Hatchery Site West is located on the east bank of the Kaskaskia River just east of the town of Carlyle, Illinois, and immediately downstream from the location of the Carlyle Reservoir dam being constructed (Fig. 1). The site is located on the crest of the T-2 terrace where the river is currently cutting into this old terrace system. The old terrace bends to the east just north of the site while the present river channel continues north resulting in a broad, low T-0 terrace formation just north of the site. The terrace on which the site is located is very broad and not heavily dissected and extends eastward joining with a fairly good-sized hill formed of glacial till concentrated in what was probably crevasse fill.

The accurate definition of the boundary of the site was made possible by the controlled surface collection and is shown on Fig. 1. The total area of the distribution of ceramics definitive of the site is approximately 1.23 acres. Within this distribution there are two distinct concentrations of ceramic debris (see Fig. 2). In addition, there was a recognizable difference in the color of the soil within the area defined by the ceramic distribution with two distinct areas of heavy humus content and brown coloration. These two areas corresponded well with the areas of high ceramic concentration. These two areas will be described separately since, as will be shown, they represent distinct components of the several present on the Hatchery West site.

The initial problem in approaching an archaeological site as a structural entity is to define the boundaries of the sampling universe within which the investigator must work. In the case of Hatchery West, it was hoped that by following the procedures mentioned earlier (having the fields especially plowed, waiting until there had been sufficient rain to settle the newly plowed earth, laying out a grid over the entire plowed area, and collecting all items exposed on the surface and bagging them by grid square) the analysis of the resulting data would permit definition of the boundaries, as well as something of the internal structure of the site. Aside from the obvious advantages of defining the universe in which we would work, we hoped to investigate the nature of the relationship between the structure of the site as defined by the surface distribution of cultural items and the structure of the site as defined by the spatial configuration of sub-surface cultural features. This kind of investigation was prompted by an awareness that certain general principles of site selection as well as guide lines to selecting the locations on a site for excavation have been traditionally employed for many years. It seems fair to say that archaeologists have tended to select for excavation sites with ihe greatest yield of cultural items from surface collections and have tended to excavate areas on such sites selected in terms of two general criteria: 1) where there are above ground indications of cultural features, e.g., such things as mounds, house depressions, walls, etc.; and 2) where there is the greatest density of cultural items observed in the surface reconnaissance. I have previously addressed myself to the problem of obtaining an adequate and representative sample of both cultural items and cultural features (Binford 1964). The work reported here is an attempt to evaluate objectively the assertion that traditional methods of selecting locations for excavation could lead to inadequate and misrepresentative data about the site being investigated.

It should be pointed out that, in general, cultural features are types of facilities representing a major investment of a social unit and as such are extremely important indicators of the nature of the activities conducted at a given location. For this reason we planned our recovery program so as to obtain as much information about cultural features as possible. The following analysis has been conducted with the aim of maximizing information on functional and temporal differences between the formal variants of the recovered sample.

One of the most significant aspects of the 1963 field season in the Carlyle Reservoir was the discovery of a housetype previously unknown in the Middlewest. There were four such structures located in the southernmost end of the West Field of Hatchery site. They were “keyhole” in shape, and consisted of a round, semi-subterranean floor and an extension, subrectangular in shape, which angled off toward the east-southeast.

A detailed description of the structures follows. The description will in turn be followed by a comparison between and an interpretation of the four structures. The order of description does not proceed on the basis of feature numbers or inferred chronology of occupancy. The descriptions follow the order of completeness of information, beginning with the structure about which the most data was obtained and ending with the least well known.

It has previously been argued (Binford 1964) that because of the differences in the formal properties of cultural items (as distinguished from cultural features) that there are different problems associated with obtaining an adequate and representative sample of both items and features. Excavation and recovery techniques which might be adequate to the recovery of a fair sample of cultural items might not be sufficient for supplying similar information regarding a population of cultural features.

The investigation of Hatchery West was conducted with an eye to this problem, and the methods employed were aimed toward insuring an adequate and representative sample of both cultural items and cultural features. These sampling procedures allowed us to investigate the relationship between these two classes of archaeological data and to assess the degree to which the formal-spatial structure of a population of items is correlated with the formal-spatial structure of a population of features.

In this section we move from the analysis of various formal classes of data defined by non-historical criteria to an anlysis of categories of artifacts treated historically, i.e., their spatial and formal characteristics within the remains of single communities.

The discussion of these occupations, although suspected of being rather numerous, must by virtue of the nature of the data be very skimpy. The relevant data are amost exclusively obtained from the surface collections. Their manner of clustering suggests that there was some range of variation in the activities carried out.

Even a cursory examination of the projectile points recovered in the surface collection indicates the presence of multiple preceramic occupations. Types represented include Dalton points, Modoc Expanding Stem points, Faulkner Side-Notched points, and points of the Saratoga and Boaz type clusters. The detailed typological analysis of these points, as well as of other chipped stone materials, should make possible a more rigorous definition of the number of components. At present our understanding is limited by the nature of the data thus far analyzed.

We are dealing here only with the data from cultural features; our analysis and conclusions cannot be considered final until the information from plant and animal remains and the analysis of cultural items are complete. We have attempted to isolate differences attributable to functional variability in the range of activities represented by the cultural features; we have also attempted to delimit those particular features representing different historical episodes in the occupation of the site. The results of this analysis were the delineation of 4 major discrete occupations to which the vast majority of cultural features could be referred. Each occupation can be structurally defined in terms of its composition in functionally differentiated types of features.

We now turn our attention to the analysis of these components as units to be compared in terms of a limited number of attribute classes, each largely dependent upon the taxa resulting from the analysis discussed above. With respect to cultural features, we can compare the components with respect to the following broad attribute classes: 1) form, spatial arrangement, and history of domestic structure; 2) form, spatial arrangement, and history of burials; and 3) form and spatial arrangement of classes of pits.

Utilizing these criteria, we shall compare the communities recognized in the hope of arriving at some insights into the settlement systems of the socio-cultural units represented and into the correlated forms of social organization.