Ethnoarchaeological research has shown that the assumptions formerly guiding the investigation of hunter-gatherer site structure are invalid, at least as general rules. Further research is now required to identify the determinants of site structure and assess their effect. Here I describe those aspects of Alyawara behavior that shape site structure at residential base camps. I then compare certain features of the Alyawara case with those of the !Kung and the Nunamiut. I tentatively infer that most of the variability in site structure in these cases is a function of differences in the degree of reliance on food storage, seasonal variation in weather, household population size, and the length of time activity areas are in use. Predator pressure and the relative importance of interhousehold food sharing may also be involved. The apparent importance of food storage suggests that variation in site structure may parallel Binford's (1980) forager-collector continuum. Further research is required to determine if this is so, and whether other aspects of behavior also affect site structure. These observations have important implications for the conduct of archaeological research, especially for the spatial scales at which analyses of site structure are undertaken, and for the sizes of refuse items considered in such analyses.

The Early Central California assemblage was described on the basis of four sites in the delta region of California's central valley. The original reports interpret the sites as residential villages with associated cemeteries. In this discussion I argue that the evidence for residential occupation is very weak and that the sites are specialized mortuary mounds. The implications for California archaeology are discussed.

The data presented here, by providing examples of resource diversity in Belize, add support to the current consensus that a basis indeed exists in the Maya Lowlands for placement of research emphasis on elucidating patterns of local or intercommunity exchange as a step toward understanding socioeconomic integration. The point also is made that any generalized model of exchange will fail to shed light on lowland integration unless patterns are defined for exchange that operated at the local or inter-community level. It is suggested that these patterns be sought against a backdrop of differential resource distribution.

Zooarchaeologists have used several methods to assess the relative dietary contribution of species found at archaeological sites. The most common methods are either based upon the assumption that bone weight is a fixed percentage of total body weight, or require estimating an "average" body size for identified taxa. In fact, the relationship between parameters of bone and body mass is generally allometric and can be described by linear regression. Use of allometric models places original body mass predictions on a more sound biological basis and makes calculations of "average" weight unnecessary. The potential of allometry is discussed and objections addressed with the goal of encouraging others to develop allometric formulae and to use them in their research.

This paper discusses a method for the analysis of spatial patterning which is based on aspects of the study of spatial autocorrelation. By means of computer simulation it is shown that the join count statistic (Cliff and Ord 1973; Moran 1950) has a wider validity than was presumed originally. Some archaeological applications from the analysis of a late Mesolithic cemetery in Karelia, USSR, are presented.

If research on the Initial Variant of the Middle Missouri tradition is to progress, more attention must be given to the development of a processual understanding of the groups represented. This article provides a discussion of the Mill Creek culture as it has been formulated through the culture-history approach. The background information is provided as a means of gaining insights into factors of processual significance that can be investigated deductively. The paper then presents a model of Mill Creek culture process and suggests hypotheses that are believed to have broad implications for understanding the causes of change and interaction within the Initial Variant.

In Bordes's typology for the Lower and Middle Paleolithic, he defines four major classes of side scrapers: single, double, convergent, and transverse forms. Data from three Middle Paleolithic assemblages, one from Iran and two from France, suggest that these scraper classes may be interpreted as representing stages in reduction of flake blanks through continued reuse and remodification.

The use of dry metallic powder is suggested as a means of enhancing photography of obsidian or glassy stone tools.

Bituminous coal, previously suspected as a contaminant in radiocarbon samples, has now been demonstrated to be a source of dating error in archaeological and geological samples. Standard laboratory procedures are inadequate to recognize or remove coal from samples, but techniques for identifying coal particles include paleo-botanical and reflectance analyses, and possibly accelerator dating.

The use of architecture as an artifact gives archaeologists a data set that we have rarely employed to examine cultural evolutionary questions. In this paper I consider the determinants of some architectural forms and propose some causes of architectural change, specifically in the American Southwest. Using ethnographic material from around the world, I demonstrate the relationship between structure seasonality and two of the most common southwestern architectural forms, pit structures and pueblos. The ethnographic record also suggests that the conditions surrounding the southwestern use of pit structures, and to a lesser degree of pueblos, are rather different from what many archaeologists have supposed. The differences in the conditions under which pueblos are used rather than pit structures form the basis for a theoretical framework that begins to explain the transition between the two kinds of structures in the prehistoric Southwest.

Twenty years ago excavation began on Marmes Rockshelter at the mouth of the Palouse River in southeastern Washington. Radiocarbon determinations from cave and adjacent floodplain sediments indicated an established pattern of human use by 10,800 B.P. However, a scarcity of dates between 6000 and 1300 B.P. suggested that either these strata were undated or the cave was unused. Examination of all Marmes radiocarbon ages showed agreement between shell and charcoal dates, but also some sediment mixing. Recent radiocarbon dating of charcoal, shell, and bone verifies previous determinations and supplies data between 6000 and 1300 B.P.

Methods for recognition of site formation processes are slowly being developed. Ethnoarchaeological data on refuse disposal suggest that comparison of the distribution of large and small objects across a site may be an interpretive method because size is often a factor in determining whether an object becomes primary or secondary refuse. To test this in the field, I carried out microdebitage analysis on soil samples from a stone circle site in the Northern Plains. Comparison of the microdebitage and macroflake distributions indicates the utility of microdebitage analysis for study of use and disposal in lithic tool manufacture areas.

To enhance the recovery of dense, nonfloating botanical remains, we used a siphon subsequent to the normal flotation process. The siphon vacuums the heavy fraction of a flotation sample and retrieves most of the non-floating botanical material. Experiments show an improvement averaging 9.9% and 11.3% by count and weight, respectively, for the recovery of high-density botanical remains using the siphon technique.

Some of the first recorded archaeology done in North America was accomplished by John Rowzéé Peyton in 1774. With two companions, Peyton escaped from jail in Santa Fe. Using a compass and a copy of the Delisle map of 1703, he made his way to an Osage camp on the Missouri River and thence to St. Louis. En route, he stopped to excavate in a burial mound. His interpretation of the mounds he saw is one of the first recorded instances of the Mound Builder myth. There is reason to suspect that Peyton may have been responsible for the origin of the myth, as an oral tradition among the educated elite of the American colonies.

Failure to distinguish clearly between human behavior and cultural behavior, as well as inattention to procedures for evaluating inferences about the past, undermine some recent efforts in archaeological interpretation. Examples from the archaeological literature of the American Southwest show how analytical confusion may arise when research strategies obscure cultural variability. We are especially concerned about instances in which archaeologists assume that variability in archaeological assemblages derives primarily or exclusively from variability in human behavior (rather than cultural behavior) or from noncultural processes that are instrumental in forming the archaeological record. Suggestions for modifying research strategies to avoid these problems are offered.

The settlement area that encompasses all the refuse-generating activities that contribute to a single midden may be referred to as a midden's "catchment." Two related issues are explored with respect to the problem of measuring midden catchment: 1) the relation between settlement density, midden size, and distribution, and midden catchment; and 2) the relation between the heterogeneity of a midden's contents and its catchment.

Determination of whether a particular age profile derived from a sample of prehistoric bones represents catastrophic or attritional mortality depends largely on the frequencies of age classes. Analyses of mortality profiles derived from samples of cervids (Odocoileus hemionus and Cervus elaphus) that died catastrophically in the Mount St. Helens blast zone indicate that a minimum sample of 30 individuals is required for a reliable inference of mortality pattern when a newborn individual of the taxon of concern has an average life expectancy of about three years. This conclusion is corroborated by drawing 16 random samples of various sizes from two paleontological faunas, each consisting of one taxon, one with a life expectancy similar to and the other with a life expectancy different from the Mount St. Helens cervids. Pre-eruption survivorship of the Mount St. Helens cervids was controlled by an estimated 20% annual harvest rate. Comparison of simulated survivorship in nonhunted cervid populations with archaeologically documented survivorship may provide insights to effects of prehistoric hunting on animal populations.

Environmental data are often compiled for archaeological sites; however, the natural distributions of environmental variables are seldom investigated. A study of Ceramic Period coastal settlements in the Boothbay region of Maine used a simulated site distribution of random locations to obtain an unbiased representation of the available physical environment. The site locational data were then compared to the random locational data to test specific hypotheses. Examples show the importance of non-site environmental data in evaluating cultural hypotheses.

We show that a structure built by the Inkas at the entrance to a cave at Machu Picchu was designed to admit the light of the rising sun only for a brief period before and after the December solstice. This structure is precisely constructed, unique at Machu Picchu, and may have played a part in the annual Inka ritual, Capac Raymi.

A five day conference on Apachean ceramics was held at the University of Colorado, Boulder from July 29 through August 2, 1985. The format of this conference consisted of a workshop with a series of open discussions rather than formal papers. Results of this conference included a reorganization of Brugge's (1982) Apachean ceramic taxonomy. The participants concluded that more than one source was responsible for the development of Southern Athapaskan ceramics. Furthermore, the development of this technology occurred, in most cases, as a complex in conjunction with a horticultural and/or pastoralist system.