Long-term creep experiment on some rocks observed over three years

Since August 1974 the authors have been carrying out an experiment on creep by bending three small granite beams and three gabbro beams. While making measurements, an optical flat is set to produce interference fringes of Na-D light upon the upper polished surface of the beam, which was previously bent convex upward. By analyzing the fringes the profile of the surface is determined with in an accuracy of one-tenth of a wavelength. The routine determination of profiles gives a change in the amount of bending with time.A similar experiment on two large beams of granite was carried out for 21 years by Kumagai and Itô. It took 10 years in order to find out the secondary creep of granite. However, in this three-year experiment it has been found that the secondary creep of the granite specimens gives a viscosity of about 1 · 10²⁰ poise, which is comparable to that obtained over 20 years in the previous experiment (Kumagai and Itô). As for the gabbro specimens, the present authors cannot yet definitely recognize the secondary creep, because gabbro creeps much more slowly than granite.Both this and the previous experiments have revealed that the creep of rocks does not show a steady and monotonous progress, but a repeated “turn back” with various periods, some of which are longer than one year. In order to explain the creep accompanied by the “turn back” phenomenon, the authors venture to hypothesize that the elastic constants of a test-piece slightly vary with time during the advance of creep.