At the 4th Eurobitume Symposium a fatigue life definition wasintroduced, which was based on the dissipated energy concept (Hopmanet al, (1989)). This new definition is based on a change in amaterial related aspect (dissipated energy per cycle) insteadof a specimen property. It was shown that, using the dissipatedenergy law and this new fatigue life definition, the results wereexchangeable for strain and stress controlled fatigue tests evenat different frequencies. These experiments were only carriedout in a 4 point dynamic bending test. In this paper the resultsare presented of a comparison of fatigue lives measured in 2 and4 point dynamic bending tests. In spite of the new fatigue lifedefinition, the results are not comparable for both tests. Thismay be due to the composition of the mix at issue and the smalldimensions of the specimen in the 2 point bending test.

Also the healing phenomenon is studied in the 4 point bendingtest. Rest periods between loadings ought to enlarge the fatiguelife considerably. After a resting period the stiffness modulusis largely restored. If instead of resting periods (with zeroloading), load blocks are used with a smaller load amplitude,it is shown that during these pseudo rest periods the stiffnessmodulus will increase. Depending on the duration of the pseudorest period and the amplitudes of the loadings, this increasewill be followed by a decrease in stiffness modulus. These findingsconfirm the opinion that healing already takes place during thefatigue process and should be taken into account in the interpretationof the fatigue measurements and the characterization of the fatigueproperties. Based on the dissipated energy concept a model isproposed taking into account the healing phenomenon. This modelshould enable a more fundamental description of the asphalt fatigueproperties including healing.