[94]. To-day I cannot give a more precise account of the final result of my survey. Though all my collections are safely stored in the Museum, every practical zoölogist understands that a critical examination of more than eighty thousand specimens cannot be made in less than several years.—L. A.
[95]. The rest of this letter is omitted, as its substance is contained in Chapter XIII., on the Physical History of the Amazons.
[96]. See Chapter III. p. [86].
[97]. The name consecrated by De Saussure to designate certain rocks in Switzerland which have had their surfaces rounded under the action of the glaciers. Their gently swelling outlines are thought to resemble sheep resting on the ground, and for this reason the people in the Alps call them roches moutonnées.
[98]. See “Glacial Phenomena in Maine,” Atlantic Monthly, 1866.
[99]. The atlas in Martius’s “Journey to Brazil,” or the sketch accompanying Bates’s description of these hills in his “Naturalist on the Amazons,” will give an idea of their aspect.
[100]. Bohn’s edition of Humboldt’s Personal Narrative, Chap. II. p. 134. Humboldt alludes to these formations repeatedly: it is true that he refers them to the ancient conglomerates of the Devonian age, but his description agrees so perfectly with what I have observed along the banks of the Amazons and the Rio Negro that there can be no doubt he speaks of the same thing. He wrote at a time when many of the results of modern geology were unknown, and his explanation of the phenomena was then perfectly natural. The passage from which the few lines in the text are taken shows that these deposits extend even to the Llanos.
[101]. I am aware that Bates mentions having heard that at Obydos calcareous layers, thickly studded with marine shells, had been found interstratified with the clay, but he did not himself examine the strata. The Obydos shells are not marine, but are fresh-water Unios, greatly resembling Aviculas, Solens, and Arcas. Such would-be marine fossils have been brought to me from the shore opposite to Obydos, near Santarem, and I have readily recognized them for what they truly are,—fresh-water shells of the family of Naiades. I have myself collected specimens of these shells in the clay-beds along the banks of the Solimoens, near Teffé, and might have mistaken them for fossils of that formation had I not known how Naiades burrow in the mud. Their resemblance to the marine genera mentioned above is very remarkable, and the mistake as to their true zoölogical character is as natural as that by which earlier ichthyologists, and even travellers of very recent date, have confounded some fresh-water fishes from the Upper Amazons, of the genus Pterophyllum (Heckel), with the marine genus Platax.
[102]. As I have stated in the beginning, I am satisfied that the unstratified clay deposit of Rio and its vicinity is genuine glacial drift, resulting from the grinding of the loose materials interposed between the glacier and the solid rock in place, and retaining to this day the position in which it was left by the ice. Like all such accumulations, it is totally free from stratification. If this be so, it is evident, on comparing the two formations, that the ochraceous sandy clay of the valley of the Amazons has been deposited under different circumstances; that, while it owes its resemblance to the Rio drift to the fact that its materials were originally ground by glaciers in the upper part of the valley, these materials have subsequently been spread throughout the whole basin and actually deposited under the agency of water. A survey of the more southern provinces of Brazil, extending to the temperate zone, where the combined effects of a tropical sun and of tropical rains must naturally be wanting, will, I trust, remove all the difficulties still attending this explanation. The glacial phenomena, with all their characteristic features, are already known to cover the southernmost parts of South America. The intervening range, between 22° and 36° of south latitude, cannot fail to exhibit the transition from the drift of the cold and temperate zone to the formations of a kindred character described above from the tropical zone. The knowledge of these deposits will definitely settle the question; and either prove the correctness of my generalizations or show their absurdity. I feel no anxiety as to the result. I only long for a speedy removal of all doubts.
[103]. I would here remind the reader of the terraces of Glen Roy, which indicate successive reductions of the barrier encasing the lake, similar to those assumed to have taken place at the mouth of the Amazons.