The shells of the clam Meretrix lusoria, excavated from 12 archaeological sites (4 shell layers in shell-middens, 8 shell deposits in dwelling pits, 9 shell deposits in other pits), were seasonally dated by counting the daily growth lines. These data are summarized and used here to discuss the intensity of the shell collecting activities of prehistoric people in Japan.

As a first step in the analysis, the distribution of collection seasons was used to discriminate the seasonally-stratified deposits from seasonally-mixed deposits. All the shell layers sampled in the large horseshoe-shaped shell middens at Soya Locations A and B and Kamitakatsu Locations A and B were determined to be seasonally-mixed deposits. Although seasonal clusters were vaguely distinguishable in some sampling blocks from these sites, it was difficult to follow the seasonal sequences. Moreover, the middens were usually on a slope and were probably reworked naturally. Two shell deposits, of the 9 that were found in pits, also seemed to be secondary deposits. For example in the case of pit 2 at Soya Location A, the lower layers of the shell deposit consisted of shells having been collected during single seasons. Contrarily, the upper layers were mixtures of clams that had been collected during various Seasons of the year, suggesting secondary deposition including shells falling into the pit from neighboring deposits. The seasonally-mixed deposit was found in pit 3 at Soya Location A. This deposit was probably composed of shells derived from neighboring shell deposits when the fireplace found over the shell deposit was made. It should be noted that some of these secondary deposits were thought to be primary deposits at the time of excavation, but seasonal dating of the shells revealed the mixed stratification expected of secondary deposits.

Most of the shell deposits in abandoned dwelling pits or other pits seemed to be primary deposits in which the shells of a particular collection season were clustered stratigraphically. The stratification of these primary deposits is illustrated in columns in Fig, IV-1. Some of the deposits can be subdivided into layers along the clear stratification observed during the excavation. In all these primary deposits, only 2 depositions, those of Kode dwelling pit 413 and Soya Location A pit 3, were determined to contain clams that had been collected during one 1/8-year collection season. There were layers consisting of the shells which had been collected in one season: the lower layer of Nishinoyato dwelling pit J-5, the upper fill of Kowashimizu dwelling pit 179 and the upper fill of Miyamotodai pit 1-2. The remaining layers contained shells which had been collected in more than one season, 5 or 6 seasons in the majority of cases, and each layer showed an uninterrupted seasonal sequence changing successively with the calendar from the lower to the uppe part of the layer. For example, the collection seasons in the shell deposit in Kode dwelling pit 407 changed from early summer at the lower part of the deposit to late autumn at the upper part.

The second step in the analysis was to estimate the duration of the accumulations by studying the seasonal sequences of the primary deposit. Unfortunately, seasonal dating by daily growth lines makes it difficult to assign the collection seasons to a specific year. Even if the lower part of the shell layer is a spring deposit, for example, and the next is a summer deposit, they are not necessarily from in the same year. It is possible that the layers were deposited several years apart, but using this technique it is ditficult to determine the differences. In the case where a distinctive pattern of growth increments, especially in the winter band, is detected frequently among analyzed clams from a deposition, this pattern can be useful for estimating contemporaneous collection in the same year. The estimated duration of the accumulations in cases where seasonal sequence changes with the calendar was detected are therefore usually of minimum value.

For clarity, the dates were summarized in histograms (Fig. IV-2). Successive seasonal sequences appear as unbroken histograms, while discontinuous sequences show up as interrupted histograms. Among the unbroken or consecutive histograms, the shell deposits in Kode dwelling pit 405 and Natsumidai dwelling pit 6 are assumed to span more than one year. The shell deposit in Natsumidai dwelling pit 6, in particular, was excavated three-dimensionally, and the orientations of all seasonal layers were found to be very similar (Koike, 1978). The clams in the deposit span from the early winter in the lower fill to the spring of the next year on the upper surface. Judging from the inclination of seasonal layers, they were thrown into the abandoned dwelling pit from the northern corner. This continuity suggests that the prehistoric people did not dispose of shells at random in their settlement; rather, they continuously used the depressions of abandoned pits or dwelling pits as refuse places.

The accumulation rate of the primary deposits was calculated from the volume of the shell deposit in the largest 1/8-year season in each deposit (Fig. VI-3). For example, in Kode 407 the shells which had been collected during the main season in early autumn were 50% of the total shells analyzed. The volume of the shell deposit was estimated to be 2,1001. These data gave an accumulation rate of 1,050 ± per 1/8-year season, or 50% of the volume.

The accumulation rates in the primary deposits ranged from 16 dm³ per 1/8-year season in the shell deposit in Kode dwelling pit 413 to 1,050 dm³ per 1/8-year season in the shell deposit in Kode dwelling pit 407. This wide variation is thought to be due to differences in the frequency of disposal in a single 1/8-year season. All the clam samples obtained from the shell deposit in Kode dwelling pit 413 had been harvested within ±8 days of the 153rd day after the annual growth minimum, with the variation probably due to a counting error. In other words, the shell deposit was formed in a very short period of time, possibly at one time, and the amount may be regarded as a single disposal unit.

In general, the shell deposits of the Jomon period showed much more rapid accumulation rates than those of the later ages, the 3 shell deposits of the Kofun period and the 2 shell deposits of the Edo period (Fig. IV-3). Some of the Jomon shell deposits were formed at an average rate of 10 dm³ per day during the main season of shell collection. The data suggest that shell collecting was one of the routine exploitation activities of the Jomon period.

Seasonal intensity of shell collecting activity for each site was generalized by aggregating the seasonal composition of each deposit in proportion to the volume of the shell deposit (Fig. IV-4). The black histograms are those of shell-midden sites where shell deposits were seasonally mixed. These are thoguht to be preferable for estimating a general seasonal pattern of shell collecting activities. The white histograms are those of sites where all the sampling blocks were from shell deposits in dwelling pits or other pits. These usually had a more particular seasonality.

Most seasonal intensity patterns indicate the continuation of collecting activity throughout the year, suggesting that these sites were used as permanent settlements.

At Soya Locations A & B and Kamitakatsu Locations A & B, shell collecting activity had been intensive during spring, but had gradually decreased in summer and was slack during late autumn and winter. A similar pattern was also recognized for the Nishinoyato, Kowashimizu, and Natsumidai sites, where the duration of accumulation exceeded a cycle of one-year. This pattern is thought to be characteristic of the shell-collecting activity of Jomon people. Clams are easily collected during the spring season when tidal flats are exposed for a longer time at ebb tide during the spring equinoxal tide. Beside this ecological factor, shortage of other food resources is thought to have occurred between late winter and early spring, while in autumn and early winter the people were busy gathering plant foods and migrating birds and fish.