The then newly opened prolongation of the Estrada de Ferro Dom Pedro II. to Queluz de Minas was thus affected in January, 1884, and a long bank had to be remade, causing a stoppage of traffic for some weeks. M. Liais considers the decomposition of the gneiss to be wholly due to atmospheric causes. Mr. Darwin thinks that it took place at a period of subsidence under the sea before the valleys were carved out.* Mr. C. F. Hartt, who accompanied M. Agassiz in his explorations in Rio, Bahia, and the Amazon valley, believes "that it has taken place only in regions anciently or at present covered by forest," and says, "This decomposition results, in my opinion, from the action of the warm rain-water soaking through the rock and carrying with it carbonic acid, derived not only from the air, but from the vegetation decaying upon the soil, together with organic acids, nitrate of ammonia,” etc. He states that the same phenomena is observed near New York and in the Neilgherries. M. Agassiz in his "Journal in Brazil," † lays great stress on glacial action. This is a matter which deserves a few remarks, as it is of paramount importance.

I never observed any transported blocks, or striations, or roches moutonnées, or anything that could be considered as glacial drift. M. Agassiz is obliged to admit that he never saw any striations which are such a characteristic feature of glacial phenomena, but attributes their disappearance to atmospheric decomposition. Mr. Hartt began his explorations strongly opposed to the glacial theory; but the result of his examinations and intercourse with M. Agassiz led him fully to acquiesce in the professor's theory of glaciation. I must be content with summarizing in a few words M. Liais's statement, and confess that I lean strongly to accepting his views. I may also state that Dr. G. S. de Capanema, in his work, "Decomposicão dos Penedos do Brazil," disbelieves in the glacial origin of the surface deposits claimed by Professor Agassiz and Mr. Hartt to be drift, and rather considers them to be the work of decomposition alone. Mr. Hartt lays much stress on the absence of stratification near the surface, attributing this

"Geological Observations," p. 428.

+ "Geological and Physical Geography of Brazil,” p. 25.
Published at Rio, 1866.

to the deposit of glacial drift. M. Liais explains that the action of vegetation, animals, insects, dry and wet seasons, wind, etc., are sufficient to destroy the evidence of stratification in these cases. Professor Drummond's article on the "Work of the Termites,' "* in which he claims that these insects carry on in the intertropical regions a similar work to that of the earthworms in the temperate zones, at least affords a comfirmation of M. Liais's view; and I have alluded to the parti-coloured heaps which they construct, and which forms perhaps the most noticeable feature of the campos. This proves from how many different strata even one nest is made, and the burrowing of these insects, with the cabeçudo ants (Atta cephalotes), and the armadillos, must necessarily destroy all traces of stratification to a considerable depth.

"Mr. Belt calculates that the vast amount of water abstracted from the ocean and locked up in mountains of ice around the two poles would lower the general level of the ocean about two thousand feet. This would be equivalent to a general elevation of the land to the same amount, and would thus tend to intensify the cold; and the subsidence of the ocean would produce a tract of lowland of an average width of some hundreds of miles, added to the whole east coast of Central and South America. This tract would no doubt become covered with forests as it was slowly formed, would enjoy a perfectly tropical climate, and would thus afford an ample area for the continued existence and development of the typical South American fauna; even had the glaciers descended in places so low as what is now the level of the sea.Ӡ

I would suggest that a difference of two thousand feet in the level of the highlands above the sea would not suffice for the production of the supposed glaciers. This presumption of a glacial period in the tropics presents great difficulties. First, a vast expanse of collecting ground is necessary at a considerable altitude for the névé to consolidate into glaciers at a lower level; secondly, it requires for a very prolonged period a very low temperature in the intertropical regions, which would presumably mean such an intense cold in what are now the temperate climes that no life could there exist. I cannot see that either of these * Good Words, May, 1885.

Geographical Distribution of Animals," vol. i. p. 151, A. R. Wallace.

premises is proved, especially by osteological relics. M. Liais enters into the glacial theory at great length, but I must refer my readers to his book for a full consideration of the subject. He holds that so far from the equatorial region being less warm at any past epoch, the temperature must have been higher to permit the formation of the vast amount of vapour necessary to produce the snows and ice in the temperate zone, which brought about our glacial period. Besides which, no rotation of the earth's axis. could make the equatorial region even temporarily polar. M. Liais also cites the existence of remains of animals in the caves covered by the same red clay which M. Agassiz considers to be drift. These animals must have lived prior to the supposed glacial epoch, yet they are identical with existing species now found in the same localities.

M. Liais also shows that the so-called erratic blocks are in close proximity to the virgin rock whence they have been derived. This is also the case with the angulated quartz pebbles described by Mr. Hartt as water-carried stones lying under the drift clay; dykes and veins of the same material are always in close proximity. M. Liais says, "When one is assured that these blocks come really from the region where they are met with; when one sees them sometimes still partially fixed in the decomposed gneiss with its primitive stratification" (as I have described on the Corcovado and Petropolis Railways); "when, finally, one observes the vast scale on which the decomposition of rocks by atmospheric action is carried on in Brazil;-these phenomena have a simple and natural explanation, excluding entirely the idea of transport." The rounded forms which some of these blocks present is also attributable to atmospheric action. M. Agassiz even describes rocks in situ with the same peculiarities.

It is very difficult to estimate the thickness of the gneiss formation in Brazil. In the Corcovado range, M. Liais estimates it at a thousand metres. In the Organ Mountains and the Mantiquiera range, in the province of Rio de Janeiro and South Minas Geraes, at six thousand metres.