The Consumer Product Safety Commission has become concerned that the nighttime accident rate for bicyclists greatly exceeds the daytime accident rate. The CPSC staff has been instructed to consider what changes might be made in the standard for bicycles that would reduce these nighttime accidents. As a result of these instructions, the CPSC is holding a meeting in Washington on 13 March, 1996, to consider what should be done.
The Consumer Product Safety Commission has the authority to set requirements (in the form of standards) for consumer products that apply up to the time of sale to the consumer, but not thereafter. The CPSC has no authority to set standards for the way the products are used, although the CPSC may require the manufacturers of consumer products to provide instructions for the safe use of the products that they make. The limits of the CPSC's authority are set by the Constitution, and its authority exists only because the Constitution allows the federal government to regulate interstate commerce.
Any requirements set by the CPSC preempt all requirements set by local authorities (states, cities, counties, etc.) that pertain to the same risks, at least up to the time of sale to the final consumer. That is, no local authority can require, as a condition of sale, any additional or different equipment for nighttime safety than the CPSC requires.
The CPSC does not have the authority to distinguish between different uses of the same product. If different uses require different products, then the CPSC has the authority to set different requirements for the different products based on the way that each product will be used. However, where one product, or one class of products that are indistinguishable from each other, has several uses, the CPSC has no authority to set different requirements depending on the type of use that the final consumer may choose to make of the item that he or she purchases. Furthermore, the final consumer may well not be the initial purchaser and may well have different purposes in mind.
The CPSC has stated, and its position has been upheld in court, that it cannot distinguish bicycles that will be used by adults from those that will be used by children. Equally, it cannot distinguish those that will be used for recreation from those that will be used as vehicles for transportation, nor those that will be used at night from those that will not be used at night. The inability to distinguish between these uses is not just a legal limitation that might be changed; there is no clear-cut physical way to make such distinctions.
Therefore, any requirements that the CPSC may set must be applied to all bicycles up to the time of sale. The problem is to devise a system that can be implemented within these limits that will significantly reduce nighttime accidents to cyclists more than does the present system. To do that, we need to consider the accidents that now occur and the engineering requirements for reducing their frequency, to see which ones might be susceptible to an improved system that the CPSC can properly require.
Three types of accidents to cyclists are caused by darkness:
1: Cyclist does not see where he is going and runs off the road, rides over a
road surface defect, or collides with a fixed object or pedestrian. In most
cases, the cyclist is not using a headlamp and the accident may be attributed to
that failure, although in a few cases the cyclist was using a headlamp but still
did not see the hazard.
2: Cyclist collides with another cyclist. Generally one or both of the cyclists are without headlamps and failure to use a headlamp is the cause of the accident.
3: Cyclist is involved in a collision with a motor vehicle.
Types 1 and 2 clearly require the use of a headlamp, and reflectors of any quality would not make a difference. Type 3, cyclist involved in a collision with a motor vehicle, is the only type that might be ameliorated by improving the reflector system, because it is the only type in which the object collided with possesses headlamps that might illuminate a reflector at the time required to prevent a collision. We need to investigate the types and frequencies of car-bike collisions to see which ones would be reduced by the use of headlamps and which ones would be reduced by the use of reflectors.
The only study that provides data on car-bike collisions probably caused by darkness is Cross and Fisher of 1977 (1), whose data are analyzed for this cause in Forester's Bicycle Transportation (2). Forester concludes that 79% of the car-bike collisions probably caused by darkness are of types for which a headlamp is the only satisfactory preventive equipment. In these accidents the car approaches the bicycle from ahead, either from directly ahead on the same street or from diagonally ahead on an intersecting street, generally at right angles to the cyclist's street. Of the accidents of these types that occurred during darkness, Cross ascribes 69% to the absence of a bicycle headlamp. In 21% of car-bike collisions that are probably caused by darkness, the motor vehicle approaches the bicycle from the rear on the same street. The Cross statistics show no cases in which the motor vehicle approached the bicycle diagonally from the rear, as might occur at a Y intersection.
Light radiates in straight lines practically instantaneously. It is generated by luminous sources such as the sun and lamps, and is partially reflected and partially absorbed when it falls on most surfaces.
Light varies in quantity, which our eye interprets as brightness, and in wavelength, which our eye interprets as color. What we see as white light from the sun or the slightly yellow light that we see from most incandescent lamps is actually a mixture of all colors from red to violet.
The brightness of a light source is measured in candlepower, originally the brightness of a specified candle. The brightness falls off with the square of the distance, and the intensity of light falling on a surface (including the surface of the eye) is measured in foot-candles, originally the amount of light delivered by the specified candle when observed at a distance of one foot.
Bicycle head lamps send their light out in a pattern that is determined by their mirrors and lenses. The pattern generally consists of a narrow beam of bright light and a wide arc of dimmer light. The narrow, bright beam is used by the cyclist to see where he is going. This beam has to be bright because its light must be reflected from surfaces that typically have low and similar reflectivities. The wide arc of dimmer light is used to show other people the presence of the cyclist. Although this light is dimmer than the beam, it can be seen from a far greater distance than the cyclist can see ahead, and typically over an arc of at least 70 degrees on each side of the center line.
Headlamps for motor vehicles have the same general pattern of bright beam and dimmer arc as bicycle headlamps, but because of their much greater brightness special arrangements are made to limit the top edge of the beam, making it flat-topped, so it will not glare into the eyes of other drivers. Furthermore, each car is equipped with lamps giving two sets of beams. One beam is directed straight ahead and level, for use in rural areas where no other driver is in sight, the other beam dimmer and directed downward and to the right, for use in cities and where other drivers are in sight.
The reflectors used are technically called retroreflectors because they reflect light falling on them back towards its source. That is, they not only do not generate their own light, but unlike most surfaces they also do not reflect light in a diffuse manner. Each beam of light that falls on a retroreflector is reflected back towards its source.
For reasons of optical science, a retroreflector can operate only over an arc of no more than 40 degrees, 20 degrees on each side of its center line. Beyond that arc, it totally extinguishes.
The performance of a retroreflector is measured in terms of the ratio of its apparent brightness to the intensity of the light falling on it from the same direction. That is, in candlepower per foot-candle. The brightness, in turn, depends on the efficiency of the reflector and its size. A large reflector will appear brighter than a small one because it intercepts more light from the energizing lamp. As is obvious, the brightness of a retroreflector depends directly on the intensity of the light from the motorist's headlamp beams. The brightness of a reflector falls off markedly as the reflector moves away from the center of the headlamp beams of the motor vehicle, merely because the intensity of the light from the headlamps falls off.
The retroreflectors that are used in highway situations have optical imperfections, both the unavoidable and the designed-in ones. Instead of reflecting the light directly back to its source, they reflect the light in a cone directed around the source and, in some cases, directed slightly above the source. Thus, the reflected light hits the eyes of the driver of a vehicle when his headlamp beams shine on the reflector. The narrower the cone of reflected light, the brighter the reflected beam and the greater the distance at which the reflector can be seen. However, as the driver approaches the reflector, his eyes can get above the reflected beam and, to his vision, the reflector dims and finally extinguishes. Therefore, the design of a reflector is a compromise between brightness at a great distance and brightness when close up. The appropriate design makes the reflector brightest when viewed at the distance at which a collision can be avoided.
The present CPSC system uses 10 reflectors. The 6 main reflectors are placed so that one faces forward, one faces backward, and two face to each side, one on each side of each wheel. The other four reflectors are supposedly placed on the front and back sides of the pedals, but many pedals made today are shaped so they cannot support such reflectors. Each of the main reflectors is a wide-angle reflector. It is made up of three panels set at different angles. Since each panel can reflect over about a 40-degree arc, the single reflector that is made up of the three panels can reflect over about a 100-degree arc. The reflectors are mounted on the bicycle so that each covers a 90-degree arc; forwards, right side, backwards, left side. The amount that each reflector exceeds 90 degrees allows for the inevitable misalignment of the reflectors. The idea behind this system is that whatever the direction from which a car's headlamps may shine on the bicycle, at least one reflector will be positioned to reflect light back to the driver of that vehicle. In that way, so the concept goes, the bicycle is protected against collisions from all angles.
While the states cannot impose any other requirements for nighttime protective equipment at the time of sale, they universally require different equipment to be used when any bicycle is operated at night. The essentials of the requirements of all states are a headlamp and a rear reflector, although many states also now require the other reflectors specified by the CPSC. The states require headlamps for the valid and sufficient reason that headlamps are absolutely necessary to avoid nighttime car-bike, bike-bike, and bike-pedestrian collisions. No responsible person believes that the all-reflector system can work. For those who do not understand, Figure 1 shows two of the typical car-bike collision situations that have much higher frequencies at night, in both of which the all-reflector system will not work. Situations similar to these cause 79% of the car-bike collisions that are probably caused by darkness. The inability of the motorist to see the bicycle's reflectors is particularly acute in cities, where most nighttime cycling occurs. This is because motorists in cities use their low beams that are directed low and to the right to keep them out of the eyes of motorists coming from the opposite direction.
The other item required by the states is a rear reflector and, in some states, an additional rear lamp. The rear reflector works because the motorist is coming from directly behind, or almost directly behind, the bicycle. The motor vehicle's headlamp beams shine on the bicycle for a long time before the motorist reaches the bicycle, giving ample time to avoid the collision. On a sharply-curving road the motorist's headlamp beams do not shine upon the cyclist until the motorist is considerably closer, but on such sharply-curving roads the motorist must travel much slower than normal, so there is still time to avoid the collision. The rear reflector does not have to be a wide-angle reflector because the motorist cannot be more than 20 degrees away from the bicycle's center line at any reasonable speed, and in any case his headlamp beams don't reach that far to the side so there is insufficient headlamp light beyond that angle to energize the reflector.
The engineering conclusion about appropriate nighttime protective equipment is that cyclists must use a headlamp and some form of rear device, which may be a reflector alone or a reflector plus a rear lamp. That is scientifically respectable and it is also what the states require. However, the CPSC cannot reasonably place such a requirement into its standard. I explain why.
Few of the bicycles purchased in the U.S.A. will be used for transportation, even fewer will be used at night, and even fewer will be used regularly at night. With such a small proportion that will be used at night, and therefore a small number of nighttime accidents that might be prevented, if all bicycles were fitted with headlamps the cost per accident prevented would be very high. There is also considerable doubt that fitting every bicycle with a headlamp at time of sale would actually prevent many of the nighttime accidents. This is because headlamps (or rear lamps also) that have been on the bicycle since the time of sale but have not been used for a long time will probably not work when they are needed.
The cost per accident prevented falls dramatically when only those bicycles used
at night are equipped with headlamps and rear reflectors.
The proper policy must produce two different effects of equal importance. It must get those bicycles used at night equipped with headlamp and rear reflector. It must also prevent those bicycles not so equipped from being used at night.
We know what needs to be done, but we also recognize that the CPSC can't do it alone.
The question then becomes: how to devise a plan that accomplishes these two effects as best as possible within the scopes of governmental authority and in the context of our society?
The old solution was the all-reflector system. This was devised by the Bicycle Manufacturers Association, which then persuaded the CPSC to adopt it into the CPSC's regulation. The BMA had two motives for getting this system into national law. The first motive was to get uniform requirements for nighttime protective equipment at time of sale in all states. Before this, each state had its own requirements and many were different. The industry had to supply different kits of items depending on which state each bicycle was shipped to, and this presented a real problem. The second motive was to head off any requirement for supplying headlamps on all bicycles. Besides the wider cost/benefit considerations discussed above, the BMA had to consider that the cost of good lighting equipment would double the cost of the typical bicycle that its members sold, and that after years of non-use the equipment was likely to be defective at the time when it was required. Once the all-reflector system and the companion law by which CPSC requirements preempted local laws were in place, the BMA then tried to get the states' headlamp requirement repealed (arguing that case before the Committee for Uniform Traffic Laws and Ordinances), but the states refused to allow dangerous cycling at night without a headlamp.
The national uniformity argument could have been fulfilled by a less comprehensive system than the all-reflector system. However, the anti-headlamp argument could not. While there are valid reasons for not requiring headlamps on all bicycles, as discussed above, the anti-headlamp argument could succeed only if the all-reflector system looked as though it adequately replaced the headlamp. To the ill-informed, with its all-round reflectors including its white front reflector, it indeed looked as though it did. The argument was made that the all-reflector system was cheap and maintenance-free, and that although it might not be quite as good as a proper headlamp system it was far better than a headlamp that had been installed at time of sale and had been neither used nor maintained since.
While there was a demonstration of the all-reflector system's effectiveness in reflecting headlamp light, no matter at what angle the bicycle was, there was no test or analysis at all of whether or not the headlamps would shine on the bicycle at the time a collision could be avoided. Maybe the designers and adopters of the system thought that it would, maybe they didn't consider the question at all, or maybe they knew it wouldn't but wanted it anyway; that is unknown. What is known is that nothing was done to analyze the probable effectiveness of the system in preventing nighttime car-bike collisions. By the time that I raised the question in my comments about the CPSC's regulation, it was too late for such comments to be seriously considered. Up to that time those who had the responsibility of experts in that field either thought that the all-reflector system would work or, at the worst possible interpretation, thought that their statements that it would work would be accepted by the courts and by society at large.
As it turned out, they were correct on both counts. The courts accepted their words, refusing to rescind the all-reflector system. "Reflectors appear to provide a significant margin of added safety at a relatively small monetary cost and loss in bicycle efficiency. In view of the Commission's careful balancing of the relevant factors, we do not find this standard to be irrational." (3)
The public also believes that the all-reflector system is an adequate substitute for the headlamp and rear reflector system. The court in Forester vs. CPSC agreed with the CPSC that the public would not believe this. "Forester's argument assumes that cyclists who ride at night would, but for the standard, purchase, maintain, and use headlamps. The Commission could rationally have concluded that this was unlikely; rather, that many unsophisticated or infrequent nighttime riders would otherwise do so without any protection at all."
The facts developed after twenty-five years of experience with the all-reflector system demonstrate the accuracy of my prediction and the inaccuracy of the CPSC's and the court's predictions. In the course of the Johnson vs. Derby case (4), evidence was developed by the chairman of the department of psychology at Rutgers University that 85% of New Jersey high-school students believed that a bicycle with only the all-reflector system was safe to ride at night. The respondents who felt this way typically said that because the bicycle had all those reflectors it was safe to ride at night. The CPSC's own study of bicycle accidents (5) shows that only 1/3 of those who ride at night used either a headlamp or a rear lamp, and that 50% more people added a rear lamp to the existing rear reflector than used a headlamp. Of those people who thought enough about nighttime safety to buy a lamp, many more thought that the front reflector was an adequate substitute for the headlamp than thought that the rear reflector was adequate.
In summary, the evidence is very good that the all-reflector system persuades people that it is an adequate substitute for the headlamp.
The position of the Bicycle Manufacturers Association was expressed at the last CPSC meeting on this subject. I quote from my report of that meeting.
Michael Kershow, General Counsel of the Bicycle Manufacturers' Association, presented the view of BMA, largely regarding the liability question rather than discussing the safety aspects. The average bicycle sells for less than $100 at Walmart or Toys R Us. It is a good thing that 90% of customers don't ride at night, because nighttime cycling is too dangerous. These customers wisely choose not to ride at night. The industry must not supply lights as Original Equipment because that encourages cycling at night, which is very dangerous, and presents liability problems for bicycle manufacturers. We should actively discourage cycling at night because automobiles weigh 2,000 pounds. Here is his list of 8 actions.
1: Tell the public, "Don't ride at night," but continue to require reflectors and discourage their removal.
2: The all-reflector system is a very effective passive system that requires no activation by the user; not to be replaced by complicated lights.
3: If you ride at night, you must use both the all-reflector system and lights, because these are complementary elements (in some good design?).
4: We need standards for lights.
5: No OEM lights, because these would encourage nighttime cycling, while common sense now prevails on people not to cycle at night.
6: Lights are too expensive. The all-reflector system is appropriate because it provides a high degree of conspicuity at low cost.
7: Providing lights with bicycles does not guarantee that they will be used. The consumers are too dumb to maintain lights and to use them when needed.
8: Improve the enforcement of state laws and improve education."
There is a lot of good sense in this proposal, particularly about the policy and liability aspects, but its engineering is incorrect and its predictions about behavior don't fit the facts as they have developed.
I discuss the engineering errors first. The claim that the all-reflector system is very effective is false. The all-reflector system is ineffective against all bike-bike and bike-ped collisions. It is ineffective against the 79% of the car-bike collisions that are probably caused by darkness that come from the front or from diagonally to the front on intersecting streets or driveways.
The claim that the all-reflector system and the lights system are complementary elements of some undefined system is false. If the cyclist uses headlamp and rear lamp, the only part of the all-reflector system that has any function is the rear reflector, and then merely as a backup in case the rear lamp goes out. No part of the system can substitute for the headlamp, even as a backup, because no part of the system fulfills the functions of the headlamp, even poorly.
While the all-reflector system does provide considerable conspicuity, the engineering error is the false assumption that the conspicuity that is provided is useful in preventing car-bike collisions. For example, the bouncing wheel reflectors provide great conspicuity in many situations, which makes the ill-informed public think that they are great, but the conspicuity is there only when there is no possibility of collision, unless the cyclist is just standing in the middle of the road waiting to be hit.
The BMA is very insistent that people should not be tempted to ride at night. However, the BMA argues that while headlamps tempt people to ride at night, the all-reflector system does not tempt people to ride at night. Well, consider the facts. About 3/4 of the people who ride at night do so with the all-reflector system but without headlamps (6). Whether the type of their nighttime equipment actually tempts people to ride at night is dubious, but the evidence is that the great majority of those who see the need to ride at night believe that the all-reflector system is adequate, and that the majority of those who do not believe that the all-reflector system is adequate believe that the significant inadequacy is in the rear reflector rather than in the front reflector.
The case of Johnson vs. Derby Cycle Co. that shook the industry was decided before the last CPSC meeting about nighttime protective equipment, the meeting at which the attorneys for the industry presented their position. Other car-bike collisions have occurred as a result of relying on the all-reflector system and, inevitably, some of them will result in similar suits. Because the industry has publicly stated its position, it should be more afraid of those suits now than before the last meeting, because its recommendations are untenable.
The industry wants the all-reflector system because, so it says, that system provides a high degree of conspicuity at low cost and with minimum maintenance. The industry claims that cycling at night is dangerous. The industry prefers the all-reflector system because of the cost and probable liability associated with the lights that it would otherwise install, or be required to install. If a jury were presented with that evidence after it had seen the really indisputable evidence that the claimed high degree of conspicuity cannot have been present at the time and place necessary to prevent the collision being considered, and that a headlamp probably would have been seen, there is little doubt about what the jury would find.
We need a system that fulfills two safety functions. It must:
1: Provide adequate conspicuity of the type that prevents collisions when used at night;
2: Clearly indicate that the bicycle with only the time of sale equipment is not safe to ride at night.
In addition to these two safety functions, the system must have three other characteristics. It must:
3: Be within the lawful authority of the CPSC;
4: Have the legal status to preempt local requirements at the time of sale;
5: Satisfy the need for liability protection of the manufacturers who must use it.
A system that meets these requirements can be readily available. In this system the CPSC requires only a rear reflector at time of sale and the states require both that rear reflector and a headlamp whenever the bicycle is used at night. The CPSC also specifically requires that the manufacturer provide the instruction that a headlamp must also be used for safety when cycling at night. The combination fulfills the two safety requirements.
1: It clearly provides the necessary equipment for cycling at night.
2: The presence of only the rear reflector at time of sale removes any temptation to believe that the bicycle, as sold, is properly equipped for cycling at night.
The combination also fulfills the other three characteristics.
3: Requiring a reflector on all bicycles sold is clearly within the scope of the CPSC's authority. Not having a requirement for a headlamp on all bicycles sold means that the CPSC has not issued a requirement with a very poor benefit/cost ratio and that would produce liability problems for the manufacturers who had to comply.
4: The purpose of the required rear reflector is to reduce the frequency of nighttime car-bike collisions, and that is the only purpose that reflectors serve. Therefore, no local authorities could require other or different reflectors at the time of sale.
5: Because the reflector obviously serves only to the rear, it is patently obvious that something else is required for protection from the front, both directly and diagonally. There is no additional equipment that could tempt the user to conclude that the bicycle was suitable for nighttime cycling. There is nothing that even looks as though it fulfilled the function of a headlamp.
The rear reflector should be as bright as practical over the required angular range. The present CPSC reflector is far from optimum. First, because its design philosophy was all-around protection, which is not necessary, only 1/3 of its area would reflect at any one angle. The SAE design of reflector reflects over about 20 degrees on each side of the center line, which is sufficient. Therefore, the entire area of the new reflector can reflect over the required range. That multiplies the brightness by three times. The present reflector is red, a color that reflects white light only dimly. An amber reflector is 2.5 times brighter than a red reflector of the same size. (A white reflector is 4 times brighter than a red reflector.) The rear reflector should be amber. This is not standard for the rear of a vehicle, but it is standard for roadway obstructions such as construction areas. Since in most cases the bicycle will be moving much slower than the overtaking motor vehicle, the motorist will take the same precautions whether he sees a red reflector or an amber one. In fact, amber may be safer than red of equivalent brightness. The amber will prevent the motorist from thinking that the vehicle ahead is moving fast and that he has more time to take the proper action than he actually has. White, while brighter still, indicates an oncoming vehicle and should not be used because it might cause confusion.
Combining the increase in active area and the change to amber would increase the brightness of the reflector by more than 7 times more than that now specified, for the same size reflector. This is well worth the change.
While the CPSC cannot reasonably require headlamps, it can offer interface standards that facilitate the use of lamps.
A great many lamps have been designed to use a standard bracket, which is 30 mm square and 3 mm thick. Each lamp has a socket at its back that fits over the bracket. These brackets are available in a variety of shapes that fit to different parts of the bicycle: handlebars, steering column, front fork blades, etc., as required by the type of bicycle and the desires of the user. The CPSC can reasonably specify a standard bracket and require that each headlamp manufacturer either provide a complete mounting for his headlamp or a socket to match the standard. The more manufacturers that match the standard the simpler it will be to mount lamps and replace broken ones.
A similar problem exists for generator mountings. The typical generator mounting combines the worst features of being neither permanent nor removable, and it scratches the paint and tends to come dangerously loose as well. There have been permanent generator brackets, but none became standard because of the variety of dimensions encountered. A system with sufficient adjustability has a standard bracket permanently fixed to the bicycle's frame, either front fork blade or rear stay, parallel to the center line of the bicycle, with two 5 mm holes 30 mm apart, approximately vertically spaced and about 100 mm below the tire contact point. These are the only specified interface dimensions. An intermediate adjustable plate bolts onto the standard bracket and carries the generator in the proper position and, for front fork mounting and if the user desires, carries the headlamp also in one simple removable unit. The simplest adjustable plate is made of soft aluminum that can be easily bent to place the generator at the correct lateral distance from the tire, drilled to hold the generator at the proper position, and finally trimmed to size.
Provision of standard interface dimensions for these items would promote the easy use of proper headlamps without requiring any special equipment at time of sale.
1 Cross, Kenneth D., and Gary Fisher; A Study of Bicycle/Motor-Vehicle Accidents: Identification of Problem Types and Countermeasure Approaches; National Highway Traffic Safety Administration; 1977
2 Forester, John; Bicycle Transportation 2nd ed.; The MIT Press, 1994; chap 5, chap 17
3 Forester vs. CPSC, 75-1292, Court of Appeals for the District of Columbia
4 Johnson vs. Derby, Essex County Superior, NJ, 1993
5 Bicycle Use and Hazard Patterns in the United States; U.S. Consumer Product Safety Commission; 1994
6 The all-reflector system has been on most bicycles sold in the U.S.A. for 25 years. The data in the CPSC's study indicate that about 3/4 of nighttime cyclists do not use a headlamp.
Return to: John Forester's Home Page Up: Lights