geological facts, therefore, adduced by the author were true and unquestionable, yet the conclusion that certain chains were or were not simultaneously upraised is by no means a legitimate consequence.
In the third volume of my first edition of the Principles, which appeared in April, 1833, I controverted the views of M. de Beaumont, then just published, in the same terms as I have now restated them. At that time I took for granted that the chronological date of the newest rocks entering into the disturbed series of the Pyrenees had been correctly ascertained. It now appears, however, that some of the most modern of those disturbed strata belong to the nummulitic formation, which are regarded by the majority of geologists as Eocene or older tertiary, an opinion not assented to by M. E. de Beaumont, and which I cannot discuss here without being led into too long a digression.[248]
Perhaps a more striking illustration of the difficulties we encounter, when we attempt to apply the theory under consideration even to the best known European countries, is afforded by what is called "The System of the Longmynds." This small chain, situated in Shropshire, is the third of the typical systems to which M. E. de Beaumont compares other mountain ranges corresponding in strike and structure. The date assigned to its upheaval is "after the unfossiliferous greywacke, or Cambrian strata, and before the Silurian." But Sir R. I. Murchison had shown in 1838, in his "Silurian System," and the British government surveyors, since that time, in their sections (about 1845), that the Longmynds and other chains of similar composition in North Wales are post-Silurian. In all of them fossiliferous beds of the lower Silurian formation, or Llandeilo flags are highly inclined, and often vertical. In one limited region the Caradoc sandstone, a member of the lower Silurian, rests unconformably on the denuded edges of the inferior (or Llandeilo) member of the same group; whilst in some cases both of these sets of strata are upturned. When, therefore, so grave an error is detected in regard to the age of a typical chain, we are entitled to inquire with surprise, by what means nine other parallel chains in France, Germany, and Sweden, assumed to be "ante-Silurian," have been made to agree precisely in date with the Longmynds? If they are correctly represented as having been all deposited before the deposition of the Silurian strata, they cannot be contemporaneous with the Longmynds, and they only prove how little reliance can be placed on parallelism as a test of simultaneousness of upheaval. But in truth it is impossible, for reasons already given, to demonstrate that each of those nine chains coincide in date with one another, any more than with the Longmynds.
The reader will see in the sequel (chap. 31[249]) that Mr. Hopkins has inferred from astronomical calculations, that the solid crust of the earth cannot be less than 800 or 1000 miles thick, and may be more. Even if it be solid to the depth of 100 miles, such a thickness would be inconsistent with M. E. de Beaumont's hypothesis, which requires a shell not more than thirty miles thick, or even less. Mr. Hopkins admits that the exterior of the planet, though solid as a whole, may contain within it vast lakes or seas of lava. If so, the gradual fusion of rocks, and the expansive power of heat exerted for ages, as well as the subsequent contraction of the same during slow refrigeration, may perhaps account for the origin of mountain-chains, for these, as Dolomieu has remarked, are "far less important, proportionally speaking, than the inequalities on the surface of an egg-shell, which to the eye appears smooth." A "centripetal force" affecting the whole planet as it cools, seems a mightier cause than is required to produce wrinkles of such insignificant size.