PHYSIOGRAPHY OF THE CORDILLERA DE LOS ANDES BETWEEN LATITUDES 39° AND 44° SOUTH.

BY

BAILEY WILLIS,

United States Geological Survey, Washington, D.C.

With a map.

INTRODUCTION.

The region which is to be described in the following article is that part of the Cordillera de los Andes which lies between latitudes 39° 40′ and 43° 40′. It therefore comprises a section 450 kilometres (280 miles) in length, which falls in the zone of temperate and humid climate, between Argentina and Chile. The greater part of the mountain region is Argentine, as the boundary with Chile follows the western crest of the broad Cordillera.

The geography of this mountain region was first made known in any degree of detail through the labours of DR. FRANCISCO P. MORENO and his colleagues, whose surveys were made on behalf of their country, Argentina, in the boundary dispute with Chile, which began in 1881 and continued until it was settled by the arbitration of King EDWARD in 1902.

Extensive surveys were carried on by both governments in order to procure evidence which should be produced before the Arbitration Commission, and as a result, especially of the work of Dr. MORENO, the Andean region from latitude 39° to latitude 52° has been more carefully mapped than any other large area of either Argentina or Chile. We know that it is not a mountain ridge, but a broad mountain zone. We know that the summits within the zone rise in general to a common, nearly uniform altitude above the sea. It has been determined that seven large rivers rising east of the mountain zone flow through it to the Pacific, in deep and tortuous channels. Furthermore, it is a region of numerous lakes, many of them of notable size and characterized by the features of rock basins which have been deepened by glaciation. Recent studies have shown that there were at least two distinct periods of glaciation, which were separated by a long inter-Glacial period. We have learned also that normal faults traverse the range longitudinally and are the cause of certain longitudinal valleys, which lie between the Eastern and Western Cordilleras. There is no evidence of transverse faults crossing the range.

We thus approach the detailed study of this mountain zone with quite a different concept from that of the treaty makers of 1881. We shall recognize many distinct mountain groups, which yet seem related as elements

of a broad uplifted zone. We shall have to distinguish effects of faulting, of erosion by streams, of retrogressive erosion and stream capture, and of erosion and damming by glaciers, in tracing out the development of the intricate river systems which occupy the Andean valleys and canyons.

A brief description of the rock groups and structure observed by geologists of the Comisión de Estudios Hidrológicos has been given in The Fortyfirst Parallel Survey. The following pages present the views of the writer on the physiographic history of the Andes.

PHYSIOGRAPHIC EPOCHS.

A careful analysis of the topographic features of the Andes has led to the recognition of the following physiographic epochs: (1) epoch of mature erosion and low relief expressed in the highest summits and valleys; (2) epoch of elevation and deep canyon-cutting; (3) epoch of early glaciation; (4) inter-Glacial epoch; (5) epoch of latest glaciation; (6) post-Glacial or present epoch.

The earliest of these epochs of erosion corresponds with the gravels, already referred to as probably Pliocene. We may then arrange the successive epochs as follows:

It appears from the study of the relations which the present topographic features bear to one another that the river courses have been greatly changed since the (n + 1) epoch and the continental divide has been shifted by the diversion of the headwaters of certain Atlantic tributaries to competing Pacific streams. The Quilquihue waters, the Rio Chubut, and the Rio Tecka are the streams which have lost their original heads in the area under consideration. It will be shown that these diversions were due to the normal process of retrogressive stream erosion under conditions which gave to the Pacific stream in any individual case the advantage of steeper fall and shorter course to base-level as compared with the situation of the Atlantic tributary. The captures were accomplished during the (n + 2) epoch, when the elevation, normal faulting of the mountain mass, and consequent acceleration or retardation of the rivers caused shifting of divides and gains or losses of

watersheds. The changes therefore occurred before the Glacial period and are not attributable to glaciation.

The conclusion of the preceding paragraph that the divides were shifted by stream capture before the earliest Glacial epoch of the Andes applies particularly to the diversion of Atlantic tributaries to canyons which traverse the Western Cordillera, and replaces the explanations hitherto given that attribute those canyons to faulting or to glacial erosion. Faulting is discarded except in so far as it is recognized as a longitudinal structure, not as a transverse one, on the range. Its effect was to retard, not to accelerate the rivers which crossed the faults. Glacial erosion could not have opened the canyons across the Western Cordillera, since the conditions of a high head of ice and confinement in a narrow gorge, which are essential to erosion by a glacier, could not exist on a dividing ridge. The elimination of faulting and glacial erosion as causes of the transverse canyons in the Western Cordillera is a negative which agrees with the positive evidence of retrogressive erosion in establishing the latter as the real cause.

Notwithstanding the conclusion that glaciation did not occasion the transfer of Atlantic tributaries through the Western Cordillera to the Pacific, it is recognized that glaciers did interfere with the river courses in the Cordillera and in some instances established the continental divide on glacial moraines east of the Eastern Cordillera. They transferred certain drainage areas from the Atlantic watershed to the rivers of the Central Cordilleran valley and thus to rivers that had previously been diverted to the Pacific. These glacial diversions are due to damming, to the formation of lakes, and to the cutting of canyons by new lake outlets. The evidences are clear and unmistakable, and having been observed by various explorers, have led to the conclusion that the entire diversion in any one case was due to one and the same cause, glaciation. A more careful analysis distinguishes the earlier effects of retrogressive erosion from the later interferences by glacial damming and shows that the processes which have transferred the continental divide to the eastern side of the Cordilleran plateau, or to the Pampas still east of it, were the normal processes of retrogressive erosion and glacial obstruction; the earlier and greater work of stream capture was accomplished before the later and minor effects of glaciation were added to them.

Since writing the above I have received the excellent article by QUENSEL1 in which he states his observations on the glaciation of the southern Andes. His observations in the south and mine farther north lead us to agreement on the following essential points:

(1) The present continental divide, where it now lies east of the Cordillera, commonly follows morainic ridges, which were built across wide preGlacial valleys by ice tongues protruding eastward from the mountains. Canyons corresponding to the head-waters of these valleys were profoundly deepened and sculptured by the ice descending into them from the Andean

1 On the Influence of the Ice Age on the Continental Watershed of Patagonia; Bull. Geol. Institute of Upsala, Vol. IX.

heights. On the retreat of the ice the basins enclosed by the moraines across the valleys were filled with lakes, which found low outlets to the west.

(2) The Cordillera was traversed by low passes (valleys of erosion) in pre-Glacial time, and these had a marked influence in determining the courses of the rivers that flowed from the glacial lakes to the Pacific.

Mr. QUENSEL is inclined to lay more stress on the deepening of the passes in the Western Cordillera than I would where they cross divides, as the conditions for glacial erosion were not favourable at those passes where I have seen them, whereas the relations of the streams are those which should result from capture by retrogressive erosion. This fact QUENSEL does not appear to have recognized. I quote the following passages from his interesting paper:

66 FORMATION OF SUB-ANDINE LAKES."

"Before proceeding farther south, I will here briefly dwell on the formation of the depressions of the typical sub-andine lakes, as found in Lagos Buenos Aires, San Martin and Argentino. In the first place, an interesting fact is the near correspondence in height between the largest lakes: Lago Argentino 200 m., Viedma 250 m., San Martin 200 m., Pueyrredon 199 m., and Buenos Aires 217 m. I have several times already emphasized that the eastern parts of those lakes seem to be a direct continuation of the transversal valleys. Imagining the glacial deposits, which head these valleys, removed, the level of the lake would be greatly reduced and the valley would stretch farther westward. It is therefore necessary, when seeking to explain the formation of these lakes, to keep this in mind, as far as their eastern part is concerned. On the other hand the western fiords of the lakes with their perpendicular shores present an aspect so totally different as to claim another explanation. They have every resemblance to the Alpine lakes of other mountain landscapes and without doubt ice has played a principal part in moulding their present form. Descending the steep mountain sides, the glaciers reach their maximum of erosive power. But farther eastward the glacier, which had once filled the whole depression, has not had the same power to erode, the fall there being insignificant, if any, and in this case the ice would spread out more as in a piedmont-glacier. Of course a series of soundings in these lakes would be most interesting and doubtless gives a definite answer as to the mode of formation of these depressions. There is, however, every reason to suppose that the deeper part of the lakes is to the west, the whole topography of the country indicating this to be the case, especially by analogy with the very similar inlets farther south, where the depths are known. I will return to this question later on after having dealt with the general formation of the southern inlets. In trying to solve the formation of the sub-andine lakes, it is therefore necessary not simply to classify them, as Ameghino and Hatcher do, as tectonic depressions, but assume that in their eastern parts they represent the continuation of the erosion valleys of the Pampas, whereas their western parts have without doubt to a great extent been moulded by glacial erosion. In pre-glacial time the watershed would then have lain along the eastern slopes of the Cordillera."