Taxonomía del género Arachis (Leguminosae)

Almost 100 years elapsed between Linnaeus' naming the then lone species of Arachis (A. hypogaea L.), known to Europeans, and the first taxonomic treatment of the genus by Benthamin 1841. During the next 100 years five to ten additional species descriptions appeared,assigning different species to the same names,different names to the same species. By mid20th century, it was impossible to examine anyherbarium collection of Arachis and assign any epithet with any assurance to any specimen (which was not a type collection) except to A. hypogaea, A. guaranitica, A. tuberosa and A. villosulicarpa. In our treatment the Iiterature of this botanical chaos in Arachis is reviewed in detail and an assessment is made of the foundations for its occurrence. It is shown that the bases for the confusion lay in the combination of the esoteric nature of the differentiating morphological features of Arachis, the fragmentary early collections and the representation of species by seedling specimens. Also, it is related how, in 1959, we decided to re-explore the type locality of each species then known, collect therein complete plant specimens and thereby resolve the problem. Thirty five years, two generations of plant collectors and around 2000 collections later we present here 69 species descriptions of Arachis, species distributed in South America east of the Andes, south of the Amazon, north of La Plata and from NW Argentina to NE Brazil. We soon discovered that the most significant characters of Arachis lay in their underground structures, including their fruits, rhizomatous stems, root systems and hypocotyls. We showed that these defining characters tended to cluster the collections into groups which were associated with generally different geographic areas and ecological features. We drew a sample of 100 collections representing these c1usters, areas and features and arranged them in a hybridization diallel and showed, in crosses between collections representing different c1usters of characters, areas and features, a remarkable number of complete failures to cross-fertilize and in those hybrids which were recovered a high degree of F1 hybrid infertility. When these cross-incompatibilities and pollen infertilities were combined with the data on character-clustering, the nine distinct sections of the genus presented here then crystallized. Figures imposed upon maps of South America iIIustrate the geographic distributions of these sections. The collections were then assigned