (Lacandones indians) reported that other trees of such cacaos are found scattered; this cacao has a somewhat scandent stem and produces small, elongated, pointed, slightly 10-ridged and rugose fruits. Standley and Steyermark (FI. Guat. 426) say that in the lowlands of Guatemala sometimes cacao is found more or less wild in the forest, especially in Alta Verapaz, and that it is not improbable that it is native in the wet North Coast region. Many forms of the cultivated cacaos have received local or regional names which after the many introductions of cultivars from one country to another have brought confusion to the complicated system of cacao types. Morris was the first to make a classification of cultivars, arranging them in two main groups: Criollo and Forastero. Hart (see page 396) modified this, making three groups (Criollo, Forastero, and Calabacillo), later separating T. pentagonum (cultivar "Lagarto") into another group, as a different species from T. cacao. The classification of Hart has been followed by many authors of cacao treatises until recent times when van Hall, simplifying, went back to the basic two groups of Morris: Criollo and Forastero. Pittier, a botanist with much experience both in Central and South America, recognized the existence of two different, original, spontaneous forms of cacao, Criollo with elongated, ridged, pointed fruits and white cotyledons, and Forastero, with short, roundish, almost smooth fruit and purplish cotyledons; he believed that they corresponded with two different species, T. leiocarpum generally spread throughout tropical South America, where it is still found spontaneous, and T. cacao, spontaneous in Central America and which was the source of the prehistoric cultivation and selection of cacao, all of the Criollo type, in Mexico and Central America. Introduction of Criollo types in South America and West Indies and conversely of the smooth type into West Indies and Central America created the cross-varieties which multiplied with the years. Although there is much speculation in this, the theory is a workable and reasonable one, because the available historical data favors the recognition of the earlier Central American and Mexican cacaos as being of the Criollo type. The finding of spontaneous cacao of this type in Chiapas and British Honduras also supports this theory. Another favorable fact is the uniformity of the Venezuelan Criollo, supposed to have been introduced from Central America to Venezuela in the earlier times of the Spanish conquest. Soria, after visiting (1961) important plantations in Mexico where new Forastero types have been abundantly introduced in recent times, recognized that in Mexico before 1900 varieties of the Criollo type almost exclusively were cultivated. He observed in large, old (more than 50 years) plantations a great variation in the Criollo type, but the seeds were always white and the pods, variable in shape, were always pointed, with surfaces varying from tuberculate to rugose, from light green, through green, to reddish; the trees were small, slow growing, and often with fewer branches (5-3 in each whorl) than is normal in the species; the petal-laminas are bright yellow. Pittier's theory was very much welcomed by botanists and cocoa experts; Chevalier supported it, and Ducke also in its basic idea. Cheesman adhered to it at first (1932), but later (1944) developed a new theory that all cultivated Criollo cacaos came "from an offshoot of a general cacao stock in the headwaters of the Amazon carried over the Andes into southern Colombia, and developed many of their present characters in association with man." But historical knowledge at present can only closely relate cacao to Central American man, especially to the Mayans and not to the South American Indians. Central American Indians undoubtedly developed the art of planting and selecting of cacao through several thousands of years, finally obtaining the high quality produce which the Spaniards found at the time of the conquest. It may be assumed that in early times a natural population of Theobroma cacao was spread throughout the central part of Amazonia-Guiana, westward and northward to the south of Mexico; that these populations developed into two different forms geographically separated by the Panama isthmus; and that these two original forms, when isolated, had sufficiently consistent characters to be recognized as subspecies. As they intermingle readily by crossing, giving fertile and robust hybrids, they cannot be considered distinct species. Both types in cultivation have originated many mutations, some of them persistent, thanks to ecological adjustment, selection, and isolation. Because the cultivation and selection has been very active for some thousand years in the Central American-Mexican area, it is in these areas where we find the richest variety of forms. When the subspecies (with their different forms) interbred the products gave the great and confusing variety of existing types. Another theory is that most of the stable forms of cacao might have originated by mutations from a widespread, uniform original specific type. In such a way the forms pentagonum, leiocarpum, Criollo, Cundiamor, Angoleta, Nacional Ecuador, Trimtario, etc., and their variants could have originated. This theory does not exclude hybridization as a factor in the multiplicity of forms, but its influence would be secondary instead of basic, as postulated by the Pittier theory. _________________________ Synonyms.-T. pentagonum Bernoulli was characterized by the shape of the fruit and by smaller flowers. Last character is inconstant but the fruit form is a very particular one and constant, fruit always easily recognized. This type was described in vivo by Bernoulli from the Atlantic coast of Guatemala; there is no type specimen of the fruit, but it was well drawn (PL 2, fig. Ill) and the drawing may be considered the type. It is my belief that T. pentagonum is just a cultivar. It crosses easily with other forms of cacao giving intermediate products. Gross morphological study also supports this view. The pericarp in Theobroma is composed of three visible layers, one of these being more or less consistently woody; in T. cacao the woody layer is the middle one, the mesocarp. It seems that in cultivation there is a tendency for the pods to change, the shells becoming thinner in the best quality Criollos. This reduction is extreme in T. pentagonum which lacks the firm mesocarp and the fleshy endocarp, the whole pericarp being only half as thick as in other cultivars; pentagonum trees are also weaker than others. I have no doubt that pentagonum is a fixed product of mutation selected for better fruits. I must agree with Soria (1959) when he writes "Pentagona is nothing more than one of the extremes in the variability of the complex of types forming the species T. cacao." The name must be kept but only at the rank of forma. Theobroma leiocarpum was also described by Bernoulli from living specimens in cultivation on the Atlantic coast of Guatemala; there are no type specimens of the fruits recorded, and so the published drawing (PL 2, jig. II) must be chosen as the type. This plant was characterized by a glabrous ovary and smaller flowers (an insignificant feature) and by ovoid, shallowly 5-sulcate, almost smooth fruits. The color of the cotyledons was not mentioned. For a long time this form was identified with the widespread South American "Calabacillo" or "Amelonado" fruit types, especially since Pittier published his theory, it being supposed that the Venezuelan "Calabacillo," with thickshelled, rounded or ellipsoid, obtuse, slightly 10-furrowed fruits, and violet cotyledons, was T. leiocarpum. All workers followed this nomenclature, myself included. It was Cheesman (1944, p. 14) who called attention to the differences between Calabacillo and Central American T. leiocarpum. The Bernoulli cacao has thin, ovoid attenuate shells, with only 5 furrows, and plump seeds which are probably white or light violet. Preuss had written previously that the seeds were different. I saw Calabacillo in Nicaragua with very light-violaceous seeds. The recent observation by Soria (1961) in Mexico of the great variety in external form of Criollo in an old plantation of Criollo makes it clear that T. leiocarpum is a mere segregate form or mutant from the Criollo type of T. cacao, originating in a similar way to T. pentagonum. I consider it only as a cultivar. Theobroma sphaerocarpum was described by Chevalier from cultivation in Isla São Thomé, in the South Atlantic Ocean west of tropical Africa, conforms perfectly with the "Calabacillo" cultivar of Venezuela and other South American plantations. It is an extreme form of the widespread South American subspecies. The type is a preserved fruit in the Museum in Paris. This name has to take the place of the subspecies which Pittier and later authors have called T. leiocarpum. Cacao sativa Aubl. is a nomenclatural synonym of T. cacao, and cannot bo used as a substitute for the latter. Any imprecision implied by the binomial T. cacao is implied also in Cacao sativa, and Theobroma sativum. Theobroma sapidum Pittier may well have been unintentionally published as a lapsus calami for T. sativum; it corresponds to T. cacao sensu stricto of Pittier, restricted to the Criollo type. But the binomial is a nomen nudum, because it was not formally published, not being accompanied by a description and with no indication of any type. Cacao minus Gaertn. was published without mention of specimens or locality. It agrees well with some forms of the Criollo type. It cannot be T. pentagonum because in T. pentagonum the ridges are always very prominent and smooth. The type of the binomial consists of two fruits identical with the original drawing, labeled Cacao minus, Gaertner at Paris. Cacao theobroma de Tussac, T. integerrima Stokes, and T. caribaea Sweet are nomenclatural synonyms. Theobroma saltzmannianum was established by Bernoulli, the emarginate petal-laminae being the only difference from T. cacao (ligulae lamina anguste obovata, apice truncata emarginata). According to Bernoulli the shape of the ligula was a constant character in T. cacao and T. leiocarpum; having found at Kew some specimens collected by Salzmann near Bahia, in which he saw the petal-lamina emarginate, he did not hesitate to make a new species. Schumann could not identify this character in any of many flowers collected in Bahia by Salzmann, and inferred that Bernoulli had examined some exceptional, teratological flower. Kombouts in 1948 (Kew Bull. 1948: 104) studied in detail the type specimens at Kew and arrived at the same conclusion as Schumann, that Bernoulli did base his species on an accidental character. Chevalier had already expressed the same view (1946, p. 270). I also can confirm the above opinions after having examined several Salzmann collections at different times and the type specimen in 1954 at Kew. Theobroma sagittata Pavon was published by Chevalier in his revision (1946: 274) as a microspecies of the complex of T. cacao. The binomial, however, is not validly published for lack of the required Latin diagnosis. Moreover, I have identified the type specimen, which is preserved in Paris, as a 3-leaflet fragment of a leaf of Herrania nitida (Poepp.) Schultes. Theobroma hastata, a name mentioned by Chevalier in the footnote on page 273, presumably is a lapsus calami for T. sagittata. The varieties leucosperma and melanosperma published by Chevalier without reference to any type specimens are fortunately not validly published for lack of Latin diagnoses. It would be very difficult to ascribe these two supposed varieties to any recognized taxonomic entity relying only on the seed color. ______________________ Subspecific divisions.—A correct classification of all forms will only be possible after careful and extensive genetic research. In the meanwhile we cannot do more than use a provisional, conservative approach, confining the nomenclature to names already used. The summarized classification given below of cultivated varieties follows Morris, who was the first to give status to the most popular common names of cocoa cultivars; the adaptation in general use made by van Hall is followed in this paper, with few modifications. See also my reviews of Preuss (1901), and Hart (1892, 1900 and 1911) in the Historical Sketch above. The reader will find more extensive information on this subject in special treatises (van Hall, Baker in Urquhart, Hart, etc.). Key to subspecies and forms of Theobroma cacao L. 1. Fruit elongate, claviform, fusiform or ovoid-oblong, tapering and pointed, more or less strongly 10-costate or 5-costate and verrucose; pericarp moderately thick, the woody mesocarp thin; seeds ovoid or ellipsoid, usually rounded in cross section; cotyledons white or yellowish white.....7a. subsp. cacao 2. Fruit 10-costate .... 1. f. cacao, lacandonense and unnamed forms. 2. Fruit claviform, strongly 5-costate, the ridges prominent and smooth, the sides strongly verrucose, the pericarp thinner, lacking the woody mesocarp and endocarp .....2. f. pentagonum 2. Fruit ovoid, shallowly 5-furrowed, almost smooth, obtusely attenuate at apex .....3. f. leiocarpum 1. Fruit ellipsoid, almost globose or more or less oblong, rounded at both ends, smooth or slightly verrucose, more or less shallowly 10-furrowed; pericarp very thick, the woody mesocarp firm; seeds ovoid, more or less compressed; cotyledons purplish or dark violet. ..... 7b. subsp. sphaerocarpum