Page:The Journal of geology (IA journalofgeology11893univers).pdf/347

a single scratch either upon the rocks in place or upon a boulder, cobble, or pebble, that could, by any legitimate stretch of the imagination, be attributed to glacial action. And it is but just to recall the fact that both Agassiz and Hartt recognized this as the one piece of evidence, above all others, lacking for their Brazilian glacial theory. How diligently Agassiz searched for such evidence one can judge from the story of his journey as told by Mrs. Agassiz and himself, and I know that Hartt left no stone unturned and no locality unexplored that he thought might afford him the long-sought striæ. They both explained the absence of such marks by supposing that they had been obliterated by the decomposition of the rocks, and Agassiz believed that in the Amazon region there were no rock surfaces exposed.^[1] But it cannot be considered credible that glacial striæ should have been preserved in Asia, Africa and Australia since Carboniferous times,^[2] but entirely obliterated in Brazil, both from the bed rocks and from the conglomerates deposited in post-tertiary times, or as has already been mentioned, that the pitted and water-worn faces should have been preserved in these materials while the ice marks should have been obliterated.

James E. Mills, a professional geologist and a former pupil of Agassiz at Harvard, spent nearly two years in Brazil in the states of Rio Grande do Sul, Rio de Janeiro, and Minas Geraes. He expresses his views of the subject of glaciation in that country as follows:^[3] "In those portions of Brazil which came within my field of observation there is no glacial drift, and there are no glaciated rock surfaces or glacial topography or other signs of the existence of glaciers."

Agassiz points out the weakness of his own theory regarding Brazilian glaciation very nicely in his letter to Professor Pierce,