These natural features of the continent, compared with the actual distribution of languages, offer our only guides in forming an opinion as to the migrations of these various families in ancient times. Their traditions, take even the most cultivated, are confused, contradictory, and in great part manifestly fabulous. To construct from them by means of daring combinations and forced interpretations a connected account of the race during the centuries preceding Columbus were with the aid of a vivid fancy an easy matter, but would be quite unworthy the name of history. The most that can be said with certainty is that the general course of migrations in both Americas was from the high latitudes toward the tropics, and from the great western chain of mountains toward the east. No reasonable doubt exists but that the Athapascas, Algonkins, Iroquois, Apalachians, and Aztecs all migrated from the north and west to the regions they occupied. In South America, curiously enough, the direction is reversed. If the Caribs belong to the Tupi-Guaranay stem, and if the Quichuas belong to the Aymaras, as there is strong likelihood,[34-1] then nine-tenths of the population of that vast continent wandered forth from the steppes and valleys at the head waters of the Rio de la Plata toward the Gulf of Mexico, where they came in collision with that other wave of migration surging down from high northern latitudes. For the banks of the river Paraguay and the steppes of the Bolivian Cordilleras are unquestionably the earliest traditional homes of both Tupis and Aymaras. These movements took place not in large bodies under the stimulus of a settled purpose, but step by step, family by family, as the older hunting grounds became too thickly peopled. This fact hints unmistakably at the gray antiquity of the race. It were idle even to guess how great this must be, but it is possible to set limits to it in both directions. On the one hand, not a tittle of evidence is on record to carry the age of man in America beyond the present geological epoch. Dr. Lund examined in Brazil more than eight hundred caverns, out of which number only six contained human bones, and of these six only one had with the human bones those of animals now extinct. Even in that instance the original stratification had been disturbed, and probably the bones had been interred there.[35-1] This is strong negative evidence. So in every other example where an unbiased and competent geologist has made the examination, the alleged discoveries of human remains in the older strata have proved erroneous.