J. H. Pratt (see 11.654) m '855 laid the foundation in India for a generalization that has been widely accepted as an explanation of the support of mountains. He pointed out that observations on the deflection of the pendulum from the vertical indicate a defect of ma:ss in the Himalaya range and an excess of mass towards the Indian Ocean, and he concluded that highlands were upheld by differ- ences of density in the crust. Detailed work is still in progress; but numerous observations with the seconds pendulum, and by comparison of points and mercury barometers at sea, already serve to indicate a general " positive anomaly " of gravity that is, an excess of attraction in the oceanic areas. The " negative anom- alies" associated with high plateaus and mountain-chains are com- pensated by positive anomalies in adjacent plains or under adjacent seais. It is held that denser rocks, such as basaltic magmas, have ac- cumulated beneath the oceans, while mountains are typically formed of less dense material. They may thus be compared with floating bodies; they are supported by the pressure of denser masses which they have in part displaced. At a certain depth below the less dense earth-block under an elevated area, and below the denser earth- block under an adjacent area depressed below sea-level, a region of compensation must exist, where the two blocks balance one another. If of equal area, these two blocksabove this " depth of compensation " will contain the same mass. To geologists, who realize the complex- ities of intrusion, and the interlocking of various types of rock in the outer layers of the crust, the conception of columns of uniform density stretching down from the surface to the depths may seem too much like a purely mathematical expression ; but the theory of isostatic compensation is shown by Jos. Barrell (Am. Journ. Sci., vol. 48, pp. 281338) to relate to broad areas rather than to local irregularities of the surface, and it is probable that the balance is attained much more nearly under continents than under isolated mountains. J. F. Hay ford, of the U.S. Geodetic Survey, has re- viewed (1909-12) the relations of topography to gravity throughout the United States, with results extremely favourable to the isostatic theory. H. S. Washington (Journ. Franklin Inst., 1920, vol. 190, p. 812) supports Hayford's conclusions by an estimate of the average densities of the rocks underlying the areas of elevation and depression, including a consideration of the igneous rocks of oceanic islands and continental regions over the whole globe. G. Costanzi (1910) has correlated surface-relief throughout Europe with anomalies of gravity. E. Suess, in the concluding volume of his Antlite der Erde, seriously doubts if the known range of density in rocks is sufficient to account for the maintenance of the major features of relief. His most important argument is drawn from the Indian Ocean, the floor of which represents sunken continental land. Suess points out that here O. Hecker, from his marine traverses, indicates a region of strong positive anomaly.
Albert Heim (translation in Proc. Liverpool Geol. Soc., vol. 13, 1920) shows how gravity determinations in Switzerland are related to the form of the Alpine chain, positive anomalies occurring only near the Lago Maggiore and under the Black Forest. He explains the latter case by the presence of a gravity anticlinal, in which the denser matter of Suess's " Sima " layer is brought nearer than usual to the surface. The explanation of gravity-anomalies in the Himalayas and the "Gangetic trough" has been the subject of much discussion by Sir S. Burrard, H. H. Hayden and R. D. Oldham.
Great mountain-chains, as was long felt to be the case, evidently bulge both upwards and downwards, and send down crumpled and lighter matter from the upper crust that displaces the denser matter ofthe depths. This process is aided by the presence of a yielding but not necessarily molten layer, Barrell'sasthenosphere, below the depth of compensation, which lies some 80 to 100 m. (say 128 to 160 km.) below the general surface of the gepid. Barrell remarks that " the density of the crust is presumably irregular in depth as well as in distribution, but it is seen to be essentially a phenomenon of the outer fiftieth of the earth's radius."
Isostasy is not put forward as a cause of mountain-building. The relative importance of successive overfolds and of vertical upheaval in establishing these features of relief is still a matter of discussion, and geologists will probably incline more and more to the view of O. Ampferer (1906) that sliding movements in the Untergrund, Bar- rell's asthenosphere, are responsible for drag and crumpling at the surface. Melting of lower layers and consequent vertical foundering may promote extensive movements in a lateral- direction. At the surface the final overlapping of recumbent folds may be largely gravitational, a feature emphasized by Hans Schardt. A renewed appreciation of the importance of vertical uplift and vertical founder- ing leads us back to conceptions of mountain-structure that were prevalent in the early igth century. Ampferer has even pointed out the influence of notches cut by subaerial denudation on folding that may subsequently affect the surface.
In considering folding in connexion with rock-flow, C. K. Leith (Structural Geology, 1914) applies the term competent to rocks that resist crumpling, and incompetent to those that yield contortions. A competent mass under increase of pressure may of course become incompetent. The flow of incompetent rocks between competent layers obviously produces considerable changes in their relations at the surfaces of contact, and complications of this kind may be expected in any pverfolded series. Where faults result from over- folding, E. B. Bailey (1910) styles them fold-faults or slides.
The recognition of overfolding of the Alpine type in other areas has been accompanied by some criticism, and H. Schardt, himself an honoured pioneer, uttered a word of warning in 1906 when he humorously described P. Termier as afflicted with Ultra-nappismus. E. B. Bailey (Quart. Journ. Geol. Soc., vol. 66, 1910) recognized a system of recumbent folding as responsible for many features of the schistose masses in the S.W. Scottish highlands and in 1920 he des- cribed two successive nappes as resulting from overfolding from N.W. to S.E. L. Gentil (1918) traces three nappes in the structure of the coastal range of Algeria and Tunisia. In this he is strongly sup- ported by Termier, but is opposed by J.Savornin (1920). V. Uhlig, O. Ampferer, and others have studied overfolding in detail in the Hohe Tauern and Wetterstein districts of the eastern Alps, where much attention has been given to lateral shifting across the main folds from E. to W. Maria Ogilvie Gordon (1909) indicates folds and thrusts in explanation of the relations of the dolomite masses to underlying strata in the well-known Langkofel region of Tirol. P. Termier and G. Friedel trace outlying blocks, klippes, the Ger- man klippen, separated by denudation from former overfolds of the western Alps, in the southern part of the Rhone vale and even on the flanks'of the Cevennes. These outliers include a block of Urgonian strata resting on Oligocene beds and 3,700 metres long. It is held that the Card coal-field has been affected by the Alpine crumpling, and Termier recognizes older overfolded structures of Armorican (late Carboniferous) age in the eastern border of the central massif of France. On the other hand, linear mountain masses may record movements that are mainly vertical. H. E. Gregory (Am. Journ. Sci., vol. 41, 1916) thus treats the Andes as an uplifted plateau of marine and continental sediments, penetrated by igneous intru- sions. The erosion-surface here has little regard for geological structure. This is borne out by J. A. Douglas (Quart. Journ. Geol. Soc., vol. 76, 1920), who finds no overfolding in Bolivia and Peru and treats the range as a product of vertical upthrust between two resisting crust-masses. Even more recently, T. O. Bosworth records intense block-faulting as characteristic of the Cainozoic region of Peru, and a considerable Andean uplift, accompanied by a subsidence of the sea-floor, is recorded at the opening of Quaternary times. The upraised coastal band is " part of the crust-belt of the great fault." Five miles out at sea, a fault-scarp 2,000 ft. in height leads down abruptly to the depths of the Pacific. E. H. L. Schwarz points out similar features as bounding the E. coast of S. Africa, where Caino- zoic shore-deposits have been elevated to a height of 1,350 ft.
Fracturing. The study of structural geology has shown in recent years a marked return towards the recognition of lines of fracture, and founderings on a large scale, as influencing existing topographic features. The power of subaerial denudation has been justly emphasized by W. M. Davis in his development of the Huttonian cycle of erosion and his indication of traces of peneplains, where the eye is now likely to be diverted by later features of sculptural relief. But these later and secondary features, the walls of outstanding blocks and the courses of rivers across the rejuvenated country, are again and again associated with rectilinear and regularly intersecting fracture- systems. The tracing of rift-valleys, better styled trough-valleys, from the Jordan region to Nyasaland, and on a minor scale in the post-Oligocene groove of the Rhine from Basel to Mainz, has led to a general attempt to correlate faults and river-courses. The cliff-walls of elongated lakes in Finland are very probably due to late Cainozoic fracturing, and fault-scarps in the hard gneissic rocks guide the modern rivers on their way. J. W. Gregory in his book on The Origin of Fiords (1913) has collected much evidence to connect straight river-courses with more or less rectangular fracture-systems. E. B. Bailey shows the influence of downward-reaching " shatter-belts " on the valleys eroded in the region of Glencoe; and E. de Margerie gives a tectonic sig- nificance to the narrow cluses of the Juras, which unite the valleys of longitudinal and apparently subsequent streams. On the other hand, some of the surfaces cited by Gregory are con- nected with rock-folding and denudation rather than with faulting. J. Ball (Geol. Mag. 1910) points out that recent land- slides have simulated fault-scarps in the Nile valley and the Gulf of Suez; both these depressions are grooves of normal erosion and not troughs. W. F. Hume (ibid.) regards the Nile valley as on the whole connected with the erosion of the softer Cretaceous strata in its southern part, and of Middle Eocene strata in the north. It is, however, a structural feature in that these beds,, like the carboniferous limestone in the Armorican folds of south Ireland, have been eroded along the course of a synclinal. Ball states that the Gulf of Suez is guided by an eroded anticline.
One of the most powerful influences in the correlation of surface-
