A construction project in Ontario, Canada, provided the opportunity to use field data to investigate freeway capacity at long-term lane closures due to rehabilitation work. Data from two lane closures at the same construction site (eastbound and westbound) were examined. The site is located on the Gardiner Expressway in the southern part of downtown Toronto. Data were collected during 4 days, totaling around 53 h of congested traffic operations. Results showed significant variation in freeway capacity in the work zones. Despite this variation, average capacity values are reasonably close to the corresponding values provided in the Highway Capacity Manual. Four intervening variables were investigated; all exhibited significant but different effects on freeway work-zone capacity. These variables included temporal variation (which is thought to relate to driver characteristics), grade, day of week, and weather conditions. The results confirmed the pressing need for more extensive field data that will allow better identification of the effect of various control variables on work-zone capacity.