The Board of Inquiry and the Engineer’s Gavel

Seen from above, the water system of San Francisco in 1906 was first a map: a network of iron and steel pipe drawn across miles of hill and dale, from the reservoirs on the slopes down through the high-pressure mains that fed the business district, branching into the smaller service pipes that entered each building. The Spring Valley Water Company, a private corporation, maintained this network with a staff of engineers who knew every valve and hydrant. On paper, the system was a marvel of municipal engineering, capable of delivering millions of gallons daily to a large population. The paper did not show what the earthquake would reveal.

In the months following the fire, as the ruins cooled and the Army’s relief houses rose in Golden Gate Park—5, 610 redwood and fir structures designed by John McLaren, grouped in eleven camps and rented to displaced families at two dollars per month—a different kind of construction began in Sacramento.

Governor George Pardee, himself a physician trained in observing symptoms, authorized a Board of Inquiry to determine why the water had failed when the city needed it most. The board’s mandate was technical, not criminal. It would examine the physical evidence of broken mains, measure pressure at critical points, and reconstruct the sequence of hydrant failures. It would not investigate why Mayor Eugene Schmitz had issued shoot-to-kill orders, or whether Brigadier General Frederick Funston had exceeded his authority in deploying troops, or who had authorized the dynamiting campaigns that had destroyed as much property as the fire itself.

These questions fell outside the engineers’ compass. The board would answer what had broken, not who had broken it.

The hearings convened in the new year, 1907, in a temporary courtroom near the rebuilt City Hall. The first witness called was the chief engineer of the Spring Valley Water Company, who had designed the original system three decades earlier. He was seventy-three years old, white-haired, and unaccustomed to having his work questioned in public. He brought with him a complete set of plans, rolled in cylinders, and a sheaf of pressure readings taken in the hours after the earthquake. The board’s chairman, a state appointee named William Hammond Hall, asked him to explain why hydrant 307, at the intersection of Twentieth and Church Streets, had registered zero pressure at 5:20 a.m. On April 18, 1906, despite being connected to the high-pressure main from the Twin Peaks reservoir.

The engineer unrolled his map. The hydrant in question was one of fifty-four installed in the Mission District after the 1901 fire had exposed the vulnerability of the low-pressure system. It drew from a sixteen-inch main that ran along Church Street, fed by a tunnel through the hills. The earthquake, he testified, had severed this tunnel at multiple points. Water that should have flowed downhill instead flooded the subsurface strata, emerging in springs that residents would report for weeks afterward. The break was not visible from the surface; only the dry hydrant announced it.

Hall asked whether the tunnel had been constructed to withstand seismic shock.

The engineer hesitated. The tunnel had been built to the standards of 1898, he said, which did not include lateral bracing against ground displacement. The San Andreas fault ran within two miles of the reservoir. This fact had been known to geologists since the 1890s, when Andrew Lawson’s survey parties had traced its line through the Coast Ranges. It had not been known, or at least not acted upon, by the water company’s board of directors.

The questioning turned to the cisterns. San Francisco maintained 177 underground reservoirs, brick-lined chambers holding between 75, 000 and 200, 000 gallons each, distributed through the downtown and South of Market areas as emergency reserves for exactly the contingency that had occurred. Fire Chief Dennis T. Sullivan, who had died from injuries sustained when a neighboring building collapsed onto his apartment at Engine Company 2, had inspected these cisterns annually and reported them full in his last departmental review. Yet when firefighters attempted to draft from them on the morning of April 18, many proved empty or inaccessible.

The engineer explained that the cisterns were fed by the same mains that supplied the hydrants. When the mains broke, the cisterns lost their source. Some had cracked; others had been buried under rubble. The board’s own investigators had found that cistern 89, at Mission and Third Streets, contained 18, 000 gallons of water that could not be reached because the access hatch had been paved over during a recent street improvement. The fire department had no record of this change.

Here the testimony reached its crucial turn. Hall produced a telegram sent by the assistant chief at 6:47 a.m., requesting permission to dynamite buildings along Second Street to create a firebreak. The message had been relayed through the mayor’s office and forwarded to General Funston at the Presidio. Permission had been granted by telephone at 7:15 a.m. The assistant chief was dead, killed by falling debris on the second day of the fire. The board had his written reports, his diagrams of proposed firebreaks, his estimates of how many structures would need to be destroyed to save the rest. What it did not have was any record of who had actually carried out the demolitions, or whether they had followed the assistant chief’s plan.

The engineers had mapped the water system with precision. The military operations remained opaque.

—-

The Army’s chief medical officer, who had directed the Presidio’s response, appeared before the board on February 14, 1907. His testimony covered the establishment of military control, the deployment of troops to guard the Federal Mint and the Sub-Treasury, and the eventual extension of martial law across the entire city. He was precise, clinical, a man who spoke in numbered paragraphs. He confirmed that General Funston had ordered troops into the streets at approximately 7:00 a.m. On April 18, before receiving any formal request from civilian authorities. He confirmed that these troops had assisted in demolition operations, sometimes using military explosives, sometimes working alongside civilian crews. He could not say, because no records had been kept, how many buildings had been destroyed by military action as opposed to fire.

Hall asked whether he had coordinated these demolitions with the fire department.

The medical officer replied that coordination had been impossible. The fire department’s command structure had collapsed with Chief Sullivan’s injury. The assistant chief had assumed tactical command but lacked the authority to direct military operations. Mayor Schmitz had issued proclamations but remained physically absent from the command posts. In this vacuum, he said, military officers had made decisions based on immediate necessity. The concept of firebreak calculus—his phrase—had been applied intuitively rather than systematically. Buildings were destroyed when they threatened to spread the fire, or when they impeded troop movements, or when they contained supplies that might attract looting. No comprehensive map of the demolitions existed. The Army’s engineers had kept field notes, but these had not been collated or preserved.

The board pressed him on the authorization for dynamiting. The medical officer produced a memorandum from Funston dated April 20, directing that all necessary measures be taken to arrest the progress of the conflagration, including the destruction of private property when such destruction appeared essential to the public safety. The memorandum cited no statutory authority. Funston had later claimed that his actions fell under the general provisions of the Insurrection Act, though no insurrection had been declared. The board noted this legal ambiguity but declined to pursue it. Its mandate was technical.

What the medical officer could provide was physical evidence. The Army’s engineers had surveyed the ruins of the Palace Hotel, the Grand Opera House, the Examiner building, and other structures destroyed by military explosives. Their findings, entered as Exhibits 47 through 62, showed that many demolitions had been executed without regard to the fire’s actual path. The Palace Hotel, for example, had been dynamited on April 19 despite being separated from the active fire zone by two city blocks. The explosion had scattered burning debris across adjacent structures, possibly expanding the fire’s reach. Similar patterns were documented on Kearny Street, on Market near Third, and along the waterfront where military crews had cleared a fire lane that subsequently burned anyway.

The medical officer defended these decisions as the best available under conditions of extreme uncertainty. The board’s engineers, reviewing the same evidence, reached a different conclusion. Their supplemental report, appended to the official findings, noted that the employment of explosives as a firefighting measure, while theoretically valid, was in practice attended by circumstances that frequently resulted in the extension rather than the arrest of the fire. The engineers came no closer to criticism than this sentence. They did not name the officers responsible. They did not recommend disciplinary action. They simply recorded that the method had failed, and that no systematic review of its effectiveness had been attempted.

—-

The civilian witnesses came next, and their testimony introduced a different register of loss. A plumber employed by the water company described his attempt to reach the main valve at Fourth and Mission on the morning of April 18. The valve, which could have isolated the broken mains and preserved pressure in the undamaged portions of the system, was buried under four feet of rubble from the collapse of the California Hotel. The plumber had worked with his hands for six hours, without tools, before being driven back by approaching flames. His supervisor had ordered him to evacuate. The valve had never been opened.

A resident of Hayes Valley testified that she had observed soldiers dynamiting her neighbor’s house at approximately 9:00 a.m. On April 19, while her own house remained standing. She had appealed to the officer in charge, explaining that the fire was still three blocks away and moving in a different direction. The officer, whom she could not identify, had told her that the demolition was precautionary. Her house had burned anyway, ignited by sparks from the explosion. She had received no compensation; the Army’s claims process required documentation of military authorization, which she could not provide.

These individual stories accumulated without cohering into a pattern that the board could recognize. Each witness described a specific failure—a dry hydrant, a premature demolition, a blocked street—but the board’s methodology required generalizable findings. It could document that 23, 000 linear feet of water main had ruptured, that pressure in the high-pressure zone had dropped to zero within forty minutes of the earthquake, that the cistern system had failed to function as designed. It could not, within its mandate, assess whether these failures resulted from engineering error, natural catastrophe, or institutional neglect.

The testimony of the hydrologists brought this tension to the surface. A physician who had conducted independent measurements of the water system’s performance during the fire presented data collected from surviving pressure gauges and eyewitness accounts. His findings showed that adequate pressure had existed in the Twin Peaks reservoir throughout the critical period. The water was available; it simply could not reach the hydrants. The breaks in the distribution network were not random. They clustered along the line of the San Andreas fault trace, where ground displacement had been greatest, but they also appeared at points of known engineering weakness: joints between different pipe materials, intersections where mains crossed under streetcar lines, sections that had been repaired after previous earthquakes and never tested under stress.

The physician’s analysis implied a system designed for normal conditions and never audited for resilience. The Spring Valley Water Company had resisted calls for seismic upgrades, citing cost. The Board of Supervisors had accepted these assessments because rate increases were politically unpopular. The state Railroad Commission, which had authority over water utilities, had focused on pricing and service quality, not disaster preparedness. Each decision had been reasonable in its context. The aggregate result was a network that shattered under stress.

The board’s counsel asked whether better engineering could have prevented the fire’s spread. The physician replied that no water system could have suppressed a fire of that magnitude once established, but that adequate pressure in the first hours might have contained the initial outbreaks before they merged into a general conflagration. The board recorded this as hypothesis, not proof.

—-

The military testimony resumed in March 1907, with the appearance of the senior commander of the Division of the Pacific, Funston’s superior. He had not been present in San Francisco during the disaster; he arrived on April 22 to assume formal command of relief operations. His role before the board was to explain the chain of authorization that had allowed Funston to act independently in the crucial first days.

His testimony was brief and carefully constructed. Funston had been the senior officer present. The telegraph lines to Washington had been severed by the earthquake. Under Army regulations, he had authority to protect Federal property and, by extension, to assist civilian authorities in maintaining order. The extension of this authority to encompass demolition operations, the establishment of refugee camps, and the de facto governance of the city was, he acknowledged, an interpretation of considerable latitude. He defended it as necessary and successful. The fire had been stopped. The city had been preserved. The Army’s role had been widely praised in the national press.

Hall asked whether any review had been conducted of the specific decisions made during the emergency. The commander replied that a comprehensive after-action report had been prepared by Funston’s staff and submitted to the War Department. This report, he believed, remained classified. The board requested a copy. The War Department, responding through the Adjutant General’s office, declined to release it, citing operational security. The board noted this refusal without protest. Its mandate was technical, not military.

What the board did obtain, through persistent inquiry, were the engineering surveys conducted by the Army’s own personnel. These documents, prepared for internal use rather than public release, contained assessments that contradicted the optimistic narrative of successful firefighting. An Army engineer who had surveyed the burned district in May 1906 had mapped the progression of the fire against the locations of military demolitions. His findings, entered as Exhibit 89, showed that in seventeen of twenty-three major demolition sites, the fire had already passed or would have passed regardless of the structures’ removal. In four cases, the demolitions appeared to have accelerated the fire’s spread by creating channels of burning debris. In only two cases was there clear evidence that demolition had arrested the fire’s progress.

The engineer had appended a recommendation that the Army develop standardized protocols for emergency demolitions, including coordination with fire officials, pre-positioning of explosives in vulnerable districts, and post-action review of effectiveness. No such protocols had been adopted. The board included this recommendation in its final report, again without commentary.

—-

The board’s deliberations extended through April 1907, the anniversary of the disaster. Hall and his colleagues worked from a conference room in the rebuilt Merchants Exchange Building, surrounded by maps, testimony transcripts, and the physical artifacts of failure: sections of ruptured pipe, fragments of exploded dynamite casing, the charred remains of pressure gauges recovered from burned engine houses. They had accumulated a mountain of specific evidence and faced the task of distilling it into findings that would satisfy their mandate without exceeding it.

The central question was whether to name responsible parties. The engineering failures were clear: inadequate seismic design, poor maintenance of cisterns, lack of coordination between water company and fire department. These were institutional, not individual. The military operations were murkier: unauthorized demolitions, absence of record-keeping, possible excesses in the exercise of emergency power. Here individuals had made decisions, but the board lacked authority to assess their legality or assign blame.

The compromise, reached in late April, was to separate findings from recommendations. The findings would be purely technical: the water system had failed due to earthquake damage exacerbated by design limitations; the cistern system had been inaccessible due to rubble and poor maintenance; military demolitions had been conducted without systematic coordination and their effectiveness had not been evaluated. The recommendations would be prospective: future water systems should include seismic redundancy; cisterns should be mapped and regularly inspected; emergency demolition authority should be defined by statute with clear chains of command.

No individuals were named. The Spring Valley Water Company was criticized for inadequate provision for seismic resistance but not for negligence. The Army was commended for prompt and effective response while its demolition practices were noted as requiring review. Mayor Schmitz and General Funston appeared only as recipients of telegrams, authors of orders, figures in the administrative record. Their decisions were described, not judged.

The board’s final report, submitted to Governor Pardee on May 15, 1907, ran to 340 pages of text and 87 appendices—the most comprehensive technical analysis of an American urban disaster to that date. It established beyond dispute the mechanical causes of the water system’s failure and the operational chaos of the firefighting response. It provided the documentary foundation upon which all subsequent assessments would build. And it carefully, deliberately constructed a narrative in which no one was at fault for anything that could be attributed to the earthquake itself.

—-

The report’s publication was timed to coincide with the opening of the new municipal water system. In November 1906, San Francisco voters had approved a $5 million bond issue to construct a public water supply, ending the Spring Valley monopoly. The first stage of this system, drawing from the Hetch Hetchy valley in the Sierra Nevada, would not be complete for twenty years, but the political decision had been made. The board’s findings provided technical justification for this transition without requiring any official to acknowledge that private utility management had contributed to the disaster.

In the Army camps at the Presidio, where 20, 000 refugees still lived in McLaren’s redwood houses, the report circulated in summary form. The chief medical officer, now promoted to brigadier general, issued a memorandum noting that the board had confirmed the necessity and propriety of military intervention during the emergency. He did not mention the findings on demolition effectiveness. The relief houses continued to be rented at two dollars per month, the camps gradually clearing as families rebuilt or departed.

The physical evidence gathered by the board was stored in a warehouse at the corner of Spear and Mission Streets, available to researchers who obtained permission from the State Engineer. Most of it was never requested. The ruptured pipe sections rusted; the testimony transcripts yellowed; the maps became outdated as the rebuilt city grew in different directions from the one the board had surveyed.

In 1909, the city launched the fireboat Dennis T. Sullivan, named for the chief who had died in the first minutes of the disaster. It was the most powerful firefighting vessel on the Pacific Coast, capable of pumping 10, 000 gallons per minute. The dedication ceremony featured speeches about progress and resilience. No one mentioned that Sullivan had spent his last years warning about exactly the failures the board had documented—or that his warnings had been filed without action.

The official, sanitized technical verdict was now a public document, creating a definitive factual record upon which arguments about liability and reform must now be built—or founder.