Market Street subway dreams

Hard for some of us San Franciscans of a certain age to think of the Muni subway under Market Street as a part of history. Because that means that we ourselves…well, you know.

But it’s true. The shiny new Muni Metro subway opened to carry passengers on its first line, the N-Judah, on February 18, 1980, bringing a long-held dream to reality, and, to cynics, starting an operational nightmare that has recurred repeatedly since. 

The subway, and all Muni rail service, was shut down in Spring 2020 as part of the response to the Covid-19 crisis, with buses using surface routes replacing the subway lines. The subway gradually reopened through 2021, but it is still not operating at nearly the volume, either in passengers or trains, as it did before the pandemic. The shutdown, though, provided an opportunity to rethink how best to use this multi-billion-dollar asset to effectively move the most San Franciscans along the City’s most travelled corridor. One option might provide opportunities to get more value from the City’s historic streetcars as well.

This story covers the vision and creation of a subway under Market Street: 75 years of dreaming, planning, and building. Separately, we report on the subway’s first 40 years of operation, providing insights into why it has never worked nearly as well as promised.

Subway stirrings

New York City famously opened the first full subway in America, the IRT, in 1904. (Boston had dug a short streetcar subway in 1897.) On Christmas Day 1904, the San Francisco Chronicle gave big play to the vision of an engineer who’d been to New York, suggesting that Market Street needed one of those things too, along with an array of other rapid transit (including a steel train tube under the Bay to Oakland, presaging BART). 

But like many transit “plans” over the decades, this was just a concept, with no official backing. The aftermath of the 1906 Earthquake and Fire soon refocused San Francisco’s attention, but as recovery from the cataclysm brought growing transit ridership, dreams grew again as well. 

In 1913, with the new San Francisco Municipal Railway less than a year old, City Engineer M.M. O’Shaughnessy decided to design the East Portal (at Castro Street) of the yet-to-be-built streetcar tunnel under Twin Peaks “to permit the continuation of the tunnel as a subway under Market Street with but slight alterations.” The tunnel, which opened in 1918, succeeded in its goal of opening up the southwestern quadrant of the city to residential development, though it took several decades for the build-out. 

When the tunnel opened, city-owned Muni and its private competitor, soon to be known as Market Street Railway (MSRy), began jousting for passengers all along Market Street on adjacent pairs of tracks (which came to be known as “The Roar of the Four”). The City’s official position called for public ownership of all utilities, including transit. As the expiration date of MSRy’s most important operating franchises neared, voters were asked to buy out the private company in 1925, but overwhelmingly refused. Those franchises expired in 1929 but were extended short-term. 

Two 1930s plans

O’Shaughnessy began preparing a “subway report” in 1930 that showed an ambitious two- and four-track subway under Market, costing an estimated $21 million. The first phase ramped two tracks down from the surface just east of Valencia Street (with the notation that “in the future” the subway could be extended to connect directly to the Twin Peaks and Sunset Tunnel). Two other tracks stayed on the surface all the way to the Ferry. These two subway tracks were joined by two more down a ramp from McAllister Street. From the drawings, we infer that this portal would serve both MSRy’s busy 5-McAllister line and its 21-Hayes line. 

The four-track subway stretch marched east, joined at O’Farrell Street by tracks that descended into their own subway at Larkin Street. These were for Muni’s Geary lines. At Sutter, two of the subway tracks ramped up to the surface, finishing the trip to the Ferry Loop next to the MSRy’s Sutter line tracks, the only stretch of four tracks retained on Market’s surface. The other two subway tracks terminated underground in a loop where Bush, Battery, and First Streets met Market. This anticipated the construction of the Transbay Transit Terminal a block south. O’Shaughnessy said trying to dig a subway east of that point would be ruinously expensive because of poor landfill, old ships, and a high water table. (He was proved right.)  

Subway stations were spaced very close together, at Third, Fourth, Sixth, Ninth, and Van Ness. 

Unfortunately, his work was for naught, since by the time the plan was released in 1931, voters had already approved extending Market Street Railway’s franchises another 25 years. The private company thus held a veto over the big capital project, as they didn’t have to invest a dime in it, and it wasn’t feasible without incorporating MSRy lines. (Remember, at this time, MSRy carried 75% of the city’s transit passengers, more than three times Muni’s share.) 

But the subway dream lingered. A consultant’s report in 1935 recommended a more extensive system, with a two-track Market Street subway running from Fremont to Church, surfacing to let the J and N lines out, and then continuing on the surface to the Twin Peaks Tunnel. The report also called for a subway line through the Mission District, surfacing at the Bernal Cut (later BART’s alignment) and one under Geary Street from Divisadero to Market, then turning under Montgomery to Columbus, linking the B and a relocated F-Stockton line. It retained surface operation of the outer ends of the existing streetcar lines, like Muni Metro, but pledged it would be designed to allow an eventual upgrade to true grade-separated subway operation. 

Unlike O’Shaughnessy’s 1931 study, this plan made it to the ballot in 1937, with a price tag just under $50 million. The City’s Public Utilities Commission, which oversaw Muni, promoted it with the promise of construction jobs as the Depression lingered, and the fear of losing out to the suburbs. “With [the new San Francisco-Oakland and Golden Gate] bridges increasing traffic and providing easy outlets to suburban areas, the adoption of this rapid transit plan is vital. It is needed to build industry, stimulate growth, and prevent the loss of population and industry. It is needed to protect realty values and safeguard tax revenues.” 

The bridges were indeed flooding city streets, including Market, with commuters’ cars, causing growing traffic jams. Yet voters still turned down the bond issue that would have funded the three subways. With the bridges’ openings, ferry service withered, and lower Market Street businesses with it. Most streetcars were diverted to the Transbay Transit Terminal at First and Mission Streets, which opened in 1939. Though the automobile was already cutting into rail ridership around the nation, the new Bridge Terminal started with three privately owned interurban passenger rail companies serving it, including the Sacramento Northern, much of whose right-of-way through Contra Costa County would later be used by BART. But within two years, economics forced two of the three companies off the bridge, with the last one, the Key System, hanging on until 1958, when its tracks on the bridge were replaced by more automobile lanes. 

More plans without results

As Muni prepared to merge with MSRy in 1944, the Board of Supervisors asked City PUC boss E.G. Cahill what his plans were for improving the combined systems. Having been beaten at the polls on his 1937 subway plan and facing strong labor opposition on reducing rail crew size, he knew how to respond, saying there would be no proposals “for subways, elevated railways or other types of grade separation…until, by the employment of modem equipment, the most efficient use is made of the City’s street surfaces.” Instead, Muni won voter funding in 1947 to buy hundreds of trolley buses to convert streetcar lines. 

That goal accomplished, subway dreams resumed. A 1949 subway study by De Leuw, Cather & Co. for the city planning department included an underground loop to serve Transbay Terminal. Modernized PCCs would be used to start. There would be stations every block on Market between Second and Fifth. A second subway started from its own tight loop close to Transbay Terminal under Second Street and Post Street, to replace the Geary streetcars. 

This plan likewise went nowhere, but just three years later, at the request of Mayor Elmer Robinson, the City’s Public Utilities Commission, which then oversaw Muni, commissioned yet another study, this one entitled, “Rapid Transit for San Francisco: Monorail, Elevated, Subway? A Report of Possibilities”. It was authored by consultant Marmion D. Mills, infamous among Muni history buffs for savaging the streetcar system in the late 1940s and recommending a plan, which succeeded, to cut the cable car system in half in 1954. Mills’ plan, too, called for terminating the system at Transbay Terminal, and considered alternatives to a subway. It apparently wasn’t taken seriously (we couldn’t find anything in a newspaper search of that period), but how about those cool drawings?

And then, BART

Probably the main reason the Mills study went nowhere was the blossoming of a bigger dream: an integrated Bay Area-wide rapid transit system. After World War II, the counties around San Francisco boomed in population. The city itself didn’t, at least in terms of people who slept there every night. But the number of workers and shoppers—and the cars they drove—jammed the downtown area, leading to one-way streets feeding new parking garages in the city (and helping doom the O’Farrell cable car and Geary streetcar). 

Regional planners and business leaders in the 1950s assumed that downtown San Francisco would remain the epicenter of the Bay Area white-collar employment and retail business. They revived yet again the dream of a Market Street Subway, but on a far grander scale, as part of a high-speed passenger rail network ringing the Bay Area.

After politics, costs, and squabbles caused Santa Clara, San Mateo, and Marin Counties to drop out, Bay Area voters passed a $792 million bond issue in November 1962 to build the 75-mile, three-county Bay Area Rapid Transit (BART) system, using what was billed as space-age train technology. (The initial system ended up costing twice that).

To lure San Franciscans’ votes, the subway under Market Street would have two levels, BART on the lower, Muni on the upper; an underground echo of the old “Roar of the Four” surface tracks. In February 1962, BART officials and San Francisco Mayor John F. Shelley’s “Transportation Council” agreed that the subway level constructed for Muni would allow all five existing streetcar lines to use it. So, San Francisco voters who approved BART believed single-seat streetcar rides downtown would continue, with faster service in the new subway.

“What do we do now?”

In the 1972 film The Candidate, a California Senate candidate played by Robert Redford, who had vacillated between various positions during the campaign, comes from behind at the last minute to win. With a dumbstruck look on his face, he turns to his campaign manager and utters the film’s final line: “What do we do now?”

In the real world, Muni leaders and city politicians found themselves asking that same question, over and over, after the voters approved BART at the end of 1962. 

The BART district had complete control over the design and construction of its system, including the Muni subway level, which, as approved by voters, allowed all five existing streetcar lines to access it. The bond issue extended the subway on the west beneath West Portal Avenue to surface at St. Francis Circle. BART would own the subway and designed Muni’s level to be convertible to BART trains in the future if needed.

In the voter-approved plan, the easternmost San Francisco station, for both BART and Muni, was Montgomery Street. Like every subway plan made since the Bay Bridge opened, BART designers scorned lower Market Street and the Ferry Building area (by then cut off by the double-deck Embarcadero Freeway and economically deteriorating). They did include an underground track loop east of Montgomery to allow single-end Muni vehicles to turn around.

But while BART would design the Muni subway and stations, it had nothing to do with what kinds of vehicles Muni would run, nor would it pay for them. And BART had no say in how Muni would operate it. Frankly, Muni didn’t know either.

The Simpson-Curtin Report

Transit leaders on both sides of the Bay recognized that the building of BART would greatly impact all existing transit operations. They commissioned a detailed study of all aspects of transit affected by the advent of BART, which was undertaken by the Philadelphia firm of Simpson & Curtin. Early on, the project team recognized that the plan to operate all five Muni streetcar lines in the new subway was problematic.

Citing existing surface-subway streetcar operations in Boston and Philadelphia as examples, the Simpson-Curtin Report stated:

Where street cars are operated both on surface and below grade, headways between vehicles are not dependable inasmuch as they are a function of peak hour traffic congestion for the surface portion of the service. This characteristic is particularly significant during rush hours where the combination of street traffic and the need for close headways cause a progressive deterioration in the transition from sur- face to private right-of-way—scheduled speed is reduced accordingly.

– “Coordinated Transit for the Bay Area, Now to 1975”, Simpson and Curtin, p. 120

The report clearly stated that to keep the subway from backing up with congestion, streetcars on the individual lines had to run at least 7 1/2 minutes apart, “a considerably lower quality service than presently is being operated on the streetcar lines at grade.”

In its initial assessment of how to operate the new subway level, Muni planners looked at using its existing single-end PCC streetcars in the subway, running left-side so their right-hand doors could use the center platforms. The turning loop BART promised at the east end of the line made this option feasible, so long as the platforms were built at a low level (which, however, would have made the streetcars inaccessible to wheelchair users. This option also would have required a dangerous crossover for the streetcars somewhere to get them to the opposite track of the subway.

Though the Simpson-Curtin report was still in preparation, Muni and City officials knew that its recommended approach would be to scrap the existing streetcar system in favor of a completely grade-separated subway system, like New York’s. This, the experts on the team said, was the way to truly transform Muni’s rail transit. It should include replacing the M-Ocean View with true subway trains running underground all the way to San Francisco State College, replacing the N-Judah with a subway extension to 19th Avenue and Irving Street, and replacing the rest of the streetcar network with feeder buses connecting to the subway. The report would also propose a new subway under Post Street to Cathedral Hill, continuing under Geary Boulevard to 45th Avenue, to serve Muni’s highest-ridership corridor. (This last measure provided partial atonement for ripping out the B-Geary line in 1956.)

Believing the Simpson-Curtin team’s recommendation to be the best solution, the Board of Supervisors unanimously put on the November 1966 ballot a $96.5 million bond issue to fund the subway extension to San Francisco State, plus the construction of a new maintenance and storage facility nearby, and purchase of new high-level subway-style trains. The bond issue also dangled the prospect the N-line and Post/Geary subways as part of an overall vision, but did not fund them. Almost 58% of voters said ‘yes’, but it needed 2/3 approval, and so failed. Thus, when the Simpson-Curtin report was finally issued publicly in 1967, its key recommendation to make the Muni subway work well had already been rejected by San Francisco voters.

Back to the Drawing Board

As Muni planners regrouped after the 1966 bond issue defeat, a powerful member of the Board of Supervisors, John Barbagelata, led the fight to scrap the already-funded subway extension under West Portal Avenue, to avoid construction disruption on the street where his real estate office was located. He won, and that money was shifted to build an additional BART-Muni station under Market near the waterfront. (Embarcadero Station proved transformational for lower Market, and it is hard to see how Montgomery Station could possibly have handled rider volumes on its own.)

The bond issue defeat meant no money for new equipment, sending Muni back to the PCC plan until a new mayor, Joseph Alioto, decreed early in 1967 that Muni would buy European-style trams that could be operated in trains and load from either side in the subway. This led to Muni’s acceptance of a proposal by BART, under extreme financial pressure by now, to build a simple “stub-end” terminal for Muni railcars at Embarcadero Station rather than the turnaround loop promised and funded in the initial BART bond issue. That was the end of any talk of putting PCCs into the subway, since double-end vehicles would now be required. 

Analysis of the failed 1966 bond issue indicated that many users of the existing streetcar lines didn’t want to give up their single-seat ride to get downtown. This was no surprise to Muni leaders. In the late 1940s, an attempt to replace J-line streetcars with trolley buses on Church Street flopped when Noe Valley residents made it clear they wanted the streetcars to stay. And budget-cutting moves to replace K, L, and M streetcars with feeder buses west of the Twin Peaks Tunnel nights and weekends in 1951-52 aroused rider ire.

Muni management still explored multiple options, but by early 1969, finally committed to putting all five lines in the subway.

Reinventing the streetcar

To fit all five lines into the subway, Muni’s consultants, Louis T. Klauder and Associates (LTK), recommended having single cars run on the outer ends of the lines, and then couple together into trains at the subway portals to go downtown: the K, L, and M lines at West Portal Station, the N and J lines at Duboce and Church. LTK also repeatedly advocated for an underground loop to be built near Embarcadero Station, believing (correctly) the stub terminal would prove inadequate.

To carry this off, Muni needed new streetcars that could be coupled into trains, were double-ended, and could load from either high or low platforms, an unprecedented combination for an American-built streetcar. When Muni put LTK’s initial design out to bid in 1971 (a design that relied on PCC-type motors and some other components), bids came in at twice Muni’s estimate, unaffordable. 

Boston needed to buy new streetcars, too, and wanted a Düwag model designed for Hanover, Germany’s subway-surface Stadtbahn. But the Nixon Administration wouldn’t fund any of it, instead pushing a “Buy American” program that would help repurpose factories producing materiel for the Vietnam War. The Urban Mass Transit Adminstration took the LTK Muni design, worked with Muni and Boston to adjust it, and dubbed it the “US Standard Light Rail Vehicle”, wanting to leave the term “streetcar” behind. 

The design was skinnier than Muni wanted and had a slanted front end, both required to navigate the 1897 Boston trolley subway. This made the front doors unusable in Muni’s subway and limited boarding to two doors per car, on both sides. The bid was won by Boeing-Vertol, the helicopter division of the aircraft maker. 

It must be noted, San Diego, building its own light rail system from scratch, passed up the lure of federal funding and purchased proven German vehicles from Düwag (later acquired by Siemens). Those Model U2 vehicles remained in service for 40 years with a mean-distance-between-failure rate fifteen times lower than Muni experienced with the Boeing cars. (Many of San Diego’s U2s have gone on to a second life in Argentina.)

But given the parsimony of San Francisco voters in supporting Muni capital programs, federal money (strings and all) seemed the only way Muni could get new vehicles for its new subway.

Delays in vehicle procurement, along with other factors, pushed back the projected opening date for the Market Street Subway again and again. The subway itself was physically finished by early 1975, but it took five more years to begin service. The aging PCC fleet soldiered on, the venerable streamliners kept operating beyond all expectations by a dedicated maintenance crew, using the Twin Peaks Tunnel and “temporary” ramps (still there) to surface at Market and Castro, and a detour via 17th, Church, and Duboce to avoid the cut-and-cover subway construction on upper Market. And one of the PCCs, Work Car 1008 (today fully restored to its original 1948 passenger condition), pinch hit for the late-to-be-delivered LRVs in testing out the subway clearances and wiring, using a newly-installed pantograph.

The need to keep the PCCs running precluded redeveloping the PCCs’ Geneva Division for the new LRVs; a good thing, since it was too small anyway. Instead, Muni converted its ramshackle Elkton streetcar shops and Ocean Division bus yard across the street, next to BART’s new Balboa Park Station, into a new light rail maintenance and storage facility, later named for Muni general manager Curtis Green.

The way Muni navigated through the process of preparing for the Market Street Railway was, in the words of Anthony Perles, author of the definitive Muni history, The People’s Railway, “convoluted”.  

Indeed it was. What would happen when they actually tried to operate it? Here’s that story.

---

Monorail mania

Lest you think the 1950s Muni-monorail-above-Market was wacky, it should be remembered that it was seen in many places as the sexy new technology that would lure people out of their automobiles.  Walt Disney thought so, and partnered with Alweg of Germany to design a system for Disneyland. Monorail was briefly considered early on as the technology for the regional system too. As with the Muni study, consideration of monorail technology for what became BART was encouraged by San Francisco business leader Marvin Lewis, a driving force on mid-century Bay Area transportation issues (and grandfather of Salesforce Founder and CEO Mark Benioff, whose company’s name is on the sleek replacement for the old Transbay Transit Terminal). At the end of the day, only one monorail transit line was built in North America: in Seattle, for its 1962 World’s Fair. It’s still operating.

We at Market Street Railway not only work to preserve and support historic transit in San Francisco, we also research and write about how transit has shaped our city in the past – for better or worse – so that these lessons might be applied to make informed decisions for the future. We would appreciate it if you could join or donate to our nonprofit.