11-03.-2003   G Harman tells us: "Near me there is a roundabout which has an area intended for cyclists marked with a different colour surface and white line border around part of its periphery. Today I seen one of the entrance roads has gained a new electronic sign,positioned adjacent to the Give way line to the right of a joining vehicle." Hmm. Interesting. Do keep us posted on further news. My cynical view is that they wouldn't have spent the money for the sign unless there was a "real" safety problem. safety problems only become "real" if there are actual dead and/or injured cyclists (or motorists ete.). I think that improvements only get installed if the money saved in dead etc. cyclists is likely (according to the knowledge in cycling matters of those who installed the sign) to be "repaid" in just a year or so. I think a dead cyclist counts as about 1.2 million pounds nowadays, a seriously injured cyclist as about 1/10 a dead cyclist, and a slightly injured cyclist as about 1/80 a dead cyclist. Adults and children count the same, as far as I know. I wonder how much the sign cost. The line-round-the-edge-route round a roundabout is described in rule 62 of the Highway Code, which contrasts it to what rule 61 calls the "correct" procedure for cyclists to use a roundabout. It would be interesting to know what turns the signs on, and how many seconds warning it gives motorists to take action. Jeremy Parker

Cyclists are 14-16 times more likely to be involved in a crash on a roundabout than a car driver (Maycock and Hall, 1984) and a full 50% of these collisions are due to a driver entering a roundabout and hitting a cyclist who is already circulating and supposedly has 'right of way'. A further 10 % of collisions are due to drivers exiting a roundabout cutting across the path of a cyclist and an additional 10% are due to a
cyclist being run down from behind. (Layfield and Maycock 1986). The best available evidence suggests that as many as 20% of all cycle crashes occur at
roundabouts, though for reported 'accidents' the figure is nearer 10%.

Roundabouts are 4 - 8 times more dangerous for cyclists then traffic signal controlled junctions.
Roundabout safety can be improved by adopting 'continental' styled roundabouts. However, no
treatment is as effective as replacing a roundabout with a properly designed signal controlled junction with advanced stop lines for cyclists. No cycle specific facilities have been identified that improve the attractiveness or safety of roundabouts for cyclists and many, including cycle lanes marked around the periphery of a roundabout, may actually increase the hazards faced by cyclists.

Roundabouts and the 'Highway Code'
Although the public roads system is supposedly provided for the benefit for all road users, it is
very apparent that in the case of roundabouts the convenience of motor vehicle users is given precedence over the safety of cyclists (and indeed motor cyclists). The 'Highway Code' rovides further evidence that roundabouts do little more than undermine the cyclists right to use the public road.

The 'Code states:
Rule 61. ..'Roundabouts can be hazardous and should be approached with care.

Rule 62. 'You may feel safer keeping to the left on the roundabout or dismounting and walking your cycle round on the pavement or verge'. One is drawn to wonder whether highway engineers would get away with installing features that necessitated
advice to the effect that drivers would be advised to
avoid using the highway...

The 'advice' given with regards cyclists 'keeping to
the left' is possibly one of the most dangerous pieces
of 'safety' advice given in the 'Code. All other
vehicles executing a right turn are advised to
approach and circulate on the right hand side and
staying on the left increases the chance that a driver
leaving the roundabout will cut across your path. In
general, the best approach when riding on a roundabout
is to claim ones road space, keep to the right if
turning right and to avoid riding near to the entry
points in case a driver enters the roundabout into
your path.

Research on Roundabouts and Cyclists
One of the best sources of information on cyclists and
roundabouts is still  the report 'Cyclists and
Roundabouts. A review of Literature' commissioned by
the Cyclists Touring Club in 1990 and compiled by the
Consulting Engineers Allott and Lomax. This report
conclusions and recommendations include:


'Greater awareness of the serious safety problems for
cyclists and motor cyclists at roundabouts is needed.'

'Current roundabout design advice and practice gives
insufficient weight to safety factors, particularly
for cyclists and motor cyclists. Standards should be
improved. In some cases roundabouts will be
inappropriate.

'A substantial programme of remedial measures to
improve safety at roundabouts is justified. This
should include speed reducing traffic calming
techniques.'

'In general roundabouts pose very considerable
accident risks to cyclists both in absolute terms and
when compared to other junction types. The decision to
install a roundabout instead of traffic signals is
based on criteria which give insufficient weight to
the needs of cyclists, motor cyclists, pedestrians and
bus users.'

'Vehicle entering/cyclist circulating accidents are by
far the most common type of cyclist accident at
roundabouts. Given that circulating traffic has
priority over entering vehicles, it would appear that
drivers are predominantly at fault.'

'The cycle facilities introduced at roundabouts to
date do not suggest that a widespread 'solution' in
terms of a specific facility is possible...'


Other items of interest in the report include the
following;

Small roundabouts (5-40m in diameter) with flared
entries are the most dangerous for cyclists.

Large roundabouts (up to 70 m in diameter) are most
feared by cyclists but the statistical evidence is not
available to show the nature of the relationship
between the type and occurrence of accidents and
roundabout size. The report notes that such a lack of
data is unsatisfactory and that further research
should be done to show the precise relationship
between accidents and the geometry of a roundabout
including its size.

'Mini' roundabouts are 'safer' then other roundabouts
having a similar recorded accident rates to 4 arm
traffic signal junctions. (However, see the statistics
section and the section on mini roundabouts below.*)

Roundabouts affect the safety, enjoyment and route
choice of cyclists with 28% of cyclists reporting
having avoided using a route because it involved using
a roundabout. (Watkins 1984) Roundabouts may also
deter cyclists from choosing to travel by cycle at
all.

Cycle specific facilities at roundabouts 'may have
site specific benefits but do not provide a generally
applicable solution'. For example, segregated left
turn lanes increase the hazards faced by cyclists
wishing to go straight ahead or make a right turn.

'No satisfactory method has been found for reducing
the risk to cyclists once they have entered the
circulatory system.'

Studies conducted in Hertfordshire showed that the
outer 1.5m of a roundabout were most hazardous for
cyclists. This study suggested hatching out the outer
areas of a roundabout so as to encourage cyclists to
ride away from these areas. (A recommendation which
those 'traffic engineers' who think peripheral cycle
lanes are a good idea seem to have overlooked...).

Roundabouts may be created by developers purely
because of the 'image' they convey. 'Roundabouts imply
freedom for drivers whereas traffic signals imply
control of drivers and delay, even though this
perception may be untrue when objectively measured.'
A very important point as it goes a long way to
explain the attraction of roundabouts to drivers and
car orientated highway engineers.

(That roundabouts aid traffic flow is often a fallacy
is well illustrated by the installation of traffic
lights on heavily used roundabouts. Where the traffic
flow is dominantly in a single direction other feeder
roads may find it very hard to get access to a
roundabout without signalised control. In addition, as
traffic speeds on roundabouts tend to be high their
capacity may be surprisingly low as vehicles cannot
risk entering the roundabout until a large gap in the
traffic appears. In comparison, traffic signals allow
large 'blocks' of vehicles to be moved in turn. The
report itself notes that roundabouts do not give any
control over traffic flow whilst linked traffic
signals that respond to changes in traffic patters and
demand can be used to maximise traffic flows.)

Roundabouts generally do not cater for pedestrian
movements.

Design features intended to maximise vehicle flows and
speeds are directly responsible for increasing the
hazards faced by cyclists. These include the use of
flared entries, having excessive entry capacity,
having wide circulatory carriageways and the use of
large roundabouts. The use of low entry angles (ie.
below 40 degrees) also forces drivers to attempt to
merge by looking over their right shoulders or even by
looking in their mirrors. This again increases the
risk of 'not seeing' a cyclist already circulating the
roundabout.

With regards off-carriageway cycle lanes the report
notes:

'Cyclists travelling straight ahead are less likely to
divert from the main carriageway due to the number of
stops and give way manoeuvred necessary....the need to
give way when crossing entry and exit arms poses
problems when motor vehicle flows and speeds are high.
Furthermore, peripheral cycle tracks are unlikely to
be suitable for urban sites where frontage development
and high pedestrian flows exist.'

Roundabouts and accident frequencies
As has already been noted the risks to cyclists posed
by driver behaviour at roundabouts is undeniable.
Maycock and Hall, (1984) showed that cyclists are
14-16 times more likely to be involved in a crash on a
roundabout than a car driver. Layfield and Maycock
(1986) showed that a full 50% of these collisions are
due to a driver entering a roundabout and hitting a
cyclist who is already circulating and supposedly has
'right of way'. A further 10 % of collisions are due
to drivers exiting a roundabout cutting across the
path of a cyclist and an other 10% are due to a
cyclist being run down from behind.

Downing (1985) states that 8% of reported cycle
'accidents' occurred at roundabouts. The DETR Traffic
Advisory leaflet 'Cyclists at roundabouts: Continental
Design geometry (9/97) gives a figure of 10% of all
reported accidents with 11% of these involving fatal
or serious injuries to the cyclist. However, other
studies show that the actual number of cycle accidents
occurring at roundabouts is higher then this. This
reflects the gross under-reporting of cycle crashes,
especially those  involving slight injuries.  (Pedder
et al [1981] reports that as many as 87% of minor
cycle crashes go unreported.)

Lawson (1989) found that 22% of accidents involving a
cyclist occurred at roundabouts. Watkins (1984)
conducted a survey of 2000 C.T.C. members who had been
involved in a crash and found that 13.8% of accidents
had occurred at 'normal' roundabouts and a further
5.2% at mini roundabouts, figures in line with
Lawson's study. (Quoted in the C.T.C. report 'Cyclists
and Roundabouts, second edition, 1993).

Whilst working within the Traffic Management section
of the East Riding of Yorkshire Council I was very
concerned that approval was being given to the
creation of new roundabouts without any consideration
being given to the danger this would create for
cyclists. As all the available research was
effectively dismissed out of hand (one officer doing
so on the basis that the research wasn't done in the
East Riding of Yorkshire!)  I was forced to look at
the authorities own accident figured for support.

I found that there were 38 traffic controlled
junctions in the East Riding with a similar number of
'urban' roundabouts. Naturally, there were also a
number of roundabouts on main rural roads but my main
concern was with those roundabouts found in built up
areas as cyclists had little choice but to use these
when making utility journeys.

I collected data from the KEY police accident computer
for the 3 years preceding 19 January 2001 and found
the following:

Total number of cycle casualties.                    
 406.    Fatal 6,    Serious 61,    Slight 348.

'Non junction' cycle casualties.                     
    144.    Fatal 5,    Serious 27,    Slight 116.

All 'junction' cycle casualties                      
      262.    Fatal 1,    Serious 34,    Slight 232.

'Junction' casualties (Roundabouts and
traffic signal controlled junctions excepted)     
216.    Fatal 1,    Serious 29,    Slight 186.

Cycle casualties occurring on a roundabout.        46.
  Fatal 0,    Serious 5,      Slight 44.

Cycle casualties occurring at a
traffic signal controlled junction.                  
         2.   Fatal 0,    Serious 0.      Slight 2.

A closer analysis of the data for the 'fatals' shows
that the most common cause of a cyclist fatality was a
driver running down a rider from behind whilst they
were cycling on a rural road, with 4 of the fatalities
being due to this. This is perhaps to be expected the
large rural area the authority covers, the high speeds
drivers travel at on rural roads and the fact that
many drive far above any speed which would enable them
to stop within the distance they could see to be
clear.

As is to be expected the majority of cycle crashes
occur at junctions. However, it needs to be borne in
mind just how many 'T' junctions, Cross Roads and so
on there are in an area the size of the East Riding
and its constituent town such as  Beverley,
Bridlington, and Goole. Given this it is immediately
apparent that roundabouts account for a
disproportionately large number of casualties. It is
also clear that the 'roundabout' and 'signalised
junction' data contrasts strongly.

Practically every urban roundabout in the whole
authority had a recorded injury accident within the
previous 3 years. (One can only wonder as to the
number of unrecorded crashes). One roundabout had 4
cycle casualties, one 3 cycle casualties and four had
2 cycle casualties. Where further data was recorded it
was apparent that the actual cause of these crashes
followed national trends with the largest number  of
crashes being due to drivers entering a roundabout and
hitting a cyclist who was already circulating.

Although the 35 roundabouts constituted a very small
minority of all the junctions in the county, they
accounted for 11.3% of all cycle casualties. (A figure
in line with Downings 1985 figures for reported cycle
crashes). With a total of 46 casualties the injury
frequency per roundabout was 0.43 cycle casualties per
junction per year.

Eleven (24%) of the reported 'accidents' occurred at
mini roundabouts
In stark contrast the authorities 38 traffic
controlled junctions accounted for only 2 slight
injuries, both the result of a driver skidding their
car at the junction on a wet road. With 2 casualties
the injury frequency per signal controlled junction
was 0.02 cycle casualties per junction per year. This
means that the accident frequency for roundabouts was
26 times higher then for traffic signal controlled
junctions. In addition, 5 serious injuries occurred on
a roundabout but not one at a traffic signal
controlled junction. Even if all the authorities
roundabouts were included in the figures it is very
apparent that cyclists are far more likely to be
injured whilst using a roundabout then they are when
using a traffic signal controlled junction.

It is also the case that many of the available
comparative statistics for roundabouts and traffic
signals (including those above) relate to 'old style'
traffic signals which do not have features such as
advanced stop lines for cyclists. As such features can
make signalised junctions even safer for cyclists then
they already are it is obvious that the size of the
'safety gap' between roundabouts and signalised
junctions is often understated as the effect of ASL's
and so are are often missing from the data. (And of
course, no design feature has been shown to improve
the appalling casualty rates for cyclists at
roundabouts.)

Of course all this data had no effect whatsoever on
the desire of the Highways section to build yet more
roundabouts. Even less was it going to persuade anyone
to remove any roundabouts removed in order to replaced
them with traffic signal controlled junctions.
Thankfully, at least Hull City Council was pressing
ahead with replacing some of its roundabouts with
properly designed traffic signal controlled
junctions...


Roundabout Design
This section considers roundabout design in greater
depth in order to understand just why they have such a
poor safety record for cyclists. (Other than drivers
failing to drive in a manner that gives due regard to
cyclists safety of course!).
  
Features of a UK roundabout
Primary design consideration is to maximise vehicle
speeds and capacity of roundabout. This is done by:

1) Having multiple entry and exit lanes.

2) Having multiple circulatory lanes  (which may or
may not be marked out).

3) Vehicles approach the roundabout at a low angle of
deflection. This means they do not have to 'turn' onto
the roundabout.

The above design considerations create the following
problems:

1) Vehicle speeds through the roundabout are high, so
reducing the margin for error and increasing the
severity of any crash that does occur. (Naturally,
this is of particular significance for vulnerable road
users such as cyclists.)

2) The multiple entry, exit and circulatory lanes mean
that there is the potential for vehicles to come into
conflict, either through changing lanes or encroaching
on the path being used by another vehicle. Many
vehicles will use the full width of the circulatory
carriageway in order to straighten their path through
the roundabout and maximise their speed.

3) The low entry angle makes it difficult for drivers
to make and adequate observation to the right prior to
entry and encourages a high entry speed.

4) The  use of multiple entry lanes creates the
possibility that  the view of a vehicle in the
'inside' lane may be obscured by one in the 'outside'
lane. (In the diagram above the yellow car is
obstructing the view of the red car).

5) Cyclists are especially at risk when making a right
turn as they must either remain towards the outside of
the roundabout, and so risk being cut up by drivers
leaving the roundabout, or remain towards the centre
of the roundabout and risk being struck by vehicles
coming up from behind at a higher speed.

Continental' roundabout design
The diagram above shows a much improved roundabout
layout which has been shown to markedly improve the
safety of cycle users, albeit at a 'cost' of reducing
the capacity of the roundabout. The significant
features are as follows:

The inner kerbs have been built out (marked in black -
4), the central island has been moved forward (2) and
hatching laid in the centre of each lane (3) . In
addition the size of the central island has been
increased (1). These changes have the following
effects.

1) The roundabout now has only a single entry lane (5)
and a single circulatory lane. This reduces the
possible number of different paths that can be taken
on the approach to and through the roundabout.

2) The single entry lane means that a cyclist wishing
to make a right turn can position themselves more
easily and will also be well over to the right before
they enter the roundabout.

3) The single entry lane means that the view of a
vehicle waiting to enter the roundabout cannot be
obscured by another vehicle.

4)  The buildouts and hatching have increased the
entry angle markedly so that vehicles must slow and
'turn onto' the roundabout. vehicles also are in a
much better position to make an effective rightwards
observation. Note position of the red 'car' (6).

The image below shows such a modified roundabout to be
found on Kingston Road/ Willerby road in Hull. a fine
scheme this saw injury accidents fall from 59 in 3
years to 1 in 18 months. And it was the work of the
East Riding of Yorkshire Council's Highways section!
The drivers complained of course but, for once, a job
well done...

On-Highway Cycle Facilities
Generally speaking, no on (or indeed off) road cycle
facility has been shown to be effective in reducing
cycle casualties at roundabouts. They may well
increase the risks posed by cyclists.  However, it is
often necessary to integrate a roundabout into a
scheme which includes cycle lanes. Although each
situation is likely to be unique due to local
conditions the following general 'rules' apply.

1) Cycle lanes running around the periphery of the
roundabout are A VERY BAD IDEA. They encourage
cyclists to use the most dangerous part of the
roundabout and cause particular problems for cyclists
wishing to make a right turn on the safest part of the
roundabout (ie. the right hand lane) as drivers may
resent cyclists who fail to use the 'facilities' as
provided.

 Where  it is felt necessary to link on road cycle
lanes or to mark a coloured surface cycle lane through
the roundabout (for example, to signal to drivers that
cyclists might be encountered), the lane should be
laid along the line actually used by cyclists. This
may mean leaving a 1.5 m or greater gap between the
'Give way' line and the inside edge of the cycle lane
so as to move cyclists away from the danger posed by
vehicles entering the roundabout and in order to give
the cyclists the straightest possible run though the
roundabout.


2) Where cycle lanes are laid on the approach to a
roundabout 2 strategies, neither ideal, may be
adopted.

a) Discontinue the lane prior to the roundabout so
legitimising the cyclists use of the whole approach
lane and so making it easier for cyclists wishing to
make a right turn (potentially the most hazardous
manoeuvre) to position themselves correctly.

b) Continue the lane up to the roundabout so that
cyclists going left or straight ahead have a clearly
delineated area of the road which is to some degree
protected from encroachment by vehicles. This will
also allow a wider entry so preventing cyclists from
being 'squeezed' by any vehicles alongside them whilst
at the same time ensuring that drivers are encouraged
to enter the roundabout at a high angle.

Off-Highway cycle paths
It may be recalled that the report 'Cyclists and
Roundabouts' (1993)  had the following to say of
off-carriageway cycle lanes:

'Cyclists travelling straight ahead are less likely to
divert from the main carriageway due to the number of
stops and give way manoeuvred necessary....the need to
give way when crossing entry and exit arms poses
problems when motor vehicle flows and speeds are high.
Furthermore, peripheral cycle tracks are unlikely to
be suitable for urban sites where frontage development
and high pedestrian flows exist.'

Of course, Traffic management 'experts' often see no
need to follow such advice. The following highlights
the shortcomings of a typical roundabout off highway
cycle 'facility'.  The roundabout concerned is to be
found on Springfield Way in Anlaby, just to the west
of Hull. The cycle path is filled with turns so tight
one cannot negotiate them without running over onto
the grass. A large section of it is also marked as
being a cycle path only. (Not that this dissuades all
the local pedestrians from using it...)

To make matters even worse the off highway cycle path
on this roundabout was laid using non approved hazard
warning tiles which have closely spaced round ribs
rather then the low widely spaced and flat ribs found
on approved cycle path tactile paving slabs. These non
approved tiles are very slippery to ride over,
especially in the wet and are unfortunately common on
cycle routes. Despite repeated requests to replace
these tiles with the approved type no action has been
forthcoming from the East Riding of Yorkshire Council.
However 'Deaf Eric' has show its concern for the
safety of cyclists by producing a leaflet that pointed
out all the locations where cyclists had to give way
to motor vehicles...

Haltemprice roundabout cycle path
The ERYC leaflet points out the 10 locations where
cyclists must give way to motor vehicles whilst using
the off-road cycle path. (In reality there are another
5 locations where one must give way as one must also
give way at the points where the cycle path runs out
and you have to return to the road). A cyclist simply
going straight ahead has to give way at 5 points. A
cyclist making a right turn would have to give way at
7 locations. Then again you could just stay on the
road and give way just the once...

Such is the concern of the ERYC for the safety of
cyclists at this roundabout that the leaflet even make
it clear to cyclists who has priority. (And in so
doing makes the fatal flaw of such off- highway paths
very clear).

Mini roundabouts
Although  it is sometimes claimed that a mini
roundabout is 'safer' then a traditional roundabout,
it might be more accurate to say they are less
dangerous overall than normal roundabouts and that
accidents occurring on a mini roundabout are less
likely to be fatal. As the statistics section above
shows between one third and one quarter of all
'roundabout' crashes occur at mini roundabouts.

Mini roundabouts, in common with some other traffic
calming features such as contraflow gateways may help
reduce vehicle speeds. However, they may also increase
the level of conflict between cyclists and motor
vehicles so that cyclists are forced to adopt their
riding style so as to avoid collisions. For example,
many drivers (as high as 49 in every 50) will fail to
give way to a cyclist at a contra flow gateway, even
where the cyclists supposedly has priority.  Drivers
also frequently fail to give way to an approaching
cyclist at a mini roundabout and consequently cyclists
frequently feel compelled give way to other vehicles
even when they supposedly have priority themselves.
Such 'adaptive behaviour' may well mask the true
effect of installing a mini roundabout (or indeed
other road feature). As ever casualty figures only
reveal a small part of the impact of traffic schemes.

One important factor which may mask the true impact of
a feature such as a mini roundabout is the gross
under-reporting of cycle crashes, with as many as 87%
of minor cycle crashes going unreported. (Pedder et al
1981).  Given the speed reducing effect of mini
roundabouts it is to be expected that injuries
sustained by cyclists using mini roundabouts are
likely to be slight and so less likely to be reported
in any case. A similar situation may well exist for
the drivers of motor vehicles as a  low speed
collision between two motor vehicles is most likely to
result in vehicle damage only and so will not be
recorded on the official accident statistics which
record injury accidents only.

Very significantly, local design consideration such
the number of arms on a roundabout and the degree of
deflection required to circulate the roundabout have a
significant bearing on a given mini roundabouts
'safety'. In addition a mini roundabout may simply be
poorly designed. For example, at the mini roundabout
below vehicles approaching the roundabout in the same
direction as the camera are not deflected by the
central island and may pass though the 'roundabout'
without slowing down at all. This may increase the
risk to vehicles (and especially cyclists) approaching
from the right who have priority. In comparison
vehicles travelling towards the camera have the
roundabout directly in their path and so may be
encouraged to slow down in order to pass round it. 

That traffic calming features frequently increase the
hazards for cyclists seems incontestable. However, it
is often claimed that this is justified because of the
speed reduction that such features supposedly achieve.
The problem with this is that drivers will do almost
anything to avoid having to slow down when
encountering traffic calming features.

For example, the East Riding of Yorkshire Council
created some ill conceived traffic cushions though the
village of Melton. However, they left a gap in the
middle of the road so large that the cushions could be
avoided by driving down the middle of the road. As a
result drivers travelling in each direction play out a
little 'car ballet', with each vehicle being driven
out to the middle of the road and back over to the
left at each cushion. With two oncoming streams of
traffic the effect is quite comical to watch but how
the cushions contribute to road safety is hard to work
out.

In the case of a mini roundabout many drivers simply
choose to drive over the roundabout or at least part
of it, especially those in 4x4's. (The manufacturers
of some 'Chelsea Tractors' have even gone so far as to
advertise their vehicles are the idea means by which a
driver may pass over speed calming features as though
they weren't there). In general, any traffic calming
feature needs to be very carefully thought out if it
is not to create an undue hazard for cyclists and is
actually going to reduce traffic speeds.

Although  traffic calming schemes may show a reduction
in casualty rates it needs to be remembered that a
calming effect is often achieved by bringing vehicles
into potential conflict with one another, for example,
by using contraflow gateways. Given that drivers are
unlikely to give way to cyclists in such situations,
even when the cyclist has priority, it might seem
strange that such schemes may reduce casualty rates at
all. In reality any fall in casualty rates is often
not due to the skill of the highway engineer but they
reflect the desire of the cyclist to stay alive and
avoid injury. Consequently, the cyclist is compelled
to adapt their behaviour to the hazard created, even
when this means giving way in situations where they
supposedly have priority. Unfortunately, such
'adaptive behaviour' may include the cyclists seeking
alternative routes which may be objectively more
dangerous or even ceasing to cycle.
--------------------------------------------------------------------------------
Howard Peel
December 2002

This page.
http://www.thebikezone.org.uk/thebikezone/campaigning/roundabouts.html