sodium to be retained in the sediments in insoluble form, as shown by their average composition, the salt of the sea requires the erosion of a shell of igneous rock 2,200 feet thick enveloping the globe. If the erosion be regarded as restricted to the continents, and their area be taken as 28 per cent of the area of the earth, the equation would require the erosion of 7,900 feet of rock. The actual average erosion from the present land areas has probably been between these two figures, since ancient lands such as Appalachia, which have supplied much sediment, are now in part submerged.

The erosion, however, has been very unequally distributed. No igneous rock has been removed from the broad interior of the United States since pre-Cambrian times, and the same is true of other wide areas. Therefore there is room for the removal of many miles of igneous rock from mountain axes, not counting the erosion and re-erosion of the sedimentary formations. But allow for the demands of Paleozoic, Mesozoic, and Cenozoic erosion of igneous rocks, and the balance left to be assigned to pre-Cambrian erosion is materially reduced. Under the most favorable assumptions it can hardly be much over a mile in thickness, and if allowances are made for later continental fragmentation, it may not be much over half a mile.

Now the metamorphism of the basement complex is universal, and the batholithic exposures cover enormous areas. From such regions the older lavas have been removed, and erosion has bitten deeply into the underlying granites. But how deeply? For any particular region it may be five miles, ten miles, or even more, but considering the limitations placed by the salt in the sea, it would seem conclusive that the average erosion which has exposed the Archean zone of anamorphism and rock flow has not been more than a mile. Even the metasediments of the ancient cover are counted into this estimate, for although their removal would not add a fresh supply of salt, their original formation had given it, and by just so much had diminished the measure of later erosion. It would appear fairly certain that the profound metamorphism of the Archean complex does not imply the necessity of formation beneath from five to ten miles of cover.

This line of evidence, although indirect, is perhaps one of the strongest of many which go to show that the conditions of regional metamorphism are bound up with

batholithic invasion. How, then, shall the geologic evidence be reconciled with the laboratory demonstrations of Adams? This is not a conflict between facts and an imperfect theory, but rather between two categories of facts. Experimental work as conducted by Adams requires temperatures and pressures for rock flowage far in excess of the conditions implied by the geologic evidence. It is the business of theory to explain the paradox.

The theory of rock flowage, in the form stated by Van Hise, calls for an intermediate level. What adjustments of this theory are needed to bring harmony between it and the apparently diverging facts? It appears that the most probable solution may be found by invoking the high temperatures and abundant crystallizers which must be present in batholithic roofs.2 Let crustal compression operate at the same time and stress-differences arise through the crust. But the molecules under strain are more soluble, energy is released by their solution, and they are precipitated free from strain. The energy absorbed in deformation is transformed into heat and assists further in the process of recrystallization.

The phenomenon of rock granulation is suggestive of a very considerable depth of cover, yet even for this the geologic evidence does not support the view that any such profound depth is necessary as the laboratory work of Adams would seem to imply. It occurs especially in the granite gneisses, massive and resistant formations. In part it seems to be related to a last stage in crystallization · accompanied by movements in a very viscous magma. The granulation of phenocrysts in rhyolite surface flows offers a suggestion of the process. In part the presence of foliation at right angles to the pressure may prevent shearing on major diagonal thrust planes and compel each layer to yield independently by intramineral shear.

(To be continued)

Daly has presented, in a paper on metamorphism in volume 28 of the Bulletin of the Geological Society of America, a similar argument that the depth of pre-Cambrian metamorphism was not great; he suggests that it is to be explained by a steeper thermal gradient in the early history of the earth. This is one factor in the present suggestion, the other being emanations from the same magma that produced the steeper thermal gradient.

ART. XII.-On the Permian of Coahuila, Northern Mexico; by EMIL BÖSE.

In 1913, Doctor Erich Haarmann published his discovery of marine Permian beds in northern Mexico.1 This discovery was rather surprising, as everywhere in northern central Mexico the lowest exposed rocks had seemed to be of the Upper Jurassic. Furthermore, there did not appear to be any indications of large dislocations, which alone could apparently be the explanation of an exposure of lower beds, as the Jura-Cretaceous series seemed to be uninterrupted at every point of this region where the geology was known. Haarmann on a very rapid journey found the Permian beds on the Hacienda de las Delicias, 68 kilometers north of San Pedro de las Colonias, at the foot of the high Sierra del Sobaco, in the southern part of the state of Coahuila. The geographic position of these Paleozoic beds made the occurrence still more surprising, since a little south of San Pedro the Upper Jurassic is exposed in many places (Concepción del Oro, Mazapil, Sierra de Ramirez, San Juan de Guadalupe, etc.). Moreover, west of San Pedro de las Colonias, the same conditions had been observed at San Pedro del Gallo, Durango, and as it is well known that the JuraCretaceous series exists in many parts of eastern Chihuahua, it therefore seemed very improbable that any Paleozoic rocks were exposed in this region. On the other hand, the discovery could scarcely be doubted, especially since Haack published an account of the little fauna of Permian age collected by Haarmann.

Haarmann distinguished two different Paleozoic formations at Las Delicias. He considered as the older one a series of at least 2000 meters thickness, the lower portion of which is largely volcanic in nature and consists mainly of pebbles and cemented sands; upward the pebbles grow smaller and decrease in quantity; thick beds of sands derived from volcanic rocks predominate and might at first view be taken for weathered volcanics; still higher follow dark to black marly shales and marls, which contain geodes and beds of dark limestone. Haar

'Erich Haarmann, Geologische Streifzüge in Coahuila. Zs. deutsch. geol. Ges., 65, 1913, Monatsberichte.

Wilhelm Haack, Ueber eine marine Permfauna aus Nordmexico nebst Bemerkungen ueber Devon daselbst. Zs. deutsch. geol. Ges., 66, 1914.

mann called this series the Delicias beds and considered it as pre-Permian, stating that remains of coral reefs were lying unconformably on the Delicias beds in several places, and that this higher series consists of a yellowish gray, solid, massive limestone, especially well exposed at the Pichagua. The limestone contains the faunule described by Haack, composed of the following species: Cyathaxonia girtyi Haack, C. sp., Cladopora spinulata Girty, Streptorhynchus (?) sp. 1, S.(?) sp. 2, Richthofenia permiana Shumard, Spiriferina haarmanni Haack, S. hilli Girty, Hustedia meekana Shumard, Dielasmina guadalupensis Girty, Dielasma cf. biplex Waagen.

Haack also found fossils in rocks of the Delicias beds. One of these he considered to be Gypidula aff. pseudogaleata, and others are a goniatite apparently related to Sporadoceras, and a tabulate coral nearly related to Alveolites goldfussi Billings. In thin slides remains of Foraminifera were seen. Relying on his determinations of the few fossils gathered from the Delicias beds, Haack referred the latter to the Devonian, but thought that beds of other age might be contained in the same series.

From the very beginning, it was evident that the discovery of Haarmann was of the greatest importance for either the tectonics or the stratigraphy of northern Mexico. I therefore had a great desire to visit the locality myself and to study the relations between the Permian and the younger rocks, as well as that between the Permian and the supposed Devonian. Unfortunately, for many years the Hacienda de las Delicias has been the seat of revolutionary chiefs and a visit to it was practically impossible. However, when political conditions in that part of the country improved, during the early part of the summer of 1920, I made two trips to the Hacienda de las Delicias and made the field studies above outlined. Of my results, which differ somewhat from those of Haarmann and Haack, the following statement gives a preliminary account.

My first visit was to the limestone cliff known as the Pichagua, lying about 15 kilometers north of the hacienda buildings and in the foothills of the Sierra del Sobaco. I collected the fauna described by Haack, and other forms, and soon gained the impression that the Pichagua cliff was nothing other than a lenticular thickening of a limestone bed intercalated in the Delicias series. A second visit confirmed me in this belief; I followed this limestone

toward a small creek, where it could be easily observed that it was normally lying between the sandy and shaly beds of the Delicias formation. I then resolved to study the relation between the cliffs and the Delicias beds in another portion of the valley where I could also find sufficient drinking water, and where the Delicias beds are much better exposed than in the Arroyo de Wenceslao. This place is a few kilometers farther south, near the Noria de Malascachas, where a well dug in the Delicias beds contains plenty of good water, a rather scarce commodity in this region.

Here the Delicias beds compose the lower half of the Sierra del Sobaco and show in several places cliffs of dark limestone. I tried to make a cross-section and to collect as many fossils as possible.

A little east of Noria de Malascachas the Delicias beds. are cut off by a north-south fault which toward the east brings the series into contact with a block of Cretaceous limestone. The lowest portion of the Delicias exposed here consists mainly of dark, thin-bedded or laminated clays, alternating with thin to thick beds of greenish to yellowish sandstone, mostly of igneous detritals. Sometimes the clays contain concretions of dark limestone with scattering remains of undeterminable brachiopods, and often Fusulina. The beds show a strong dip (45 and more) toward the west. These rocks continue for about 150 to 200 meters higher stratigraphically; then the sandstone becomes coarser and passes into a conglomerate, mainly of igneous material, the pebbles being large and well rounded. Upward in the conglomerate occur calcareous concretions which also pass into the dark clay and a few meters higher consolidate into a bed of limestone. This forms a small cliff on the crest of a spur of the mountain but continues in the creeks on both sides of the spur, without any interruption, as a solid bed of dark limestone, becoming somewhat thinner with distance from the lenticular swelling of the cliff. Higher stratigraphically the bed of limestone dissolves into a great number of rounded blocks. These apparently are only concretions imbedded in dark clay, and about half a meter above the solid bed of limestone the dark clay predominates, reproducing here, as elsewhere, intercalated beds of greenish sandstone. There can be no doubt that the limestone is interbedded in the clay, sandstones, and conglomerates of the Delicias beds, and that the little cliff