Geographical variation and population biology in wild Theobroma cacao.

Abstract

Geographical variation in wild Upper Amazon populations of cocoa (Theobroma cacao) was studied with a view to improving germplasm collecting strategies for cocoa. Morphological descriptors, which covered pod, seed, flower, flush and tree habit characters, were assessed in 419 accessions from the Amazon Region of Ecuador, 136 accessions from the Rio Caqueta in Colombia, and about 50 clones of wild origin from Peru. The Ecuadorian and Colombian material was collected in the course of the London Cocoa Trade Amazon Project (LCTAP), and data were acquired both at the time of collection and in the field genebank at San Carlos, Napo Province, Ecuador. The Peruvian material was examined in the genebank at Pichilingue, Ecuador. There were clear differences between the Ecuador, Caqueta and Peru accessions, but there was relatively little variation within Ecuador, so that the initial sampling strategy used to survey variation in Ecuador effectively captured most of the variation in that target region. Using cost data for LCTAP, it is argued that marginal costs for additional accessions are higher than is generally assumed, so that it is not cost-effective to acquire many similar accessions. Field observations from LCTAP collecting trips as well as data from the San Carlos field genebank and from a study population of wild cocoa at San Carlos were used to build a picture of the population biology of Theobroma cacao in its natural habitat. Cocoa was consistently associated with certain soil and terrain types in Ecuador. Observed rates of seed and vegetative reproduction were low, but adult mortality was also low; cocoa could be characterized as a shade-bearing, K-adapted species, although short-term increases in seed production and growth rates resulted from shade reduction. Seed dispersal by mammals resulted in most seedlings germinating in groups close to parent trees, with high early mortality resulting from mammalian herbivory; after three months, mechanical damage was the most frequent cause of seedling mortality. Approximate tree ages were estimated from the relationship between basal area and growth rate in the study population over a two year period. The age distribution showed a peak in the age class 60-80 years, with fewer older or younger trees. Possible causes of disturbance in the past were reviewed. Observations made in the course of this project, combined with published information on Upper Amazon ethnobotany, support the view that Theobroma cacao is a genuinely wild species whose present-day distribution in the Upper Amazon has been only marginally affected by human activities. It is argued that the Vavilovian centre of diversity-centre of origin model is not appropriate for cocoa, and that the Upper Amazon cocoa gene pool is valuable to breeders because it is distinct from the gene pool of cultivated cocoa.