Mg alloys have a high potential for reduction in CO2 emission because of their high specific strength and stiffness [1]. For more applications of Mg alloys, it is desirable to improve creep resistance because Mg alloys often show poor creep resistance. It has been reported that Mg alloys containing Ca showed high creep resistance and elevated temperature strength [2, 3]. Recently, Bohlen et al. [4] suggested that dynamic recrystallization (DRX) occurred due to particle-stimulated nucleation (PSN). Recrystallization due to PSN tends to occur in metals containing large-sized particles more than 1 lm [5]. Insoluble second phases such as Al2Ca and Mg2Ca are present in Mg–Al–Ca system alloys, so that DRX is expected to be enhanced due to the PSN mechanism in Mg–Al–Ca system alloys. Several studies [6–8] showed that DRX occurs during hot deformation in Mg–Al–Ca system alloys. However, DRX in Mg–Ca alloys has not been understood sufficiently. In the present paper, compression tests are conducted on Mg–6Al–2Ca–2RE (in mass%) alloy at 523–573 K with 10–1 s and its DRX behavior is investigated. A Mg–6Al–2Ca–2RE (in mass%) alloy ingot was prepared by die-casting (Mitsui Mining & Smelting Co., Ltd.). The chemical composition of the alloy is listed in Table 1. Annealing was carried out at 683 K for 108 h for homogenization. Microstructure of the annealed alloy is shown in Fig. 1. The grain size of the alloy was approximately 20 lm. The second-phase particles were observed at the grain boundaries. Energy-dispersive X-ray spectroscopy in a transmission electron microscope showed that the precipitates were (Al, Ca, RE) compounds. The cylindrical specimens with 10-mm diameter and 12-mm height were cut from the ingot after homogenization. Compression tests were carried out at 523 and 573 K with constant true strain rates of 10–1 s. Microstructure of the specimens after the compression tests was observed with an optical microscope. It took less than 5 seconds for the specimens to cool off to room temperature after stopping the test. The grain size was measured by the line intercept method (d = 1.74L, where d and L are the grain size and the line intercept length, respectively). Microstructures of the specimens deforming to e = 1.6 at 573 K with 10 1 s are shown in Fig. 2. DRX was completed throughout the specimen and the grains were almost equiaxed in these specimens. The average grain sizes were 8.8, 6.7, 7.1, and 7.1 lm at the true strain rates of 10, 10, 10, and 1 s, respectively. In Mg–Zn– Y–Zr alloy [9], the DRX grain size depended on the strain rate. In the Mg–Al–Ca–RE alloy, however, effect of strain rate on the DRX grain size was small. It appears that the distribution and size of the particles did not depend on the strain rate. Microstructure of the specimen deforming to e = 1.6 at 523 K with 10 s is shown in Fig. 3. DRX occurred at 523 K as well as 573 K, but non-recrystallized regions were partially observed at 523 K. The grain size of the recrystallized regions was 5.5 lm, which was a little lower than the grain sizes at 573 K. Some studies [10, 11] showed that the DRX grain size of Mg alloys is governed by the Z-parameter (=_ eexp(Q/RT)), where _ e is the strain rate, Q is the activation energy for M. Hakamada (&) A. Watazu N. Saito Materials Research Institute for Sustainable Development, National Institute of Advanced Industrial Science and Technology, 2266-98 Anagahora, Shimo-shidami, Moriyama-ku, Nagoya 463-8560, Japan e-mail: masataka-hakamada@aist.go.jp