Abstract
Many old documents show that the Ainu in the Edo period (1603-1867) had exploited surrounding resources by fishing, hunting and collecting and that they were migratory people. The movement pattern of the Ainu in the Edo period has generally been recognized that they had migrated seasonally from the fixed home ground. This means that the members of the home ground were stable. The purpose of this paper is to point out that many households in the Mitsuishi district of Hokkaido (Fig. 1) moved their home grounds to other settlements from 1856 to 1869 and that their residential groupings were open, flexible and variable in composition. The documents which listed the inhabitants' names in 1856, 1858, 1864, 1865, 1868 and 1869 were used for the present study.
The analysis of the movement of the home ground was derived from tracing the names of heads of households in each settlement at intervals over a year. If the name of the head of household in settlement A in 1856 was found at settlement B in 1858, he and his family were recognized as having moved their home ground from A to B.
Seasonal migration means that the Ainu migrate from the inland home ground near rivers to the seaside for fishing in spring mainly under the management of the Japanese and they return to their home grounds in autumn. In the south-eastern part of Hokkaido, most of the households had at least one seasonal migrant in 1858. Most of the households had the remainders at the home ground and they probably lived there from spring to autumn. This means that the location of the home ground did not change at least within a year. However, in the Mitsuishi district from 1856 to 1869, many households moved their home grounds to other settlements (Table 2). Both the number of houses and the location of each settlement changed largely (Table 1). When moved in each of five terms (I-V), households moved independently in most cases and two households moved together in some cases. The largest number of households which had moved and also stayed in all terms as if they had been one unit, was two. Such examples were merely five. The members of each settlement changed largely (Fig. 2). Then, the residential groupings of each settlement were open, flexible and variable in composition.
The thirty-nine persons were recognized as having moved to the other households (Table 3). The thirty of them moved by marriages and the rest by adoptions, etc. After mar-riages, the fifteen persons formed the independent new households with their spouses re-spectively, but the rest fifteen persons lived with bridegroom's parents or with bride's parents in the same houses respectively. We focus on the households that have kinship relationships to any other households through mainly the above thirty-nine persons. In many cases, when the households moved their home grounds to other settlements, the close kins (parent, child, sibling) meeted again at each destinations (Table 4, Table 5 and Table 6). The members of the households, newly moved into settlement Y from different settlements, had close kinship relationships with old members of settlement Y. Also, the members of new households, formed at settlement Y mainly by newly-married couples, had close kin-ship relationships with old members of settlement Y (Fig. 3). Then, the members of house-holds, moved out to the other settlements from settlement B, had also close kinship relation-ships with the members of each destinations respectively (Fig. 4).
The forming mechanism of the fluid residential groupings was recognized as follows The members living together in the same households (family of orientation) parted out to various settlements by marriages and etc., and then later they with each their new families (family of procreation) moved their home grounds to meet again at the same settlement (Table 8).
