The simultaneous effect of mechanical loading and exposure to aggressive environment was studied on reinforced concrete (RC) elements. The elements comprised a single reinforcing 16 mm bar (Swedish Grade Ks500) embedded in 640 mm long concrete prisms with square cross-sections. The cover thicknesses were set to 2 and 3 times the bar diameter. The elements were precracked to control the spacing and widths of tensile cracks as well as to reduce the influence of time-dependent effects of creep and rate of loading. During the entire test the elements were loaded to either of two load levels (around 25 kN and 45 kN) giving average crack widths of approximately 0.05 mm and 0.15 mm for the respective load level. The cracks expose the interior of the elements to the aggressive environment. Here the environmental attack was created by covering an exposed cracked surface by a 3% sodium chloride solution that was allowed to freeze and thaw during one day, with the temperature varying between +20°C and -20°C. The exposure test went on for 56 days in accordance to the Swedish Standard SS 13 72 44, so-called Borås method. Tests were also performed on 150x150x50 mm slabs. In all cases the top faces were exposed. From durability aspect, the concrete tightness and porous system are decisive. Two different concrete mixtures were used: Concrete (A) and (B) with water-to-cement ratios 0.70 and 0.45. All together, 12 RC elements were tested. Based on the amount of scaling on slabs, concrete (A) had a poor frost resistance, whereas concrete (B) had a good (or even excellent) resistance. Despite the marked differences in durability between the two concretes, they had quite similar mechanical properties, both in compression and in tension. The observed differences between scaling in RC elements and slabs were more conspicuous for concrete (A), implying that cracks are more detrimental for a poorer concrete. The rate of scaling increases also in areas that are under local high stresses, and when the element is being bent. After 56 days exposure, higher concentrations of chlorides were found in the concrete at the location of the reinforcing bar for a larger crack (approximately 0.15 mm) than in a smaller crack (approximately 0.05 mm). For longer period of exposure the chloride concentration may even out. This does not necessarily mean that the service life of the structure is expired, even if corrosion processes have initiated at the location of a crack. Consequently, better criteria for assessment of service life of cracked RC structures are needed. The experiments have shown a great complexity with respect to both physical degradation of concrete (due to frost cycles and transport mechanisms in the concrete) and the experimental conditions (in terms of loading and the temperature variations in the climate chamber). This study thus serves as a step to further illuminate the problems to be faced when trying to correctly characterise actual conditions in field.