An experiment was made to determine the degree of alternative-route selection from a rural freeway that occurs because of implementation of a traffic-responsive variable message signage system in a work zone. The message signs gave real-time estimates of travel time to the end of the work zone. Alternative-route selection was measured through extensive volume counts on the freeway and on parallel arterial streets, both before and after implementation of the signage system. In addition, traffic volumes were measured at a variety of locations where alternative-route selection was likely to occur and where it was unlikely to occur. Analysis of peak period data found that alternative-route selection rates were between 7% and 10% of the freeway traffic, depending on the location and the day of the week.