Firstly a short survey is given of the phenomenological relations between the ultimate properties and the relaxation spectra of polymers. It turned out that in many cases the temperaturetime positions of brittle-tough transitions agree quite well with those of certain relaxation maxima. Further, it is demonstrated that the ultimate properties (in particular the extensibility of poiymers) are essentially determined by the growth of flaws. A description is given of an indirect method to examine the average flaw growth during tensile tests. By means of this method, it is sho~ that in the case of PVC the secondary (short time and low temperature) relaxation mechanisms cause a first brittle-tough transition by diminution of the macroscopic stress. A second transition occurs when the microscopic notch stresses are diminished by the main (long time and high temperature) relaxation mechanism.
* The wide gap between the main transition and the main relaxation maximum at long times is, at least partly, due to the fact that at high extensions, as a consequence of the so-called cold flow of the material, the real temperature of the specimens rises considerably above the surrounding temperat ure.~ l t * That means that the so-called stress intensity factor K <~s) of thermoplastics can be, apart from the external stress and the geometric size of the crack, a function also of the time and the temperature.