Grace Rajnovich's (1983) thesis describes excavation of two Selkirk (Late Woodland) habitation structures at the Spruce Point site (DjKq-1) in the Kenora District of western Ontario.  The structures were defined by a color and texture contrast between the interior and exterior sediments.  Both had interior features.  Neither structure was completely exposed.  Rajnovich provided inferred outlines of the structures.  She estimated their dimensions to have been about 4m in width x 6 or 7m in length.  The plan view of Structure A is shown to the right.  The plan view of Structure B is here.

Above: Structure A (from Figure 8 of Rajnovich 1983:98).  This will be Structure 2235 in the database.

Continuing the Ontario theme . . .

Ash Rapids East

A 1988 paper by C. S. "Paddy" Reid contains illustrations of several prehistoric structures.  One of those is this structure from the Ash Rapids East site (DjKq-4, Kenora District) identified as an Archaic house.  The focus of the paper (Reid 1988) is largely methodological, and few details are provided about the structures.  This structure appears to have been a shallow basin, defined by the contrast between the basin fill and the surrounding matrix.  No information is given on how the structure was dated (the only entry for the Ash Rapids East site I could find in the Canadian Archaeological Radiocarbon Database pertained to a much later Blackduck occupation). A quick search online did not produce any later papers or reports that discuss this structure, but that doesn't mean they don't exist.  If you have more information about this structure, please let me know. I estimate the size of the structure to be about 3.5 x 4m based on the plan map.

Smoothwater Lake

The second Ontario Archaic structure of the day is a possible "tent ring" from the Smoothwater Lake site (CiHd-1) in the Timiskaming District.  The possible structure is described in John Pollock's (1975) M.A. thesis (available online here):

     "Certainly intriguing but hypothetical is a feature of large spaced rocks that may represent a tent ring (Fig. 12).  The large rocks that comprise the ring were plotted on the upper 3-inch level and thence removed, even though several extended into the second 3-inch level.  In hypothesizing this feature, data from all levels were considered and three main criteria were used in reconstructing it.  The rocks are all over 5 inches in diameter, a central hearth is present, and eleven of the seventeen projectile points excavated were concentrated within the rock feature at CiHd-1. These points all pertain to the late Shield Archaic period" (Pollock 1975:51,56).

Pollack (1975:178) estimates the footprint of the structure to have been about 10' x 12' (~ 3.0 x 3.7m).

Above: Figure 9 from Reid (1988:211).  This will be Structure 2232 in the database.

Above: Figure 12 from Pollock (1975:55). This will be Structure 2233 in the database.

I stumbled upon a paper by Reid and Rajnovich (1991) that describes three Woodland structures from the Ballynacree site in the Kenora District of western Ontario.  The structures are about 8m x 4m, defined by both postmolds and contrasts between interior and exterior sediments.  At the time the paper was written, the three structures from the site constituted the majority of known, excavated Laurel houses.  I do not know if that is still the case.

Three radiocarbon determinations suggest that House 1 dates to about AD 1250. This is, of course, a relatively late age for something that is called "Middle Woodland."  These houses, like others that strain the Midwest/Southeast-centered cultural-historical framework that I'm currently using, are encouraging me to rethink the organization of the database.  Since the relationships between our temporal and cultural chronologies are not consistent across the eastern woodlands, it may be a good move to separate them in the organization of the database. I'm considering simply dividing "time" into discrete blocks (say 200, 500, or 1000 years in duration) and then several columns to categorize the cultural-historical placement of the site (e.g., Archaic - Late Archaic - Brewerton).  This would allow one to plot groups of contemporaneous structures within thicker or thinner slices of time.  Something to think about.

Above:  Figure 3 from Reid and Rajnovich 1991:196.  The caption is "The Ballynacree site excavations revealed three houses with their associated features--a complete Laurel village.

Today I gave a short presentation on the EWHADP to the DINAA (Digital Index of North American Archaeology) workshop that is taking place at the University of Tennessee.  I couldn't go in person, so I missed being able to meet the participants face-to-face. I also missed out on the sandwich and salad bar lunch, which, according to the workshop agenda, included "your choice of a brownie or a cookie + two gallons of extra coffee."  That sounds about right.  Anyway, I think the presentation went well and I hope that we can figure out how to made the EWHADP data accessible through the DINAA database and vice versa.  That could be pretty sweet. The presentation is here (small version without animations).

Daniel Zulandt's M.A. thesis (Zulandt 2010 - available online here) contains plan maps of the 1957 and 1958 excavations at the Zencor Village site (33-Fr-8) in Franklin County, Ohio.  Those excavations exposed the remains of at least three post structures dating to somewhere in the period A.D. 500-1000.  Raymond Baby and colleagues mentioned the existence of these structures in several short papers (e.g., Baby 1971; Baby and Shaffer 1957; Mays and Baby 1958), but with scant details:

"The remains of at least three houses . . . The structures represented by the postmolds were 25 to 36 feet in diameter, with an overlapping wall serving as an entrance.  The postmolds averaged 0.6 feet in diameter and were set in the ground 1.5 feet deep and 2.5 feet apart" (Baby 1971:196).

The plan maps reproduced by Zulandt (2010:Figures 5, 6, and 7) show the post structures and their relationships to features at the site.  I don't know that these maps were not available elsewhere prior to Zulandt's thesis, but this is the first place I have seen them and I think it is useful to point out their existence. 

I already had an entry for the Zencor Village site in the database.  In the next issues of the database, however, there will be one entry for each structure.  The dimensions of each structure will be estimated from these drawings.

Figure 7 from Daniel Zulandt's thesis (Zulandt 2010:25), showing remains of structure in Area A' (this will be Structure 2227 in the database).

Figure 6 from Daniel Zulandt's thesis (Zulandt 2010:24) showing remains of structures in Area B (Structure 670 in the database) and Trench 1 (Structure 2212 in the database).

I produced the map above for my presentation.  It shows the current number of listings in the database by state/province.  The large numbers of structures at many Mississippian settlements produce the high counts in Illinois and several states in the Southeast (figure to right).  The dataset is weakest all along the Atlantic coast and on the western side of the Mississippi River. 

It would be interesting to know to what degree we can attribute the absence of database listings to (1) the absence of structures and/or (2) the absence of excavated structures.  In other words, does the absence of structures in the database reflect a real absence of structures or, alternatively, simply the absence of information about those structures? 

The EWHADP will not be able to answer that question by itself.  First, it is far too early in development of the database to know if a lack of reported structures from an area means that structures have not been identified there.  There are a lot more data to be entered and a lot more tree shaking to be done before we can really see where there are spatial gaps.  Second, because the EWHADP is only concerned with positive evidence (i.e., sites with structures present), it cannot answer questions about negative evidence (i.e., sites with no structures present).  Linking to efforts like DINAA will be required to allow one to confidently discriminate between the absence of evidence and evidence of absence.  Such a linking will also allow the spatial distribution of structures to be systematically examined in relation to the spatial distribution of other kinds of archaeological remains.  Sounds like a winner to me.