#43 – Shared-Use Arrow

Gainesville, Florida

William W. Hunter, Senior Research Scientist, UNC Highway Safety Research Center


A bicycle lane stripe provides a lateral positioning reference for both motorists and bicyclists, and the presence of the stripe, as well as signs, informs motorists that bicyclists are typically present upstream. In contrast, the absence of bicycle-specific pavement markings in wide outside lanes (also known as wide curb lanes), another widely acknowledged way to accommodate bicyclists, obviously means that there is no reference for lateral positioning, or a visual cue to the existence of upstream bicyclists.

Another argument put forth is that bicycle lanes are clearly marked spaces for bicyclists that have been shown to draw riders off of adjacent sidewalks and onto the roadway, a desirable outcome given the inherent dangers of sidewalk riding. On the other hand, because there are no bicycle-specific markings in wide outside lanes, they are not recognized as an on-road bicycle “facility” by many bicyclists, resulting in a higher incidence of adjacent sidewalk riding than could otherwise be the case.


Figure 1. Shared-use ARROW.

Figure 2. Actual lateral placement.

The shared-use ARROW is a symbol placed on the roadway with a stencil and is used to indicate proper positioning for a bicyclist in a shared travel lane. The shared-use ARROW (figure 1) was developed with the intention of addressing the deficiencies of wide outside lanes mentioned above. Furthermore, for situations at which sufficient pavement width exists to choose between striping a bicycle lane or leaving a wide outside lane, the shared-use ARROW may offer a third option, “bridging the gap” between the two existing treatments. Unlike a bicycle lane stripe, the shared-use ARROW does not restrict bicyclists and motorists to separate areas of the roadway, thus addressing several potential problems of bicycle lanes. The shared-use ARROW also requires less pavement marking materials than a bicycle lane stripe, and the ARROW reinforces the correct direction of travel, an issue of great importance for bicycling safety.
The original shared-use stencil was developed by James Mackay, the Bicycle and Pedestrian Planner for the city and county of Denver, CO. The city of San Francisco, through Manito Velasco, assistant transportation engineer, has also used the stencil. They elongated it from 1.3 m (4.25 ft) to 1.8 m (6 ft) and also altered the placement specifications. The current ARROW builds upon these efforts by establishing a widened opening along its centerline in an effort to channelize and make it more obvious to bicyclists to track down the centerline of the symbol.

Lateral placement was proposed at 0.8 m (2.5 ft) from the curb face, which was based on the local conditions of a 4.6-m (15-ft)-wide lane with no gutter pan and preliminary BEFORE measurements which showed bicyclists riding 0.5 m (1.6 ft) on average from the curb. Furthermore, with this specified spacing, it was expected that motor vehicle tires would be less likely to track over and wear out the marking. However, earlier paving over the old gutter pan had left a seam about 0.6 m (2 ft) from the curb. Thus, instead of at 0.8 m (2.5 ft) from the curb face, the ARROW was placed at 1.1 m (3.5 ft) by Gainesville Public Works (Figure 2).

Evaluation and Results

A before and after evaluation was conducted. Four locations along 13th Street (US 441) in Gainesville, FL, were examined using videotaping equipment to record bicycles and motor vehicles. In this study area 13th Street has four lanes with wide outside lanes in both directions. The street has a 30 mph speed limit and carries about 35,000 vehicles per day. Sites 1-3 were acceptable for all data that was to be collected, while one site (Site 4) was not acceptable for spacing measurements.

Seventeen videotaping sessions about two hours long were used to gather data both before and also after the ARROW was installed for a total of 34 sessions. Concurrent with installation of the device, about one week of public awareness was conducted. A press release was prepared, and television crews filmed bicyclists riding along the stenciled street. Information about the stencil was widely disseminated to University of Florida students, faculty, and staff through normal channels. The videotapes were examined by HSRC personnel. Three lateral spacing measurements were made using Jandel Scientific SigmaScan Pro Image Measurement Software on still images of the videotape captured by Snappy Version 3.0. The measurements were bicycle to curb, bicycle to motor vehicle, and motor vehicle to curb.

Before the ARROW was placed, 39.3 percent of bicyclists rode in street, with traffic. After the ARROW was placed, the proportion of bicyclists riding in street, with traffic increased to 45.3 percent. Comparing in street, with traffic with all other positions and directions combined (a 2x2 table, chi-square test) yields a statistically significant increase (p<.05) toward riding in the street with traffic after the placement of the ARROW.

Bicycle-to-curb measurements were made to determine if the ARROW was associated with a change in the lateral positioning of bicyclists. The difference between the before measurement of 0.5 m (1.6 ft) and the after of 0.6 m (1.8 ft) (about 76.2 mm (3 in.)) was statistically significant (p<.01). However, this small difference was not considered to be practically significant.

Bicycle-to-motor vehicle measurements were made when a motor vehicle with a driver with unobstructed view was directly next to the bicyclist, the front wheels of the motor vehicle and bicycle in line. The mean bicycle-to- motor vehicle measurement in the before period was 1.8 m (6 ft) (n=92). The mean bicycle-to-motor vehicle measurement in the after period was 1.9 m (6.1 ft) (n=83). The difference was not statistically significant.

The motor vehicle-to-curb distance was measured from the outside edge of the front tire (or in some cases the rear tire) to the curb face when there were no bicyclists nearby to influence the drivers’ positioning. The difference between the before mean of 1.9 m (6.3 ft) and the after of 2 m (6.4 ft) was not statistically significant.

There was an interesting difference in the distributions of the measurements that were made, and the difference was associated with the Bicycle-to-Curb distance. There was increased spread in the lower end of the distributions in the after period, such that the proportion of bicyclists riding 0.5 to 0.8 m (1.8 to 2.5 ft) from the curb increased substantially, in effect increasing their safety margin.

Conclusions and Recommendations

There were no practical differences in the average lateral spacing measurements of bicycle-to-curb, bicycle-to-motor vehicle, and motor vehicle-to-curb. However, the proportion of bicyclists riding 0.5 to 0.8 m (1.8 to 2.5 ft) from the curb showed a substantial increase, giving them a larger safety margin. There was a statistically significant increase in the proportion of bicyclists riding in the street after placement of the ARROW. This shift from the sidewalk to the street should increase safety by putting cyclists where they are more visible to motorists and out of conflict with vehicles entering or exiting driveways that cross sidewalks, as well as reduce the conflicts with pedestrians.

The 13th Street corridor was chosen because there were enough bicyclists riding on a daily basis to make data collection efficient. In retrospect, however, the number of cyclists may be a factor that mitigates against possible shifts in the distance measures of effectiveness. It is certainly possible that motor vehicle drivers on this route are well attuned to the presence of bicyclists, and thus may already have shifted their traffic lane location away from the curb to account for the space needs of bicyclists before the ARROW was installed. However, the shift in the lower end of the Bicycle-to-Curb measurement which yielded more riding space for bicyclists is compelling enough to “keep the jury out” on this shared lane treatment a bit longer. More trials in other locations are recommended and should result in more conclusive findings.

Costs and funding

Approximate costs were the following:

Labor: $500
Trucks and arrow board: $216
Paint and stencil: $118
Total: $834


William Hunter
UNC Highway Safety Research Center
730 Martin Luther King Jr Blvd, Suite 300
Chapel Hill, NC 27599-3430
(919) 962-8716

Brian Kanely
City of Gainesville
Public Works - Engineering
P.O. Box 490
Gainesville, FL. 32602-0490
(352) 334-5074

The modification (shared use arrow) that is the subject of this case study is not compliant with the Manual on Uniform Traffic Control Devices, but a version of this marking (bike symbol followed by chevrons, shown on page 279, Figure 2) is being considered for inclusion. Accordingly, it is imperative that any jurisdiction wishing to utilize the shared use arrow (or any other non-approved traffic control device) should seek experimental approval from the Federal Highway Administration. For information on how to do so, please visit this Web site: http://mutcd.fhwa.dot.gov/kno-amend.htm.