The extra signals are positioned on both sides of the roundabout, at the points where Cycle Superhighway 2 enters the junction. There are three sets on each side, two at the front and one further back to control the "early start" for cyclists entering via the filter lane. All the miniature lights do is repeat the signal atop the pole above them, they're nothing particularly special, except for the hope that cyclists will see them more clearly and react appropriately.
And that wasn't all that was unusual at Bow Roundabout during Monday's rush hour. The police were out in force, standing in all sorts of places around the junction, along with a lot of the contractors who've been responsible for installing the lights in the first place. Also present was at least one TV news cameraman, his big lens poised to film cyclists whizzing round the blue stripe. No misbehaviour or red-light-jumping ensued. But these were entirely atypical conditions, with every rider enduring the beady eyes of scrutiny, so not the best time to determine whether the lights were having the intended effect.
So I went back during the evening rush hour, after every scrutineer had disappeared, to carry out three small experiments of my own. They're entirely unscientific too, based on ridiculously tiny sample sizes, so should in no way be taken as indicative of how cyclists' behaviour might have changed. But that won't stop us jumping to conclusions, right?
Experiment 1: Do westbound cyclists obey the lights?
Unfortunately this experiment has a sample size of only one, because there aren't many westbound commuters during the evening rush hour. But down Stratford High Street he came, resplendent in professional pink hi-vis jacket, so I had high hopes. The first low-level cycle light was green, so through he sailed, while the traffic queued alongside patiently waited. But the 'cycle early start' design dictates that at least one of the pair of lights is always red, if not the first then the second. So the next low-level cycle light was red, because the main lights were red, to prevent anyone from entering the junction prematurely. Success - this cyclist duly stopped and paused at the red light. One down for the new system, I thought. Except then he spotted a gap in the traffic, a whopping great gap to be fair, and off he went. A slow burst at first, then more quickly, and across to the safety of the segregated blue lane before the next lorry came roaring round the roundabout. Conclusion: No, he didn't obey the lights. Westbound cyclists still jump the lights when it suits.
Experiment 2: Do eastbound cyclists obey the lights?
This experiment has a sample size of three, because there are more eastbound than westbound cyclists in the evening peak. Cyclist number one acted much as their westbound counterpart had done, cycling through the first green signal as intended, than failing to pause for the regulation amount of time at the second red. Cyclist number two was perfectly behaved, filtering into the early start zone and then waiting for the main lights to turn green before proceeding, safely ahead of the following traffic. With the scores level at 50-50, along came cyclist number three. This time the low-level light in the filter lane was red, signalling for cyclists to stop while the main body of traffic entered the roundabout. Cyclist number three wasn't having that, he rode straight through at speed and up to the second set of lights. These were on the turn and had just switched from green to red, signalling that now was the time for everyone to stop. And this message was totally ignored too. He zoomed on, with no attempt to brake, to cross the roundabout before the next pulse of traffic arrived. Again all was carefully judged, and at no time did any collision look possible, but just one left-turning truck could have caused a very different outcome. Conclusion: No, only a minority obeyed the lights. Many eastbound cyclists still jump the lights when it suits.
Experiment 3: Do eastbound cyclists take advantage of the updated infrastructure at the Bow Roundabout?
This experiment has a sample size of twenty, which is almost scientific. That's ten consecutive cyclists near the start of the evening rush hour, plus ten consecutive cyclists near the end. I watched them approach from Bow Road and checked whether they rode down to the roundabout or whether instead they took the flyover. It's not necessarily easy to get to the Bow Flyover on a bike, you have to pull out into the traffic and cross a couple of lanes to get there. But thirteen of my twenty cyclists did this, speeding across and up and over the flyover, leaving just seven to follow the approved route down to the Bow Roundabout. I was surprised how unpopular the blue stripe was, but no, the majority of eastbound cyclists shunned all the improvements and chose to bypass over the top instead. Most won't have realised that the low level cycle lights had been switched on, it being Day One, and there being no signs to this effect. But a few new lights don't really make a material difference to cycle safety, so I'd expect most cyclists to continue to shun the junction despite all the trumpeted tweaks that TfL have made. Conclusion: No, two-thirds avoid the roundabout. Most eastbound cyclists choose the flyover instead.
It strikes me that these new low level lights aren't really about cycle safety, because all they do is reinforce existing signals. Instead they're about trying to get cyclists to stop, and to use the cycle early start in the way the engineers intended. What's been created at the Bow Roundabout is a junction that's "always red" for cyclists, and an over-complex system that many riding through either fail to understand or choose to ignore. Regular cyclists at Bow have already learnt a series of bad habits, or chosen to bypass the set-up altogether, with the low level lights merely a cosmetic tweak that makes it look like something has been done. I fear that the Bow Roundabout is a poor place to trial the country's first set of low level lights, given that the existing set-up is already over-complicated, and they'd have been better off installed somewhere that bad behaviour hadn't already set in. Although they may look like an improvement, I doubt they'll make this killer junction any safer.