mits the use of basins of both streams for storage purposes. This condition exists on Salt River, where the Roosevelt Dam is located, immediately below the mouth of Tonto Creek, and also at the San Carlos dam site, which is located on the Gila, a short distance below the mouth of the San Carlos River.

Several other excellent dam sites are found along the Gila, both below and above the San Carlos Dam site; those below at the Buttes, at Riverside, and possibly at Sentinel; those above at Guthrie and Red Rock, on the Gila proper; and at Alma, on the San Francisco. The three sites above San Carlos may form a solution of the difficult silt problem, which now seems seriously to limit the life of the proposed San Carlos project. If the construction of these reservoirs would so regulate the flow of the Gila as to prevent destructive floods, which now invariably erode and carry down the river large sections of cultivated land, it seems that, while the silt problem would not be entirely solved, the life of the proposed reservoir might be increased to compare favorably with that at Roosevelt.

The area drained by the Gila and its tributaries west of Florence does not lend itself favorably to the construction of storage works, owing to the absence of reservoir sites and the great depth to bedrock, although one reservoir site, located near Gila Bend, has received considerable attention. This region is, however, admirably adapted to the storage of underground water in the voids of immense deposits of débris. It is safe to predict that considerable development may be expected in this section through pumping. Over`a considerable portion of this area, however, the deposits are very deep, and the water-bearing gravels lie at too great a depth for economical pumping.

The rich alluvial bottoms of the valleys about Florence and the broad, fertile plains to the west only require the application of water to make them of great agricultural value.

CLIMATE.

The great difference in elevation existing in the Gila River watershed induces great climatic variations. Along the headwaters of the Gila, at elevations ranging from 5,000 to 7,000 feet, the climate is comparable to that of Virginia. The average maximum temperature is 100 degrees, the average minimum is -8 degrees.

In the Solomonville Valley, which includes San Carlos, and which is situated at an elevation of 2,000 to 3,000 feet, the average maximum temperature is 106 degrees, the average minimum temperature is 8 degrees, while the normal precipitation is 13 inches.

From Florence to the mouth of the Gila at Yuma the elevation drops from 1,500 feet to 140 feet above sea level, and true desert conditions prevail. Here the average maximum temperature is 113° F., the average minimum temperature is 21° F., while the normal annual precipitation does not exceed 10 inches. It is this section that is responsible for Arizona's reputation for intensely hot summers; rather unjustly so, however, as a temperature of 120° F. at Yuma, owing to the low humidity, is scarcely more uncomfortable than 90° F. at Washington. D. C.

149700-19-VOL 2-2

Throughout the entire watershed radiation is very great and produces a wide range in temperature from day to night, averaging 30° F. in the upper region and 10° F. in the lower valley.

There are two wet seasons, one in summer (July and August), the other in winter (October to March), which are separated by much drier periods. The summer rainy season is characterized by frequent violent local showers, which are due to convectional atmospheric currents and are usually accompanied by severe electrical

storms.

The winter is marked by steady, gentle rains and by snowstorms at elevations above 4,000 feet. The winter precipitations are due to general widespread cyclonic disturbances.

Rainfall records have been kept at a number of different points in the Gila Basin. These records have been published in several previous reports, notably in Water-Supply Paper No. 33, page 20, and in the more recent report of the Army board elsewhere referred to, published as House Document No. 791, Sixty-third Congress, second session.

The local summer storms give rise to and are frequently accompanied by heavy winds, but during the remainder of the year very little wind occurs. An average wind movement at Phoenix is given at 2.41 miles an hour.

The atmospheric humidity over the entire area is unusually low. According to United States Department of Agriculture Bulletin 235, page 17,"Relative humidities of less than 10 per cent are often recorded in June, the annual average for four years at Phoenix being 35 per cent."

Owing to the prevailing high normal temperatures and low humidity, evaporation even in the absence of wind is very rapid and must be taken into consideration on planning any irrigation project. But little observed data are available. Deductions must be drawn from results obtained in regions where similar conditions prevail. In the report of the Army board the conclusion has been made that the mean annual evaporation at San Carlos is 60 inches.

VEGETATION.

The great range in elevation encountered in the Gila watershed, and the consequent variation in climate and rainfall, bring about. striking changes in vegetation.

The upper region, above the elevation of 5,500 feet, are clothed in a magnificent forest of pine, fir, spruce, and juniper. Below this elevation, to 3,000 feet, the mountain sides are blanketed with a dense thicket of brush, scrub oak, etc., while occasional groves of juniper are found. Below 3,000 feet the usual desert vegetation is in evidence, consisting of mesquite, paloverde, ironwood, greasewood, sage, and the several varieties of cactus. The stream courses of the upper regions are fringed with cottonwood, walnut, and willow. The lower Gila is bordered by dense mats of "water motes" and willow, while the uncultivated tracts of the valley land are generally covered with thick groves of mesquite.

1 Arizona Sta. Bul. 20, p. 36, 37.

Formerly, during the late spring and early summer, practically the entire unforested area of the watershed was covered with a thick carpet of various wild grasses and alfilaria. This, owing to overgrazing or other causes, has largely disappeared, and is not now sufficient appreciably to retard the run-off. As a result, the run-off from the lower regions is rapid, and in some localities considerable erosion has taken place, contributing still more to the muddy waters of the Gila.

CROPS.

While a considerable variety of natural vegetation occurs in the different portions of the Gila watershed, no great variety of crops are grown in the irrigated districts along the Gila River, and, except in the extreme upper portion of the watershed, the irrigation season extends throughout the year when sufficient water is available. In general the climatic conditions are such that plants which are grown. throughout the year must be able to stand excessive heat and must also be able to endure some frost. The principal plant of this nature is alfalfa, and ordinarily five crops per annum are cut, yielding a total of 4 to 8 tons per acre. Wherever sufficient water is available throughout the Gila River Basin, alfalfa is the staple product.

When two different kinds of products are planted during the year, the first or spring crop must be of a variety favored by a long, hot, growing season, while the second or fall crop must be able to withstand some frost. Ordinarily small grains, such as wheat, oats, and barley, are planted in the fall and harvested in the late spring, after which the ground is planted to corn, sorghum, cotton, watermelons, squash, or other hot-weather crops.

The yield of grain crops, under favorable conditions, varies from 1,900 to 2,500 pounds per acre for barley, while a wheat crop varies from 1,500 to 2,400 pounds per acre. Of the summer crops, Egyptian cotton yields from 400 to 1,000 pounds per acre, while the corn crop averages 1 ton to the acre.

Subtropical fruits, such as dates, olives, oranges and lemons, figs and pomegranates are successfully grown in the lower Gila valleys, while in the upper valleys the fruit trees are of the deciduous variety.

HYDROGRAPHY.

The topographic and physiographic conditions obtaining in the Gila watershed are so varied that general statements covering the stream as a whole are, as a rule, impossible.

The watershed forms a complete river system, and each tributary presents its own individual problem. The streams comprising the system are, in order, from the source of the Gila to its mouth, the Upper Gila, San Francisco, Eagle Creek, San Carlos, San Pedro, Santa Cruz, Salt River, Agua Frio, and Hassayampa.

Reference to the map No. 1 of the watershed, accompanying this report, shows that all but two of these tributaries (namely the Santa Cruz and San Pedro) enter the Gila from the north, and that the more important of them have their source in the high mountain region in the vicinity of Mounts Thomas and Ord, in eastern Arizona and western New Mexico, of the two streams entering from the south,

the San Pedro only is of sufficient importance to warrant investigation.

The Salt River, Agua Frio, and Hassayampa join the Gila below the western limits of the Gila River Indian Reservation, and as the flow of the Salt River is largely constrolled by the Salt River Reservoir, and the flow of the Agua Frio and Hassayampa are of minor importance, and have only an indirect bearing on the proposed San Carlos project, the study of these streams has not received the attention required on the upper tributaries.

The headwaters of the Gila and its principal tributaries are typical mountain streams, while the Gila throughout its lower course is a characteristic desert river flowing over a shifting sandy bed, except where it cuts through the mountains that separate the several valleys.

The Gila, owing to the occurrence of the two rainy seasons already mentioned, is subject to two annual flood periods. The winter flood is the longest, usually continuing from December to the 1st of May, and, owing to melting snows in the mountains, reaching its maximum in March. The summer flood extends from the 1st of July to the 1st of September, or nearly October, and is the result of heavy rainfall concentrated over small areas along the Gila and its tributaries.

The discharge of the Upper Gila and its tributaries is fairly uniform, owing to the retarding effect of snowfall, and the heavily forested areas through which they flow. This flow is, however, largely dissipated through evaporation and seepage losses, before the lower valleys are reached. The territory contiguous to the lower Gila is sparsely covered with vegetation, and much of the region, aside from the valley proper, is of a steep and rocky nature, offering favorable conditions for an exceptionally large and rapid run-off. As a result the flow of the lower Gila often diminishes to a negligible quantity during low water, while during rainy seasons it is subject to violent and erratic floods. The current during flood periods, as it follows the tortuous channel of the Gila, often impinges against the soft silt banks and washes away considerable areas of cultivated land, which is later deposited lower down the stream. The Gila also possesses a characteristic common among western streams of deepening its beds during flood periods. A deepening of the channel from 10 to 15 feet often occurs with a rise of only a few feet. It is this fact that renders maintenance of diversions and protection work along the Gila so difficult, as there appears to be no economical means of preventing undercutting.

As set forth in the accompanying Appendix A, irrigation was practiced by the Pimas, and possibly other tribes along the Gila and its tributaries, long before the coming of the white man. At that time a considerable area was irrigated by the Pimas in the vicinity of Sacaton. Irrigation was commenced by the whites in the vicinity of Florence, and later was extended to include the Solomonville, Safford, and the Duncan Valleys. The small amount of water available during the low-water season placed a definite limit on the amount of land it was practicable to irrigate, and in general the irrigation of additional land higher up the stream was made at the expense of that lower down. This shortage was seriously felt at Florence and by

the Pima Indians on the Gila River Indian Reservation. Early ef forts were made by the settlers to increase the low-water supply and the situation was investigated as early as 1894 by the United States Geological Survey.

The great value of water for irrigation, and the possibility of storage along the Gila, also attracted the attention of parties interested in the development of the region. Experienced engineers were engaged and several reports prepared, which contain valuable information respecting the flow of the Gila.

The United States Army board, in the report already referred to, assembled and collated the data from the several sources. Unfortunately, the records have not been continuous at any one point, and an effort was made to supply the missing information by means of various assumptions based on correlated data.

As stated above, the low-water flow in the lower valleys is considerably affected by losses through diversions for irrigation purposes along the upper reaches of the streams. The exact amount of such losses is difficult to determine, since there are several vital factors that must be taken into consideration which are more or less obscure, namely, seepage losses in the wide, sandy stream channel, evaporation, transpiration from plant growth, return water from the canals, and possible inflow from hidden springs and other sources. relating to these factors are of great importance in determining the amount of water legally available where conflicting interests exist along the course of a stream, such as the Gila. As previously stated, an investigation to obtain this information, as well as to determine the influence of the several tributaries on the flow of the Gila was made by Mr. C. C. Jacob, district engineer of the Water Resources Branch of the Geological Survey, whose report appears in Appendix B.

Data

The problem was attacked from a different angle by Mr. H. K. Palmer, of this service, who arrived at certain theoretical conclusions respecting the effect of the diversions along the upper Gila, on the low-water flow of the lower valley. His arguments are presented in Appendix B, accompanying this report. The conclusions are based on logical deductions respecting the effect of transpiration and evaporation losses on the amount of water diverted in the canals above the San Carlos Box Canyon, in the Solomonville-Safford Valley. Not effort was made to include the diversion in Duncan Valley or on the upper tributaries, as the effect of these diversions on the low-water flow at Florence was involved and the information was insufficient to permit of a definite result.

While the conclusions are theoretical and are therefore necessarily subject to a wide interpretation, they appear to be reasonable, and at least give an idea of the low water conditions prevailing along the Gila before the coming of the whites.

Appendix B, accompanying this report, includes an analysis of all the assembled data. In the volume of maps accompanying this report are charts showing graphically the flow of the Gila at Florence and San Carlos, and by means of mass curves, the possibility of storage at the latter point.

The analysis shows that the average low-water flow of the Gila at San Carlos, exclusive of the amount diverted above that point for irrigation, is 82 feet per second. Including that amount, with the