At eight stations in and around shrine forests in Kyushu and Tokyo metropolitan area, long-term use patterns of natural tree hollows and nest boxes by the Japanese giant flying squirrel, Petaurista leucogenys, was studied during 1972-1981, together with an experiment in captivity. The results obtained are summarized as follows: 1) The flying squirrel seemed to remember the locations of the tree hollows and nest boxes within its home range, and paid frequent visits to them at night. 2) Gnawed marks were often found at the entrance of tree hollows and nest boxes. Distribution of the marks had no relation to the use ratio of nest boxes, but this gnawing habit appeared effective in securing many refuges in the woods 3) The animal often rotated the primary nest among several nests within its home range. Duration of sttlement in a nest ranged from several days to several monthes. Occasionally, the animal mode one day's lodging in one of the secondary nests which were rarely used ordinarily. 4) Nest materials were always found in both the tree hollow nests and the nest boxes in which the animal settled for several days, although an animal might lodge for only one day even in the nests devoid of the materials. 5) The same individual was never found in a given nest over two years. Thus, the home range seemed to shift within two years at the longest. 6) A given nest was used by several individuals one after the other except for a mother and her young. An animal which had returned to a nest befor others had priority and prevented an invader from entering it. 7) Captive individuals also refused to lodge with a stranger, but accepted it in daytime probably because of a decline of the activity. 8) Intra- or inter-specific competition for a nest and occurrence of external parasites in a nest seemed to be external factors causing the shift of nest. A rainfall in the nighttime made the animal take a secondary shelter nearby. 9) Even after becoming independent, young still lived together their mother with repeated separations for several months. 10) The use ratio of tree hollow nests which had been utilized for many years averaged 32% (7-64%) which showed roundly almost no regional difference. 11) The use patterns of nest boxes at Okawachi was similar to that of the tree hollow nests at other areas in that they were used as primary nests by a certain individual for lodging consecutively to some extent. On the other hand, at Hikosan Biological Laboratory (HBL), only adult males used boxes, and then mainly as secondary nest. These males had wider home ranges than females, and the males invaded the study area at the periphery of their ranges. 12) The mean use ratio of nest boxes was 0.7% at HBL (60 boxes) and 5.4% at Okawachi (24 boxes). At HBL big trees are fewer than at neighboring areas, and the flying squirrels appeared to use the site only as feeding sites at night and to move to adjoining areas to rest in daytime; accordingly, the boxes did not facilitate an increase of the number of individuals. 13) The use ratio of nest boxes at HBL became relatively high during the period of May to September. Frequent changes in nest by males sexually active, or the shift of home range owing to the seasonal change of food items seemed to be the main cause of such a fluctuation in use ratio.