It seems to be influential that the process of age-hardening of Al-Cu alloy is associated with the congregation of Cu atoms on the (100) planes of crystal as suggested by Desch (1934) and Preston (1938). As to the cause of artificial age-hardening of the alloy, there are precipitation theories as suggested by Fink and others, but several authors assume that the phenomenon is associated with the congregation of solute atoms within the solid solution. The present author investigated the hardening phenomena of Duralumin, Al-Cu alloy, Lautal and Al-Mg2Si alloy by tempering these aged specimens at various temperatures (180°, 220°, 260° 300° and 350°) for various durations (from 1/1000 to 80 hr.) and measured the hardness and impact value directly after tempering and repeatedly measured subsequent changes in hardness of the specimens with the lapse of time at room temperature. When naturally aged specimens are tempered at above mentioned temperatures, the first effect is the sudden reduction of the hardness accompanied with the increase of impact value and the decrease of electrical resistance. On ageing these softened samples at room temperature, they recover the hardness to the same degrees as that of original specimens accompanied with the decrease of impact value and the increase of electrical resistance, and the author called this phenomenon under the name of “secondary hardening” (denoted by II in Figs.). The phenomenon of sudden reduction o ` hardness on short time tempering may be explained that the congregated solute atoms have gone back into solution before the precipitation of excess solute begins, and the subsequent re covert of hardness is attributed to the re-congregation of solute atoms as the quite same process of natural ageing. Softening and hardening processes may be repeated several times as indicated in the Fig. 20 by the repetitiozl of short time tempering and natural ageing. On the prolonged tempering, the increase of hardness is observed as denoted, by III in Figs. and the author called, this hardening under the name of “temper hardening” according to the suggestion of Prof. K. Honda (1939). With the rise of hardness due to the “temper hardening” boundary precipitations become visible under the microscope, and then the hardness reaches maximum. (III) On further duration of times of tempering, the hardness gradually decreases and the lamellar structure becomes visible under the microscope. On more prolonged tempering, it finally loses the regularity- of lamellar structure and alters its form into small spheroidal particles. The lamellar structure seems to be so called “intermediate” phase as indicated by Guinier and his collaborators (1938), but the maximum hardness takes place prior to the appearance of lamellar structure. The experimental data on age-hardening of Duralumin alloys have led to the following conclusions: 1. There are three hardening phenomena in Duralumin alloys as follows: (a) Primary hardening (natural age-hardening). (b) Secondary hardening. (c) Temper hardening (artificial age-hardening). 2. Primary hardening may be explained by the theory of congregation of solute atoms which are supersaturated by quenching. 3. The mechanism of secondary hardening may be explained by the same theory as that of primary hardening. 4. During the primary and secondary hardening, the reduction of impact value and the rise of electrical resistance are similarly observed. 5. “Temper hardening” occurs on the basis of precipitation of “intermediate” phase or of compounds, but the maximum hardness is attained prior to the appearance of the lamellar structure under the microscope.