In part 2 of The Past, Present and Future of Metropolitan Line Services we finished with details of frequencies for the Metropolitan Line once the Sub-Surface Lines upgrade was complete. With plans for a overview of the ultimately intended service pattern for all the Sub-Surface Lines in mind, we now look at the Circle Line.

Never Run Your Trains In A Circle!

In the 1960s and 1970s London Transport had a flourishing international consultancy arm which made money by advising other cities on on how to go about setting up and running a metro service with a particular emphasis on advising far eastern countries how to plan their fledging metros. Rumour has it their first bit of advice was always: Never, ever run your trains in a circle!

Given that London Transport knew that running trains in a circle was a seriously bad idea it begs the obvious question of why on earth didn’t it follow its own very sensible advice. With that in mind, in part 1 we look at the 19th and 20th century history of the Circle Line with an emphasis on the nature of its operation.

The Circle Line was created – reluctantly


Mansion House station in 1896 when the Circle Line had already been completed.
Note the reference to “Circle Train” and not “Inner Circle”.

In 2013 one hardly needs reminding that the Metropolitan Railway opened as far as Farringdon in 1863. On the south side of today’s Circle Line its bitter rival, the District Railway, had opened as far as the woefully misnamed Mansion House station, where it had a four platform terminus, as early as 1871. The station was located at the junction of Cannon Street and the fairly recently-built Queen Victoria Street. This was not the originally intended location for this station. It should not be hard to guess where that was.

It is difficult to imagine nowadays in the gloom of the current day Mansion House station that for thirteen years this station would have been a busy London terminus complete with short sidings on the western side to assist with locomotive changes. Although it originally had four platforms it only had three tracks into them (rather like current day Uxbridge and Cockfosters) although a fourth track was added later in its life and the platform layout changed.

Meanwhile the Metropolitan Railway too was pushing eastward. Moorgate, long planned, was reached in 1865 just two years after the original opening as far as a temporary station at Farringdon. Like Mansion House station, the multiple platforms suggest that the Metropolitan, having reached the city, was in no hurry to continue eastwards beyond Moorgate. Indeed, land and costs generally would start to get really expensive. It took a further 10 years to continue the short distance to Liverpool Street and a yet another additional year to get to Aldgate, the next station beyond Liverpool Street – again a station built with four platforms and giving the impression that it was intended to be the end of the line.

If it had been up to the two rivals things might have remained this way. A joint service from Mansion House to Aldgate was run and although it was referred to as the “Inner Circle” it was in fact more like a horseshoe. It was called “Inner” because at various times there was also a “Middle” and an “Outer” Circle service. Confusingly, when reading about Circle Line history, “Inner” and “Outer” were many years later used to refer by staff to anti-clockwise and clockwise services on the “Circle Line”. By then the other services no longer existed so there was no confusion amongst themselves as to what service was meant.

The two rival companies could have probably seen that completing the circle was going to be expensive hard work which was probably very convenient for passengers but not something that was going to bring in much additional revenue. The also probably did not want to progress with something that would extend the need for a high-degree of day-to-day co-operation. They may have even appreciated the difficulties that running such a service would cause.

The two companies may not have wanted a Circle Line but parliamentary committees in both 1863 and 1864 came out strongly in favour of the idea. Indeed, the success of getting bills through parliament to further the Metropolitan and District’s aspirations below highly contentious land in central London was probably possible in no small measure to parliament’s dream of seeing the circle completed.

When the two companies seemed unwilling to progress further a third independent company, the Inner Circle Completion Company, attempted to do the job instead. The stick to get the two unruly siblings to co-operate had been provided. The carrot would be an extension east to Whitechapel to link up with the East London Railway which, due to its financial state, was ripe for takeover. In this way the markets of South East London would be opened up.


Whitechapel Station – a major objective that provided interchange with the East London Railway. Again this is an 1896 photo. Note that the station was a Metropolitan District station which, confusingly, was the full title of the District Railway.

In 1879 the Metropolitan and District Railways (City Lines and Extensions) Act was passed authorising the extensions to Whitechapel from Aldgate and Mansion House. Furthermore, the act not only authorised the very short link between District and Metropolitan Lines between Aldgate and modern-day Tower Hill (to complete the circle) – it required a Circle Line service to be provided. Clearly parliament were well used to the antipathy of the two companies involved and were determined that they would be forced to provide a Circle Line service.


Mark Lane was the forerunner to Tower Hill station. The station seemed to be fairly basic.
The absence of a railway company name prominently displayed is surprising.

It is not clear how the Circle Line fared in its early days. In later years it caused a lot of problems that no other line experienced. The problems with running trains in a closed loop are many. Operationally the biggest problem has always been that if you are late you can’t make up time by reducing turnaround time at the terminus. In the early days the lack of a terminus meant that you had the added problem of having to change locomotives in something rather equivalent to a modern pitstop and then, in the confined space available, quickly re-water and re-fuel that engine so that it could be ready to take over from another locomotive in just a few minutes.


An early 20th century Underground Map.
Note the total absence of any reference to the Circle Line.

Operating Problems: You cannot lose time

On the London Underground you can’t give the Circle Line trains extra time because they share the track with other trains on different lines. Indeed there were just two places where they could wait in a platform to lose time without affecting other trains – Aldgate and Gloucester Road westbound (clockwise). Edgware Road did at least provide some further saving grace as there was also a bit of leeway because the only service affected was the Hammersmith & City which ran at the same frequency and should have been at least four minutes behind the Circle Line train. Older readers may well remember that Circle Line stops at Edgware Road would always seem to involve an interminable delay waiting in a train carriage with all four double doors open in one of the few platforms of the Circle Line that is directly exposed to the cold winter air.

It follows from the inability to have much “slack” that it is important that circle line trains run to time. Unfortunately in London the Circle Line encounters a number of flat junctions and all are intensively used. These are Praed St (near Paddington), Baker Street, Aldgate, Minories (north and south of Aldgate station respectively) and Gloucester Road. Fairly self evidently these junctions are only a problem on the outer rail (clockwise) service. The inner rail (anti-clockwise) service has no conflicting junctions at all during normal working.

If losing time on a circular line is difficult then gaining or recovering time is next to impossible. Skipping stops is not a realistic option on the London Underground for anything other than the Metropolitan Line. More aggressive driving can achieve a limited effect but is generally only condoned on ATO lines when it is controlled centrally. Terminating short of destination does not work on a circle line and neither does taking advantage of the built-in recovery time at a terminus. A basic rule of thumb: Continuous Circling Compounds Delays, Terminating Trains Absorb Delays.

Other Operating Problems

Other problems can include the issue of the driver needing a physical needs break (go to the toilet), what to do with the train if a relief driver does not appear at a crew change and uneven wear on the wheels if a train consistently goes round in one direction. In London problems are compounded by it being impractical for the depot to be on the route of the line itself.

A further problem is that it is very difficult to change the number of trains providing a service during the day so off-peak and peak hour services are run with the same number of trains. One might expect peak hour frequencies to be the same as off-peak frequencies but it is actually worse than that because it takes longer to make the journey in the peak. Depending on the timetable one has to offer a fixed number of trains per hour and realistically, at best, one can go to the odd half a train per hour. So in its later years, when running as a circle, the Circle Line off-peak service at 7.5 trains per hour (tph) was actually better than the peak service (7tph).

If all the above were not bad enough, the continuous Circle Line was a very unpopular duty made worse by the fact that the journey was all most entirely in tunnel and very repetitive without a break. This led to union agreements about the maximum number of circles that could be done without a break and the maximum that could be done in a day which then constrained the duty roster for the timetable. So not only was the line difficult to run, it was also difficult to timetable.

If it is a bad idea – why do it?

An obvious question to ask therefore is why, if it was known to be such a bad idea, did the Circle Line remain as a service? The answer seems to be that, despite its known horrors, it was just too difficult to get rid of. In fact one could argue with some evidence that London Underground and its predecessors spent an entire century trying unsuccessfully to get rid of the Circle Line.

Frank Pick’s dislike of the Circle Line

Frank Pick was one of the people, perhaps the main guiding light, whose influence even today is largely responsible for making the London Underground an organised well-designed integrated product. Back in 1919, as commercial manager of the London Electric Railways Group, he had already grown to dislike the Circle Line. No doubt the awkwardness and difficulty of operating the Circle Line would have been totally contrary to the simplicity and sense of order that he strove for. More particularly, in 1919, he disliked the fact that the Circle Line was taking up running slots that he believed could more usefully and more profitably used to bring in what we would now call commuters from further afield.

In 1919 the District Line to the west would have had branches to Wimbledon, Richmond, Hounslow, Ealing Broadway and South Harrow. It could have been worse but at least by now the Middle and Outer Circle no longer ran. One could have understood Frank Pick’s frustration. It seems that, in essence, he thought that by abolishing the Circle Line you could replace each Circle Line train by two starting from a District Line branch in the west. One of these trains would be a District Line one taking up the southern slot via Victoria. The other would be one from the still-rival Metropolitan Railway which would go via High St Kensington and the northern part of the Circle Line and continue to East London. Nothing came of the idea which may have been, in part, because the plan may have required the repeal of the clause in the 1879 act requiring a Circle Line to be run.

Looking at an old idea in a modern context

It probably worth looking at Frank Pick’s idea in a little more detail because in a modern content it actually reveals an awful lot about the Circle Line and ideas in planning underground lines in general.

The modern equivalent of Frank Pick’s idea would be to send a Circle Line train starting from Hammersmith (Hammersmith & City and Circle Line station) then continuing around the circle via Aldgate but from Gloucester Road it would leave the current Circle line to divert to Earl’s Court and then down to Wimbledon. It would then return to Earl’s Court and go via High St Kensington to Edgware Road taking up a Circle Line slot for its critical path through Praed St junction.

The crushing flaw in the plan would appear to be that altering the Circle Line like this would increase the number of sub-surface trains through Earl’s Court by about 50% which nowadays is simply not possible. One wonders if Frank Pick had any plans to deal with this or whether it would not have been a problem in 1919 or he had simply overlooked the issue?

What is possibly more interesting is to look at a variant of the scheme where one avoids Earl’s Court by sending trains Wimbledon trains to and from the south part of the Circle Line via a new link (possibly via King’s Road, Chelsea). Then one has an opportunity to bring many more people into the centre of London. Nowadays this however would be considered irresponsible because one is increasing capacity into London without increasing capacity within London.

The modern approach would probably be to reduce the number of Underground branch lines by providing a new central section. This cannot be done for the District Line alone because there is only one branch to the east. One could however take over the branch of another line such as the Central Line. So a line from Wimbledon to Epping would neatly free up extra capacity on the District Line and the overcrowded Central Line. This is, of course, a variation of the proposed Chelsea-Hackney Line which for around forty years was seen as the way forward – until Crossrail 2 proposals came along.

The other reason that Frank Pick’s idea is so interesting is that it really enables us to give an good answer to the question “What is the purpose of the Circle Line ?”. Turning Frank Pick’s idea on its head, one could argue that the purpose of the Circle Line is to use otherwise unusable capacity between Aldgate and Edgware Road via Victoria.

The idea of getting rid of the Circle Line will not go away

In 1933 the London Passenger Transport Board was formed and the Metropolitan Railway and the District Line came under the same organisation. Frank Pick was its Vice-chairman. Today we would probably have called his post “chief executive”. As well as day-to-day issues the board quickly got themselves involved in initiating solutions to outstanding major transport issues. We have already seen their initiation of a scheme for new tunnels between Finchley Road and Baker Street and for the Bakerloo Line to take over the hopelessly overcrowded Stanmore branch and most inner suburban stations of the Metropolitan Line.

Another thing the board did surprisingly quickly was to get the clause in the 1879 act requiring a Circle Line to be provided to be revoked. This they had done within 16 months of being in existence. It would not be hard to guess who probably was the driving force in getting the relevant clause repealed.

In April 1934 Frank Pick was already writing to Kensington Council supplying details of a proposal to run a shuttle service from South Kensington to Edgware Road and promising the reconstruction of South Kensington station to provided cross-platform interchange with the District Line there. This was no doubt do-able but would have been enormously disruptive and expensive with a need to keep a service running most of the time.

It is not surprising that nothing became of this proposed South Kensington – Edgware Road shuttle and it is difficult to see any real tangible benefit in providing it. At first sight one can put this down to one man and his obsession. It may well be though that the reason for pursuing this idea had changed. We know in that by 1928 the section of the Underground between Mansion House and South Kensington was the busiest two-track section and on it during the rush hours an incredible frequency of 40tph were timetabled. This would have been very difficult to sustain especially as usage went up and it may be that Pick was not planning to run more trains but fewer. In other words this section was probably running beyond its sustainable capacity and the least painful way to sort it out was simply to stop running the Circle Line trains through it.

It seemed that at the time it was only Frank Pick who was determined to see the elimination of the Circle Line happen and he was already not in the best of health. When he died in 1941 it seemed that the this scheme died with him. Obviously there was a war on and any further consideration would have to wait until the cessation of hostilities.


Frank Pick in 1922.
Even he appeared to find the problem of eliminating the Circle Line beyond his capability.

The Post-War Plan

Frank Pick’s plan may have died but the idea of getting rid of the Circle Line had not. In 1941 relocated Circle Line platforms opened at King’s Cross. Previously they had been inconveniently sited next to what is now the former King’s Cross Thameslink platforms. Included in the rebuilding was an extra platform between the running lines to conveniently terminate trains from the west. The original plan was that is would be used for extending some of the off-peak trains that then terminated at Baker Street. This did not really make much sense. Why bother? Alternatively why not terminate at Liverpool Street or Aldgate? It was now proposed to make use of this for a “turn and a half” Circle Line service.


The original Metropolitan Railway station at King’s Cross photographed in 1934 shortly after the Metropolitan Railway was taken over by the London Passenger Transport Board.
Note the diamond in the station name sign which was uniquely a Metropolitan Railway feature and the name on the board was King’s Cross & St Pancras which was the Metropolitan Railway name for this station.

The idea this time was to start at Putney Bridge, go to Edgware Road and from there to a complete circuit of the Circle Line before continuing to King’s Cross. No explanation was offered as to how Earl’s Court would cope with the extra trains. On the Baker St – King’s Cross it was proposed to introduce speed controlled signalling to handle 40 trains per hour. The speed controlled signalling could be thought of as very crude Automatic Train Protection system with the control mechanism being track based rather than train based.

It is not surprising that the turn-and-a-half scheme came to nothing. For starters it had the feel of someone starting off by trying to make use of an unused recently-built potential asset (that platform at King’s Cross) rather than solving a problem. It was also quite expensive in cash-strapped post-war Britain as it involved making the terminating facility at King’s Cross fit for use, extra rolling stock and signalling changes. In retrospect it is hard to believe that it wouldn’t cause as many problems as it solved.

A further complication to the original turn-and-a-half proposal which was originally made in 1948 was that proposed railway plans for a post-war London were being optimistically firmed up and one of the proposals was a route ‘C’ from Walthamstow to Brixton. This would clearly compete for funds and management time. There was also a proposal for a variant on route C to go to Fulham Broadway after Victoria and from there take over the District Line to Wimbledon. This was now seriously complicating the issue. With the Victoria Line serving King’s Cross it was questionable how useful additional trains on the Circle Line would be and, as well as that, the impact on of route C going to Fulham Broadway would now have to be taken into account. The turn-and-a-half scheme was dropped and revived again then died a death in this particular form for a final time. The scheme followed the 1949 Bakerloo Line extension to Camberwell in a list of “what might have been” for the period.

Further Ideas

The last decade of the 20th century saw some further half-hearted schemes. A couple of them seemed more intent with trying of find a use for the much-unloved 1983 stock that was replaced on the Jubilee Line in preparation for the opening of the Jubilee Line Extension. A simpler, more sensible was the idea of simply getting the Hammersmith & City Line to do a turn-and-a-half. This was actually put into practice for the Notting Hill Carnival in 1990, 1991 and 1992.

The idea of the turn-and-a-half Hammersmith & City for daily use was ultimately rejected on two grounds. The first was because of the limited ability to recover from major delays. This does sound surprising because it surely had to be better than the arrangement then in existence. It also seems completely at odds with what eventually happened which is really remarkably similar in concept although operated as two separate services. The second reason for rejection is that it was too complicated and too confusing for passengers. Apart from possibly suggesting that Carnival attenders must have higher than average intelligence, this seems inconsistent with experiences abroad – as we shall see in part 2. This reason for rejection also seems inconsistent with what actually subsequently happened which is very similar in nature.


Yet another 1896 photo. This one is of Cannon Street station. This station, along with Monument and Mark Lane, was then jointly owned as is clear by the signage.

It is clear that, despite every attempt to break the continuous circle being either stillborn or regarded as unworkable, London Underground had not given up the idea of getting rid of a continuous Circle Line. To quote the final paragraph of the book “The Circle Line” (see below):

Finally we have Mr Derek Smith, then managing director of London Underground, addressing the London Regional Passengers Committee on 24th November 1999 – ‘The Circle Line’s contribution to the totality of the network is not large, and it makes control of shared lines difficult. It is not the right design, and in the long term the question is whether it should continue’.

In part 2 we look at how London Underground tackled the issue of the Circle Line in the 21st century. For comparison we also look a few other cities and see how they deal with this issue.


This picture was taken in 1902 and features in many books. It may well be the only known photograph to actually explicitly reference the Inner Circle. It is a picture of Aldgate Junction taken from the north end of Aldgate station. Note the very limited space to service the locomotives. The locomotive in the siding has a board that says NEW CROSS as for a few years the Metropolitan ran through services to New Cross (SER).

The bulk of the background for this article was either obtained from or verified using The Circle Line by Desmond Croome. Like all books in this series it packs an awful lot of information into its 80 pages and is highly recommended for further reading although it pretty much stops at the end of the 20th century. Despite the title it also includes coverage of the Hammersmith & City Line from Hammersmith to Paddington.

Our thanks go to the LT museum for allowing us to use their photos.

jump to the end
There are 116 comments on this article
  1. Greg Tingey says:

    I note you refer to the”Inner & Outer Circle” as pertaining to the direction of travel, & later, correctly refer to “Inner & Outer Rail” ( Anti-clockwise & clockwise, respectively.)
    The Middle & Outer circles are another story entirely, of course. Let’s not go there today, shall we?

    The rivalry wasn’t so much between the railways as their chairmen, of course, which also intimately affected, & still affects travel in SE England, especially Kent & the London suburbs in that direction.
    The Manchester, Sheffield & Lincolnshir Railway metamorphosed into the Great Central, & had Sir Edward Watkin as chairman. He was also chairman of … The Metropolitan Railway, the South-Eastern Railway, the Channel Tunnel company & was on the board of the Nord, in France.
    A man with a “VISION”
    His bitter personal rival was James Staats Forbes, Chairman of the Met-Dist & of the LCDR … he comes across as a bit of a twister, at this distance in time, but he lasted longer, dying in 1904.

  2. Castlebar says:

    I posted this on the current “Old Oak Common part 3” thread a few says ago.

    “I agree that the Circle line could easily be dispensed with. How ?? Easy!! Richmond trains run to Ealing Broadway, clockwise around the Circle. Ealing Broadway trains run to Richmond anti-clockwise around the Circle. Job done at the eastern end of the Circle.

    Trains from Upminster/Dagenham run alternately clockwise/anti-clockwise around the Circle.

    ADDED BONUS: Trains and crew no longer end up at the wrong end of the line after a days work in the event of “service disruption”.

    It needs some meat on these bones, but I’m sure everyone can get the basic concept of this. Personally, I think this would even out traffic flows and be less disruptive than previously tried ideas dating back for generations. I think this need to be tried.”

    Just an idea, but I hope it can be discussed as a possible solution, or, at least a part of one.

  3. Henry Fowler says:

    ‘Frank Pick —– “whose” influence’ etc
    [Now fixed. Thanks.PoP]

  4. lmm says:

    If it’s about spare capacity Aldgate – Edgeware Road, wouldn’t the simplest answer be for Metropolitan line trains to continue around the loop to Edgware Road?

  5. mr_jrt says:

    My solution to the Circle problem is simple: Basically, start from a 50/50 split on the northern Circle between the two lines, then adjust as demand requires ( I suspect a 2:1 in favour of the Met would be more likely though). Crossrail will be taking away all the short suburban services on the GWML, leaving the former H&C platforms (13&14) free to revert back to the H&C from NR. Utilising the central roads for terminating services is conflict-free, so it makes sense to turn more back here than Met services at Baker Street, which do conflict – badly (incidentally, I wonder why they haven’t knocked through platform 4 so platform 3 can be used for non-conflicting terminal services…) This operates the northern Circle line at capacity between Praed Street Junction to Aldgate East, the Hammersmith branch, as well as the Met line to Harrow-on-the-Hill.

    Next, we have to remove the bottleneck between Praed Street Junction and Edgware Road, this then enables the line down to High Street Kensington to be used to capacity, with trains terminating at Edgware Road conflict-free. A further alteration west of Earls Court (as discussed on the Earls Court article) to permit the Wimbledon trains to have exclusive use of platforms 1 & 4 then gives a segregated line from Wimbledon to Edgware Road that can operate at capacity.

    Finally, the District Line.

    In order to make paths for District Line services through Aldgate East, H&C or Met services would have to be turned back somewhere on the northern Circle, and the obvious place with the least capacity wastage is Aldgate. So for every District train that runs through Aldgate East, a H&C train would terminate at Aldgate. I say H&C as it will have to operate using the short S7 trains, whilst the Met uses the longer S8 trains, so will be able to provide more capacity out to Upminster using the long platforms already present on this section. Likewise, for every H&C/Met train that operates through Aldgate East, a District service will terminate at Tower Hill. That then gives you Tower Hill back through to Turnham Green operating at capacity (assuming the Olympia branch remains defunct).

    So in summary, action items from that:
    * Restore H&C terminal platforms at Paddington
    * Widen Praed Street Junction to Edgware Road.
    * Alter flyover west of Earls Court.
    * (optional) Knock through platform 4 at Baker Street

    …and then the only section of line that lose service from having the circle withdrawn are High Street Kensington to Gloucester Road and Aldgate to Tower Hill. Both resolved by interchange using the more frequent services at Earls Court and Aldgate East, respectively.

  6. Philip Wylie says:

    Interesting article as usual. I wouldn’t object to the Circle being withdrawn but what it does do beautifully (being selfish here) is to provide a quick link from Victoria to High Street Kensington and Notting Hill Gate, journeys I make frequently. Certainly, the Vic/Central offers ‘unthinking’ frequency albeit with a change at Oxford Circus if bound for Notting Hill Gate.

    Both bus alternatives (148 and 52) are potentially good but dogged by congestion especially the 52 in its traverse of Knightsbridge and at the junction of Kensington Church Street awaiting the right-hand filter.

    As for Earls Court, it’s not difficult to ‘go round the corner’ by changing there. I do like the idea of a dedicated platform 1 for all Edgware Road bound trains and I rarely see it occupied by one. The majority all seem to come into 2 and most clog the platform up for either crew changes or awaiting a District to move out of 1 to clear the junction ahead. Up to 4/5 mins is not unknown and seems an incredible waste of time.

    Finally, would it be sensible to knock through Plat 3 or 4 at High Street so that you can accommodate two northbound trains without one being stuck in tunnel until the other clears? You’d still have a conflicting move southbound though.

  7. Anonymous says:

    The problem is that the western end of the former Metropolitan and District railways have seven branches (Watford, Amersham/Chesham, Uxbridge, Hammersmith, Ealing Broadway, Richmond, Wimbledon) but the eastern end has only one (Upminster) hence turnback or other such facility is required at the eastern end of the city.
    I rather like Castlebar’s suggestion of trains from the eastern turnbacks (Upminster, Dagenham and Barking) running round the Circle in alternate directions returning to eastern turnbacks whilst trains from western termini running round alternately returning to western termini. Unfortunately whilst this would result in traincrews being close to depots at eastern end it would possibly result in traincrews being remote from depots at western termini if the Watford and Amersham branches were included in the extension. So perhaps it would be preferable to only include the current District branches.
    Another idea would be to transfer the Piccadilly service to Rayners Lane/Uxbridge back to District whilst Piccadilly take over the Richmond or Ealing Broadway branches. The Trains from Uxbridge and Rayners Lane could run round the Circle in alternate directions via Harrow, Baker Street, Aldgate, Earls Court, Acton. and vice versa.

  8. Castlebar says:

    @ Anonymous 12:03

    Exactly right about the Met.

    That’s why I only mentioned the District

  9. Lost in Middlesex says:

    Here’s my idea for uncircling the circle – I hope I haven’t repeated someone else’s suggestion. It basically involves swapping the end destination of the current circle line with the end destination of the current District line Wimbledon – Tower Hill service. So there would be a service that starts at Hammersmith, goes three-quarters of the way round the circle, then to Wimbledon. And another service from Edgware Road to Tower Hill via the west and south sides of the circle. It seems like win-win to me, but I’m sure someone will spot a flaw that I haven’t yet spotted!

  10. Si says:

    Interestingly, the SLL route through Whitechapel doesn’t have the full compliment of trains – only 3 services: 2x District and 1x H&C, compared to the 4 services on the northern and southern circles (1xDis terminate at Tower Hill, the Circling is one on each side and there’s 2xMet terminating at Aldgate).

    The SSLs are endless wonder for rehashing service patterns, and with little bits of work (Earls Court junction, 4 tracking Praed St – Edgware Road, changing the whole Barbican-Liverpool St area to create a junction, etc) there’s even more possibilities. I could, like many others, wax lyrical for ages on this. However I will content myself on picking up a comment made in the post:
    “One could however take over the branch of another line such as the Central Line. So a line from Wimbledon to Epping would neatly free up extra capacity on the District Line and the overcrowded Central Line.”

    While this strikes me as a very good solution, the issue is that both it and Upminster would only be able to have 2 services each (the easiest way to do this is to have the Met go to Upminister and the District to Epping, with the H&C terminating at Aldgate). However, in doing this, the service between Mile End and Barking is reduced, which isn’t good. However, some route to Epping is a good use of some of the 5 (or 4 if you up the service through Whitechapel) services that die before reaching the east end.

  11. Martin Smith says:

    “by now the Middle and Outer Circle now longer ran.”
    Should be “no longer”.
    [Now fixed. Ta.PoP]

    Thanks for a fascinating article, look forward to Part 2

  12. StephenC says:

    The Circle is a weird one, interesting to read some of the history. I doubt we will see many changes in the next few years as the “teacup” seems to have solved the immediate pain of circling (by introducing hassle for customers).

    I suspect that there is only likely to be change if there is a larger scheme led by development money. The obvious location is redevelopment of Aldgate led by the City of London. As I’ve remarked before, Aldgate was touted in an exhibition last year as the next area of the square mile for a major makeover. If done well, that could involve digging up the whole area, allowing a new station south of the extsing Aldgate with two District line platforms directly above two Met platforms with full turnback sidings to the south to handle 30tph. This would then allow the Met to be extended in the future to South of the River (eg. via City Hall and Bermondsey to Hayes).

    Potentially this would result in completely separate District and Met services (no more Minories/Aldgate junctions), and no Circle or H&C swapping tracks in the east. A Gloucester Road to Edgware Road shuttle would fixup the west (interesting to see its been proposed before!). The H&C lost link at Aldgate is tempered by the new interchange station and Crossrail serving Whitechapel to Liverpool Street/Farringdon.

    Again, this isn’t overly pie in the sky. There are distinct possibilities that arise from a major Aldgate area redevelopment, and there are few cheaper ways of getting a cross-London route to the south-east.

  13. Long Branch Mike says:

    In the second paragraph of the ‘Frank Pick’s Dislike of the Circle Line’ section, there is the following typo “It could have been worse but at least by now the Middle and Outer Circle _now_ longer ran.” It should be “no longer ran”.
    [Now fixed. Ta.PoP]

  14. Anonymous says:

    You say that the inner circle has no flat junction conflicts. What about at Gloucester Road where it merges with the Eastbound District?

  15. Anonymous says:

    .. and indeed after Aldgate?

  16. Will says:

    Anonymous @ 14:55 & 14:56 – Conflicting movements are where train paths cross on the flat (e.g. at Aldgate Junction, an outer-rail Circle line train crosses the path of a westbound H&C.) The cases you mention are of paths merging rather than crossing. The only conflicting movement for an inner-rail Circle train is when the terminating platforms at Moorgate are used.

  17. Anonymous says:

    How you other readers return to London Reconnections posts and go straight to JUST the comments they haven’t read?

    Is there an easy way?

    LR gurus: Would you be able to code cookies and write them on to our computers?

  18. Fandroid says:

    Has anyone seen a chart of loadings on the Circle? It would be particularly interesting to see how many travel on the High Street Ken – Gloucester Road and Tower Hill – Liverpool Street links, compared with the main east-west legs to north and south. Personally, nearly all my journeys involve crossing the Circle (in reality a distorted flattened ellipse) from below the southern midway point (Waterloo) to one of the northern mainline termini, or from Paddington along the northern leg.

    I’ll restrain myself and comment on the Edgware Road terminators in part 2.

  19. Anonymous says:

    I only ever look at my comments. Isn’t that what this site is all about? You write something and then you read it. Over and over. Then you feel good all over. I like you like this site a lot.

  20. Centrifugal Kitten says:

    There’s an interesting history of Circle Lines in London which goes beyond the LU Circle Line we know today: it would be interesting to read more about these former Outer and Middle Circle Lines. As illustrated by this map on Wikipedia these lines incorporated bits of the North London, Dudding Hill and West London Lines. I would be interested to learn more about these old routes and find out if re-instating any of the dismantled junctions could enhance today’s Underground/Overground…. worth a thought.

  21. PeteD says:


    I’ve always thought that continuing the Met to the east or southeast of London would be a good idea. Before Crossrail got the go ahead I thought that the Met to Canary Wharf would give a cheap alternative.

  22. @Lost in Middlesex

    Flaw no 1: You are increasing the number of trains going through Earl’s Court by approximately 25%. The station cannot handle that.

    Flaw no 2: You are increasing services by approximately 25% between Gloucester Road and Tower Hill. This was virtually the identical problem (except it was too many trains South Kensington to Mansion House) that Frank Pick was trying to sort out. We don’t want to re-introduce it.

    I am saving the details for a future article but basically the sub-surfaces lines are, or will be, pretty much optimised to eke the absolute maximum capacity out of them. It is a recurring theme of mine but I never understand why people never seem to think that the planned service, thought out involving a lot of deliberation by people with backgrounds in operational research and with access to the necessary data and expert services like civil engineers for their opinion, is the best one and always want to change it in some way.

  23. Lost in Middlesex says:

    Pedantic of Purley:

    What I propose doesn’t increase the number of services on any part of the network at any time. The circle line is currently 6 tph, and the Wimbletower (have I just coined a new term?) is 6 tph. The two services run on the same tracks between Gloucester Road and Tower Hill. What I’m suggesting is that trains originating at Wimbledon, instead of terminating at Tower Hill, should continue on to Hammersmith (Anyone for Wimblesmith?). And trains originating at Edgware Road, instead of going all the way to Hammersmith, should terminate at Tower Hill (Towerware? Edgtower?). It won’t solve all the world’s problems, but it will sort out the current confusion at Paddington and Edgware Road.

    And while we’re at it, let’s rename Edgware Road after the street it’s actually located on – Chapel Street. Then the two lines terminating there could be called Wimblechap and Towerchap.

  24. Anon5 says:

    From a purely aesthetic point of view are there any modern maps showing the tube without the Circle? It would certainly take some getting used to

    If the Circle were to disappear could the Wimbledon to Edgware Road and Wimbledon to Upminster services be coloured yellow?

  25. ChrisMitch says:

    I’d like the old continuous Circle back please! It is mildly annoying that there are no trains from High Street Ken that continue beyond Edgware Road. And Edgware Road is pretty inconvenient for changing trains.

  26. Anonymous says:

    @Fandroid 16:34
    Trains between Gloucester Road and High Street Kensington always seem busy, all seats are usually taken off-peak so this service provides a well used link.
    Trains between Tower Hill and Aldgate seem much less well used, you can always get a seat off-peak. I do often wonder if this service is really needed. Central Line to Liverpool Street from Bank is faster then waiting for a Circle Line from Monument.

    I like the idea of removing the triangular junction at Aldgate and operating all 30tph from the east (Upminster/Barking etc) through to Victoria and Gloucester Road. When Crossrail opens at Whitechapel there will be an easy interchange there for Liverpool Street/Moorgate/Barbican/Farringdon/Paddington so serving some of the Hammersmith and City Line destinations currently possible as a direct journey. The 30tph departing Liverpool Street for Whitechapel/Aldgate could then be diverted into a new tunnel running south under the Thames, perhaps to London Bridge. The resulting reduction in conflicting movements would improve reliability.

  27. The other Paul says:

    The Yamanote line in Tokyo seems to run a reliable circular service. How do they manage it?

  28. timbeau says:

    “basically the sub-surfaces lines are, or will be, pretty much optimised to eke the absolute maximum capacity out of them.”
    Indeed – without changes to the infrastructure each leg is probably running at capacity, and all that can be done is swap what goes where. For example, there are essentially four different services on the south side of the Circle, running in rotation (Circle, Richmond, Wimbledon, Ealing) and likewise on the north side (Amersham/Chesham, H&C, Uxbridge, Circle) . Similarly, feeding through Earls Court are four services from the west (including two from Wimbledon), three of which go to Gloucetser Road and one to HSK., and feeding into the Edgware Road/Baker Street complex, there are eight – two from Hammersmith, two from High Street Ken, and four from the Met (two Uxbridge, one Watford, one Chesham/Amersham), of which only four in total can continue east of Baker Street. You can swap these over as much as you like – for example if you want anything to run from High Street Ken to Kings Cross, either a Hammersmith or a Met main service will have to terminate at Edgware Road or Baker Street to make room for it.

    Incidentally, in drawing this out I was struck by the topological resemblance between the complete network that will be operated by the S7 stock and the map of the Northern Line as it was between 1940 and 1975. (Turn it through 90 degrees). Only the two short links unique to the circle are missing)

  29. Malcolm says:


    Yes, this (Aldgate crossing) is what Stephen was suggesting, except that he carefully made the idea a little opaque by temporarily renaming Aldgate. But in fact he has looked more carefully at the map, and it is indeed not London Bridge, but rather Tower Bridge, that the Met is pointing at. He called it City Hall, which is not a name that I favour, but is probably less confusing than calling it Tower Bridge (unless it had platforms under the river with an entrance at each end). Then take your pick from south-east London.

    But of course ideas like this will have to get in a jolly long queue, unless an angel with very deep pockets turns up.

  30. Will says:


    If I read you correctly, and given that the Circle Line operates a 6 tph service, this seems to suggest that both the northern and southern sides of the circle operate at 24 tph. Isn’t this rather unambitious? Is it the flat junctions which constitute the main constraint to a more frequent service?

  31. timbeau says:

    @will – 24 tph is a train every two and a half minutes: and yes, the flat junctions are the constraint. You might squeeze a few more through Baker Street if you terminate all Met services there (essentially plain-lining the Circle/H&C route), but they’ll still get unstuck at both Praed Street and Aldgate junctions, unless you are going to resort to drastic measures like running Paddington-Hammersmith as a shuttle and closing one side each of the Aldgate and Gloucester road triangles. (And even then you’ve still got a flat junction connecting the other two sides of each triangle).
    You can shuffle the pack how you like: but if you do so, you will have to move other things in consequence
    for example Upminster – Circle – Upminster and Ealing – Circle -Richmond, – this would require the Hammersmith trains will have to terminate at Edgware Road, (or Aldgate if no Met trains go beyond Baker Street) and all Wimbledon line trains would have to go round the south side of the Circle (to Barking and Tower Hill alternately)

  32. Graham Feakins says:

    @ The Other Paul – The Tokyo Yamanote Line indeed runs a reliable ‘Circle Line’, with 11-car trains and at 2-min. intervals during the peaks at that, whilst many other routes join and part all around its route. However, the main difference between that and London’s version is that all those other lines have their dedicated tracks and the Yamanote has flying junctions everywhere where there’s a conflict of route. In other words, apart from depot workings, there are no flat junctions and the Yamanote trains simply follow one another.

    Having said that, I am not sure that I follow all this scaredy cat approach to flat junctions as discussed in the main article. There are more flat junctions encountered on the Southern from London Bridge to West Croydon via Tulse Hill and Streatham (at least 11) and so on replicated throughout the South London suburban area.

    For many a year, London Transport had recognised the problem that running the Circle Line as a simple service, station after station, without a break anywhere would eventually conflict at one or other of its junctions and give the crew no break. If I remember correctly, the timing was 53 minutes to complete the Circle but some seven minutes was written in to afford a form of layover, especially at Aldgate, where the line is on its own. Edgware Road was used as more of a ‘timing correction point’ to regulate the service that side. On the Yamanote, on-train crew breaks do not happen as the crews simply rapidly swap over, normally without a word between them and the train sets off as if nothing had happened. Not even a second of delay or faffing about.

  33. Ian J says:

    Thanks for a really interesting article. A couple of questions that occurred to me on reading it:

    This was not the originally intended location for [Mansion House station]. It should not be hard to guess where that was

    So was the Circle originally intended to go via Bank? How would it have worked beyond that – would it have ducked down to Tower Hill, or gone a totally different route to meet the Metropolitan?

    The modern approach would probably be to reduce the number of Underground branch lines by providing a new central section

    Isn’t that what was planned at the time, with the proposed deep level tube underneath the District Line? In the end that scheme was merged with the Piccadilly Line scheme, so the District was able to lose two Western branches to the Piccadilly.

  34. Will says:

    So is it beyond the wit of man, or the funding of TfL, to sort out the junctions? I appreciate that grade-separated junctions don’t come cheap (and you also have to factor in the disruption while they are being built). But if this is the main technical obstacle to having a 30+ (40?) tph service … seriously, what are the options? Stick with the present hobbled system, and artificially stifle demand? Build an entirely new line, at vastly greater cost? One way or another, this is a nettle which is going to have to be grasped.

  35. Malcolm says:

    Grade-separating the circle line junctions would only work if they were all done. And that does seem to get into the ball park marked “too expensive”. Remember that they are all rather near the surface. And I suspect that there would also have to be quite a lot of work done enlarging many of the stations to handle bigger flows of passengers – even with the current frequency they do seem to get pretty thronged. A whole new line somewhere might even come cheaper.

    The other reason the circle line had to have pauses was because the total journey time is not an exact multiple of the frequency, if you see what I mean. So the proper number of trains to run a non-pausing circle was not a whole number.

    Changing drivers “without fuss” is quite possible. But it is one more thing which can go wrong.

  36. Malcolm says:

    It might be worth noting that, in theory, if you get the timetable right, flat junctions need not constrain the frequency at all. You “just” ensure that a train from A to B passes a train from B to A on the junction, immediately followed by a train from A to C passing a train from C to A.

    Naturally real life is not quite like that. But one does wonder whether you could let a computer loose on the timetable, and somehow make it all work. How do the Swiss do it?

  37. stevekeiretsu says:

    Not sure about the Aldgate regeneration theory – there’s already lots of development in the vicinity (Aldgate Tower, One Commercial Street, Altitude, Goodmans Fields, Cityscape) and none of it seems to have come with reconfigure-the-tube-network money or masterplanning attached.

  38. Ian J says:

    An ironic point about London Underground’s advice and Frank Pick is that London Underground staff were heavily involved in designing the route network for the Moscow Metro in the early 1930s, which included a circular line. Pick even got a Soviet award from Stalin for his help. The Soviets must have got on OK with their circle line, because they in turn helped Mao’s China plan the Beijing subway, which had a ring line too – and now has two concentric rings.

  39. Greg Tingey says:

    Stephen C
    You will still need to sort out Praed St Jn & the W end of Earl’s Court.
    The latter, IMHO, is the greatest problem, as discussed in the “cocooning the WLR” thread ….

    A surprising number go round the corner: Liverpool St – Aldgate – Tower Hill etc…. More convenient, especially now the tea-cup service is more reliable. Quicker & less hassle than fighting your way down to the Central, then through interchanges (typically) @ Holborn or Tottenham Ct Rd. Sub-surface stock, too, not a deep-level cram-box.
    So sorry, anonymous @ 21.34 – 03/10/13 but no thanks – for reasons given above.

    Pete D
    It was called the “St Mary’s curve” & was finally lifted, when the most recent ELL conversion to OvergrounD took place.

    Flat junction smooth working:
    Seems to work very well between Bethnal Grn N Jn & Liverpool St.
    Weirdly, though, Clapton Jn does not seem to go so well!

    Ian J
    Ah, but, according to legend, the Moskva circle line came about, because Stalin put his tea-glass down in the middle of the stylised map – so they had to build it!

  40. timbeau says:

    @Graham F
    “There are more flat junctions encountered on the Southern from London Bridge to West Croydon via Tulse Hill and Streatham (at least 11)”
    Yes, indeed, but they are only trying to run 2 or 4tph on those routes, not 24.
    By the way, I can only find nine junctions.

    “So is it beyond ….the funding of TfL, to sort out the junctions?”
    Probably, yes

    “Build an entirely new line, at vastly greater cost?”
    A new line through virgin territory (clay) is probably cheaper than remodelling existing layouts. At least that’s the justification for HS2.
    In fact, that’s exactly what tfL are doing. It’s called Crossrail

  41. Walthamstow Writer says:

    @ Malcolm 0022 – why do you imagine a computer hasn’t been let loose on the timetable? I don’t think the LU Schedules department is devoid of computers. From my travels Circle Line trains typically are scheduled to pass each other at Gloucester Rd and at Aldgate thus ensuring smooth movement across flat junctions. Clearly if you have an intensive service level it does not take very much to cause perturbations which can cause delays at key points. It will be interesting to see if ATO under the Cityflo signalling system provides “smoother” transition across key junctions on the Sub Surface network. It will certainly be a challenge given how many services “interlock” with each other at key points, some remote from the centre, to ensure that trains arrive at the right time at these key central area junctions.

  42. ngh says:

    I was going to hold of commenting till part 2 but…

    Introducing the new S7 stock on the H&C, District and Circle and subsequently new signalling will allow a slight increase in frequencies on a small number of routes, barring that no major changes are likely until Crossrail has settled down and frequencies west of Paddington have been increased from the initial infrequent service. i.e. there is only so much crystal ball gazing that can be usefully done the further you look into the future.

    If the HEx service is discontinued post agreement end in 2023 and CR gets 8tph to Heathrow, the frequency on the Piccadilly (and hence number of units and paths required for the Piccadilly line to serve Heathrow) could possibly be reduced and those units and paths through London could go to taking on a branch of the District – Ealing Broadway would make the most sense?

    One might think that the H&C services between Whitechapel and Paddington would be likely to get a bit quieter on some sections in the short term after CR opening but CR might just generate more passengers swapping on to H&C at Paddington and the same at points east too. Also lots more passenger getting off Thameslink Services at St Pancras post 2018 as the frequency and length of trains goes up. And in the longer term how to dissipate passengers from HS2? (Or CR2 at Victoria on the southern part of the loop etc.)

    And for the Crayonistas:
    I quiet like the idea mentioned by a few others of extending the Met South of the Thames – via Fenchurch Street to London Bridge in deeper tunnels with possibly a new Met Aldgate platforms below the current one and grade separated junctions would be my precise route suggestion. Providing relief for Northern Line / Bank which will still have capacity issues even after the current planned works. (As not all London Bridge service will call at New Cross or New Cross gate there are limits to what the ELL can do

  43. c says:


    Are you suggesting that if 8tph Crossrail went to Heathrow, that the Piccadilly line would have Uxbridge, Ealing Broadway and Heathrow trains? 4 branches there, would be tough.

    And what about Chiswick Park?

    I think the current arrangement of District and Piccadilly is probably the best.

    I’d personally like more frequency down to Heathrow on the Picc. The Northfields terminators could be looked at. And faster journey times with new stock and signalling would be good too.

  44. Graham H says:

    @Malcolm and others too numerous to mention individually. There’s plenty of software (of varying quality) that enables you to simulate the operation of an particular timetable and assess the impact pf perturbations. Railsys is the one often used (tho’ NR have their own Total Route Planning stuff these days, and LU, no doubt, something else.) I don’t believe the Swiss, despite their usual efficiency, have to face 24-30tph through very many flat junctions, although the way their timetable is constructed, there is a spike in movements at nodes around each hour/halfhour and a lot of slack in between. The issue with flat junctions is the loss of a path every time something crosses over the “other” track – and this remains an issue, however well the service is regulated. Ultimately, the limitation is one of how quickly a point machine can move and the signalling reset; most operators would argue for something in excess of 60 seconds.

    @Greg T – isn’t that the Commie version of the Tsar’s thumb accounting for the kink in the Petersburg-Moscow line?

  45. Pedantic of Purley says:

    @Graham H and others too numerous to mention individually

    London Underground have their own very old Fortran in-house simulator. That might sound behind the times but Fortran is the sort of language wierdos operational researchers love because of its power and mathematical ability. The fact that it is in-house means any deficiency (e.g. no support for SDO) can be quickly rectified and you don’t have time delays whilst updating manuals, getting customer approval etc. What little I have seen of it looks very impressive. Apparently they almost claim they don’t really need it because there calculations are correct but it is useful for showing MPs and others why their suggestions won’t work as they can visualise something that they cannot understand in words.

    You can be pretty certain all the scenarios suggested (and many more) have been played out on the simulator but timbeau put it better than I did: each leg is probably running at capacity, and all that can be done is swap what goes where.

    Sorry Lost in Middlesex if I misunderstood but that is pretty much the revised answer to your suggestion. At the end of the day we can shuffle what is available.We probably all have our answer as to what is best and doubtless we won’t all agree. One thing is pretty much for sure – we cannot create capacity by re-organising the Circle Line.

    I thought actually the reference to computers was about real-time ones predicting any clashes at junctions and altering the speed of trains (in ATO) to maximise capacity. This will happen, will increase capacity but doesn’t happen now on the sub-surface lines.

    We are now lightly treading into areas intended for future articles. But it is interesting that the Swiss get a mention as that is what I planned to touch on as a point of comparision.

  46. Pedantic of Purley says:

    “So is it beyond ….the funding of TfL, to sort out the junctions?”
    Probably, yes

    Another of my mantras is that you have to think in 3D when thinking of railways (especially the London Underground). It is not so much beyond the funding as something that is often physically impossible. I cannot see how Baker St can be anything other than a flat junction. At other places (e.g. Aldgate triangle) the gradients would be way, way too step if indeed possible. Of course you could look at lowering stations but, even assuming that was possible, the expense would be so great for so little achieved that, as others say, it would be more cost effective to start again with a new line.

  47. Castlebar says:

    Hi all,

    I would like to re-emphasise the points I made on what is just the second comment on this thread. My ideas (some call it a ‘panhandle’ routing), do not involve any track layout changes whatsoever. I do not proposed re-building, adding or taking out any junctions. The scheme does not require new build of junctions, ‘flying junctions’, and certainly no crayonistic line extensions to any line. I have simply proposed a plan that does away with “The Circle” which, IMHO, will improve service all around the District, with knock-on added benefits to the northern part of the Circle too. If it would be possible for discussions to focus on using existing infrastructure, I think that is a separate issue to any possible new build projects, which would of necessity be many years into the future. My idea would alleviate a problem that seems to exist today, and because it uses only existing trackwork, could possibly be solved tomorrow.

    Whereas futuristic ideas are nice, they will take many years, and will not immediately solve any problems that seem to exist now. Furthermore, my suggestion would be quite cheap to implement whereas any track alterations require money that might be difficult to find.

  48. JamesC says:

    Interesting artical yet again. Readers may be surprised that this morning at around 7am for a first time in what may well have been a while a H&C train (an S7 for that matter) came very close to running on the district line. After we got stuck ontop of the points between aldgate east and liverpool street they managed to get back off them and after a very long wait werr about to reverse back j to aldgate east with the intention of running the first passenger service on an S7 down the district line to edgeware to rejoin the h&c in the other direction. Not sure ifan s stock has actually run down the district yet, but they seemed fo unjam the pojnts just as we were moving back to aldgate east.

  49. ngh says:

    Re c
    10:33, 4 October 2013

    Install 4 points to swap the Ealing Broadway and Uxbridge Picc line trains from the inner tracks to / from the outer ones just east of Chiswick Park (or just use the existing points at Hammersmith) if you don’t mind using up a bit of potential capacity on the remaining district branch in the short term.

    And on the Picc Heathrow branch operate the service as the traditional T4 then T123 loop with a T123 to T5 shuttle. (That leaves 3 branches).

  50. Centrifugal Kitten says:

    I always like to imagine in a sort of crazy Heath Robinson kind of way, unfettered by the constraints of budgets or engineering, how sensible it would be to just dig a tunnel under the whole circle line and move all anticlockwise trains of all hues down to a lower level. I know it’ll never happen but it would be great to free up all that capacity and unclog the junctions. Maybe it could be done on just one stretch of the circle, just to alleviate congestion….

  51. StephenC says:

    Apparently I was dopey yesterday re Aldgate/Aldwych (perhaps a moderator could fix?).
    [Done. And easily done. I once wrote a whole article on the South London Lines (SLL) and only realised at the last minute that all my references were to SSL (sub-surface lines).PoP]

    Examining developments further, the Trinity EC3 building is the key one here. If built as is, then the Met/District junction will never be fixed and the Met will always terminate at Aldgate. Perhaps someone from the City of London/TfL is reading and wants to re-examine that proposed building?

    @PoP, ngh, I’m ambivalent about exactly where the Aldgate cross would join in. When I walked the site, I reckoned that if you closed the existing Aldgate and the H&C link to Aldgate East, then the entire length of the current Aldgate station could be used for diving (terminate the Met at Moorgate during the works?). With a realignment of the District a little further south than the current Minories Jn more space is created (again all this is possible only if cut-and-cover working from the Trinity site). About 180m length total can be found which should be enough to get a 4.5m height difference. The potential of a cheapish cross-river link is one worth a proper examine.

  52. Philip Wylie says:

    @JamesC if you count S7s running round the Circle and thereby on the District between Tower Hill and Gloucester Road and on to Edgware Road sharing with Wimbleware, there are indeed quite a few in service now. All change for those who have got used to getting in or out of a certain door nearest the stairs/exit. For exit at Notting Hill Gate, use the first door of the second car 🙂

  53. ngh says:

    Re Pedantic of Purley
    10:59, 4 October 2013

    And it being in fortran means it will be completely transparent compared to off the shelf tools.

    The recent TSGN ITT documents on what was permitted or not in excel models to be submitted with the bid was highly entertaining!
    Post West Coast franchise… Obviously none of the relevant people in DfT had read the finance industry textbook “Practical Financial Modelling” [with excel/access] until after the enquiry as the verboten list comes straight from there. 😉

  54. timbeau says:

    @ Castlebar
    I understand the concept, but note the consequences. Consider the throughput on each leg.
    Firstly – the section High Street Kensington to Paddington and, crucially, Praed Street Junction. If your two “panhandle” services run through this leg, then they have to replace not only the Circle but also the Wimbleware (probably by diverting it to Tower Hill or Barking). Similarly, as the panhandle services continue to Baker Street and beyond, paths must be available for them over the flat junctions at Baker Street and Aldgate, so either all trains from Hammersmith must terminate at Edgware Road, or all Met main line trains must terminate at Baker Street (or possibly half and half, which would be even more confusing).
    However you cut it – Barking to Tower Hill and Wimbledon to Wimbledon, Upminster to Ealing via Victoria, Kings Cross and Victoria again plus Upminster to Wimbledon via Kings Cross, Hammersmith to Ealing via Aldgate, Watford to Wimbledon, Rickmansworth to Richmond, Uxbridge to Upminster, you can’t stuff any more trains through each leg.
    If you want to provide a simpler service,m something has to give. Abandon the Aldgate-Tower Hill and Gloucester Road-High Street Kensington sections, and cut all the met services back to Baker Street, and what’s left is remarkably similar to the pre-1975 Northern Line: and we all know how easy that is to operate!

    (To spell it out – Earls Court is Camden Town, Edgware Road is Moorgate (although there was never any possibility of through running off the Highbury branch as there is off the Hammersmith line) , Aldgate/Aldgate East is Kennington, and yes, I know it’s the wrong branch that can reverse there) .
    (Indeed, if we include redundant tracck, Ive just noticed that if Shepherds Bush is Finsbury park, Hammersmith would equate to Highgate!)

  55. timbeau says:

    Sorry for double post – I can’t keep up with new ones
    @Centrifugal Kitten
    ” how sensible it would be to just dig a tunnel under the whole circle line and move all anticlockwise trains of all hues down to a lower level.”
    It would have to be a very long way down – there are lots of other lines below the Circle, all crossing under and over each other. In particular, a diveunder at Baker Street is precluded by the Bakerloo and Jubilee Lines being in the way

    @ StephenC
    ” then the entire length of the current Aldgate station could be used for diving” Remember you’ve got to get under the DLR tunnel as well as the District Line .

  56. Castlebar says:

    @ timbeau

    Interesting, and Thank You for your valid and respected opinion

    but I think IMHO you are slightly overcomplicating things as these will take the ‘slots’ of existing Circles, and as these will have a start point which should be less erratic, there should be no problems dovetailing these in with existing Met services on the northern section.

  57. Castlebar says:

    @ spinning moggy 12:44

    The reason your idea of having the two Circle lines at different altitudes would fail absolutely, would be because if a train had to be reversed, how do you get it from one level to t’other??

  58. NG says:

    I used to work in an office where a substantial proportion of the staff started work early and stayed late just to avoid peak-hour congestion. This was as true for those using cars as trains.

    What happens when you provide more capacity in the peak?

    Well, at least some of those former peak-avoiders will now choose to travel in the peak.
    That makes spend to address the peak crush on existing lines very poor value for money.

    Example: I took a train from Waterloo to Putney around 5:45 the other evening and was surprised that it was only four cars and with a seating layout not suited for crush loading so that the station stops were artificially prolonged as people struggled to get on and off at Vauxhall, Clapham Jc etc. Not a nice travel experience.

    South West Trains is going out to tender for more stock as well as introducing re-worked 458 stock and transferred 456’s. This extra stock will provide peak hour capacity but stand idle for the rest of the day as there is already adequate capacity for travel in the off-peak.

    The success of Overground has given rise to a belief that ‘build it and they will come’. But rail infrastructure costs lots of money and then operating the system costs even more and so schemes have to compete not just over the infrastructure cost but the continuing cost of operation. The key points seem to be that the successful schemes are the ones that are game-changers for the areas served and will have good all-day base loadings.

    Expensive alterations to the Circle infrastructure won’t give payback. Nor, in my view, will tube trains to Hayes.

    If the tax-payer subsidises workers travel costs through below-cost fares, the employers in the West End, City and Canary Wharf can get the staff they need at lower cost. You can see why there is a move to complete the circle and get those employers to subsidise new infrastructure. But maybe it shouldn’t stop there and a French-style levy to subsidise operating cost of public transport would make the money-go-round more transparent?

  59. Philip says:


    The hope is that Thameslink and Crossrail passengers will mutually interchange at Farringdon and take some of the interchange load off Kings Cross St Pancras.

    More generally, the Ringbahn on the Berlin S-Bahn is another currently operating continuous circle, which unlike the Tokyo and Moscow examples also has other services partly using it. Does anyone know if there are flying junctions?

  60. ngh says:

    re Philip
    14:30, 4 October 2013


    The hope is that Thameslink and Crossrail passengers will mutually interchange at Farringdon and take some of the interchange load off Kings Cross St Pancras.

    “Hope” – TfL won’t find out till 2018 and have any sensible data to base any SSL changes on till late 2019. Walking distance / time to ultimate destinations being less certain.

    I expect there will be a few surprises…

    Especially if people move to an area in anticipation of changes to the transport and developers build lots of new rabbit hutches homes in anticipation as well.

  61. Pedantic of Purley says:


    The Ringbahn is an interesting one. We will cover that in part 2.

  62. Anonymous says:

    From a quick glance at that ever-reliable source that is wikipedia, the Ringbahn appears to operate 12tph in the peak. I suspect that makes quite a big difference…

  63. SAINTSMAN says:

    Our dear old friend the Circle (or snail). It used to be convenient shorthand for central London. Since zones first appeared “on the map”, zone 1 has practically replaced the yellow bottle. It became a great marketing tool back in the day, showing a practical orbital link connecting (most original) terminals.

    From a marketing perspective its days are surely numbered. With the DLR joining the tube map the centre spread out (as it did on the surface). Then the orange lines of London Overground appeared, offering new orbital choices. Finally Crossrail (s) will appear in a dominant place.

    You may be wondering what cartography and marketing have to deal with real trains. In essence people want to get from A-B via the quickest (and least crowded) route without having to think too hard. The Circle is prominent “on the map” but only offers 7-8tph. The Circle as a “Line” is a confusing relic in a world where peak central underground services of 24tph are seen as low frequency and we push beyond 32tph elsewhere [the short W&C is the exception which proves the rule]. Sharing tunnels leaves the uninitiated confused by multiple names/destinations. Regular users end up with a serious bit of mental planning, waiting for and staying on the Circle, may not always be the quickest option. Smart devices help massively but when its busy who stops to check the quickest route, we trust our instincts and hope for the best.

    As a first step when Crossrail 1 signage gets rolled out, I’d consign the Circle to history showing it as a branch of an H&C (new colour and name?) system (a system like the Met or District).

  64. Graham H says:

    @saintsman – I don’t think people are that stupid…

  65. Whiff says:

    @NGH, Philip – what do passengers coming in from the north on Thameslink and heading west do now? Do many go the extra stop to Farringon for the slightly easier interchsnge with the SSL.

  66. ngh says:

    Re Whiff
    23:23, 4 October 2013

    They don’t and that was kind of my point – what TFL hope and what passengers actually do can be different. Post 2018 if they are heading to Paddington then probably Farringdon to change to Crossrail but otherwise change at King Cross St Pancras like they do at the moment.

    Passengers appear to have won’t the battle over circulation patterns above ground at Kings Cross with NR too.

  67. Robert Butlin says:

    The Berlin Ringbahn has plenty of flying junctions, It also is further out from the centre of town than the circle line and has rather easier curves.

  68. Whiff says:

    @ngh – Drifting slightly off-topic this goes back to one of my bugbears and that is that TFL do not do enough to inform passengers about the varying nature of the different interchanges (although there are some excellent private apps filling the gap in the market). After all how many people realise that when changing between Thameslink and SSL, Farringdon is quicker than Kings Cross St Pancras and can save time even if you have to double back.

  69. The other Paul says:

    To be fair, I’m not sure how easy it is for TFL, or anyone, to indicate that whilst KXSP is a good place to interchange between Thameslink and Eurostar, Southern, Great Northen, Northen, Victoria and Piccadilly, and also a good place to interchange between the Circle, H&C and Metropolitan to the same, it isn’t such a great place to change between Thameslink and Circle/H&C/Met. Particularly as Thaneslink is rarely identified by TfL separately from “National Rail” or possibly “First Capital Connect”.

    If the Thameslink Programme provides the route with the separate identity it needs, along with a place on the standard tube map, this might be easier to deal with.

  70. timbeau says:

    The Thameslink programme will make matters worse in at least one respect – some trains for the East Coast Main Line will be leaving from St Pancras Low Level. Will passengers on either concourse wanting to go to Stevenage, for example, be made aware that some services will go from next door? I appreciate that passengers for the outer reaches of the Metropolitan/Chiltern have always had to manage this problem, but they have had 114 years’ experience and have probably got used to it by now.

  71. Si says:

    The other Paul
    “KXSP is a good place to interchange between Thameslink and Eurostar, Southern, Great Northen, Northen, Victoria and Piccadilly”

    Southern? You mean Southeastern (High Speed).

    And surely those last three are not good places to interchange, but sadly the only place (other than Kentish Town/Finsbury Park) to interchange. From the Thameslink platforms, you are almost better off walking to Euston to get the Victoria line!

    TfL can indicate a “don’t change here for this service” if there’s a couple of options – for instance in-car diagrams showing interchange between the Met and Circle at Liverpool Street, and not at Aldgate. Or Bakerloo/Northern and Charing Cross (is that still done?)

    I think the most stupid one is at Amersham on Chiltern heading northbound saying “change for London Underground”. Southbound is a bit better (though a 23 min wait), however you’d be better changing at Chalfont. I know there are fast peak trains that skip stops south of Amersham, but surely northbound that isn’t valid routing without a ticket valid in zone 9 (and who would have that to get to Chalfont in the peak?) and not having Met passengers is what Chiltern are trying to achieve with these non-stop services?

  72. The other Paul says:


    Southern? You mean Southeastern (High Speed).

    Yes absolutely

  73. Ameecher says:

    How does the Chicago L cope with circular services? I know there are only 2 flat junctions on that but there are a lot of lines and the flat junctions are properly space confined and have some nice conflicting movements. Perhaps I should wait until Part 2 before asking this?

  74. Jeremy says:

    @Ameecher: I don’t believe the Chicago L has ever had circular services just going round the loop. Rather the loop has just been used as a turning circle for services off the outer lines.

    Somewhere I picked up the impression that the early LPTB proposal was a ‘loop the loop’ service – essentially combining the H&C and Circle, with trains running Hammersmith (H&C) – Kings Cross – Aldgate – Victoria – High St Ken. – Kings Cross – Aldgate East – somewhere further East.

    Looking at Barker and Robins, the LPTB 1934 Act authorised a flying (or diving) junction at Cromwell
    Curve/Gloucester Road. I am always somewhat surprised that a set of flying junctions for the various Circle Junctions have not (or have seemed not) to be on LT/LU’s list of future projects (even if towards the bottom), with at least semi-prepared designs ready if the time ever seemed right. While these would be very expensive, if combined with platform lengthening to permit S8 trains throughout the SSL network, how much extra capacity would there be – and how would the price compare with other schemes giving similar increases..

  75. NG says:

    If the fares went up to cover the cost of running the services, how many of these fancy-schmancy ideas for new lines, new junctions would be necessary?
    The only reason that trains are so full in the rush hour is that the alternative is worse and the fare affordable.
    Put up the fares.
    Take out the seats.
    Let people choose, you can sit down for £XX or stand up for £X.
    It is not sensible to build a society that depends on a transfer of wealth from non-travellers to travellers.

  76. Anonymous says:


    Are you the guy from that ‘think tank’ who advocated this type of system in the Guardian some months ago?

    It was torn to pieces then, and deserves to be done so again.

  77. RicP says:

    To deviate a bit from our Circle Line(s), I was amused by comparison of the machinations of our blogs here on LR, DD (and there are other well known blogs on railways available, as one has to say on TV Chat-Shows these days). I was looking up something else in relation to transport operations named SilverLine.
    I harked back to the Jubilee Line being renamed by Horace Cutler’s Tory GLC, to celebrate HRH’s Silver Jubilee – remember those 25 Silver real Routemasters? We could not have a simple name like the Fleet Line, oh no, even though Ma’am herself was not likely to be a regular user, and of course the Vic Line was carefully aligned to avoid Buck House.

    To cut a long story short if you would like a quick look at the issues affecting those on the other side of the big pond, look at these commints from the Washington Post! …
    I hope this works: for the Blog and then: for the new and old Metro Maps.
    It is so interesting to see that over there, they have the same sort of hassles as we have over here. The similarity goes back to when the H&C was separated out from the Met, and also more recent discussions on Met services to Aldgate. This Washington blog argued over how many trains should run over the various branches of the Orange, Blue and new Silver Lines, with it’s future extension to Dulles Airport.
    Sounds all rather familiar to me …..!

  78. RicP says:

    They do work, I’ve just tried them. Two additional points, on the first link, near the top of the page, click on the blue chevron marked Back to get the Washington Post article, and click on all comments. They even mention London!
    Happy Reading! Cheers R

  79. The other Paul says:


    blah blah… it is not sensible to build a society that depends on a transfer of wealth from non-travellers to travellers.

    It’s hard to respond to this nonsense without being rude, but I’ll have a go.

    Basically, several million people travel into and around London by train every day. The majority of these journeys are made by people in gainful employment to get themselves from their place of residence to their place of work. They are journeys which they have no choice but to make, since they require employment and cannot afford to live close to it.

    If any of these people are successfully “priced off” the trains that take them to their work, there are two likely consequences. One is that many will attempt to drive; this is highly likely to cause complete gridlock on London’s crowded roads, leading to pollution and inconvenience in spades, but also hours of lost productivity, costing the economy significantly. The other is that they will have top give up their jobs altogether, similarly costing both them and the economy significantly.

    Mr NG, it is actually not sensible to transfer wealth from people who travel daily to earn and generate it, to non-travellers who sit around at home and expect the rest of the world to function, to feed them, to clothe them, to provide them power, telecoms, technology and a chance to write pontificating rubbish on the internet, without their making a material contribution in the form of travelling somewhere to do something useful.

  80. Graham H says:

    @the other Paul – well said! And while Mr NG is at it, why not advocate stopping subsidy to all those who are uneducated, sick, poor or unemployed – a drain on the wealth of those who are actually earning money?

    Back to topic – @RicP – and worse still, it didn’t even open in the Jubilee year; it was just a blatant attempt to divert attention from the harm that had been done in return for Cutler’s knighthood.

  81. Milton Clevedon says:

    @the other Paul, @Graham H
    Hear hear! The underlying business case for many transport schemes these days (not just railway ones) is the Gross Value Added by enabling a more efficient, higher capacity, better connected economy – and with it the growth in personal and corporate wealth which can then be taxed to pay for uncommercial purposes desired by society as a whole. I also happen to think that a better public transport system is a civilising element in society. Even if the Circle Line, as a complete circle, isn’t necessarily the most useful railway around, it is very handy indeed for trips from quadrant to quadrant.

  82. timbeau says:

    @NG and The Other Paul

    By NG’s argument, it should also follow that no roads should be paid for by the government, as non-travellers are subsidising those people who use them to get to work, or make a living from transporting goods or passengers, or providing on-calls services. Nor, by NG’s reasoning, should government subsidies be made available to deliver the Internet to remote communities.

  83. marek says:

    A thought strikes me about the topology of circular lines. If the problem (to massively oversimplify the discussion) is in large part the lack of availability of ‘spare’ platforms to regulate the service, does that mean that it is not actually the circularity of the Circle line that is the underlying issue?

    The two overground services to Clapham Junction each has a single terminating bay platform. Physically, of course, the trains reverse, but topologically, it is as if there were just a loop (and also equivalent to the Kennington loop, if you ignore the siding): a second train cannot arrive, let alone pass, until the first train has moved out.

    So how far would reinstating four platform working at either or both of Gloucester Road and South Ken, converting one or both bays at Mansion House and Tower Hill and giving the potential to through running for all four Edgware Road platforms (truncating Wimbleware at High Street Ken if necsseary) provide enough operating margin to fine tune line running?

    I realise that doesn’t address the flat junction problem, but I am struggling to see why a circular line should be intrinsically harder to regulate.

  84. Graham H says:

    @marek – a friend of mine who used to operate the District a few years ago told me that he couldn’t understand the LU obsession with trying to undo the Circle; it was straightforward enough to operate from the point of view of regulation; the flat junctions were certainly a performance risk, but that applied equally to the other lines that used the same tracks.

    @Milton clevedon – yes, the Circle caters for a good many journeys that would otherwise be tiresome, such as Victoria to Kensington, etc.

  85. Alan Griffiths says:

    NG @ 22:14, 5 October 2013 link

    “a society that depends on a transfer of wealth from non-travellers to travellers.”

    It’s not wealth, its income.
    If it was wealth, the implication would be that land further out was worth more than land further in; it would be impossible for houses in (olympic) Stratford (centre of World attention) to be worth more than houses in Romford. They are!
    Income is transferred every time a payment of any kind is made.

  86. Pedantic of Purley says:


    It is certainly true that if you had somewhere (actually preferably two or three places) in each direction where you had an island platform with the ability of trains to stop on either side you could then hold the service there for a short while. Effectively you would be giving the circle some elasticity and overcome some of the reasons not to run trains in a circle. The trouble is that although such potential places exist (e.g. Tower Hill, Edgware Road, Aldgate) they are needed for reversing. Other potential sites are constrained by the length of modern stock (e.g. Mansion House).

    There are still two fundamental problems. The first is that you are sharing the line with other trains so every train (28-32tph) gets delayed for the sake of 6-8tph Circle Line trains and in any case would passengers really find the delay acceptable? Remember you are probably standing in a crowded train. Furthermore the extra cost of rolling stock alone could well make this a very expensive solution.

    The other problem is that you may still haven’t got over the need to have an exact number of trains per hour. If the sub-surface lines were frequent everywhere and self-contained that wouldn’t manner. However north of Harrow on the Hill (Chesham especially) they need to operate to a timetable as they are not frequent. You also have the situation where you share with National Rail services (Richmond, Watford Junction from 2016) as well as the Piccadilly Line so you can’t “just do your own thing”.

    @Ian J

    Sorry not to have replied to this before. I don’t think the District Railway really cared about completing the circle. Certainly as built Mansion House did not look well designed for future extension so I don’t think they were at all keen to go east. Only today I came across a comment in a book which echoed my thoughts on the layout of Mansion House station so I am not the only one.

    I don’t think they ever wanted to go via Tower Hill but that was more or less forced on them once they wanted to go to Whitechapel which is where they really wanted to go. Completing the circle was just a nuisance. The Circle Line protrudes quite a bit at this point and this was one of the problems. It was often quicker and cheaper to take an omnibus or even walk from Mansion House to Liverpool Street or Moorgate. So a lot of extra expense with little extra revenue.

    As for going via Bank why not? They were first. There would be nothing in the way to stop them going straight on to Liverpool St.

  87. timbeau says:

    The problem with a circular service, which we also see to a lesser degree on the Kingston Loop, is that even if you have a convenient layby to do it in, any regulation pauses have to take place on a train in service, which annoys the passengers. Somehow, people accept sitting on a train at a terminus waiting for “time” (or indeed sitting on a platform waiting for a train to turn up in the first place) rather more than they do when waiting on a train en route. Perhaps because comparatively few people join, and therefore wait on, a train at the outer London termini.

  88. Mark Townend says:

    @timbeau 19:04, 6 October 2013

    You express beautifully the problem from a passenger point of view. Whilst a short wait at a station for regulation is always preferable to an unexpected stop en-route, a long wait by default is rather annoying.

  89. Ian J says:

    @Timbeau, Mark Townend: the example of this on the Underground would be the way Piccadilly Line trains have to wait for a while at Heathrow Terminal 4 because the loop means there is no conventional layover. In that case it is tolerable because there is a separate direct service to Terminals 1-3, but it must be annoying for people who don’t realise this and find themselves stuck waiting at T4 with a plane to catch.

    @Greg: the nice thing about the cup of coffee urban myth on the Moscow Metro is that it “explains” the choice of line colour as well as shape.

    @Pedantic: Thanks for the explanation – interesting to reflect on how nearly the Circle never happened at all (and the Middle and Outer circles were never “completed”). But the power of a complete circle as an idea is a strong one with obvious political appeal – this must have been a factor in getting money to “complete” the Overground through to Clapham Junction – though TfL are wise enough not to attempt to run a full circle service on it.

  90. The other Paul says:

    I’m not sure that Londoners are that accepting of waiting long on platforms or even at termini for tube trains. It’s all a matter of expectation isn’t it really? Expectation on the tube is that a train will arrive quickly, and once arrived will take you to your stop without any delays at intermediate stations. If the train sits at a station for more than a couple of minutes, people start looking uncomfortable.

    Contrast to the trans-siberian, where a journey takes up to a week and passengers look forward to the longer 20 minute stops at stations as a chance to stretch their legs and buy snacks!

  91. timbeau says:

    I think part of the difference is that people expect the Tube to be a turn up and go, every couple of minutes, service, and don’t expect it to run to a timetable – so it can’t be “early”. “Waiting for connections” is an understandable concept on the main line services (at least to passengers, in the privatised world it seems to be less of a priority for the operators for whom their own punctuality is paramount over that of any passenger requiriung the connection)

    It is equally annoying to be sitting on a bus, usualy one stop from the terminus where everyone wants to go, because the bus is “ahead of time”, thus turning a five minute journey into a ten minute one (or more, if a train is missed as a result) . The passengers sitting on the stationary bus are not ahead of time of course, because if the service had been running to schedule they only missed the previous bus because it too was ahead of time at the point they should have caught it.

    The problem is that regulators see the service as simply an end to end operation, with the intermediate calls as a minor inconvenience

  92. Mark Townend says:

    @Ian J 00:51, 7 October 2013

    Before T5, with only the loop service via T4, the Piccadilly layover was at CTA, where the twin platforms meant an eastbound train could be waiting at one or the other most of the time. Hence T4 passengers going to central London often faced a longish wait at their first stop after departure.

  93. Graham H says:

    Reflecting on the thread so far, let me lob a firework into the discussion: “Why Wimbledon- Edgware Road”? It seems to be a relic of the days when (a) Kensington was a major shopping destination with several department stores, and (b) – probably – something to do with the District getting to Paddington. These days, Paddington apart, the line doesn’t serve any major traffic destinations (stress “major”) . Removal of the service, and some improvements to interchange at Earls Court to soothe pain of Wimbledonians trying to get to Paddington, would not only release resources for redeployment elsewhere on the District, but also remove a substantial number of conflicting moves.

  94. timbeau says:

    I recall this, but the wait at T123 seemed less of a problem because, firstly, passengers from T4 were less likely to be in a hurry (they wouldn’t be trying to catch a plane, anyway). Any who were really pushed for time had the option of crossing over to the other platform if their train got into T123 before the previous one left.

    The next train displays and destination displays on the trains do their best to point people in the right direction (i.e not to use a T4 train unless they are actually going there) by displaying “T 1,2,3 & T5” or “T4 & T1,2,3” as appropriate. It is also explained on the displays that T4 trains can wait there for some time – although it is not mentioned that it is only worth getting a T4 train for T123 if there is no direct train for at least five minutes – and you are quite likely to get there faster even if the gap is nearer ten. (And it is rarely more than this)

  95. timbeau says:

    @Graham H
    The “Wimbleware” service is actually a post-1933 extension of the District Railway’s Wimbledon – Kensington HS service deep into Metropolitan territory, which ran at different times to Edgware Road, Kings Cross, Moorgate, etc. Presumably, its introduction required a reduction in some other services on the west (and north) side of the circle to fit them in – probably the Met’s services to South Kensington.
    I’m not quite sure what you are proposing – if the Wimblewares go, you either have to reduce the frequency on the Wimbledon branch, or send all Wimbledon trains over the south side of the Circle, at the expense of some other service. Likewise, you would have to double the Circle Line’s frequency if you were to keep the same level between HSK and Paddington – and my occasional experiences of that section suggest that both the Circle and Wimbleware services are in general well-loaded.
    If the Wim-ERd service were to go, improvements in interchange at Earls Court would not “ease the Wimbledonians’ pain” – they would have to change at Gloucester Road (as would any other District and Picadilly passengers from the west for Paddington, unless they wanted to run the gauntlet of the traffic on Hammersmith Broadway)
    Don’t forget also that HSK has a comparatively big hinterland – especially now that Olympia is less well connected than it was: and HSK is the easiest way to get to and from the Royal Albert Hall – reducing services to HSK would inevitably lead to even more crowding at South Kensington (which is marginally closer but has much poorer bus connections) than there is already.

  96. Will says:


    “CTA”? Central Texas Airport? (IATA) Catania? Or maybe Cairo Transportation Authority? Pretty big diversions for the Piccadilly line, in any case. 🙂

    OK, I’ve now learned it means “central terminal area”. So it’s been an educational experience. But boy, these abbreviations don’t half make life hard work!

  97. Graham H says:

    @timbeau – sorry, I meant Gloucester Road, of course. What you do with the resources released is the nub of the discussion – one possibility is to add (oh dear!) another branch to the District – Mansion House, say, to HSK or EdR. I am not convinced that Wimbledon isn’t “over-trained” these days – whenever, I (admittedly) sail through on SWT at peak times, there invariably seem to be 3 Districts in the platforms, another one leaving with perhaps ten passengers per car, and one waiting to enter (load ditto). Linking Wimbledon (Id put this in italics if I knew how…) to Edgware Road seems a mere operational convenience, whereas better southside Circle frequencies to Paddington seems a better prospect and serves a bigger market.

    I take your point about the enlarged HSK catchment area and the Albert Hall (haven’t we all struggled from the AH to catch the last train from Waterloo etc?) – but surely the AH traffic has a very short and spiky busy time? And Olympia has had too erratic a service over the years to develop a reliable clientele?

    @Will – offtopic – welcome to the UK rail industry and its obsession with TLAs. [BTW as a very minor entertainment during the BR privatisation, when I had to sign 2500 station leases and about 6000 access agreements, each in sextuplicate, one of the games for lawyers was to try and spot as many three letter stations (TLSs) as possible – – I believe the count got to 8.]

  98. Pedantic of Purley says:

    @Graham H

    It is <i> to put put text in italics and </i> to end the italics.

    Welcome to the world of HTML (in the computer biz we tend to use ETLAs)

    ETLA = Extended Three Letter Acronyms which is taken to mean four letter acronyms.

  99. MJG says:

    @ James C

    While walking on the West Cromwell Road bridge over the WLL/District on Friday I saw an S stock train leaving Olympia. The destination board said it was heading to Tower Hill. First District line service running with S stock, or a strange anomaly? (Or was I seeing things?)

  100. timbeau says:

    @Graham H
    Although a significant number of people take the District all the way to/from Wimbledon, judging the number of trains needed on the branch only by observing the number of people at the terminus is to fall into the same trap that I mentioned in my 09:12 post – I don’t know which line you use when you pass through Wimbledon on the way to Waterloo, but doubtless the number of people going all the way to Alton, Chessington, Dorking, Exeter, Hampton Court, Portsmouth, Shepperton, Windsor or Weymouth would not, on their own, justify the number or length of trains provided to those destinations (and the number of passengers going all the way from Waterloo to Basingstoke or Guildford on a stopping service must be vanishingly small!). On the District’s Wimbledon branch, the trains fill up very rapidly, especially after they cross the river.

    Doubtless a better south/west Circle frequency would be welcome by some. Little alteration to interchange arrangments would be necessary, as all necessary interchanges for journeys between between Earls Court and HSK can be made cross-platform at either South Kensington or, in some cases, Gloucester Road. It would also take the guess work out of whether it would be quicker to do two sides of the triangle or wait for a direct service. But doubling the frequency on the south side of the Circle means reducing the frequency on the District – and unless you’re going to thin the service to Wimbledon even more than the halving that the loss of the Edgware road service would require, that probably means closing the Richmond branch or reducing Ealing Broadway to a shuttle.

    Put another way, the trains from Paddington have to go somewhere after HSK, and there isn’t room for all of them on the route through South Kensington, so they have to go to Earls Court. Given that there are faster routes from Paddington to Hammersmith (and Ealing Broadway), it makes sense to send these trains to Wimbledon. This also means that the south side of the Circle can send equal numbers of trains to the three western District destinations (and HSK), whilst also maintaining equal headways both through Hammersmith and Fulham.

  101. timbeau says:


    An Olympia service in daylight hours on a weekday? You must have been imagining it!

  102. Graham H says:

    @timbeau – I quite understand that the trains fill up after Wimbledon – in cost terms, it would be better therefore to thin out the frequency south of, say, Parsons Green, therefore. Not sure whether it’s inevitable , or desirable that each District branch should necessarily get the same level of service (I can see some merit in an ABC(D) destination pattern, but that rarely occurs in practice).

    I am surprised that there aren’t any further paths available GRd- Man Ho; a quick look at Dist 143 suggests that there are 12 eastbound trains in the half hour 1800 to 1830 (alright, 13 if you go to 1830 1/2); is it really down to 24-26 tph?

  103. Long Branch Mike says:

    @Graham H

    “sail through on SWT at peak times” – South West Trains? – Swindon Working Timetable?

    “BTW as a very minor entertainment” – By The Way? – British Thames Works?

    Can everyone please write out all TLAs & ETLAs in their first use to be clear? Your posts lose a lot of clarity and effectiveness when readers are still wondering what you are referring to.

  104. timbeau says:

    @ Graham H, If anything, westbound should be busier than eastbound in the evening peak, and the morning peak is even busier, as it is more concentrated.

    According to the public timetable, in the busiest half hours there are fourteen trains each way at Embankment (including three Circles), which is 28tph. Given the dwell times at the busier stations I doubt if you could squirt many more through. (Will S7s, with fewer but wider doorways than D stock – 7×3 instead of 6×4), make matters better or worse?)

  105. Graham H says:

    @timbeau – Ho hum! The implication of what you say is that it’s is virtually impossible to alter the District service (and therefore by implication, the Circle and H&C) any more. Which is a pity… Dr Pangloss rules OK! [We all knew that any way!].

    A propos doors, I imagine LU have undertaken the requisite modelling using Legion or similar to come up with the reversion to 2 door format. Anecdotal observations suggest that a wider single leaf door was really no better than a normal single leaf – it’s still one at a time boarding/alighting; two normal doors seem to allow up to 2 at a time depending on whether there are “sentinels” on either side of the door setback. On that basis, S7s ought to be noticeably quicker but I haven’t yet been out to watch.

  106. ngh says:

    re timbeau
    15:35, 7 October 2013
    Will S7s, with fewer but wider doorways than D stock – 7×3 instead of 6×4), make matters better or worse?

    Not as simple as that:
    The D stock has single leaf doors and the S stock double leaf doors so the S stock doors should open and close quicker.
    Also announcements on which side the doors open (Greg will be along shortly) for less regular travellers and having a uniform stock with the same door locations not the current C/D stock mix at some stations will help regulars.
    S Stock has less seat padding and more room in aisle so should have better passenger flow.

    There are probably plenty of others.

  107. mr_jrt says:

    Regarding the discussion of overcapacity on the Wimbledon branch…

    …the relevant parts of my proposal for unlocking capacity are to remove the bottleneck between Praed Street and Edgware Road to enable to Wimbleware service to terminate conflict-free there and to alter the junction west of Earls Court so the Wimblewares have dedicated use of platforms 1 & 4. Given a segregated line all the way from Wimbledon to Edgware Road that could thus run a maximum capacity (that is apparently far higher than what is already deemed oversupply), could you then make an argument for building the grade separation required to take the District over/under the SWML and into platforms 9 & 10, thence down to Sutton (new terminal platforms on the other side of the road bridge perhaps?). That frees up paths through Mitcham and Sutton for more mainline services via Epsom and free Thameslink of its Wimbledon loop shackles. Tooting is the only real loser I suspect…not sure what can be done there short of the CR2 dogleg or a Tramlink extension from Wimbledon to make the double-back interchange negligible (or you extend the Tramlink extension to the mainline at Mitcham to use the enhanced services calling there.).

    …these increased mainline services via Mitcham could well even be Thameslink ones, a restoration of the Guildford branch perhaps? 🙂

  108. Ian Sergeant says:

    If it were possible to run Wimbleware trains with their own tracks from Praed Street to Edgware Road, why would you not use the two southernmost platforms to avoid any possibility of conflicts with H&C and Circle trains? I know you could dive under, but why bother?

  109. Will says:

    @Ian Sergeant: The only reason I can see is if one had it in mind to provide cross-platform interchange. And to do that properly would require a turnback siding (or two) to the east of the station.

  110. mr_jrt says:

    @Ian Sergeant, using the southernmost platforms is actually what I though I implied. The flyover I was referring to for alteration is the one west of Earls Court, and the new flyover/under I’m proposing is north of Wimbledon. Not to mention, though I think they’ve been used for something else now, the former sidings would have made excellent terminal bays (or just a handy 3rd platform).

    That said – the obstacle to 4-tracking the line may well be the width of the road above. If there were the case a solution suggested previously was to divert the line from Bishops Road to meet the Circle closer to Edgware Road, probably by following the canal basin. Another alternative could be to excavate from Praed Street to Edgware Road’s throat and have the H&C run along underneath the Circle (or vice versa). I suspect there would be gradient issues that would make that a non-starter, but worth a muse if nothing else.

  111. Will says:

    @mr_jrt: If we’re getting the cheque book out to the extent of contemplating a new route for the H&C between Bishops Rd and Edgware Rd, wouldn’t it make sense to carry it on to Baker St, if at all possible? Services from HSK could terminate at Baker St (in the existing platforms 5 & 6) instead of Edgware Rd, with the new H&C tracks to Baker St merging with the Metropolitan east of Baker St Jn – thus eliminating conflicting movements at both Praed St and Baker St, and also incidentally enabling interchange between Wimbleware services and the Metropolitan and Jubilee lines.

  112. timbeau says:

    ………………………..have the H&C run along underneath the Circle (or vice versa).
    Interesting idea, but I’m not sure whether the Bakerloo and/or Crossrail tunnels would be in the way. But it would almost certainly require the Praed Street line to terminate in a deep level dead end at Edgware Road (or possibly Baker Street): there is no width available between EdRd and Baker Street to join them up again.

  113. Alan (28481k) says:

    I have thought long and hard at Castlebar’s suggestion of treating the Circle Line as the Chicago ‘L’ Loop where trains from either West or East London loop around the circle and reach back to their respective origins. I have even drawn a diagram of such possibilities, but even without timetabling, I can see several disadvantages against this idea.

    1. Chicago Loop is a 1.79 miles (2.9km) double track loop whereas the original Circle Line is 14.5 miles (23.3km). This means unlike the ‘L’ trains looping around the Loop in relatively short time, these District and Circle Line trains will spend up to or even more than an hour looping around the Circle Line, which massively increases the service distance of these services and require more trains to serve the circle. Even at 6tph per branch (let’s suppose Ealing Broadway-Circle-Richmond and Upminster-Circle-Barking), this would require so many more trains (let’s say journey time increased by 30-40 minutes to route around the circle so at 6tph you need 3-4 trains more per direction at least, which the East London Circle would require more trains significantly due to the distance) because each service is longer than used to be and unless S7 stocks fully plenish the SSL I doubt there are enough trains (both C and D stocks) to cover all these diagrams.

    2. With so many crossovers at flat junctions (except the Earl’s Court diveunder from High Street Kensington), it increases the probability and propagation of delays due to points failure or even operational delays.

    3. After thinking a way to fulfill Hammersmith service, I tried to do either Hammersmith-Edgware Road-Kings Cross St Pancras-Aldgate-Embankment-Victoria-Gloucester Road-Earl’s Court-West Brompton-Wimbledon, or the current T-cup service. It seems that the former service will require Wimbleware services to connect the missing gap and simply putting Pread Street Junction into superdrive. Whereas the T-cup service is neat and not requiring Wimbleware to fill in the missing service gap, but if that’s the case then T-cup service can be served alongside Wimbleware service without altering District Line services.

    4. I haven’t even introduced Metropolitan Line into the mix yet, and adding this to the mix would mean 30-32tph between Baker Street and the Minories if Metropolitan Line are service 6-8tph beyond Baker Street. (This presumes the remaining 24tph are distributed as follows: 6tph Hammersmith and City, 6th Wimblesmith via Circle or T-cup, 6tph West London and Circle, 6tph East London and Circle) You can imagine that section of the track would be the worst timetabling and signalling nightmares.

  114. timbeau says:

    End to end Richmond- Ealing or Barking – Upminster journey times would be, I guess, about 30 minutes longer than the existing Upmister- Ealing journeys.

    With an an Ealing-Aldgate – Richmond and Upminster – HSK – Upminster service as castlebar suggest, and existing frequencies over each leg of the SSL network, existing capacity and stock availaibity suggests a Wimbledon – Tower Hill service, with half the trains extended to Barking (or you could swap the eastern termini, e.g run all Wimbledon trains to Upminster, and a service from Barking via Baker Street to Tower Hill).
    This leaves the Hammersmith and Met main services. Basically, if you run the two loop services suggested by CB through Edgware Road and Baker Street, it has to be at the expense of something else. The bottleneck is not, actually, Praed Street Junction – if the double track section from there to Edgware Road can cope, it would be able to do so even if all trains were to run beyond Edgware Road. Trains terminate at edgware Road because something has to stop short of the junction east of baker Street. You could therefore run either the Hammersmith or Met main services to Aldgate, but the other would have to terminate at Edgware Road or Baker Street respectively if the north side of the Circle is not to be overloaded.

  115. Dstock7080 says:

    @Pedantic of Purley
    “Note the diamond in the station name sign which was uniquely a Metropolitan Railway feature”
    i understand from a recent LURS presentation that the East London Railway pre-dated the MET with the use of a green diamond.

  116. Pedantic of Purley says:


    I don’t doubt you may well be correct. Perhaps I should have said red diamond but it would have been a bit silly to state “Note the red diamond” in a black & white photograph!

    What I really meant was before the formation of the LTPB the underground was basically a combination of the Metropolitan Railway and the Underground Electric Railways Company of London (UERL) (est 1902). The latter adopted the roundel so familiar today. So from early in the 20th century you could be pretty sure that any picture with a diamond in it was then or earlier in Metropolitan Railway ownership.

Leave a Comment

In order to make LR a pleasant place for discussion, please try to keep comments polite and, importantly, on topic! Comments that we feel do not meet these criteria, or that contain language that could cause some people trouble at work, may be moderated or deleted.

acceptable tags

* (This won't be shown, but you can link it to an avatar if you like)

Recent Articles

Friday Reading List – 24 March


As anyone looking to properly understand London’s transport needs and network knows, context, background and best-practice are important. As readers might imagine, behind the scenes here at LR Towers we thus spend a lot of time sharing links and reading

Read more ›

Friday Reading List – 17 March


As anyone looking to properly understand London’s transport needs and network knows, context, background and best-practice are important. As readers might imagine, behind the scenes here at LR Towers we thus spend a lot of time sharing links and reading

Read more ›

LR Magazine Issue Five: Overgrounded


With print copies now being prepped for dispatch to subscribers at LR Towers, London Reconnections Magazine Issue 5: Overgrounded is now available to purchase in our online store. Transport is politics, politics is transport You don’t get transport without politics.

Read more ›