It is especially in studying the vast and largely unexplored regions watered by the upper streams of the mighty Amazon, that one is yet at a loss to bring the native inhabitants into ethnic order. Of the various explorers and travellers who have visited that territory, few have paid attention to the dialects of the natives, and of those few, several have left their collections unpublished. Thus, I have been unable to learn that Richard Spruce, who obtained numerous vocabularies along the Amazon and its branches, gave them to the press; and there were in the hands of Von Tschudi more than a hundred vocabularies collected by the German naturalist, Johannes Natterer, in the interior of Brazil,[224] most of which I learn are still in manuscript. In default of material such as this, the classification of the tribes of Brazil must remain imperfect. It is also a matter of much regret that no copy can be found of the work of the celebrated missionary, Alonso de Barcena, _Lexica et Precepta in quinque Indorum Linguis_, published at Lima, in 1590--if, indeed, it was ever really printed. It contained grammars of the Kechua, Aymara, Yunca, Puquina and Katamareña, (spoken by the Calchaquis). Of the two last mentioned idioms no other grammar is known, which makes the complete disappearance of this early printed book particularly unfortunate. Another Jesuit, Father Guillaume D’Étré, wrote out the catechism and instructions for the sacraments in eighteen languages of eastern Peru and the upper Orinoco;[225] but this, too, seems lost. Of late years no one has paid such fruitful attention to the relationship and classification of the South American tribes and languages as M. Lucien Adam. Although I have not in all points followed his nomenclature, and have not throughout felt in accordance with his grouping, I have always placed my main dependence on his work in the special fields he has selected--the three great South American families of the Amazon region, the Arawak (called by him the Maypure), the Carib, and the Tupi.[226] The general plan which I shall adopt is rather for convenience of arranging the subject than for reasons based on similarities either of language or physical habitus. It is that which allows the presentation of the various stocks most in accordance with their geographic distribution and their historic associations. It is as follows: I. The South Pacific group. 1. The Columbian region. 2. The Peruvian region. II. The South Atlantic group. 1. The Amazonian region. 2. The Pampean region. I. THE SOUTH PACIFIC GROUP. 1. THE COLUMBIAN REGION.