[[p. 510]] This thick volume of 620 pages is the result of two visits to Brazil, the first with the Thayer Expedition, the second during a vacation holiday of "some months." The author has proposed to combine with his own personal observations all the information on this subject obtainable from other sources, and thus give a complete view of the present state of our knowledge of the geology and physical geography of this vast and interesting region. The design is an admirable one, but the execution of it is, in some respects, disappointing.

    The first great fault of the book is, that it has been swelled by the introduction of much irrelevant matter. Mr. Hartt's own journeys were mainly along the coast, from Rio Janeiro to the Amazon, with occasional trips of a hundred miles or so into the interior, and he inflicts upon us pages of unimportant detail on the topography of small rivers, creeks, and harbours, which have no bearing on the geology or physical geography of the country. Detailed descriptions of the marine animals and fossils collected would also have been better in an appendix than in the body of the work where they are given. The arrangement of the book, too, is faulty, since it treats of the provinces of Brazil in succession, and makes no attempt to indicate the great physical divisions of the country, and there is not a single geological or physical map of Brazil, or of any part of it; the maps alluded to in the title-page being mere outline or sketch maps of small districts, or plans of harbours and mouths of rivers. Another strange defect is the absence of all measurement of heights. The author travelled without barometer or aneroid; he, consequently, everywhere roughly estimates his heights, and gives no sections, but a few "ideal" ones. Notwithstanding the bulk of this volume, it does not complete the geology of the voyage, for we are informed that Mr. St. John, another geologist attached to the expedition, who travelled more in the interior of the country, will give the results of his observations in a separate work.

    But although we have thus plainly indicated the defects of the book, there is much valuable matter to be found in it. The author has been very diligent in examining all the chief authorities on Brazil, and has extracted from them most of their geological matter; and among the extracts from Spix and Martius, Prince NeuWied, Darwin, Gardner, Halfield, and others, are to be found many interesting passages descriptive of the peculiarities of the scenery and geology of the country. The chapters on the coral reefs of the Abrolhos and on the gold mines of Brazil, the account of the exploration of the bone-caves by Lund, and the appendix on the Botocudos Indians, will furnish some interesting matter for the general reader, while the student of science will obtain (though with some difficulty) a notion of the general physical and geological characteristics of an almost unexplored region.

    The most striking geological feature of tropical South America east of the Andes is the enormous extension of gneissic rocks, which appear to form the whole foundation and much of the surface of the country, from the cataracts of the Orinooko to Paraguay and the southern frontier of Brazil. All the great mountain tracts of Brazil and Guiana, as well as the low plain which separates the watersheds of the Orinooko and Amazon, are of this rock, which is considered to be of Laurentian age. Its characteristic features are the great dome-like masses and the conical peaks or pillars, generally of more or less smooth and rounded outlines, a peculiarity dependent on the decomposition of all exposed surfaces, which fall away in concentric flakes. Great hemispherical domes up to a thousand feet in diameter are one of the results of this decomposition wherever a more resisting mass has occurred. Still more extraordinary are the vertical pillars of rock, that rise up at intervals out of the forest to some hundreds, or, in the case of the Pedra lisa, in the province of Rio de Janeiro, to more than three thousand feet high. Similarly formed peaks or pillars in Fernando, Noronha, and St. Helena have been formed by injections of fluid felspathic lava. What an enormous amount of denudation do these isolated pillars indicate!

    In South Brazil a few tracts of Silurian and Carboniferous rocks occur, but the next formation of any extent is the Cretaceous, which consists of sandstones, generally upheaved and fractured. Other sandstones, which cover an immense extent of country, and form the ranges of flat-topped hills from one to nearly three thousand feet high, called taboleiros, are in perfectly horizontal strata, and as these lie unconformably on the cretaceous rocks they are presumed to be tertiary, although no fossils have yet been found in them.

    We now come to a very wide-spread, yet recent and superficial deposit, which is at once the most puzzling and the most interesting feature in Brazilian geology. This is a layer of clay or loam, varying in thickness from a few feet to one hundred, and wrapping in its folds hill and valley, over vast tracts of country, including the steep slopes and summits of some of the highest mountains. All Rio de Janeiro, and all the coast provinces visited by our author, were thus covered. It has been described in Minaes Geraes and San Paulo, and Prof. Agassiz has observed it in all the northern provinces as far as the Amazon valley. It covers alike the gneiss and the tertiary formations. This clay is of a red colour, and is evidently formed of the materials of the adjacent and under-lying rocks, but ground up and thoroughly mixed. There is never the least sign of stratification throughout its mass, although it very frequently rests on a thin layer of quartz pebbles. It contains, scattered through it, rounded and angular boulders of quartz, gneiss, and other rocks and the surfaces upon which it rests are always more or less smooth and rounded. Our author always speaks of this formation as "drift," and he agrees with [[p. 511]] Prof. Agassiz that its peculiarities are such as unmistakeably to indicate its glacial origin.

    This is truly a startling conclusion, and one which has hitherto been received in this country with some incredulity. Prof. Agassiz was thought to be glacier-mad; but if we separate his theories from his facts, and if we carefully consider the additional facts and arguments adduced by Mr. Hartt in this volume, we shall be bound to conclude that, however startling, the theory of the glaciation of Brazil is supported by a mass of evidence which no unprejudiced man of science will ignore merely because it runs counter to all his preconceived opinions.

    Mr. Hartt's facts and deductions have the more weight, because he is evidently not very enthusiastic on the subject, and because he fairly states the sources of error in observation, and fully discusses such other modes of explaining the facts as have been proposed. He acknowledges that in some cases the decomposed gneiss cannot be distinguished from the "drift;" but he shows that in the former the materials remain in situ, especially the quartz veins, while in the latter all are mixed and ground up together; and wherever the two are seen in contact for any distance, the sudden cutting off of the quartz veins at the drift, and other well-marked characters, render it impossible to confound them. He also adduces several other phenomena which are strongly indicative of a glacial origin. Both in the Orang Mountains and in Bahia there are valleys with no outlet, and in Alagoas there are many deep lakes in rock-basins. In the province of Bahia there are extensive bare, elevated, rocky plains thickly strewn with angular blocks of stone, some of which are erratics, and exactly resembling the drift-covered plains of the north. On similar elevated plains, far removed from any higher land, Mr. J. A. Allen (another member of the Thayer Expedition) found numerous deep and smooth pot-holes worn in solid gneiss. They were of various sizes, the largest seen being elliptical, eighteen feet long by ten wide, and twenty-seven feet deep. Similar pot-holes are known to be formed by glacial waterfalls, and they are found over the glaciated regions of New Brunswick and Nova Scotia. Heaps of débris, exactly resembling glacial moraines, have also been found both in the south and north of Brazil. Mr. Hartt is satisfied of their resemblance to true moraines in the valley of Tijuca near Rio, and Prof. Agassiz has described others still more perfect in Ceara, only four or five degrees south of the equator. After describing these in detail, he concludes: "I may say that in the whole valley of Hasli there are no accumulations of morainic materials more characteristic than those I have found here, not even about the Kirchet; neither are there any remains of the kind more striking about the valleys of Mount Desert in Maine, where the glacial phenomena are so remarkable; nor in the valleys of Loch Fine, Loch Awe, and Loch Long, in Scotland, where the traces of ancient glaciers are so distinct." It can hardly be maintained that the discoverer of glacial phenomena in our own country, and who has since lived in such a pre-eminently glaciated district as the Northern United States, is not a competent observer; and if the whole series of phenomena here alluded to have been produced without the aid of ice, we must lose all confidence in the method of reasoning from similar effects to similar causes which is the very foundation of modern geology. The only objection of any weight that has been made to this interpretation of the phenomena, is the fact of the absence of glacial striæ but Mr. Hartt states that no striæ or polished surfaces have yet been found even in the extreme south of the continent where the glacial phenomena are undisputed. It is at best a negative argument, and as such cannot neutralise those of a positive nature. We must also remember that we have no indication of the age of the Brazilian or southern glacial epoch. It may have occurred much earlier than that of the northern hemisphere, and the greater lapse of time, combined with the powerful decomposing and denuding agency of tropical rains, may have obliterated most of such marks.

    The physical difficulty of the conception of tropical glaciers may be resolved into the question of whether a subsidence to the extent of ten or twelve thousand feet may not have subsequently occurred; since a greatly increased elevation at a time when a severe glacial epoch reigned in the south temperature zone, would probably lead to the glaciation of the southern tropics down to what is now the sea-level.

    A much more serious objection seems to be the biological one. If the whole surface of what is now Brazil was recently covered with a sheet of ice, whence has arisen the wonderfully rich and varied, and, in many respects, peculiar flora and fauna that now inhabit it? Judging from the map of the Atlantic given in Maury's "Physical Geography of the Sea," a rise of twelve thousand feet would only add a belt of about four hundred miles in width to tropical America; but a great part of this might have enjoyed a truly tropical climate, just as the valleys of Switzerland have a warm summer though in the immediate vicinity of glaciers. It seems probable, also, that the glaciation was a southern one, and did not extend far north of the equator, if it even reached so far, so that the whole of Venezuela and Guiana, with the additional belt of land due to elevation, might have been even more luxuriant and more densely populated than at present. There would thus have been an ample surface to support the ancestors of the existing fauna and flora of Brazil during the glacial epoch, just as there was sufficient land in Europe to support the ancestors of the existing European fauna and flora even when so much of the present surface was covered by a thick mantle of ice.

    It must be stated that Mr. Hartt does not accept Prof. Agassiz's extraordinary hypothesis (which rests on a very slender basis of fact) of a great Amazonian glacier. He believes that the wide-spread beds of clays and sandstones, which Prof. Agassiz classes as glacial, are marine, and states that they agree perfectly with the tertiary beds in other parts of Brazil. The patches of drift, with erratics in the Amazon valley, may well have been produced by detached masses from the glaciers of the Andean and Brazilian highlands, which melted and deposited their load of drift in the warm waters of the ancient Amazon.

    We have devoted so much of our space to this question of the Glaciation of Brazil, in the hope of attracting the attention of geologists to a country which offers such an interesting subject of inquiry, and which it is so easy and agreeable to explore. The facts, as stated by two careful observers, both thoroughly experienced in glacial phenomena, are undoubtedly such as to warrant the main conclusion drawn by them; and it is to be hoped that [[p. 512]] geologists will not ignore the facts because the conclusions seem improbable, as they so long ignored facts proving the antiquity of man for no other reason.