« how universal is transit's geometry? | Main | guest post: scott johnson on "over-constrained projects" »


TrackBack URL for this entry:

Listed below are links to weblogs that reference rail-bus differences: premise or conclusion?:


Feed You can follow this conversation by subscribing to the comment feed for this post.

Otso Kivekas

You omit the one fundamental difference that does exists: Capacity.

Busses carry 50-80 passengers, in some cases over hundred. A tram of hundred passengers is small, 6000 big. Metros and local trains have capacity of hundreds to thousands of passengers.

Of course you can always just increase the amount of cars on a lane/track. But there are limits how often they can go. And of course each separate driver is an additional cost.

Nevertheless, your point is valid: in many ways the key question is not the mode of transport, but the level of service.

David in Ottawa

I'd go with fewer columns unless there is an extremely compelling reason to stick to what you've got above. Look at all the trailing whitespace - that's happening because the boxes are too narrow to fit much in, so content keeps getting bumped to new lines.

How are you going to class the light rail systems and some BRT systems that grade separate sometimes (i.e. at major roads) but not always? Calgary and Edmonton come to mind in Canada, but I'm sure there are others elsewhere. These systems, despite being "Class A-B", have greater capacity than Class A busways like Ottawa (which isn't fully grade-separated anyway, even outside downtown) because they're rail-based and not bus-based. This "Class A-B" region is where a lot of transit systems operate in the real world.

I understand what you're trying to illustrate but one result is a false impression of equality of capability between bus and rail in any one class, but especially in Class B and the region between Class A and Class B. Buses require more grade separation than do trains for a given capacity in a given corridor due to the higher number of more random vehicle movements. Cross street traffic that can be interrupted once every couple of minutes on a schedule by a train might not be able to be interrupted by a bus coming along randomly on average every 30 seconds. In the real world, the train gets preemption but the bus gets a traffic light, possibly with priority, because preemption is a practical impossibility due to the numbers and the randomness. Accordingly, the designer of a light rail system will seek to avoid grade separations and put them in only where genuinely needed, whereas the designer of a bus system will seek to avoid level crossings as much as possible. This is why Calgary and Ottawa's initial systems cost about the same - while Calgary's was rail-based, it had far less grade separation than Ottawa's. The notion of BRT being cheaper is a myth because that myth conveniently ignores the fact that BRT needs to grade separate where LRT does not - but Ottawa's bus fanatics will always compare them on "equal" terms (i.e. grade separation for both) and then conclude that BRT is "cheaper".

A generation on, Calgary's "Class B+" C-Train has proved more successful that Ottawa's "Class A-" Transitway.


People tend to go down lists and checklists. Maybe turn that to your advantage here by making the columns Rail and Bus, and let people go down class A, B, C and so on, seeing how *similar* they are between bus and rail instead of how different?


As usual, you downplay the real-world experience with busways - that they have been and continue to be far more susceptible to being nibbled away (allow cabs, allow turning cars, allow carpoolers, heck, now we might as well allow cars) in a way that reserved lanes for light rail has not been.

Jeffrey Jakucyk

I can understand why Jarrett wants to put this topic to rest for a while, as it's been beaten to death over a number of articles. Here's what I think needs to be said though. You can argue that bus and rail systems can all be built to the same level of service until you're blue in the face, but it doesn't address the fact that in practice they usually aren't.

As the previous articles and their comments have shown, the differences are not all cultural/perceptual/misattributed. For instance, getting Class A service out of a busway is very likely more expensive than doing so with rail. The vehicles are shorter so more drivers are needed, the structure is wider and more expensive to build with strong guardrails at the side, the buses wear out sooner and so does the concrete. If it wants to be non-polluting it needs more complicated double-overhead wiring. By the time you get anywhere close to that point, buses don't make much sense anymore. All that extra width of the busway and expense and complication just to use a vehicle with a steering wheel and rubber tires.

That's just one example, but there's hundreds more for all modes. While it is worthwhile to discuss how buses can be made to operate more like rail, the real question is why isn't it done more? Even those "misattributed" differences between bus and rail are there for a reason after all.


While disagreeing with David on his references to "bus fanatics"--a pejorative which is not necessary in this forum--his distinction between "separate right of way, level crossings, priority over cross traffic" and the same thing without signal priority, is an important one. In the Portland context, the MAX aligment along SW Baseline Road in Beaverton/Hillsboro, the MAX alignments in the median of streets such as N Interstate or E Burnside, and the MAX aligments along Yamhill/Morrison downtown, are all "Class B"; but have vastly different performance characteristics. In the Beaverton/Hillsboro stretch, MAX functions like a freight railroad, with level crossings guarded by gates, and traffic forced to stop whenever a train approaches--the train only has to stop at stations in this stretch. MAX enjoys signal priority in the Interstate and Burnside sections, but frequently finds itself waiting at orange horizontal bars (the train signal meaning "stop") for cross traffic. And downtown, MAX enjoys little signal priority at all.

I would add two other categories to the first table, each of which are relevant only to one technology or the other. I'd add a "class A+" for driverless, grade-separated systems like SkyTrain--at the current level of technology, these can only be rail. And I would split Class C into two subclasses--fixed guideway, mixed traffic (streetcar and trolleybus), and free-running, mixed traffic (regular bus).

WRT the layout issue--I'd simply swap rows for columns, so you have one column each for rail and bus (or trolleybus if you wish to add it as a separate category), and as many rows as you like.

Sean Nelson

In addition to the capacity distinction so rightly pointed out by Otso Kivekas, there's also frequency. Class A rail can be automated, while Class A bus cannot - that leads to the likelihood that a Class A rail system will run more frequent service because frequency is not directly constrained by labour costs.



I'm not saying that these are the only distinctions that matter. I'm simply trying ways of illustrating the separability of the rail-bus distinction from the other distinctions that are often confused with it.

The distinctions mentioned in this post are examples of the idea, not a claim or complete list. The complete list is here:


Joseph E

@ Jarrett, do you intend for "Class A" rail to include "at-grade" tracks with gated crossings? I would consider these class A. For example, the Blue Line here in Los Angeles has a "Class B" section in downtown Long Beach, where the right-of-way is exclusive, but the train stops at red lights, and can only go as fast as 35 mph (55 kph) due to being in a street median.

However, between Los Angeles and Long Beach, the Blue Line has a wider right-of-way with gates at every crossing. Trains can go 55 mph or more (100 kph), and never stop for signals, normally. This section is just as fast and almost as reliable as a subway or elevated rail line (though occasional accidents at grade crossings are a real problem).

Perhaps this would be "Class A minus", and grade separated rail or busways could be "Class A plus"? For the transit rider the two services are almost the same (unless the "A+" service is a grade-separated, AUTOMATED driverless train, which can offer more frequent off-peak service)


HOT lanes are NOT fully exclusive guideways for bus. Sure, congestion pricing can promise you travel times similar to a Class-A facility. But once you introduce other users, you risk being at their mercy, when future negotiations happen.


@all. Great comments over at Streetsblog as well:


Several commenters there claim that there's no point in using buses for Class A and B, to which I replied:

Thanks for the link and comments. I don’t agree that rail is always the right answer in Class A and Class B settings, partly because I think there should be a lot more Class B, reaching down to lower levels of ridership than are needed to justify rail.

Note too that the main virtue of busways in Class A-B situations is that you don’t need to build the entire line to all the destinations you want to serve. Busways can be implemented incrementally, and can be used in situations where you want service to flow through to several possible off-line destinations. This is extremely germane to a number of new starts situations. The developed-world leader on this kind of busway is Brisbane, Australia. Lots of posts on that at Human Transit, but you could start here:


Another example is the on-street bus lane, which if placed in the median is effectively Class B, while if placed on the curb (and thus prone to interference by turning traffic) would have to be called B-minus. In many cases you need this bus lane only for certain segments but want to flow through to many others. The Portland Transit Mall, for example, is a Class B facility, and has been since 1978.

As for cost/passenger issues, yes, there’s a level of capacity need where rail is essential, but also lower levels where it’s not, but where there’s plenty to be gained by using Class A and B facilities where you can get them.

There are many reasons to seek Class A and B facilities even when planning for buses, which is why all six cells in my table are well-populated.


It's obvious that for coming up with mobility solutions, the distinctions in the columns are considerably more important than the distinctions in the rows. But the problem is that the distinctions are not entirely orthogonal in theory, much less in practice. Rail and bus can each provide services that the other can't, and as was mentioned, rail can provide an A+ level of service where bus just can't (and bus can provide a C- level of service where rail can't). And in practice, it gets even messier, because "bus is cheaper than rail" is often code for "bus is easier to downgrade (in terms of columns) than rail". Now, this sort of downgrading can happen because of sound engineering concerns, like it not making sense to have expensive exclusive ROW way out into the suburbs. But it's just as likely to happen because of political factors like NIMBY complaints. Rail, being more resistant to downgrading (you can't add grade crossings to a metro line, for example), tends to make it more likely that those arguing against the downgrade will prevail. With a bus, the default is presumed to be Class C, and so it's more likely that the ROW-class will be downgraded and the resulting service worse.


My preferred refactoring, also posted at Streetsblog:

Class A+: Fully grade-separated ROW, automatic vehicle control.
Rail: Driverless metro (SkyTrain)
Bus: Not yet possible with current technology.

Class A: Fully grade-separated ROW, manual operations.
Rail: Most subways, els, highway-running LRT systems.
Bus: Grade-separated busways (Brisbane, Curitiba)

Class A-: Segregated ROW, absolute transit priority at level crossings; transit vehicles only need to stop at stations under normal operation.
Rail: Many LRT systems (ie MAX Blue Line between Beaverton and Hillsboro)
Bus: Some busways

Class B+: Segregated or exclusive ROW, limited signal priority (i.e. transit may need to wait at level crossings).
Rail: Many LRT systems; much of Hong Kong New Territories light rail.
Bus: LA Orange Line

Class B: Exclusive ROW adjacent to traffic, limited signal priority.
Rail: Median running LRT/Streetcar
Bus: Portland Transit Mall, median BRT.

Class B-: Exclusive ROW adjacent to traffic, no signal priority and/or possible interference from turning or parking cars.
Rail: Yamhill/Morrison segment of MAX
Bus: Curbside bus lanes

Class C+: Flexible (variable-route) mixed-traffic operation with signal priority
Rail: Not possible with present technology.
Bus: “Rapid bus”, i.e. LA Metro Rapid service

Class C: Flexible mixed-traffic operations, no signal priority.
Rail: Not possible with present technology.
Bus: Standard local bus service

Class C-: Fixed-route mixed traffic operations.
Rail: Local-service streetcar
Bus: Local-service trolleybus.

As a note for anonymouse, under this refactoring, bus can provide "C" and "C+" service, whereas the only class C service rail can provide is C- service. And perhaps the mobility difference between mixed-traffic, fixed guideway service (like streetcars) and mixed traffic flexible service (like POBS, "plain old bus service") is sufficient that we ought to banish rail from Class C altogether and put streetcars and trolleybusses in Class D; with trolleybusses getting a D+ because they can at least maneuver around limited obstructions without dewiring. Doing that is left as an exercise for the reader.

In Brisbane

Hi Jarrett,

This is fantastic! I have been writing about Class of Right of Way for a while now trying to get people to understand the difference between mode and ROW. Maybe some pictures would help.

If you take buses and trams off streets and place them into ROW class A you get BRT and LRT (with a little fiddling of the stopping distance of course).

How about adding a comment about $$$ as well? Class A is the most expensive, takes the longest to construct and is most disruptive. However it also has the highest benefits. Class C is the least expensive, fastest to implement but has the lowest benefits. And this makes sense- higher quality costs more.

In Brisbane

Just one other thing- Paris metro. Is that rapid transit? I've heard that the average speeds on that is about 20 km/hour or thereabouts. How is that "rapid".

Jarrett at HumanTransit.org

In Bris.  I think the metro is better than that but it does have very close station spacing.  Note that I didn't tie specific speeds to the three classes, because there are so many other variables (signal priorities, acceleration).

In Brisbane

Jarrett, I think this is great. Do you think it help if the left column (Bus, Rail) were deleted?

One other thing- what about ferries like CityCat and Bicycles? These often get forgotten. Class of ROW also extends to these as well. CityCat runs in Class C ROW. Bicycles can be Class A (dedicated bikeway), Class B (painted lane) or Class C (yellow bicycle stencil-man on road!) as well.

Joseph E


Your finer distinctions may be too complicated for the general public, but I love it! And everyone (in the USA, at least) understands A/B/C with pluses and minuses.

Is it really impossible to have a "Class C+" train? I know this does not currently exist, but why couldn't it? The Metro Rapid buses here in Los Angeles are Class C+, because they have transponders that will sometimes hold the light green, or change it from red to green for the bus. Couldn't the same exact technology be used in a streetcar?

Of course, I think all "Streetcars" should be at least Class B-, else they are not much of an improvement over local buses.


There's a lot of parameters here which can be varied sort of independently, mapping them onto a linear scale isn't always easy. Yes, mixed-traffic streetcars can have signal priority. Does a mixed traffic streetcar with signal priority fare better than a mixed-traffic bus without? Depends on how readily the things in front of it get out of its way.


And of course class of ROW extends to cars too, with Class A being freeways, Class B being things like arterials with relatively few, synchronized lights, and Class C being local streets with stop signs.


What Class A busways exist in Los Angeles? The Harbor and El Monte transitways are both shared with private carpools (becoming quite congested with these at times), and are therefore Class C.

This sort of discussion seems to inevitably lead to rail advocates exaggerating the inherent benefits of rail while bus advocates exaggerate the quality of existing implementations of buses. Plenty of rail projects have been downgraded due to cost-cutting and NIMBYism, the East LA Gold Line (originally proposed as an extension of the heavy rail Red Line, but implemented as light rail that waits at stoplights) being a prime example. The existence of things like Rotterdam's ParkShuttle makes me think that automated buses might not be as far off as others seem to. Triple-articulated buses are feasible on Class A lines, and begin to compete with light rail capacity. And the flexibility of buses (e.g. for through-running) really can outweigh other drawbacks in many situations. On the other hand, every existing implementation of bus transit I have seen has been thoroughly disappointing. All of the systems Jarett lists as "Class A" include at least a few signalised intersections (which do not appear on any existing "heavy rail" system), and some even involve major mixed-traffic segments. I know of no busway that uses crossing gates. And as noted by others, counting things like HOV/HOT lanes as Class A is ridiculous.

But I guess it's too much to hope for people arguing on the internet not to overstate the strength of their positions.

Ben Smith

For the second chart, I would make the first column commuter and long distance at 3/4+ mile interstation, rapid at 1/2-3/4 mile interstation, and local at under 1/2 mile interstation. These numbers are averages, and could be more or less depending on urban structure and density.

Alon Levy

@EngineerScotty: you should collapse A and A+ to one class, and the same for C and C-. Automation does not always represent a higher class of service: in Singapore, automated people movers feed non-automated subway lines. And trolleybuses are not a lower class of service than buses; in the BMT's classification they were a higher class of service, and have acceleration and pollution advantages.

In Brisbane

"When you think about transit technologies, how do you categorize them? And why?"

Hmm... I have re-read this and actually, the ROW and stop spacing really have nothing to do with technologies at all. Buses can run in Class A, B and C and so can rail. So this is more of a service characterisation than a characterisation of technologies.

The technology could be flying saucer or magic carpet for all we care. But if it is in Class C, whatever it is is going to get stuck in congestion and provide bad service.

Jarrett at HumanTransit.org

In Brisbane:  That's my whole, and only, point.

In Brisbane

Good stuff!
I KNEW you would say that!



"Note too that the main virtue of busways in Class A-B situations is that you don’t need to build the entire line to all the destinations you want to serve. Busways can be implemented incrementally,"

And they can be de-implemented incrementally too, while rail right-of-ways cannot (or cannot anywhere nearly as easily). Again, you de-emphasize the most critical disadvantage of buses in practice in the 1st world - it's easier for motorists to try to incrementally grab some of that space back.


Why not add a row for cost? Roughly:
A 5 M€/km
B 10 M€/km
C 50 M€/km

Or something like that. Complete grade separation costs a lot. And I don't see why it would matter whether we are talking rail or bus.

This BTW is why, and I don't know anything about Vancouver so in abstract, you and Condon are having a non-discussion IMO. Basic tram costs a fraction of what fully grade separated system would cost and as such they are not mutually exclusive. And of course it is possible to have the tram use that A class track in some segments and C class in others, actually that would be ideal.

Dan W

The missing factor I would add here is, of course, cost.

There is the capital costs of constructing infrastructure, and then subsequent operating costs per passenger.

Jarrett at HumanTransit.org

Again, there are many missing factors.  My point here is the logical independence of many of the factors, from each other and from the bus-rail distinction.


The Runcorn Busway in north west England is a good example of a class B category BRT. Built 35 years ago, it is a 22km network of bus-only roads and was built to serve the new town from it's inception. Grade crossings with traffic light pre-emption, elevated approaches to town centre.

I think an advantage of exclusive bus right of ways is that they employ simple highway building techniques that are understood by all local authorities, and road building contractors, who the authority will be used to dealing with can be employed on construction. Rail on the other hand can be an unknown to areas without previous experience.

The capacity limitations of bus based rapid transit should not be a problem in smaller cities/medium towns where passenger flows are not overly heavy.

Runcorn Typical stop design:


Tour of the busway, at 0.17 you can see the town centre stop (with enclosed shelter) on the elevated section. At 1.19 a typical grade crossing.



It is a good analysis of technical concerns, but classes A B or C do not show how the design works socially, architecturally, aesthetically, or as an economic development tool.
How does this interact with humans and with existing cities?
As you describe the differences, it is obvious you could in theory have a bus that performs all of the functions.
For instance, here in St. Louis there is a regional planning agency that pretty well takes a one size fits all approach to transportation planning. It is hard to distinguish the city transit from suburban transit. The transit design does not take into account advantages of denser, older city areas and the especially the possible redevelopment of them. (Hong Kong uses redevelopment to help pay for the system)

Just as important is the impact of layers (whether it be a mix or bus only). This means different ways to arrive at the same destination. Subways for instance tend to be rapid and have fewer stops as they connect to the surface. Travel options help make transit successful, thus variations of A to C work together as a unit as needed and not separately. Hong Kong has many layers and is a good example of this. (Should bikes, taxis, walking, and other options be included?)
The creation of public space, the design of commercial centers and proximity of housing are some major factors in the preference from A to C.
If transit works hand in hand with cities and regions to help them prosper, A to C must relate to the design of the city and region.
Getting back to St. Louis, the city proper transit is typical unimaginative bus lines with several train lines running to suburbs (regionalism).
Adding a layer in the city proper, for arguments sake, let's say a rapid transit bus circle, going quickly from monument, to market to public space, to downtown and so on, essentially using transit to enhance older public spaces and creating new ones. At the same time this establishes an alternate route with the use of visible landmarks to create user identity with transit supported activities along the way.
The decision picking A to C becomes intertwined with other values in the society.
The rapid bus circle would be cost effective in this instance.

Still the ideal solution should be looked at before cost and political considerations.
Otherwise the danger is to be sucked into designing a system based only on technical concerns.

A to C is a great toolbox, except it is hard to understand how decisions can be made without strategic goals for the city/region. Transit is a tool of city building. A tool many regions ignore and it is why their transit systems struggle and are often not valued by the public at large.

Mr Wizard

Another dimension of the matrix that is, in fact, NOT determined by the technology but seems to be is: comfort. That is, rail = newer, actual seat cushions, air conditioning that works; bus = old, smelly, very noisy, welfare class of passengers, and knock your teeth loose jarring on potholes. These things greatly affect the willingness of people to use the service -- at least those with options.


You'd think that the basic highway building aspects of a busway are "simple", but unless you've dealt with having lots of bus traffic, they're not. Buses tear up pavement much more than cars, which is why things like concrete pads at bus stops are necessary. The Orange Line busway in LA had to have sections repaved after something like two years due to poor design.

And then there's cities that just can't maintain their streets. I was in San Francisco a couple days ago, and as I was riding the 49 bus on Mission Street, it felt like the main limiting factor on its speed was not traffic or double-parked cars or poorly-timed lights, but pavement so awful that the bus could only go about 15 mph lest the bus and its passengers be rattled to bits then and there.


Having read all the comments (which are mostly both interesting and valid), "the answer" comes back to what you are trying to highlight.

I agree with In Brisbane, that you seem to be trying to highlight the relationship between the right of way configuration (e.g. grade separated) and resultant service performance. If you are focused on this then you should be able to highlight the resultant SERVICE performance (capacity, reliability, frequency) that are possible from different rights of way. You may note the differences between different technologies at this point (e.g. rail vs bus frequency).

Much of the discussion seems to be the other way around, that is, what mode technologies can deliver certain levels of service which is fine, if this is what you are trying to highlight, but the focus on the right of way configuration is then lost.

Finally, IMO it is important to be consistent between using the theoretical capability and actual implementations apart. Great discussion though :)

In Brisbane

I think theoretical capacity is unhelpful. The practical capacity has more to do with how the system is designed rather than what it is.

A car lane on a freeway can carry 2000 people/hour. All public transport modes have far more capacity, far in excess of this number. And yet 80% or so of trips in Brisbane is done by car- how? You just design the freeway to have six lanes and voila- you have LRT/BRT-like capacity of 12 000 persons/direction/hour on a freeway using car.

Theoretically, there is no reason why you can't have 4 lane busways (Bogota?) or 4 track railways (Tokyo Yamanote Line??). So I think it should be called UNIT theoretical capacity- to point out that you can get higher capacity if you design the system that way.

Jonathan Parker

"...the ROW and stop spacing really have nothing to do with technologies at all. Buses can run in Class A, B and C and so can rail. So this is more of a service characterisation than a characterisation of technologies."

In Brisbane, this is not true for heavy rail/metro systems, which needs fully exclusive type A ROW for its operation for safety reasons due to the electrified third rail. However, it's the only technology that I can think of which has this rigid ROW requirement. I think in some ways this rigidity (some might say purity?) is part of the reason (also capacity) that heavy rail enjoys "top billing" in the transit technology hierarchy. Of course, there are only so may places on the planet where that type of capacity can be justified.

Alon Levy

@In Brisbane: yes, practical capacity matters. But you should distinguish a few flavors of it.

1. Two tracks or lanes are standard, pretty much everywhere. More are warranted only when there's a severe capacity issue - for example, in Bogota and Tokyo - and this leads to a proportional increase in costs. (By the way, the Yamanote Line is two-tracked, but for nearly the entire route, it runs next to either the Keihin-Tohoku Line or the Saikyo + Shonan-Shinjuku Lines).

2. Some places report practical vs. theoretical capacity based on real-world vs. theoretical capabilities, in which case the practical figure is the only one that matters. I don't recall seeing this in official agency publications, though.

3. Sometimes, theoretical vs. practical capacity is an issue of rolling stock availability. In other words, the platforms may be long enough for longer trains, or the signals may allow shorter headways, but current equipment imposes a lower limit. To go beyond practical capacity would require spending more money on rolling stock, which typically costs far less than new urban infrastructure. This is the issue for Calgary Transit, as well as the Washington Metro's Red Line.

4. Many travel patterns don't come close to capacity. This includes suburb-to-suburb car travel, and small-city buses.

4a. Cars have the advantage of flexibility, i.e. once the road network is paved, you can go from anywhere to anywhere without schedule coordination. When demand is far below capacity, transit can't compete. That said, it's a two-way feedback mechanism: a large transit network will encourage the growth of a dominant CBD, while a large freeway network will encourage the growth of low-density office parks.

4b. Jarrett's point about bus vs. rail is that in terms of service, buses can provide (almost) anything rail can, so merely replacing buses with rail will not lead to ridership increases. Therefore, he says, on routes that don't exhaust the capacity of buses, rail does not add much.

In Brisbane

I'm pretty sure that you can run heavy rail on class C right of way, if you really, really want to: (I wouldn't recommend it though).


And surely not all metro systems must have third rail. I think that is more to do with the design not mode. Is this "common agency practice" rather than an intrinsic limitation of the mode?

@ Alon Levy
"4b. Jarrett's point about bus vs. rail is that in terms of service, buses can provide (almost) anything rail can, so merely replacing buses with rail will not lead to ridership increases. Therefore, he says, on routes that don't exhaust the capacity of buses, rail does not add much."

I'm not sure about this- and TransitPlannerMunich seems to disagree on this too. Watch this video, a train has been placed down a freeway median busway in Perth and this is what happened to the patronage:


"It is carrying 50 000 passengers per day,
the bus system in that corridor was carrying 16 000 per day.
(3:50 minutes)" - Dr Peter Newman

Patronage went up, not down, and this is despite the train replacing the bus and introduction of "forced" interchange that I often hear decried on the basis of "transfer penalties".

In Brisbane

"In Brisbane, this is not true for heavy rail/metro systems, which needs fully exclusive type A ROW for its operation for safety reasons due to the electrified third rail."

Sorry, perhaps there was some misunderstanding on my part... please allow me to clarify: when I say "rail" I mean the family of rail-guided modes such as metro rail, light metro, light rail, streetcar, tram-trains all that kind of thing.

So rail in Class C would be streetcar/tram (maybe not in Zwickau?)

I think the grade separation is more to do with high frequency of metro systems rather than the electric rail. Think about it- if you have a car intersection and a train closing that intersection every ~ 3 minutes during peak hour, that intersection is going to be effectively closed because the trains are so frequent that the boom gate would be down for the whole of peak hour. That would be unacceptable, so grade-separation is done.

The same on busways in Brisbane- the buses run at very high frequency, so they are separated from the cars by bridges and tunnels (Class A ROW).


It's perfectly possible to have level crossings with third rail. People shouldn't be wandering on the right of way in the first place, so the incremental danger from the third rail is actually small.

Alon Levy

@In Brisbane: yes, there's plenty of disagreement on what the rail bias in ridership is. Jarrett argues that it's entirely cultural, and often dependent on service issues like level boarding, etc. TPM argues for very high figures - he said 43% in one example, and some regional services in Germany using good ROWs have more than doubled ridership after railstitution.

Full-fat rapid transit can have level crossings, which could be closed for very long periods of time if it's very busy - for example, the Chuo Line west of Mitaka. But the traffic disruption then justifies the expense of full grade separation, both economically (the same cost is spread across more users) and politically (the local NIMBYs would stand to gain from the project).

Class-C rapid transit doesn't exist anymore. There used to be some commuter lines running down streets, and there's still one Amtrak line that does that, but it's not normally done on modern transit systems. Even Class B is not very common on main lines.

In Brisbane

I think you can have Class C rapid transit, but it is rather rare.

In this case I think it is more to do with the stop spacing. In Brisbane we have the CityGlider rapid bus and it is claimed that this is 'Bus Rapid Transit' (and I agree with them on that)


It operates in Class C ROW (a short section is Class B at the single busway station it serves, The Cultural Centre Busway).


@In Brisbane:
I really like the Pearth-example...
Indeed I am a bit tired of the whole discussion. From the European perspective it is so totally retro, so "50ies" and "60ies"... let's do the time warp again! (I know that sounds terrible arrogant now, but that is probably what you expect anyways from a German engineer).

Excuse me, when I repeat myself now: even if Jarretts arguments would be theoretically totally right, for me it counts what I see in reality. People love rail and stay away from bus system (except people who cannot ride a car, they go by bus anyways). If it would be the other way around I would be the first person demanding to tear out the tracks (they make my engineering life anyways only more difficult)!

But even in cities like Bogota (the famous Transmilenio-BRT-system) they want now to built a rail system, noticing the limitations of bus. And the BRT system there has two bus lanes in each direction...

In Brisbane

TransitPlannerMunich, I think it is more to do with the lack of detailed published data/info on this phenomenon.

This could well be a cultural feedback effect or maybe just because of ride quality and increased legibility. It would be interesting to see if the same effect were operating on trolleybuses.


@Alon: The South Shore Line in Michigan City is Class C commuter rail, and the Capitol Corridor in Oakland arguably is as well. I would include lines that have to share with significant freight rail traffic as Class C as well, or at best class B; waiting at a junction is the same regardless of the type of vehicle you are waiting to pass.

@In Brisbane: If "rapid transit" can be applied to a bus in mixed traffic then it has lost all meaning. In Alon's post I believed he was using the "rapid transit" to mean "metro"/"heavy rail", as it meant historically.

Alon Levy

@In Brisbane, Anon256: yes, I'm using the more standard and historical terminology, in which unmarked, unqualified "rapid transit" is by definition not a bus or light rail; those would be bus rapid transit, or light rapid transit.

I didn't think of lines that have to share tracks with freight. However, I'd be weary of classifying any rail line as urban transit if it didn't have the sort of frequency that would force the tracks to be passenger-dedicated or at least passenger-primary. The South Shore Line has worse than hourly off-peak frequency, and the Capitol Corridor, which is the Amtrak line I was referring to, don't count. There's nothing that counts as mainline or commuter rail in North America that I'd be comfortable calling urban transit in the same category as CityRail, the RER, and any S-Bahn.

@TPM: I don't think it's arrogant, but there's wide disagreement on how big the rail bias factor is. The FTA thinks it's 20%, for what it's worth. Modern-day LRT ridership projections in the US tend to be unbiased - that is, they may not be accurate, but they aren't wrong in a consistent direction.


I am not trying to be ignorant here. I just don't understand how the specifics of an A to C transit choice without considering its relation to the physical urban planning can be possible.
Should there be another column for potential applications, high speed rail, fewer stops means large public square for instance? I should also mean more density.
Trolley is the form of a smaller public space and its focus. It means less density.
Every transit decision should have a balancing, architectural, urban design factor. I do not understand how it can be another way if transit is going to effective.
In the case of Perth, Dr Newman did not react to the physical impacts until the end of the first and into the second tape. All of the projects he cited were reactive, not proactive.
There were seven stops I believe, what if each one was used in some way to create and support public spaces, squares, plazas, centers or other points of social importance? Urban design is integral partner with transit planning. More important, creating and enhancing public space with transit becomes the tool to convince doubters of transits value.
Dr. Newman was lucky, the public supported transit. In America the constant propaganda machine precludes rational discussions, it is all about what corporations want.
A partial remedy is to present the transit system as a tool for physical and economic development. This requires the inclusion of planning concepts.

Rail is probably the inherently and instinctively best approach in many situations. Nevertheless the use of the bus as an experimental tool should not be overlooked. It means duplicating the rapid routes of subways or trains to better understand future permanent installations of rail(as an example).

Then there is the flexibility of the design. What kind of overlap should you expect if you have an A plus B plus C system? does that overlap help the system succeed, and why or why not?
I don't know, maybe there is a column D for cities and E for suburban uses and F for rural factors?

Ted King

@GMichaud - If you are looking for an example of overlap then take a look at San Francisco, CA, US. The city is served by an "A" (BART), a "B" (Caltrain), a hybrid "A+B+C" (SFMuni's Metro), and two heritage beasts - the F-Market+Wharves (a "B+C" hybrid) and the cable cars (a "C"). Great fun if you're a railfan but a headache for transit planners and operators.

How did we get into this mess ? Like other tangles (e.g. New York City) it's part evolution and part accretion. The roots of this mess go all the back to the 19th century. The Market Street RR got started in 1860. The cable cars got started in 1873. Passenger service on the peninsula (early version of sprawl) dates back to around 1870. So the pot has been cooking for around a hundred and fifty (150) years. "Mystery Meat" and "Mell-of-a-Hess" warning tags should be prominently displayed on that pot.

NB - Initial service dates are used in the above.



For Class B rail I would substitute "some" in
"Parts of some European....tram systems" with "most". And for Class C I would say "some" European trams. Since the second world war virtually all Western European tram systems have been upgraded to majority running on reservation (PRW). This can be grassed reservations in medians:
Or paved like the Dutch "trambaans" (dedicated tram lane with paved surface slightly raised above adjacent traffic lanes):
All the new generation of French tramways are virtually 100% in priority lanes:

Alon Levy

@GMichaud: the A/B/C classification is intended for urban transit, and breaks down otherwise. It's not really appropriate for intercity rail, where you'd ask about average speed and reliability first.


Transit activism of a technical nature in Detroit, Lego style: http://www.youtube.com/watch?v=egc_MwazUWo (thanks to fellow PTer Chris Smith for the link).


@Engineer Scotty, I watched the film, and have been thinking. Even bolts holding track are social. The shear, bending strength and so on are to support the train of course, but ultimately it is to prevent accidents so riders are not injured by a failure.
In the same way, the film talks sparingly about the social implications of their technical description. (As you point out)
The trains running down the center are described as being able to avoid the congestion of trains along the curb. Even though implied, it also means riders can get to their destinations in a timely fashion,make connections, be somewhere at a predictable time.
I am thinking out loud now, but what if the central tracks formed a plaza in the shape of a large roundabout that does not connect. (There is vacant and underutilized land in old cities, not sure about here, it has been years since I have been to Detroit)
A plaza could contain movable vendors, events, or permanent buildings.
This is only technically possible with tracks down the center.
In summary here is what this means to me.
1. Technical problems all have a social connection and meanings; sometimes minor, but also major.
2. Technical decision making is effective if the social connection is taken into account. (The film for instance, talking more on how the central route benefits riders in a way that helps public acceptance.)
3. City planning and the social design of the problem can led to technical solutions.

Daniel Sparing

I agree that third rail is only an illustration but not part of the definition. third rail level crossings were already mentioned -- to give the other example, Barcelona metro (completely underground and segregated) uses pantographs.

As for the Zwickau TramTrain - stadtbahns and tramtrains are not a new class but lines with sections of different class (A-B and A-C, respectively, where we don't distinguish heavy rail A and metro A now).


Here's the problem: if you're building a Class A or Class B system, spending money on a busway is totally wasteful, and *you want rail*.

The buses are lower capacity (wasting your investment in a high-capacity system), they wear out faster, the roadway wears out faster than tracks wear out, you need more ROW than you would need for rail, etc.

So yeah, your columnar analysis is good, and natural, but the fact is:

CLASS A AND B SYSTEMS ARE RAIL SYSTEMS. Attempting to use buses for class A and B systems is simply an ERROR.

Ottawa discovered this the hard way, by building a busway and then ripping it out for a railway.

So your columns in the first example show us that there are two types of systems:
(1) Rail;
(2) Class C, which might be rail or might be bus.

What does this say? This says that rail is inherently superior for most applications!


"Another dimension of the matrix that is, in fact, NOT determined by the technology but seems to be is: comfort."

Yeah, except that, again, it is determined by the technology.

I've ridden in brand new buses with all the mod-cons, and I've ridden in old broken-down trains on very badly maintained track.

Which was more comfortable? Hint: I get motion sick in buses and not in trains. So it was the train.

These technology-neutral evaluations are great! But if you do them honestly, you conclude that you usually want trains. If Jarrett wishes to point out that Lind and Condon are viewing things from a different point of view, that's fine.

But even from Jarrett's point of view, he should conclude that *trains are better*, simply because the columnar choices drive you towards trains most of the time. Only if you're going for the cheap, low-speed, low-capacity, low-frequency option do they point towards buses. (If you consider other technologies, sure, sometimes the columns point towards bikes, walking, gondolas, or even private cars.)

The fact that he hasn't concluded that *trains are better than buses* makes me suspect that something funny is going on in his head.


Jarrett: "I don’t agree that rail is always the right answer in Class A and Class B settings, partly because I think there should be a lot more Class B, reaching down to lower levels of ridership than are needed to justify rail."

You haven't done your homework. If you're simply taking an existing road and restriping it, in that case, sure, I agree, Class B bus lanes are great.

However, if you actually have to BUILD a lane, class B service with a bus is more expensive to build, maintain and operate than class B service with rail. Period. Do your homework.

"Note too that the main virtue of busways in Class A-B situations is that you don’t need to build the entire line to all the destinations you want to serve."

This is a great virtue of railways in Class A-B situations, too, in countries which didn't rip out their railway networks; you don't need to build the entire line, you can just run your tram onto the existing freight tracks.

Again, all you are saying is "there is value to leveraging whatever already exists". This is true, but it does not mean what you seem to think it means; you seem to think it provides an inherent advantage for buses, and it doesn't.

For instance, applying this principle *correctly*, the Pittsburgh East Busway would NEVER have been built, as it involved ripping out existing railway tracks in order to lay asphalt -- a total failure to recognize that leveraging what already exists is the way to save money.


Jarrett: "My point here is the logical independence of many of the factors, from each other and from the bus-rail distinction."

And I guess *my* point is that although these factors are logically independent (you're right, they are), they are *not* actually independent of the bus-rail distinction.

The technical differences between bus and rail mean that rail is *consistently* appropriate and cost-effective for providing some of these factors while bus is consistently *not*.


....while in contrast, the supposed "flexibility" of buses in going onto non-separated roads, which Jarrett touts as fundamental, is not fundamental to the technology at all. It is something which depends on the installed base which exists in some cities and not others.

In those others, trains may have the "flexibility" of going onto tracks already in use by other trains, while buses can't.

This leaves you with a question: build further on your installed base (cheaper) or change it? There's one good reason to change it: if it's going to become too expensive to maintain. Guess what: as oil prices go up, asphalt is going to become too expensive to maintain. So.... there is actually a good reason *not* to expand on the installed base of roads.



A partial rebuttal, over at Portland Transport.

The comments to this entry are closed.

the firm

Jarrett is now in ...

Related Posts Plugin for WordPress, Blogger...
Related Posts Plugin for WordPress, Blogger...