In the north of Holland, Germany, and the countries bordering on the Baltic, enormous fragments of granite and syenite are scattered within certain limits. According to Humboldt, it seems to be now proved, that they have been carried southward, with a distribution like that of radii from a centre, from the Scandinavian peninsula, during some of the ancient revolutions of our globe, and that they have not originally belonged to the granitic chains of the Hartz and Saxony, which they approach without, however, actually attaining their basis[374]. Born, says Humboldt, on the sandy plains of the Baltic, and until the age of eighteen, not knowing any other rock than these scattered blocks, I could not but feel curious to know whether the new world presented any thing of a similar nature. I was surprised not to find a single block of this description in the _Llanos_ of Venezuela, although the immense plains were immediately bordered to the south by a group of mountains entirely granitic[375], and which presents, in its broken and almost columnar peaks, traces of the most violent action[376]. Towards the north, the granitic chain of the Silla of Caracas and of Portocabello is separated from the Llanos, by a range of mountains which are schistose between Villa de Cura and Parapara, and calcareous between the Bergantin and Caripe. I was equally struck with the same absence of blocks upon the banks of the Amazon. La Condamine had already affirmed, that from the Pongo of Manseriche to the strait of Pauxis, not the smallest stone was to be observed. Now, the basin of the Rio Nigro and of the Amazon is also but a Llano, a plain like those of Venezuela and Buenos Ayres, the difference consisting only in the state of the vegetation. The two Llanos, situated at the northern and southern extremities of South America, are covered with gramineæ; they are Savannas destitute of trees. The intermediate Llano, that of the Amazon, exposed to almost continual equatorial rains, is a thick forest. I do not remember to have heard that the Pampas of Buenos Ayres or the Savannas of the Missouri[377] and New Mexico contain granitic blocks. The absence of this phenomenon appears general in the new world. It is probably equally so in the Sahara in Africa; for we must not confound rocky masses which pierce the soil in the midst of the desert, and of which mention has often been made by travellers, with mere scattered fragments. These facts seem to prove, that the blocks of Scandinavian granite, which cover the sandy plains on the southern side of the Baltic, in Westphalia, and in Holland, are owing to a particular debacle which proceeded from the north, to a purely local catastrophe. The old conglomerate (grès rouge), which covers a great part of the Llanos of Venezuela and of the basin of the Amazon, contains, without doubt, fragments of those same primitive rocks of which the neighbouring mountains are composed; but the convulsions of which these mountains present undoubted