water from storage. There is no reason why, in such circumstances, under a suitable agreement between the United States and Mexico, this should not apply beyond the United States boundary, so that Mexico, too, may benefit by the use of stored waters whenever an adequate amount of water shall have been provided by the construction of large storage reservoirs in the United States. There is all the more reason for such an arrangement, due to the effect which storage on a large scale would have in reducing the peak of the flood discharge on the lower river. Any reduction of the flood menace thus secured will be of direct benefit to Mexico as well as to the United States.

The levee system which prevents the floods of the Colorado River from taking a northerly course and upon which Imperial Valley is dependent for protection against the menace of destruction, as above stated, is briefly described as follows: 1. The California Development Co. Levee, the so-called C. D. Levee, was built in 1906 and 1907 by the California Development Co. It extends from Hanlon's just north of the international boundary, southerly along the right bank of the Colorado River a distance of 7 miles and thence southwesterly bearing away from the river an additional distance of some 8 miles, being thence continued under the name of the Saiz levee to a point some 7 or 8 miles northeasterly from the area origially occupied by Volcano Lake.

2. The Volcano Lake Levee, on which the first work was done in 1908, and which has been raised from time to time and strengthened and extended. This levee extends from the base of Cerro Prieto easterly for 6 or 7 miles, across the bad lands at the north of Volcano Lake, and thence northeasterly to a connection with the embankment of the Inter-California Railroad. Its total length is about 16 miles.

3. The Ockerson Levee, constructed in 1911 for the primary purpose of returning the flow of the Colorado River back into its original course to the Gulf of California. The river had in 1909 abandoned its channel about 2 miles below the Arizona boundary line and taken a westerly course into the Volcano Lake region along the general route of the Abejas high-water channel which it enlarged to river dimensions. This levee was breached at the head of the Abejas and also at several other places soon after its construction. It has not been repaired. The river still flows into the Volcano Lake region as elsewhere explained.

In 1906 and 1907 the Colorado River was twice turned from an inland course at points about 4 miles south of Pilot Knob by the construction of rock-fill barriers, which now appear as parts of the C. D. or river levee.

The cost of all the above protection work has been large, probably in excess of $5,000,000.

Within the last year the Volcano Lake Levee has been raised 3 feet, at the expense of Imperial irrigation district, and portions of the Saiz Levee have been raised 4 feet. More work is to be done in the near future on these levees, because the menace from the south is growing.

DRIFT

The amount of drift carried by the Colorado River is very large. It is of all kinds. Large logs, 20 to 40 feet and longer, are numerous. The caving banks supply brush and litter of all kinds and also many cottonwood and willow trees which, being green, float nearly or completely submerged. The drift ordinarily runs near the middle of the stream in the river's straight reaches but close inshore along concave banks. Some is sure to find its way into any outlet from the river into which there is considerable flow of water.

At the Laguna Dam some of the larger type of drift can be seen lodged above the dam and here and there hanging to the crest of the structure. By far the largest portion of the drift shows the effect of having been buffeted about and of having been long submerged. Much has been several seasons in traveling from its place of origin to the lower river.

This drift has thus far given no trouble at the Yuma Canal diversion works. Only a negligible amount follows the water into the desilting channel. The little that gets to the head gates is readily removed when the sluice gates are opened for the usual sluicing operations.

When the Yuma Canal is enlarged there will be more drift at the head gates. It may then be a more difficult problem to dispose of it. Booms may be required to deflect the drift from the entrance of the desilting channel.

THE DRIFTING SAND

This board has found no practical way of determining how much sand will blow into an open canal on the 10-mile stretch in which the location of the allAmerican canal is through a region of drifting sand. The sand has been piled by the wind in dunes with crests at some points 100 to 200 feet in height above the general surface of the mesa over which the sand is traveling. These dunes, while much broken, are yet more or less connected in ridges at right angles to the longitudinal axis of the sand-hill area. Between these ridges are depressed areas the limit of the depression being the surface of the mesa of which areas of various size are here and there exposed between the drifting dunes. The mesa itself is built up of detrital sand, but a sand of quite different texture from that found in the sand hills. The sand of the mesa contains a large proportion of coarse sand grains and fine gravel with a little coarser material up to small cobble size, and just enough of very fine material to be slightly cohesive. The wind has apparently whipped out of its surface much of the fine material which is now adrift in the dunes.

The prevailing wind to which the dunes owe their shape and their grouping is from northwest to southeast. When a strong wind is blowing there is a steady stream of sand over the crest of each sand wave and over the tops of the dunes. These dunes are more or less crescent shaped, with flat-sloping, gently rounded backs to the windward, i. e., toward the northwest, and steep slopes generally about 1.75 to 2 feet horizontal to 1 vertical to the leeward or southeast. The momentary condition is subject to considerable variation, depending upon the direction of the wind and the amount of moisture in the sand. Some shrubs and plants that were found growing on the backs of the dunes were deeply rooted, having pushed their growth upward as layer upon layer of sand was piled on by the wind. Their persistence under adverse conditions indicates, however, that the change in surface elevation is probably slow.

Several years ago a plank road was built through these sand hills. Every storm that blew piled some sand on this road. Some difficulty was experienced in keeping it open. And yet, the amount of sand which the wind deposited on this road was relatively small. The attention of one to two men with scraper outfits sufficed to keep the road passable except for short periods during severe wind storms several days in duration. It would signify but little if the amount of sand removed from the road should fall into a large canal. Unfortunately, however, the deposited sand on any such area as a road surface does not give any dependable clue to the sand which would be trapped by a canal in the same location. Some sand is blown across the road in both directions which does not lodge there. A canal across the sand hills would have a water surface some 30 to 50 feet below the surface of the mesa, and if no precautionary measures are taken the sand blown across the top edge of either bank would fall into the water and would lodge within the canal as an obstacle to be artifically removed.

In recommending a large open canal through the sand hills this board recognizes that the sand may prove troublesome. When it does a remedy must be applied. The following is suggested as one that certainly would be successful. One or more suction dredges can be kept at work in the canal making delivery of the excavated material into basins between the sand hills and covering the sand-hill slopes to such extent that these slopes will become fixed. Furthermore, pumped water can be used for the irrigation of the sand-hill slopes until they are so covered with vegetation that for half a mile or more on each side of the canal the drifting sand will be checked.

In view of such possibilities, the cost estimate which we submit is based on the construction of an open cut with comparatively flat slopes and a wide berm. Some protection will also be secured by a proper disposal of the excavated material, with which some of the blow sand can be blanketed.

In the eastern portion of the sand-hill area the drifting sand is of a finer texture than in the westerly portion. It is noticeable that in the eastern two-thirds, approximately, the dunes are in motion from base to crest and that the sand of which they are composed is uniformly fine. It has a very pale, slightly reddish tinge. There is very little vegetation. A sample taken at random, but fairly representative, shows the composition noted below in the table.

On the western one-third of the area darker spots are noticeable on the surface of the dunes. There is a little more vegetation, the vegetation appears more permanent, and the crests are more broadly rounded. Probably more than one-half of the area is of this darker color, which on observation is found to be due to a coarser, grayer sand. This coarser material does not run with the wind

as freely as the finer material, and it is noticeable that secondary dunes of an intermediate grade of sand of comparatively small height are traveling over the more permanent gray sand. The last or most westerly ridge of sand to be cut by the canal is of the coarser texture, and there is therefore less to be feared here from sand drifts than farther east.

The following comparisons will give some idea of the difference in texture of the sands as they appear near the surface:

In determining that the all-American canal should be planned for a capacity of 9,000 second-feet, consideration was given to the following facts:

On the 586,000 acres within Imperial irrigation district it is estimated by those supposed to be best informed that 515,000 acres will ultimately be under cultivation and therefore irrigated. The west-side canal can be extended and constructed on lighter grade, whereby some 40,000 acres more west-side land not now within the district can be added to the irrigated area.

On the east-side mesa there are about 200,000 acres of land, of which it is estimated it is feasible to irrigate 140,000 by gravity flow.

Northward from what is generally known as the east-side mesa, on the eastern slope of the valley and extending to Coachella Valley, there are about 115,000 acres of land that will be commanded by an extension of the mesa canal, of which probably 75,000 acres net should be considered as being suitable for irrigation. Still farther north lies the Coachella Valley, where irrigation might be extended to possibly more than an additional 75,000 acres. Of pumpage area there are about 30,000 acres on the east-side and about 25,000 acres on the west-side mesa. Without at the present time including lands so remote as those to the westward of the northerly end of Salton Sea, the net irrigable areas to be considered aggregate 900,000 acres. It is understood, however, that whenever it shall become apparent that water is available for a larger area and the demand for extension is made the canal can be enlarged by dredging and the irrigated area can be extended beyond the limit named as the basis for the project.

Based on the experience in the Imperial Valley and in the Yuma project, it has been found that a delivery of about 3.33 acre-feet of water per acre per year and a diversion from the river of about 4.44 acre-feet for each acre of the irrigated area should adequately meet the requirements of general farming as the same has been practiced in these regions. It is known that the losses of water from the Imperial Canal in its course through Mexico in a distance of about 50 miles are from 15 to 20 per cent of the amount diverted from the river, or in round numbers, according to the canal stage, about 7.5 to 20 second-feet per mile of canal. It is also known that the losses in the secondary mains in Imperial Valley are light-probably not much in excess of 10 per cent.

The all-American canal will be subject to ordinary canal losses from the Laguna Dam until it gets beyond Pilot Knob. Thence westerly for nearly 20 miles the water surface of the canal will be deep in the ground. The canal bottom in this stretch will be at an average depth of about 50 feet below the surface of the The canal may for a time, at the beginning of operation, lose water here as it would if in shallow excavation, but presently the lower layers of the mesa formation will be filled with water, and the canal losses will be restricted to what

mesa.

these layers can carry away. This amount will probably be small-probably not one-half of the amount which is lost by the Imperial Canal per unit of length. Allowing, however, an equal loss in amount per mile, or from 72 to 20 secondfeet for the all-American canal to the point where delivery into the head of a mesa canal will be made, the loss would be only about 62 to 72 per cent.

This

At the headworks there will be waste due to the desilting operation. may, as elsewhere stated, amount to about 5 per cent. It is believed, under these circumstances, that an allowance of about 15 per cent for all losses down to where the canal will deliver into the east-side mesa canal, will be adequate and that the figures above noted relating to the per acre requirements, based on experience, which allow for 33 per cent more water diverted than is applied to the land, may be accepted as a safe guide in determining the amount of water required to supply the needs of 900,000 acres of cultivated lands.

It is, of course, quite possible that after some years of operation, when the canal beds have been rendered less pervious by deposits of mud and silt, the more sanguine estimate of canal loss as noted in Mr. Preston's report will be realized.

The demand for water by months will be about as shown in the following table, which is based on the assumption, not strictly true, that the canal loss will be proportional to the amount diverted. The percentage of loss will in fact be somewhat less when the canal is flowing full than when at a low stage, because the loss is more nearly proportional to canal water-surface area in the canal than to volume of flow, and the water-surface area increases at a less rate than does the volume of flow.

The maximum demand for irrigation water will determine the canal capacity. The maximum demand is expected in June and July. In these months the diversion from the river should be in such amount that if all the water were delivered without loss in transit it would cover the entire cultivated area in each of these months to a depth of 0.60 foot or 7.2 inches. The duty of each secondfoot of water diverted would, if expressed otherwise, be 100 acres. The canal to supply the desired amount of water to a cultivated area of 900,000 acres in Imperial Valley should have a capacity of 9,000 second-feet.

For the Yuma project, with some allowance for expansion beyond the limits generally indicated, a maximum diversion of 1,600 second-feet has been assumed. This is the amount for which the Reclamation Service desires that capacity be provided and for which the original project canal was constructed.

The required canal capacity, therefore, of 10,600 second-feet results, from the Laguna Dam down to the point at the siphon drop where the Yuma water is to be delivered to a separate canal and 9,000 second-feet for the all-American canal below this point to the upper mesa canal heading where the main supply for the east mesa would be taken off.

The board, upon request of the Directors of Imperial Irrigation District, submits as an alternative the estimated cost of constructing the canal at first at a capacity of only 6,000 second-feet to meet the immediate needs of the district. The structures for this alternative have, however, been planned at full capacity to meet early future requirements. The power plants would be the same as for a full-capacity canal.

METHODS OF CONSTRUCTION AND UNIT COSTS

The work of excavating the canal should be done to the maximum extent possible by machinery. The climatic conditions are not favorable for securing good results with ordinary labor. For work of the magnitude of an all-American canal, particularly for the deep uniform excavation westerly from Pilot Knob, special methods or types of machinery and appliances may be designed to do the work expeditiously and economically. It is possible that such methods or machinery can be perfected and brought into successful operating condition, and that thereby the cost of the work can be somewhat reduced. Nevertheless, it would not be safe to assume other methods of construction as the basis for a -cost estimate than such as have been standardized by experience.

It would be useless to speculate as to what the operating and plant costs of special types of such machinery would be when, perhaps after considerable experimentation, they have been brought into a condition to operate successfully. This board believes that it would be unwise to permit any experimentation which does not at the outset give almost certain promise of success, and it has, therefore, adopted unit costs based on methods of work which experience has standardized and which will give assurance of successful accomplishment within a time limit that can be predicted with some certainty.

It is known that practically all of the work can be done with such standard equipment as drag-line excavators and steam shovels and that such work, under payment for electric energy or the generation of power at the drops or with oil as fuel, should be done at a cost of about 20 to 30 cents per cubic yard when no powder is required to loosen the material, and 75 cents per cubic yard for rock. Wherever the conditions are such that the excavated material must be transported for a considerable distance lengthwise, steam shovels are well adapted for the work. They will load upon cars, and trackage will have to be provided to send the excavated material to selected dumps or to some place where its deposit will be of some use.

For the rock cuts at and adjacent to Laguna Dam, steam shovels with the addition of a clamshell or orange-peel excavator and a hydraulic dredge for the excavation of the desilting basin are contemplated. The material from the rock cuts will be hauled for use in the training jetty of the settling basin and to riprap the levee, which affords protection to the upper end of the main canal. Some of this material may be placed in the reservation levee of the Yuma project, but the haulage costs for this work would not be charged to the all-American canal.

For the enlargement of the 10 miles of the Yuma project main canal to siphon drop large size drag-line excavators are contemplated. These will handle all the material with the exception of a few short stretches where the main canal hugs the base of the mesa bluffs. In these stretches a medium-sized steam shovel will be used.

From siphon drop to Araz practically all the work is suited for drag-line excavators, and from Araz to Pilot Knob for steam shovels.

At Pilot Knob, too, practically all the excavation will be made with steam shovels. Space for spoil along the line of the canal, which will be but a few hundred feet from the international boundary, is limited. The excavated material must nearly all be transported for a considerable distance lengthwise of the canal. It is proposed to have the steam-shovel track system connected with the track on the "River" levee and to extend this track the full length of the Saiz levee and also to put a track on the Ockerson levee, connecting with this line and extending to the head of the Bee River. It will thereby be made possible, at the cost of some extra haul, which is small, to send the material from the canal to the "River," Saiz, and Volcano Lake levees which need raising and strengthening. A small proportion of the steam-shovel excavation on the south slope and a large proportion of that on the east slope of Pilot Knob will be rock, suitable for river bank and levee facing. When once loaded upon cars this material should be placed, so far as it is economically feasible, where it will be useful in affording protection against the flood menace.

In the board's estimates of cost no addition has been made for long haul when material is thus disposed of, it being assumed that some provision will be made to meet this small additional cost out of special funds that will have to be raised from time to time for protective work.

Excavation of the sand-hill stretch of canal by electrically operated shovels is contemplated. Power for the shovels will be furnished from the power station at Pilot Knob. Some grading and construction of track and the establishment