This paper presents experimental modeling that explores the congestion relief effect of transit in Melbourne, Australia. The method adopts a four-step model and assumptions on transit user diversion to auto travel when all or part of the transit system is removed in the a.m. peak. When all transit is removed, diversion to autos causes more than 1,000 additional road links to become congested and increases congestion by 30%. More than 1.3 million additional auto vehicle kilometers of travel occurred and some 97,000 vehicle hours were added under this assumption. Congestion delay increased by more than 150%. Analysis of the separate contributions of transit trips to the central business district (CBD) and greater-CBD areas shows that they contribute to about 40% and 53% of congestion relief in total. When individual transit modes are separately removed, heavy rail use manifests the greatest effects on congestion relief across all suburbs (inner, middle, and outer). Bus effects are second in middle and outer Melbourne, with similar effects to rail use in outer Melbourne. Tram use shows little effect in outer suburbs, but it is significant in inner areas. The greatest percentage change in congestion measures (vehicle hours/delay vehicle hours) is observed in inner Melbourne. However, absolute changes in measures are substantially higher in middle and outer Melbourne, as a result of the bigger spatial scale of these areas. Overall findings increase the understanding of how transit acts to make cities perform more efficiently. An important outcome is the wider non-CBD-based effects on congestion relief, challenging the traditional view that transit and its effect on congestion is an issue only for central cities.