The automobile accounts for more than 95 percent of all person miles traveled in the United States, whereas transit accounts for less than three percent of all trips. Between the private automobile and traditional transit, niche markets exist for other transportation services, such as airport and transit feeder shuttles and carsharing. Commuter-based carsharing, by which individuals share a fleet of vehicles linked to transit, could potentially fill and expand one such niche; complement existing services, mainly transit and feeder shuttles; and develop into a viable transportation alternative. A brief overview is presented of transit feeder services, particularly in Northern California. The CarLink I field test is described–a commuter-based carsharing model developed and tested in the San Francisco Bay Area in 1999. Key behavioral findings are highlighted. Recommendations are made for strengthening the viability of this innovative service, mainly on the basis of experience from the CarLink I field test. Empirical evidence suggests that CarLink could be sustainable if certain conditions exist: strong cooperation with local businesses and transit is established; parking limitations and congestion are prevalent; system feasibility is demonstrated; incentives are used (e.g., preferred parking and transit pass subsidies); higher user fees are generated; and technology and management costs are lowered. Further, CarLink’s long-term potential and viability could be strengthened by a combination of approaches, including cost reduction strategies, policy incentives, linkages to other complementary services (e.g., ridesharing), and local program support (e.g., demonstration grants).