Introduction

Cootes Paradise is a large, fresh-water marsh located at the western tip of Lake Ontario, in the south-central region of the province of Ontario, Canada (Figures 1 and 2). Because of a complex interplay of historic, topographic, hydrologic and climatic factors, the region at the western end of Lake Ontario in which Cootes Paradise is situated is one of the more environmentally diverse in the Lower Great Lakes area of northeastern North America. This region was home to native communities from the end of the last Pleistocene glaciation (c. 9,000 B.C.) through to the period of earliest contact with Europeans (between A.D. 1600 and 1650). Throughout most of the 11,000-year time span, settlement in the region at the western end of Lake Ontario can be characterized as relatively dispersed, and exploitation of the marsh at Cootes Paradise was sporadic and seasonal. An exception to this pattern occurred between A.D. 500 and 1000, when settlement became focused on the marsh, and only a few sites are found in other parts of the region. The purpose of this paper is to explore reasons for the focus of settlement at Cootes Paradise that occurred between A.D. 500 and 1000, and to set this pattern into the broader context of economic and social transformations that occurred throughout south-central Ontario during this period.

The chronology and cultural taxonomy of the Woodland period in southern Ontario is the subject of some debate (Williamson and Watts 1999), and a detailed discussion is beyond the purview of this short paper. For purposes of this discussion I use a compromise scheme of periods, stages and sub-stages (see also Smith 2000). The Woodland period is divided into Early (ca. 1000 - 400 B.C.), Middle (ca. 400 B.C. - A.D. 700) and Late (ca. A.D. 500 - 1650) stages, with significant chronological overlap between Middle and Late Woodland evidenced by calibrated radiocarbon dates (Smith 1997a). The Late Woodland stage can be divided into three sub-stages: Late Woodland 1 (ca. A.D. 500 - 1000), Late Woodland 2 (ca. A.D. 1000 - 1550) and Late Woodland 3 (ca A.D. 1550-1650) (see Mason 1981).

The focus of this paper is the Late Woodland 1 sub-stage, which is represented in south-central Ontario by the Princess Point Complex (Stothers 1977; Fox 1990; Smith and Crawford 1997). Princess Point sites are distributed over an area that includes the western end of Lake Ontario, the north shore of Lake Erie east of Long Point, the Lower Grand River Valley, and the Niagara Peninsula (Figure 2). Within the Princess Point region, significant clusters of sites occur in the Lower Grand River Valley, at Long Point on the north-central shore of Lake Erie, and at Cootes Paradise.

Considered in isolation, the apparent settlement anomaly at the western tip of Lake Ontario may be thought of as an interesting local deviation from the traditional pattern of settlement location. It takes on broader significance, however, when we realize that it coincides with a period of similar settlement coalescence in the rest of south-central Ontario. At the same time, there were pervasive changes in economy, settlement system and social organization that occurred throughout the Lower Great Lakes area. Important among these shifts was the transition to resource production, based on the introduction of maize, followed by beans, squash, sunflower and tobacco. Maize first appears in the archaeological record by about A.D. 500; by A.D. 1000, fully horticultural subsistence was practiced. Prior to A.D. 500, settlement pattern was organized according to a seasonal round of scheduled mobility. After A.D. 500, site situations became more focused on particular locales, leading to the first semi-sedentary villages between A.D. 900 and 1000. Associated with these changes in subsistence and settlement is an inferred transition from the unranked, band level social organization of mobile foragers toward the ranked, tribal form characteristic of the later Iroquoian societies that inhabited the southern Ontario. The nature and rate of these transitions are, however, not well understood.

Because the term 'Princess Point' refers to a geographic feature, a site and a cultural group, to avoid confusion in this paper the following conventions will be used. The Princess Point Complex will be referred to as the PPC, the promontory will be referred to as the PPGF (Princess Point Geographic Feature), and the site name will be spelled out in full.

Environmental Setting

The region at the western end of Lake Ontario is one of the most environmentally diverse in northeastern North America. This diversity arises from a number of circumstances that are discussed in some detail elsewhere (Smith In press a), and I will only summarize them here. The general climate of the region is humid continental (Brown et al. 1980). Late Woodland 1 times fall within what is called the ‘little climatic optimal’ (100 B.C. - A.D. 1250), when mean temperature was a few degrees warmer than today (Bryson and Padoch 1981). The region lies at the northern edge of the eastern Deciduous Forest Zone of North America which, prior to Euro-Canadian clearance, was dominated by beech-maple climax forest and associated Carolinian flora and fauna (Rowe 1972; McAndrews 1994:180).

Compared to other regions of south-central Ontario, the study region is topographically and hydrologically more complex. Because it is difficult to describe this diversity in words alone, I have prepared two Digital Terrain Models (DTMs) that convey a sense of the variation in three-dimensional space. Figure 3 shows the geographic zone where the Lake Ontario basin meets the Niagara Escarpment (contour interval 15 m). Figure 4 displays the topography in the locale immediately surrounding Cootes Paradise (contour interval 5 m).

The most prominent topographic feature in the region is the Niagara Escarpment, a scarp face that runs roughly north to south across south-central Ontario (Figure 1). The land 'above' the escarpment (i.e., on the higher elevations) slopes gently to the scarp face. In the vicinity of Cootes Paradise, the elevation on the heights of the escarpment is roughly 200 m above sea level (asl), and drops to about 120 m asl at the base. The land below the escarpment to the north of Cootes Paradise and Hamilton harbour slopes more abruptly to the shore of the Lake Ontario, going from 120 to 75 m asl in five km or less. The western end of the Lake Ontario basin meets the Niagara Escarpment at the Dundas Valley (Figure 3), which was cut into the escarpment by late Pleistocene outwash and scouring.

The landscape both above and below the escarpment is cut by literally dozens of low-order streams, but relatively few second-order streams. (Some of these low-order streams have been channeled underground, and appear to have no beginning or end in Figure 3.) Only Grindstone, Spencer, Twenty Mile and Red Hill Creeks can be considered to be relatively major waterways, and none of these can be classed as a river. None have produced broad alluvial valleys in the study region. In addition, Cootes Paradise is the only major wetland in the region.

Cootes Paradise is situated at the base of the Dundas Valley behind an early post-glacial lake bar (the Iroquois Bar) that separates the wetland from Hamilton Harbour (Figure 4). The land around Cootes Paradise is about 100 m asl. There is a 25 to 30 m drop to the water at roughly 75 m asl); this drop is relatively gentle in some areas, and precipitous in others. Stream valleys and forelands border Cootes Paradise on the both the northern and southern shores, and ravines cut several of the promontories.

Although several minor creeks drain into the wetland, the primary source of water is Spencer Creek, which flows into the marsh at its western end (Figures 3 and 4). The natural outlet from the marsh was Hendrie Valley, a crescent-shaped channel through the Iroquois Bar. Beginning in the nineteenth century A.D., Euro-Canadian modifications altered the drainage patterns in the marsh. A canal was dredged through the wetland, a channel was blasted through the Iroquois Bar to form a new outlet, and Hendrie Valley was partially filled in by railway and roadway development. Drainage from Cootes Paradise flows into Hamilton Harbour, a separate basin that, in turn, flows into the main Lake Ontario basin. Current water level in Lake Ontario is roughly 75 m asl (Coakley and Karrow 1994: 1618).

Today, Cootes Paradise is a cat-tail marsh in its western third, and a body of shallow, open water in its eastern two-thirds. Pollen cores document that the floral regime was much different in the past. The pollen cores show that, during PPC times, aquatic grasses dominated the marsh flora, at least in the eastern part of the wetland (Smith et al. 1998). Examination of the surface structure of fossil grass pollen by scanning electron microscopy shows that the grass was wild rice (Zizania sp.) (Lee et al. 2000). The pollen cores also show that the wild rice stands declined rapidly between ca. A.D. 1000 and 1200 (i.e., immediately following the PPC occupation of the wetland) to be replaced by cat-tails.

To summarize, the environmental setting in the region at the western end of Lake Ontario is characterized by topographic, hydrologic and biotic diversity. This stands in contrast with other areas inhabited by PPC communities. For example, more than half of the 85 known PPC sites are located in the Lower Grand River Valley (Figure 2). Here, the Grand River flows across a wide, flat clay plain. The valley is characterized by a system of extensive river bars and alluvial terraces, both of which were favoured by PPC communities as site locations. In contrast, most of the stream courses at the west end of Lake Ontario do not feature extensive alluvial bars and terraces. The wetland at Cootes Paradise is the environmental zone in this region that most closely approximates the conditions in the Lower Grand River Valley.

Late Woodland 1 Settlement System in South-Central Ontario

As mentioned above, the Late Woodland 1 stage is manifested in south-central Ontario by the PPC. The PPC is distinguished by a material culture assemblage that includes cord-wrapped stick decorated pottery, a flake-based lithic assemblage, Levanna type triangular projectile points, and a poorly developed bone tool industry. Subsistence was based on mixed hunting and foraging, supplemented by limited cultivation of maize (Smith and Crawford 1997). The earliest maize yet recovered in the Lower Great Lakes region comes from the Grand Banks site, a PPC component in the Lower Grand River Valley (Crawford et al. 1997; Smith 1997a). Maize was also recovered at Cootes Paradise from the Princess Point site and the Bull's Point site. Although maize likely constituted a relatively minor component of PPC subsistence (Smith and Crawford 1997), the requirements of its cultivation may have influenced choice of PPC settlement location.

PPC settlement system signals a departure from patterns that prevailed during the Middle Woodland stage. Whereas Middle Woodland communities located their sites in a variety of environmental zones, PPC communities focused primarily on specific lacustrine, wetland and, especially, riverine habitats (see Figure 2). At present, there are two models that have been proposed to explain this anomaly. One is a total-replacement migration scheme proposed by Fiedel (1991; 1999), Snow (1992; 1995; 1996) and Bursey (1995). It argues that the PPC originated in groups of Iroquoian-speaking horticulturalists who migrated into south-central Ontario from regions to the south at ca. A.D. 500. These Iroquoian migrants displaced resident Algonkian-speaking hunter-gatherers, and settled into the water-oriented habitats mentioned above. The other hypothesis suggests that the PPC settlement system developed from the preceding Middle Woodland (Chapdelaine 1993; Smith 2000). This model proposes that the previously mobile Middle Woodland settlement-subsistence regime was transformed into a more sedentary pattern.

In truth, there are insufficient data available at present to definitively support either the migration hypothesis or claims for continuity. I have argued, however, that if there was continuity instead of migration, it can only be understood within the context of congruence of a particular set of material, social and ideational parameters (Smith 2000). The material parameters are the climatic, topographic and hydrologic conditions that prevailed in southern Ontario after A.D. 500. Mean temperature was a few degrees warmer than earlier (during the ‘little climatic optimal’), and lake levels were relatively constant. The latter contributed to the stability of the river bars and alluvial terraces in the flood plain system of the Lower Grand River Valley. It was these bar and terrace combinations, in particular, that PPC communities preferred. By A.D. 500, maize was likely present in small populations that were becoming acclimatized to the growing conditions in the Northeast (Hart 1999), an adaptation that may have been ameliorated by the 'little climatic optimal'. The bars and terraces exploited by PPC groups are characterized by rich, alluvial soils that would have been conducive to garden cultivation of maize.

Social conditions were likely as important as the material criteria. During the Middle Woodland stage, a mobile settlement-subsistence strategy appears to have been seasonally scheduled around spring-summer macro-band congregation and fall-winter micro-band dispersal (Spence et al. 1990). Ferris and Spence (1995: 99-100) argue that, later in the Middle Woodland stage, the macro-bands were favoured for longer periods; Wilson (1994; 1999) proposes that the macro-band settlements in some riverine habitats were virtually sedentary. I suggest that, at the beginning of the Late Woodland stage with the formation of PPC communities, the tendency toward focus on one locale was completed, and the macro-band became permanent. The resultant communities were not yet, however, the nucleated villages that characterize the later Iroquoian settlement system in the Lower Great Lakes area (cf. Snow 1996:792). Current evidence suggests that PPC settlements were loosely aggregated hamlets centred on one locale. These communities engaged in seasonal activities, including limited garden cultivation of maize, but the entire community did not disperse into other parts of a band territory on a seasonal basis.

Finally, I propose that changing ideational attitudes toward landscape were also a contributing factor. Elsewhere I have argued that concepts of 'home' would have been altered in significant ways (Smith 2000). Mobile foraging communities, all members of which circulated throughout a band territory on a seasonal basis, would have viewed the entire territory as their home, with no environmental niche being more important than others. As the tendency toward increasing sedentism developed, the concept of home became more focused on one locale, likely the location favoured by the spring-summer macro-band. This transition was completed at the beginning of the Late Woodland 1 sub-stage, when the macro-band became permanent. Home was centred on a single location, such as the terrace-floodplain situations in the Lower Grand River Valley or, as I will argue below, a wetland locale such as Cootes Paradise.

Cootes Paradise may represent the closest analogue in the region at the western end of Lake Ontario to the conditions in the Grand Valley that were so attractive to PPC communities, but it represents a departure in significant ways. It does not have massive low-lying alluvial bars where maize could be grown, nor does it have an extensive terrace system. It is also somewhat isolated from other PPC communities. Why, then, would a PPC community choose to settle in this locale at all?

Archaeology at Cootes Paradise

Archaeological work has been conducted at Cootes Paradise over the past 40 years. In 1961, Frank Vallee of McMaster University conducted salvage excavations at the Old Lilac Gardens site, located at the eastern end of the wetland. In 1968 and 1969, William Noble and David Stothers of McMaster University, conducted excavations at the Princess Point site, which is located on the promontory of the same name (Figure 2). Site survey by Stothers in 1969 recorded the locations of five additional PPC sites around the marsh (Stothers 1969; 1977). In 1984, Dean Knight of Wilfrid Laurier University carried out a site assessment at the Old Lilac Gardens site (Knight 1984). In 1995, the author and colleagues began research at Cootes Paradise to investigate the size and nature of the PPC components, to establish their environmental context, and to explore the social and economic relationships among them. From 1995 to 1998, the author directed short-term excavations at the Bull's Point 2, Bull's Cove and Sassafras Point sites (Smith and Crawford 1997; Smith 1997b), and at the Princess Point site in 2000. Site survey discovered three additional PPC sites in 1995 and 1998 (Smith 1997b; Smith et al. 1998).

The work noted above provides evidence for the presence of human groups at Cootes Paradise from Late Archaic times (ca. 2000-1000 B.C.) to the Late Woodland 3 stage (Historic Iroquoian). Table 1 summarizes the archaeological data from the wetland for all cultural periods. The most intensive occupation at the marsh, however, occurred during PPC times. The locations of PPC sites at Cootes Paradise are illustrated in Figure 4; data available for these sites, as well as interpretations of site type, are summarized in Table 2. The site sizes are estimates inferred on the basis of Stothers' (1969) survey data and recent test excavations. Attributions of presence or absence of hearths and middens are based on excavations at the Princess Point, Old Lilac Gardens and Sassafras Point, and test pitting at Bull's Point 2, Bull's Cove and Andrea (Stothers 1977; Smith 1997b). A question mark indicates lack of evidence to make a judgment.

Based on type of location, there appear to be four different locational categories at Cootes Paradise: (1) major promontories, (2) minor promontories/terraces, (3) islands, and (4) ravines. The Princess Point and Sassafras Point sites are situated on low-lying promontories on the south shore. The Old Lilac Gardens and Arboretum Nursery sites lie on short headlands that may be better described as terraces, the former on the western side of the Iroquois Bar at the eastern end of the marsh and the latter on the eastern north shore. The Rat Island and Hickory Island sites are located on small islands off the north shore. The Bull’s Point 1, Bull’s Point 2 and Bull's Cove sites lie in the lower ends of ravines that cut the southwestern edge of the Bull’s Point promontory, and the Andrea site is in the lower end of a ravine on the northwest side of Arnott's Point. Information on internal site structure is available for only Princess Point, Sassafras Point, Bull's Point and Old Lilac Gardens, and this in limited quantity.

Major Promontory Sites: Princess Point and Sassafras Point

The Princess Point site is situated on a low-lying peninsula (the PPGF) at the southeastern corner of the marsh (Figure 4). The surface area of most of the PPGF is now relatively flat, and lies at about five meters above the present water level. In the past, a relatively shallow 'saddle' across the point separated it into southern and northern sections; this saddle was filled with trucked-in soils to level the ground earlier this century. At the northern end of the PPGF the surface slopes down to a low-lying crescent-shaped area one to two meters above the current water level. The excavations by Noble and Stothers focused on the northern half of the peninsula (Stothers 1969:7-10; 1977:25-33). An excavated area of approximately 300 square meters along the northwestern edge of the peninsula uncovered pits and hearths attributed to the PPC. Stothers described this area as a habitation site (1969:8), although no post moulds were found. Smaller units were dug in the northwestern interior and in the low-lying crescent on the northeastern side. Stothers (1977:29-33) reported a stratified midden in the crescent, with a PPC stratum overlain by later components (Early, Middle and Late Iroquoian).

During the summer of 2000, test pits and one-meter squares were excavated in the northern section of the site on both the higher ground and the low-lying crescent. This work established that cultural deposits extend over an area of roughly 4,000 square meters in the northern section of the promontory. One-meter test squares uncovered a midden extending over 300 to 400 square meters of the low-lying crescent. These deposits yielded diagnostic material from Middle Woodland, Late Woodland 1 (PPC) and Late Woodland 2 components (Early, Middle, and Late Ontario Iroquoian), but none of the profiles revealed to date show distinct stratification.

The multi-component nature of the Princess Point site hinders interpretation, but there are several general points that can be made. The overall extent of the PPC component is estimated at roughly 3,500 square meters. The pits and hearths associated with PPC artifacts recovered during the 1968 and 1969 excavations on the northwestern side of the promontory suggest a more than transitory habitation, although Stothers refers to it as a small seasonal camp (1969:9). The presence of extensive midden deposits also supports the inference that the PPC component may have been a more substantial settlement. It can be hypothesized that the elevated section at the northern end of the point was a PPC habitation area, the occupants of which disposed of their garbage over a number of years on the low-lying crescent.

In 1997 and 1998 excavations were conducted at the Sassafras Point site on the south shore of Cootes Paradise to clarify the nature of the PPC component reported by Stothers at the end of the Sassafras Point promontory (Stothers 1969; 1977). This foreland consists of two separate sections. Close to the base of the peninsula is a high plateau at roughly 85 m asl. About halfway toward the tip of the foreland, the elevation falls to about 80 m asl, or roughly five meters above the present water level in the marsh. The lower area toward the tip of the point is a relatively level expanse of approximately 7,000 square meters. Test excavations documented three cultural components at Sassafras Point, one from the late Archaic (ca. 4000 to 3000 B.P.), one PPC and the third dating to Middle Ontario Iroquoian times (600 to 700 B.P.). The PPC component extends over roughly 2,500 square meters of the level ground. A total of 22 square meters was excavated in this section (Smith 1998; In press b). Numerous post moulds and a few features were uncovered, but separation of these between Late Woodland 1 and Middle Ontario Iroquoian could not be accomplished. No midden deposits and no hearth floors were found. Unfortunately, because of the multi-component nature of the deposits, little else can be said about the PPC occupation of the Sassafras Point site.

Minor Promontory / Terrace Sites: Old Lilac Gardens and Arboretum Nursery

The Old Lilac Gardens site is situated on a small headland on the west side of the Iroquois Bar at the eastern end of Cootes Paradise (Figure 4). The eastern edge of the headland was impacted by the construction of a highway along the bar. There are no surviving maps or notes to indicate where Vallee excavated in 1961 (Knight 1984). In 1969, Stothers excavated one 5-foot square near the tip of the headland (Stothers 1969). In 1984, the site was subjected to test excavations by Knight (1984). He excavated 156 square meters total, including a two meter wide trench, 34 meters long, running east-west across the site. The other testing consisted largely of two meter square units over rest of the headland. Knight reports two middens and a possible burial mound. On the basis of this work, Knight concluded that highway construction had impacted the site on its eastern edge, but the main part of the site on the headland had escaped destruction. A total site area of roughly 1,000 square meters can be estimated.

The Arboretum Nursery site is located on a low terrace toward the eastern end of the north shore of Cootes Paradise (Figure 4). It was recorded by Stothers in 1969 on the basis of artifacts recovered from the site by local collectors. In addition, Stothers excavated one five-foot unit. The artifact assemblages included Archaic bifaces and cord-wrapped stick decorated rim sherds. On the basis of Stothers map of the site, the site size is estimated to be about 1,000 square meters (Stothers 1969). No information about intra-site structure is available.

Island Sites: Rat Island and Hickory Island

Rat Island and Hickory Island are situated off the north shore of Cootes Paradise (Figure 4). Stothers discovered PPC sites on both islands in 1969. Rat Island, the larger of the two, is roughly 5,000 square meters in size. There is a level area of about 3,000 square meters on the southern two-thirds of the island. Stothers' work on Rat Island included two five-foot excavation units in the center and at the east end of the level area. In addition, artifacts were collected from an eroding bank on the south shore of the island. It is difficult to calculate the size of the site, but a rough estimate would place it at about 1,000 square meters.

Hickory Island is a long narrow islet, about 900 square meters in total size. The surface is less than a meter above the present water level at its western end, but slopes up to about two meters above the water toward its eastern end. Stothers excavated two five-foot square units in 1969, one on the high point of the island and the other at the base of the slope on the southwestern side. He reports that the latter of these units revealed shallow midden deposits. From Stothers' map of the island (1969) a site size of rough 400 square meters can be estimated.

Ravine Sites

The Bull's Point promontory is a high point of land jutting into the water on the north shore of Cootes Paradise. In contrast to the major headlands on the south shore, Bull's Point is surrounded on all sides by fairly precipitous slopes of 25 to 30 meters above the present water level. The point is cut in various places by dry ravines. In 1969, Stothers discovered a site located at the water's edge in the bottom of one of these ravines. He mapped the site and determined that it was about 500 square meters in size. In addition, he conducted limited excavations, and reported cultural deposits dense enough to be termed a midden (Stothers 1969:25). In 1995, the author returned to the Bull's Point promontory in search of this site, and duly found one in the bottom of a ravine. Unfortunately, it now appears that the ravine site discovered by Stothers was destroyed in the early 1990s. The site found in 1995, now called Bull's Point 2, is another ravine location. Two additional PPC sites in ravines have since been discovered. Bull's Cove is located on the southwest side of the Bull's Point promontory within 50 meters of the Bull's Point site, and the Andrea site was found in a ravine on the northwest side of Arnott's Point on the south shore (Figure 4). The situation and size of both the Bull's Cove and Andrea sites is very similar to that of Bull's Point, although the artifact densities in both cases appear to be lower. Two more ravine sites were found around the Bull's Point foreland; although neither of these yielded ceramics nor other diagnostic artifacts, they may also date to PPC times.

In 1995 and 1996, excavations were conducted by the author at the Bull's Point 2 site (Smith 1997a). At the lower end of the ravine, a level area roughly 15 meters wide extends inland for about 20 meters. The site covers all of the level section, and also extends up the ravine for another 10 meters; the total site area is estimated to be approximately 400 square meters. It was initially suspected that the lower reaches of the site may have been inundated by rising water levels over the past 1,000 years, but testing of marsh sediments at the mouth of the ravine during a low water period in 1998 yielded no cultural material. An area of 34 square meters excavated in the center of the level ground uncovered a small, rectangular structure 4.0 by 3.8 meters in size (Smith 1997b:9). Only one shallow pit was found inside the structure, and no hearths were uncovered. The artifact assemblage from within and around the structure includes pottery, chipped lithics and floral remains. Few ground stone tools and almost no faunal remains were recovered. The pottery sherds represent about five vessels; the lithics include formal tools, cores, informal flake tools and debitage; the floral material includes maize as well as fleshy fruit species (bramble, etc.). The AMS radiocarbon assay on maize of cal. A.D. 980 (1040) 1220 dates to end of the PPC sequence (Smith 1997b).

The ravine sites represent a settlement type virtually unknown in the rest of southern Ontario, although it is now clear that they are relatively common at Cootes Paradise. All appear to be small (400-500 square meters), seasonally occupied camps, but their purpose remains difficult to infer. The midden at the Bull's Point site, the structure at Bull's Point 2, and the suite of artifacts recovered from the latter site indicate that the ravine sites were more than overnight way-stations. It is possible that they were camps for wild rice collection, occupied by a family for the several weeks in the late summer and early fall when this cereal is harvested (see Vennum 1988, Chapter 5, for a discussion of family camps for harvesting wild rice among the Ojibway). To date, no wild rice macro-fossils have been recovered from flotation samples, however, and this hypothesis remains unconfirmed at present.

The PPC Settlement System at Cootes Paradise

Previous interpretations of settlement at the marsh argued that PPC exploitation was primarily seasonal (e.g., Stothers 1977:122-123). It is possible, for example, that PPC groups made their way overland from other more permanent sites such as from those in the Grand Valley. An alternative hypothesis, based on the data provided above, is that one PPC macro-band community occupied Cootes Paradise between c. A.D. 500 and 1000, and that the wetland locale was the permanent focus of this community. The Princess Point site appears, at roughly 3,500 square meters, to be the largest and, with midden deposits, pit features and hearths, the most intensively occupied. I propose that the Princess Point site was the main base for the community, and probably had some people inhabiting it year-round. During the summer, however, the macro-band separated into smaller groups that occupied the second tier sites. The second tier consists of the Sassafras Point, Old Lilac Gardens, Arboretum Nursery and Rat Island sites. These range from 1,000 to 2,500 square meters in size. Only the Old Lilac Gardens site had cultural deposits dense enough to be termed a midden; there is no evidence for hearths at any of the second tier sites. All of these locations are within a few hundred meters of each other, so this separation is not the same as seasonal dispersal. Seasonal marsh resources were exploited over the entire wetland at this time of the year. The third tier sites include Hickory Island, Bull's Point 1, Bull's Point 2, Bull's Cove and Andrea, much smaller stations estimated at either 300 or 400 square meters. None of these has yielded evidence for either midden deposits or hearths. The third tier sites are special purpose stations occupied for short periods to exploit specific resources. Such a specific resource may have been wild rice, although we have yet to recover macro-fossil remains of this aquatic grass archaeologically. At the onset of colder weather, the community re-congregated at the Princess Point site.

This model, then, claims that Cootes Paradise was not simply a seasonal station for a small group of mobile foragers. Rather, it was the economic, social and ideational centre for a community undergoing the transition to a more sedentary settlement regime. This community chose to focus their activities at the marsh, not simply because of abundant wild resources, but also because of a need to maintain the social integrity of the macro-band group, and because they identified the locale, not the surrounding territory, as 'home'. In this, Cootes Paradise was similar to the occupation centres in the Lower Grand River Valley. The community was somewhat isolated because there were none of the habitats that PPC groups preferred within 20 km of Cootes Paradise. It is also possible that the occupation centre at the marsh was more environmentally constrained than was the case elsewhere within the PPC area. Thus, more of the annual activities of the PPC community are represented within the boundaries of the Cootes Paradise locale than appears to be the case at analogous centres elsewhere.

Conclusions

I have argued in this paper that the focused settlement regime at Cootes Paradise during Late Woodland 1 times was not simply a local anomaly, but, rather, an expression of the broader PPC settlement system. The hypothesis of a single, macro-band level community inhabiting the Cootes Paradise locale for an extended period of time is an application of a comprehensive model that views PPC settlements as year-round occupation centers. Thus, the site cluster at the wetland was a response to material, social and ideational conditions that prevailed throughout south-central Ontario between A.D. 500 and 1000. These conditions were bound up in the economic transition from foraging to resource production, a social transformation from mobile, band-level groups to semi-sedentary communities, and changing ideational perceptions of the landscape.

This viewpoint brings a perspective to bear upon the origins of resource production in northeastern North America that sees the transition as more than a simple shift in subsistence practices. I propose that, if we are to understand the transformation here, and elsewhere, we will have to take into account a complex interaction of multi-dimensional parameters. Detailed study of settlement systems and patterns at specific locales such as Cootes Paradise provides us with one avenue of insight.

Acknowledgements. The research on the Princess Point Complex by the University of Toronto team from 1993 to 2000 was supported by grants from the Social Sciences and Humanities Research Council of Canada, the National Geographic Society, Earthwatch, the Ontario Heritage Foundation, and the University of Toronto. The work at Cootes Paradise from 1995 to 2000 was conducted with the permission and assistance of the Royal Botanical Gardens. I would like to thank Frank Dieterman and Elizabeth Major for their invaluable help with the Digital Terrain Models, and Gary Crawford and Tony Davis for comments on an earlier draft of this paper.

References Cited

Chapdelaine, C.

1993 Sedentarization of the Prehistoric Iroquoians: A Slow or Rapid Transformation. Journal of Anthropological Archaeology 12:173-209.

Coakley, J. P. and P. F. Karrow

1994 Reconstruction of Post-Iroquois Shoreline Evolution in Western Lake Ontario. Canadian Journal of Earth Science 31:1618-1629.

Crawford, G. W., D. G. Smith and V. E. Bowyer

1997 Dating the Entry of Corn (Zea mays) into the Lower Great Lakes Region. American Antiquity 62:112-119.

Ferris, N. and M. W. Spence

1995 The Woodland Traditions in Southern Ontario. Journal of American Archaeology 9:83-138.

Fiedel, S. J.

1991 Correlating Archaeology and Linguistics: The Algonquian Case. Man in the Northeast 41:9-32.

1999 Algonquians and Iroquoians: Taxonomy, Chronology and Archaeological Implications. In Taming the Taxonomy: Toward a New Understanding of Great Lakes Archaeology, edited by R. F. Williamson and C. M. Watts, pp. 193-204. Eastendbooks, Toronto.

Fox, W. A.

1990 The Middle Woodland to Late Woodland Transition. In The Archaeology of Southern Ontario to A.D. 1650, edited by C. J. Ellis and N. Ferris, pp. 171-188. Vol. 5. Occasional Publication of the London Chapter, Ontario Archaeological Society.

Hart, J. P.

1999 Maize Agriculture Evolution in the Eastern Woodlands of North America: A Darwinian Perspective. Journal of Archaeological Method and Theory 6:137-180.

Knight, D. H.

1984 Test Excavations at the Lilac Garden Site. Report on file, Ontario Ministry of Citizenship, Culture & Recreation, Toronto, Ontario.

Lee, G.-A., A. M. Davis, D. G. Smith and J. H. McAndrews

2000 Identification of Fossil Wild Rice (Zizania aquatica) Pollen at Cootes Paradise, Ontario, Canada. Ms in possession of the authors.

Mason, R. J.

1981 Great Lakes Archaeology. Academic Press, New York.

McAndrews, J. H.

1994 Pollen Diagrams for Southern Ontario Applied to Archaeology. In Great Lakes Archaeology and Paleoecology: Exploring Interdisciplinary Initiatives for the Nineties, edited by R. I. MacDonald, pp. 179-195. Quaternary Sciences Institute, Waterloo, Ontario.

Rowe, J. S.

1972 Forest Regions of Canada 1300. Department of the Environment, Canadian Forestry Service, Publication, Ottawa.

Smith, D. G.

1997a Radiocarbon Dating the Middle to Late Woodland Transition and Earliest Maize in Southern Ontario. Northeast Anthropology 54:37-73.

1997b Recent Investigation of Transitional and Late Woodland Occupations at Cootes Paradise, Ontario. Ontario Archaeology 63:4-16.

1998 The 1997 Erindale College Field School at the Sassafras Point Site (AhGx-3). Ninth Annual Archaeological Report, Ontario:78-82.

2000 Long-Term Transition or Short-Term Migration: Explaining Settlement Shifts at the Middle To Late Woodland Interface in Southern Ontario. Ms. In possession of the author.

In press a Middle Woodland and Initial Late Woodland Cultures in the Lower Great Lakes Region. In The Encyclopedia of Prehistory, edited by P. N. Peregrine and M. Embree. Kluwer.

In press b The 1998 Erindale College Field School at the Sassafras Point Site (AhGx-3). Tenth Annual Archaeological Report, Ontario.

Smith, D. G. and G. W. Crawford

1997 Recent Developments in the Archaeology of the Princess Point Complex in Southern Ontario. Canadian Journal of Archaeology 21:9-32.

Smith, D. G., A. M. Davis, T. L. Ormerod and M. Peros

1998 Prehistoric Native Ecology at Cootes Paradise, Royal Botanical Gardens, Hamilton, Ontario. In The Edge and the Point: Proceedings of the 1997 Leading Edge Conference, edited by S. Carty, R. Murzin, S. Powell and D. Ramsay, pp. 277-281.

Snow, D. R.

1992 L'Augmentation de la Population chez les Groupes Iroquoiens et ses Consequences sur l'Etude de leurs Origines. Recherches Amerindiennes au Quebec 22:5-12.

1995 Migration in Prehistory: The Northern Iroquoian Case. American Antiquity 60(1):59-79.

1996 More on Migration in Prehistory: Accommodating New Evidence in the Northern Iroquoian Case. American Antiquity 61:791-796.

Spence, M. W., R. H. Pihl and C. R. Murphy

1990 Cultural Complexes of the Early and Middle Woodland Periods. In The Archaeology of Southern Ontario to A.D. 1650, edited by C. J. Ellis and N. Ferris, pp. 125-169. Occasional Publication of the London Chapter, Number 5, Ontario Archaeological Society.

Stothers, D. M.

1969 An Archaeological Survey of Dundas Valley, Hamilton, Ontario. Ms on file, McMaster University, Hamilton, Ontario.

1977 The Princess Point Complex. National Museum of Man, Archaeological Survey of Canada, Mercury Series 58.

Vennum, T.

1988 Wild Rice and the Ojibway People. Minnesota Historical Society Press, St. Paul.

Williamson, R. F. and C. M. Watts (editors)

1999 Taming the Taxonomy: Toward a New Understanding of Great Lakes Archaeology. Eastendbooks, Toronto.

Wilson, J.

1994 The Racher Site (AfHi-141): More Evidence Concerning Large Riverine Middle Woodland Sites Along the Middle Thames River Drainage. Kewa 94(4):2-17.

1999 The Pocock Site: Exploring the Early Woodland/Middle Woodland Transition in Southwestern Ontario. Kewa 99(1):2-27.