In a previous chapter this law was given as follows: "Of any two strata, the one which was originally the lower is the older;" the general truth of the law depends upon the fact that except under very exceptional circumstances the strata are deposited upon the surface of the lithosphere, and not beneath it. There are occasions where strata may be deposited beneath the lithosphere, but as a general rule the geologist will not be misled by such occurrences. In caverns, accumulations often occur which are newer than the strata over the cavern roof, and so long as caverns are formed in ordinary sedimentary rocks, no great difficulty will result from this exception to the law of superposition. When caverns occur beneath masses of land ice, the order of superposition may be misleading. A deposit may be formed on the surface of the ice, and subsequently to this a newer deposit may be laid down in a sub-glacial or englacial cavern; upon the melting of the ice the newer deposit would be found with the older one resting upon its surface. Apart from these exceptional cases, the law as stated holds good, but the reader will notice the insertion of the word 'originally' which requires some comment. A geologist speaks of one bed lying _upon_ another not only when the beds are horizontal, but when they are inclined at any angle, until they become vertical, so that until beds have been turned through an angle of 90° by earth-movement the test of superposition is applicable, but when they have been turned more than 90°, the stratum which was originally lower rests upon that which was originally above it, and in the case of these _inverted_ strata, the test of superposition is no longer applicable. It was formerly supposed that cases of inversion were comparatively rare and local, and that the test of superposition could therefore be generally applied with confidence, but it is now known that though this is generally true of such strata as have been subjected only to those widespread, fairly uniform movements which are spoken of as _epeirogenic_ or continent-forming, where the radius of each curve is very long, inversion is a frequent accompaniment of the more local _orogenic_ or mountain-forming movements, where the radius of a curve is short. Though orogenic movements are limited as compared with those of epeirogenic character, they often affect large tracts of country, in which case the apparent order of succession of the strata need not be the true one, and examples of inversion may be frequent[9]. [Footnote 9: For a discussion of the principles of mountain-building see Heim, A., _Untersuchungen über den Mechanismus der Gebirgsbildung_, and Lapworth, C. "The Secret of the Highlands," _Geological Magazine_, Decade II. vol. x. pp. 120, 193, 337.]