by Jon Wilkman
Henry King, a mechanic from Saugus, ran the “grizzly” sieve for about a year and a half, classifying the gravel-and-sand aggregate. He refused to say for sure whether clay contaminated the mix, but he told the jury, “There was considerable amount of dirt in the gravel taken out of the main canyon below the dam site … it stands to reason that some got through.” Before he quit in September 1926, King watched as the concrete was poured against the conglomerate on the west abutment. “Looked to me like they poured the concrete against the rock, didn’t go into it,” he said.34
Frank LeBrun saw the same thing. LeBrun lived in Newhall and lost friends in the flood. His family had sold some of their San Francisquito Canyon ranchland to the city for the St. Francis Reservoir. During construction, he worked with a pick and shovel as a “pit man” and sometimes ran the steam shovel. When asked why he thought the dam collapsed, he answered that there were no “wings” cut into the abutment to secure the dam in place. He testified that the hillside was “hydrosluiced and then the[y] removed the top just a little bit and poured the concrete over it right next to the hill.”35
Henry Ruiz, who lived in the Powerhouse 1 employee village above the reservoir, survived the flood, but eight members of his family died. Ruiz drove kids to school in the canyon and often passed the dam site. He told the jury that both sides of the dam were leaking: “Leakage on the westerly side there, sort of got on my nerves for a while, leaking badly toward the last, and there was a leakage on the eastern side too. Made me feel uneasy.”36
Many of the construction workers seemed uncomfortable on the stand and qualified previous testimony. Jim Erratchuo was born in the canyon. He worked as a laborer on the dam doing “all kinds of work, what they told me to do.” Under oath, Erratchuo refused to elaborate on a story he’d told before about how he and Henry Ruiz had been shot at in the canyon, and a second time when they went to hide in Hollywood. He didn’t hesitate to say he saw concrete put directly against the hill. “[They] just washed the dirt off, just like putting your finger against a board.”37
William Hoke, the retired owner of a heavy-equipment company, erected the temporary steel structures used during construction of the St. Francis Dam and considered himself an amateur geologist. Familiar with the area since he was six years old, he had done some gold mining in San Francisquito Canyon a few years before, and even had a house there for a while. When Hoke was called to the witness stand, he was asked to identify samples of red conglomerate he brought with him. He borrowed Assistant District Attorney Dennison’s eyeglasses and confirmed the rocks were the ones he found below the west abutment. Hoke told the jury that the samples softened and became “brittle [and] awfully easy to break” when exposed to water.38 In a dramatic demonstration that became popular with the press, a piece of San Francisquito Canyon conglomerate was dropped into a glass of water. In a few minutes the rock emitted air bubbles as it dissolved into mush, surrounded by a bloodred slurry.39
Hoke was certain there wasn’t “any rock or formation in that canyon would support a dam below Powerhouse 1.” As early as April 1927, he’d seen landslides along the road about a mile above the dam. He also saw cracks with leaks between the dam and the west abutment. “There was water, not ebbing, but running out,” he told the jury. Despite this, he never expressed his concerns to William Mulholland or other DWP employees. “I am not a meddlesome person,” he explained.40
A fragment of red conglomerate from the west abutment beside a glass of water. Prosecutors showed that the rock dissolved when it was immersed in the water. (Author’s collection)
Cracks and leaks concerned dam keeper Tony Harnischfeger and many visitors to the St. Francis Dam, but they didn’t bother William Mulholland. He knew when concrete dries or cures, it forms shrinkage-contraction cracks. By 1928, other engineers knew this too. It was becoming increasingly common to relieve stresses as a concrete dam dried by including predetermined contraction joints during the construction process. Later they were filled with grout, ideally from the upstream side. The Chief preferred to let the cracks occur naturally and then stuff them with oakum; in the case of the St. Francis Dam, which was already partially filled, on the downstream face, like a sailor sealing the hull of a ship against the sea.
As the Coroner’s Inquest continued, much of the testimony was easy to understand by reporters and even newspaper readers, but a more obscure engineering term was mentioned repeatedly: “uplift,” or “up-thrust.”
The phenomenon of uplift had been acknowledged and discussed since the 1880s.41 It occurs when water is allowed to collect under the foundation of a dam. When the underlying geology is saturated, it swells and pushes out in all directions. This can lift and literally float a concrete dam. When the effective weight of such a structure is lessened by uplift, it becomes weaker and less secure. Between 1900 and 1915, American dam designers experimented with a number of design methods to counteract these dangers.42
The use of an arched design is one way to confront hydrostatic forces like uplift. From 1889 onward, a number of large arched American dams were built, beginning with Bear Valley Lake in Northern California. The most influential structures were constructed by the U.S. Bureau of Reclamation. Three dams mentioned during the Inquest were Roosevelt (1911) in Arizona, Arrowrock (1915) in Idaho, and Elephant Butte (1916) in Texas.
With no direct experience constructing concrete gravity dams, arched or not, William Mulholland and his DWP staff turned to existing examples as models and consulted engineering textbooks to validate their designs. The most consulted sources, The Design and Construction of Dams, by Edward Wegmann, and Charles Morrison and Orrin Brodie’s High Masonry Dam Design, suggested mathematical approaches to alleviating uplift, but also left final decisions to an engineer’s evaluation of the needs of the site.
In 1924, when the St. Francis Dam was built, the 1916 concrete gravity Elephant Butte Dam was considered state of the art in the application of dam-safety techniques. Located on the Rio Grande River near El Paso, the 301-foot-tall barrier was built by U.S. Bureau of Reclamation designers. It used an extensive system of internal drainage wells that extended into the dam’s bedrock foundation. To further protect against seepage, a “grout curtain” was employed, consisting of a parallel series of holes drilled into the upstream side of the dam and injected with cement-based filler, creating a fencelike moisture barrier. An especially innovative feature of Elephant Butte was a drainage “gallery” embedded into the body of the dam, allowing inspectors to directly observe seepage passing internal drains.43
Not all dams built in the 1920s employed the extensive uplift-relief features included in the Elephant Butte Dam, but by 1928 a consensus on “best practices” was growing. Independent and self-confident Bill Mulholland never consulted any of the engineers who worked on these projects, even those with offices in Los Angeles. To deal with uplift in the design of the St. Francis Dam, the Chief concentrated drainage “bleeders” in the center section of the structure, where he rightfully thought they were most needed. As for other safety features included in the Elephant Butte Dam, Mulholland apparently considered them unnecessary.
When District Attorney Keyes questioned the Chief’s decision to limit the drainage system, the Chief replied, “There was no water in the formation. That formation is dry as a bone.”
“We have the fact that the dam fell,” Keyes snapped.
Mulholland refused to be rattled. “Don’t imagine for a minute that I would throw you off the scent,” he assured the jury. “I am willing to take my medicine like a man. If there is anything I can say that will help you in your disclosures, I will be the very first to point it out if I see it first. I have nothing to conceal.”44
Mulholland remained steadfast in his conviction that the dam he built was safe. The Chief’s tough, some said arrogant self-assurance had been shaken, but he refused to acknowledge second thoughts. When Assistant D.A. Dennison bluntly confronted him with a statement that “the best minds in engineering” believed th
at the site of the St. Francis Dam was unsafe, Mulholland didn’t abandon his convictions. “I would not have built it if I thought that.”45 Dennison persisted. Knowing what happened, would he build a dam at the same site again? “Not in the same place,” the Chief finally admitted. Why? Mulholland hesitated. “It fell this time and there is a hoodoo on it. That would be enough for me.”
Hoodoo? The intellectual snickering among Mulholland’s critics and a generation of university-trained engineers was almost audible. What did the Chief mean by “hoodoo?” His answer was chilling. “It is vulnerable against human aggression, and I would not build it there.”46
A sympathetic juror suggested the seventy-two-year-old Chief might have been “letting down” in recent years and delegating more. Mulholland straightened in the witness chair. “I believe I have been working harder than I ever did in my life … the only time I have taken a vacation in fifty years [was] through the Panama Canal to New York. I am the first up in the morning, and the last to go to bed … There are a very few [who] beat me in the office in the morning … As far as letting up is concerned, I wish I could. I believe I will have to very shortly. This thing has got away with me,” he said, then broke down in tears. When Mulholland regained his composure, he looked squarely at the men in the jurors’ box. “If there is an error of human judgment, I was the human. I won’t try to fasten it on anybody else.”47
Frank Raggio Jr., whose father owned a venerable San Francisquito Canyon ranch inundated by the St. Francis flood, had ambitions to become a lawyer. Frank Sr. knew the Chief and arranged for his son to attend the Inquest. As the young man watched the old engineer confront tough questions on the witness stand and struggle with his emotions, he saw a culprit and a victim. “As he was building the dam he thought he was doing a great thing for the City of Los Angeles to store the water,” Raggio remembered nearly seventy years later. “They brought him up so high as Chief Engineer. Consultants from all over the world came to consult with him. They built up his ego … at the Inquest, Bill Mulholland was reduced to a very small man.”48
Whatever William Mulholland’s fate, as the Inquest continued and other engineering investigations neared completion, strands of conflicting evidence were tightening. However, a full explanation of the St. Francis Dam disaster remained incomplete, and who or what should be blamed had yet to be determined.
There was time for surprises.
Mulholland, third from left; Assistant D.A. Dennison, second from right; Coroner Nance, far right (Los Angeles Department of Water and Power)
11.
Rewinding Time
In the midst of a cacophony of engineering explanations and political and personal invective, investigators continued to gather evidence in the relative quiet of San Francisquito Canyon. Although the flood had washed away much of the terrain around the dam site, DWP Field Engineer Ralph R. Proctor and surveyor Harold Hemborg found bench marks used to align and locate the St. Francis Dam during construction. There also were U.S. Geological Survey markers a mile north of the emptied reservoir, and another downstream near the Harry Carey Ranch. The bench marks installed on April 23, 1926, along the crest of the main dam were especially useful, including one located on the west end of the wing dike and another four hundred feet downstream. A marker that survived on an upper step of the Tombstone was perhaps the most valuable of all.1
Using a geometric procedure called triangulation, surveyors measured known distances and angles between all of these markers and compared them to determine differences between positions before and after the failure. This could indicate whether the dam had moved from its original location. The process was repeated with different sets of benchmarks to home in on an answer. Investigators wanted to know: if the Tombstone had moved, could such a shift help explain how or why the structure failed?
While surveyors set up triangulation comparisons, high above, specially designed biplanes owned by the Spence and Fairchild aerial photo companies crisscrossed the canyon from the dam site to the ocean. During the 1920s, aviation pioneer Robert Spence was popular with the Los Angeles Chamber of Commerce, which was eager to use aerial photography to show off the City of the Angels’ impressive growth. On these flights there wasn’t much to be proud of.
To survey the floodpath from the air, the city hired Fairchild Aerial Surveys. Founder Sherman Fairchild had perfected an aerial camera that used a special shutter and custom-designed aircraft. Inside the Fairchild plane, thousands of feet in the air, a photographer looked straight down through a viewfinder and clicked pairs of overlapping images. To assure accuracy and precision, the pilot maintained a steady rate of speed and followed a precise back-and-forth pattern—like mowing a lawn from the sky.2 When studied side by side in a stereopticon viewer, examples of the twin images offered a 3-D view of the landscape below. Altogether, the aerial imagery covered the dam site and as far as a mile and a half downstream, where fragments were strewn along the canyon floor. The largest piece, found 1,500 feet from the Tombstone, was nearly sixty feet tall and weighed ten thousand tons. How it traveled so far was another mystery to be solved.
The largest downstream fragment of the dam (Ventura County Museum of History and Art)
A chart showing the location of fragments and their original position in the St. Francis Dam created for the California Governor’s Investigation Report (Author’s collection)
The original downstream surface of the St. Francis Dam was constructed with stepped ledges five feet high. Because the structure was curved and slanted, the width of each step was unique. Investigators could measure shapes, angles, and variations in the height and width of surviving steps and determine where a block originally fit in the downstream face of the dam.
To begin the process, each fragment was precisely located and identified with a number. State of California investigators and DWP surveyors labeled forty-four individual pieces or blocks. Exactly where each block was found, and whether one piece was over or under another, contributed to an emerging scenario. Pieces discovered farther downstream were likely to have been carried away during the first rush of the flood, when water volume and pressure were greatest. If one piece was on top of another, that could indicate it arrived after the fragment below.
Employing this information, forensic engineers reassembled the structure on paper, like drawing a jigsaw puzzle. With a mix of detailed descriptions, precise measurement, applied physics, informed speculation, and best guesses, St. Francis Dam investigators attempted to rewind time. If the St. Francis Dam failure was an engineering puzzle, understanding where these pieces came from and determining when they arrived downstream were critical to assembling a complete picture of what happened, and perhaps why.
Fragments of the St. Francis Dam, strewn downstream (J. David Rogers)
During the initial surveys, there was one piece investigators couldn’t find. It came from a section of the dam near the east abutment. All the enormous fragments that remained below the Tombstone were readily identified. Where was the concrete from this mysterious gap? The unaccounted-for fragment was dubbed “the missing section.” Although not considered important at first, it would turn out to be a surprisingly significant clue.
Along with a visual record of the location of fragments, the Fairchild photos captured another prominent feature of the floodpath—continuous light-colored bands along both sides of the canyon. They were scour lines, left as surging water stripped away foliage, leaving deposits of sand and silt that defined the changing path and levels of the flood.
When measurements from the air and on the ground were combined with other evidence, including eyewitness accounts, reports of power outages, and even mud-encrusted watches recovered from victims, investigators were able to calculate the torrent’s speed. The highest scour lines were just downstream from the dam, 140 feet above the creek bed. It was determined that about five minutes after the collapse the flood reached a maximum volume of seven hundred thousand cubic feet per second.3 At its widest, around Bard
sdale, the torrent was eight thousand feet—one and a half miles—across. At its fastest, between the dam and Powerhouse 2, the flow was clocked at eighteen miles per hour.4
The list of fragments, or blocks, from the St. Francis Dam started with the Tombstone—more than two-hundred-feet tall, one hundred feet across at the top, and eighty-five feet broad at the base. Why it survived was a major unanswered question. When two DWP employees scaled the water-stained monolith, they recovered one of the most useful and debated sources of information about how and why the St. Francis Dam failed—the Stevens Continuous Water Stage Recorder, or Stevens Gauge for short.
A drawing showing the original locations for dam fragments. The “missing section” is highlighted in gray. (J. David Rogers)
Light-colored scour marks left by the flood above Powerhouse 2 (Author’s collection)
Contained in a small metal shed, the device was connected to a seventy-five-foot-long, twelve-inch-diameter pipe attached to the upstream side of the dam. As water from the reservoir filled the pipe, a pencil resting on a float drew a line that recorded level changes on a slow-moving roll of graph paper. The gauge was controlled by a counterweight-driven clock mechanism that was wound once a week. An important part of dam keeper Tony Harnischfeger’s job was to check the recorder graph, keep the mechanism running, and verify its accuracy.5
The path of the pencil tracing represented a possible timeline for the St. Francis failure. After the reservoir was filled to capacity five days before the collapse, the line appeared to begin a barely noticeable descent. Around two o’clock on the afternoon of March 12, after William Mulholland and Harvey Van Norman had finished their examination of the new leak on the west abutment and returned to their downtown Los Angeles office, the pencil mark descended more noticeably, seeming to record an ominous 0.36-inch drop in the level of the reservoir, a loss of as much as seven million gallons. In the final forty minutes before the collapse, an accelerated decline of twelve inches began, ending with a precipitous fall. The line’s sudden descent apparently mirrored the abrupt drop in the level of the reservoir as 12.4 billion gallons of water escaped the confines of the St. Francis Dam and surged downstream.
