The Journal of San Diego History
Winter 1975, Volume 21, Number 1
James E. Moss, Editor

By Robert L. Sperry

Images from this Article

I am a slave. I have not always been a slave nor will I be one forever. I was born free on high ice and snow covered mountains and went my unhampered way league after league through canyons and mountains and deserts to the lowlands and the sea. And then man made a slave of me. Today I do the bidding of man, go where he directs and do his work. But it will not always be that way.

It was in the lowlands where man first started to place fetters on me when he tapped part of my resources and sent them north. I was happy with the route I was following to the south, which I had been following for some time, but I was also familiar with the one to the north, since I had more than once traveled that way for long periods of time. So, reflecting that the way to the north was easy, and part of my forces were diverted, I decided to close the southern route and go north. Man was displeased and immediately set to work to force me to take all but a small portion of my forces south.

I had never been one to be hampered by obstacles, consequently, when an attempt was made to force me to reverse my direction I became even more determined to follow the new course. Man was equally determined, and the resistance I made, being forced to go where he chose, made stories for his newspapers and magazines as well as two years of hard work for many men and much of their machinery. But man won, and for the first time I was burdened by the chains of his making. Men were happy and congratulated themselves, saying I would never again be free to go where I wished. But they were wrong, even though it may be a long time, as man measures time, before I will again be free to slash my way through ranges of granite mountains and build deltas across an arm of the sea.

Man knows me as the Colorado River and believes I will be a slave forever because he has locked me in my canyon behind great dams of his making and uses my forces as he sees fit. Man does not realize how patient I am, that to me a hundred years is no longer than a tick of his clock. I am working to regain my freedom even as he devises new tasks for his slave. Within ten thousand years, or perhaps just one thousand, I will have filled the reservoirs above the dams with silt and started to remove the dams themselves. Grinding away the dams will be no harder, nor will it take as long as it took me, at an earlier time, to cut through the granite backbone of the earth to form the Grand Canyon and the other canyon lands of the Southwest. Yes, I will be free again, after being a slave to man for what he thinks will be forever, but to me, with my way of measuring time, it will not be very long.

The future looked especially promising to those pioneering families arriving in Imperial Valley in Southern California during the years from 1900 to 1904. It was the American Dream all over again; to push back the frontier, to build a new life in a new land. There would be land to develop and a legacy to build for their children. The desert was an inhospitable land, but it would be changed. Summers so hot as to be almost beyond belief were viewed as only a slight discomfort. Living conditions in the new Valley presented only temporary inconveniences. Their attitude was that shared by other pioneers who had crossed the plains in covered wagons a few decades earlier. It was a continuation of the drive and the spirit that expanded our nation across a vast continent, from a few tiny colonies along the Atlantic to the shores of the Pacific.

Hope rested on water to be drawn from the Colorado river and delivered to the Valley by a canal many miles in length, most of it in Mexico.

When it became known that the area was going to be developed, many people moved into the desert valley. Five townsites were surveyed and thousands of acres were made ready for farming and crops planted. Several orchards were planted by farmers looking toward the future. In the new towns businesses were started and everything was getting off to a good start. During 1904, a branch railroad was extended from Old Beach to Calexico, not only opening distant markets for Valley products but making the area seem less remote.

The Fall of 1903 brought the first indication that progress might not continue as smoothly as had been hoped. It was not in the form of an overwhelming disaster, but appeared only as a small cloud on the horizon as warning of an approaching storm.

A chain of events, some caused by man, and others brought about by the wild and unpredictable Colorado river, would change the wonderful dream of many into a nightmare.

The California Development Company, headed at the time by Engineer-Financier George Chaffey, had built the “Chaffey Gate” to admit Colorado river water to the main canal. Chaffey, drawn into the project because of his experience in irrigation and his willingness to supply funds for continuing development, replaced C. R. Rockwood as Chief Engineer during parts of 1901-1902. The Chaffey Gate, built of wood, was located one mile north of the Mexican border beside a small, once volcanic mountain known as Pilot Knob. The canal started below the gate, extending directly south across the Mexican border a short distance until it joined an ancient overflow channel of the Colorado known as the Alamo river. This channel, used as the main canal, curved southwest around Pilot Knob and continued west for approxi­mately fifty miles, passing the south end of the ridge of sand hills that cross Imperial Valley from north to south. A wooden control structure was built in the main canal ten miles east of Calexico to divert and regulate water flow to the lateral and smaller canals of the distribution system. This essential structure was named Sharp’s Heading and was to play an important part in the irrigation system for many years.

In securing permission to use the ancient overflow channel in Mexico as the main canal, Rockwood and officials of the California Development Company resorted to means that the Mexican Government considered devious and underhanded. Mexico’s Constitution, then as now, prohibits foreigners from owning land within 100 kilometers of the border. When Mexican officials learned the California Development Company had purchased one hundred thousand acres of land on both sides of the Alamo from Mexican owner Guillermo Andrade, they were surprised, disturbed and resentful. Official protests to the United States Government forced Rockwood’s Company to organize a Mexican Company called “Sociedad de Irrigación y Terrenos de la Baja California” to hold title to the land. In addition to money paid Andrade for the land, he received commitments for all water necessary for irrigating over six hundred thousand acres of other land in which he had an interest.

Officials of the Development Company found themselves in conflict with the Governments of both the United States and Mexico. The Colorado had been classified as a navigable stream, and Mexico claimed that restricting the stream or diverting water was a treaty violation. Government officials in Washington contended that it was not. Rockwood spent much time in Washington trying, without success, to convince officials the Colorado should be reclassified, contending navigation was of no importance whatever and irrigation, on the other hand, of very great importance. As time went on other events added to the tension between the United States and Mexico, particularly with the people who had settled in Imperial Valley. Mexico, mindful of what had happened in the early and middle part of the nineteenth century as a result of disputes with its neighbor to the north, was moderate and restrained in protests, fearing, no doubt, if the situation became too tense it might lose Baja California as it had lost other big blocks of territory on earlier occasions.

Pioneer ranchers organized a number of mutual water companies to manage the distribution of water to the various parts of the Valley. Each landowner was assessed for water stock which permitted him to receive water for a specified piece of land. Funds from this source paid costs of construction of the many miles of canals and the structures of the rancher-owned distribution system. In addition to being obligated to buy water stock, ranchers were expected to pay for water in the amounts used. These funds went to the Development Company as payment for delivering the water.

Things had gone quite well during most of 1903, but toward the end of the year the river became lower and lower until water would no longer flow through the Chaffey Gate into the main canal.

According to Rockwood, this wooden gate, installed in the summer of 1901 by Engineer-Financier George Chaffey, was intended as a temporary expedient to control water entering the canal. Rockwood admitted the gate was sturdily built and would have served for years had it not been placed five feet too high in the river bank. It seems strange that a man with Chaffey’s record of accomplishments in the field of irrigation would have made such an error. One explanation is that the gate was set properly, but because the river was high at the time, the engineer in charge placed several removable flash boards in the bottom of the gate. The bottom of the canal sanded up to the level of the flash boards, and Rockwood may have mistaken the top of the boards for the bottom of the gate.

An improvised and perhaps illegal scheme caused the river to rise enough to enter the canal on the first occasion, but the process could not be repeated. Consequently another, but dangerous, means was resorted to at subsequent times of low water. As explained by Rockwood, “due to the fact that the floor of the gate was left above grade, we found it necessary, in the Falls of 1902, 1903 and 1904 to cut a by-pass around the gate to the river, and it was through this by-pass then, during these three years, that water was obtained at low water for the irrigation of the Valley.”

On the first occasion of low water the ranchers’ crops suffered. When they could see the flow of water in the river headed for the Gulf of California, explanations for the empty canals made no sense. The Development Company was under great presure and law suits were threatened by numerous ranchers charging the Company had failed to live up to the terms of the agreement to supply water. Rockwood was a sincere man and must have understood and shared their concern.

With the main canal carrying no water, and the situation in the Valley becoming more tense each day, Rockwood left his post at the river and went to Imperial to discuss the problem with Anthony H. Heber, President of the Company. Heber, concerned because the ranchers threatened lawsuits against the Company for not having delivered water as agreed, ordered the engineer to return to the river and dynamite the bottom of the gate. Back at the canal heading, Rockwood was reluctant to take the hazardous step of breaking out the bottom of the gate, and instead of following orders approached the problem in another way.

Sending helpers out to all the Indian villages in the area to recruit workers, he soon had large crews busy cutting arrow weeds and tying them tightly with wire into arm load size bundles. Arrow weeds, straight stemmed woody plants, grow profusely near streams where their roots can reach moist soil. Their stems are from one-half inch to an inch in diameter and they grow to a height of four to eight feet at maturity. Small branches, bushy and leaf covered extend over the upper third of the plant. They had long been used by the Indians for making arrow shafts and as building materials. Properly bundled and secured in place, mats made of arrow weeds are excellent for controlling erosion along the banks of streams.

After several days work by the Indian crews, there were thousands of bundles of arrow weeds stacked adjacent to the half-mile wide river. The water was shallow, probably no deeper than a few inches, when the workers were set at piling the bundles side by side on the sandy river bottom to build a brush dam across the stream just below and to the south of the canal inlet structure. The river silt, sliding and rolling along the bottom, deposited in and on the bundles, quickly anchoring them in place.

As the brush dam, or weir as it is more properly called, was extended across the wide river bed, a sandbar formed over the mass of brush and raised the water level. Within an hour it was high enough to flow into the canal and on its way to the thirsty fields of Imperial Valley. The day was saved. The farmers’ crops had been given another lease on life. However, Rockwood must have realized the brush weir was but a temporary measure, and would wash away with the first small flood.

The next day Heber visited the river and was at first delighted to see water again flowing down the main canal. However, he was furious when he learned how it had been accomplished and that Rockwood had failed to carry out his orders to dynamite the bottom of the gate. He demanded the gate be blasted as ordered and the brush weir removed as well, pointing out that placing it was a violation of Federal Law, the Colorado river having been classified as a navigable stream. When the engineer refused to do either, relations between the two became very strained.

Water continued to flow down the Alamo canal to the new ranches, and while all was not serene, pressure on the Development Company lessened to some extent. Fortunately the Federal Government did not choose to make an issue of the brush weir at that time.

An ever-present source of discord which divided the thinking of the early settlers was the effort being made by the Federal Government to take control of the California Development Company and the land of Imperial Valley and combine it with Yuma to make a large reclamation project of both areas. Many were in favor of this, but an equal number, who treasured their independence and freedom from governmental domination and management, opposed it violently.

The United States Bureau of Reclamation made a point of informing the farmers in Imperial Valley that the California Development Company was inadequately financed, lacked funds for any emergency, and it was questionable if the company was qualified to continue. The adverse publicity weakened confidence and hurt the credit of individual farm owners as well as the Development Company.

Officers of the Development Company started negotiations with the Government of Mexico for permission to construct an inlet from the river south of the border. It was reasoned there would be numerous advantages to such an intake, not the least of which would be the elimination of the constant threat of the Bureau of Reclamation to take over the project.

The level of sand in the first four miles of the canal was building up at an alarming rate, and it was obvious that unless something was done this first section would soon be useless and would carry water only during times of high water in the river. Rockwood watched the situation ceaselessly and tried several schemes to make the sand move down the canal. None of the measures was effective until, almost in desperation, he made a wide cut in the river bank at one end of the Chaffey Gate. He reasoned that a great volume of water rushing through the canal could not fail to sweep the sand along with it and out onto the Mexican desert a few miles downstream through a cut he made in the canal. This scheme failed and when the cut had been closed, it was found the level of sand on the canal bottom was even higher.

Rockwood must have realized the only answer to the sanding of the canal was the use of one or more large floating dredges. At the same time he surely was aware that his Company had neither the funds nor the credit to acquire this expensive equipment. It must have been a discouraging time for Rockwood. It had been his vision to change the barren and forbidding Salton Sink into the greatest irrigation project the world had seen. He had spent more than twelve years on the venture and was wagering his professional and financial future on its success. Before the turn of the century he had journeyed to most of the financial centers of the world seeking funds to carry on the work. He made many trips to Washington and to Mexico City to confer with officials of both Governments about the Colorado river and the complex problems of diverting water from the river in the United States, running it into Mexico through a canal over fifty miles in length, then bringing it back into the United States. He had wrangled with both sets of officials over navigation, treaties and rights of way. With Government harrassment added to the problem of a sanded canal, it must have seemed almost too much to bear.

Perhaps it was on this account, when word came that the Mexican Government was ready to sign an agreement for an intake in Mexico, Rockwood and his Development Company acted too quickly. Mexican officials, being aware of the sanded canal and the urgency the situation created, were in a strong bargaining position and dictated terms accordingly. Provisions regarding supplying water to Mexican lands were far more stringent than the earlier agreement with Andrade. Water rates were set and the Company was obligated to supply half of all water diverted when enough Mexican land was developed to a point when it was needed. At the time there appeared to be no reason to expect extensive development on the Mexican delta. However, allocation of water for Mexican lands grew into a complex problem not fully resolved even as the last quarter of the century approaches.

The agreement with Mexico permitting the Mexican intake effectively restricted the actions of the California Development Company. It was held responsible for all flood control measures, but any levee or control structure that was to be built required prior approval by Mexican Government engineers. Mexico was not obligated to pay expenses of any necessary work. It was difficult to comply with the agreement. The stringency of the contract perhaps entitles Mexico to a share of blame for the disaster that occurred after Rockwood diverted water from the river in Mexico.

Upon receiving a telegram from Mexico City authorizing the river cut, Heber ordered the Chief Engineer to do so immediately after first excavating a channel some sixty feet wide and three thousand feet long to join the Alamo canal below the sanded section. Rockwood was further directed to design a control gate for the inlet and submit the plans to engineers in the Mexican Capitol, but not to install it until after it had been approved.

An open river cut was against Rockwood’s engineering judgement, but direct orders from Heber and the ever increasing demands of ranchers that he restore deliveries of water were difficult pressures to resist.

The channel joining the new inlet and the Alamo canal was completed within three weeks, during which time Rockwood spent many hours in Yuma studying flow records of the Colorado and Gila rivers. What he learned was encouraging in that records of near thirty years recorded several winters with no floods, and never a winter during which more than one flood came down the Gila. With this data he felt a little more secure, and proceeded to make the cut, installing safeguards where possible with the funds and material available.

No one foresaw it at the time, but when Rockwood’s men made the cut in Mexico and a tiny trickle of water began to flow down the Alamo canal, at that instant a no-quarter war began between man and river that raged for many years. It was a vicious struggle, and like all wars, left a trail of wreckage and heartbreak in its wake. Two great scars were gouged across the face of Imperial Valley, and Valley settlers watched helplessly as years of effort and wonderful hopes for the future crumbled. The struggle brought bankruptcy to the California Development Company and put the vast resources of the Southern Pacific Railroad to a severe test. The cost of controlling the river was so great that it was many years and a generation later before people of the Valley could consider the war won.

There were two or three minor skirmishes followed by several major battles, each more extensive that the preceding, and all except the one man thinks of as the last, were won by the river. Six times the river defeated man’s best engineers and his machinery, but in the seventh battle the river was beaten, and since that day has been double locked behind great concrete dams. The Gila too, has been made a prisoner, and no longer disgorges vast floods of mud and debris into the lower Colorado.


The year 1905 was unusual in that there were several winter floods. The first heavy flood came in early February but surprisingly did not widen the Mexican inlet a significant amount. On the contrary, when the high water had passed it was discovered that sand had been deposited in the inlet channel to such an extent it was necessary to dredge to restore the water flow to the Alamo canal.

Within a few days another heavy flood roared down the river with results similar to the first, leaving the inlet undamaged but filled with sand and it again became necessary to dredge to maintain service to the Valley ranchers.

In March, still another flood, confirming that it was an unusual season, swept down the Colorado. At that time of year the water level of the river was normally high enough to flow out the original diversion through the Chaffey Gate. Anticipating an even greater water flow with the summer flood season approaching, it was decided to close the Mexican cut for the time being and restore the Chaffey outlet gate to use.

To close the Mexican cut, the same procedure was employed that had served effectively in closing cuts beside the Chaffey Gate on earlier occasions. The sides of the cut were “rip-rapped” with bundles of arrow weeds secured in place by barbed wire and tightly stretched steel cables. Pilings were driven deeply into the mud bottom a few feet apart across the width of the cut. A little above the top of the river bank several strong steel cables were stretched tightly across the cut and loaded with several layers of arrow weed bundles woven together with barbed wire and steel cables to form a single heavy brush mat. Tons of heavy boulders were then piled on top of this bridge of arrow weeds. Dynamite charges were placed in such a way that when detonated both ends of all the cables would part at once and the mass would drop in the cut ahead of the pilings. It was a proven method which worked on several previous occasions and very likely would have again had not a fourth flood come down the river just before the closing device was finished. As it was the whole thing washed away within a matter of hours. With the threat of more flooding the engineers became increasingly alarmed and immediately started another dam.

It was discouraging and frightening when the fifth flood of the winter destroyed the second dam almost before it was started. It was evident that the Development Company and the ranchers faced a serious problem. Two attempts to close the cut had failed, and while as yet a relatively small part of the river’s waters were entering the Alamo canal, it was obvious the anticipated summer floods would bring added problems.

In mid-June, 1905, the river was high, although not at flood level. However, Rockwood considered the water too high for his men to continue trying to close the cut, which by that time the rushing water had widened to one hundred fifty feet. As the river started to drop, the banks of the cut began to cave in and disappear down the canal. Also, the banks of the by-pass started caving and washing away. It was extremely alarming, but nothing could be done except to watch. On August 9 the entire river abandoned its old course to the Gulf and began rushing down the Alamo canal to the Imperial Valley. Probably at no time had an engineer confronted a more difficult and challenging situation than Rockwood on that occasion. It was difficult because his Company was on the verge of bankruptcy when the river started pouring all its water into the Valley, and challenging because in the history of engineering no engineer had ever faced just such a problem. When the river abandoned its old course and made a 180 degree turn toward the Salton Sea, something new was added to the experience of man.

When the Colorado changed direction and started filling the Salton Sink, the great river was doubtless doing as it had many times in the distant past. Geologists report that this was the case when the silt burden of the river, through countless centuries, caused it to alternate between flowing directly into the Gulf and into a great, sometimes fresh water, lake to the north. Perhaps the time was approaching for it to make a switch to the north anyway, when man, in a more or less bumbling way, helped make it come about. As recently as a few hundred years ago it would not have mattered, but with man on the scene with his cities, his railroads, and the countless ambitious and eager people who wanted to farm the land that would be at the bottom of the lake, as well as other land that would be ruined for man’s use, it meant a great deal indeed. It was just too important and vital to man for him to let it happen. Its importance went far beyond the financial ruin and heartbreak of a few thousand pioneer ranchers. It went beyond the flooding of a salt plant in the bottom of the Salton Sink and the flooding and re-building of many miles of railroad track. The real loss, the community loss, would have been unimaginably great both to the United States and to Mexico. Damage to the geology of the areas of southeastern California, western Arizona and southern Nevada would have been beyond description had the river not been controlled. A great canyon, perhaps more than a hundred feet deep would have continued the cut-back across Imperial Valley to the Colorado itself, and on back up that great river past the present towns of Blythe and to the lower reaches of the Grand Canyon. The city of Yuma and the irrigation project there would have washed away as would smaller projects farther north. Bridges, dams and other improvements would have been destroyed and Southern California would not have developed as it has. It is fortunate that men won the battle.

Repeated failures to close the Mexican cut had shaken the already weak financial structure of the California Development Company. Banks refused to advance needed funds under any circumstances. Only a small amount of financial assistance came to the Company from the mutual water companies.

It is ironic that the same people who had schemed to avoid the threat of intervention by the Bureau of Reclamation now appealed to the Federal Government for help. The appeal was quickly rejected because the break was in Mexico.

The Southern Pacific Railroad was more than an interested bystander since the company had more at stake than any other organization or individual. As the Salton Sea started rising it was necessary to move many miles of track to a higher contour. At the time of the first move it was thought the river would be quickly thrown back into the old channel to the Gulf. But it was not to be. The grade in use by the railroad at the present represents the fifth time the track was moved, and still another grade, at a higher elevation was made ready and can be seen today.

Heber and Rockwood went together to San Francisco to lay the river control problem before E. H. Harriman, President of the Southern Pacific Railroad. It had been reported the Railroad was considering filing suit against the California Development Company for flooding their tracks crossing the Salton Sink. If that was the case, Harriman may or may not have been surprised when he was asked for a loan to turn the river. However, because he was not satisfied with the management of the Development Company, the loan was refused. A short time later, following a reorganization of the Development Company and the naming of Railroad Company men as officers, with other conditions favorable to the Railroad, a $200,000 loan was granted.

To secure the loan Harriman required that 6,300 shares of the capital stock of the Development Company be held by a trustee named by the Railroad and that the Southern Pacific be given full management of the operations. At the annual meeting in May, 1905, Epes Randolph, Assistant to the President of the Southern Pacific, was named President of the Development Company and all the other offices of the Company were filled by railroad men. It was a full and complete Southern Pacific take over.

The main offices of the Development Company were moved from Imperial to Calexico and C. R. Rockwood was appointed Assistant General Manager in full charge.

At Harriman’s request, Randolph made a visit to the river break to appraise the situation and make an estimate of the probable cost of turning the river. Impressed with the magnitude of the task, he immediately wired Harriman that it would take much more than the $200,000 already agreed to. When asked for an estimate of what it would cost, Randolph replied that it would take $750,000. When pressed for his assurance that it could actually be done even for three quarters of a million dollars, Randolph told the Southern Pacific president he was sure it could be done for that amount. Harriman then ordered him to proceed.

Much as a military commander calls his generals together to plan a campaign of battle, Harriman called his engineers together in the summer of 1905, to form a plan to force the river back into its old channel.

According to H.T. Cory, who was named Chief Engineer a year and a half later, there were numerous plans suggested, many having no merit whatever, and only four worthy of consideration. They were, the Laguna Weir Plan; the Concrete Headgate Plan; the Rockwood Headgate Plan; and the Barrier Dam Plan.

The Laguna Weir Plan bears special mention. It was advanced by the Reclamation Bureau which was then building a diversion structure nine miles upstream called Laguna Dam which was to divert water for the Yuma Project. This plan probably would have succeeded in eventually turning the river back into the Gulf, however, it also meant the river would continue to flow into Imperial Valley for three more years causing the Salton Sea to continue to rise and adding to the problems the new lake already had created. Practically all of the irrigation systems of the Valley would have been destroyed, bringing financial ruin to the area. In short, had this plan been followed, the battle would have been won, but the war may have been lost.

The Laguna Plan proposed that all attempts at closing the break and turning the river be abandoned until after the Laguna Dam was completed and a large canal, with capacity to carry the entire river, be excavated from the Laguna Dam to connect with the Alamo canal just below the Chaffey Gate, a distance of nine miles. The entire river would then be sent down the Alamo canal into Imperial Valley and the Salton Sea. With the lower section of the river carrying no water it was believed that building dams across the break, clearing the old channel and adding to the levees would be relatively simple. However, after considering all the implications, Harriman and his engineers rejected the Laguna Weir Plan.

The Rockwood Headgate plan called for the installation of a wooden flash-board type headgate in the riverbank above the Mexican break. The gate was to be sufficiently large to carry the entire river while the break was being closed during the expected low water period.

The Barrier Dam Plan, advanced by Railroad Company Engineer F.S. Edinger appeared as perhaps the quickest and by far the cheapest possible method of bring the river under control long enough to close the break. The plan consisted of building a long dam of driven pilings and brush mats across the channel above the break forcing the river to flow along the east side of the long sandbar that had formed opposite the Mexican break. The sandbar was referred to as “disaster island” by those striving to control the river. Edinger’s crews worked on the dam from early October, 1905, and succeeded in forcing a substantial flow to follow the east channel to the Gulf. However, on November 29 a heavy flash flood of 110,000 second feet came roaring down the river. Within a matter of hours every trace of the dam disappeared and the entire river was again rushing into Imperial Valley.

Because the Chaffey Gate had to be replaced by a more satisfactory structure regardless of the method used to close the break, engineer James D. Schuyler was directed to design a suitable concrete structure to be installed in its place. The gate was constructed by Carl Leonardt, who contracted to install it for $55,221.08. This gate was used until 1917, when it was replaced with one of a different design which, it was thought would admit less sand to the main canal.

With the new concrete gate started, it was decided to proceed at the same time with installing the Rockwood Gate. It was Rockwood’s plan to place this wooden structure in undisturbed ground adjacent to the river a short distance upstream from the Mexican break. However, when initial excavation revealed that soil conditions at that site were unsuitable, the plan was temporarily abandoned. An alternative devised a short time later involved dredging a new intake channel a short distance farther south into which the river was diverted. With the river flowing through the new cut, construction of the wooden headgate began in the bottom of the old break. Coffer dams were built above and below the site, the water was pumped out, excavation began and pilings were driven.

Rockwood supervised construction on the new gate until late in 1905 when he asked to be relieved of work at the river so he could devote full time to the business affairs of the company. Edinger was placed in charge of the remaining construction. In April, 1906, Rockwood resigned as Chief Engineer, but retained a position as Consulting Engineer until October 1, 1906, when he severed his connection with the company.

When Rockwood stepped down as Chief Engineer, H.T. Cory was named to fill that position and immediately took charge. At that time Randolph and Cory made perhaps the most far reaching decision yet regarding turning the river. They decided to build a Railroad from the mainline at Hanlon siding to the river break. Before the branch line was in, all materials used in the several attempts to check the break had been transferred from rail cars to barges and floated down the river to the break. It was dangerous, cumbersome and very expensive. Construction of the branch line started July 1, 1906, and the first trainload of material passed over it on August 15.

In mid-April, 1906, the river rose daily, indicating the summer flood season had started. Consequently, it was decided to postpone all efforts to divert the river and fill the break until after the flood season passed and the river was at its lowest.

The 1906 summer season was one of general high water. During March the stream carried from 5,000 to 7,000 second feet, and there were two flash floods during the month when the flow exceeded 70,000 second feet. In June 100,000 second feet flowed into the Valley. After each of these flash floods it was apparent that the width of the break had increased, and that the magnitude of the task of turning the river was becoming greater and greater.

The Rockwood Headgate was completed on April 18, 1906, at a cost of $122,500 in addition to the costs of the new channel and other necessary excavations. When the $100,000 fruitlessly spent on Edinger’s Barrier Dam and the $80,000 for a new dredge are added, it is obvious that a great deal of money had been spent, but the river was still out of control. It seems remarkable that at this point Harriman agreed to advance an additional $250,000.

The time spent waiting for low water was not wasted. Men, material and machinery, were marshalled for an all-out effort when conditions were right. Maintaining a dependable labor force was a critical problem. A constant stream of laborers came to work for a few days and then moved on because of the extreme summer heat. At the peak of employment there were in excess of 1,000 men working on the job. An arrangement that helped solve the labor problem was worked out with the Federal Government and with Mexico. The men, women, and children of six Indian Tribes were moved to the area. The Pimas, Papagos, Maricopas and Yumas from Arizona, and the Cocopahs and Dieguenos from Mexico were established in a new village of over 2,000 persons where they lived together in harmony, and the men worked together without friction. The Indian encampment provided about 400 dependable workmen, and as stated by Cory, “Indian Labor was very satisfactory, and, indeed, just what other arrangement could have been made is problematical. Under intelligent foremen who understood their peculiarities, chief of which is lack of assurance and consequent timidity in going ahead with work, they are quite satisfactory.”

A rock quarry was developed at Pilot Knob and a clay pit was opened close by. A huge gravel deposit owned by the Southern Pacific Company at Mammoth, about fifty miles to the west, was made ready for shipping gravel. In addition, spur tracks were constructed near Pilot Knob to serve as a marshalling yard for material as it reached the area. Large drums of one-half inch and three-eighths inch steel cable arrived at the site along with hundreds of coils of wire to be used in weaving brush mats. A barge, the Silas J. Lewis, over one hundred feet long and thirty five feet wide, was decked over and equipped like a giant loom for weaving a continuous brush mat approximately two feet thick and one hundred feet wide, interlaced and bound together with the steel cable and wire. Several heavy duty pile drivers were brought to the job from distant points of the Southern Pacific Railroad system. Also, and of utmost importance, experienced pile driver operators arrived with the machines, as did men skilled and experienced in building pile bent railroad trestles. Car after car loaded with 90 foot lengths of piling, and twelve inch by twelve inch cap timbers, and eight by sixteen inch stringers, arrived to wait on the rail siding ready for use.

Repeated flash floods during the Spring and Summer of 1906 had widened the river break to 2,700 feet and the entire flow rushed into Imperial Valley. With the wooden Rockwood Gate finished, it was planned to divert the river through the gate by means of a diversion dam across the river and then close the break.

The barge equipped to weave brush mats was positioned at the north bank and the crew proceeded to lay a double layer of the one hundred foot wide mats across the river. As the mats became saturated with the silt in the water they settled heavily to the bottom. At the same time trestle construction crews built a pile bent railroad bridge above the mat. Less than a month was required to place the mats and build the trestle, complete with rails and ready for use. With the bridge finished, a train of flat cars loaded with heavy boulders was pushed out on the trestle, and the boulders dumped into the water coming to rest on the brush mats on the river bottom. Trainload after trainload of rock was dumped until the water gradually started to rise. When the level was raised six feet, water started flowing through the Rockwood gate.

There was considerable concern about the Rockwood gate being able to carry the amount of water flowing in the river at the time, so it was watched carefully. By October 10 it was carrying in excess of 10,000 second feet and there appeared to be some water going under it. In an attempt to prevent the gate from being washed out, a railroad trestle was hurriedly built just ahead of the structure so that rock could be dumped to raise the water level instead of depending on the flash boards which the design called for. However, when the first cars of rock were pushed out on the bridge, three of the bents settled and the train was wrecked. It was fortunate that no one was seriously injured. Frantic measures were taken to save the gate, but at 2:30 p.m. on October 11 it buckled in the center section and went out with a crashing roar.

Not only did the Rockwood gate fail, but the resulting rush of water and the debris from the gate destroyed a section of the railroad bridge across the channel one hundred yards downstream. A quick thinking engineer moved an engine and train off the trestle with seconds to spare.

Loss of the Rockwood gate was a serious setback. The gate had been a key factor in the plan to control the river, and the engineers realized they would have to begin again. Studying the now high and dry diversion dam, they came to the conclusion that dams could be successfully built without the use of brush mats, provided rock and gravel was dumped rapidly enough.

The sixth attempt at turning the river back into its old channel to the gulf was made by building three dams in series. The first, under engineer T.C. Hinds was to be a diversion dam below the new concrete headgate extending far enough to direct the flow into the old channel to the east. After repairing and strengthening the damaged trestles across the channel both above and below where the Rockwood gate had been built, rock was dumped, creating a second dam. A third trestle was built across the old break to the south and all made ready to dump rock for the third dam. By October 29, after many thousand of tons of rock and gravel had been dumped, these dams raised the water level enough that a flow started down the old channel. By November 4 the entire river flow of 9,200 second feet was following the old channel and blocked from flowing into the Salton Sea.

The break was closed at last, but the disturbing thing at the moment was that very little water was passing through the concrete gate and down the canal to meet the irrigation needs of Imperial Valley. However, much blasting and some dredging in the first section of the canal created a flow of 300 second feet which prevented any serious hardships in the Valley.

Previous to the sixth attempt to close the break, there had been practically no water in the old channel to the gulf since the river had turned north in August, 1905. During that time the old river bed had grown up with willows which would restrict the water flow and raise the water level higher than it had been. In addition, the passage of two hot summers with no water in the lower river caused the adobe soil of the river bank to dry and crack. To avoid any problems that might be created by the growth of brush in the old channel and to give the river a good start down its old course, prior to the completion of the rock dams, crews of Indian workers cleared the brush from a strip over one hundred feet wide and several miles in length, and mule teams pulling fresno scrapers scooped out a shallow channel the same distance.

As a secondary precaution, it was determined that levees would be built adjacent to the old river channel extending for several miles south of the dams. Ordinarily, the first step in levee construction is the excavation of what levee engineers call a muck ditch, a relatively narrow ditch beneath the surface to be occupied by the levee itself. This ditch is to make sure there are no cracks, crevasses, or voids, in the soil beneath the levee. The muck ditch is deepened until examination shows it to be well below all fractures and cracks in the soil. The ditch is then back filled and the levee built above it. Unfortunately, for some reason, digging a muck ditch, the accepted and proven method of water proofing levees, was not done in this instance.

Cory and the other engineers must have experienced a deep feeling of satisfaction that the wild river had at last been controlled. There was but a moderate flow in the river when Cory assigned seventy-five men to constantly watch the new rock dams and the levees extending to the south. With this precaution he took the train to Calexico to look after the somewhat neglected business end of his position as General Manager of the Company.

On December 5, the men patrolling the river were surprised and shocked when a great flood came out of the Gila, rapidly raising the level of water against the rock dams and the new levees. Numerous leaks under the levees prompted the men to hastily summon Cory by telegraph. Cory and Hinds made a hurried trip by night to the river, where by the first morning light they saw three distinct breaks. There were so many leaks that there was nothing to do but watch as the rushing water rapidly widened the largest of the breaks. The banks started caving in and washing away, and within twenty-four hours the Colorado was again flowing into the Salton Sea. It happened so quickly the crews south of the breaks were marooned, and when the steamer Searchlight was sent to rescue them, the little craft was grounded on the dry river bed.


The magnitude of the task of permanently controlling the river was emphasized to the Southern Pacific Company officials by the second break. After a conference between Harriman, Randolph and Cory, when an analysis was made of the financial aspects of the problem, the people of Imperial Valley were notified that the Railroad Company would not continue without a guarantee of reimbursement for sums already expended and other funds that would be needed to continue the effort.

Since little money could be raised locally, the problem was presented to President Theodore Roosevelt. The President took the position that because the break was in Mexico, the United States government could do nothing. It was then pointed out to him that the problem was indeed in Mexico, but the appeal was for help to save American farms and property in the United States. His response was little changed except to add that Congress was not in session. Finally, after days of haggling, President Roosevelt told Harriman to go ahead and control the river and he would see that the Southern Pacific was reimbursed for what it cost. The result was a telegram from Randolph to Cory ordering him to control the river at all cost.

By the time the Company had decided to continue, the situation at the river was worse than it had ever been. Within a short time the break had become a crevasse 1,100 feet wide and 40 feet deep through which the river rushed with a thundering roar. Flash floods of over 30,000 second feet again caused the cut back in New River which soon reached Volcano Lake. This frightened even the Reclamation Bureau because of the likelihood of losing their million dollar Laguna Dam if the flooding continued.


The situation at the river and the inevitable exodus of all the people from the Valley if the river was not controlled soon, became a matter of national concern. The disaster of the San Francisco earthquake became of secondary interest in the news reports of that day.

It was clear to the engineers and to the public that the next battle with the river would either bring victory or result in utter and complete defeat. With this in mind Cory used the authority extended by Harriman and followed his instructions to control the river at all cost.

Because the rock quarry at Pilot Knob was almost exhausted, all quarries accessible to the Southern Pacific or connecting rail systems within a radius of 500 miles were activated and equipment readied for loading rock. The Southern Pacific’s fleet of 300 “battleships,” the steel-sided dump cars, each capable of carrying over fifty tons of material, were all brought to the site. Train after train of flat cars, loaded at distant quarries with large rock, boulders too big to be handled in the “battleships,” rolled into the siding at Pilot Knob. East into New Mexico and north as far as San Francisco, trains of rock were given the right-of-way even over passenger trains. Night and day, trainloads of rock and gravel and clay rolled in until hundreds of cars stood on the spur tracks waiting for use.

Trestle crews built two heavy pile bent bridges across the crevasse fifty feet apart. Two pile drivers started work at each side of the crevasse and another, mounted on a barge, started work in the center of the channel, anchored in place by long steel cables. Working in the swift current of midchannel was very dangerous, as there was the constant hazard of the barge overturning when positioning the 90 foot pilings.

Several boats were manned and kept in constant readiness downstream to rescue any workmen who fell into the rushing water. There were numerous accidents, but luckily few fatalities. There were three near-disasters when flash floods bearing great quantities of floating debris tore out wide sections of nearly completed bridges. When this happened, even the success of getting the bridges in place seemed in doubt, but the work went on day and night. Cory seemed to be everywhere and never appeared to sleep. He was constantly walking, patrolling and watching over the job at all hours, relaying orders to the workmen through his assistants, Clark and Hinds.

Credit for completing the bridges is due to the experienced bridge foremen and workers who had built many hundred similar bridges over the vast Southern Pacific system. The bridges were finished and the last rail spiked in place at five p.m. on the evening of January 27, 1907. The workmen stood back and cheered as the first train of rock was eased out on one of the bridges. By dawn next morning 145 carloads of rock had been dumped into the crevasse. From the time the first car was dumped until the end of the job the work hardly stopped. Crew replaced crew at the end of a shift without break, and great searchlights illuminated the weird scene at night.

It was necessary to lift the bottom of the Colorado eleven feet just to get it back to the level of the old river bed. Trainload after trainload of rock seemed to disappear into nothingness after being dumped. Gradually the rock dams began to take shape below the trestles and trains of “battleships” dumped their great cargoes of gravel and clay to fill the voids between the boulders. More trains of rock, carefully spaced across the whole width of the crevasse were followed by still more trains of gravel and clay.

Gradually the dams rose and on February 11, 1907, water again started flowing down the old channel to the Gulf. The space between the two rock dams was filled with gravel brought from the deposit at Mammoth Wash, and clay came from the pit near Pilot Knob and other deposits. As the section was filled the material was washed into place and consolidated by workmen directing high pressure streams of water. When the fill was completed to the top of the trestles, the rails, ties and stringers on one bridge were removed and salvaged, but those on the other were left in place.

Thirty miles of levees were constructed, this time with muck ditches under them and a railroad on top. Train loads of rock stood waiting to be moved to where trouble might start. Thousands of sand bags were waiting to be used at the first sign of a leak in a levee. The river could not be trusted for a minute. A telephone line was built along the levee with a call station each mile.

The battle was over, but for many years no one could be sure it would not break out again. The years of peace that followed were uneasy ones in many ways.

Two years later, in 1909, the Colorado abruptly changed its course several miles below the levees and started flowing into Volcano Lake on its way to the Gulf. Millions of tons of silt were dumped in the old lake which careful checking revealed was building up at the rate of one foot per year. It would have been a matter of a short time until the river would again turn toward Imperial Valley, so steps were taken to prevent a recurrence of that tragedy. A cut was made turning the river into a low area called the Pescadero Basin to the southeast and in turn along the Sonora Mesa to the east.

More levees were required to cope with periods of high water. However, there were periods when the river became so low it was necessary to resort to brush weirs installed below the intake to get even a small part of the water needed in the Valley. There were three years when the river was completely dry for several weeks at a time.

It might be said that the Colorado never gave up, but kept fighting until Hoover Dam stabilized its flow, and the All American canal simplified delivering water to American farms in Imperial Valley.

The Southern Pacific Company put in a claim to the United States Government for $3,113,677 for turning the river back into the Gulf. Twenty-two years later, in 1930, the company was reimbursed with a check for $1,012,665.

1 Acre (Unit of Area)…………… 43560 Square Feet
1 Acre-foot (Unit of Volume)…… 43560 Cubic Feet
1 Cubic-foot………………….. 7.48 U.S. Gallons
1 Second-foot (Unit of Flow)… 1 Cubic foot per second
    For l Minute……………………. 60 Cubic Feet
        1 Hour…………………………. 3,600 Cubic Feet
        1 Day (24 hours)……………. 86,400 Cubic Feet
(646,272 U.S. Gallons)



Cory, Harry T., Imperial Valley and the Salton Sink. (San Francisco: Dodd Mead & Co., 1915).

Farr, F.C., History of Imperial Valley. (Berkeley: Elms and Franks, 1918).

Henderson, Tracy, Imperial Valley. (San Diego: Neyenesch Printers, 1968).

Hendricks, William O. “Developing San Diego’s Desert Empire.” Journal of San Diego History, Vol. XVII, No. 3, Summer, 1971.

Hundley, Norris, “The Politics of Reclamation: California, the Federal Government, and the Origin of the Boulder Canyon Act-A Second Look,” California Historical Quarterly, Vol. LII, No. 4, Winter, 1973.

McKelvey, Nat, “The Indomitable Epes Randolph,” Trains Magazine, July, 1950.

Tout, Otis B., The First Thirty Years—1901-1931: History of Imperial Valley, Southern California, U.S.A. (San Diego: Otis B. Tout, 1932).

Wood, Madelyn, “The Battle to Save Imperial Valley,” Cornet Magazine, May, 1952.

Woodbury, David O., Colorado Conquest. (New York: Dodd Mead & Co., 1941).

Historic Salton Sea and Imperial Irrigation District, El Centro, Imperial Irrigation District (c 1965).

“How the Southern Pacific Saved Imperial Valley,” Southern Pacific Press Release, January, 1950.

Robert L. Sperry owned and operated an extensive ranching and agribusiness in Imperial Valley for many years with headquarters near Calipatria. His engineering innovations and mechanical improvements of farming equipment contributed greatly to the development of the Imperial Valley.

Mr. Sperry served on the Calipatria School Boards for 15 years and was President of these bodies when the North End School District was absorbed to form the first Unified School District permitted under a new state law. He served as Chairman of the Calipatria Planning Commission and President of the Calipatria Community Church Board of Trustees. In 1954 he was elected to the Pioneers Memorial Hospital District Board of Directors serving for 9 years, 5 years as President, and he was an officer in the State Association of District Hospitals.

Mr. and Mrs. Sperry retired from active business and moved to Rancho Santa Fe in 1963 where they continue to be active in community affairs. Mr. Sperry is a member of numerous historical organizations and committees, and has written several articles published by the San Diego Corral of Westerners including “My Fifty Years in Imperial Valley,” Brand Book Number Three, San Diego Corral of Westerners (1973).

Mr. Sperry has a genuine interest in history and feels that the recording of local events is of utmost importance. As he expresses it, “local history restores the flesh and life to the bare bones of history!” Experienced in engineering and construction work, he feels a close kinship to the men who fought the raging Colorado river as described in his article published here.