In several parts of the continent they have been examined by competent observers and the question of their date approximately ascertained. I need not say this differs widely, for these refuse heaps of ancient villages or stations were of course begun at wide intervals.
Long ago I called attention to the singular size and antiquity of those I found in Florida and along the Tennessee River;[13] and the later researches of Professor Jeffries Wyman would, in his opinion, measure the age of some of the former by tens of thousands of years.[14]
Further to the south, in Costa Rica, Dr. Earl Flint has examined the extensive artificial shell deposits which are found along the shores of that republic. They are many feet in height, covered by a dense forest of primeval appearance, and are undoubtedly of human origin.
In Brazil such shell-heaps are called _sambaquis_, and they are of frequent occurrence along the bays and inlets of the coast. Some of them are of extraordinary dimensions, rising occasionally to more than a hundred feet in height. The lower layers have been consolidated into a firm, stony breccia of shells and bones, while the surface stratum, from six to ten feet thick, is composed of sand and vegetable loam supporting a growth of the largest trees. Yet even the lowest layers of this breccia, or shell-conglomerate, yield tokens of human industry, as stone axes, flint arrow-heads, chisels, and fragments of very rude pottery, as well as human bones, sometimes split to extract the marrow. The shells are by no means all of modern type. Many are of species now wholly extinct, or extinct in the locality. This fact alone carries us back to an antiquity which probably should be counted by thousands of years before our era.
At that remote period not only did a fishing and hunting race dwell along the Brazilian coast, but this race was fairly advanced on the path to culture; it was acquainted with pottery, with compound implements, and with the polishing of stone. We further know that this race was not that which occupied the land when the whites discovered it; for the human skulls disinterred from the sambaquis are, craniologically, almost diametrically opposite those of the Botocudos and the Tupis. Yet if we can trust the researches of Dr. Lund in the caverns of Brazil, the oldest skulls in these deposits, found in immediate connection with the bones of extinct mammalia, belonged to the ancestors of these tribes. Markedly dolichocephalic, they present an entire contrast to the brachycephalic type from the sambaquis.[15]