Frost resistance of high-strength concrete (80-100 MPa) was studied by subjecting 44 concrete mixes to freeze-thaw cycles in water (ASTM C666, standard procedure A) and to scaling tests in the presence of deicer salts (ASTM C672, standard). The test programme was designed to analyze the effects of the water/cement ratio, the type of cement, the type of coarse aggregate, the duration of curing, and the air-void spacing factor. Results demonstrate that the water/cement ratio limiting value, below which entrained air is no longer necessary to protect concrete against freeze-thaw cycles, is sometimes higher than 0.30 but is more often below 0.25. This limiting value is affected most by cement characteristics: type 30 cement yielded much more durable concretes. Laboratory scaling tests demonstrated that when water/cement ratios are less than 0.30, the resistance deicer salt is generally very good, no matter what are the type of cement, the silica fume content, or the air-void spacing factor of the concrete.