Fig. 65—Regional diagram to show the physical relations in the coastal desert of Peru. For location, see Fig. 20.
If the reader will examine Figs. 65 and 66, and the photographs that accompany them, he may gain an idea of the more important features of the coastal region. We have already described, in Chapters V and VII, the character of the plateau region and its people. Therefore, we need say little in this place of the part of the Maritime Cordillera that is included in the figure. Its unpopulated rim (see p. 54), the semi-nomadic herdsmen and shepherds from Chuquibamba that scour its pastures in the moist vales about Coropuna, and the gnarled and stunted trees at 13,000 feet (3,960 m.) which partly supply Chuquibamba with firewood, are its most important features. A few groups of huts just under the snowline are inhabited for only a part of the year. The delightful valleys are too near and tempting. Even a plateau Indian responds to the call of a dry valley, however he may shun the moist, warm valleys on the eastern border of the Cordillera.
Fig. 66—Irrigated and irrigable land of the coastal belt of Peru. The map exhibits in a striking manner how small a part of the whole Pacific slope is available for cultivation. Pasture grows over all but the steepest and the highest portions of the Cordillera to the right of (above) the dotted line. Another belt of pasture too narrow to show on the map, grows in the fog belt on the seaward slopes of the Coast Range. Scale, 170 miles to the inch.
The greater part of the coastal region is occupied by the desert. Its outer border is the low, dry, gentle, eastward-facing slope of the Coast Range. Its inner border is the foot of the steep descent that marks the edge of the lava plateau. This descent is a fairly well-marked line, here and there broken by a venturesome lava flow that extends far out from the main plateau. Within these definite borders the desert extends continuously northwestward for hundreds of miles along the coast of Peru from far beyond the Chilean frontier almost to the border of Ecuador. It is broken up by deep tranverse valleys and canyons into so-called “pampas,” each of which has a separate name; thus west of Arequipa between the Vitor and Majes valleys are the “Pampa de Vitor” and the “Pampa de Sihuas,” and south of the Vitor is the “Pampa de Islay.”
The pampa surfaces are inclined in general toward the sea. They were built up to their present level chiefly by mountain streams before the present deep valleys were cut, that is to say, when the land was more than a half-mile lower. Some of their material is wind-blown and on the walls of the valleys are alternating belts of wind-blown and water-laid strata from one hundred to four hundred feet thick as if in past ages long dry and long wet periods had succeeded each other. The wind has blown sand and dust from the desert down into the valleys, but its chief work has been to drive the lighter desert waste up partly into the mountains and along their margins, partly so high as to carry it into the realm of the lofty terrestrial winds, whence it falls upon surfaces far distant from the fields of origin. There are left behind the heavier sand which the wind rolls along on the surfaces and builds into crescentic dunes called médanos, and the pebbles that it can sandpaper but cannot remove bodily. Thus there are belts of dunes, belts of irregular sand drifts, and belts of true desert “pavement” (a residual mantle of faceted pebbles and irregular stones).
Yet another feature of the desert pampa are the “dry” valleys that join the through-flowing streams at irregular intervals, as shown in the accompanying regional diagram. If one follow a dry valley to its head he will find there a set of broad and shallow tributaries. Sand drifts may clog them and appear to indicate that water no longer flows through them. They are often referred to by unscientific travelers as evidences of a recent change of climate. I had once the unusual opportunity (in the mountains of Chile) of seeing freshly fallen snow melted rapidly and thus turned suddenly into the streams. In 1911 this happened also at San Pedro de Atacama, northern Chile, right in the desert at 8,000 feet (2,440 m.) elevation, and in both places the dry, sand-choked valleys were cleaned out and definite channels reëstablished. From a large number of facts like these we know that the dry valleys represent the work of the infrequent rains. No desert is absolutely rainless, although until recently it was the fashion to say so. Naturally the wind, which works incessantly, partly offsets the work of the water. Yet the wind can make but little impression upon the general outlines of the dry valleys. They remain under the dominance of the irregular rains. These come sometimes at intervals of three or four years, again at intervals of ten to fifteen years, and some parts of the desert have probably been rainless for a hundred years. Some specific cases are discussed in the chapter on Climate.
The large valleys of the desert zone have been cut by snow-fed streams and then partly filled again so that deep waste lies on their floors and abuts with remarkable sharpness against the bordering cliffs (Fig. 155). Extensive flats are thus available for easy cultivation, and the through-flowing streams furnish abundant water to the irrigating canals. The alluvial floor begins almost at the foot of the steep western slope of the lava plateau, but it is there stony and coarse—hence Chuquibamba, or plain of stones (chuqui = stone; bamba = plain). Farther down and about half-way between Chuquibamba and Aplao (Camaná Quadrangle) it is partly covered with fresh mud and sand flows from the bordering valley walls and the stream is intrenched two hundred feet. A few miles above Aplao the stream emerges from its narrow gorge and thenceforth flows on the surface of the alluvium right to the sea. Narrow places occur between Cantas and Aplao, where there is a projection of old and hard quartzitic rock, and again above Camaná, where the stream cuts straight across the granite axis of the Coast Range. Elsewhere the rock is either a softer sandstone or still unindurated sands and gravels, as at the top of the desert series of strata that are exposed on the valley wall. The changing width of the valley is thus a reflection of the changing hardness of the rock.
There is a wide range of products between Chuquibamba at 10,000 feet (3,050 m.) at the head of the valley and Camaná near the valley mouth. At the higher levels fruit will not grow—only alfalfa, potatoes, and barley. A thousand feet below Chuquibamba fruit trees appear. Then follows a barren stretch where there are mud flows and where the river is intrenched. Below this there is a wonderful change in climate and products. The elevation falls off 4,000 feet and the first cultivated patches below the middle unfavorable section are covered with grape vines. Here at 3,000 feet (900 m.) elevation above the sea begin the famous vineyards of the Majes Valley, which support a wine industry that dates back to the sixteenth century. Some of the huge buried earthenware jars for curing the wine at Hacienda Cantas were made in the reign of Philip II.
The people of Aplao and Camaná are among the most hospitable and energetic in Peru, as if these qualities were but the reflection of the bounty of nature. Nowhere could I see evidences of crowding or of the degeneracy or poverty that is so often associated with desert people. Water is always plentiful; sometimes indeed too plentiful, for floods and changes in the bed of the river are responsible for the loss of a good deal of land. This abundance of water means that both the small and the large landowners receive enough. There are none of the troublesome official regulations, as in the poorer valleys with their inevitable favoritism or downright graft. Yet even here the valley is not fully occupied; at many places more land could be put under cultivation. The Belaunde brothers at Cantas have illustrated this in their new cotton plantation, where clearings and new canals have turned into cultivated fields tracts long covered with brush.
The Majes Valley sorely lacks an adequate port. Its cotton, sugar, and wine must now be shipped to Camaná and thence to Mollendo, either by a small bi-weekly boat, or by pack-train over the coast trail to Quilca, where ocean steamers call. This is so roundabout a way that the planters of the mid-valley section and the farmers of the valley head now export their products over the desert trail from Cantas to Vitor on the Mollendo-Arequipa railroad, whence they can be sent either to the cotton mills or the stores of Arequipa, the chief distributing market of southern Peru, or to the ocean port.
The foreshore at Camaná is low and marshy where the salt water covers the outer edge of the delta. In the hollow between two headlands a broad alluvial plain has been formed, through which the shallow river now discharges. Hence the natural indentation has been filled up and the river shoaled. To these disadvantages must be added a third, the shoaling of the sea bottom, which compels ships to anchor far off shore. Such shoals are so rare on this dry and almost riverless coast as to be a menace to navigation. The steamer Tucapelle, like all west-coast boats, was sailing close to the unlighted shore on a very dark night in April, 1911, when the usual fog came on. She struck the reef just off Camaná. Half of her passengers perished in trying to get through the tremendous surf that broke over the bar. The most practicable scheme for the development of the port would seem to be a floating dock and tower anchored out of reach of the surf, and connected by cable with a railway on shore. Harbor works would be extraordinarily expensive. The valley can support only a modest project.
The relations of 65 , representing the Camaná-Vitor region, are typical of southern Peru, with one exception. In a few valleys the streams are so small that but little water is ever found beyond the foot of the mountains, as at Moquegua. In the Chili Valley is Arequipa (8,000 feet), right at the foot of the big cones of the Maritime Cordillera (see Fig. 6). The green valley floor narrows rapidly and cultivation disappears but a few miles below the town. Outside the big valleys cultivation is limited to the best spots along the foot of the Coast Range, where tiny streams or small springs derive water from the zone of clouds and fogs on the seaward slopes of the Coast Range. Here and there are olive groves, a vegetable garden, or a narrow alfalfa meadow, watered by uncertain springs that issue below the hollows of the bordering mountains.
Fig. 67—Irrigated and irrigable land in the Ica Valley of the coastal desert of Peru. |
Fig. 68—The projected canal to convey water from the Atlantic slope to the Pacific slope of the Maritime Cordillera.[19] |
In central and northern Peru the coastal region has aspects quite different from those about Camaná. At some places, for example north of Cerro Azul, the main spurs of the Cordillera extend down to the shore. There is neither a low Coast Range nor a broad desert pampa. In such places flat land is found only on the alluvial fans and deltas. Lima and Callao are typical. 66 , compiled from Adams’s reports on the water resources of the coastal region of Peru, shows this distinctive feature of the central region. Beyond Salaverry extends the northern region, where nearly all the irrigated land is found some distance back from the shore. The farther north we go the more marked is this feature, because the coastal belt widens. Catacaos is several miles from the sea, and Piura is an interior place. At the extreme north, where the rains begin, as at Tumbez, the cultivated land once more extends to the coast.
Fig. 69—A stream of the intermittent type in the coastal desert of Peru. Depth of water in the Puira River at Puira, 1905. (Bol. de Minas del Perú, 1906, No. 45, p. 2.)
Fig. 70—A stream of the perennial type in the coastal desert of Peru. Depth of water in the Chira River at Sullana, 1905. Data from May to September are approximate. (Bol. de Minas del Perú, 1906, No. 45, p. 2.)
These three regions contain all the fertile coastal valleys of Peru. The larger ones are impressive—with cities, railways, ports, and land in a high state of cultivation. But they are after all only a few hundred square miles in extent. They contain less than a quarter of the people. The whole Pacific slope from the crest of the Cordillera has about 15,000 square miles (38,850 sq. km.), and of this only three per cent is irrigated valley land, as shown in 66 . Moreover, only a small additional amount may be irrigated, perhaps one half of one per cent. Even this amount may be added not only by a better use of the water but also by the diversion of streams and lakes from the Atlantic to the Pacific. Figs. 67 and 68 represent such a project, in which it is proposed to carry the water of Lake Choclococha through a canal and tunnel under the continental divide and so to the head of the Ica Valley. A little irrigation can be and is carried on by the use of well water, but this will never be an important source because of the great depth to the ground water, and the fact that it, too, depends ultimately upon the limited rains.
The inequality of opportunity in the various valleys of the coastal region depends in large part also upon inequality of river discharge. This is dependent chiefly upon the sources of the streams, whether in snowy peaks of the main Cordillera with fairly constant run-off, or in the western spurs where summer rains bring periodic high water. A third type has high water during the time of greatest snow melting, combined with summer rains, and to this class belongs the Majes Valley with its sources in the snow-cap of Coropuna. The other two types are illustrated by the accompanying diagrams for Puira and Chira, the former intermittent in flow, the latter fairly constant.[20]
The noble proportions of the Peruvian Andes and their position in tropical latitudes have given them climatic conditions of great diversity. Moreover, their great breadth and continuously lofty summits have distributed the various climatic types over spaces sufficiently ample to affect large and important groups of people. When we add to this the fact that the topographic types developed on a large scale are distributed at varying elevations, and that upon them depend to a large degree the chief characteristics of the soil, another great factor in human distribution, we are prepared to see that the Peruvian Andes afford some striking illustrations of combined climatic and topographic control over man.
The topographic features in their relations to the people have been discussed in preceding chapters. We shall now examine the corresponding effects of climate. It goes without saying that the topographic and climatic controls cannot and need not be kept rigidly apart. Yet it seems desirable, for all their natural interdependence, to give them separate treatment, since the physical laws upon which their explanations depend are of course entirely distinct. Further, there is an independent group of human responses to detailed climatic features that have little or no connection with either topography or soil.
The chief climatic belts of Peru run roughly from north to south in the direction of the main features of the topography. Between 13° and 18° S., however, the Andes run from northwest to southeast, and in short stretches nearly west-east, with the result that the climatic belts likewise trend westward, a condition well illustrated on the seventy-third meridian. Here are developed important climatic features not found elsewhere in Peru. The trade winds are greatly modified in direction and effects; the northward-trending valleys, so deep as to be secluded from the trades, have floors that are nearly if not quite arid; a restricted coastal region enjoys a heavier rainfall; and the snowline is much more strongly canted from west to east than anywhere else in the long belt of mountains from Patagonia to Venezuela. These exceptional features depend, however, upon precisely the same physical laws as the normal climatic features of the Peruvian Andes. They can, therefore, be more easily understood after attention has been given to the larger aspects of the climatic problem of which they form a part.
The critical relations of trade winds, lofty mountains, and ocean currents that give distinction to Peruvian climate are shown in Figs. 71 to 73. From them and 74 it is clear that the two sides of the Peruvian mountains are in sharp contrast climatically. The eastern slopes have almost daily rains, even in the dry season, and are clothed with forest. The western leeward slopes are so dry that at 8,000 feet even the most drought-resisting grasses stop—only low shrubs live below this level, and over large areas there is no vegetation whatever. An exception is the Coast Range, not shown on these small maps, but exhibited in the succeeding diagram. These have moderate rains on their seaward (westerly) slopes during some years and grass and shrubby vegetation grow between the arid coastal terraces below them and the parched desert above. The greatest variety of climate is enjoyed by the mountain zone. Its deeper valleys and basins descend to tropical levels; its higher ranges and peaks are snow-covered. Between are the climates of half the world compressed, it may be, between 6,000 and 15,000 feet of elevation and with extremes only a day’s journey apart.
Fig. 71—The three chief topographic regions of Peru.
Fig. 72—The wind belts of Peru and ocean currents of
adjacent waters.
Fig. 73—The climatic belts of Peru.
Fig. 74—Belts of vegetation in Peru.
In the explanation of these contrasts we have to deal with relatively simple facts and principles; but the reader who is interested chiefly in the human aspects of the region should turn to p. 138 where the effects of the climate on man are set forth. The ascending trades on the eastern slopes pass successively into atmospheric levels of diminishing pressure; hence they expand, deriving the required energy for expansion from the heat of the air itself. The air thereby cooled has a lower capacity for the retention of water vapor, a function of its temperature; the colder the air the less water vapor it can take up. As long as the actual amount of water vapor in the air is less than that which the air can hold, no rain falls. But the cooling process tends constantly to bring the warm, moist, ascending air currents to the limit of their capacity for water vapor by diminishing the temperature. Eventually the air is saturated and if the capacity diminishes still further through diminishing temperature some of the water vapor must be condensed from a gaseous to a liquid form and be dropped as rain.
The air currents that rise thousands of feet per day on the eastern slopes of the Andes pass again and again through this practically continuous process and the eastern aspect of the mountains is kept rain-soaked the whole year round. For the trades here have only the rarest reversals. Generally they blow from the east day after day and repeat a fixed or average type of weather peculiar to that part of the tropics under their steady domination. During the southern summer, when the day-time temperature contrasts between mountains and plains are strongest, the force of the trade wind is greatly increased and likewise the rapidity of the rain-making processes. Hence there is a distinct seasonal difference in the rainfall—what we call, for want of a better name, a “wet” and a “dry” season.
On the western or seaward slopes of the Peruvian Andes the trade winds descend, and the process of rain-making is reversed to one of rain-taking. The descending air currents are compressed as they reach lower levels where there are progressively higher atmospheric pressures. The energy expended in the process is expressed in the air as heat, whence the descending air gains steadily in temperature and capacity for water vapor, and therefore is a drying wind. Thus the leeward, western slopes of the mountains receive little rain and the lowlands on that side are desert.
A series of narrow but pronounced climatic zones coincide with the topographic subdivisions of the western slope of the country between the crest of the Maritime Cordillera and the Pacific Ocean. This belted arrangement is diagrammatically shown in 75 . From the zone of lofty mountains with a well-marked summer rainy season descent is made by lower slopes with successively less and less precipitation to the desert strip, where rain is only known at irregular intervals of many years’ duration. Beyond lies the seaward slope of the Coast Range, more or less constantly enveloped in fog and receiving actual rain every few years, and below it is the very narrow band of dry coastal terraces.
Fig. 75—Topographic and climatic provinces in the coastal region of Peru. The broadest division, into the zones of regular annual rains and of irregular rains, occurs approximately at 8,000 feet but is locally variable. To the traveler it is always clearly defined by the change in architecture, particularly of the house roofs. Those of the coast are flat; those of the sierra are pitched to facilitate run off.
The basic cause of the general aridity of the region has already been noted; the peculiar circumstances giving origin to the variety in detail can be briefly stated. They depend upon the meteorologic and hydrographic features of the adjacent portion of the South Pacific Ocean and upon the local topography.
The lofty Andes interrupt the broad sweep of the southeast trades passing over the continent from the Atlantic; and the wind circulation of the Peruvian Coast is governed to a great degree by the high pressure area of the South Pacific. The prevailing winds blow from the south and the southeast, roughly paralleling the coast or, as onshore winds, making a small angle with it. When the Pacific high pressure area is best developed (during the southern winter), the southerly direction of the winds is emphasized, a condition clearly shown on the Pilot Charts of the South Pacific Ocean, issued by the U.S. Hydrographic Office.
Fig. 76—Temperatures at Callao, June-September, 1912, from observations taken by Captain A. Taylor, of Callao. Air temperatures are shown by heavy lines; sea temperatures by light lines. In view of the scant record for comparative land and water temperatures along the Peruvian coast this record, short as it is, has special interest.
The hydrographic feature of greatest importance is the Humboldt Current. To its cold waters is largely due the remarkably low temperatures of the coast.[21] In the latitude of Lima its mean surface temperature is about 10° below normal. Lima itself has a mean annual temperature 4.6° F. below the theoretical value for that latitude, (12° S.). An accompanying curve shows the low temperature of Callao during the winter months. From mid-June to mid-September the mean was 61° F., and the annual mean is only 65.6° F. (18° C.). The reduction in temperature is accompanied by a reduction in the vapor capacity of the super-incumbent air, an effect of which much has been made in explanation of the west-coast desert. That it is a contributing though not exclusive factor is demonstrated in Fig. 77. Curve A represents the hypothetical change of temperature on a mountainous coast with temporary afternoon onshore winds from a warm sea. Curve B represents the change of temperature if the sea be cold (actual case of Peru). The more rapid rise of curve B to the right of X-X′, the line of transition, and its higher elevation above its former saturation level, as contrasted with A, indicates greater dryness (lower relative humidity). There has been precipitation in case A, but at a higher temperature, hence more water vapor remains in the air after precipitation has ceased. Curve B ultimately rises nearly to the level of A, for with less water vapor in the air of case B the temperature rises more rapidly (a general law). Moreover, the higher the temperature the greater the radiation. To summarize, curve A rises more slowly than curve B, (1) because of the greater amount of water vapor it contains, which must have its temperature raised with that of the air, and thus absorbs energy which would otherwise go to increase the temperature of the air, and (2) because its loss of heat by radiation is more rapid on account of its higher temperature. We conclude from these principles and deductions that under the given conditions a cold current intensifies, but does not cause the aridity of the west-coast desert.
Fig. 77—To show progressive lowering of saturation temperature in a desert under the influence of the mixing process whereby dry and cool air from aloft sinks to lower levels thus displacing the warm surface air of the desert. The evaporated moisture of the surface air is thus distributed through a great volume of upper air and rain becomes increasingly rarer. Applied to deserts in general it shows that the effect of any cosmic agent in producing climatic change from moist to dry or dry to moist will be disproportionately increased. The shaded areas C and C’ represent the fog-covered slopes of the Coast Range of Peru as shown in Fig. 92. X-X’ represents the crest of the Coast Range.
Curves a and b represent the rise of temperature in two contrasted cases of warm and cold sea with the coastal mountains eliminated, so as to simplify the principle applied to A and B. The steeper gradient of b also represents the fact that the lower the initial temperature the dryer will the air become in passing over the warm land. For these two curves the transition line X-X’ coincides with the crest of the Coast Range. It will also be seen that curve a is never so far from the saturation level as curve b. Hence, unusual atmospheric disturbances would result in heavier and more frequent showers.
Fig. 78—Wind roses for Callao. The figures for the earlier period (1897-1900) are drawn from data in the Boletín de la Sociedad Geográfica de Lima, Vols. 7 and 8, 1898-1900: for the latter period data from observations of Captain A. Taylor, of Callao. The diameter of the circle represents the proportionate number of observations when calm was registered.
Fig. 79—Wind roses for Mollendo. The figures are drawn from data in Peruvian Meteorology (1892-1895), Annals of the Astronomical Observatory of Harvard College, Vol. 30, Pt. 2, Cambridge, Mass., 1906. Observations for an earlier period, Feb. 1889-March 1890, (Id. Vol. 39, Pt. 1, Cambridge, Mass. 1890) record S. E. wind at 2 p. m. 97 per cent of the observation time.
Fig. 80—Wind roses for the summer and winter seasons of the years 1911-1913. The diameter of the circle in each case shows the proportion of calm. Figures are drawn from data in the Anuario Meteorológico de Chile, Publications No. 3, (1911), 6 (1912) and 13 (1913), Santiago, 1912, 1914, 1914.
Turning now to local factors we find on the west coast a regional topography that favors a diurnal periodicity of air movement. The strong slopes of the Cordillera and the Coast Range create up-slope or eastward air gradients by day and opposite gradients by night. To this circumstance, in combination with the low temperature of the ocean water and the direction of the prevailing winds, is due the remarkable development of the sea-breeze, without exception the most important meteorological feature of the Peruvian Coast. Several graphic representations are appended to show the dominance of the sea-breeze (see wind roses for Callao, Mollendo, Arica, and Iquique), but interest in the phenomenon is far from being confined to the theoretical. Everywhere along the coast the virazon, as the sea-breeze is called in contradistinction to the terral or land-breeze, enters deeply into the affairs of human life. According to its strength it aids or hinders shipping; sailing boats may enter port on it or it may be so violent, as, for example, it commonly is at Pisco, that cargo cannot be loaded or unloaded during the afternoon. On the nitrate pampa of northern Chile (20° to 25° S.) it not infrequently breaks with a roar that heralds its coming an hour in advance. In the Majes Valley (12° S.) it blows gustily for a half-hour and about noon (often by eleven o’clock) it settles down to an uncomfortable gale. For an hour or two before the sea-breeze begins the air is hot and stifling, and dust clouds hover about the traveler. The maximum temperature is attained at this time and not around 2.00 P. M. as is normally the case. Yet so boisterous is the noon wind that the laborers time their siesta by it, and not by the high temperatures of earlier hours. In the afternoon it settles down to a steady, comfortable, and dustless wind, and by nightfall the air is once more calm.
Fig. 81—Wind roses for Iquique for the summer and winter seasons of the years 1911-1913. The diameter of the circle in each case shows the proportion of calm. For source of data see Fig. 80.
Of highest importance are the effects of the sea-breeze on precipitation. The bold heights of the Coast Range force the nearly or quite saturated air of the sea-wind to rise abruptly several thousand feet, and the adiabatic cooling creates fog, cloud, and even rain on the seaward slope of the mountains. The actual form and amount of precipitation both here and in the interior region vary greatly, according to local conditions and to season and also from year to year. The coast changes height and contour from place to place. At Arica the low coastal chain of northern Chile terminates at the Morro de Arica. Thence northward is a stretch of open coast, with almost no rainfall and little fog. But in the stretch of coast between Mollendo and the Majes Valley a coastal range again becomes prominent. Fog enshrouds the hills almost daily and practically every year there is rain somewhere along their western aspect.
Fig. 82—The wet and dry seasons of the Coast Range and the Cordillera are complementary in time. The “wet” season of the former occurs during the southern winter; the cloud bank on the seaward slopes of the hills is best developed at that time and actual rains may occur. |
Fig. 83—During the southern summer the seaward slopes of the Coast Range are comparatively clear of fog. Afternoon cloudiness is characteristic of the desert and increases eastward (compare 86 ), the influence of the strong sea winds as well as that of the trades (compare 93 B) being felt on the lower slopes of the Maritime Cordillera. |
During the southern winter the cloud bank of the coast is best developed and precipitation is greatest. At Lima, for instance, the clear skies of March and April begin to be clouded in May, and the cloudiness grows until, from late June to September, the sun is invisible for weeks at a time. This is the period of the garua (mist) or the “tiempo de lomas,” the “season of the hills,” when the moisture clothes them with verdure and calls thither the herds of the coast valleys.
Fig. 84—Cloudiness at Callao. Figures are drawn from data in the Boletín de la Sociedad Geográfica de Lima, Vols. 7 and 8, 1898-1900. They represent the conditions at three observation hours during the summers (Dec., Jan.) of 1897-1898, 1898-1899, 1899-1900 and the winters (June, July) of 1898 and 1899.
During the southern summer on account of the greater relative difference between the temperatures of land and water, the sea-breeze attains its maximum strength. It then accomplishes its greatest work in the desert. On the pampa of La Joya, for example, the sand dunes move most rapidly in the summer. According to the Peruvian Meteorological Records of the Harvard Astronomical Observatory the average movement of the dunes from April to September, 1900, was 1.4 inches per day, while during the summer months of the same year it was 2.7 inches. In close agreement are the figures for the wind force, the record for which also shows that 95 per cent of the winds with strength over 10 miles per hour blew from a southerly direction. Yet during this season the coast is generally clearest of fog and cloud. The explanation appears to lie in the exceedingly delicate nature of the adjustments between the various rain-making forces. The relative humidity of the air from the sea is always high, but on the immediate coast is slightly less so in summer than in winter. Thus in Mollendo the relative humidity during the winter of 1895 was 81 per cent; during the summer 78 per cent. Moreover, the temperature of the Coast Range is considerably higher in summer than in winter, and there is a tendency to reëvaporation of any moisture that may be blown against it. The immediate shore, indeed, may still be cloudy as is the case at Callao, which actually has its cloudiest season in the summer but the hills are comparatively clear. In consequence the sea-air passes over into the desert, where the relative increase in temperature has not been so great (compare Mollendo and La Joya in the curve for mean monthly temperature), with much higher vapor content than in winter. The relative humidity for the winter season at La Joya, 1895, was 42.5 per cent; for the summer season 57 per cent. The influence of the great barrier of the Maritime Cordillera, aided doubtless by convectional rising, causes ascent of the comparatively humid air and the formation of cloud. Farther eastward, as the topographic influence is more strongly felt, the cloudiness increases until on the border zone, about 8,000 feet in elevation, it may thicken to actual rain. Data have been selected to demonstrate this eastern gradation of meteorological phenomena.
Fig. 85—Temperature curves for Mollendo (solid lines) and La Joya (broken lines) April, 1894, to December, 1895, drawn from data in Peruvian Meteorology, 1892-1895, Annals of the Astronomical Observatory of Harvard College, Vol. 49, Pt. 2, Cambridge, Mass., 1908. The approximation of the two curves of maximum temperature during the winter months contrasts with the well-maintained difference in minimum temperatures throughout the year.