On Variation of Assemblage Size in Late- Pleniglacial Hokkaido: Sampling Error or Behavioral Signature? Masami Izuho and Yuichi Nakazawa Assemblage size, here defined as the amount of artifacts recovered from a single geological layer, has provided a basis for discussing spatial and temporal variation of hunter-gatherer site occupational history and land use (Clark and Straus 1983). In Hokkaido, northern Japan, more than 400 late Pleniglacial sites typologically attributable to the Upper Paleolithic, dated to ca. > 27,000– 24,000 to 11,000 RCYBP, have been found in both surface and buried contexts since the beginning of Paleolithic research (Izuho and Akai 2005). In spite of intensive excavations mostly conducted under contract projects on the large island of Hokkaido (77,981 km 3 ), excavated sites remain a small fraction of the known total (< 15 percent). Of these, late Pleniglacial sites tend to be concentrated on large plains and basins (e.g., southern Ishikari Lowland, Tokachi Plain, Kitami Basin), where modern population growth has acceler- ated city extensions and road construction. Considering limitations of small-scale probes (Brown 1975) and a potential bias in site discovery in Hokkaido (Nakazawa et al. 2005), here we address the issue of how much assemblage size can be explained by the degree of occupa- tional intensity versus the effect of sample size. We examined the effect of trench size on a sampled population size (Orton 2000), which is normally treated as assemblage size. We performed an analysis of the relationship between excavated area (m 2 ) and assemblage size as represented by the number of recovered artifacts distributed in an intensively researched region. We wished to see if assemblage size is affected by the amount of excavated area and site location. Given the assumption that assemblage size is a reflection of human occupational intensity, if occupational intensity was spatially uniform within a geographically defined space, such as a plain or river basin, it is expected that assemblage size is correlated with the size of excavated areas. In other words, the larger the area of excavation, the more the artifacts recov- ered. Conversely, if occupations were clustered/clumped or scarce in specific locations within a region, assemblage sizes will not be correlated with amounts of excavated areas. Intensively occupied locations will have larger sites than less intensively occupied locations. Based on site densities and the high proportion of excavated sites, we selected a geographic area that late Pleniglacial hunter-gatherers would have intensively occupied—the southern Ishikari Lowland in southwest Hokkaido (Nakazawa et al. 2005:Figure 1a). Masami Izuho, Sapporo Buried Cultural Property Center, Minami 22, Nishi 13, Chuo-ku, Sapporo, 064-0922, Japan; e-mail: izuhom@serenade.plala.or.jp Yuichi Nakazawa, Department of Anthropology, University of New Mexico, Albuquerque, NM 87131; e-mail: ynaka29@unm.edu 4 IZUHO/NAKAZAWA Special Focus