The aerodynamic development of cars is conventionally carried out in the wind tunnel with the working section floor representing the ground surface. While recognising that the boundary conditions are false it has generally been assumed that this basic ground simulation is adequate for road cars. This situation is currently being reappraised. In this investigation a simple 1/4 scale car model has been tested in the MIRA Model Wind Tunnel, using both moving and stationary belt to represent the ground surface. The body and wheel drag contributions were measured independently. The influence of rear body shape, front spoilers, cooling airflow, and underfloor roughness on the aerodynamic characteristics, for both ground conditions has been assessed.