Commenter Mike recently laid out a nice explanation of the line numbering system in Aachen, Germany, and then asked, fatefully:
How do professionals assign line numbers?
The answer is: Much as geeky amateurs do, when drawing imaginary networks. It’s a process of (1) imagining beautiful systems of order, and (2) willing them in to being. Unfortunately, real-world professionals have to proceed through the additional steps of (3) clashing with proponents of competing systems, (4) enduring the derision and sabotage of anarchists, and finally (5) resigning to a messy outcome where only traces of beauty remain, visible “between the lines” so to speak, for those still capable of enchantment.
All this is visible, for example, in a slice of the bus network in San Francisco:
Look at the numbers of the east-west lines, from top to bottom, focusing on the right part of the image between Fillmore and Van Ness avenues, where the pattern is clearest. The sequence is: 1, 2, 3, 38, 38L, 31, 5, 21, 6, 71. Here, obviously, is a kind of Parthenon of line numbering, a ruined but still recognizable system of order: At one time, starting with Line 1 and proceeding south, there was a series of lines numbered 1, 2, 3, 4, 5, 6, 7, in more-or-less geographical order. The numbers 31, 38 and 71 were added later, by those anarchists I mentioned. (UPDATE 2015: With Neoclassical visionaries back in charge, line 71 has been renumbered back to 7, restoring a bit of the previous order.)
The number 21 is a trace of a different system of order. Originally, the one-digit numbers referred to lines that flow into downtown, mostly along Market St., while the 20-series referred to lines that cross Market, generally running perpendicular to the first group. The north-south lines 22 and 24 in this image still tell that story, and as I understand it, the 21 used to flow across Market, though it was revised long ago to flow into Market and thus ruin the beautiful pattern.
Line numbering, in short, is really a dialogue between three impulses:
- Grand Synthesizing Visionaries, who imagine schemes where each number will not just refer to a line, but reveal its exact position and role in the network. For example, these visionaries may think up schemes that recall the patterns of numbered streets and avenues in many North American cities, or the similar numbering of the US Interstate system.
- Anarchists, who need a number for a new line, don’t care about the vision, and pick whatever number comes to mind.
- Conservatives, who believe that once a line number is assigned it should never be changed, no matter how offensive it may be to the Visionaries. Conservatives are responsible for the permanence of various reckless numberings made by Anarchists over the years.
Of course, there are really four or more characters in the dialogue, because there’s usually more than one Visionary, and Visionaries, by their nature, prefer their own visions to other people’s.
The most common vision of line numbering is to categorize lines by location. In small networks of lines all radiating from a point, it’s common to see numbers assigned sequentially going around the circle. In Portland, for example, these radial lines used to be numbered clockwise starting with North Portland, and very astute eyes can still see traces of that largely ruined pattern.
Networks that have always been grids will sometimes be numbered according to the grid pattern. Thus, for example, a quick glance at the Las Vegas network map shows that the east-west lines, in order from south to north, are 201, 202, 203, 213, 204, 206, 207, 215, 208, 209, 210, 211, 218, 219 — a reasonable effort to hold back the anarchists.
I started out life happy to number lines in geographical order, but over time I’ve realized that people need to understand what kind of service a line provides even more than they need to know where it goes. So I generally advocate line numbering systems that reflect crucial distinctions in either
- frequency and span (is the service running when I need it?) or
- rapid vs local stopping pattern (is the service designed to be ridden long distances or short ones?)
For example, I always recommend a numbering scheme for peak-only commuter express services that distinguishes them from any all-day services in the same area, because peak service tends to be more complex than all-day service and can therefore tend to obscure it, whether on a map or on a numerical list of lines.
But above all, line numbering is a lesson in the impermanence of all things, and especially of visions of the perfectly ordered city.
Any nominations for the best, and worst, line numbering systems? I am currently tussling with that in Dubai, which I am sure is a long way off the worst in the world but shows much conservatism and anarchy and not much vision.
Too many letter prefixes and suffices, some lines with both a prefix and a suffix (ugh!), and Express lines which sometimes have an X suffix, sometimes an X prefix, and sometimes neither …
The system in Brisbane, Australia uses a “regional” system, in which the first digit of a three-digit number indicates the region a particular route serves, regardless of whether it’s all-stops, limited-stops, express, full-time or peak only. Besides not indicating what type of service a route is, such a system doesn’t cater for through-routing either.
I think Hong Kong has the most complicated system, if it even has a system at all. A mess of bus network companies, many different letter prefixes, it’s all quite messy.
Vancouver has system also based on the region. “1-20” are electric trolley buses, which reside in the City of Vancouver. “1-80” are buses in Vancouver, where the number generally represents the street it primarily uses, i.e. 49 bus on 49th Avenue. “90s” are B-Line buses (express buses). Then for triple digit numbers, the first digit represents the city/region it travels in i.e. 100s for Burnaby or 400s for Richmond.
How often to big cities really get to remake their numbering systems, since most will have had services running continuously for more than a century at this point? Many of Baltimore’s bus routes — which follow no discernable numbering system, really — have the same numbers as the streetcar lines they replaced in the 1950s.
How one numbers routes depends really on whether you want the number to carry some meaning beyond a shorthand for the name of the route. Jarrett advocates frequency or stopping pattern, yet many places with any systematic numbering use geography.
Examples from Ontario, Canada:
* York Region groups its numbers by type (1-99 = regular routes, 100s = routes going into Toronto, 200s = shuttles, 300s = expresses, 400s = school runs, 500s = community buses, 600s= BRT routes).
* Durham Region nominally groups its line numbers based on the municipality served (with cross-region routes getting their own group), although really they are based on which of its predecessor system the route was inherited from.
* GO Transit (interegional operator) roughly groups its route numbers by corridor.
* Toronto uses 1-4 interanlly for subway routes, 5-169 for regular routes (with express routes numbers vaguely related to their ‘normal’ version), the 190s for shuttles, 300s for nightbus routes, and 500s for streetcar routes.
* Other systems in southern Ontario seem to just number haphazrdly.
Same with DC – many of the bus lines carry the same numbers as the streetcar lines they replaced.
What makes DC’s system even worse is that the numbering systems are still in place from the various different transit agencies, each with their own system, that were combined together to form Capital Transit.
Since then, of course, there have been lots of service changes. However, it’s remarkable that some bus routes haven’t changed at all from the streetcars they replaced – the 42 Bus, for example, still follows almost the exact same route as the 42 Streetcar – up Connecticut Ave from downtown, through Dupont Circle, up Columbia Road into Adams Morgan before turning around in Mount Pleasant.
Some explanation of the line numberings:
http://www.washingtonpost.com/wp-dyn/content/article/2007/12/29/AR2007122901438.html
“There were four private companies before Metro, privately owned,” Erion said. “Each had its own numbering system.”
D.C. Transit, which operated in the District, Montgomery County and a little piece of Prince George’s, used two naming systems. The old streetcar lines, later replaced by buses, had numbers with no letters. Bus routes were identified by a letter and a number.
There was some relation to geography. The streetcar lines started their numbers in Northwest Washington and swept clockwise around D.C. Transit’s territory. The bus route names went the opposite way, starting in Southeast Washington and sweeping counterclockwise.
Many lines will have various spurs in outlying areas, but will share a common trunk route through the core of the city – these are often grouped and signed together – i.e the 30s (32, 34, 36, and 39) series all follow the same route through downtown along Pennsylvania Ave, the S series (S1, S2, S4, S9) all use 16th Street, etc. Some trunk lines are stronger than others, however. The D routes run crosstown and don’t really share much of a common segment at all, for example.
So, as long as you know you’ll be traveling along that trunk segment of the line, you can grab any one of those buses. The savvy rider can use that as a kind of indication for service frequency, but it’s not entirely obvious and requires knowledge of the system.
The other thing is the addition of Metro’s limited stop services. To date, Metro’s tried to identify these by making sure they all end in ‘9’ – the 39, 79, S9, etc. This is good, but it’s not perfect. There is one express service that also shares the 30s line, but obviously can’t also be the 39, so it’s the 37. On the other side, there’s the 89 route which isn’t an express service at all.
A visionary example is Seoul’s new (since 2004) bus colour codes and numbering system (which went with wholesale reforms of the regulatory arrangements). See http://www.lifeinkorea.com/information/trans/seoul-trans.cfm#Numbering for more on the numbering.
“… each bus in the municipal transit system will now employ one of four colors: blue, green, red or yellow, depending on its function. Now, passengers can easily identify the starting point and the destination of a bus at a single glance. The entire area of Seoul is divided into eight zones, which are clearly indicated in the numerical designation of each bus.”
I don’t know if the Anarchists have started to mess this up yet.
Maybe you can do a post on line styles and colors for system maps too? Or maybe something on creating a simple map of the downtown area, which is usually challenging due to the huge number of routes that go there.
A good question asked in another thread, but worth repeating here: Is it wise for similar (but different) routes to have similar numbers; especially when the number differs only by a letter suffix? Such may imply that one service substitutes for another…
My preference, if any, is for different routes to have different base numbers; with letters used to indicate different levels (or times) or service–X for an express, etcetera. The practice in many places of using letters for branches annoys me; especially if you want to get somewhere that’s out on a branch.
Really, the important question is: What information should be communicated by a route number? Should they be no different from random numbers–serving only to distinguish one route from another? Or should the inform the user of additional details?
Hey Paul,
Seoul’s system is still the same. Visited in the winter.
But I think there are some bus routes that have two or three spurs or something because I’ve seen bus numbers like 110A and 110B…
But the system remains intact.
Los Angeles, whose bus lines originally had the number/letter system of the old streetcar and interurban systems, underwent a “Great Renumbering” in the early 1980’s. Since then, MTA routes (and former RTD routes, now part of other operators such as LADOT or Foothill) are numbered as follows:
1-99: All stop local routes radiating from Downtown Los Angeles. The route numbers sweep counter-clockwise from Downtown:
1-10: Hollywood area
14-20: Mid-City to Wilshire Bl. corridor
25-39: South of Wilshire (Olympic, Pico, Adams, etc.)
40-44: South West LA (Crenshaw, South Bay, etc.)
45-59: South LA (Watts, Compton, etc)
60: Long Beach Bl
61-69: East LA, Southeast LA County
70-79: San Gabriel Valley
80-89: North LA (Eagle Rock , Glendale, etc.)
90-99: San Fernando Valley Rd. corridor
Routes numbered 100-199 are east-west crosstown routes (not serving Downtown). Such routes south of Central Los Angeles are generally numbered between 100 and 149; the San Fernando Valley gets 150-169; and the San Gabriel Valley gets 170-190.
Routes numbered 200-299 are north-south crosstown routes (again, not serving Downtown).
200-220 are for the area west of Downtown; 221-249 for the San Fernando Valley, and 250-299 for the routes through areas east of Downtown, such as East LA, Southeastern LA County, and the San Gabriel Valley.
Lines in the 300 and 700 series are respectively limited-stop and Rapid Bus (a special case of limited-stop, really). Most of these are numbered by adding 300 or 700 to the local (1-99) bus service number, or changing the first number of the crosstown bus route from 1 or 2 to 3 or 7….
400’s are freeway express routes, again allocated on a counter-clockwise bases (400’s and 410’s going north of LA, 420’s and 430’s to the Westside, 440’s to the South Bay, etc. ending with 480’s and 490’s to the San Gabriel Valley). Most of these are now run by LADOT or Foothill not MTA.
500’s are “intersuburban” express routes that do not serve Downtown. There were never that many of these to begin with, and their route numbering doesn’t really follow any particular rhyme or reason.
600’s are “special” routes; mostly short-distance shuttles or special event lines (Hollywood Bowl, etc.)
The rail lines (Red, Blue, Green and Gold) have internal line numbers in the 800’s, but these do not appear on the vehicles.
Of course, there are a few exceptions here and there, mostly due to through-routing or other issues too complicated to get into right now. But look at the MTA maps and schedules:
http://www.metro.net/around/maps/
It seems that the four most common colors for colored transit lines (usually rapid transit; because bus lines swiftly become too numerous to demarcate with colors), are the primary ones: Red, blue, yellow; oh, and green (a primary color of light though not of pigmentation).
When MAX opened its second line in 2001, necessitating some way of distinguishing between lines; the original line (including its Westside extension) became Blue; the airport line, Red. The third line–Interstate (2004)–is Yellow; the fourth, from PSU to Clackamas (2009), Green. The Milwaukie MAX line, scheduled to break ground next year, is labelled on most maps in orange; though it still is not clear whether it will be a fifth distinct line, or an extension of the Yellow or Green lines (the Yellow makes far more sense than the Green).
Right now, the current Portland Streetcar route is officially the Amber line (though it isn’t presently signed as such); when the new Streetcar Loop opens, it will be the Aqua line. Plan is to name Streetcar lines after gemstone colors–a development which cause Streetcar advocate Chris Smith to jokingly complain that “MAX gets all the good colors”.
We are starting to get the push now in Los Angeles for train service to get letter names instead of color names. With the 30/10 initiative gaining traction, it seems that we might actually have mutiple rail services here within ten years, that are too numerous for color designations.
The colors can stay on the maps but we need letters for the services. Colors for lines (tracks), letters for services (trains). Some people might think this is blind aping of New York City’s subway, but there is precedent in Los Angeles for this. The Los Angeles Railway (yellow streetcars) used letter designations for their routes. I forget how many they had, but I think it was over 20 routes at maximum.
We need to do this before we start being forced to use silly colors, whose names are not even translatable. Single letters for titles of services ARE translatable.
I agree. My rule: When your color-based numbering starts contemplating Aqua or Teal lines, you've run out of colors!
Thanks for the interesting article on route numbering! But I think you may be crediting Muni with more rationality than it really has, because I don’t think the 21 Hayes ever really was a crosstown route. The 1913 Report on the Improvement and Development of the Transportation Facilities has a map of the city’s transit as it was in 1864-5, and it looks like even then the Hayes line was a branch off of Market. Quite a bit later, but still a long time ago, here is a map from 1931 showing the 21 Hayes running to the Ferry Building.
There was a project at Muni in the early 1950s to rationalize and renumber the entire system. But it was never put into practice, and the numbering stayed mostly unchanged even when the “modified grid” rationalization was introduced in the early 1980s.
What’s the most number of colors used by a transit agency?
Chicago’s L uses 8: Red, Blue, Brown, Green, Orange, Purple, Pink, Yellow.
LA also has 8, counting the not-yet open Expo Line. It has Red and Purple (Metro), Blue, Gold, and Green (LRT), and Orange and Silver (BRT). The color for the Expo line has not yet been determined–black, aqua, and rose have all been suggested.
Of course, over in the UK there are 12 colours used on the London Underground. Brown, Red, Yellow, Green, Orange, Pink, Grey, Magenta, Black, Blue, Light Blue, and Turqoise–though it’s my understanding that while these colours are standardised; common practice is to refer to the lines by name (“Bakerloo”, “Picadilly”) and not by colour.
Any systems beat that?
Orange is no longer used by the London Underground (it’s now the color of the Overground, of which the East London Line is now part), and nobody ever refers to the lines by colors anyway. I think Chicago is the current winner in terms of the most colors in a purely color-based system. I think what will finally push LA from colors to letters is when the Regional Connector opens and through service becomes possible with more complex patterns. A plain pairing up of destinations northeast and southwest won’t work due to relative levels of demand during rush hour.
some observations…
AC Transit in the East Bay recently combined the routes of two lines, 7 and 9. The new line is numbered 49. Cute.
The same line is also a loop line. One direction is called “counterclockwise” the other “clockwise”. It’s been pretty confusing for everyone, especially since there is a small segment where both run in parallel.
On the topic of loop lines… the 49 runs once every half hour. However at certain points on the line it has an effective headway of 15 minutes if one is going to the opposite side of the loop, as one can take either direction.
In Taipei the loop lines are all prefixed with a ‘0’. Cute.
There are also some bus lines that run their entire route on a single street, and are named after that street, without a number. A friend of mine who couldn’t read Chinese said “This is a bus I would never be able to take.”
SamTrans (San Mateo County, California) has a fairly informative numbering system. Taking loosely from wikipedia, three digit designations are for routes with a rail connection; the first digit of these indicate what kind of rail connection it is. The second number (in rail connections) or first number (in non-rail connections) indicate which city the bus covers.
I’ve only used SamTrans a few times–and never since I learned what the numbers mean–but it seems very user-friendly.
I used to play around with fantasy numbering schemes for the NYC subway. I won’t go into details but the schemes I came up with were immensely pleasing to my sense of order.
These days, I tend to think that clever numbering schemes are a complete waste of time, because the average rider doesn’t have enough of a grasp of the complete system that would be required for the scheme to make any sense. The one exception I find somewhat useful is the common practice of using lower numbers for urban routes and higher numbers for the outer regions. Otherwise, I think the usefulness of encoding some sort of “meaning” in route numbers is highly overrated to anyone other than us geeks.
Interesting question for US-based readers:
The Interstate Highway System has a a fairly consistent numbering scheme. The main trunk routes have one or two digit numbers; for these, predominantly N-S highways have odd numbers and E-W highways have even numbers. Numbers of trunk routes generally increase as you move east or north. Principal Interstates have numbers divisible by 5 (such as I-5 or I-90); local or regional ones generally have other numbers (I-66 or I-17). Three digit numbers designate a spur or bypass route, and the lower two digits indicate the trunk that the 3-digit route is a spur or bypass of. An even numbered hundreds digit indicates a bypass or loop; an odd number indicates a spur. The even/odd rule for direction does not apply for spurs and bypasses.
While there are exceptions to the rules–quite a few places where highways are out of order, a few where they are grossly so, and the infamous case of I-238 south of Oakland, CA (there is no I-38 for it to branch off of)–in general, the scheme laid down nearly six decades ago is intact.
The question is–how many motorists know, or care? The numbering scheme is well documented in motorist literature–but the only people I know who care about this are roadgeeks. This is even moreso in the day where cars routinely have GPS navigation systems, and little preparation is needed for wayfinding. Roadgeeks get all bent out of shape about violations of the rules–there’s plenty of sites on the web dedicated to the sheer and unmitigated evil that is I-238 or I-99 (I kid you not)–but other than that; I’m not sure how useful the messaging properties of the Interstate numbering system are. To many motorists, highway numbers are like airline flight numbers–arbitrary designations which only serve to distinguish different things from each other.
Of course, navigating a city is different than navigating a country the size of the United States–but still; there may be a point to the argument that transit architects who obsess over line numbers are wasting their time.
I have never really paid much attention to number systems before, but there’s something very interesting about that sequence, and I’m probably only seeing it because I recently started a part-time job shelving books at a library.
1, 2, 3, 38, 38L, 31, 5, 21, 6, 71.
See it? With the exception of 21 and 31, it looks like a decimal system. In other words, if you imagine a decimal point in front of the number and you get an order (again, with the exception of 21 and 31). .3 goes before .38, which goes before .4, etc.
I’m not saying this has anything to do with why these numbers were actually assigned, but I couldn’t help but notice the pattern. I guess when the anarchist needed to come up with the number for a Geary line, they wanted to pick a two-digit number that started with “3”, perhaps to keep some semblance of the existing number pattern.
It’s a shame that “Pink” has been unofficially reserved for the future line thru West Hollywood, as it is a better name and more distinct color than “aqua” (or turquoise), the current provisional color for the Expo line in Los Angeles. If we had not used Silver and Orange on the BRT routes we would have a few more easily distinguished colors.
My wife, who is a linguists, tell me that not every language has every color, and additional colors tend to come in a certain order, according to a study of the world’s languages: [Most Common > Least Common]
White & Black [All languages!] > Red > Yellow = Green > Blue > Brown > Orange = Purple = Pink = Gray
Black and white tend to be reserved for background and text colors on maps, leaving Red, Yellow, Blue and Green as the best choices, followed bly Brown and Orange, with Purple, Pink and Gray as your last distinct choices. It is interesting that Pink is considered a distinct color by so many languages, but it does stand out well, much better than light yellow, and also better than light blue or light green.
I would suggest that the optimal system would use no more than 9 distinct colors for lines. The next choices would be light blue/green (Aque or Teal), and light green/yellow (Lime), and Black. After that, which gives 12 options, colors really start to blend together. It’s a good thing London only has 12.
One thing to remember with San Francisco is that it is based on two numbering schemes: the original Market Street Railway (numbers) and the San Francisco Municipal Railway (letters). (Cable cars originally were just named based on their route.)
When streetcar lines were abandoned, Muni replaced with numbers (B Geary became 38 Geary…in San Francisco, the vernacular is to use the line number and letter together). When MSRy went bankrupt, Muni took over almost all routes and kept the numbers in order to avoid confusion. There was some consolidation, especially following the war.
The cable car lines are numbered – 59/60 is Powell/Hyde/Mason (I can never remember which one is which) and 61 is California. In fact, what helped trigger the “save the cable cars” movement of the late 1950s-early 1960s was publication of planning maps that used those route numbers for “bustitution”.
Special owl routes (combinations of multiple routes) are numbered in the 90s (90 and 91 currently). Some other lines also run 24 hours, but they are not
If you have serious interest in the numbering scheme of Muni, contact the Western Railway Museum. They have extensive archives from both MSRy and Muni. Additional information can be found in the various Interurban Specials that discuss the history of transit in San Francisco.
In the East Bay, the old Key System used lettered routes for the rail lines, which is why the AC Transit Transbay routes are still lettered today. (Remember that originally the lower deck of the Bay Bridge served three rail lines – 50 Internet Points if you can name the other two).
Down south, check out the San Fernando Valley Transit History Page (do a Google Search) for an extensive history of bus routes, including number usage. For example, you can see that routes 14-16 at one time were routes that were taken over from Burbank Transit after the war, and routes in the upper 80s were originally Asbury Transit routes in North Hollywood and Burbank. Route numbers in the lower 80s were Pacific Electric routes.
@ Tom West
* Other systems in southern Ontario seem to just number haphazrdly.
Not So! Hamilton’s bus routes are numbered by geographic area
(not all numbers currently in use)
1-14 Lower City
15-19 Ancaster or Waterdown
20-39 Hamilton Mountain North-South
40-49 Hamilton Mountain East-West
50-59 Dundas or Stoney Creek
99 summer only waterfront shuttle
Ari at the http://amateurplanner.blogspot.com offered this comment:
A few comments on the comments:
1. The Chicago colors only came in to play in 1994—before that the lines were referred to by their termini or the neighborhoods they traversed, such as the Ravenswood Line or the Skokie Swift. As far as I can tell, no one calls the Skokie Swift the “Yellow Line.” It will probably take some time.
2. I took Samtrans once—when BART had shut down after my delayed plane got in, and, as I browsed the schedules waiting for the bus, found the numbering scheme made some sense. Some.
3. The Interstate system’s numbering scheme matters far less than a transit system because of the scale. While the numbers are generally useful, there is usually only one way to get from point A to point B. It’s a long-distance driving scale. In a city grid it wouldn’t make sense for the streets to be numbered 89th, 87th, 99th and 79th (in that order). But it’d be hard to mistake Central Pennsylvania for New York City. A transit system is far more human-scale, and having numbers make sense when lines are less than a mile apart (generally) is good for pedestrians and riders.
4. Then there’s Minnesota. The Twin Cities’ transit system developed out of each downtown, and while four lines ran between both cities, each city had its own numbering system. Until around 2000! This meant that there were two route 4s, and two route 10s, et cetera (lines never crossed, though). For example, this portion of an archived map of the system (from 2000) shows a section which has both Minneapolis route 22 AND Saint Paul route 22! Talk about confusing. (Although, wow, airport service has drastically improved. To Minneapolis, from a 40 minute ride with 20 minute headways to a 20 minute ride with 7.5-10 minute headways, to Saint Paul from a 25 minute ride with 30 minute headways to a 20 minute ride with 15 minute headways.) The switch was made during 2001, in which every Saint Paul route which didn’t share a number with a Minneapolis Route (the 16 and 21) was renumbered. Not all at once, actually, but pretty quickly.
For several years afterwards, however, bus schedules were marked “63 (old route 3)” and such. It’s possible that this was the largest wholesale overhaul of a transit numbering system in recent memory. Surprisingly, however, the numbers did not carry over from streetcar operations (Mpls, St Paul) although all streetcars were referred to by name (Selby-Lake, Como-Harriet, Minneapolis-Saint Paul) rather than number. But, boy, were those numbers sensible on he west side of Saint Paul where they ran 1, 2, 3, 4, 5, 6, 7, 8, 9, north to south.
CTTransit in Hartford, CT went through a complete changeover from letters to numbers in the last three years. I’m pretty sure it made the system less legible and harder to use.
Under the old system each trunk route had a letter designation and number suffixes referred to branches. Now all local routes have two digits and sometimes a letter suffix for a branch. Since there are fewer digits than letters, the first digit in the route number doesn’t necessarily indicate the trunk route. For instance, 60, 62, 64, and 66 are all Farmington Avenue routes, but 61 goes a different direction entirely. Under the old system, you knew an E bus would serve Farmington Avenue.
It seemed like they were using 90s for crosstown routes, but then they renamed the O Glastonbury (and branches O1, O2, and O3) as the 95 (with branches 95P, 95C and 95H). As you can see, they didn’t even cut down on the use of suffixes much.
There aren’t patterns for north-south and east-west, or even clockwise and counterclockwise. I’m at a loss to explain why they bothered with the whole scheme. CTTransit New Haven and Stamford still use letters.
I am sorry that I don’t have more time to talk about this — I am going out of town today.
There is a distinction between the kind of numbering system that is useful for people looking at the transit system as a whole, and the type of numbering system that is useful for most passengers.
Most people do not look at the transit system as a whole and try to figure out the way it is designed. They are interested in figuring out their particular trip and then taking it once they get there.
Numbers are primarily an aid to memory. Which bus am I supposed to take?
We had for years at AC Transit a situation where you’d show up at Fremont BART and your bus choices were 211, 212, 213, 214, 215, 216, 217, and 218. It made sense — everything going to Fremont BART began with 21. But if you actually lived in Fremont and mostly took those buses, that meant nothing to you. Instead, because those numbers are so similar — and people’s perception of numbers is logarithmic; 215 seems more like 216 than 5 seems like 6 — it was easy to show up at Fremont BART and not remember whether it was the 213 or 215 that they were supposed to take.
Additionally, because transit lines change all the time, any system you come up with today is probably going to end up being broken tomorrow. Something about a line changes and all of a sudden it’s mostly the same except not — a “north-south” route is given a substantial east-west section at one end, or a bus line that used to serve one area now ends up being extended to another, or whatever. So you then have to either change the numbers even though the bus line isn’t changing that much (which can be a costly process, not least in terms of public understanding — where did my 15 go?), or breaking the “sensible” system.
Moreover, most of the ways people want to lay out numbering “sensibly” leads to lots of exceptions. Odd numbers go east-west and even ones go north-south, but what happens to diagonal routes? Locals are 0-99 and expresses are 100-199, except due to budget cuts one of those expresses is really the local bus line for its area? (Like San Francisco’s 8X)
The most useful thing a bus line number can be is *different from all other bus lines in the area, so people can remember them*. Which is likely to cause more confusion — if the two lines running down your street are 15 and 16, or if they are 15 and 38? In which situation are people likelier to get on the wrong bus by accident?
Using transit is hard, and making a mistake has serious consequences — you’re stuck far away from where you need to be. We should be building information systems that make mistakes harder to make (and transit systems that make the consequences less serious, where we can).
Totally agree on all points…. On occasional visits to NJ I used to arrive at Port Authority Bus Terminal and never recall whether I was supposed to take the 163, 165, 169, etc.
I very much admire the L.A. “Great Renumbering” system we’ve had since the 1980s. It takes a few minutes to learn and for the most part, it’s very intuitive.
I’m a Visionary partisan myself, and I think any system that designs an intuitive route numbering framework shows that it has something of a grasp on service quality.
What’s intuitive? That all depends on the area, the number of routes and types of services.
Generally, a small service with fewer than a dozen routes won’t need any kind of numbering system. A system wouldn’t be needed.
Large services ideally should have some sort of system that denotes variety of lines (local, express, limited/rapid, non-downtown, etc.), and if possible, have an easy-to-understand pattern.
It could be like giving the bus number to the street number (6 runs along 6th Street, etc.) or even basing the line numbers on your city’s postal compass (the dividing lines of north, south, east and west directions).
I like a system suggested for Chicago’s CTA. Chicago has a famous wayfinding mechanism built into its street grid. Chicago’s postal compass is based at State and Madison, and generally streets eight blocks from this intersection are one mile.
Routes would be given a letter N, W, S or E. Then they would be given the first 1, 2 or 3 digits to correspond with the blocks.
Going west from downtown, you’d have W8 on Halsted, W16 on Ashland, W20 on Damen; north of downtown you’d have N8 on Chicago, N12 on Division and so on.
Here in New Orleans, numbers appear less important to riders than their names for bus lines due to our streetcar heritage. Thus people will think of the Broad as the Broad instead of the #94. This historical approach does present challenges when a line like Louisiana runs only on a portion of that street.
Hey great post
I barely come on your blog and recently i was browsing through Auto Shipping Network http://www.autoshippingnetwork.com/ website while i was finding a way to ship one of my automobile overseas and i happen to stumble upon your blog, great post, very informative. Cheers
I remember when several routes in Queens were renumbered to eliminate the use of letters, around the time of the Archer Avenue extension. Route Q3A became Q83, Q4A became Q84, and Q5A became Q85. We also changed some routes that came out of Brooklyn Division depots but mostly served Queens from B to Q – but sometimes the number was taken, so the number changed too (e.g., B53 became Q54). Several Staten Island 3-digit routes were renumbered heuristically (a fancy way of saying if you know the scheme, you’ll be able to remember the number – the best example being the crosstown Manhattan routes, the M86 is on 86th Street). My problem was that I could never remember the heuristics!
The similar-number problem happened recently in Seattle too. On 15th Ave NE at the UW, there were five local trolleybus routes: 7, 9, 43, 44, and 70. There was also a diesel route 48. Around five years ago, the 7 was split and this portion was renumbered to 49. I’d expected it to become the 6 because that number had earlier been vacated, but it didn’t. The 9 was simultaneously truncated so it didn’t come to this area any more.
It used to be easy to distinguish the 7, 43, and 48 two blocks away (and of course the 48 was diesel which made it easier). But now it’s hard to distinguish between 43 and 49, and more than once I’ve gotten onto the wrong bus when I didn’t look closely enough at the number.
An interesting article and follow-up comments, but I’ve often wondered if those of us trying to put order into transit systems forget the everyday rider. How many different routes do people use on a regular basis? One, two… maybe three at most? They only need to know the route numbers of the routes they take and the rest may be meaningless (Aaron Priven made a similar point above).
As a transit planner, I want to see a logical system so operations run smoothly, but as a rider I only need to keep track of my routes.
Of course, similar numbers serving the same stops will confuse a new rider. If I’m waiting for #22 and #22B comes by, can I take it instead? Maybe… but I also might end up missing my destination. Once you’ve cracked the system, however, and ridden it successfully a few times, that confusion should disappear.
“f I’m waiting for #22 and #22B comes by, can I take it instead? Maybe… but I also might end up missing my destination.”
A solution to that would be a timetable and route map at the bus stop. To encourage a passengers to view the network as a whole, you might add info about connecting routes along the way.
Joe, it’s the transit planner’s responsibility to come up with a framework for the entire system, not the rider.
If you frame it around the rider, and he or she only uses a handful of routes, make sure the system is easy enough to take away just the information needed.
Strive for “economy of information”. Think Twitter: Force your thoughts into 140 characters. A system that dumbed down to the lowest common denominator will end up being as confusing as one that tries to convey way too much information.
As an aside, I personally hate the use of number-letter combinations for route designations.
(One exception is the Chicago system I alluded to earlier, where the letter denotes a position relative to the street grid).
The reason why I hate routes like 22/22B is that more often than not, transit agencies fail to make clear what the suffix denotes and how major of a routing difference it is. Also, there’s no degree of consistency among suffixes, with the exception of letters L and X, which are generally known as limited or express routes.
Some cities will just use the alphabetical order. Other cities will give phonetic suffixes “33E or 33J” for Elm or Juniper street running, for instance.
Mostly, though, agencies don’t explicitly point out the differences and why they are relevant. The printed information and/or the bus stop and headsign signage doesn’t say all the changes.
The number-letter combo is a relic in the day of the modern electronic headsign. Letters were needed because it was much harder to draft a new Mylar rollsign for every change; the electronic sign solved that problem more than 30 years ago.
I’d challenge the assumption that all passengers just use one or two lines. In a city like San Francisco, a lot of people use transit as their primary method of transportation. Therefore, over time they use multiple lines. They probably use one or two lines most often, but periodically need to use other lines. At those points, the most easily understood line numbers are helpful to them. From my understanding of how the mind works, it’s easier to grasp a pattern than remember a random number table. This is where the value of organized numbers for the transit system overall merges with what’s best for individual passengers.
The conservatives, actually have a point, though they tend to overplay it. When an agency cycles rapidly through line numbers and route patterns, people get confused. When people get confused, they may stop riding. While you don’t want to keep a route number because you had it in 1928, there needs to be some recognition of the need for stability.
San Francisco, of course, being a good example of a transit system with no recognizable pattern to the route numbers.
Of course, it’s easier to recognize a pattern of numbers, but only if that pattern is both consistent and meaningful in context. The common “Lines 60-69 operate in Fredville” pattern is pretty useless once you get to Fredville and actually need to figure out which bus you want, which is usually a lot more important than figuring out which of all the buses in your system run in Fredville.
Moreover, most of the time some bus goes to Fredville and Sallyville both, and you have to decide which set of numbers it falls into. So all the buses in Fredville are between 60-69, except the 75. Or you run out of 60-69 numbers and then what do you do?
Geography is, in general, very complicated and hard to sum up in a designation that’s usually three or at most four characters.
As for conservatism — Is it the route numbers or the routes themselves people need stability for? Changing a route significantly but leaving the number the is a recipe for confusion. People will, naturally, assume the 16 bus they got on last week is the same one they’re getting on this week.
(Well, changing it *away* from something. Simply extending a route doesn’t cause any confusion because nobody will get on thinking it goes somewhere it doesn’t go.)