Why Did The Dam Burst?

SHORTLY BEFORE MIDNIGHT ON MARCH 12, 1928, a carpenter named Ace Hopewell rode his sputtering motorcycle through the darkness of San Francisquito Canyon, north of Los Angeles. With his headlight sweeping over the rough road, he drove cautiously alongside San Francisquito Creek, past a sleeping settlement around a hydroelectric station, and ascended a grade up the canyon wall. In the darkness ahead a monolith loomed 200 feet above the canyon floor.

Just two years old, the St. Francis Dam had only recently reached full capacity. Stretching 700 feet between the canyon walls like an empty amphitheater, the concrete dam stored 12 billion gallons of Sierra Nevada snowmelt piped through the aqueduct that had transformed Los Angeles into a boomtown and turned deserts into blooming citrus groves. By now some workers and ranchers had begun saying that the dam might be unsafe. That morning yet more leaks had appeared on both abutments.

But on this night Hopewell saw nothing to cause alarm. The dam and reservoir slept peacefully in the dim moonlight. Hopewell passed the dam’s east abutment and continued along the edge of the reservoir. A mile up the road, he later testified, he heard a strange noise over the growl of his motorcycle. He dismounted and left the engine idling while he smoked a cigarette and listened to the rumbling behind him. Then he climbed back on the bike and continued on his way. He was the last person to see the St. Francis Dam intact and live to tell about it.

Just before midnight the dam collapsed. Twelve billion gallons of water roared along the twisting canyon, carrying concrete blocks weighing up to 10,000 tons a mile downstream. A mile and a half below the dam, a 110-foot wave engulfed a sleeping hamlet at a hydroelectric station. It razed a 65-foot concrete powerhouse. One man woke his wife in time to hustle her and their infant son through the window just before the flood carried him off along with the rest of their children; his wife clawed her way up the canyon wall with her baby. Another man awoke just before the wave slammed into his house. He grabbed for his wife but instead found himself somersaulting through a sea of mud, pounded by debris and slashed by tangles of barbed wire. Somehow he popped to the surface and hauled himself onto a floating rooftop. When the house crashed into a canyon wall, he leaped to dry land and called in vain for his wife and three children.

The flood thundered down the canyon behind a battering ram of houses, concrete, rocks, trees, and barbed wire. It hit Saugus, overflowed the dry Santa Clara River bed, and turned west. At Castaic Junction 19-year-old George McIntyre and his father stood outdoors, listening to the strange noises and watching the flashes in the sky from the destruction of power lines. Suddenly the flood wave knocked them off their feet. Mclntyre caught a glimpse of his little brother Billy scrambling out a window. Mclntyre and his father grabbed a utility pole. As the cataract carried them along, his father groaned, “Oh my God! I’m hurt. Goodbye.” Then he slipped under. Miles downstream George hauled himself into a cottonwood tree. He was Castaic’s only survivor.

The flood crashed into the work camp at Kemp. Tents with loose flaps turned into smothering shrouds of death. The men who had closed the flaps fared better; their tents floated to the surface like bubbles. Of the 150 men at the camp, 84 died. One World War I veteran said the carnage was worse than anything he’d seen in the Argonne.

The water raged into Ventura County and through Piru, Fillmore, Santa Paula, and Saticoy. It knocked down bridges, carried away houses and cars, and flooded thousands of acres of citrus groves. By the time the flood reached the Pacific Ocean, 50 miles downstream, more than 500 people were dead or missing. Survivors were rescued from roofs, trees, and telephone poles. Many of the dead were buried in silt or washed out to sea. Bodies and debris washed up on the beaches of San Diego 150 miles to the southeast.

The St. Francis Dam was the worst American civil engineering disaster of the twentieth century. The failure of one of the first generation of gravity arch dams caused a tremor in the industry and led to regulations that remain on the books to this day. It brought a tragic end to the previously enchanted career of William Mulholland, the chief architect of the water system that gave birth to modern Los Angeles.

But this greatest of civil engineering disasters also lingered as one of the most mysterious. A number of separate investigations looked into it, only to produce inconclusive and often contradictory findings. Most agreed that the design had lacked sufficient review by independent experts and that the dam had rested on a poor foundation. But they differed on the exact reasons for and sequence of the collapse.

Now modern analysis suggests that a variety of human and natural factors conspired in the disaster. After studying it for more than two decades, the geological engineer J. David Rogers has offered a revolutionary and controversial theory for the calamity.

The story of the St. Francis Dam begins nearly 50 years before the disaster, when William Mulholland arrived in Los Angeles. Born in Belfast, Ireland, in 1855, he went to sea when he was 14 and landed in New York five years later. After holding a few laboring jobs on the East Coast and at lumber camps in Michigan, he set out for California. Following an unsuccessful stint as a prospector, he finally settled in what was then a sleepy town that drew its water from the Los Angeles River and a series of primitive ditches, which he got work tending.

Mulholland lacked formal education, but he had a voracious appetite for books and a restless curiosity. While working on a well-drilling crew, he made a discovery that changed his life. “When we were down about six hundred feet, we struck a tree,” he recalled. “A little further we got fossil remains and these things fired my curiosity.” Mulholland went to the library and found a book on the geology of California. “Right there I decided to become an engineer,” he said.

A driven worker who barely took a vacation during his 50-year career, Mulholland quickly rose through the private Los Angeles City Water Company and in 1886 became its head. When the city acquired the company in 1902, he retained his job, in part because his prodigious memory and sloppy records made him indispensable. In his first three years as a public employee, he rebuilt an outdated water system, halved domestic water rates, and made a $640,000 profit. In the third year he announced a project that helped turn Los Angeles into a boomtown.

By the turn of the century, Los Angeles was nearing the limits of the local water supply. In 1904 Mulholland unveiled a plan to build an aqueduct running to the city from Owens Valley in the Sierras, nearly 250 miles across desert and mountain range. It would be the longest aqueduct in the world, more than four times the length of its grandest Roman predecessor. Los Angeles voters approved the issuance of $24.5 million in bonds to pay for the project—a staggering expenditure that could have put the city in receivership had it failed.

The antithesis of the office-bound bureaucrat, Mulholland tramped into the field to oversee nearly every aspect of the work. He relied on an almost intuitive sense of construction and grew impatient with written reports. During the building of the aqueduct, he once told a construction superintendent that he would rather get a single shovelful of muck out of a tunnel than acquire all the cost data. “Mulholland was above all a practical man, a man of action, who liked to get things done rather than merely to talk about doing them,” wrote his longtime friend J. B. Lippincott. “He seldom studied cost and progress reports. During the latter part of his life he was inclined to assume grave responsibilities alone.”

He finished the 233-mile aqueduct ahead of schedule and within budget. The first Sierra waters cascaded into the San Fernando Valley on November 5, 1913, in a ceremony attended by tens of thousands of people. “We have the fertile lands and the climate,” Mulholland told the crowd. “Only water was needed to make this region a rich and productive empire, and now we have it.” The immense success of the project elevated Mulholland to almost mythic stature. Admirers urged him to run for mayor. His response was typically salty and trenchant: “Gentlemen, I would rather give birth to a porcupine backwards than be mayor of Los Angeles.”

When the aqueduct began flowing, it delivered 258 million gallons per day, four times the city’s domestic use. But the water fueled a growth spurt that created demand for more. Between 1900 and 1920 the population of metropolitan Los Angeles quintupled. By 1920 Los Angeles exceeded 500,000 residents, surpassing San Francisco, and growth, drought, and troubles in the Owens Valley were forcing the city to increase its reservoir capacity. Owens Valley ranchers had turned openly hostile and begun sabotaging the aqueduct, staging armed takeovers of water facilities and kidnapping city officials. After building a series of smaller dams, Mulholland planned a much larger one to round out total reservoir capacity to a one-year supply. He settled on an area he had surveyed two decades before, in the rugged San Francisquito Canyon.

At the head of the canyon, the Los Angeles Aqueduct emerged from the mountains above the Powerhouse 1 hydroelectric plant. Below the plant the canyon opened southward into a broad, relatively flat plain before narrowing into a twisting gorge four miles downstream. During construction of the pipeline a labor camp had been located in the flat canyon, and Mulholland had had ample opportunity to survey the area, which seemed ideally suited to a dam. A steep formation of mica schist, a hard laminated rock that tends to cleave along parallel planes, ran down the steep east wall of the canyon, under the streambed, and partially up the west side. There it gave way to a formation of red conglomerate, a crumbly sandstone on the west side. Between the two rock formations lay a fault line, a common feature in Southern California, even at dam and reservoir sites. “You can scarcely find a square mile in this part of the country that is not faulty,” Mulholland said later. “It is very rumpled and twisted everywhere.” He had workers sink shafts and tunnels into the hillsides and run water percolation tests. The results convinced him that the narrow pass at the south end of the broad plain would make a suitable foundation for a dam.

St. Francis would be Los Angeles’s most ambitious reservoir project, accounting for about half the storage capacity around the city. Original plans called for a reservoir of more than 30,000 acre-feet and a 175-foot-tall concrete curved gravity dam built in a series of narrowing steps.

St. Francis was the city’s second concrete curved gravity dam. A gravity dam is one that holds back water by its dead weight. In contrast, an arch dam has insufficient mass to hold back the water and transfers the burden onto the abutments by arch action. A curved gravity dam (also known as a gravity arch dam) like St. Francis combines features of both types: Although its weight should be sufficient to hold back the water, its designers add the arch for extra strength.

Smaller arch dams had been built as early as Roman times and had been constructed in Europe for centuries, but the first arch dam in the United States had gone up only in 1884. Curved gravity dams began to appear in the United States in the late nineteenth century, such as the 170-foot San Mateo Dam, opened in 1889 as part of the San Francisco water system. Mulholland built 19 dams in his career, most of them earthen structures. He opened his first concrete curved gravity dam, later named the Mulholland Dam, in 1925 above Hollywood, and its design closely resembled that of the St. Francis. Although concrete offered the advantages of being less permeable and less subject to erosion than earthen structures, it had an unforgiving precondition: Masonry dams require hard, stable rock foundations.

The St. Francis Dam project was subject to little outside scrutiny. By now Mulholland’s judgment went almost unquestioned. One engineering journal later noted that “the reason there was no intervention from above and no effective protest from below found specific expression in the statement: ‘Mr. Mulholland was personally overseeing the work.’”

Donald C. Jackson, a leading bridge and dam historian, notes that Mulholland took advantage of a loophole in California’s 1917 dam safety law that exempted dams built by major municipalities from review by state authorities. He points out that by contrast the Littlerock Dam, built 50 miles away between 1922 and 1924, endured years of delays before getting approval from the state engineer. “Mulholland really separated himself from the rest of the hydraulic engineering community,” says Jackson, an associate professor at Lafayette College, in Easton, Pennsylvania. “We see that reflected in the key issue: Why is he the one who has sole authority over the dam?”

Construction crews began pouring concrete in August 1924. While that was under way, a long drought created pressure for greater reservoir capacity. Mulholland had raised the dam in the design phase by 10 feet; during construction he raised it another 10. When completed, the south-facing dam stood 195 feet above the canyon floor, augmented by a wing dike, a long wall allowing the reservoir to rise above the natural crest of the west ridge. Behind the dam sat a three-mile-long reservoir of 38,000 acre-feet. Although the dam had been raised by 20 feet, the evidence suggests the designers had failed to strengthen the structure. After the collapse city officials presented the governor’s inquiry board with designs showing a 176-foot-thick base. But the historian Charles Outland, whose 1963 book Man-Made Disaster remains a landmark work on the St. Francis Dam, later raised the possibility that the base was in fact 20 feet thinner, because four bottom steps were missing in construction photos. The record remains uncertain. Perhaps the bottom steps were added later onto the original monolith (a questionable move at best); perhaps the bureau misled investigators. If the steps were missing, that meant that the height of the dam had been raised by 20 feet without any thickening, a perilous move for a structure supposed to hold back water by dead weight.

The first aqueduct water flowed into the canyon on March 1, 1926, while the dam was still two months from completion. One month later the water reached the fault line between the schist and conglomerate and immediately began leaking through. The reservoir filled rapidly, rising 1.8 feet per day during the first three months. By the following year the surface of reservoir was within 3 feet of the spillway. During the filling, two sets of cracks had appeared across the face of the dam. Mulholland dismissed them as a natural result of concrete curing and ordered them sealed with oakum (twisted fibers impregnated with tar).

Over the next two years leaks bubbled forth intermittently. But Mulholland called them typical for so large a dam. “It was the driest dam of its size I ever saw in my life,” he said at the coroner’s inquest. All dams leak to one degree or another. Several investigators later concluded that some of the cracks had resulted not from curing of the concrete but from movement of the foundation. Still, the expectation of routine leakage may have led the water bureau to overlook more serious problems.

On March 12, 1928, the day of the disaster, the reservoir had stood three inches from the spillway for five days, and the city had stopped diverting aqueduct water into it. A brisk wind blew down the canyon. Waves splashed over the top of the dam and soaked the downstream steps. That morning the dam keeper, Tony Harnischfeger, called Mulholland to report muddy water leaking from the west abutment. Muddy water was a red flag suggesting erosion of the foundation. Mulholland and his assistant Harvey Van Norman hurried up from Los Angeles to inspect the dam. They were relieved to discover clear water leaking from the west side and picking up sediment only as it cascaded down the canyon wall. Then they spotted another leak across the canyon where the east abutment joined the mountain. The engineers inspected this one, too, before pronouncing the dam safe and returning to Los Angeles.

That night a surge-chamber attendant sleeping above the canyon was jarred awake by what he assumed was an earthquake shaking the floors and rattling the windows. He clambered down the canyon wall in the darkness alongside the penstocks to Powerhouse 2, about a mile and a half downstream from the dam. Near the bottom he stopped. The pipes disappeared into a muddy river. The settlement—power plant, houses, and school—was gone.

Mulholland and his engineers were back at the site hours before sunrise. When the first rays of light filtered over the hilltops, they saw the full extent of the destruction. The dam Mulholland had walked on the day before was a pile of rubble. Both the east and west wings had been carried away. Only the center section remained standing, like a solitary tombstone. A photographer caught the chief’s image as he stood above the wreckage that morning, his eyes distant, his face blank with disbelief.

No eyewitnesses survived to testify about how the dam collapsed. The flood scoured the hillsides to the bedrock, carrying away vital evidence. Years later Outland observed that no two people had experienced the disaster in the same way, no two engineers had drawn the same conclusions, and no jury had agreed on criminal negligence. “The evidence in the St. Francis Dam story is not unlike the Bible; one can prove or disprove any point of view he wishes,” Outland wrote.

The disaster riveted a nation. The 1889 Johnstown flood in Pennsylvania had killed more than four times as many people, but that was an earthen dam, and concrete was supposed to be stronger. A few weeks afterward the Literary Digest wrote of the fear created by the first large masonrydam failure: “Here the highest embodiment of modern dam-building science crumbled in ruin, taking a total of hundreds of lives as the price of mistaken confidence in its strength… . Often bitter protest has been made against the erection of a dam above populous communities. In every instance engineering science answered the protest and gave assurance that the waters would be safely controlled. The destruction of the St. Francis Dam challenges that assertion.”

More than a dozen investigations probed the cause of the failure. Critics were quick to point out the dam’s dearth of important safety features: It had no contraction joints to reduce cracking caused by the curing of concrete; the city had neglected to seal the foundation with pressure grouting, a procedure in which cracks in the rock foundation are stopped up with cement; and the St. Francis also lacked a sizable cutoff wall beneath the ground to divert water seepage under the foundation. Although the dam had drainage under the main section, the flanks were without wells to relieve the buoying effects of water seepage in the foundation.

Most observers, blaming an inadequate foundation for the collapse, concluded that the failure had begun in the red conglomerate on the west side of the dam. After the floodwaters had stripped away the hillsides, the contact line between the schist and red conglomerate on the west abutment stood out in sharp relief, and many investigators immediately honied in on it. The blocks found farthest downstream were initially thought to have come from the west side. Experts also noted the tendency of the red conglomerate to swell when wet, and post-disaster surveys showed that the west wing dike had risen slightly. Moreover, the west side had been leaking a few hours before the collapse.

Gov. Clement C. Young appointed a commission, which produced a 79-page report within five days. It concluded that the dam had failed because of water percolation and erosion on the west abutment near the contact point between the schist and conglomerate. The commission believed that water then swirled toward the west side of the dam, causing erosion landslides on the east side. A panel of experts assembled by the Los Angeles district attorney reached similar conclusions. But most historians have dismissed the governor’s commission report as a superficial account undermined by hastiness and a lack of crucial evidence.

Another panel offered a very different opinion. Two engineers, C. E. Grunsky and E. L. Grunsky, and a Stanford geology professor emeritus, Bailey Willis, said that the failure had begun on the east side of the dam. Hired by Santa Clara River valley ranchers to report on the failure, they argued that the east abutment rested on an old slide that had begun to move when lubricated with water. They concluded that the cracks had not resulted from curing, as Mulholland had thought, but instead were evidence that the east mountain had begun to shift. They observed that the mica schist on the east side had been disrupted by previous movement and was bedded at a severe angle that nearly paralleled the slope of the canyon wall. The east abutment had slid downstream when water seeped into the cleavages, the main section of the dam tilted into the breach, and moments later the west side gave way. The report acknowledged the tendency of the west side to swell and concluded that it would have eventually failed even if the east side had held. In hindsight the investigation of the Grunskys and Willis explains much of the evidence left unaddressed by the government reports. But their study, commissioned by an obscure ranchers’ water agency, attracted less attention because it came out after the governor’s commission, the district attorney’s panel, and the coroner’s grand jury published their own conclusions.

The earlier attacks on the aqueduct in the Owens Valley led some to suspect that saboteurs had blown up the dam. Mulholland vaguely alluded to this possibility during the inquest but stopped short of making an outright charge without any real basis. Investigators reviewed seismic records from the California Institute of Technology and found no evidence of blasts or earthquakes the night of the disaster.

Many perplexing questions remained. A gauge attached to the still-standing section recorded a sharp drop in the reservoir level a few hours before the collapse. At first some investigators cited this as evidence that the dam was leaking before it failed. But the 3.6-inch drop recorded by the gauge would equal a discharge of 935,000 cubic feet of water. Such a huge outpouring would almost certainly have been noticed by the people at Power-house 2 downstream. Where did the water go?

On one point all observers agreed: The dam had not had sufficient review. The coroner’s-inquest jury concluded that “the construction and operation of a great dam should never be left to the sole judgment of one man, no matter how eminent … checking by independent experts will eliminate the effect of human error and insure safety.”

The jury at the coroner’s inquest found Mulholland and the city’s water bureau responsible but recommended no criminal penalty. Mulholland broke down and wept on the witness stand. “Don’t blame anybody else, you just fasten it on me,” he said. “If there is an error of human judgment, I was the human.”

His career shattered, Mulholland retired later that year as chief engineer of the Los Angeles Department of Water. During the coroner’s inquest he said, “The only ones I envy about this thing are the ones who are dead.” He remained haunted by the disaster until his own death in 1935.

Meanwhile, Los Angeles and the dam-building industry moved as quickly as they could to leave the St. Francis behind. With the Boulder Dam project pending before Congress, they were eager to reassure the public that concrete gravity dams were safe. The city quickly accepted responsibility and paid out death claims with little legal wrangling. Utilities kept records secret for decades afterward. “A past generation did an effective job of sweeping the dirt under the rug,” Outland wrote in the preface to his book, “and present personnel have little knowledge of, or desire to learn, what was thus put out of sight.”

In a few years the failure dwindled in significance as the St. Francis was eclipsed by more spectacular projects, such as the Boulder and Grand Coulee Dams. Gradually it became little more than a footnote in the annals of dam building in the western United States.

But it was a durable footnote. For decades the St. Francis Dam remained a matter of debate among historians and geologists. Its site drew pilgrimages from geology students, who would arrive by busloads every year. One of them, J. David Rogers, became captivated by the mystery. As a young man, he read Outland’s Man-Made Disaster . He went on to earn a Ph.D. in geology and geotechnical engineering from the University of California at Berkeley, serve as a naval intelligence officer, and start a consulting firm specializing in engineering failures.

Rogers enjoyed one advantage over his predecessors: newer science. He combined cutting-edge failure-analysis techniques with a historian’s dogged immersion in old records. He pored over transcripts of the coroner’s inquest and autopsy photos. He ran computer models of the collapse and tested a variety of failure scenarios. “The dam collapsed for a whole number of reasons,” Rogers says. “I’ve analyzed it now by six different modes of failure. It failed by all six.”

Rogers believes the dam was indeed built against an ancient landslide on the east wall of the canyon, as the ranchers’ investigation had suggested. In amassing evidence for a step-by-step reconstruction of the collapse, he has taken the theory a step further by identifying a landslide complex a mile and a half long on the eastern slope of the canyon, far greater than contemporary investigators—or certainly Mulholland—observed. This paleo-megalandslide would actually explain why Mulholland recognized the canyon as an excellent reservoir site. Rogers believes that in the Pleistocene epoch an ancient landslide choked off the canyon and created a lake 10 times the size of the St. Francis reservoir. He thinks this ancient dam also failed catastrophically, hundreds of thousands of years ago, unleashing a flood that dwarfed the 1928 one.

As the reservoir filled, water seeped into the eastern abutments and lubricated the schist. The slide began to wedge under the dam, forming the cracks that Mulholland blamed on curing. Meanwhile, water percolated underneath the foundation. The main section was equipped with drainage, but the abutments were not. Over time, as the water seeped under the foundation and into cracks, it buoyed the dam by a phenomenon called uplift.

Rogers’s theory holds that the moving mountain created another potentially fatal Achilles’ heel at the same time. The activity on the east abutment drastically increased pressure on the dam’s center section, tipping it downstream and forming a crack in the upstream heel. When water soaked into the fissure, uplift diminished its dead weight by 45 percent, tilting the main section slightly forward before the collapse, which would explain why the gauge recorded a drop in the reservoir level. The water did not drop; rather the main section of the dam, and the gauge with it, rose from the water. “That’s so bad it’s hard to even describe,” he says. “Any kind of crack in the upstream heel is big-time bad news.”

Would these clues have been noticed the night the dam burst? Would the tilting of the massive block of concrete have made noise? Some of the last people to pass the dam recalled seeing lights at its bottom. The body of Leona Johnson, the dam keeper’s common-law wife, was discovered fully clothed pinned under a block of concrete at the dam. Perhaps she—and presumably her husband, whose body was never found—had been worried enough to be investigating something in the middle of the night.

Rogers thinks the failure began on the lower side of the east abutment and precipitated a rapid chain reaction: The pressure of the reservoir forced water through the cracks with an aerosol effect that may explain why one survivor awoke to a thick mist just before the flood hit. The first block blew out on the lower east abutment. Water quickly eroded the foundation. A massive landslide of 500,000 cubic yards of schist slid into the breach, but the rushing water quickly carried away the 877,500 tons of earth. One witness described the initial flood as “liquid mud.” Turbid water has greater buoyancy than pure water, and Rogers calculates that the ooze would have reduced the effective weight of the concrete chunks to only 10 or 20 percent of their dry weight, floating them downstream like icebergs. Then, Rogers says, the central section of the dam shifted toward the gap, enlarging cracks on the west side. The reservoir had dropped about 85 feet by the time the weakened west side collapsed.

Rogers calculates that the flood reached a peak flow of 1.7 million cubic feet per second. Just over an hour later the first parties to reach the reservoir found it empty.

Rogers cites a compelling array of evidence to support the landslide theory: Some witnesses reported cracks in the road on the east canyon wall two days before the disaster, and one of the last people to travel past the dam told Outland that the road had dropped about 12 inches, evidence that the mountain was beginning to slip. Then the full force of the slide let loose. Power lines above the dam toppled. Debris washed up on the west shore of the reservoir 4 feet above the high-water mark, suggesting that a landslide splashed into a full lake. Victims at the Powerhouse 2 camp were found with lungs and stomachs clogged with silt, indicating they died in a flow of mud.

Although his diagnosis focuses on the east abutment slide, Rogers also has pinpointed 13 design deficiencies in the dam and shown that it could have collapsed by at least six different modes of analysis. For example, the dam had to fail by cantilever action (bending downstream) when water got within 7 feet of the top. Another analysis shows that arch loads would have become dangerously high once the water reached 10 feet below the crest.

Having spent much of his career analyzing the shortcomings of the dam, Rogers paints an unusually sympathetic portrait of its creator. He argues that Mulholland fares better viewed in the historical context of the 1920s, when dam building was undergoing rapid changes. Standards for technologies to counter uplift—cutoff walls, contraction joints, grout curtains, uplift relief wells, and thermal cooling mechanisms—were in embryonic stages at the time. Dam builders had yet to fully appreciate the concept of uplift, and the ancient landslide remained invisible not only to the majority of contemporary investigators but to later generations of geologists as well. After all, he says, more than 100 dams in the United States have been unknowingly built against ancient landslides.

“How much should you blame a guy for not appreciating geology at the dam site when thousands of geologists after him—including myself—went out there and didn’t recognize the landslide?” Rogers says. “I’m not nearly as hard on Bill Mulholland as I used to be.”

But historians like Jackson disagree. Although impressed with Rogers’s forensic work, Jackson doubts Mulholland can be exonerated. He cites discussion of uplift in published journals, most notably a lengthy article in the Transactions of the American Society of Civil Engineers in 1912, and says that cutoff walls, grout curtains, and relief wells were already in wide use by the 1920s. Indeed, contemporary critics were at least familiar enough with uplift to point out the dam’s lack of technology to reduce it. A man of Mulholland’s stature, Jackson says, should have been familiar with uplift and taken far more care to guard against it. “In this case Rogers is not correct,” says Jackson. “This was an issue that was very much a part of what dam engineers thought about well before St. Francis.”

On the larger issue of culpability, Jackson maintains that Mulholland must bear responsibility for the collapse. He cites a list of mistakes that courted disaster, such as raising the height of the dam without any appreciable thickening and failing to solicit independent consultants. He sees Mulholland as a man whose reach exceeded his grasp, an autocrat who separated himself from the hydraulic engineering community, lacked academic training, and avoided outside review in his haste to finish his dam. One of the primary lessons to emerge from the collapse, says Jackson, was that one engineer never again should bear so much responsibility.

“In retrospect the responsibility that was placed on William Mulholland was deserved,” Jackson says. “I don’t see any reason why history should be any kinder. I don’t think the weights on the scale have changed since then. It was harsh, but deservedly so.”

Although Mulholland became a lightning rod for the blame, the disaster caused reverberations throughout the dam-building industry. The dam safety act passed by California the following year placed all nonfederal dams under the regulatory authority of the state engineer. No longer would municipalities like Los Angeles escape outside review. In the final analysis the St. Francis Dam and its victims became martyrs for much more stringent safety standards.

Today the valley that once held the Sierra snowmelt has returned to its natural state. A small stream trickles down San Francisquito Canyon past huge blocks of crumbling concrete. A road winds up the twisting canyon past housesized chunks of rubble and straightens out in what once was the floor of a vast lake. Occasionally the dry earth yields up a watch or a child’s toy as a reminder of the disaster. Geologists and historians still make pilgrimages to climb over the slabs of concrete and peer up the steep walls of mica schist and sandstone. And this dry parchment of geology whispers the clues to the mystery of the St. Francis Dam.