London Underground Platform Numbering

London Underground and its predecessors are 'Metro' type railways where simplicity and consistency of approach might be thought highly-valued parameters. What could be simpler, and lend itself to consistency of approach, than the mere numbering of platforms? The evidence suggests that despite the introduction of very simple rules in fact the corporate memory has failed, or that the people responsible for such matters don't feel such things are important. It would be interesting to know. Anyway, the author has set out below what is known aboutthe subject.

Why number platforms?

The obscure subject of platform numbering should not detain anybody for very long. However, it is a point of some slight interest that when railways began there were no numbers but at some point a need arose to differentiate one platform from another. The need seems to have been more urgent at terminal stations, once they had evolved beyond provision simply of an arrivals and a departure platform. Intermediate stations were usually simple affairs with no more than two platforms, each usually dealing with traffic in only one direction, enabling one to be distinguished from another by the simple expedient of referring to them as the ‘to such and such’ direction, though staff would use the official terminology ‘up’ or ‘down’, representing the ‘to London’ or ‘from London’ directions. Once stations became larger, with three or more platforms, or with complex train workings, then the need arose to distinguish one from another more readily.

There is a certain amount of evidence that main line railway practice at simple stations became to number the ‘up’ platform No 1 and ‘down’ platform No 2, but consistency is much less evident at larger stations and quite unreliable at termini.

Early platform numbering

On the Underground railways, most stations were of the simplest design and the train services suited the ‘to such and such’ method of directing traffic to the relevant platform. No definite evidence has been discovered that any Underground stations prior to about 1911 had platforms that were numbered, though there may have been isolated examples.

The District Railway was certainly using platform numbers by 1931, as a photograph of Earls Court shows. Signage at Putney Bridge also suggests platform numbers had already been used there for some years. Official documents also refer to platforms at Ealing Broadway being numbered by 1930. These are unlikely to be the only sites, and documents setting out lengths of platform accommodation in 1930 hint that platforms may already have been numbered at Whitechapel and Hounslow West. The 1932 reconstruction of the west end of the District resulted in Northfields and Acton Town being numbered. At the east end of the Piccadilly, Cockfosters was numbered when opened in 1933. High Street Kensington was certainly numbered by 1936. Edgware was certainly numbered prior to 1932 when it was enlarged, and probably had numbers from opening.

On the Metropolitan Railway the earliest use of platform numbers may well have been at Baker Street when the ‘main line’ platforms were rebuilt around 1912, though there is no indication that Circle Line platforms were then numbered. It is probable that Aldgate, Moorgate, Aldersgate (now Barbican), Farringdon and Kings Cross may have been early candidates. Stanmore was numbered when the branch opened in 1930 (but not intermediate stations), and certainly Uxbridge was numbered by 1930. Wembley Park was numbered by 1932. Edgware Road was certainly numbered when rebuilt in 1926.

One might infer from the evidence that terminal stations were the first to be given platform numbers, probably from the 1920s, as there was a clear need to be able to direct passengers to one of several platforms for their next train. This was followed by allocation of platform numbers to ‘complex’ stations on a ‘needs’ basis where there were multiple platforms on the same line, or where trains to the same destination could leave from more than one platform. It is at least probable that a few busy central London interchange stations were also numbered by the mid 1930s, but evidence is not yet forthcoming.

Need for numbering consistency

It seems that by the late 1930s there was some discussion within London Transport about the need for some consistency in numbering as additional stations were identified to receive numbers. By early 1939 specific instructions were set out, with some straightforward rules attached. Most stations comprise just two platforms. The rule is that platforms serving westbound or northbound trains are numbered ‘1’ and those serving eastbound or southbound trains are numbered ‘2’.

This system means that there is consistency throughout the network, and staff, knowing how the system works, may thereby be able to give helpful information to passengers wherever they might be going. The system is stated to have resulted from meetings during 1939 when it was agreed between the Assistant Publicity Officer (then responsible for information provision) and the Operating Manager (then responsible for station operations). It was then endorsed by the signal engineer, responsible for physical provision and maintenance of signage.

Although it is inferred from the correspondence that most underground stations hadn’t previously had platform numbers, unfortunately no distinction is made between those that already did, and those to which the instructions suggested needed to be equipped next.

It is clear that consistency in numbering was intended, but that this created a number of issues.

1. Where possible, simple stations were to be numbered 1 (NB or WB) and 2 (SB or EB).

2. It appears that true ‘island’ platforms were left out of the system, presumably on the basis there was only one platform and other signage provided sufficient information (this lack of numbering applied mainly to through stations where trains did not reverse).

3. Some stations already had platform numbers. Evidence suggests these were left alone, even where they did not conform to the system.

4. Interchange stations that were not already numbered were treated to a separate process. The book 'Rails Through the Clay' indicates that some large central London interchange stations had their platforms numbered in 1936 (and suggests the above system was used); this probably explains some of the gaps in the 1939 system.

5. There existed a number of multiple platforms (that were not interchanges) where the ‘simple’ system needed adaptation.

Complex stations

Dealing with the last problem in the above list first, the ‘rule’ appears to be that the extreme northbound (or westbound) platform was numbered ‘1’, and remaining platforms numbered ‘2’ upwards in logical sequence. This was explicitly set out at Finchley Road, Willesden Green, Neasden (1938), Wembley Park, Harrow-on-the-Hill, Mansion House, Kennington and High Street Kensington (noting the inner rail circle is regarded as Westbound). Many of these stations had already been numbered this way.

The rule was subsequently followed at: Tower Hill, Gloucester Road, Amersham, Chalfont & Latimer, Moor Park, South Ealing, Acton Town, Turnham Green, Ravenscourt Park, Hammersmith, Barons Court, Dagenham East, Plaistow, Uxbridge, Northfields, Cockfosters, Arnos Grove (observing that when built these last two stations were regarded as ‘Northbound/Southbound’ rather than 'Eastbound/Westbound' as they are today), High Barnet, Finchley Central, East Finchley, Golders Green, Edgware, Camden Town, Loughton, Leytonstone, White City. The stations referred to in these two paragraphs relate to virtually all instances where there are multiple platforms (other than odd examples already numbered).

The instructions to staff (covered by the Signal Engineers Instructions to Drawing Office Staff No 10/39, dated April 1939) sets out a specific platform numbering regime for interchange stations, where the above regime could not easily be applied. This raises a number of questions, not least of which is that there is some departure from the straightforward system already described. The schedule sets out the following scheme; for brevity the word ‘normal’ is used where numbering conforms with the rules set out above, and ‘as expected’ means it is a logical application of the rule, even where not explicitly stated:

Baker Street – Platforms 1-4, Met main (normal), 5 Circle EB, 6 Circle WB, 7 Bakerloo SB and 8 Bakerloo NB. In each case one might have expected both 5/6 and 7/8 to apply to tracks in other direction, but as already explained 5 and 6 follow the Met’s existing approach where lower number applied to the ‘up’ line. It is not clear the Bakerloo platforms were ever numbered thus, as provision had to be made for the new Stanmore branch SB platform. As far as it has been possible to ascertain, the two SB platforms were numbered 7 and 8, and the NB platform was 9.

Kings Cross – Platforms 1-4, Circle, 5 Piccadilly WB, 6 Piccadilly EB, 7 Northern NB and 8 Northern SB (all as expected, but note that Circle Line numbers are the opposite way from former Met practice). Platforms 2 and 3 though provided were never used. When the Victoria Line opened in 1968 the opportunity arose to renumber platform 4 as platform 2, releasing the numbers 3 and 4 for the new line.

Moorgate – Platforms 1 (EB Circle) and 2 (WB Circle), thence 3-8 (bay roads, working north), 9 NB Northern, 10 SB Northern, 11 NB Northern City and 12 SB Northern City, all as expected. However the tube platforms were not in fact numbered like this as earlier instructions dated with effect from 1 November 1937 indicate Northern City Line was numbered ahead of the Northern Line, at the lowest level. The 1937 Instructions require Northern City SB to be 9 and NB to be 10 and Northern Line SB to be 11 and NB to be 12. Whether this was actually carried out is hard to tell, but the 1939 instructions suggest platforms were actually numbered Northern City NB 9, SB 10, and Northern Line NB 11 and SB 12, the reverse of what was set out previously. After station reconstruction in the 1970s the Northern Line was renumbered 7 (NB) and 8(SB) in 1976, taking up what had become spare numbers. The fact that the upper station appears to follow earlier Met practice suggests station might have carried numbers already.

Liverpool Street - 1-3 Circle, but with No1 EB rather than the bay, in conformance with Met practice (probably numbered many years previously). Central 4 (Westbound) and 5 (Eastbound). This raises an issue about what the ‘rule’ actually was. In the case of east-west routes, was it that the westbound line should be an odd number, or the lower number of a pair?

Whitechapel – District platforms 1-4 normal, with 5 East London NB and 6 East London SB, as expected. As mentioned previously, these platforms seem to have been numbered many years previously, but the other way around (platform 1/2 being eastbound and 3/4 being westbound); the evidence suggests that the existing numbers were just left alone. Recent reconstruction has perpetuated the irregularity (1 east and 2 west).

Swiss Cottage – 1 and 2 Metropolitan Line (normal) and 3 and 4 Bakerloo Line (normal). The Bakerloo numbers were retained for many years after Metropolitan station closed and renumbered 1 and 2 only in the 1970s.

Monument (Bank) – District Line 1 and 2 (normal), Northern Line 3 and 4 (normal) and Central Line 5 and 6 (normal).

Holborn – 1 and 2 Central Line (normal), 3-5 Piccadilly Line numbered from 3 being Westbound as expected and branch platform is 5. A photo of December 1937 clearly shows the northbound (eastbound) platform numbered as 3, so this was presumably altered in 1939.

Tottenham Court Road – 1 and 2 Central Line (normal), 3 Northern Line SB and 4 Northern Line NB, the reverse of what is expected. The numbers today are 3 (NB) and 4 (SB), with no evidence of change. Either the instructions were not carried out or they were recognized as an error; the likelihood is that the platforms were already numbered and were simply not changed.

Oxford Circus – 1 and 2 Central Line (normal), 3 Bakerloo Line NB and 4 Bakerloo Line SB, as expected. In fact Bakerloo platforms were numbered the reverse of this (4 NB and 3 SB). When Victoria Line opened, Victoria Line took 6 (NB) and 5 (SB) also the reverse of what is expected. The likelihood is that the Bakerloo platforms were already numbered and were simply not changed.

Charing Cross – 1 and 2 District Line (normal), 3 and 4 Northern Line (normal), 5 and 6 Bakerloo Line (normal). (It appears these platforms were once numbered differently but were renumbered thuswise on 17th December 1939, though platform 5 was numbered such by 1931).

Waterloo - 1 and 2 Northern Line (normal), 3 and 4 Bakerloo Line (normal).

Elephant & Castle - 1 and 2 Northern Line (normal), 3 and 4 Bakerloo Line (normal).

Kennington – Platform 1 NB via Charing Cross, and remainder numbered through to 4 SB via Charing Cross, as expected. In fact, Kennington was numbered 1 NB via Charing Cross, 3 NB via City, 2 SB via Charing Cross and 4 SB via City. This was logical from an operating point of view, but rather unusual.

In the majority of cases the numbering is as expected, and to the extent any rule is obvious it is that the highest level platforms were numbered lowest.

It may be noted that the schemes was not entirely carried out as set out above, perhaps because of the onset of war or the feeling that the cost of changing signage was not warranted where platforms were already numbered.

A few stations are not on the list, notably:

Leicester Square where 1 and 2 are Piccadilly Line (normal) but 3 is Northern SB and 4 is Northern NB (reverse of that expected). This station was renumbered at some point as a 1934 photo shows platform 3 Piccadilly WB and 4 EB, as expected.

Piccadilly Circus where 1 and 2 are Bakerloo Line (normal) but 3 is Piccadilly EB and 4 is Piccadilly WB (reverse of that expected). These were numbered this way from 26 November 1939 but there is a hint in the instructions that they had been numbered in a different way from an earlier date. Having said that a 1934 mock up of station signage at Piccadilly uses the later numbering. More photos needed.

Earls Court The early numbering of District platforms has been noted, but a 1934 photo shows the Piccadilly numbered 5 (EB) and 6 (WB), the opposite to that expected. This endures today.

Euston where 1 and 2 are Northern via Charing Cross, and 3 and 4 are Northern via City (all as expected).

Old Street, as expected but with Northern City branch numbered after Northern Line.

Paddington where Bakerloo is numbered 3 and 4 as expected, Circle Line taking 1 and 2 (in accordance with Met practice).

Putney Bridge, numbered from 1-3 upwards from EB platform. Probably early numbering scheme.

Morden, numbered upwards from wall road on east side (Probably numbered from 1926)

Watford (Met). Renumbered 1 (NB line) and 2 (SB Line) 10 September 1951. Previously opposite way around (in accordance with Met practice).

Wood Lane, numbered such that loop (inner platform) numbered 1 and outer platform 2; not known which of the Ealing platforms were 3 and 4.

The schedule of numbering omits reference to the Hammersmith & City and Circle Lines, except as already referred to.

There is some evidence that platforms were already numbered in the old Metropolitan Railway system which followed main line practice, that is with the ‘up’ or London-bound platform taking the lowest number.

Under this scheme stations Hammersmith to Ladbroke Grove and Notting Hill Gate to Paddington Circle were numbered 1 (up line, or eastbound), with Edgware Road EB (northern track) numbered 1 and Baker Street No 5. The corresponding WB platforms were numbered 2 (4 in case of Edgware Road and 6 in case of Baker Street).

Other stations around north side of Circle that might reasonably have been expected to have platform numbers for many years owing to their complexity also follow this pattern, viz: Farringdon, Aldersgate, Moorgate, Liverpool Street and Aldgate. This clears up a number of inconsistencies in the data quoted earlier. If one takes this into account, plus the fact that some stations referred to earlier probably had historical numbers, then one can see that in the vast majority of cases the simple rules set out were followed.

Later position

In 1980 there were still some platforms not numbered (according to some official documentation).

Metropolitan Line

West Harrow

These were all later correctly numbered in conformance with the scheme.

District Line

Ealing Common
Chiswick Park
West Kensington

These were all later numbered in conformance with the scheme. When the main line stations east of Bow Road were taken over from BR in the 1960s they were all renumbered correctly.

Piccadilly Line

Hounslow East
Boston Manor
Sudbury Hill
Sudbury Town
Caledonian Road
Holloway Road
Manor House
Turnpike Lane
Wood Green
Bounds Green

Oakwood, Holloway Road, Sudbury Town and Sudbury Hill were correctly numbered 1 (westbound) etc, whilst Southgate, Bounds Green, Wood Green, Turnpike Lane, Manor House, Arsenal and Caledonian Road were numbered 1 (eastbound). The apparent reverse of the expected position north of Kings Cross is almost certainly related to the publicity people regarding the directions of traffic on this section as northbound and southbound until about 1990 when east and west was adopted. The numbering had therefore been entirely correct at the time, and only redenominating the line direction later has created what appears to be an inconsistency today. 

Northern Line

Angel (island Platform). This apparently remained unnumbered until platform was duplicated in the late 1990s, during which process they were numbered incorrectly, northbound being No 2.

Hammersmith & City Line

These were all numbered, but the platforms inherited from British Rail in 1970 at Royal Oak and Westbourne Park were renumbered correctly in accordance with the scheme, but are therefore out of kilter with all the others on the branch. The new station at Wood Lane, on the other hand, conforms with others on the section, but not with the ‘scheme’. Paddington suburban retained its BR numbers.

Victoria Line

Agrees with the general scheme, except as referred to.

General Rule for Platform Numbering as derived from above.

Platforms are numbered in geographical order beginning with the lowest number (usually 1) as the extreme WB or NB Line as case may be. Where there is more than one line then the upper level is numbered first, the other lines being numbered in increasing depth (but treating lines at approximately the same depth as a single level). Local exceptions are possible where this would be more logical to passengers (eg at Bank where Northern Line [3 and 4] was intermediate along the walking route between Central and District, and takes the intermediate numbers even though below Central Line).

What are platform numbers used for?

On the whole platform numbers have two primary purposes. First they are a fast way of explaining to passengers how to get to the platform they need. To say 'follow the signs to Platform 6, for example, is quicker than saying they should proceed to the Hammersmith & City Line westbound platform. This is important when staff are dealing with a constant steam of enquiries but the point is now somewhat blunted by the corporate design people insisting that platform numbers are only put on signs on the platforms themselves and at the final bifurcation point. To me this largely defeats their object.

The second obvious use is to pass messages to those occupying trains at large stations where the message has to be absorbed by those on the intended train and on no other. The obvious example might be 'All change platform 5' (a message I myself hear too frequently). This is all well-intentioned, and it is hard to see how else such messages can be passed from a station public address system. Unfortunately this really only works with regular passengers who are intimately familiar with platform numbering at the station concerned. All platform number signs are placed in positions where they can be seen by those entering the platform (and to a limited degree by those waiting on the platform) which makes them invisible to those on the trains who are expected to react to the message. The outcome is a selection of studied silence (this can't apply to me!), an invitation to get up and peer out to see what clues there might be, and only in the case of regulars, to obey. This has always seemed to me a very strange way to go about things and I have often thought that the platform number should appear on the larger station signs, at least at stations where station staff often feel impelled to pass messages over the public address system.

There is a third reason for numbering platforms, which is purely for the benefit of staff who need a concise, shorthand way of referring to platforms in reports and documents and the like, we are not really concerned about this.

So there we have it. My suggestions are therefore:

  • Be consistent about platform numbering, a perfectly good system exists;
  • Use the platform numbers and place them earlier in the signage chain;
  • At appropriate stations, put the platform numbers on the larger station signs where they can be seen from the trains.