Subway Cars Should Be Like Centipedes

Most of the world has figured out this simple solution to overcrowding. Why does the U.S. resist it?

Singapore MRT
An empty open gangway train on Singapore’s MRT rail network.

Jonathan Chiang/Getty Images

The New York City subway is crowded, have you heard? Subway ridership is higher than at any point since the Truman administration. Delays rose 45 percent between 2013 and 2014 and were up another 20 percent by July.

Why so many delays? Because the trains are so crowded. Why so crowded? Because there are so many delays.

There are big solutions to this problem, like new tunnels and new signals. New York’s Second Avenue Subway, the world’s most expensive per-mile subway, is expected to open next year and carry 200,000 passengers per day. Billions of dollars in systemwide signal upgrades will allow trains to run closer together, easing congestion.

But there’s another, less extravagant solution to overcrowding, one so obvious that the rest of the world figured it out decades ago: open gangway trains. Often described as centipede or accordion trains, these interconnected cars have increased train capacity in cities such as Toronto by up to 10 percent. In New York, that extra space could contain the system’s past five years of record ridership growth. We could ride like it was 2011 again!

So why don’t we have open gangway cars in New York—or anywhere else in the United States? The answer lies partly in the uncertain science of transit design, a murky domain where 150 years of experimentation has yielded surprisingly few universal solutions.

Open gangways, as it happens, may be one of the more widely used elements of subway design. You will find trains like these on every subway system in China, India, Spain, and Germany, as well as in Dubai, Singapore, Tokyo, Rio de Janeiro, Paris, and Toronto. According to research by planner Yonah Freemark, open gangway trains run on 3 out of every 4 subway systems in the world. Mexico City hasn’t bought separate-car trains in two decades.

Yet open gangway trains are nowhere to be found in the United States. They will debut in Honolulu in 2018. New York City might request 10 of them—in an order of 950.

Why have we squandered so many opportunities to implement a technology that’s standard in Lima, Peru; Tokyo; and elsewhere? Open gangway trains have a number of advantages. Not only do they carry more people, but they’re also lighter (because they have fewer walls) and therefore use less energy. They make it possible for straphangers to move between cars, evening out congestion and enabling an escape from an unpleasant situation. 

That American transportation authority leaders are reluctant to embrace the concept reflects a couple of facts about how they do business. First, they clearly don’t spend enough time using the systems they run. Second (and relatedly), they are conservative about change. They are willing to let culture, or their perceptions of it, dictate design—rather than the other way around.

The most convincing explanation for the absence of open gangways in the United States is that planners feel “amenity-conscious” (or “choice”) riders would find them unpleasant. The enhanced mobility open gangways grant to beggars, merchants, and buskers has been cited as a potential problem with the model. That shouldn’t be sufficient reason to keep riders stuffed in like sardines. New Yorkers don’t take the subway because it’s pleasant, but because it gets them to work on time. The MTA could aid that cause by ordering a hundred—or a thousand!—open gangway train cars.

But it would be wrong to suggest that in the rest of the world, the subway car, like a shark or a horseshoe crab, has reached its final state of evolution. The shaping of a train car is not a search for a Platonic form, but a compromise among conflicting priorities.

The fewer seats in a train car, the more people it fits, and the better it functions at high capacity. According to a 2011 report by planners at Melbourne, Australia’s Monash University, when Melbourne tried four new train car models with less seating, the city found that the more seats were removed, the faster the cars entered and left the stations. That improvement, in what transit planners call “dwell times,” is the most significant variable in a train’s journey—and has a cascading effect in a busy and necessarily efficient system like the New York subway, Washington’s Metro, or San Francisco’s BART. In Stockholm, which moved from a 48-seat car to an experimental 26-seat car in 2008, 3 in 4 commuters would rather arrive on time than have a seat.

The more seats you scrap, however, the greater the burden you place on the elderly, the disabled, and long-distance commuters. In Paris, according to a 2007 paper by the then–director of transport for RAND Europe, not having a seat on suburban train journeys produced an effect on commuters equivalent to 27 additional minutes of travel time.

That awareness is informing an ongoing worldwide shift from transverse seating (like an airplane) toward latitudinal seating (like a limousine). The latter cars fit more people because they allow for more standers but still preserve plenty of accessible seats.

Passengers ride in a New York City subway car that’s not an open gangway train.

Spencer Platt/Getty Images

New York started replacing transverse cars with latitudinal cars in the early 2000s and hasn’t looked back. Unfortunately, the middle seats on those benches are almost universally despised, and according to New York transportation officials, seats aren’t 90 percent full until cars are 20 percent over seating capacity.

Further complicating matters, when Chicago tried that shift on its “5000” series, introduced in 2011, riders revolted against “New York-style” seating. The new cars did not increase capacity, in part because riders extended their legs into potential standing room. So the CTA will make a partial return to airplane-style seating with its “7000” series in 2019.

The research on Melbourne’s system affirms the intuitive idea that rush-hour passengers prefer trains with more standing room and fewer seats. This is why Paris has snap-back seats; each car can lose 24 seats at peak times to make room for more passengers.

Boston took a more radical approach. In 2008, the MBTA rolled out two “Big Red” cars at the center of some rush-hour trains that more closely resembled boxcars than Pullman cars—they had only a couple of seats available. Eight years later, administrators haven’t expanded the trial. A similar planned experiment in New York, which would have increased capacity by 18 percent, does not seem to have left the station.

So you can see why transit designers are perennially fiddling with layouts: There is no such thing as the right train car.

But if there were one, it would have open gangways.