Fig. 128—Composition of slopes and profiles in the Peruvian Andes. By superimposing the cross profiles of typical valleys as shown in 127 a restoration is possible of the longitudinal profiles of the earlier cycle of erosion. The difference in elevation of the two profiles gives less than the minimum amount of uplift that must have occurred. Case A represents a valley in which recent cutting has not yet reached the valley head. Below the point 1 the profile has been steepened and lowered by erosion in the current cycle. Above point 1 the profile is still in the stage it reached in the preceding cycle. In case B the renewed erosion of the current cycle has reached to the valley head. Case C represents conditions similar to those in the preceding cases save that the stream is typical of those that lie nearest the steep flexed or faulted margins of the Cordillera and discharge to the low levels of the desert pampa on the west or the tropical plains on the east.
Further proof of recent and great uplift is afforded by the deeply intrenched streams. After descending the long graded slopes one comes upon the cliffed canyons with a feeling of consternation. The effect of powerful erosion, incident upon uplift, is heightened by the ungraded character of the river bed. Falls and rapids abound, the river profiles suggest tumultuous descents, and much time will elapse before the river beds have the regular and moderate gradients of the streams draining the mature surface before uplift as shown in the profiles by the dotted lines representing the restored valley floors of the older cycle. Since the smooth-contoured landscape was formed great changes have taken place. The streams have changed from completely graded to almost completely ungraded profiles; in place of a subdued landscape we now have upland slopes intersected by mile-deep canyons; the high-level slopes could not have been formed under existing conditions, for they are being dissected by the present streams.
Since the slopes of the land in general undergo progressive changes in the direction of flatter gradients during a given geographical cycle, it follows that with the termination of one cycle and the beginning of another, two sets of slopes will exist and that the gradients of the two will be unlike. The result is a break in the descent of the slopes from high to low levels to which the name “topographic unconformity” is now applied. It will be a prominent feature of the landscape if the higher, older, and flatter gradients have but little declivity, and the gradients of the lower younger slopes are very steep. In those places where the relief of the first cycle was still great at the time of uplift, the erosion forms of the second cycle may not be differentiated from those of the first, since both are marked by steep gradients. In the Central Andes the change in gradient between the higher and lower slopes is generally well marked. It occurs at variable heights above the valley floors, though rarely more than 3,000 feet above them. In the more central tracts, far from the main streams and their associated canyons, dissection in the present erosion cycle has not yet been initiated, the mature slopes are still intact, and a topographic unconformity has not yet been developed. The higher slopes are faced with rock and topped with slowly moving waste. Ascent of the spur end is by steep zigzag trails; once the top is gained the trail runs along the gentler slopes without special difficulties.
It is worth noting at this point that the surface of erosion still older than the mature slopes herewith described appears not to have been developed along the seventy-third meridian of Peru, or if developed at one time, fragments of it no longer remain. The last well-developed remnant is southwest of Cuzco, 130 . I have elsewhere described the character and geographic distribution of this oldest recognizable surface of the Central Andes.[41] Southern Peru and Bolivia and northern Chile display its features in what seems an unmistakable manner. The best locality yet found is in the Desaguadero Valley between Ancoaqui and Concordia. There one may see thousands of feet of strongly inclined sediments of varying resistance beveled by a well-developed surface of erosion whose preserval is owing to a moderate rainfall and to location in an interior basin.[42]
The highest surface of a region, if formed during a prolonged period of erosion, becomes a surface of reference in the determination of the character and amount of later crustal deformations, having somewhat the same functions as a key bed in stratigraphic geology. Indeed, concrete physiographic facts may be the only basis for arguments as to both epeirogenic and orogenic movements. The following considerations may show in condensed form the relative value of physiographic evidence:
1. If movements in the earth’s crust are predominantly downward, sedimentation may be carried on continuously, and a clear geologic record may be made.
2. Even if crustal movements are alternately downward and upward, satisfactory conclusions may be drawn from both (a) the nature of the buried surfaces of erosion, and (b) the alternating character of the sediments.
3. If, however, the deformative processes effect steady or intermittent uplifts, there may be no sediments, at least within the limits of the positive crustal units, and a geologic record must be derived not from sedimentary deposits but from topographic forms. We speak of the lost intervals represented by stratigraphic breaks or unconformities and commonly emphasize our ignorance concerning them. The longest, and, from the human standpoint, the most important, break in the sedimentary record is that of the present wherever degradation is the predominant physiographic process. Unlike the others the lost interval of the present is not lost, if we may so put it, but is in our possession, and may be definitely described as a concrete thing. It is the physiography of today.
Even where long-buried surfaces of erosion are exposed to view, as in northern Wisconsin, where the Pre-Cambrian paleo-plain projects from beneath the Paleozoic sediments, or, as in New Jersey and southeastern Pennsylvania, where the surface developed on the crystalline rocks became by depression the floor of the Triassic and by more recent uplift and erosion has been exposed to view,—even in such cases the exposures are of small extent and give us at best but meager records. In short, many of the breaks in the geologic record are of such long duration as to make imperative the use of physiographic principles and methods. The great Appalachian System of eastern North America has been a land area practically since the end of the Paleozoic. In the Central Andes the “lost interval,” from the standpoint of the sedimentary, record, dates from the close of the Cretaceous, except in a few local intermont basins partially filled with Tertiary or Pleistocene deposits. Physiographic interpretations, therefore, serve the double purpose of supplying a part of the geologic record while at the same time forming a basis for the scientific study of the surface distribution of living forms.
The geologic dates of origin of the principal topographic forms of the Central Andes may be determined with a fair degree of accuracy. Geologic studies in Peru and Bolivia have emphasized the wide distribution of the Cretaceous formations. They consist principally of thick limestones above and sandstones and conglomerates below, and thus represent extensive marine submergence of the earth’s crust in the Cretaceous where now there are very lofty mountains. The Cretaceous deposits are everywhere strongly deformed or uplifted to a great height, and all have been deeply eroded. They were involved, together with other and much older sediments, in the erosion cycle which resulted in the development of the widely extended series of mature slopes already described. From low scattered island elevations projecting above sea level, as in the Cretaceous period, the Andes were transformed by compression and uplift to a rugged mountain belt subjected to deep and powerful erosion. The products of erosion were in part swept into the adjacent seas, in part accumulated on the floors of intermont basins, as in the great interior basins of Titicaca and Poopó.
Since the early Tertiary strata are themselves deformed from once simple and approximately horizontal structures and subjected to moderate tilting and faulting, it follows that mountain-making movements again affected the region during later Tertiary. They did not, however, produce extreme effects. They did stimulate erosion and bring about a reorganization of all the slopes with respect to the new levels.
This agrees closely with a second line of evidence which rests upon an independent basis. The alluvial fill which lies upon all the canyon and valley floors is of glacial origin, as shown by its interlocking relations with morainal deposits at the valley heads. It is now in process of dissection and since its deposition in the Pleistocene had been eroded on the average about 200 feet. Clearly, to form a 3,000-foot canyon in hard rock requires much more time than to deposit and again partially to excavate an alluvial fill several hundred feet deep. Moreover, the glacial material is coarse throughout, and was built up rapidly and dissected rapidly. In most cases, furthermore, coarse material at the bottom of the glacial series rests directly upon the rock of a narrow and ungraded valley floor. From these and allied facts it is concluded that there is no long time interval represented by the transitions from degrading to aggrading processes and back again. The early Pleistocene, therefore, seems quite too short a period in which to produce the bold forms and effect the deep erosion which marks the period between the close of the mature cycle and the beginnings of deposition in the Pleistocene.
The alternative conclusion is that the greater part of the canyon cutting was effected in the late Tertiary, and that it continued into the early Pleistocene until further erosion was halted by changed climatic conditions and the augmented delivery of land waste to all the streams. The final development of the well-graded high-level slopes is, therefore, closely confined to a small portion of the Tertiary. The closest estimate which the facts support appears to be Miocene or early Pliocene. It is clear, however, that only the culmination of the period can be definitely assigned. Erosion was in full progress at the close of the Cretaceous and by middle Tertiary had effected vast changes in the landscape. The Tertiary strata are marked by coarse basal deposit and by thin and very fine top deposits. Though their deformed condition indicates a period of crustal disturbance, the Tertiary beds give no indication of wholesale transformations. They indicate chiefly tilting and moderate and normal faulting. The previously developed effects of erosion were, therefore, not radically modified. The surface was thus in large measure prepared by erosion in the early Tertiary for its final condition of maturity reached during the early Pliocene.
It seems appropriate, in concluding this chapter, to summarize in its main outlines the physiography of southern Peru, partly to condense the extended discussion of the preceding paragraphs, and partly to supply a background for the three chapters that follow. The outstanding features are broad plateau areas separated by well-defined “Cordilleras.” The plateau divisions are not everywhere of the same origin. Those southwest of Cuzco (Fig. 130), and in the Anta Basin (Fig. 124), northwest of Cuzco, are due to prolonged erosion and may be defined as peneplane surfaces uplifted to a great height. They are now bordered on the one hand by deep valleys and troughs and basins of erosion and deformation; and, on the other hand, by residual elevations that owe their present topography to glacial erosion superimposed upon the normal erosion of the peneplane cycle. The residuals form true mountain chains like the Cordillera Vilcanota and Cordillera Vilcapampa; the depressions due to erosion or deformation or both are either basins like those of Anta and Cuzco or valleys of the canyon type like the Urubamba canyon; the plateaus are broad rolling surfaces, the punas of the Peruvian Andes.
There are two other types of plateaus. The one represents a mature stage in the erosion cycle instead of an ultimate stage; the other is volcanic in origin. The former is best developed about Antabamba (Figs. 122 and 123), where again deep canyons and residual ranges form the borders of the plateau remnants. The latter is well developed above Cotahuasi and in its simplest form is represented in 133 . Its surface is the top of a vast accumulation of lavas in places over a mile thick. While rough in detail it is astonishingly smooth in a broad view (Fig. 29). Above it rise two types of elevations: first, isolated volcanic cones of great extent surrounded by huge lava flows of considerable relief; and second, discontinuous lines of peaks where volcanic cones of less extent are crowded closely together. The former type is displayed on the Coropuna Quadrangle, the latter on the Cotahuasi and La Cumbre Quadrangles.
So high is the elevation of the lava plateau, so porous its soil, so dry the climate, that a few through-flowing streams gather the drainage of a vast territory and, as in the Grand Canyon country of our West, they have at long intervals cut profound canyons. The Arma has cut a deep gorge at Salamanca; the Cotahuasi runs in a canyon in places 7,000 feet deep; the Majes heads at the edge of the volcanic field in a steep amphitheatre of majestic proportions.
Finally, we have the plateaus of the coastal zone. These are plains with surfaces several thousand feet in elevation separated by gorges several thousand feet deep. The Pampa de Sihuas is an illustration. The post-maturely dissected Coast Range separates it from the sea. The pampas are in general an aggradational product formed in a past age before uplift initiated the present canyon cycle of erosion. Other plateaus of the coastal zone are erosion surfaces. The Tablazo de Ica appears to be of this type. That at Arica, Chile, near the southern boundary of Peru, is demonstrably of this type with a border on which marine planation has in places given rise to a broad terrace effect.[43]
THE Western or Maritime Cordillera of Peru forms part of the great volcanic field of South America which extends from Argentina to Ecuador. On the walls of the Cotahuasi Canyon (Fig. 131), there are exposed over one hundred separate lava flows piled 7,000 feet deep. They overflowed a mountainous relief, completely burying a limestone range from 2,000 to 4,000 feet high. Finally, upon the surface of the lava plateau new mountains were formed, a belt of volcanoes 5,000 feet (1,520 m.) high and from 15,000 to 20,000 feet (4,570 to 6,100 m.) above the sea. There were vast mud flows, great showers of lapilli, dust, and ashes, and with these violent disturbances also came many changes in the drainage. Sixty miles northeast of Cotahuasi the outlet of an unnamed deep valley was blocked, a lake was formed, and several hundred feet of sediments were deposited. They are now wasting rapidly, for they lie in the zone of alternate freezing and thawing, a thousand feet and more below the snowline. Some of their bad-land forms look like the solid bastions of an ancient fortress, while others have the delicate beauty of a Japanese temple.
Not all the striking effects of vulcanism belong to the remote geologic past. A day’s journey northeast of Huaynacotas are a group of lakes only recently hemmed in by flows from the small craters thereabouts. The fires in some volcanic craters of the Peruvian Andes are still active, and there is no assurance that devastating flows may not again inundate the valleys. In the great Pacific zone or girdle of volcanoes the earth’s crust is yet so unstable that earthquakes occur every year, and at intervals of a few years they have destructive force. Cotahuasi was greatly damaged in 1912; Abancay is shaken every few years; and the violent earthquakes of Cuzco and Arequipa are historic.
On the eastern margin of the volcanic country the flows thin out and terminate on the summit of a limestone (Cretaceous) plateau. On the western margin they descend steeply to the narrow west-coast desert. The greater part of the lava dips beneath the desert deposits; there are a few intercalated flows in the deposits themselves, and the youngest flows—limited in number—have extended down over the inner edge of the desert.
The immediate coast of southern Peru is not volcanic. It is composed of a very hard and ancient granite-gneiss which forms a narrow coastal range (Fig. 171). It has been subjected to very long and continued erosion and now exhibits mature erosion forms of great uniformity of profile and declivity.
From the outcrops of older rocks beneath the lavas it is possible to restore in a measure the pre-volcanic topography of the Maritime Cordillera, In its present altitude it ranges from several thousand to 15,000 feet above sea level. The unburied topography has been smoothed out; the buried topography is rough (Figs. 29 and 166). The contact lines between lavas and buried surfaces in the deep Majes and Cotahuasi valleys are in places excessively serrate. From this, it seems safe to conclude that the period of vulcanism was so prolonged that great changes in the unburied relief were effected by the agents of erosion. Thus, while the dominant process of volcanic upbuilding smoothed the former rough topography of the Maritime Cordillera, erosion likewise measurably smoothed the present high extra-volcanic relief in the central and eastern sections. The effect has been to develop a broad and sufficiently smooth aspect to the summit topography of the entire Andes to give them a plateau character. Afterward the whole mountain region was uplifted about a mile above its former level so that at present it is also continuously lofty.
The zone of most intense volcanic action does not coincide with the highest part of the pre-volcanic topography. If the pre-volcanic relief were even in a very general way like that which would be exhibited if the lavas were now removed, we should have to say that the chief volcanic outbursts took place on the western flank of an old and deeply dissected limestone range.
Fig. 129—Composition of slopes at Puquiura, Vilcabamba Valley, elevation 9,000 feet (2,740 m.). The second prominent spur entering the valley on the left has a flattish top unrelated to the rock structure. Like the spurs on the right its blunt end and flat top indicate an earlier erosion cycle at a lower elevation.
Fig. 130—Inclined Paleozoic strata truncated by an undulating surface of erosion at 15,000 feet, southwest of Cuzco.
Fig. 131—Terraced valley slopes at Huaynacotas, Cotahuasi Valley, at 11,500 feet (3,500 m.). Solimana is in the background. On the floor of the Cotahuasi Canyon fruit trees grow. At Huaynacotas corn and potatoes are the chief products. The section is composed almost entirely of lava. There are over a hundred major flows aggregating 5,000 to 7,000 feet thick.
The volume of the lavas is enormous. They are a mile and a half thick, nearly a hundred miles wide, and of indefinite extent north and south. Their addition to the Andes, therefore, has greatly broadened the zone of lofty mountains. Their passes are from 2,000 to 3,000 feet higher than the passes of the eastern Andes. They have a much smaller number of valleys sufficiently deep to enjoy a mild climate. Their soil is far more porous and dry. Their vegetation is more scanty. They more than double the difficulties of transportation. And, finally, their all but unpopulated loftier expanses are a great vacant barrier between farms in the warm valleys of eastern Peru and the ports on the west coast.
The upbuilding process was not, of course, continuous. There were at times intervals of quiet, and some of them were long enough to enable streams to become established. Buried valleys may be observed in a number of places on the canyon walls, where subsequently lava flows displaced the streams and initiated new drainage systems. In these quiet intervals the weathering agents attacked the rock surfaces and formed soil. There were at least three or four such prolonged periods of weathering and erosion wherein a land surface was exposed for many thousands of years, stream systems organized, and a cultivable soil formed. No evidence has been found, however, that man was there to cultivate the soil.
The older valleys cut in the quiet period are mere pygmies beside the giant canyons of today. The present is the time of dominant erosion. The forces of vulcanism are at last relatively quiet. Recent flows have occurred, but they are limited in extent and in effects. They alter only the minor details of topography and drainage. Were it not for the oases set in the now deep-cut canyon floors, the lava plateau of the Maritime Cordillera would probably be the greatest single tract of unoccupied volcanic country in the world.
The lava plateau has been dissected to a variable degree. Its high eastern margin is almost in its original condition. Its western margin is only a hundred miles from the sea, so that the streams have steep gradients. In addition, it is lofty enough to have a moderate rainfall. It is, therefore, deeply and generally dissected. Within the borders of the plateau the degree of dissection depends chiefly upon position with respect to the large streams. These were in turn located in an accidental manner. The repeated upbuilding of the surface by the extensive outflow of liquid rock obliterated all traces of the earlier drainage. In the Cotahuasi Canyon the existing stream, working down through a mile of lavas, at last uncovered and cut straight across a mountain spur 2,000 feet high. Its course is at right angles to that pursued by the stream that once drained the spur. It is noteworthy that the Cotahuasi and adjacent streams take northerly courses and join Atlantic rivers. The older drainage was directly west to the Pacific. Thus, vulcanism not only broadened the Andes and increased their height, but also moved the continental divide still nearer the west coast.
The glacial features of the western or Maritime Cordillera are of small extent, partly because vulcanism has added a considerable amount of material in post-glacial time, partly because the climate is so exceedingly dry that the snowline lies near the top of the country. The slopes of the volcanic cones are for the most part deeply recessed on the southern or shady sides. Above 17,500 feet (5,330 m.) the process of snow and ice excavation still continues, but the tracts that exceed this elevation are confined to the loftiest peaks or their immediate neighborhood. There is a distinct difference between the glacial forms of the eastern or moister and the western or dryer flanks of this Cordillera. Only peaks like Coropuna and Solimana near the western border now bear or ever bore snowfields and glaciers. By contrast the eastern aspect is heavily glaciated. On La Cumbre Quadrangle, there is a huge glacial trough at 16,000 feet (4,876 m.), and this extends with ramifications up into the snowfields that formerly included the highest country. Prolonged glacial erosion produced a full set of topographic forms characteristic of the work of Alpine glaciers. Thus, each of the main mountain chains that make up the Andean system has, like the system as a whole, a relatively more-dry and a relatively less-dry aspect. The snowline is, therefore, canted from west to east on each chain as well as on the system. However, this effect is combined with a solar effect in an unequal way. In the driest places the solar factor is the more efficient and the snowline is there canted from north to south.
THE culminating range of the eastern Andes is the so-called Cordillera Vilcapampa. Its numerous, sharp, snow-covered peaks are visible in every summit view from the central portion of the Andean system almost to the western border of the Amazon basin. Though the range forms a water parting nearly five hundred miles long, it is crossed in several places by large streams that flow through deep canyons bordered by precipitous cliffs. The Urubamba between Torontoy and Colpani is the finest illustration. For height and ruggedness the Vilcapampa mountains are among the most noteworthy in Peru. Furthermore, they display glacial features on a scale unequaled elsewhere in South America north of the ice fields of Patagonia.
One of the most impressive sights in South America is a tropical forest growing upon a glacial moraine. In many places in eastern Bolivia and Peru the glaciers of the Ice Age were from 5 to 10 miles long—almost the size of the Mer de Glace or the famous Rhone glacier. In the Juntas Valley in eastern Bolivia the tree line is at 10,000 feet (3,050 m.), but the terminal moraines lie several thousand feet lower. In eastern Peru the glaciers in many places extended down nearly to the tree line and in a few places well below it. In the Cordillera Vilcapampa vast snowfields and glacier systems were spread out over a summit area as broad as the Southern Appalachians. The snowfields have since shrunk to the higher mountain recesses; the glaciers have retreated for the most part to the valley heads or the cirque floors; and the lower limit of perpetual snow has been raised to 15,500 feet.
Fig. 132—Recessed volcanoes in the right background and eroded tuffs, ash beds, and lava flows on the left. Maritime Cordillera above Cotahuasi.
Fig. 133—The summit of the great lava plateau above Cotahuasi on the trail to Antabamba. The lavas are a mile and a half in thickness. The elevation is 16,000 feet. Hence the volcanoes in the background, 17,000 feet above sea level, are mere hills on the surface of the lofty plateau.
Fig. 134—Southwestern aspect of the Cordillera Vilcapampa between Anta and Urubamba from Lake Huaipo. Rugged summit topography in the background, graded post-mature slopes in the middle distance, and solution lake in limestone in the foreground.
Fig. 135—Summit view, Cordillera Vilcapampa. There are fifteen glaciers represented in this photograph. The camera stands on the summit of a minor divide in the zone of nivation.
These features are surprising because neither Whymper[44] nor Wolf[45] mentions the former greater extent of the ice on the volcanoes of Ecuador, only ten or twelve degrees farther north. Moreover, Reiss[46] denies that the hypothesis of universal climatic change is supported by the facts of a limited glaciation in the High Andes of Ecuador; and J. W. Gregory[47] completely overlooks published proof of the existence of former more extensive glaciers elsewhere in the Andes:
“... the absence not only of any traces of former more extensive glaciation from the tropics, as in the Andes and Kilimandjaro, but also from the Cape.” He says further: “In spite of the extensive glaciers now in existence on the higher peaks of the Andes, there is practically no evidence of their former greater extension.”(!)
Whymper spent most of his time in exploring recent volcanoes or those recently in eruption, hence did not have the most favorable opportunities for gathering significant data. Reiss was carried off his feet by the attractiveness of the hypothesis[48] relating to the effect of glacial denudation on the elevation of the snowline. Gregory appeared not to have recognized the work of Hettner on the Cordillera of Bogotá and of Sievers[49] and Acosta on the Sierra Nevada de Santa Marta in northern Colombia.
The importance of the glacial features of the Cordillera Vilcapampa developed on a great scale in very low latitudes in the southern hemisphere is twofold: first, it bears on the still unsettled problem of the universality of a colder climate in the Pleistocene, and, second, it supplies additional data on the relative depression of the snowline in glacial times in the tropics. Snow-clad mountains near the equator are really quite rare. Mount Kenia rising from a great jungle on the equator, Kilimandjaro with its two peaks, Kibo and Mawenzi, two hundred miles farther south, and Ingomwimbi in the Ruwenzori group thirty miles north of the equator, are the chief African examples. A few mountains from the East Indies, such as Kinibalu in Borneo, latitude 6° north, have been found glaciated, though now without a snow cover. In higher latitudes evidences of an earlier extensive glaciation have been gathered chiefly from South America, whose extension 13° north and 56° south of the equator, combined with the great height of its dominating Cordillera, give it unrivaled distinction in the study of mountain glaciation in the tropics.
Furthermore, mountain summits in tropical lands are delicate climatic registers. In this respect they compare favorably with the inclosed basins of arid regions, where changes in climate are clearly recorded in shoreline phenomena of a familiar kind. Lofty mountains in the tropics are in a sense inverted basins, the lower snowline of the past is like the higher shoreline of an interior basin; the terminal moraines and the alluvial fans in front of them are like the alluvial fans above the highest strandline; the present snow cover is restricted to mountain summits of small areal extent, just as the present water bodies are restricted to the lowest portions of the interior basin; and successive retreatal stages are marked by terminal moraines in the one case as they are marked in the other by flights of terraces and beach ridges.
I made only a rapid reconnaissance across the Cordillera Vilcapampa in the winter season, and cannot pretend from my limited observations to solve many of the problems of the field. The data are incorporated chiefly in the chapter on Glacial Features. In this place it is proposed to describe only the more prominent glacial features, leaving to later expeditions the detailed descriptions upon which the solution of some of the larger problems must depend.
At Choquetira three prominent stages in the retreat of the ice are recorded. The lowermost stage is represented by the great fill of morainic and outwash material at the junction of the Choquetira, and an unnamed valley farther south at an elevation of 11,500 feet (3,500 m.). A mile below Choquetira a second moraine appears, elevation 12,000 feet (3,658 m.), and immediately above the village a third at 12,800 (3,900 m.). The lowermost moraine is well dissected, the second is ravined and broken but topographically distinct, the third is sharp-crested and regular. A fourth though minor stage is represented by the moraine at the snout of the living glacier and still less important phases are represented in some valleys—possibly the record of post-glacial changes of climate. Each main moraine is marked by an important amount of outwash, the first and third moraines being associated with the greatest masses. The material in the moraines represents only a part of that removed to form the successive steps in the valley profile. The lowermost one has an enormous volume, since it is the oldest and was built at a time when the valley was full of waste. It is fronted by a deep fill, over the dissected edge of which one may descend 800 feet in half an hour. It is chiefly alluvial in character, whereas the next higher one is composed chiefly of bowlders and is fronted by a pronounced bowlder train, which includes a remarkable perched bowlder of huge size. Once the valley became cleaned out the ice would derive its material chiefly by the slower process of plucking and abrasion, hence would build much smaller moraines during later recessional stages, even though the stages were of equivalent length.
Fig. 136—Glacial sculpture on the southwestern flank of the Cordillera Vilcapampa. Flat-floored valleys and looped terminal moraines below and glacial steps and hanging valleys are characteristic. The present snowfields and glaciers are shown by dotted contours.
There is a marked difference in the degree of dissection of the moraines. The lowermost and oldest is so thoroughly dissected as to exhibit but little of its original surface. The second has been greatly modified, but still possesses a ridge-like quality and marks the beginning of a noteworthy flattening of the valley gradient. The third is as sharp-crested as a roof, and yet was built so long ago that the flat valley floor behind it has been modified by the meandering stream. From this point the glacier retreated up-valley several miles (estimated) without leaving more than the thinnest veneer on the valley floor. The retreat must, therefore, have been rapid and without even temporary halts until the glacier reached a position near that occupied today. Both the present ice tongues and snowfields and those of a past age are emphasized by the presence of a patch of scrub and woodland that extends on the north side of the valley from near the snowline down over the glacial forms to the lower valley levels.
The retreatal stages sketched above would call for no special comment if they were encountered in mountains in northern latitudes. They would be recognized at once as evidence of successive periodic retreats of the ice, due to successive changes in temperature. To understand their importance when encountered in very low latitudes it is necessary to turn aside for a moment and consider two rival hypotheses of glacial retreat. First we have the hypothesis of periodic retreat, so generally applied to terminal moraines and associated outwash in glaciated mountain valleys. This implies also an advance of the ice from a higher position, the whole taking place as a result of a climatic change from warmer to colder and back again to warmer.
Fig. 137—Looking up a spurless flat-floored glacial trough near the Chucuito pass in the Cordillera Vilcapampa from 14,200 feet (4,330 m.). Note the looped terminal and lateral moraines on the steep valley wall on the left. A stone fence from wall to wall serves to inclose the flock of the mountain shepherd.
Fig. 138—Terminal moraine in the glaciated Choquetira Valley below Choquetira. The people who live here have an abundance of stones for building corrals and stone houses. The upper edge of the timber belt (cold timber line) is visible beyond the houses. Elevation 12,100 feet (3,690 m.).
But evidences of more extensive mountain glaciation in the past do not in themselves prove a change in climate over the whole earth. In an epoch of fixed climate a glacier system may so deeply and thoroughly erode a mountain mass, that the former glaciers may either diminish in size or disappear altogether. As the work of excavation proceeds, the catchment basins are sunk to, and at last below, the snowline; broad tributary spurs whose snows nourish the glaciers, may be reduced to narrow or skeleton ridges with little snow to contribute to the valleys on either hand; the glaciers retreat and at last disappear. There would be evidences of glaciation all about the ruins of the former loftier mountain, but there would be no living glaciers. And yet the climate might remain the same throughout.
It is this “topographic” hypothesis that Reiss and Stübel accept for the Ecuadorean volcanoes. Moreover, the volcanoes of Ecuador are practically on the equator—a very critical situation when we wish to use the facts they exhibit in the solution of such large problems as the contemporaneous glaciation of the two hemispheres, or the periodic advance and retreat of the ice over the whole earth. This is not the place to scrutinize either their facts or their hypothesis, but I am under obligations to state very emphatically that the glacial features of the Cordillera Vilcapampa require the climatic and not the topographic hypothesis. Let us see why.
The differences in degree of dissection and the flattening gradient up-valley that we noted in a preceding paragraph leave no doubt that each moraine of the bordering valleys in the Vilcapampa region, represents a prolonged period of stability in the conditions of topography as well as of temperature and precipitation. If change in topographic conditions is invoked to explain retreat from one position to the other there is left no explanation of the periodicity of retreat which has just been established. If a period of cold is inaugurated and glaciers advance to an ultimate position, they can retreat only through change of climate effected either by general causes or by topographic development to the point where the snowfields become restricted in size. In the case of climatic change the ice changes are periodic. In the case of retreat due to topographic change there should be a steady or non-periodic falling back of the ice front as the catchment basins decrease in elevation and the snow-gathering ridges tributary to them are reduced in height.
Further, the matterhorns of the Cordillera Vilcapampa are not bare but snow-covered, vigorous glaciers several miles in length and large snowfields still survive and the divides are not arêtes but broad ridges. In addition, the last two moraines, composed of very loose material, are well preserved. They indicate clearly that the time since their formation has witnessed no wholesale topographic change. If (1) no important topographic changes have taken place, and (2) a vigorous glacier lay for a long period back of a given moraine, and (3) suddenly retreated several miles and again became stable, we are left without confidence in the application of the topographic hypothesis to the glacial features of the Vilcapampa region. Glacial retreat may be suddenly begun in the case of a late stage of topographic development, but it should be an orderly retreat marked by a large number of small moraines, or at least a plentiful strewing of the valley floor with débris.
Fig. 139—Glacial features on the eastern slopes of the Cordillera Vilcapampa.
The number of moraines in the various glaciated valleys of the Cordillera Vilcapampa differ, owing to differences in elevation and to the variable size of the catchment basins. All valleys, however, display the same sudden change from moraine to moraine and the same characteristics of gradient. In all of them the lowermost moraine is always more deeply eroded than the higher moraines, in all of them glacial erosion was sufficiently prolonged greatly to modify the valley walls, scour out lake basins, or broad flat valley floors, develop cirques, arêtes, and pinnacled ridges in limited number. In some, glaciation was carried to the point where only skeleton divides remained, in most places broad massive ridges or mountain knots persist. In spite of all these differences successive moraines were formed, separated by long stretches either thinly covered with till or exposing bare rock.
In examining this group of features it is important to recognize the essential fact that though the number of moraines varies from valley to valley, the differences in character between the moraines at low and at high elevations in a single valley are constant. It is also clear that everywhere the ice retreated and advanced periodically, no matter with what topographic features it was associated, whether those of maturity or of youth in the glacial cycle. We, therefore, conclude that topographic changes had no significant part to play in the glacial variations in the Cordillera Vilcapampa.
The country west of the Cordillera Vilcapampa had been reduced to early topographic maturity before the Ice Age, and then uplifted with only moderate erosion of the masses of the interfluves. That on the east had passed through the same sequence of events, but erosion had been carried much farther. The reason for this is found in a strong climatic contrast. The eastern is the windward aspect and receives much more rain than the western. Therefore, it has more streams and more rapid dissection. The result was that the eastern slopes were cut to pieces rapidly after the last great regional uplift; the broad interfluves were narrowed to ridges. The region eastward from the crest of the Cordillera to the Pongo de Mainique looks very much like the western half of the Cascade Mountains in Oregon—the summit tracts of moderate declivity are almost all consumed.
The effect of these climatic and topographic contrasts is manifested in strong contrasts in the position and character of the glacial forms on the opposite slopes of the range. At Pampaconas on the east the lowermost terminal moraine is at least a thousand feet below timber line. Between Vilcabamba pueblo and Puquiura the terminal moraine lies at 11,200 feet (3,414 m.). By contrast the largest Pleistocene glacier on the western slope, nearly twelve miles long, and the largest along the traverse, ended several miles below Choquetira at 11,500 feet (3,504 m.) elevation, or just at the timber line. Thus, the steeper descents of the eastern side of the range appear to have carried short glaciers to levels far lower than those attained by the glaciers of the western slope.
It seems at first strange that the largest glaciers were west of the divide between the Urubamba and the Apurimac, that is, on the relatively dry side of the range. The reason lies in a striking combination of topographic and climatic conditions. Snow is a mobile form of precipitation that is shifted about by the wind like a sand dune in the desert. It is not required, like water, to begin a downhill movement as soon as it strikes the earth. Thus, it is a noteworthy fact that snow drifting across the divides may ultimately cause the largest snowfields to lie where the least snow actually falls. This is illustrated in the Bighorns of Wyoming and others of our western ranges. It is, however, not the wet snow near the snowline, but chiefly the dry snow of higher altitudes that is affected. What is now the dry or leeward side of the Cordillera appears in glacial times to have actually received more snow than the wet windward side.