A Lake through Time: Archaeological and Palaeo-Environmental

Ontario Archaeology No. 93, 201352

A Lake through Time: Archaeological and Palaeo-EnvironmentalInvestigations at Lake Temagami, 1985–1994

Diana L. Gordon

(with contributions by John H. McAndrews and Ian D. Campbell)1

Surveys for precontact and historic sites, excavation of the multi-component Three Pines (CgHa-6) andWitchPoint (CgHa-7) sites, and pollen coring of three bogs provide wide-ranging information on the changingcultural and natural history of Lake Temagami. Increased knowledge of changing palaeo-shorelines modifiesinterpretation of known sites and alters survey methodology. Survey first along and later inland from modernshorelines produced precontact campsites, lithic workshops, vein quartz quarries, and nineteenth- to twentieth-century Teme-Augama Anishnabai traditional use sites. Because of the topographic constraints of LakeTemagami, hunter-gatherers favoured those limited locations with well-drained, flat ground; protection fromcold winds; and ease of shoreline access. Three Pines has shallow, compressed soils that are typical of precontactsites on the Canadian Shield, but a modified Harris Matrix analysis allows insights into the stratigraphicsequence of occupations from the Archaic to Modern periods. Site-significant landscape changes seem to haveaffected spatial patterns of site usage and potentially led to the loss of early components. In contrast to the resultsfrom the Three Pines site, preliminary results from the Witch Point site show deeper deposits, greater artifactdensities, numerous cobble features, and lithic reduction activities. These structural differences relate to variationin site landscape, elevation, seasonality, and cultural usage. LateWoodland ritual behaviour, suggested by a dogburial and red ochre at theWitch Point site, is examined in the context of early post-contact accounts whereinrituals reinforce group alliances.

Introduction

This paper presents a chronological overview ofarchaeological and palaeo-environmental researchconducted at Lake Temagami between 1985 and1994 (Figure 1). Fieldwork involved the 1986excavation of the multi-component Three Pines(CgHa-6) site (Gordon 1986, 1987, 1989, 1990a,1990b, 1990c, 1991a, 1991b), palynological andpalaeo-hydrological research on Lake Temagamiin 1987 and 1991 (Gordon 1989, 1992; Gordonet al. 1992; Gordon and McAndrews 1992), andexcavations at the Witch Point (CgHa-7) site in1993 and 1994 (Gordon 1993, 1994a, 1994b,1995a, 1995b).

What ties this diverse work together? The firsttie is the research narrative: the interweaving of

clues that connect questions arising from oneproject to further investigations in the next. Forexample, a Harris Matrix stratigraphic analysis ofthe Three Pines site indicated that some Archaicperiod occupations were marginally closer to theshore than nineteenth-century ones, suggesting

1 The analysis and writing of this paper were conductedlargely by Gordon, but the paper incorporates fieldinvestigations and analysis by McAndrews in thesections entitled “Mid-Holocene Lake Transgression:Evidence from the Three Pines Bog” and “Vegetationand Climatic History of Lake Temagami” and byMcAndrews and Campbell in the section entitled“Palaeo-hydrological Investigations at Lake Temagami:Implications for Precontact Site Location.”

Diana L. Gordon A Lake through Time 53

changing water levels. This interpretation wasconfirmed by a pollen core from the adjacentThree Pines Bog, which showed a substantial lake

level rise of 4 m over the past 7,500 years. Thisfinding led to the field-testing of a predictivemodel of palaeo-shoreline location, involving the

Figure 1. Northeastern Ontario Drainage Systems and Key Archaeological Sites(Inset: Locations of (A) Lake Temagami Study Area (B) Lac Washadimi, Qué. (C) North Caribou Lake).


coring of three additional bogs as well as someinland archaeological survey. It also led tochoosing a higher-elevation site, the Witch Pointsite, for excavation, to add to the developingsequence of occupation for Lake Temagami.

The second interconnecting factor is the“landscape context/stratigraphic approach”(Gordon 1991a, 1991b), which integrates themethodological work of Edward Harris (1979a)in stratigraphy, Karl Butzer (1982) ingeoarchaeology, and Michael Schiffer (1987) insite formation processes. This approach focuses onthe natural and cultural formation processes thathave interacted to create the archaeological siteand its landscape setting over time andincorporates such geoarchaeological studies asgeology, geomorphology, pedology, andmicrostratigraphy. These broader lines of inquirycomplement formal artifact analyses.

The discussion also integrates ethnographic,historical, and oral history information to aid inthe interpretation of cultural formation processesin the precontact, historic, and modern periods.Gordon’s archaeological research at LacWashadimi in northwestern Québec (Chism1978; Gordon 1980; Tanner 1978a, 1978b) andat North Caribou Lake in northwestern Ontario(Gordon 1985, 1988a, 1988b) was conducted inconjunction with ethnographic andethnoarchaeological studies, working with bothCree and Ojibway families (see Figure 1 inset).The insights gained into the subsistence,locational, and material usage strategies of thesemodern hunter-gatherers have strongly influencedthe overall project approach to the archaeologicalrecord of Lake Temagami. In 1994, former Chiefand Teme-Augama Anishnabai Elder William (Bill)Twain of Bear Island showed Gordon manytraditional use sites, sharing his knowledge of thehistory of the “Deep Water by the Shore People”.This information has helped interpret siteselection criteria based on the landscape contextof surveyed sites and elucidate differences inseasonal use between the Witch Point and ThreePines sites. It has also suggested possibleinterpretations of social factors underlyingevidence for Late Woodland ritual behaviour atthe Witch Point site.

These different projects together reveal aresearch trajectory, an expanding focus of inquiryin space and time. Here a visual analogy is useful,one of ever-broadening, concentric rings from anartifact in its stratigraphic context (Three PinesHarris Matrix analysis); to the archaeological site,first in its local landscape context (Three Pinesbaymouth bar and bog); then in its regionalenvironmental context (palaeohydrology of LakeTemagami); and, finally, hunter-gathererbehaviour in the context of the human ecosystem(site selection criteria; Late Woodland ritualevidence and alliance formation).

Natural and Cultural Setting

Lake Temagami is a deep lake with good fishingand a diversity of shoreline features, and it wouldhave been attractive to precontact and historichunter-gatherers. It is connected to three drainagesystems and to north–south travel routes (Figures1 and 2). From Diamond Lake to the north, oneenters the Montreal River drainage into LakeTimiskaming and the Ottawa River system. Onecan also reach Lake Timiskaming eastward throughthe Matabichuan system. Through Cross Lake tothe south and the Temagami River, there is accessto the Sturgeon River drainage, Lake Nipissing, theFrench River, and Lake Huron. Several of theseroutes were well known in the early post-contactand fur trade periods (Heidenreich 1971; Mitchell1977). The height of land marking the drainagedivide between the Great Lakes and Hudson Bay is130 km to the north.

Lake Temagami’s shoreline is relativelyundeveloped in terms of mining, hydro-electricgeneration, and construction. It is mostly Crownland with restricted road access. Cottage buildingis confined to the 1,200 islands. During thisfieldwork, development was slowed by legalcautions on land titles filed in 1973 by the Bearlsland Foundation (Court Judgments 1984; Hall1990; Potts 1989). Land claims had also generatedmuch local interest in heritage matters andsupport of archaeological and ethnoarchaeologicalresearch. While archaeological resources had beendocumented (Boyle 1900; Conway and Conway1989; Dewdney and Kidd 1967), there was, andstill is, great potential for in-depth exploration.


Figure 2. Lake Temagami - Key Locations (Note: only the largest of the 1,200 islands are shown).


Bedrock GeologyBedrock and surficial geology define theconfiguration of the lake and its varied landforms.Table 1 presents a lithologic column for LakeTemagami (Gordon 1990a) and possible primaryand secondary sources of lithic raw materials usedby precontact tool-makers, identified bysedimentary geologist Geof H. Burbidge (1988).

Lake Temagami lies at the junction of all threestructural provinces of the Precambrian CanadianShield: the Superior, Southern, and Grenvilleprovinces (Burbidge 1988; Hewitt and Freeman

1978; Simony 1964). Oldest are the SuperiorProvince rocks (Archean >2.5 bya) formed in a“marine environment with volcanism andsedimentary activities” (Hewittt and Freeman1978:75). In the Temagami area, SuperiorProvince rocks comprise mostly metamorphosedvolcanics.

In the Early Proterozoic (2.5 bya), Huronianrocks of the Southern Province formed. Thesesedimentary rocks underwent repeated folding,faulting, and metamorphosis (Hewitt andFreeman 1978). Two of the four Huronian

Table 1. Lithologic units, Lake Temagami area (after Gordon 1990a). Numbers in parentheses refer to lithic raw materialsources for tools recovered from the Three Pines site. (1) light grey, fossiliferous “local” chert nodules in glacial deposits; (2)well-rounded cobbles of hard granitoid and sandstone, pottery temper; (3) mafic gneiss; (4) vein quartz, pottery temper;(5) metamorphosed olive-grey quartz siltstone; and (6) mudrocks (mudstones, mudshale, claystones, clayshales), wackesand low metamorphic grade sandstone.

PhanerozoicCenozoic

Palaeozoic

PrecambrianLate Proterozoic

Early Proterozoic

Late Archean

Quaternary PeriodGlacial deposits: unconsolidatedsand and gravel (1, 2)-unconformity-

(Lake Timiskaming)Temiscaming Outlier: limestone,dolostone (1)Silurian Period

Grenville Province: gneisses,metagranitoid rocks (3)-Grenville Front-

Nipissing quartz diabase (4)-intrusive contact-Southern ProvinceHuronian Supergroup

-unconformity-

Superior Province: metagranitoid,metavolcanic, metasedimentaryrocks

Thornloe Formation (1)

Cobalt GroupBar River Formation (5)Gordon Lake Formation (5)Lorrain Formation (2, 5)Gowganda Formation (6)

Firstbrook MemberColeman Member


Supergroup formations outcrop in and near LakeTemagami. These are the Lorrain and GowgandaFormations. The slightly metamorphosedsedimentary rocks (mudrocks, sandstones, slates,and conglomerates) of the Gowganda Formationcomprise more than half of Lake Temagami’sbedrock exposures (Burbidge1988:1).

Around 2.2 billion years ago, molten igneousrock from the mantle intruded into the Superiorand Southern province, creating NipissingDiabase dikes and sills (Hewitt and Freeman1978).

In Lake Temagami, Nipissing Diabase,which occurs in the form of dikes ofvarious sizes up to several hundredmetres is the second most exposedbedrock type after the GowgandaFormation. Because the NipissingDiabase is relatively more resistant toerosion than the softer rocks of theGowganda Formation, it forms most ofthe high land, ridges, lookouts and steepwaterside cliffs [Burbidge 1988:2].

The highest peak, Mount Ferguson, at 582 m asl,is an example of a dike, as is High Rock Island(Simony 1964). Nipissing Diabase includes veinsof milky quartz, several of which are found on theeast-central and west-central mainland of the lake(Simony 1964).

About 1 billion years ago, in the LateProterozoic, the Grenville Province formed. Theuplifted Grenville Front, which marks theboundary with the Southern province, lies justsouth of Lake Temagami (Hewitt and Freeman1978). “Composed primarily of hard competentgneissic and granitic rocks, the Grenville provincerocks do not outcrop in Lake Temagami but arefound in the lakes and waterways to the south,such as the Temagami River” (Burbidge 1988:2).

At the head of Lake Timiskaming, 70 km tothe northeast, is the Temiscaming Outlier (seeFigure 1) of Palaeozoic (440 mya) sedimentarylimestones, dolomites, and shale rock that underliethe Hudson Bay Lowlands and the Clay Beltnorth of Lake Temagami (Lovell 1977; Lovell andCaine 1970; OMNR 1978). This outlier containsa grey-and-tan-coloured, soft fossiliferous chert in

its Silurian Thornloe Formation (Hewitt andFreeman 1978). Glacial ice moved south–southwest in this area and deposited material fromthe Temiscaming Outlier onto the Precambrianbedrock. Large pockets of carbonate-rich till arefound around the Northeast Arm of LakeTemagami and in the lower Montreal River(Veillette 1989). “Pebbles and cobbles of a porous,light coloured fossiliferous chert are quitecommon on beaches in Lake Temagami […] andmay come from this Temiscaming outlier”(Burbidge 1988:2).

Surficial GeologyIn contrast to the Little and Great Clay Beltsfarther north, Lake Temagami is on the uplandCanadian Shield, characterized as “moderatelyrolling thin till over bedrock with local morainicdeposits and outwash” (Boissonneau 1968:98).The continental ice sheet eroded the PrecambrianShield, leaving glacial tills and moraines(Boissonneau 1968; OMNR 1978). As the iceretreated from Lake Temagami about 12,000 yearsago (Veillette 1988), meltwaters reworked thesematerials. Glaciofluvial deposits, such as outwashplains and eskers, consisting of well-sorted sandand gravel, cover a relatively small proportion ofthe region (OMNR 1978:14). According toSimony (1964:17) north–south trending eskers,which resemble tributaries of south-flowingstreams, are the most important Quaternarydeposits on the lake. The 1985 aerialreconnaissance of Lake Temagami confirmed thatthe distribution of sandy soils is limited. Thebroad outwash plain at Ferguson Bay at the northend, the Sand Point peninsula on the west-centralmainland, the Witch Bay esker on the east-centralmainland, and sandy deposits on islands, such asBear Island and Temagami Island, are unusualfeatures along otherwise rocky shorelines (Gordon1990a).

Modern DrainageLake Temagami is a narrow lake, oriented north–south (Figure 2). Its five long arms, 20 km each,radiate out from the locally named central “Hub.”The mainland shoreline measures 592 km, and thelake’s 1,200 islands add another 320 km (OMNR


1973). Total water area is 128 km2, with amaximum depth of 110 m (OMNR 1969).Interestingly, Lake Temagami drains in twodirections. Before dam-building in the nineteenthcentury (Hodgins and Benidickson 1989), rapidsat Sharp Rock Inlet provided a northern outlet.Water flowed into the Diamond Lake–LadyEvelyn Lake–Montreal River–Lake Timiskaming–Ottawa River drainage. The major modern outletis south, through Outlet Bay and Cross Lake intothe Sturgeon River–Lake Nipissing–Great Lakesdrainage. The lake is 293 m asl, which can beraised to a maximum of 294.2 m asl by threemodern control gates at the south end of CrossLake (OMNR 1973; Randy Plant, personalcommunication 1988). During the 1991 survey,shoreline flooding was evident at the south end,with snags occurring in Cross Lake. Remnants ofa 1900s dam (1.2 m high) were noted at theentrance to Outlet Bay (Gordon and McAndrews1992).

VegetationLake Temagami lies 50 km south of the northernboundary between the Great Lakes–St. LawrenceMixed Forest and the Boreal Forest (Liu 1990;OMNR 1978). Classified as the “TimagamiForest Region,” the predominant species are whitepine, red pine, scattered white birch, and whitespruce. Also common is a mix of birch, pine, andspruce with balsam fir and aspen. In certain locallyprotected areas, there is a scattering of hard maple,red maple, and yellow birch (OMNR 1973).While a skyline forest reserve has been maintainedaround the perimeter of Lake Temagami, theinterior forest has been logged extensively.

Modern Cultural SettingSince the cessation of mining and logging, LakeTemagami is primarily used for recreation:canoeing, boating, fishing, numerous designatedcampsites, portages, year-round fishing lodges, aprovincial park, and summer youth camps(Hodgins and Benidickson 1989). The small townof Temagami is located where Highway 11, theOntario Northland Railway and the TransCanadaPipeline pass the Northeast Arm (Figure 2).

The Bear Island Reserve is located in the

Central Hub. Established as a reserve in 1971, itis home to many Teme-Augama Anishnabai. Anolder settlement, the Austin Bay site (CfHa-28),once occupied the South Arm (Pollock 1992;Gordon 1995b). Bill Twain identified the peopleas Algonquin, speaking a slightly different dialectthan is spoken on reserves on Lake Nipissing andLake Huron. He said that the people used to live“all over the lake,” travelling extensively to suchplaces as Mattagami and Gogama to the north,Lake Nipissing to the south, and beyond (Gordon1995a:11). Craig Macdonald’s (1993) HistoricalMap of Temagami, based on interviews with manyolder Teme-Augama Anishnabai, identifies a vastnetwork of traditional place names, landmarks,and seasonal travel routes (Jenish 2006). In 1913,anthropologist Frank Speck (1915a, 1915b)interviewed Bear Island residents for hisethnographic monographs on family huntingterritories in the Temagami-Temiskaming area(Feit 1993). Through genealogies, oral histories,and historical documents (Potts 1989), some ofwhich were compiled for the Bear IslandFoundation court cases (Court Judgments 1984,1991, 1999; Hall 1990; McNab 1999, 2009;McNeil 1990), the Teme-Augama Anishnabai haveshown a long-time connection to the lake andsurrounding area, which they refer to as n’DakiMenan—Our Land.

Previous Archaeological Research

At the time of my fieldwork, in the vast areabetween Lake Nipissing and Lake Abitibi, only adozen precontact site excavations had been fullyreported (Figure 1). Researchers focused on thelargest lakes—Nipissing, Timiskaming, andAbitibi—following Frank Ridley’s pioneeringsurvey and excavation work between 1948 and1962. Major, documented excavations include theFrank Bay site (CbGw-1) on Lake Nipissing(Brizinski 1980; Brizinski and Savage 1983;Ridley 1954) and the Montreal River site (CgGu-1) on Lake Timiskaming (Knight 1977; Ridley1956, 1957, 1966). Ridley also excavated severalsites on the western half of Lake Abitibi, includingthe Ghost River Garden and Abitibi Narrows sites(Ridley 1956, 1958, 1963, 1966). This large lake


in the Great Clay Belt straddles the Ontario–Québec border. It is known for its quarry sites,such as the Jessup site (DdGw-2) (Kritsch-Armstrong 1982), associated with the Mt.Goldsmith quarry (Pollock 1984), and a largehabitation site in Québec, the Joseph Bérubé site(DdGt-5) (Marois and Gauthier 1989). Pollock’s(1975, 1976) excavations on several smaller lakesincluded the Smoothwater Lake site (CiHd-1), theDuncan Lake site (CiHf-2), and the Pearl Beachsite (DaGv-1) on Larder Lake (Noble 1982;Pollock 1976). Hanks (1988) dug the Foxie Ottersite (CdHk-3) on Fox Lake in the Spanish Riverdrainage to the southwest.

Questions of cultural identification andcultural chronology through descriptions ofartifact variability characterize much of this work(Brizinski 1980; Hanks 1988; Knight 1977;Pollock 1976). Syntheses of the Shield Archaictradition (Wright 1972a), of archaeologicalsequences for northeastern Ontario (Pollock 1975;Ridley 1966), and of Larder Lake (Noble 1982)have been presented. The authors themselvescaution that these are based on limited data andare subject to revision (Noble 1982; Pollock 1975,1976, 1984; Wright 1972a, 1972b, 1979).

The first published finds from the LakeTemagami area include reports of a weatheredknife, a slate axe, and a birch bark canoe (Boyle1900:6, 1904:63, 1905:27), as well as rockpainting sites (Dewdney and Kidd 1967:92-93;Phillips 1907). Between 1975 and 1985, surveyswere conducted by the provincial government’sregional archaeologist for northeastern Ontario,Thor Conway. Based on archaeological site recordsmaintained by the Ontario Ministry of Tourism,Culture and Sport, Conway registered 48 sites: 22rock paintings, 13 precontact lithic sites, and 13sites dated to the nineteenth to twentieth centurysites. These latter sites include the Hudson’s BayCompany post on Bear Island, burial grounds onTemagami Island and High Island, historic ormodern camps, a sugarbush site, religious sites,and an area of defensive pits used against theIroquois. Additional rock paintings, quarry sites,and habitation sites were recorded on ObabikaLake just west of Lake Temagami (Conway 1984;Conway and Conway 1989:35). Short-term

salvage work was done at the Sand Point site(CgHa-1) in 1978 and 1981 (Conway 1986).From salvage excavations at theWitch Point site in1982, Conway identified five components,including Middle Archaic, Late Archaic, MiddleWoodland, Late Woodland, and Historic periodoccupations (Conway 1982; Smith 1983).Subsequent fieldwork has included the excavationof the Lake Temagami site (CgHa-2) at FergusonBay (Carscallen 1994a, 1994b), heritage resourcesurveys for the Teme-Augama Anishnabai (Pollock1992; Pollock and Koistinen 1993) and rock artstudies (see Zawadska, this volume).

Fieldwork at Lake Temagami, 1985–1994

In July 1985 a preliminary boat survey of theCentral Hub and a Cessna overflight of the entirelake were conducted (Figure 2). The Three Pinessite on the west-central mainland was excavatedover eight weeks in July and August 1986. Six newsites, comprising lithic scatters, a vein quartzquarry, and an associated workshop site, were alsorecorded (Gordon 1986, 1987, 1990a, 1990b,1990c, 1991a).

In June 1987, McAndrews and Gordonextracted a pollen core from the Three Pines Bog.In August of 1987, Gordon collected off-site soilsamples at Three Pines, which were analyzed byBurbidge (1988), a sedimentary geologist alsoconducting doctoral research on Lake Temagami.A short survey at Cross Bay added two new lithicsites (Gordon 1989, 1990a, 1990b, 1991a).

To test palaeo-hydrological reconstructions byCampbell (Gordon et al. 1992), the 1991fieldwork focused on two predicted palaeo-outlets.In July, a pollen core was extracted from Bear Bognear Sharp Rock Inlet to the north. At this time,minor archaeological survey and a brief visit to thesouthern Outlet Bay were also conducted. InAugust, short interior surveys were undertaken atthe north end of the lake (Barnac Lake andFerguson Bay) and, on contract, along theTransCanada Pipeline in the town of Temagami.In November, two more pollen cores wereobtained from Baseball Bog and Jessie Fen at theend of the Northeast Arm (Gordon 1992; Gordonet al. 1992; Gordon and McAndrews 1992).


Excavation of the large Witch Point site on theeast-central mainland occupied five weeks in 1993(Gordon 1993, 1994a, 1994b) and seven weeksin 1994 (Gordon 1995a, 1995b). With theguidance of Bill Twain, rock painting sites, quartzveins, the old Austin Bay settlement, nineteenth-century cemeteries, and a nineteenth-centurytrading outpost were further documented(Gordon 1993, 1994b, 1995a, 1995b).

The Three Pines Site: Applying a HarrisMatrix Stratigraphic Analysis

Excavation and analysis of the Three Pines site wasundertaken as doctoral research at McMasterUniversity (Gordon 1990a, 1991a). At that time,there was no example of a Harris Matrixstratigraphic analysis (Harris 1979a) of a shallow,northern forest, precontact site. According to JohnTriggs (personal communication 2012) it hassubsequently been adopted as a standardexcavation and analysis technique by ParksCanada, the Ontario Heritage Trust, and WilfridLaurier University Field School. The Harris

Matrix is now used worldwide, often combinedwith computer-based analyses (e.g., Harris et al.1993). However it is still not well known as ananalytical technique for precontact Ontario sites.For this reason, the methodology is explained indetail. Natural versus cultural modifications to thesoils are distinguished through a comparison ofon-site and off-site soil samples. This comparisonis followed by a step-by-step description of howthe Harris Matrix was specifically applied to ThreePines site, with full examples of 3 of the 24stratigraphic columns. A sequence of occupationsis described “top down” as the site itself isencountered, from the modern present to theprecontact past. Evidence for lithic raw materialchange between the Archaic and Woodlandperiods is also summarized.

Modern Landscape SettingThe Three Pines site is located on the west-centralmainland, near the entrance to the NorthwestArm (Figures 2–5). Here a broad peninsula,composed of sand and gravel to the north andsand to the south, juts into the lake (Simony

Figure 3. Locations of the Three Pines site (CgHa-6), Three Pines Bog and Sand Point site (CgHa-1), west-CentralHub of Lake Temagami.


Figure 4. Three Pines site Test Pits, Elevations and Excavation Grid.

Figure 5. Three Pines site – View from the lake north-west to lower terrace and excavation on upper terrace,showing west and central erosional gullies.

1964). The site is found at the southwest corner ofthe peninsula. It comprises two flat sand terraces,separated by a 0.5 m erosional face (Figures 4 and5). The lower terrace extends from the sandybeach to the erosional face. The upper terraceslopes from 1.9 m (above the lake) at the erosionalface northward to 3.6 m, then it drops sharply.Immediately west of the site, the land rises to thewest sand hill (5 m above the lake). To the east isa low-lying sphagnum peat bog. Farther east is ahigh bedrock dome, then a triangular sand spit,on which is situated the Sand Point site (CgHa-1).A 5 m wide sand beach stretches the 288 m fromthe west hill to the rock dome, where it becomesa narrow cobble strand (see below, section titledBaymouth Bar Formation).

Upper Terrace DescriptionInitial test pits on the west hill and eastern beachberm were negative. However, surface finds oflithic flakes were visible on the upper terrace.Along the lake side of the upper terrace is a grass-covered clearing measuring 25 m by 12 m. Threeerosional gullies, used as pathways from the beach,


cut into the upper terrace, while shrubs havestabilized other sections. The clearing is encircledby three 150-year-old white pine trees, which havebeen scorched by forest fire, along with shrubs,conifers, and hardwoods. Marked as a “Day UseOnly” recreational area, the site is used for picnics,swimming, boating, and occasional overnightcamping. The first impression was that thesemodern activities and natural erosion may havedestroyed precontact traces. However, beneath thegrass and coarser sand of the grey-brown“overburden” were discrete areas of compactedgrey ashy sand and/or fire-reddened sand, incontrast to the looser texture of the yellow-brownsand substrate. These were labelled as featuresbecause they contained high concentrations ofprecontact artifacts (Figure 6).

Excavation MethodsOn shallow sites where I had worked previously,one metre square units were employed forhorizontal control and natural soil horizons for

vertical control (Gordon 1980, 1985). Whereas,researchers in northeastern Ontario have used 1.5m units (5 ft squares) for horizontal control andused arbitrary levels ranging from 4 cm to 12 cmdeep for vertical control (Brizinski 1980; Knight1977; Pollock 1976; Ridley 1950-53). For theThree Pines site, excavation procedures andstandardized recording forms followed thosedeveloped by C.S. “Paddy” Reid for the OntarioHeritage Branch in northwestern Ontario (e.g.,Hamilton 1981), a methodology learned at theForestry Point (EgKl-1) site (Pelleck 1983). Theyare described in detail, as this precise recordingmade it possible to later conduct a formal HarrisMatrix stratigraphic analysis of the Three Pinessite.

The 1 m squares were subdivided intoquadrants. Where specific features were designatedin the field, quadrants were further subdivided asinside or outside the feature. Units were openedin a checkerboard fashion, allowing for verticalprofiles of features bisected by grid lines.

Figure 6. Three Pines site - Distribution of Field-designated Features (Detailed analyses of F15N/17, F7 and F4 intext).


Additional profiles were taken at regular intervals.Vertical control was in 3 cm increments below

surface (Level I: 0–3 cm; Level II: 3–6 cm). These“levels” were not analytical units; they were simplya method to plot and bag recoveries by square,quadrant, feature, and 3 cm level. Standard squareplans were used for each level. Soil colour, soiltexture, and fire-cracked rocks were mapped andphotographed at the surface of each 3 cm level.Any soil changes within the level were indicatedin the square plan notes. All lithics, ceramics,faunal bone, and other finds were individuallyplotted using a standard legend. Spot depth belowsurface was measured for each lithic tool andceramic rim using a string and line level. Allexcavated soil was screened through ¼" (6 mm)mesh. Screen recoveries were counted andrecorded in the notes.

Excavation units were dug down until onecompletely sterile level was removed, usually 12–15 cm deep. To confirm site depth, three test pitswere dug within the main block to a depth of twometres. All three were sterile below 15 cm, with asandy C horizon (parent material) at 40 cm deep;no buried soil horizons, bedrock, or ground waterwere encountered.

Table 2. Three Pines site distribution of artifacts bysubcategory.

Faunal bone

Lithics

Pottery

Modern

Historic

Total

calcinednon-calcined

debitagetools

sherdlets/exfoliatebody sherdsrims

metalglassplastic/otherceramic

clay ball pipesgunspalls/gunflintsdecorative metal

f.(6,584)6,5813(6,095)5,856239(1,377)1,02732624(956)5202701579(22)1732(15,034)

%(43.8)

(40.5)

(9.2)

(6.4)

(0.1)

(100.0)

Table 3. Three Pines site distribution of faunal boneidentifications (after Prevec 1987).

Clam

Reptile

Bird

Mammal

Class unknownTotal

Unionidae

pond turtle

common loonruffed grouse

beaverwolf or dogblack bearmartenwhite-tailed deermooseunidentified mammal

f.

2

3

12

77857141105,722126,584

Table 4. Three Pines site distribution of lithic toolclasses.

ScraperBifaceRetouched flakePebble netsinkerFlake knifeProjectile pointUtilized flakeLarge utilized flakeCeltUniface edgeHammerstoneBattered itemGraverUtilized cobble/celtAxeAbraderGrinding stoneUtilized pebblePolished pebbleTotal

f.663228191815141397432222111239

%27.613.411.78.07.66.35.95.43.82.91.71.30.80.80.80.80.40.40.4100.0

Over eight weeks, a total of 82.5 excavationunits were dug on the upper terrace, as well as 47test pits (0.5 m x 0.5 m) around the main block.Including other test pits at the west hill (13),


beach berm (10), and lower terrace (2), a total areaof 101.5 m2 was opened. A total of 15,034cultural items and 543 rock, soil, or organicsamples were recovered (Tables 2–9).

Natural and Culturally Modified SoilsTo determine cultural modification processes andnatural soil development, 21 samples from avariety of “on-site” upper terrace contexts (Table10) were microscopically examined and comparedwith an “off-site” soil profile and sediments fromthe west hill and lower terrace (Tables 11 and 12;Figure 7). A standard hummo-ferric podzol(Canadian Soil Survey Committee 1978) hasdeveloped on glacial outwash, which is the parentmaterial for the west hill (Simony 1964). Thistype of soil is common in the mixed coniferousand deciduous forests of northeastern Ontario(Liu 1990). The light grey Ae horizon develops asdark organic material is leached out. The yellowishbrown Bf horizon contains deposits of aluminiumand iron brought down in solution by acidicrainwater (Birkeland 1984).

On-site soils vary greatly in distribution andthickness (Figures 8 and 9). The only constantsare a surface and a Bf horizon. Podzolic soil

Table 5. Three Pines site distribution of lithic tools byraw material.

Hudson Bay Lowland chertMudrocksChertLocal chertVein quartzSandstoneSiltstoneQuartziteWackeClear quartzGranitoidNipissing diabaseGneissTotal

f.70454420161411554311239

%29.318.818.48.46.75.84.62.12.11.71.30.40.4100.0

Table 6. Three Pines site distribution of lithic debitageby raw material.

Conchoidal fracture(3,724)

Sub-conchoidal fracture(2,132)

Total

f.

Hudson BayLowland chertLocal and other“cherts”1

Vein quartzClear quartz

Miscellaneous2

Mudrocks

%(63.6)

31.8

15.6

12.63.6(36.4)

26.89.6100.0

1 includes “local” cherts, indeterminate cherts, “chert-like”fine-grained siltstone, and fine-grained sandstone2 sample identified by Burbidge contains mudrock, medium-grained sandstone, non–conchoidal fracture wacke, gneiss,granitoid

Table 8. Three Pines site distribution of analyzed pottery.

Body sherds(326)

Rims (24)Total

f.

141138291824350

%

40.339.48.35.16.9100.0

decoratedplaincord-markedindistinct decoration

Type of Fracture

Conchoidal fracture

Sub-conchoidal fracture

Total

Raw Material

HBL chertLocal and other chertsVein quartzClear quartz

MiscellaneousMudrock

UnmodifiedFlakes

1,31876536397

989458

3,990 (68.2%)

Shatter

25787365113

50883

1,413 (24.1%)

DecorticationFlakes

2644530

7020

402 (6.8%)

Core/CoreFragments

221591

22

51 (0.9%)

Table 7. Three Pines site distribution of lithic debitage by fracture category and raw material.


Rim vessels (13)

Body vessels (4)

f.

8221

1111

Design

Pseudo–scallop shellLinear stampCord-wrapped stickDentate stamp

Dentate stampPlain baseCord-wrapped surfaceLinear/dentate stamp

Vessel Identifiers

A, B, C, E, F, J, L, QD, I,G, HK

MN0P

Table 9. Three Pines site distribution of rim and body vessels.

Table 10. Three Pines site selected upper terrace soil samples.

ArchaeologicalDescriptionLitter and grassrootsGrey-brown over-burden

Dark browndecomposingorganicsDark grey sandwith charcoal

Grey ashy hard-packed sand

Bright orange, fire-reddened sandwith calcinedfaunal boneFire-reddened sand

SampleIdentifier

Sample#19

Sample#14

Sample #1

Sample#16

Sample #2

Sample#18

Context

S3E13/LI

FA8/11-Unit 4/LIII

FA12-Unit 2/LIII

FA8/11-Unit 5/LIV

FA6/LII

FA16-Unit 9/LII

Munsell Colour

Dark GrayishBrown 10YR 4/2

Dark Gray10YR 4/1

Pinkish Gray7.5YR 7/2

Gray 10YR 5/1

Yellowish Red5YR 5/8

Strong Brown7.5YR 5/6

Soil Micromorphology

L horizon

poorly sorted, medium sand;moderate rock fragments 10–20%; high organic content 40%

F and H horizons

poorly sorted, medium sand;sub-rounded; rock fragments 15–20%; organics 5–10%poorly sorted, medium sand;sub-rounded; trace rock frag-ments; clay/ash coatingpoorly sorted, fine sand;sub-angular; rock fragments 10%;organics 15%poorly sorted, fine sand; sub-rounded; rock fragments <5%;lots of haematite stain

poorly sorted, fine sand;rounded; coated with iron oxide

Figure 7. Three Pines site - West Sand Hill and LowerTerrace Soil Profiles.

horizons occur on the upper terrace, but withsignificant differences. These culturally modifiedsoils include:

1) a layer of grey-brown coarser sandoverburden with grass roots,2) localized areas of hard-packed finegrey sand with and without ash, and3) localized reddish soils with andwithout ash, with and without faunalbone.


The grey-brown “overburden” is attributed toseveral processes: a) wind-borne beach sanddeposited on the upper terrace during strongprevailing southwest winds and b) beach sandtracked up by humans or transported to smothermodern picnic fires, such as the large fire moundin the centre of the excavation block. These typesof processes would result in the burying ofunderlying archaeological sediments. In thepathways and central area, the overburden resultsfrom c) human scuffing and d) natural erosion.These processes removed upper organic andmineral horizons. All these interacting processespresumably took place in the past.

Dark brown decomposing organics (F) anddark grey humus (H) horizons are found together

in the northwest section of the excavation. Thesenatural podzolic soil horizons are associated withhistoric period artifacts. F horizons are generallyabsent in other areas of the excavation.

Discontinuous areas of hard-packed grey soilsare not simply the light colour of the leached Aesoil horizon (Fenwick 1968). These fine sandsreveal the creation and deposit of ash from high-temperature burning of organic matter, such aswood and faunal bone (Butzer 1982:82). Whenthey are viewed microscopically, it is clear that ashand/or silt coats each sand grain. This grey to lightgrey sand and some red sands tend to have a highash content. When trowelled, this compacted soilmade a rasping sound, as if the sand grains werecemented together, while adjacent soil had a looser

Thickness (cm)2.5–1.51.5–0

0–0.5

0.5–4.5

4.5–24.5

24.5–34.5+

Soil HorizonL horizonF and H horizons

Ah horizon, weaklydevelopedAe horizon; eluvial orleached horizonBf horizon

C horizon

Soil Micromorphologyforest litter: pine needles, seeds, mosspartially decomposing organic matter;95% organicsmixed organic and mineral; poorlysorted, fine sandpoorly sorted, fine sand

iron oxides precipitated out;moderately sorted, fine sandmoderately sorted, fine sand

Munsell Colour

Dark Brown 10YR 2/2

Dark Gray 10YR 4/1

Pale Brown 10YR 6/2

Dark Yellowish Brown10YR 4/6Light Yellow-Brown2.5YR 6/4

Table 11.West hill “off-site” soil profile. Sand grains are sub-angular, indicating glacial outwash as parentmaterial.

Thickness (cm)

10.0–3.0

3.0–0

0–3.0

3.0–6.06.0–15.0+

ArchaeologicalDescription/Soil HorizonLitter, grass and blow sand

Grey-brown overburden

Buried Ah horizon

Ae horizonYellow-brown sand/Bfhorizon


moderately sorted, medium sand;high percentage of rock fragmentsmoderately sorted, medium sand;40% mudstone rock fragments;baymouth bar washoverpoorly sorted, medium sand; 70%organicspoorly sorted, medium sandmoderately sorted, medium sand;hydrated iron oxide gives colour

Munsell Colour

Dark Grayish Brown10YR 4/2Dark Grayish Brown10YR 4/2

Very Dark Brown10YR 2/2Grayish Brown 10YR 5/2Olive Brown 2.5YR 4/4

Table 12. Three Pines site lower terrace soil profile. Sand grains are rounded, indicating reworking by waveaction.


texture. Artifacts move less in this hard-packedmatrix. Whenever hard-packed grey ashy soilswere encountered, precontact artifacts were alsorecovered, and these soils are interpreted as hearthand living floor remnants.

Localized areas of reddened soil are sometimes,but not always, associated with areas of hard-packed grey ashy soils. These are not the natural,iron-rich Bf horizons (Fenwick 1968). They areredder in hue, are irregularly distributed, and oftencontain calcined animal bone. Under extremeheat, hydrated iron oxide naturally occurring inthe soils may become dehydrated, creating areddened appearance. These localized fire-reddened soils are considered the best evidence forhearth remnants at the Three Pines site. Thosethat contained calcined faunal bone fragmentsmay represent the edges of former cooking and“disposal hearths” rather than the entire heartharea. The term “disposal hearth” comes fromobservations at Lac Washadimi, where Creefamilies cleaned up all animal bone and disposedof it in hearth fires, as a sign of respect for theanimals (Gordon 1980; Tanner 1979). In theintense heat at the hearth centre, faunal bone isreduced to ash, while fragments are best preservedat the cooler edges. Camp fire experiments in1986 tested and confirmed this hypothesis.

Figure 8. Three Pines site – Selected Soil Profile showing Field Feature 10 and west erosional gully.

Figure 9. Three Pines site - Upper Terrace CompositeSoil Column showing cultural modifications of naturalpodzolic sandy soils.


Natural and Culturally Modified SedimentsFor site formation processes, Butzer (1982)distinguishes among physiogenic, biogenic, andanthropogenic agents of archaeological sediments.The latter includes minerals and organics broughtonto the site by its inhabitants, “residues […]derived from alteration of human imports orhuman activity” and “minerals sediments due toacceleration of natural processes by humanactivity” (Butzer 1982:78). Along with wood ash,fire-reddened sand, and calcined faunal bone,noted above, the Three Pines site has other“sediments” brought onto the site, modified orused by precontact hunter-gatherers. Theseinclude lithic raw material, lithic knappingdebitage, finished and unfinished tools, pottery,pottery temper, cobbles, rocks, and organicmatter. The fire-cracked rock and rounded cobblesare “out-sized clasts” (Burbidge 1988), meaningthey are too large to occur naturally in this sandymatrix. Additional lithic imports include feldsparfragments, granite, and diabase flakes similar totemper found in various pottery vessels; one smallclay deposit probably introduced for potterymaking; and some red ochre, or haematitefragments.

There are also naturally produced sedimentsfrom “physiogenic agents”: lenses of wind-blownsand, terrace edge erosion, tree root growth,charred wood from forest fires, root burning, rootdecay, and tiny rootlets dropping dark humus intounderlying soils. Ground squirrels are a major“biogenic agent,” leaving visible burrows of mixedlight yellow sand, often without artifacts.

Developing a Chronological SequenceThe next question was how to organize thesenatural and culturally modified soils andarchaeological deposits into a meaningfulchronological sequence. I first considered artifactvariability and regional artifact comparison andexamined collections from a large number of siteslocated on surrounding lakes, such as Abitibi,Nipissing, and Timiskaming (e.g., Gordon 1989;Knight 1977; Marois and Gauthier 1989; Ridley1950-53, 1958, 1966; see Figure 1). Only a few ofthe Three Pines lithic tools and decorated potteryrims matched artifacts from these distant sites in

terms of size, morphology, raw material, andmanufacture. Artifact variability and inter-sitecomparison alone could therefore not offer a solidfoundation on which to build a sequence ofoccupations.

Compiled site excavation maps for the ThreePines site showed that site-wide level by level,vertical comparisons were also problematic. Nonatural soil horizons or culturally modified soilsor sediments were continuous across the site,except for the grassy surface vegetation and theyellow-brown Bf horizon. Artifact-bearing featureswere discrete. Modern disturbances werenumerous. Human trampling and erosion hadexposed subsurface layers in certain parts of theexcavation. These factors complicated attempts todetermine a chronological sequence by excavationlevel and depth.

Therefore, a new approach was undertakenusing the Harris Matrix (Harris 1979a). At firstglance, the Harris Matrix seems more suited tosites with clearly observable distinctions betweenarchitectural features and natural layers, becauseHarris (1975, 1977, 1979b) originally formulatedand applied his stratigraphic approach to a UnitedKingdom site with Roman to Mediaevalarchitectural features and occupations. Butexcavators of shell middens in New Brunswick(Black 1991) and a rock shelter site inWashingtonState (Stucki 1993) had also had some successwith this method.

The Harris Matrix method, using histerminology (Harris 1975, 1977, 1979a, 1979b,1989), involves:

1) identifying both human-made andnatural units of stratification;2) determining the stratigraphicrelationship of each unit based on theLaws of Superimposition, OriginalHorizontality, and Original Continuity;and3) translating those relationships into aschematic box diagram called a “HarrisMatrix” by applying the Law ofStratigraphic Succession.

Individual units of stratification, such asunmodified soil layers, architectural structures, or


deposits of objects, are identified and subsequentlyrepresented by a single numbered box. Harris(1977, 1979a) also includes interfaces betweenlayers, for example, the outline of an intrusive pitbecomes a unit of stratification separate from thepit fill itself, as they represent two sequentialevents. The stratigraphic relationship of a singleunit is illustrated in one of four ways: a unit isabove, is below, equals or correlates to, or has nostratigraphic relationship with other units. Theresulting sequence of stratification is based only onthe physical context of the site, that is,independent of and constructed without referenceto the artifactual contents (Harris 1979a:116).This strategy is used because formation processesare different on each site, such that thestratigraphic sequence of a site is unique andcannot be determined beforehand. It then serves asa baseline for the relative temporal sequence ofevents. Only after the Harris Matrix is constructedare the artifacts incorporated into each unit. Thestratigraphic sequence is then phased into differentoccupations by making artifact and absolutedating correlations between units within the siteand ultimately with other sites (Harris 1975,1977, 1979a, 1979b, 1989; Harris et al. 1993).

Application of the Harris Matrix to the ThreePines SiteTrying to build a single stratigraphic sequence bystarting at one corner of the excavation rapidlyproved too unwieldy. It was decided to subdividethe site into areas around each feature. These“field-numbered features” are excavation units, sodesignated because they were physicallydistinctive, discrete areas of deposition, separatedby the yellow-brown substrate (see Figure 6).“Feature Area” (e.g., FA7) is an analytical term. Itrefers to the maximum area around each field-designated feature in which units of stratificationcould be identified, correlated, and presentedwithout getting too complicated. In this way amulti-linear series of stratigraphic columns wasconstructed initially.

Step 1:Unit Identification. For each Feature Area,the analysis started with one soil profile, in whicheach unit of stratification (e.g., erosional slopes,

humus-filled pits, cobble deposits, fire-reddenedsoil etc.) was described and numbered. Adjacentprofiles and square plans were then reviewed to seewhether they contained the same units. If they didnot, new units were added where necessary. Unitidentification was assisted by additionalinformation from field notes and, in particular,photographs.

All units including “disturbances,” weretreated equally as integral parts of the cumulativetemporal record of the site. Although Harris(1977, 1979a) defines layers, features, andinterfaces—all quite useful distinctions onarchitectural sites—I simply used the generic term“units of stratification” for this shallow, sandymatrix. Numbering was also simplified. Forexample, pit interfaces were not used in thenumbering of units, but the analysis was mindfulof their role in determining spatial relationships.

Step 2: Unit Spatial Relationships. The next stepwas to determine the four spatial relationships ofeach unit to the others in the Feature Area. Doesa unit lie above, lie below, equal or correlate to, ornot stratigraphically correlate to the other units?

Step 3: Feature Area Stratigraphic Column. Thisstep created a series of stratigraphic sequences,which are combined into one master HarrisMatrix diagram for each Feature Area byeliminating clearly redundant relationships.

Step 4: Interpreting Events. Recurrent patternsbegan to emerge, making it possible to interpretwhat event or events may be represented by eachunit of stratification (e.g., ground squirrelburrows, tent drainage ditches, post moulds,intrusive pits). Human-made pits and post mouldswere more distinctive in the upper levels. Anintrusive pit was identified as two events: theexcavation and infilling of the pit.

With shallow soils and the probability ofmultiple short-term occupations, I did not assumea priori that any one unit of stratification could beequated with any single event, such as an “activityarea,” or with any one cultural or temporal periodof deposition. After years of weathering, anytextural distinctions between multiple periods of


deposition or events within one unit ofstratification may no longer be visible to the nakedeye. Also, the distinction between lower-levelsediments is as much a function of naturalweathering as cultural activities—hence thedecision not to use pit interfaces, as noted above.

Step 5: Unit Artifact Contents. The next step wasto identify the artifact contents of each unit ofstratification, using the plotted artifacts in thesquare plans.

Where a unit of stratification included morethan one 3 cm vertical level, artifacts wereseparated by level. Again the same reasoningapplies as above. A unit of stratification, whichtoday looks undifferentiated, may represent yearsof deposition; hence the artifact contents arepresented by depth. All modern material in thelower units was associated with pits, post holes, orroot holes. Occasionally small lithic flakes andceramic sherdlets were noted inside tiny, humus-filled rootlet holes in the yellow-brown Bf horizon,indicating some vertical movement of these smallitems.

Step 6: Phasing Units. During the examination ofthe contents of each unit, certain artifact patternsbegan to emerge and repeat across the stratigraphiccolumns for each Feature Area. This informationwas used to assign different units to broadtemporal divisions, based on relative stratigraphicposition, mass-manufactured modern or datablehistoric-period items, or the presence of differenttemporal pottery characteristics (e.g., MiddleWoodland Laurel and Late Woodland cord-marked pottery). Archaic was assigned to units inthe lowest stratigraphic units containing lithic rawmaterials and tool types different from those in theupper, ceramic-bearing units.

Step 7: Site Sequence. Where possible the multi-linear series of stratigraphic columns werestratigraphically correlated and phased across thesite. In this way a relative temporal sequence ofoccupations for the Three Pines site wasdeveloped, with settlement features and associatedartifacts separated out stratigraphically.

Absolute dating would be most helpful at this

step. Tiny flecks of charcoal were ubiquitous,especially in the grey ashy units, but to my eye itwas unclear whether they derived from forest fires,tree root burns, cultural activities, and/or rootletdeposits of humus from upper levels. For thisreason, samples were not submitted forradiocarbon dating, as their context wasquestionable.

Stratigraphic Column forFeature Area 15N and 17As examples of the procedures that were followed,three of the 24 stratigraphic columns are describedhere in detail. Field feature 15N is a wide expanseof hard-packed grey ashy sand with large cobbles(see Figure 6 for location in excavated area). Itmeasures 270 cm by 100 cm, ranging in thicknessfrom 3 to 25 cm. The northwest margin istruncated by modern garbage pits and a rodentburrow. Modern hearths obscure the westernmargin, while Tree B’s roots have blurred thesoutheastern margin. Field feature 17 to the east ismottled, dark to light grey with yellow sand,located beneath the tree roots. It is 12 cm thick,measuring 40 by 80 cm in area. For Feature Area15N and 17, 12 units of stratification areidentified, sequenced, interpreted, and phased(Figures 10–11; Tables 13 and 14).2

Unit 1 is the grey-brown overburden, thickerat the north end than the south. Among themodern camping and picnic debris is a sawnmoose vertebra and a .257 rifle shell casing,suggesting moose hunting and butchering. Unit 2represents the excavation and infilling of a moderngarbage pit. It truncates the layers below with anoblique interface, indicative of sharp shovel cuts.Buried up to 40 cm deep are modern boating andfood container trash.

Units 3, 4, and 5 represent soil layers in reverse

2 Units of stratification in plan view may not appearin the profile illustrations, and vice versa. The illustra-tions can show only a portion of all the data used todevelop the sequence. Plan views present the horizon-tal distribution of all artifacts from all levels. It is inthe accompanying tables that artifacts are placed intheir stratigraphic context.


order to a natural podzol, excavated out of eitherUnit 2 or an earlier pit. This is an example of how“reversed stratigraphy”, a term that Harris(1979:34) considers to be a misnomer, has beencreated by recent activities. Fortunately thesedeposits have aided in protecting the underlyingprecontact layers. Unit 6 is a clearly defined lens ofcharcoal, black sand, and rotted wood. Thicker

charcoal deposits occur in the northeast part of theexcavation, possibly resulting from forest fires.Other recent events include Unit 10, which is arecent intrusive pit of brown sand. Correspondingdips in the layers directly beneath Unit 10 mayhave been caused by this intrusive activity or theymay represent older post moulds. To simplifynumbering, Unit 12 is assigned to several vertical

Figure 10. Three Pines site - Feature Area 15N and 17 Profiles.


intrusives, mostly filled with charcoal from rootburns and subsequent tree root growth. Some mayalso be post moulds, because wood stakes from aprospector-style tent were found nearby.

Unit 7 is hard-packed grey ashy sand ofconsiderable thickness. It is one of the bestpreserved and most extensive precontact units atThree Pines. Unit 7 is interpreted as remnantcobble hearths and living floors. Notable is thecomplete absence of faunal bone. Two clusters ofsmall, rounded, unmodified pebbles occurringnear the top of Unit 7 may represent its mostrecent deposit. These are identified as net sinkers(Figure 11: #231, 232) for fishing nets (or perhapsa child’s collection of shiny pebbles).

At the north end of Unit 7 are a series of largegranitic cobbles, some piled on top of others, withpseudo–scallop shell Rim Vessel E sherds3 (Figure12: E) and very small trapezoidal or quadrilateral

scrapers of Hudson Bay Lowland (HBL) chert(Figure 12: #78, 79, 80, 49) (see below for lithicsource descriptions). This northern cobble clusteris interpreted as an interior hearth into which thescrapers and various pottery sherds were discarded.The scrapers may have been too small toresharpen. Additional Rim Vessel E sherds are alsofound at the southern end of Unit 7 in a test pit.West of the test pit, several tools are distributed ina linear fashion, perhaps indicative of former treeroot activity. Small trapezoidal and triangularHBL chert and other chert scrapers (Figure 12:#43, 47, 48) are present, along with a small side-

Figure 11. Three Pines site - Feature Area 15N and 17 Composite Plan View with Artifact Distribution (LI-VIII)and Harris Matrix Stratigraphic Sequence (See Footnote 2).

3 Throughout this paper, the term Rim Vessel refers toall individual rims sherds and associated sherds from asingle original pot, while Body Vessel refers to allindividual sherds (non-rims) associated with a singleoriginal pot base


Unit1

2

3

4

5

6

7

8

9

10

11

12

DepositGrey-brown sandoverburden, 2–5cm thick

Loose brown sandpits, 40 cm deep

Orange and brownsand, 3 cm thick

Black charcoallense, 1 cm thick

Medium grey sand,4–6 cm thick

Charcoal and rottedwood, 2–4 cm thick

Grey ashy, hard-packed sand withcobbles, 3–25 cmthick

Orange brown sand

Yellow-brown subsoil

Pit with brown andgrey sand fill

Mottled dark grey,light grey, yellowsand, 12 cm thick

Charcoal-filledvertical intrusivesand linear trench

EventMixed modern andprecontact

Excavation andinfilling of moderncamper garbage pits

Overturned depositfromUnit 2 excavation



Forest fire deposits

Precontact cobblehearths and livingfloors

Faded fire-reddenedhearth remnant

Intrusive moderncooking pit

Unit 7 extensionunder tree roots

Root burns, tree rootgrowth, post moulds

Artifact Content by Level

LI: faunal bone (1) sawn moose vertebra; debitage (11) 1 HBL, 1 chert, 1clear quartz, 1 vein quartz, 7 misc, 1 local chert core; modern (46) bottlecaps, lid, nail, screw, wire, foil, clear glass, brown glass, plastic tape,Styrofoam, cap liner

LII: debitage (5) 1 HBL, 2 clear quartz, 2 misc; modern (25) bottle cap,Alka Seltzer foil, clear glass, brown glass, plastic

LIII: modern (1) .257 shell case

modern (7) bottle cap, metal fragment, toy wheel, brown glass, plastic(also oil cans, tin cans, Smirnoff bottle, green plastic garbage bag)

faunal bone (1); tools (1) local chert utilized flake #103; debitage (7) 1clear quartz, 1 vein quartz, 5 misc; pottery (1); modern (32) foil, nail,bottle cap, clear glass, brown glass, clear glass with wire, Styrofoam

tools (6) netsinkers #232; modern (5) brass shear pins, foil, wood screw

LII: faunal bone (1); tools (10) 7 netsinkers #231, chert retouched flake #42,trapezoidal HBL scraper #43, vein quartz biface midsection #44; debitage(19) 3 HBL, 6 clear quartz, 4 vein quartz, 2 mudrock, 4 misc; pottery (3)

LIII: tools (17) 7 netsinkers #232; trapezoidal HBL scrapers #78 and#49, vein quartz projectile point #45, chert retouched flake #46,trapezoidal chert scraper #47, triangular HBL scraper #48, HBL unifaceedge #50, HBL scraper fragment #40, clear quartz utilized flake #41,rectangular mudshale biface #39; debitage (83) 20 HBL, 1 chert, 13 clearquartz, 14 vein quartz, 9 mudrock, 26 misc. including a large wacke splitcobble with platform; pottery (153) 6 plain, 2 pseudo-scallop shell; other- 1 red ochre; soil samples #10 and #11

LIV: tools (2) HBL utilized flake #77, quadrilateral HBL scraper #79;debitage (13) 3 HBL, 1 chert, 2 vein quartz, 7 misc; pottery (108) 31from rim vessel E; soil samples #8 and #9

LV: tools (1) trapezoidal HBL scraper #80; debitage (13) 1 HBL, 1 chert,1 clear quartz, 5 vein quartz, 5 misc; pottery (75) 2 plain, 4 pseudo-scallop shell and 11 rim vessel E

LVI: debitage (3) 2 vein quartz, 1 misc; 1 chert core; pottery (16) 1pseudo-scallop shell

LVII: debitage (4) - 1 mudrock, 3 misc.

Tools (1) siltstone biface #51

modern (20) foil, bottle cap, pull tab, clear and brown glass, plastic,Styrofoam

LIV: pottery (15) 2 plain, 3 worn pseudo scallop shell; modern (1) plastictent peg

LV: debitage (4); pottery (7) 2 plain, 1 of rim vessel I

LVI: debitage (1)

LVII: tools (1) siltstone retouched flake #81, debitage (1); pottery (3)

LVIII: pottery (1) plain

Table 13. Feature Area 15N and 17 units of stratification.


notched projectile point of vein quartz (Figure 12:#45) and a small vein quartz biface midsection(Figure 12: #44). A modern tree root protects thisunit from modern activities.

In Unit 11, which may be an extension ofUnit 7 that has been trapped under modern treeroots, a pottery cluster includes linear stamp RimVessel I (see Figure 22: I). Based on the localizedpottery finds and tools, Unit 7 and 11 are assignedto the Middle Woodland Laurel period.

The orange-brown sand of Unit 8 is redder inhue than that of Unit 9, but it is not as saturatedin colour as that in other hearths. Perhaps theoriginal hue has faded due to podzolization.Stratigraphically, Unit 8 predates the Laurelpottery unit. It contains a leaf-shaped biface(Figure 13: #51) of light olive green siltstone—atool unique on this site. It and the deeply buriedlarge retouched flake (Figure 13: #81) of siltstone(base of Unit 11) may be the earliest deposits inthis sequence. These tools are assigned to theArchaic period. Biface #51 resembles excurvatebifaces from the lowest stratum of the AbitibiNarrows site (Ridley 1966: e.g., Fig 4).

Stratigraphic Column for Feature Area 7Field feature 7 is a 7 to 9 cm thick area of hard-packed reddish brown sand, measuring 120 by140 cm (Figure 6). The overburden is thick on thewest half, gradually thinning to leave the feature

ArchaeologicalDescriptionGrey ashy hard-packed sand with cobbles

Hard-packed mottled sand, grey, dark grey, and reddish brown changing to strong brown

Sample Identifier

Sample #10

Sample #11

Samples #8 and #9

Sample #7

Sample #6

Context

FA15N-Unit 7/LIII

FA15N-Unit 7/LIII

FA15N-Unit 7/LIV

FA7-Unit 7/LII

FA7-Unit 7/LIII


poorly sorted, medium sand; sub-rounded; rock fragments 10%poorly sorted, fine sand; sub-angular to angular; rockfragments 5%; organics 10–15%poorly sorted, fine to mediumsand; sub-angular; trace rockfragments; organics <5%

poorly sorted, fine sand; coatedwith limonite; sub-rounded;rock fragments 5%poorly sorted, fine sand; coatedash and limonite mix; sub-rounded; rock fragments 5%

Munsell Colour

Brown 10YR 4/3

Grayish Brown10YR 5/2

Grayish Brown10YR 5/2

Yellowish Brown10YR 5/4

Strong Brown7.5YR 5/6

Table 14. Soil micromorphology Feature Areas 15N/17 and Feature Area 7.

Figure 12. Three Pines site – Unit 7 of Feature Area15N and 17. Lithic Tools and Pottery arranged instratigraphic sequence (unmodified pebble netsinkers notshown). Level II: 42 chert retouched flake; 43 HBLscraper; 44 vein quartz biface midsection LIII: 46 chertretouched flake and 47 scraper; 40, 48, 49, 78 HBLscrapers and 50 uniface edge; 45 vein quartz projectilepoint; 41 clear quartz utilized flake; 39 mudshale biface.LIV: 77 HBL utilized flake and 79 scraper; Rim VesselE. LV: 80 HBL scraper; Rim Vessel E.


exposed at the surface on the eastern edge.Ground squirrels have truncated and intrudedinto this feature with tunnels and a shaft. Thenorthern margin is defined by a tent drainageditch. Twelve units of stratification are identifiedand sequenced for Feature Area 7 (Figures 14 and15; Tables 14 and 15).

Ground squirrel activity is represented by Unit1, a mixed layer above Unit 3, a horizontal lightyellow sand burrow and Unit 4, a vertical shaft,also of light yellow sand, with a short tunnel oflight grey sand running beneath Unit 7. The shafthas caused vertical displacement of artifacts, andits infilling is dated as a modern event by thepresence of a deeply buried metal grommet. Amodern camper’s drainage ditch (Unit 5) is atright angles to a similar ditch to the east. Unit 5 isunderlain by a charcoal and dark grey sand-filledpost hole (Unit 6). Modern artifacts are few. Anineteenth-century component is represented by aball clay pipe stem on the surface, and three pipebowl fragments in Unit 2, the thicker overburdenon the west side (Figure 16). These bowlfragments fit a reconstructed pipe bowl (Figure16) found in an adjacent historic hearth(FA22/23-Unit 6).

Unit 7 is an area of hard-packed mottled darkgrey, light grey and reddish brown sand. Unit 7 isa unique settlement feature on the site. It is slightlybuilt up. Its brownish sand has a greasy textureand it contains an exceptionally high quantity ofrock fragments, broken tools and debitage. A widevariety of fauna is represented including beaver,white-tailed deer, moose, Canis sp. (wolf or dog),and black bear. All of the black bear bones, andmost of the moose and Canis sp. (wolf or dog)identified from the Three Pines site occur in thisunit (see Table 3).

Unit 8 is a cluster of siltstone flakes and asiltstone biface edge found below the fading baseof Unit 7. At the southern margin of Unit 7,where the overburden is thicker, is an area of hard-packed grey ashy sand (Unit 9), also interruptedby the rodent tunnel (Unit 3). Unit 7 appears tocut abruptly into Unit 9 and does not extendbeneath it. For these reasons it is placedstratigraphically above Unit 9. The western sectionof Unit 9 contains a concentration of mudrock

Figure 13. Three Pines Site - Archaic Period tools byraw material. Metamorphosed olive grey quartzsiltstone or quartzite from the Lorrain, Gordon Lake, orBar River Formations includes: 51 leaf-shaped biface;81 retouched flake; 169 side-notched projectile point.Gowganda Formation sandstone: 217 side-notchedprojectile point. Gowganda Formation mudshale: 144rectangular flake knife. Local Milky Vein Quartz: 220large flake or preform. Exotic unidentified Cherts: 142snub-nosed scraper with incipient side-notching; 145thick scraper fragment; 222 excurvate biface. Exoticblue Quartzite: 221 bi-pointed biface with slightlyserrated edges. Exotic Clear Quartz: 143 biface tip.

Figure 14. Three Pines site - Feature Area 7 Profile.


flakes, while the eastern section has debitage of allraw materials and some calcined bone.

The distinctive nature of Unit 7 may be anexample of redeposition (Butzer 1982; Schiffer1987), whereby more recent occupants swept upprevious deposits, then used this built-up area forcooking and disposal. In 1978 at Lac Washadimi,Gordon witnessed an Eastern James Bay Creefamily from Fort George preparing the floor of anew conical tent (michuap). They gathered all thesharp stones and debris around the sleeping areaand tossed them into a raised central hearth beforelaying down spruce-bough flooring (Gordon1980, 1988b). This possibility of redeposition isfurther supported by the fact that the adjacentsquare is mostly exposed yellow-brown Bf horizonwith few artifacts. The fact that the faunal bone islarger and better preserved hints at a more recentre-use of underlying archaeological sediments.

Based on stratigraphic position first, thensimilarities in lithic raw materials, I suggest thefollowing reconstructed sequence of events. Theoldest deposit, predating Unit 7 hearth formation,is localized knapping of mudrock, including alarge utilized siltstone flake (Unit 9; Figure 17:

Figure 15. Three Pines site - Feature Area 7 Composite Plan View with Artifact Distribution (LI-VI) and Harris Ma-trix Stratigraphic Sequence (see Figure 11 for legend).

Figure 16. Three Pines site – Historic Period Artifacts:228, 229 Gunspalls; 230 Gunflint; ball clay pipe bowlsand stems including a “Henderson Montreal” stem;Decorative Metal.


DepositGrey and yellowsand overburden,1–3 cm thick

Grey-brownoverburden, 1–3cm thick

Light yellow sandtrench, 2–8 cmthickCircular verticalshaft of light yellowsand, 63 cm deep

Grey-brown sandlinear depression, 4cm thick

Circular stain ofcharcoal and blacksand, 4–7 cm thick

Hard-packed,mottled grey, darkgrey, reddish brownsand changing tostrong brown withnumerous fire-cracked rocks, 8 cmthick

Cluster of siltstoneflakes and fire-cracked rock

Grey ashy sand,very hard-packed, 5cm thick

Slightly fire-reddened soil, 3 cmthick

Orange brown sandBf horizon

Light yellow sandC horizon

EventMixed modern andSmearing of exposedsubsurface layers overtop of ground squirreltunnel (Unit 3)Overburden, thickerover west section

Ground squirrelburrow

Excavation andinfilling of groundsquirrel burrow

Tent drainage ditchexcavation andinfilling

Post hole infilled withorganics

Remnantcooking/disposalhearths

Localizedflintknapping

Localized mudrockknapping (west) andcooking/disposalhearth remnants (east)

Hearth remnants

Artifact Content by Levelfaunal bone (30); tool (1) HBL retouched flake #15; debitage (35);modern (1) nail

Surface: historic (1) clay ball pipe stem

LI: faunal bone (34) 3 beaver; tools (4) chert scraper #59, vein quartzuniface edge #14, HBL biface edge #18, chert utilized flake #16; chertcore (1); debitage (82) mostly misc and HBL; historic (4) 1 clay ball pipestem, 3 clay ball pipe bowl fragments; modern (1) radio pickup

faunal bone (1); debitage (30); pottery (1) pseudo scallop shell; other (2)pine cone, peach pit

LI–IV: debitage (6); modern (2) cork, bottle cap

LV–VI: faunal bone (7) 2 beaver, 1 deer, 1 bear; debitage (9)

LXI–XXI: faunal bone (47) 3 beaver, 3 deer, 1 Canis sp.; debitage (51) 5HBL, 22 chert, 2 clear quartz, 1 vein quartz, 21 misc; modern (1)grommet; fire-cracked rock

LI: faunal bone (59) 10 beaver, 3 moose; tools (1) sandstone celtfragment #63; debitage (35) 2 HBL, 8 chert, 4 clear quartz, 5 vein quartz.1 mudrock, 15 misc.

LII: faunal bone (634) 97 beaver, 9 deer, 2 moose, 3 Canis sp., 3 bear;tools (6) snub-nosed HBL scraper #19, HBL biface edges #21 and #22,square chert scraper #23, siltstone retouched flake #24; quadrilateral HBLscraper #60, 1 chert core; debitage (281) 55 HBL, 53 chert, 13 clearquartz, 14 vein quartz, 8 mudrock, 136 misc; pottery (4) 1 rim vessel F;modern (1) bottle cap; soil sample #7

LIII: faunal bone (241) 26 beaver, 1 moose, 2 deer, 3 bear; tools (5) HBLutilized flake #65, sandstone hammer/grinding stone #25, rectangularmudshale biface #26, mudshale celt fragment #27, HBL uniface edge#28; debitage (166) 39 HBL, 19 chert, 4 clear quartz, 6 vein quartz, 6mudrock, 92 misc; soil sample #6

LIV–VI: faunal bone (50) 2 beaver, 1 deer; debitage (24) 7 chert, 2 veinquartz, 2 mudrock, 13 misc.

LIII: tool (1) siltstone biface edge #64; debitage (50) chert (siltstone)

LIV: debitage (11) 1 HBL, 5 chert, 5 misc; modern (1) cigarillo tip

faunal bone (36) 2 beaver, 1 deer; tools (1) siltstone large utilized flake#17; debitage (128) 39 HBL, 7 chert, 1 clear quartz, 4 vein quartz, 45mudrock, 22 misc.

faunal bone (13) 1 beaver; debitage (5)

Table 15. Feature Area 7 units of stratification.

Unit1

2

3

4

5

6

7

8

9

10

11

12


#17), and localized flaking of siltstone, includingthe siltstone biface edge (Unit 8; Figure 17: #64).Tools situated deepest in Unit 7 itself include asandstone hammerstone or grinding stone (#25)suitable for making groundstone tools, a mudshalecelt fragment (#27), and a rectangular mudshalebiface edge (Figure 17: #26). All of these tools andevents are assigned to the Archaic period.

Over top of these deposits, Woodland hearths(Unit 7) were built, including one with a MiddleWoodland occupation represented by Laurelpottery, a single pseudo–scallop shell Rim Vessel F(Figure 17: F), and a similar decorated sherd inthe ground squirrel tunnel (Unit 3). If Unit 7 is

the result of redeposition, the faunal bone mayrelate to a more recent, though undated,occupation. Remains of big game, such as mooseand bear, and medium-sized game such as white-tailed deer, wolf, and beaver were disposed of inthese hearths. These species could have beenhunted during any season of the year, includinghibernating bears, as suggested by analogy tocontemporary James Bay Cree hunting practices(Feit 2004). Also disposed of in these hearths werea variety of scrapers, uniface and biface edges,retouched and utilized flakes of HBL chert, othercherts, and siltstone (Table 15; Figure 17).

Stratigraphic Column for Feature Area 4Field feature 4 is a 4–7 cm thick area of hard-packed grey ashy sand, 90 cm by 60 cm in size(Figure 6). It is on a stabilized flatter ridge betweenthe west and central erosional gullies. Here theoverburden is thicker and overlies a remnant Hhorizon and an Ae podzolic horizon. Eleven unitsof stratification are identified, sequenced,interpreted, and phased (Table 16).

A series of historic and more recentoccupations overlie or intrude into the precontactunits (Figures 18 and 19). These include recentcamping and picnicking activities (Unit 1) andsomewhat older recreational usage as indicated bythe older style camera flashbulb, as well as rustingmetal hardware (Unit 8). An older hunting camphearth is indicated by the rock and charcoal hearth(Unit 3) containing rifle ammunition and severalpost moulds (Unit 4). A utilized mudshale flake(#141) may have been collected from elsewhere orused as found to build this modern stone hearth.Although interrupted by Unit 8, a trace of an Hhorizon of decomposing organics (Unit 9)contains a single hammer-and-chisel-cut Englishgunflint (Figure 16: #230).While Unit 8 clearlyintrudes into Unit 9 and 10, its margin with Unit3 is indistinct. It was therefore not identified asintruding into and predating Unit 3.

Unit 5 is interpreted as a remnant precontactliving floor and/or hearths. It contains severallarge, distinctive tools found in a cluster: thebroken tip of a clear quartz biface (Figure 13:#143); a complete, large, snub-nosed grey chertscraper with incipient side-notching (Figure 13:

Figure 17. Three Pines site-Feature Area 7 PrecontactLithic tools and Pottery by stratigraphic sequence (withpotential redeposition events). * = assigned to ArchaicPeriod. Unit 2: 16 chert utilized flake and 59 scraper;18 HBL biface edge; 14 vein quartz uniface edge. Unit7 Level II: 19 HBL snub-nosed scraper; 21 and 22 HBLbiface edges and 60 scraper; 23 chert scraper; 24 siltstoneretouched flake; pseudo scallop shell Rim Vessel F. Unit 7Level III: (65 HBL utilized flake and 25* sandstonegrinding stone not shown); 26* mudshale biface; 27*mudshale celt fragment; 28 HBL uniface edge. Unit 8LIII: 64* siltstone biface edge. Unit 9: 17* siltstoneutilized flake.


#142); a complete, very large, rectangularmudshale flake knife (Figure 13: #144); and afossiliferous blue-grey local chert scraper fragmentmade on a thick flake (Figure 13: #145). Thesetools suggest hide skinning, scraping, andbutchering. While all raw material types arerepresented in the debitage of Unit 5 (HBL chert,other chert, clear quartz, vein quartz, mudrock,and others), HBL chert dominates. However, onlyone HBL chert tool, a retouched flake (#146), wasrecovered. At least two separate events are thereforerepresented: the discard or loss of the larger toolsand HBL chert flintknapping.

Phasing the precontact units of this sequenceof stratification is assisted by information from the

Deposit

Grey-brown sandoverburden, 2–8cm thick

Yellow sand lenses,1–3 cm thick

Large fire-crackedrocks, decayedwood, charcoal anddark grey sand, 2–4cm thick

Intrusive pits ofdark grey sand

Grey ashy, hard-packed sand, 4–7cm thick

Rocks in yellow-brown subsoil

Yellow-brownsubsoil

Dark brown sandwith organics andcharcoal, 4–9 cmthick

Very dark brownsand anddecomposingorganics, 2–3 cmthick

Event

Aeolian deposits,recent soildevelopment

Wind-blown sandfrom active beach

Modern 20th centuryhearths within H andAh soil horizons

Infilled post moulds

Precontact livingfloors, hearthremnants, brokentool discard, HBLchert flintknapping

Precontact depositrelated to Unit 5

Modern hearths,intrusive, dippingdeep into underlyingunits

Historic (18th–19thcentury) hearth in Hhorizon

Artifact Content by Level

faunal bone (1); lithic debitage (23); modern (11) 1973 penny, nails,wood screw, zipper, clear glass, plastic

lithic tool (1) utilized mudshale flake #141; debitage (3); modern (2 ) .22bullet, bag tag

LII: debitage (21) mostly HBL chert

LIII: faunal bone (6 ) 1 beaver; tools (4) large chert scraper #142, clearquartz biface tip #143, large mudshale flake knife #144, local chertscraper fragment #145; debitage (27) - mostly HBL

LIV: debitage (28) - mostly HBL

LV: tool (1) HBL chert retouched flake #146; debitage (1)

faunal bone (5)

faunal bone (5); modern (17) - pull tabs, foil, rusty nails, wood screws,meat can key, bottle lid and cap, older camera flashbulb, tent pole cap,twist tie

tools (1) English gunflint #230

Table 16. Feature Area 4 units of stratification (see text for Units 10 and 11).

Unit

1

2

3

4

5

6

7

8

9

Figure 18. Three Pines site - Feature Area 4 Profiles.


adjacent Feature Area 7. FA4-Unit 5 correlateswith FA7-Unit 9, placing it stratigraphically belowa built-up hearth (FA7-Unit 7) containing aLaurel pseudo–scallop shell rim (Figure 17: F).The cluster of large tools therefore predates aMiddle Woodland occupation. The large snub-nosed chert scraper (Figure 13: #142) resembles adeeply buried Mattawan stratum tool from theFrank Bay site (Ridley 1950-53). The blue-greylocal chert scraper (Figure 13: #145) resembles thedark blue chert-like debitage found in greatquantities in the lower Archaic levels of theMontreal River site (Knight 1977).

A cooking and/or disposal hearth isrepresented by Unit 11, which contains fire-reddened sand, calcined beaver and other calcinedmammal bone. This unit is separated from theoverlying historic hearth (Unit 9) by a thin Ae andash lens (Unit 10). It predates the historicoccupation and is more likely associated with theprecontact occupations.

Three Pines Site Sequence of OccupationsTo develop a site-wide sequence, individualcolumns are correlated stratigraphically with eachother where possible. An example is given for theSW corner of the site, where Feature Areas 7 and4 occur (Figure 20). The result for this site is aseries of columns, with only the overburden andyellow-brown horizon present in each one. Theartifact contents are used to phase these combinedunits into temporal designations. The following isa summary of those broad temporal periods,presented from the top down. This brief summaryof the relative sequence of occupation is based oninformation from all 24 stratigraphic columns,not just the 3 presented in detail here.

Modern Period. Modern (that is, twentieth-century) occupations of the Three Pines site fallinto two categories: recreational use and huntingcamps. Boaters and overnight campers have left

Figure 19. Three Pines site - Feature 4 Area Composite Plan View with Artifact Distribution (LI-V) and HarrisMatrix Stratigraphic Sequence.


food containers, cameras parts, picnic refuse,boating refuse, and lightweight tent gear.Settlement features include tent drainage ditches,post moulds, garbage pits, and a possible roastingpit. Older hunting camp hearths contain coinsdating to 1910–1936, 1945, and 1962; rifle andshotgun ammunition; meat can keys; and oil lampchimney fragments. Thicker deposits of decayedwood, charcoal, dark grey sand, and blackenedcobbles characterize these older hearths. Beaverand moose were trapped and hunted by theseoccupants, who also used A-frame canvas tents.

Historic Period. The Historic period has only aminor representation (Figure 16). Datable artifactsinclude an eighteenth- to nineteenth-centuryblack English gunflint (Noël Hume 1980) and an

1846–1876 Henderson Montreal clay pipe stem(Smith 1986). This pipe was found together withtwo punched-out gunspalls and non-calcinedbeaver and deer bone in a hunting camp hearth.Other ball clay pipe fragments and two rusteddecorative metal items are nearby. Historic perioditems occur in thin layers of dark browndecomposing organic matter over either fire-reddened soil with large fragments of calcinedbone or very bright red soil patches. The brightred colour may be a temporal indicator. Itgradually fades as iron compounds are leached outby acidic rainwater, part of podzolic soilformation.

Faunal bone is better preserved and in largerpieces in the Historic period units. Beaver, white-tailed deer, marten, ruffed grouse, and common

Figure 20. Three Pines site - Harris Matrix for the combined southwest section of the excavation grid including FeatureAreas 7 and 4.


loon were exploited. Ruffed grouse and martenhave more fragile bones than do beaver, white-tailed deer or the heavy bones of the commonloon which enable it to dive deeply. Their presencesupports the artifact dating for these deposits.Common loon bone means a spring to fall seasonof capture, while the other species could have beenhunted during any season.

There is a strong correlation between thedistribution of remnant dark brown F and Horganic horizons and the distribution of Historicperiod settlement features and artifacts. These soilhorizons have developed in the northwest cornerof the site over at least the past 150 years. Theirabsence in other areas may be due to subsequentdestructive activities (erosion, modern camping,and trampling).

Late Woodland Period. Only a few units ofstratification could be phased to the LateWoodland period (Figure 21). Settlement tracesincluded a semi-circular line of cobbles (hearth orwindbreak), a line of rectangular stones (hearthplatform or working surface), and fire-reddenedhearths with calcined beaver and white-tailed deerbone. Associated pottery includes two Rim Vessels(G and H) with cord-wrapped stick–impresseddesigns and one Body Vessel (O) textured with acord-wrapped paddle. Although fragmentary,neither of the Rim Vessel fit descriptions ofBlackduck ware, which is widely distributed onthe Canadian Shield farther west (Gordon 1985).Blackduck is also at the Frank Bay site (Brizinski1980). All three vessels are speckled with dark greytemper. Two small, indistinct linear trailed sherdswere also found.

Lithic tools included a small, triangular, side-notched HBL chert projectile point, rectangularscrapers of other cherts and of the localfossiliferous chert, as well as HBL chert retouchedflakes. It was not always possible to separate lithictools associated with Late Woodland pottery fromthose associated with Middle Woodland pottery,especially when these were found in the samestratigraphic unit.

Middle Woodland Period. Middle Woodlandoccupations are more extensive at the Three Pinessite. Settlement traces include many grey sandwith ash and cobble deposits that are interpretedas high temperature hearths and living floorsmears. Localized areas of fire-reddened sand withcalcined faunal bone are interpreted as eithercooking or disposal hearths (Gordon 1980) or,perhaps, the cooler edges of such hearths. Onedeposit of brown clay, sand, and small stones isinterpreted as a pottery-making area. A single postmould or pit was also identified.

The extensive nature of these settlement tracessuggests considerable time depth of repeatedseasonal occupations by Middle Woodlandhunter-gatherers. Beaver, moose, and white-taileddeer are three game species clearly associated withthese occupations. The occupants may also haveexploited black bear and wolf. These species areavailable year-round. One taxon, a pond turtle, isclearly a warm season indicator. Two clusters ofunmodified pebbles, identified as netsinkers,suggest open water fishing.

Figure 21. Three Pines site – Rim and Body Vesselsassigned to the Late Woodland Period.Rim Vessels G and H - cord-wrapped stick–impressed de-signs. Body Vessel O - textured with a cord-wrapped pad-dle. (1) cord-wrapped stick decorated sherd.


Units of stratification are assigned to thisperiod based on the presence of Laurel potteryvessels (Wright 1967b, 1972b). Middle WoodlandRim Vessels include eight pseudo–scallop shellstamp, two linear stamp, and one dentate stamp(see Table 9). Body Vessels including a thickdentate stamp vessel, one combined dentate andlinear stamp vessel and a plain vessel with a conicalbase are also assigned to this period (Figure 22).One spatial pattern emerges based on potterytemper. Four of the pseudo–scallop shell vessels inthe northwest excavation contain diabase temper.In the east and southeast, two different pseudo–scallop shell vessels and one linear stamp vesselcontain granitoid temper. The speckled dark greytemper seen in all the Late Woodland vessels isonly found in one dentate stamp vessel. While

functional, temporal and/or, individualpreferences may be reflected in these differentvessels and their spatial clustering, there is no clearstratigraphic basis for temporally ordering thedifferent vessels.

Laurel ceramics were also found at the FrankBay site (Brizinski 1980). In examining Ridley’s(1950-1953, 1954) collection, I noted that thePoint Peninsula assemblage had vessels similar toThree Pines pseudo–scallop shell Rim Vessel E andto linear stamp Rim Vessel I. Vessels from Ridley’sPrimary Transitional assemblage were closer to thedentate stamp Body Vessel (M). Vessel M bearssimilarities to southern Ontario Saugeen ware(James V.Wright, personal communication 1988).

Lithic tools include small triangular scrapers,retouched flakes, and small side-notched points,all of different types of chert (HBL, anunidentified chert, and a local grey chert). Localclayshale, mudshale, sandstone, and siltstone wereused for flake knives, while cobbles and pebbleswere used for grinding stones, abraders, and celts.Vein quartz was also used for the occasional side-notched projectile point and biface.

The discontinuous nature of the units ofstratification did not allow for temporal separationof individual Middle Woodland units across thesite.

Archaic Period. Several units of stratification arephased to the Archaic period. These are thedeepest units, found stratigraphically belowMiddle Woodland; unidentified Woodland; and,in some cases, Historic period units. Thedistribution of Archaic period occupations mayreflect differential preservation. Many are locatedin areas of thick deposits with intact podzolic soilhorizons.

Settlement features include deep basins of greyashy sand with cobbles and dark orange-brownsand deposits, often associated with calcined bone.One deposit contained beaver and mammalbone.The orange-brown sand deposits are redderin hue than the yellow-brown subsoil, and somehave grains coated with ash. These may be thefaded remnants of once brighter fire-reddenedsand hearths.

Figure 22. Three Pine site - Pottery Rim and BodyVessels assigned to the Middle Woodland Period. RimVessels Q, C, B, E, F, L, J, A - pseudo-scallop shellstamped. Rim Vessels I and D - linear stamped. RimVessel K and Body Vessel M - dentate stamped. BodyVessel P - combined linear/dentate stamped designs.(Body Vessel N – plain conical base not shown).


Archaic period units contain predominantlylarger flaked core tools of distinctive raw materials,siltstone, quartzite, fine-grained sandstones, clearquartz, vein quartz, and various grey and bluecherts (see Figure 13). Tools include larger side-notched projectile points; a large stemmedprojectile point or drill; medium to large excurvatebifaces, including one bi-pointed biface; retouchedflakes; flake knives; and a large snub-nosed scraper.Inferred tool functions suggest hunting,butchering, food preparation, and hidepreparation. Tool making and primary lithicreduction were not major activities at Three Pines,judging by the low frequencies of decorticationflakes, thinning flakes, large cores, or shatter of theabove-mentioned lithic raw materials.

Several Three Pines tools have close matchesin size, shape, tool morphology, and raw materialwith identified Shield Archaic tools from LakeAbitibi and Lake Nipissing. A large snub-nosedchert scraper (Figure 13: #142); a narrow-bladed,side-notched quartzite projectile point (Figure 13:#169); and a leaf-shaped siltstone biface (Figure13: #51) all find counterparts in the loweststratum of the Abitibi Narrows site (Ridley 1966),the Joseph Bérubé site (Marois and Gauthier1989), and the Mattawan stratum of the FrankBay site (Ridley 1954; Wright 1972a). While thereis broad similarity in lithic raw materials and toolclass between the Three Pines site and the muchdeeper Montreal River site Shield Archaiccomponents (Knight 1977), there are no specificartifact matches.

Lithic Trends Through TimeStratigraphic analysis of the Three Pines sitereveals changes in tool morphology, manufacture,lithic raw material usage, and lithic procurementstrategies over time. Distinctive changes in toolmanufacture and raw material usage occursbetween the Archaic andWoodland periods. Largeunifacial flake tools and bifacially-flaked core toolsare prevalent in the Archaic units. A shift tosmaller, unifacial and bifacial flake tools is evidentin the Middle and Late Woodland units.

Vein quartz and greywacke (mudrocks) aresuggested as temporal indicators for the nearbysites of Sand Point and Witch Point (Conway

1982, 1986). No stratigraphic evidence was foundat Three Pines to support the hypothesis thatreliance on vein quartz usage is characteristic ofthe Late Woodland period or that greywacke usageis an Archaic period indicator. Local vein quartzand mudrocks occur as tools and debitage in bothArchaic and Middle Woodland contexts at theThree Pines site.

Archaic Period Raw Materials. Archaic toolmakersfavoured sub-conchoidal fracture olive greyquartzite, siltstone, sandstone, mudshale, andwacke; all derived from local bedrock formations(see Bedrock Geology section above, and Table 1).A distinctive metamorphosed olive grey quartzsiltstone or quartzite may be from the Lorrain,Gordon Lake, or Bar River Formations (e.g.,Figure 13: #51, #81, #169). A very fine-grainedlight grey sandstone (e.g., Figure 13: #217) andsofter greenish grey mudshale, mudstone, andclayshale (e.g., Figure 13: #144) come from theGowganda Formation. These materials were eitherquarried directly at bedrock outcrops or gatheredas rounded cobbles found in the glacial drift. Theonly local conchoidal fracture material used byArchaic tool makers is milky vein quartz (e.g.,Figure 13: #220). One source is the Crystal site(CfHa-34), located 7.5 km south of Three Pineson the west-central mainland. It is a large lakesidevein of particularly homogenous material that hasbeen extensively quarried and processed at theadjacent Blueberry site (CfHa-32) (see below).

Exotic to Lake Temagami are dark grey chert(e.g., Figure 13: #142), blue chert (Figure 13:#145), and light bluish-grey quartzite (Figure 13:#221). Their bedrock sources are not yetdetermined (Burbidge 1988). In the lower levels ofthe Montreal River site, Knight (1977:45)describes large quantities of thick, unfinishedpreforms made on a dark blue-grey chert. Hesuggests a local bedrock source, on LakeTimiskaming or the Montreal River. A clear quartzbiface (Figure 13: #143) is also exotic to LakeTemagami.

Woodland Period RawMaterials.Woodland flakedtools tend to be much smaller in size and are madefrom flakes, not cores, which have been unifacially


or bifacially retouched. Woodland tool makersfavoured hard, conchoidal fracture material,primarily Hudson Bay Lowland chert—acolourful, homogenous nodular material that iseasily worked and produces very sharp edges. HBLchert tools include small trapezoidal or triangularscrapers, (e.g., Figure 12: #78, 79, 80, 49),retouched flakes, and small side-notched projectilepoints. A poorer quality fossiliferous “local” chertand other unidentified cherts were also used. Veinquartz tools are less common and include a bifacemidsection and a small side-notched point (e.g.,Figure 12: #44, #45).

Woodland tool makers could only obtain chertin small nodules and pebbles from glacial fluvialand reworked beach deposits—the fossil-bearing“local” chert, originating in the TemiscamingOutlier that is found on Temagami beaches.However, HBL chert nodules are not local(Burbidge 1988). Lake Temagami lies well southof the “carbonate line” (Karrow and Geddes1987), which marks the southern extent ofreworked, glacially deposited sediments fromPalaeozoic formations in the Hudson BayLowlands. To obtain the highly desirable HBLchert nodules, Woodland tool makers eithertravelled north or traded with northern groups.While changes in tool morphology and size, aswell as a clear shift to the use of nodular cherts,have been suggested before for northeasternOntario (Knight 1977; Kritsch-Armstrong 1982;Wright 1972a), this analysis offers stratigraphiccorroboration from a shallow habitation site.

Occupations Moving InlandThe Harris Matrix stratigraphic analysis alsoreveals a subtle difference in spatial distributionover time. Historic period occupations are severalmetres farther inland than some Archaic periodones. This shift raises the intriguing possibilitythat lake levels altered over time, possiblyencroaching on the site’s front edge. Historicperiod settlement features and artifacts arelocalized in the northwest part of excavation instratigraphic units from Feature Areas 4, 5, 7,22/23, and 24 (Figure 6). These finds stronglycorrelate with buried remnant F and H soilhorizons, which tend to be absent in the rest of

the excavation. Units assigned to the Archaicperiod are more evenly distributed across theFeature Areas, but several are near the erosionalface of the upper terrace. These are FA8N-Units 5and 6, grey ashy hearths with a siltstone projectilepoint (Figure 13: #169), and the deeply buriedFA20-Unit 9, a series of grey ashy hard-packedhearths, with a cluster of distinctive tools,including an excurvate chert biface (Figure 13:#222) and a bi-pointed biface of light bluequartzite with slightly serrated edges (Figure 13:#221). Of course, the destructive effects of erosionand modern usage may have removed traces ofhistoric occupations from other parts of theexcavation. But, if this spatial difference is a truereflection of site usage, and assuming a preferencefor camping a uniform distance from the water,Lake Temagami may have been lower and the sitemay have been larger in the Archaic period.

The Three Pines Site: Landscape Context

Exploring the natural and cultural formationprocesses that created the Three Pines sitelandscape context, or site mesoenvironment(Butzer 1982), was the focus of fieldwork in Juneand August 1987. Questions of particular interestwere: how old is the site’s landscape, how did itchange over time, and how would such changeaffect site occupations? To help answer thesequestions, off-site soil samples and soil profileswere collected from the west hill and lower terraceand a trench was dug to profile the beach berm–bog interface (Gordon 1989, 1991a).

Baymouth Bar FormationAnalysis of soil micromorphology and surficialgeology indicate that the Three Pines site issituated on a baymouth bar (Burbidge 1988).Spits, or baymouth bars, extend out frompromontories formed by longshore drift (Figure23), a process whereby waves hitting the shore atan oblique angle gradually move sediments alonga beach (Bell and Wright 1987). The sub-angularsand grains of the west hill indicate unmodifiedglacial outwash. In contrast, the lower terrace sandgrains are rounded, that is, tumbled smooth bywave action. Upper terrace samples include bothsub-angular and sub-rounded clasts, showing some


reworking, but to a lesser degree (Tables 10–12).Burbidge (1988) postulates that during early

post-glacial times, a lagoon occupied the basinbetween the west sand hill and the bedrock dome.Wave action caused by southwest winds graduallyreworked and redeposited sand, forming abaymouth bar where the site is located. Thissandbar eventually reached the rock dome, cuttingthe lagoon off from the lake. As the lagoonterrestrialized, it developed into the Three PinesBog. Longshore drift continued to move sandeastward, creating a recurved sand spit extendinginto the Northwest Arm.

Modern Lake Level RiseLake level rise in the modern period is evident ina lower terrace soil profile (Table 12; Figure 7),where an Ah horizon is now buried under 10 cmof baymouth bar wash-over. This organicallyenriched soil horizon would have developed under

forest vegetation. Local visitors in 1986 said thatpine trees used to grow on the lower terrace. Aspine trees are intolerant of flooding, this indicatesthe lake has encroached on the lower terrace.Indeed, existing control gates for power generationand related activities on the Sturgeon Riversystem, as well as earlier nineteenth-centurywooden dams, can, or would have, elevated levelsabove the norm in more recent times (Gordon1992).

Mid-Holocene Lake Transgression:Evidence from the Three Pines Bog1

In order to radiocarbon date baymouth bar (berm)formation and, by extension, more specifically,date when the Three Pines site was available foroccupation, a peat core for pollen and plantmacrofossil analyses was lifted from the adjacentThree Pines Bog and analyzed by McAndrews

Figure 23. Three Pines site - Longshore Drift and Baymouth Bar Formation. This illustrates how sand from the westhill has been redeposited northeastwards to form a baymouth bar on which the Three Pines Site is located and arecurved sand spit in the NW channel on which the Sand Point site is located. With gradual lake transgression from themid-Holocene to the present, successive baymouth bars and sand spits have been inundated.


(Gordon 1990a). As no palynological work hadbeen done before 1987 on Lake Temagami, thispeat analysis is important in reconstructing localvegetation history (see below). Subsequentresearch has been done by Liu (1990), Hall et al.(1994), and Boudreau et al. (2005). One of thekey findings from the Three Pines Bog, which ledto our 1991 palaeo-hydrological investigations ofLake Temagami, is evidence for a 4 m lake levelrise from the mid-Holocene (7,500 years ago) tomodern times.

LocationThe western margin of Three Pines Bog lies just30 m east of the 1986 excavation (Figure 23). ThisSphagnum moss and Labrador tea–coveredwetland is crescent-shaped, measuring 126 m by64 m, and has a surface elevation about 20 cmabove the lake. It is separated from the lake to thesouth by a 4 m wide and 80 cm high beach berm.Using a metal rod, peat depth was probed along aneast–west transect at the inner berm edge andalong a north–south transect at the midline. Inprofile, the bog is basin-shaped, deepest in thecentre beside the berm edge and gradually sloping

upward to the east, north, and west. A 395 cmcore was lifted from the deepest peat using aLivingstone piston sampler (H.Wright 1967).Lithologic, thermal, and pollen analyses wereconducted by McAndrews (Figure 24), plantmacrofossil analysis by D. Siddiqi (Figure 25) andpollen zone correlations by K-B. Liu at the RoyalOntario Museum (Gordon 1987, 1990a, 1991a;Gordon and McAndrews 1992).

Core CompositionThe core is peat (c. 95% organic matter), exceptfor grey, silty sand in the basal 5 cm. Peat iscomposed of vascular plants and moss, whichgrow with their roots in the water. Over time,when these plants die, the water-logged conditionsinhibit decomposition of their organic material aswell as that of pollen from surrounding vegetation(Faegri and Iversen 1975). Lake sediment wasabsent, which means that the basin was notflooded by the lake. The mineral content at thebase of the core may be from sand blown up fromthe beach. The slightly increasing mineral contentin Zones 3b and 4 is also aeolian.

Figure 24. Pollen Diagram for the Three Pines Bog, Lake TemagamiThe peat core spans 7,500 years. Pine (Pinus), birch (Betula) and spruce (Picea) are the dominant tree pollen. In Zone3a white pine (Pinus strobus) dominates except at the base where jack pine/red pine (Pinus banksiana/resinosa) has asingle peak. Horsetail (Equisetum) and marsh fern (Dryopteris) type together with sedge and grass seeds indicate a nu-trient-rich fen. On the other hand, Zone 3b has much shrub pollen together with bog rosemary seeds, which indicate arelatively nutrient-poor bog comparable to the modern bog.


Radiocarbon Dates and Sedimentation RatesPeat accumulation begins at the coring location inthe mid-Holocene. Table 17 gives calibratedradiocarbon dates for the basal peat and themiddle of the core. Peat accumulation rates arecalculated from these two dates as well as the riseof Ambrosia (ragweed) pollen at 25 cm and thebog surface. The rate of peat growth was nearlyuniform, averaging 5.3 cm per century over thepast 7,334 years. It slowed slightly as the peat levelrose and the centre of the wetland becameincreasingly isolated from the surroundinglandscape, as is normal in a terrestrializingsequence. The rate of peat accumulation iscontrolled by the water table, which in turn iscontrolled by the rate of lake level rise.

Pollen ZonesThe pollen diagram was divided into three pollenzones (Zones 3a, 3b, and 4; Figure 24). Thesezones are numbered to correlate with thoseobserved in other pollen diagrams (Liu 1990)from Nina Lake, Jack Lake, and Lake Six (Figure26). The Three Pines Bog core does not span theentire period from deglaciation (12,000 cal B.P.)to the mid-Holocene. It does show forestsuccession brought by the warmer Hypsithermal(Zone 3a), followed by Neoglacial cooling (Zone3b) and continued cooling (Zone 4) resulting inthe modern Mixed Conifer–Hardwood Forest atLake Temagami. The rise of Ambrosia pollen nearthe top of the core indicates nineteenth- andtwentieth-century field clearance for farming insouthern Ontario and the Little Clay Belt ofnorthern Ontario (Hall et al. 1994; Hodgins andBenidickson 1989; Lovell 1977).

Pollen–climate transfer functions (Bartlein andWebb 1985) convert pollen data into estimatedclimatic data. They suggest this site was warmerand drier 7,500 years ago. The mean Julytemperature was 18.8 degrees Celsius, comparedwith 18.0 degrees Celsius today, and the meanannual precipitation was only 650 mm, comparedwith 820 mm today.

Wetland Plant CompositionFossil pollen and plant macrofossils from wetlandplants show a change in local vegetation (Figure25). The basin at first supported a sedge fen (Zone3a) with pools of water, succeeded by a Sphagnumpeat bog (Zones 3b and 4). The fen stage isindicated by sedge (Carex) and grass (cf. Glyceria)seeds, as well as seeds of water parsnip (Sium),arrowhead (Sagittaria), and bur-reed (Sparganiumcf. emursum). The bog state in Zone 3b is

Figure 25.Plant Macrofossil Diagram from Three Pines Bog.Indicators of a fen in Zone 3a and a bog in Zone 3b areshown here. Chamaedaphne and Andromeda are smallheath shrubs typical of acid bogs. Carex (sedge), Poaceae(grass), Sparganium emersum type (bur-reed), Sagittaria(arrowhead) and Sium (water parsnip) indicate a base-rich fen with pools of water. (D.Siddiqi, analyst).

Table 17. Three Pines Bog radiocarbon dates and sedimentation rate. Overall average rate is 0.0528 cm/yr.(47o00′09″N, 80o 07′23″W; 294 m asl)

Core Depth(cm)

25

185–190

385–393

Date(cal B.P.)

100

3829 ± 94

7334 ± 71

Date(RCYBP)

3540 ± 70

6400 ± 80

13C(%)

−28.56

−28.77

LabIdentifier

Ambrosia rise

WAT-2088

WAT-1987

SedimentationDepth (cm)

0–25

25–188

188–389

SedimentationRate (cm/yr)

0.1820

0.0436

0.0575


Figure 26. Location of Lake Temagami Pollen Cores and other key northeastern Ontario cores: Lake Six, Jack Lakeand Nina Lake (Liu 1990; Liu and Lam 1985); Lake “306” (Hall et al. 1994).


indicated by seeds of bog rosemary (Andromedapolifolia).

Mid-Holocene Lake RiseBased on the above data, the Three Pines Bogformation and its relationship to lake level changecan be reconstructed. Peat began to accumulate atthe coring location around 7,500 years ago. In themid-continental climate of northeastern Ontario,peat grows at or just above the water table(Canadian Soil Classification 1978). Because thewetland is separated from Lake Temagami by onlya narrow, permeable sand berm, the water table ofthe bog fluctuates with the lake level.Measurements of the bog water table in July 1986,June 1987, and September 1987 corresponded tovariation in lake levels and ranged from 0 to 20cm above the lake. In order for 4 m of peat toaccumulate since the mid-Holocene, acorresponding rise in the bog water table and thusin the level of Lake Temagami itself is indicated.Furthermore, peat accumulation and the lake levelhave been rising at an average rate of 5.3 cm percentury, consistent with isostatic rebound. Thetransition from fen to bog is consistent with thewetter climate indicated by the pollen–climatetransfer function as well as with increasingisolation of the wetland from nutrient-rich runofffrom the surrounding landscape.

Revised Three Pines Site Landscape FormationBased on the current slope, a minus 4 m lakesurface places the mid-Holocene shoreline at theThree Pines site 20 m farther south. Thebaymouth bar and peat bog complex could havestarted soon after deglaciation, but farther south.The absence of lake deposits indicates that the bogbasin was not an embayment as earlier posited.Instead peat began to form in a low lying basinseparate from the lake. As the lake rose, the bogspread farther inland. Longshore drift reworkednew shorelines creating successive sandbarsmoving farther and farther northward. As thesandbar approached the coring location, moresandy sediment was blown up from the beach.

Evidence for lake transgression from the ThreePines Bog, supports the hypothesis that a slightdifference in spatial distribution between Archaic

and Historic period occupations of the upperterrace at the Three Pines site relates to lake levelchange. The basal core date indicates that thebaymouth bar and bog complex was in place byat least 7,500 years ago. A shoreline 20 m farthersouth would substantially increase the area ofhabitable space available to precontact occupants.A rising lake would continuously bury lowerterrace soils, and erode upper terrace edges. Thusany early to mid-Holocene archaeologicalcomponents in these locales would now be buriedand inundated.

Palaeo-hydrological Investigations atLake Temagami:

Implications for Precontact Site Location1

Evidence of a 4 m lake level rise since the mid-Holocene at the Three Pines site in the CentralHub of Lake Temagami led to new researchquestions concerning Lake Temagami as a whole.Was lake transgression localized or did it affect theentire lake basin? What was the underlyingmechanism for lake transgression? Was it isostaticrebound altering drainage patterns, was it climaticchange, or was it some other factor(s)? Laketransgression could have an impact on both thelocation and preservation of archaeological sites.Known lakeside sites, especially low-elevation oneslike the Three Pines site, may have truncatedsequences of occupation. Early to mid-Holocenesites in other locations may now be inundated orwaterlogged (see Loewen et al. 2005 for similarfindings at Lac Mégantic, Québec).

The configuration of Lake Temagami fromearly post-glacial to mid-Holocene (12,000–7,500cal B.P.) is unknown; this period is not representedin the Three Pines Bog core, nor is it wellrepresented by northeastern Ontarioarchaeological sites. For the Archaic period, theearliest radiocarbon dates are 8488 ± 105 cal B.P.(7670 ± 120 RCYBP)4 for an early Archaiccomponent at the Foxie Otter site (Hanks 1988)and 5791 ± 275 cal B.P. (5030 ± 240 RCYBP) forthe lowest Shield Archaic stratum at the MontrealRiver site (Knight 1977). For the preceding Late

4 All dates have been calibrated using CalPal 2014.


Palaeo-Indian period, there is no conclusiveevidence north of the Great Lakes. Late Palaeo-Indian quarry and lithic reduction sites do occuron Lake Huron and Georgian Bay to the south(Greenman 1943; Julig 2002; Lee 1953; Storck1984). In northwestern Ontario, surveys ofproglacial Lake Minong yielded Late Palaeo-Indian Plano sites (Dawson 1983; Fox 1975; Julig1995). Pollock (1976, 1984) has suggested thatsurveys of proglacial Lake Ojibway-Barlowshorelines might reveal potential Planooccupations in northeastern Ontario, althoughJulig (1989) cautioned that these palaeo-shorelineswere not yet accurately mapped. These ancientlakes are usually associated with the Little andGreat Clay Belts northeast of Lake Temagami(Dyke and Prest 1987; Lovell 1977).

Thus, it is necessary to look away frommodern shorelines for evidence of early to mid-Holocene occupations at Lake Temagami. Are anypalaeo-shorelines identifiable in the LakeTemagami basin? In this densely forestedlandscape it would be time-consuming to surveyfor ancient strandlines, because no actual palaeo-shorelines have been mapped (Card et al. 1973;Grant 1964; Roed and Hallett 1979; Simony1964; Veillette 1986). Narrowing the search areawould greatly increase the chance of success inlocating early Holocene sites.

In 1991, we developed a computer model tomanipulate one important variable: the differentialeffects of isostatic rebound along Lake Temagami’snorth–south axis. We created a digital model ofthe modern lake basin and surrounding terrainfrom topographic and bathymetric maps (OMNR1969, 1982). We then applied isostatic reboundcurves, developed for the neighbouringTimiskaming basin (Lewis and Anderson 1989),to the modelled Temagami basin, essentially tiltingthe landscape in response to ice retreat. The tiltedvalues were used to identify probable palaeo-outlets and reconstruct likely palaeo-shorelines.Model predictions were tested in the field bycoring upland bogs near two potential palaeo-outlets and documenting geomorphologicalevidence for palaeo-shorelines. Bear Bog, locatednear Sharp Rock Inlet, and Baseball Bog and Jessie

Fen, located within the town of Temagami, werethen cored (see Figure 26). Results indicatedynamic post-glacial and early Holocene waterlevels, shorelines, and lake configurations (Table18). During the earliest phase, three large lakesprobably occupied the single modern basin. Someparts of the lake basin have emerged, while othershave submerged (Gordon 1992; Gordon andMcAndrews 1992; Gordon et al. 1992).

Systematic archaeological survey of thesenewly identified palaeo-shorelines has yet to becarried out. However, brief archaeological surveysnear both palaeo-outlets offered some supportingevidence for lake level change. At the north end,rock paintings at the Deer Island pictographs site(CgHa-23) are high enough up a cliff that theywould be hard to reach unless lake levels werehigher than they are at present. At the northeastoutlet, the landscape context of the inland Portagesite (CgGw-4) could potentially be indicative oflake level change (Gordon 1992; Gordon andMcAndrews 1992). Inland surveys, discussedbelow, also revealed several older hunting campsof the Teme-Augama Anishnabai.

One main reason we chose to present thispalaeo-hydrological work here is to showarchaeologists that they need not wait forgeomorphologists to define palaeo-shorelines.This predictive modelling approach can be appliedto any modern lake, and could be particularlyuseful for lakes with only a southern drainage.Digital terrain modelling software and Ontariogovernment digital elevation maps (DEM) makemapping more efficient and precise, althoughbathymetric values would still need to be addedby hand for most basins. This tilt model helpsidentify locales with a higher probability ofemergent shorelines and therefore potential post-glacial and early Holocene archaeological sites.

Previously Known Post-glacialHistory of Lake TemagamiThe Lake Temagami basin was deglaciated by12,150 cal B.P. (10,400 ± 240 RCYBP) (Veillette1988). It was immediately occupied at least in partby Lake Post-Algonquin, which extendednortheast from the Lake Huron basin to coverLake Nipissing and the southern end of Lake


Timiskaming, with a narrow channel up theMontreal River and into the Northeast Arm ofLake Temagami. This glacially-dammed lakeinundated altitudes up to 310 m asl in the southTemagami basin and 330 m asl in the north end,which may still have been partially submergedunder ice. By 11,500 cal B.P. (10,000 RCYBP),Lake Post-Algonquin drained as ice retreated fromthe Mattawa (North Bay) outlet (Veillette 1988).

Early proglacial Lake Barlow then establisheditself in the Lake Timiskaming basin, alsoextending a narrow arm along the Montreal Riverbasin and 13 km into Lake Temagami (Veillette1988). Evidence of proglacial lake deposits includethe crustaceanMysis relicta in a lake core from theNortheast Arm and laminated glacial lake clay(varves) in a backhoe excavation in FinlaysonPoint Provincial Park, near the Temagami townsite

Table 18. Palaeo-hydrological events at Lake Temagami.

Date(cal B.P.)

01,2003,8005,000

6,300

7,300

>7,600

8,0009,000

9,800

>9,80010,900

11,000

11,500

12,000

Lake TimiskamingBasin (Veillette

1988)

Montreal River site(5,800 B.P.)

ancestral LakeTimiskamingestablished

Lake Barlow beginsdecline in MontrealRiver basinLake Post-Algonquin drains;early Lake Barlowestablished

Baseball Bog,Lake Temagami

(NE)292 m asl

ancestral LakeTemagamiestablished; >289m aslproglacial lake;319–330 m asl

Ice-free

Bear Bog,Lake Temagami

(N)304 m asl

Continued lakeregression; pondchanges to fen

Lake regression;pond separatesfrom lake

ancestral LakeTemagamiestablished >300m asl; high-energyspillway at SharpRock Inletproglacial lake

Ice free

Under ice

Three Pines Bog,Lake Temagami(Central Hub)

293 m asl

291 m asl

Lake transgression;peat begins formingat 289 m asl

Ice-free

PollenZone

443b3a

3a

3a

3a2

1

VerticalDisplacement

(m)

0.22 m0.45 m

0.9 m

1.9 m

3.75 m7.5 m

15 m

30 m


(Jean Veillette, personal communication 1990). Bythe undated McConnell Phase 1, the connectionbetween Lake Temagami and Lake Barlow hadbeen severed (Veillette 1988). Proglacial LakeBarlow began to decline in the Montreal Riverbasin about 11,000 cal B.P. (9,500 BP), whileancestral Lake Timiskaming was established about9,000 cal B.P. (8,000 BP) (Jean Veillette, personalcommunication 1990).

Developing a Predictive ModelTo simplify the estimation of palaeo-shorelinelocation, we focus on basin topography,specifically, outlet sill levels controlling lake levels.Other factors, such as climatic change,evapotranspiration, and sill erosion do play a rolein lake level change. However, for this model, weassume that the region did not experience greaterevapotranspiration than precipitation (i.e., did notbecome a closed-basin, evaporitic lake) and thatthe sills presently formed by Precambrian rock didnot erode significantly. The basin topography,however, would have changed due to isostaticrebound.

Glacial isostatic rebound is the recovery of theEarth’s crust from warping caused by the weightof continental ice sheets. It is most pronouncedwhere ice sheets were thickest and caused thegreatest downward warping, which in easternNorth America is in the region of Hudson Bay(Andrews 1989). As the Late Wisconsinan icesheet melted, the crust rebounded in the directionof ice retreat, which is northeast on the centralCanadian Shield (Andrews 1989; Veillette 1988).Published isostatic rebound curves are available forthe Lake Huron, Nipissing, and Timiskamingbasins. The timing, rate, and direction of isostaticrebound is determined by radiocarbon datingraised glaciolacustrine deposits, plotting theirelevations and locations on height/distancediagrams, and using regression analysis to infer aseries of palaeo-shorelines and their gradients(Andrews 1989; Lewis and Anderson 1989;Veillette 1988).

Isostatic rebound normally follows a logisticdecay function, with the greatest rate of reboundoccurring shortly after deglaciation and the ratedecreasing through time. Along Lake Temagami’s

46 km length, the northeast portion was the last tobe deglaciated, and therefore the south end hadalready partly rebounded by the time the northend was deglaciated and started rebounding. Withthe north end therefore rebounding more than thesouth end, the net effect is one of tilting the basinat a decreasing rate throughout the Holocene.

A digitized map of the modern lake basin andlake outlets was made by plotting altitudes on themilitary grid, every 1000 m for the whole basinand every 200 m in areas with potential outlets(Figure 27: A). Altitudes were taken fromtopographic maps, which have 10 m contourintervals (OMNR 1982 1:20,000 base maps), andbathymetric maps, which have 10 foot (3.048 m)contour intervals (OMNR 1969 1:64,000fisheries maps). The precision with which palaeo-shorelines can be modelled is thus limited to about10 m vertical displacement. An algorithm wasapplied to these values based on published isostaticrebound curves for nearby Lake Timiskaming(Lewis and Anderson 1989), assuming a smoothrate of change in isostatic rebound. Locations withthe lowest sill, or pass out of the basin into outletrivers, were identified as palaeo-sills, and their levelwas used to determine the palaeo-lake level.

Model ResultsThe model predicts that 11,000 years ago, thenorth end of the basin was 30 m lower than thesouth end (Figure 27: B). The rate of crustalrebound has a half-life of about 1000 years,meaning it results in a 15 m relative verticaldisplacement at 10,000 years, 7.5 m at 9,000years, 3.75 m at 8,000 years, and so forth (Figure28). Clearly the basin was very dynamic between11,000 and 10,000 years ago. The extremedisplacement (30 m) suggests that palaeo-shorelines, particularly in earliest post-glacialtimes, may have little or no concordance withmodern lakeshores. Such a displacement wouldhave tilted the entire basin north toward DiamondLake, resulting in a greater water flow and higherwater level through Sharp Rock Inlet. Palaeo-sillsfor this ancestral water body probably occur wellnorth of the study area, perhaps into the LadyEvelyn Lake basin.

Within the Temagami basin, however, there


slower rate, lake levels would continue to declinein the northern half (lake regression) and rise inthe southern half (lake transgression). In this

would be two or possibly three lakes, instead ofthe single large lake that exists today. The entirenorthern half, including the Northwest and NorthArms, formed a lake draining through Sharp RockInlet. This northern lake would have beenseparated from the southern basin by a sill acrossthe Central Hub. This sill would have dammedhigher water levels in the South and NortheastArms. A second lake occurring in the NortheastArm would have drained through a palaeo-outletat the present-day Temagami townsite into theTimiskaming basin. The third lake comprises theCentral Hub and both south arms drainingthrough a now submerged canyon off Sand Pointinto the Northeast Arm. The major modern outletfor Lake Temagami through Outlet Bay–CrossLake to the south was not originally active as anoutlet, but it may have hosted an inflowing streambringing in water from south of the modern basin.

This configuration would have beentransitory, with the relative vertical displacementdecreasing by 15 m over 1,000 years. By 10,000years ago, the lake would have attained roughly itsmodern shorelines—that is, within the 10 m errorrange of the topographic and bathymetric maps.By this time, its modern outlets, including themain southern drainage would also have beenestablished. As rebound continued, although at a

Figure 27. Lake Temagami – a digital map of the modern lake basin (A) shows elevations plotted every 1000 m. SR =Sharp Rock Inlet and FB = Ferguson Bay. To model the configuration of palaeo-Lake Temagami (B) at 11,000 yearsago, elevations were adjusted according to published isostatic rebound curves (Lewis and Anderson 1989), resulting in a30 m downward tilt of the north end of the basin relative to the south end. Arrows show the drainage direction of thetwo predicted palaeo-outlets.

Figure 28. Relative Displacement due to isostaticrebound. As glacial ice retreated to the north, the northend of the Lake Temagami basin was slower to reboundthan the south end of the basin, so that the north end wasdepressed relative to the south end. This graph shows thedecreasing downward displacement of the North Endrelative to the South End over time. The rate of relativecrustal rebound has a half-life of 1,000 years (i.e., ineach thousand years, approximately half the remainingdifference between the north and south ends is recovered),resulting in rapid shoreline change in the early Holocene,continuing more slowly through the mid-Holocene.


model most palaeo-shorelines in the southern endwould now be flooded, because they lie below thecurrent lake level (293 m asl). Some areas in thenorthern half of the basin, where the slope upfrom the present lake level is gentle, should havepalaeo-shorelines inland from modern lakeshores.

FieldTesting the Model Predictions: North BasinThe model predicts lake regression near a palaeo-outlet at Sharp Rock Inlet in the north end ofLake Temagami. Sharp Rock Inlet was examinedand a pollen core taken was from the upland BearBog in July 1991, while the Ferguson Bay outwashplain was examined in August 1991 (Figure 29).

Sharp Rock Inlet. In modern times, the rapidsforming the northern outlet from Lake Temagamiinto Diamond Lake have been dammed, leaving a14 m wide almost dry channel. On thenortheastern side of the channel, above and inlandfrom the former rapids, are two terraces, at 297 mand 298.5 m asl, respectively. The lower terrace

consists of large, rounded rectilinear boulders 53to 156 cm in diameter. The higher terrace ismarked by smaller, rounded boulders 36 cm indiameter. On the opposite shore, boulders arepresent up to 300 m asl. These boulders andterraces indicate a former spillway with sufficienthigh-energy water movement to remove all finermaterial. The water level in this spillway wasprobably close to the upper boulder line, at 300 masl. This represents a water level 7 m above thecurrent level, consistent with the model predictionsof a higher water level at Sharp Rock Inlet.

Bear Bog Pollen Core. Bear Bog is located 2 kmsoutheast of Sharp Rock Inlet (Figure 29). It is alarge heath with black spruce snags, lying 100 to400 m inland from the lake and 11 m above it(304 m asl). On the lake side, the water table isheld up by a 6 m high moraine, or ice-push ridge,which is breached by a 1 m deep dry channel. Theland slopes down to a 2 m high break in slope withlarge exposed boulders, then continues onto a flat

Figure 29.Lake Temagami North End archaeological site locations, Bear Bog pollen core and palaeoshoreline evidence.


sandy area. Here the lake is quite shallow, with asandy silty bottom. Stands of pine with tree fallsoccupy the upper slope, and a dense growth ofsmall aspen, large birch and alder occupies the flatarea.

Using a stationary piston sampler (H.Wright1967), a 460 cm core was extracted from a location77 m west of the ridge, about 177 m from thelakeshore. The core lithology from 460 cm to 200cm deep shows a sequence of lake regression (Figure30). Basal glaciolacustrine clay represents an initialproglacial lake. It is overlain by the lake mud fromancestral Lake Temagami established by 9807 ± 98cal B.P. (8720 ± 120 RCYBP; BGS-1515). Pondmud overlies the lake mud. It must have formed inisolation from the large water body of Lake

Temagami, which would have generated wavescapable of removing pondmud. Assuming a steadyrate of accumulation of sediments, this transitionfrom lake mud to pond mud is estimated to haveoccurred by at least 7,600 years ago. The nextchange, from lily pond to fen peat, dates to 6277 ±36 cal B.P. (5,500 ± 120 RCYBP; BGS-1514).

At Bear Bog, the ancestral lake bed occurs at anelevation of 300 m asl. The water surface wouldhave been several metres higher. This confirms theminimum 300m level for Sharp Rock Inlet, datingit at 9,800 years ago (Figure 31).

Ferguson Bay Outwash Plain. A flat sandy outwashplain and wide beach form the north side ofFerguson Bay (Figure 29). At 300 m inland, a

Figure 30.Pollen Diagram from Bear Bog, Lake Temagami (47o10′20″N, 80o 09′44″W)Sediments show a transition from deep proglacial lake clays, to lake mud of ancestral Lake Temagami to isolated pondmud to a peat fen; consistent with lake regression at the north end of the lake from early to mid-Holocene times. Spruce(Picea), Pine (Pinus) and Birch (Betula) are the dominant tree pollen. Pollen Zone 1 with high white spruce (Piceaglauca) represents an early open Boreal Forest, changing to a closed Boreal Forest in Zone 2 with increase jack pine/redpine (Pinus banksiana/resinosa). Zone 3a (also seen in the Three Pines Bog core) shows a strong rise in white pinepollen (Pinus strobus) indicating the warmer Hypsithermal period.


break in slope occurs at the edge of a regeneratedjack pine stand. At 200 m inland, there is ashallow slope down to the wide sand beach. Thisarea is characterized as “peat organic terrain/lowplain wet” (Roed and Hallett 1979).

Two ridges were measured on the sandysubstrate at a distance of 280 m inland, withelevations of 299 and 302 m asl (6 and 9 m abovecurrent lake level). These are probably palaeo-shoreline remnants. As the lake graduallyregressed, peaty soils covered the shallow slope.These findings agree with the 300 m asl palaeo-shoreline at Sharp Rock Inlet and Bear Bog.5

North Basin SummaryField observations and coring results stronglyfavour the model’s prediction of early Holocenehigh water levels in the northern basin and palaeo-shorelines at the northern end, at or near theSharp Rock palaeo-outlet. Field evidence providesadditional details not predictable by the modeldue to the resolution of the topographic andbathymetric maps. The outlet channel at Sharp

Rock Inlet would have been at least 60 m wide,compared with the present width of 14 m, withwater at or slightly above 300 m asl. This waterplanes dates to 9,800 years ago. The ancestral lakein the north basin experienced reduced waterlevels by at least 7,600 years ago, separating BearBog (now 100–400 m inland) from the main lake.It abandoned palaeo-shorelines now found asboulder beaches and beach ridges between 16 m(Sharp Rock) and 300 m inland (Ferguson Bay),depending on local topography. The absence ofidentifiable palaeo-shorelines between 300 m asland the modern lake level, at 293 m asl, suggestsa smooth transition to the new, lower levels, withno still-stands of significant duration and thus nomajor influences on lake levels in this basin otherthan isostasy.

Field Testing the Model Predictions:Northeast BasinThe model predicted another outlet throughTemagami at the end of the Northeast Arm(Figure 32). There is currently an 850 m portagebetween Lake Temagami (293 m asl) and CasselsLake (289 m asl), which marks the divide betweenthe Lake Timiskaming–Ottawa River drainageand the Lake Nipissing–Great Lakes drainage.This palaeo-outlet would have consisted of severalnarrow channels draining northeast into CasselsLake.

Baseball Bog. Baseball Bog is located 250 m northof Snake Island Lake (an extension of CasselsLake; Figure 32). Situated between two steepbedrock hills, at an elevation of 292 m asl, this 60by 110 m fen is covered with grasses and dogwoodshrubs and fringed with cedar. A 570 cm long coreshows a sequence similar to that of Bear Bog(Figure 33). It reveals basal varved glaciolacustrineclay (alternating bands of light grey and dark greyfine sediments deposited in a deep proglacial lake)overlain by massive light grey clay (the lake mudof ancestral Lake Temagami), (lowering waterlevels), then lake mud (ancestral Lake Temagami),and finally peat (near or above lake level). Atlocations nearby, Veillette (1988) suggests a waterdepth of 30 to 50 m to form rhythmites similarto the varves at the base of the Baseball Bog core.

Figure 31.Sharp Rock Inlet, North End showing (Left) modernLake Temagami with a 300 m asl palaeoshoreline and(Right) reconstructed ancestral Lake Temagami with a304 m asl palaeoshoreline. Evidence from Bear Bogsuggests this ancestral water plane was at least severalmetres higher.

5 Survey in the field also showed a possible error in thetopographic maps, placing the 300 m asl contour at250 m inland, not 45 m inland as indicated on theOMNR (1982) 1:20,0000 base maps, which are com-piled from aerial photographs.


As the uppermost varves occur at 288.5 m asl, thisplaces the proglacial water surface at 319–339 masl (26–46 m above modern Lake Temagami).Massive clay above 288.5 m asl indicates alowering of water level below the point whererhythmites can form, but still well above themodern lake level. This transition is dated at10,895 ± 266 cal B.P. (9,580 ± 200 RCYBP; BGS-1543) and likely represents the establishment ofancestral Lake Temagami.

Jessie Fen. Nearby, on the south side of theNortheast Arm, Jessie Fen reveals an inconclusivesequence (Figures 32 and 33). The fen is bisected

by Jessie Creek, which flows from Jessie Lakenorth into Lake Temagami. The fen is grass-covered with some conifer snags in the centre. The220 cm long core contains glaciolacustrine varves,then massive clay with no varves, then pond mudand a peat cap. Radiocarbon dating of the basalpeat was not practical due to abundant rootpenetration from above. However, on thenortheast edge of the fen is a boulder beach, andthe fen is surrounded by smooth bedrock,reminiscent of the modern wave-washed shoreline.This boulder beach and washed bedrock lies 300m inland from Lake Temagami at an elevation of295 m asl (Figures 32 and 34).

Figure 32.Lake Temagami Northeast Arm and Town of Temagami – archaeological site locations and pollen coring locations.


Northeast Basin SummaryEvidence from Baseball Bog supports Veillette’s(1988) reconstruction of proglacial lakes Post-Algonquin and Barlow inundating the NortheastArm of the Temagami basin. It also dates theseparation of ancestral Lake Temagami fromproglacial Lake Barlow to 10,900 years ago (theMcConnell Phase 1) (Veillette 1988). As LakeBarlow declined, ancestral Lake Temagamidrained northeast, inundating at least severalmetres above the 290 m contour in the CasselsLake basin (Figure 34). Differential isostaticrebound caused this lake to regress, eventuallyclosing the northeast outlet. While the Jessie Fencore was inconclusive, evidence was found, 300 minland, of a 295 m asl palaeo-shoreline.

Archaeological Sites and Lake Level ChangeDuring the 1991 project, several archaeologicalsites were recorded in the north and northeastbasins (Figures 29 and 32), but none had clearevidence for early occupations (Gordon 1992). Asnoted earlier, the Deer Island pictographs are the

one clear archaeological indicator of higher waterlevels found in the North Arm. These are drawnon an east-facing rock face that is not accessiblefrom above, extending from 1.7 to 3.1 m abovethe modern lake level. The modern high-watermark at the site is only 40 cm, placing thepictographs at 1.3 to 2.7 m above it. In contrast,rock paintings in the Central Hub are about 1 mabove the water (Gordon 1995a). While it ispossible that the pictographs were made from asnow-drift in winter, it is more likely they weremade when the lake levels in the North Arm were1 to 2 m higher than today. The estimated rate ofrebound suggests this may have occurred between1,000 and 4,000 years ago.

The nearby South Deer site (CgHa-34) hadsurface finds of chert and vein quartz flakes in acanoe camper’s clearing 2 m above the lake (295 masl) and 7 m inland. A side-notched projectilepoint with a convex blade and rounded base wasfound along the exposed sloping bedrock near thewater. This point is thin, made on a flake, and ofhigh-quality red opaque (burnt) chert. It resemblesMiddle Woodland points from the Three Pinessite. Two palaeo-shorelines are evident behind thisclearing. A break in slope occurs 30 m inland at 3m above the lake (296 m asl), where a lag boulderbeach is found. Behind it, an upper terraceforested with birch and hazel rises to 7.5 m abovethe lake (300.5 m asl). A second break in slope,up to 305 m asl, is formed by a steep scarp.Testing of these interior palaeo-shorelines,especially those inland from lakeside precontactsites in the north basin, may prove to be a usefulstrategy in finding older components.

Within the town of Temagami, contract workfor Algonquin Associates was undertaken alongthe existing TransCanada pipeline corridor andproposed deviation (Gordon 1992). Surveyconcentrated on areas above the 300 m contour,which intersected former drainage channels forthis palaeo-outlet, including the palaeo-shorelineon the north side of Caribou Lake (Figure 34). Allerosional features were examined along a 3 kmsegment of the pipeline right-of-way with testpitting in selected undisturbed areas. Two siteswere recorded: a traditional use camp west ofCaribou Lake (CgGw-6) and a single bipolar flake

Figure 33. Composition of Cores from Baseball Bog(47o04’ 16” N, 79o 46’ 40” W; 292 m asl) and JessieFen (47o02’ 53” N, 79o 48’ 24” W; 294 m asl), town ofTemagami. These cores at the predicted northeast palaeo-outlet, show the transition from a deep proglacial lake ofvarved clays to ancestral Lake Temagami (Baseball Bog)and an isolated pond (Jessie Fen). Peatlands develop asthe lake continues to recede.


(CgGw-4) east of the Ontario Northland RailwayStation.

Designated the Portage site (CgGw-4), thisisolated find spot is in an open grassy area. An 8cm long, bipolar decortication flake (SusanJamieson, personal communication 1991) of adark brown metasediment was found in erodingsoil at the edge of a flat bedrock outcrop. Thistype of raw material is similar to the mudrocktools assigned to the Archaic period at the ThreePines site, to lithic debitage from the loweststratum of the Montreal River site dated at 5,800years ago, and to a greywacke flake cluster at theFoxie Otter site, dated to 8,500 years ago (Hanks1988; Knight 1977). Dating by raw material aloneis of course tenuous. This stretch of land betweenLake Temagami 300 m to the southwest andCassels Lake 600 m to the northeast wasmaintained by the Teme-Augama Anishnabai as awinter trail (MacDonald 1993) and may haveserved a similar function in the past. However,when ancestral Lake Temagami was at the 300 masl contour, this location would have consisted oflevel terrain, 40 m inland from the water, on thesouthwest corner of an island; commoncharacteristics of precontact sites on LakeTemagami (Figure 34).

Summary of Palaeo-hydrologyEvidence of lake transgression in the Central Hub,from the mid-Holocene to the present, led us toexamine the palaeo-hydrology of Lake Temagamias a whole. Differential post-glacial isostaticrebound was one possible explanation for lakelevel rise, with the newly deglaciated north end ofthis 46 km long lake basin, rebounding faster thanthe south end. We developed a predictive model ofpalaeo-shoreline location by computer digitizingaltitudes in the Temagami basin from topographicand bathymetric maps available in 1991. Weapplied published isostatic rebound curves fromthe adjacent Timiskaming basin to tilt thelandscape in order to identify likely former sillsand outlets and thus former water levels andpalaeo-shorelines. The model predicted amaximum 30 m relative vertical displacement,resulting in three separate lakes, draining northand northeast through two palaeo-outlets.

The model results appear to match well withsedimentary and geomorphological evidence (seeTable 18). Palaeo-shorelines for ancestral LakeTemagami found at Sharp Rock Inlet and at theNortheast Arm are consistent with the highestwater levels predicted by the model.Lithostratigraphy of pollen cores from the BearBog (N) and Baseball Bog (NE) shows deepproglacial lake inundation, findings in accord withVeillette’s (1988) reconstruction of post-glaciallakes Post-Algonquin and Barlow in theTimiskaming basin. Shorelines for these ancientlakes have little or no correspondence withmodern lakeshores. Their altitudes are estimatedfrom 310 m asl to 339 m asl, or 17 to 46 m abovethe current elevation of Lake Temagami.

The establishment of ancestral LakeTemagami dates to 10,900 years ago in thenortheast lake basin, with a multi-channel outletinto the Timiskaming basin, and to 9,800 yearsago in the north lake basin, with a high-energyspillway out through Sharp Rock Inlet. Associatedpalaeo-shorelines were observed as far inland as400 m (Bear Bog), ranging from 2 to 9 m abovethe modern lake depending on slope. As reboundcontinued, the ancestral lakes regressed from theseshorelines, causing lake transgression in other partsof the basins, eventually forming one large lakebody with a southern outlet. While climaticchange, as evidenced by the pollen cores from BearBog and Three Pines Bog, may have caused somelake level fluctuation, we suggest that it did nothave as strong an impact on the Temagami basinas did isostasy.

Our research into reconstructing the palaeo-hydrology of Lake Temagami has severalimplications for precontact site location. Assumingthat early colonizers preferred lakeshoreencampments, early Holocene habitation siteswould be distributed in substantially differentlandscape locations than are likely to be found byarchaeological surveys along modern lake and rivershores. However, archaeologists need not wait forgeomorphologists to define palaeo-shorelines. Thispredictive model reduces the potential search areato those locales with a higher probability foryielding early sites, that is, emergent and ofteninland palaeo-shorelines. This method could easily


be applied to other lakes. Lake Temagami todayhas a shallow sill across the Central Hub and drainsboth to the north and the south. In earlier times,this central sill may have divided the lake intoseparate basins, draining exclusively to the norththrough two palaeo-outlets. With modern digitalmodelling software and digital elevation maps,greater accuracy can be achieved than was possiblewith the topographic base maps used here.

Knowing which areas have submergingshorelines also informs modern shoreline surveysand archaeological interpretation (Loewen et al.2005). Timing archaeological survey to periods ofnaturally lower water (April) or co-ordinating withhydroelectric companies might be feasible. Certainlandforms are differentially affected by laketransgression. A substantial area of the low-elevation Three Pines site has been inundated. Incontrast, rising lake levels may have had a lessdestructive impact on theWitch Point site, locatedabove a steep slope. Furthermore, rising waterlevels would make different landscape featuresmore attractive for occupation at different timeintervals. Reconstructing the local effects ofchanging water levels can add a rich complexityand time depth to our reconstructions ofprecontact land use. Altered drainage patterns,travel routes, portage locations, flooded bedrockresources, changing wetlands, and lake depthswould all have impacted subsistence, settlement,and other procurement strategies of precontactinhabitants.

Vegetation and Climatic Historyof Lake Temagami1

Analyses of the pollen cores from Bear Bog (Figure30) and Three Pines Bog (Figure 24) togetherprovide the entire vegetation history of LakeTemagami from soon after deglaciation to thepresent (Table 19). These correlate well with otherpollen cores (Figure 26) from Nina Lake 110 kmsouthwest; Jack Lake 130 km northwest (Liu1990; Liu and Lam 1985); and “Lake 306,”located near Bob Lake, 7 km west of Bear Bog(Hall et al. 1994).

Zone 1 represents an early open Boreal Forest,established soon after glacial retreat. It isdominated by white spruce with balsam fir, aspen,and tamarack, perhaps along with more deciduoushardwoods (oak, elm) than the modern BorealForest growing on the newly formed rich soils (Liu1990). The nearby ice front created a periglacialclimate, with warmer, drier winters and cooler,windier summers than the climate of today’sBoreal Forest (Hall et al. 1994; Liu 1990).Charcoal data from “Lake 306” suggest a lowincidence of forest fire in this patchy, open forest(Hall et al. 1994).

In Zone 2, jack pine increases while sprucedecreases, reflecting a closed Boreal Forest muchlike that around Jack Lake today (Liu 1990).Along with jack pine, spruce, birch, poplar, andalder dominate. In Zone 3a, a strong rise in whitepine pollen, along with oak and elm, signals thereplacement of the Boreal Forest by a Great Lakes–

Table 19. Regional vegetation and climatic history.

PollenZone

443b

3a

21

Nina Lake, Sudbury,Date (cal B.P.)(Liu 1990)

8,200–4,500

10,000–8,20011,000–10,000

Date (cal B.P.), ThreePines Bog (TP)/Bear Bog (BB)

100–0 (TP)1,200–100 (TP)3,800–1,200 (TP)

8,200–3,800 (TP)>6,300 (BB)

>9,800 (BB)

12,000

Forest Composition

Ragweed/birch from cleared landMixed Conifer–Hardwood ForestBoreal Forest/Mixed Forestecotone stabilizes north ofTemagami (decrease in white pine)Great Lakes–St. Lawrence Forestspreads 140 km farther north(increased white pine)Closed Boreal Forest (jack pine)Open Early Boreal Forest(white spruce)Tundra?

Climate

Modified continentalCooling continuesNeoglacial cooling

Hypsithermal

Warming climatePeriglacial climate

Ice-free


St. Lawrence Forest. In this Hypsithermal Period,with temperatures 1 to 2 degrees higher thanpresent, the Boreal Forest–Mixed Forest ecotonemigrated 140 km north to the edge of theCanadian Shield (Liu 1990).

The onset of Neoglacial cooling is indicatedby Zone 3b, with declining white pine, increasingjack and red pine, birch, and slightly increasingspruce, as seen in the Three Pines Bog. This zonereflects the southward expansion of boreal forestspecies. Compared with Jack Lake, the ThreePines Bog shows higher white pine and lowerjack/red pine values in the upper levels of Zone3b, supporting a stabilized Boreal Forest–MixedForest ecotone, which is now 50 km north of LakeTemagami (Liu 1990).

In the Three Pines Bog Zone 4, white pineslightly decreases, with an increase in spruce andfir. A Mixed Conifer–Hardwood Forest profilecharacterizes this zone. Increased ragweed andbirch, seen near the top of the core and in the“Lake 306” core, are indicative of humandisturbance of the vegetation. These colonizingspecies reflect forest clearances for farming in theLittle Clay Belt and for mining and logging on theupland Canadian Shield since the mid-1800s(Hall et al. 1994; Hodgins and Benidickson 1989;Mitchell 1977).

Thus, soon after deglaciation at LakeTemagami (12,000 B.P.), an early open BorealWoodland was established, followed by a closedBoreal Forest. Climate warming in theHypsithermal allowed white pine to expand,creating a mixed conifer–hardwood forest withmore white pine than today (Liu 1990).Neoglacial cooling introduced more boreal forestspecies into this mixed forest. Cooling continuedto the present, with white pine graduallydecreasing while white spruce, fir, and birchincreased. Lake Temagami remains at the northernedge of the Great Lakes–St. Lawrence Forest today.

Surveyed Sites and Settlement Patterns

Precontact, historic, and traditional use sites werefound by survey in the Central Hub, the northend, the northeast end, and the south end of LakeTemagami. Several previously registered sites were

also re-examined. This section broadens theresearch to the next level of analysis. Previoussections dealt with artifacts in their stratigraphiccontext and a single site in its local context, or sitemicroenvironment, to use Butzer’s (1982)terminology. The landscape context of multiplesites, or sites’ mesoenvironment, is now the focus.

To understand hunter-gatherers’ seasonaleconomic round, ethnographic information ispresented first. Then, the landscape context ofsurveyed sites with precontact components areexamined to determine what site selection criteriamay be operating (Gordon 1990a). Data are alsocompared with site characteristics at NorthCaribou Lake, a closed boreal forest lake innorthwestern Ontario (Gordon 1985, 1988a,1988b) and Lac Caniapiscau, an open borealforest lake in north-central Québec (Hanks 1983).Criteria such as southern exposure, protectionfrom cold winds and storm tracks, well-drainedsoil, and accessible shorelines are clearly preferred.The differential impact of lake level change onprecontact Lake Temagami sites from the northend (lake regression) is compared with those fromthe south end (lake transgression).

Many precontact locales continued to beoccupied in the Historic period (pre-twentiethcentury). Archaeological findings are comparedand contrasted with campsite and trailinformation in Macdonald’s (1993) HistoricalMap of Temagami, an important ethnographicresource for understanding the relationshipsbetween habitation sites and their connectingnetworks of open water and winter trails. Thelandscape context of several twentieth-centurytraditional use sites is also discussed.

Archaeological Survey LocationsArchaeological survey between 1985 and 1994documented a wide variety of precontact,nineteenth-century, and traditional use sites(Gordon 1986, 1987, 1989, 1990a, 1990b,1991a, 1992, 1995a; Gordon and McAndrews1992). In 1985, a Cessna overflight was takenaround the entire lake. Landscape features such asprotected bays, points of land, sand beaches, andriver outlets were noted on topographic maps andphotographed as potential precontact site areas.


Raised water levels at Cross Lake due to thetwentieth-century dam appeared to extend intothe South Arm. For this reason, surveyconcentrated on the Central Hub of LakeTemagami, particularly the undevelopedmainland, as many islands contained cottages andrecreational camps.

In 1986 and 1987, modern shorelines to thenorth and south of the Three Pines andWitch Pointsites were surveyed by boat, to compare sitelandscape and usage with these large, multi-component sites (Figure 35). The focus was onlakeside locales with enough flat ground to pitch atent. These were surface-examined and test pittedusing 0.4m squares. Precontact habitation sites werefound at six locations with lithic scatters (CfHa-31,-33, -35, -36, -37, -38). A vein quartz quarry (CfHa-34) and an associated lithic workshop (CfHa-32)were also identified. The three previously registered

sites that were re-examined include the Sand Pointsite (CgHa-1), the High Rock Island site (CfHa-9),and the Lake Temagami site at Ferguson Bay(Carscallen 1994b; Conway 1986; Gordon 1986,1987, 1989, 1990a, 1991a).

Brief inland survey was conducted in 1991 atthe north and northeast ends of the lake, inconjunction with the palaeo-hydrologicalfieldwork described above (Figures 29 and 32).Survey was too limited in scope to uncover cleararchaeological evidence of early palaeo-shorelineoccupations. However, at the north end, three newlithic sites (CgHa-33, -34, -35), one potentialbedrock quarry, and an early twentieth-centuryhotel (CgHa-35) were found at the lakeside.Farther inland, two traditional use sites (ChHa-3and 4) were recorded. During contract work alongthe TransCanada Pipeline, a sawmill site (CgGw-2) at Owaissa, north of Temagami; a precontact

Figure 34. Northeast End of Lake Temagami showing peat coring locations at Jessie Fen and Baseball Bog in the townof Temagami. (Upper Left) Modern lake configuration and (Bottom Right) reconstructed ancestral Lake Temagami witha palaeoshoreline at 300 m asl . While this multi-channelled palaeo-outlet drained northeast, the Portage site wouldhave been on an island.


find spot within town (CgGw-4); and a traditionaluse site (CgGw-6) west of Caribou Lake weredocumented (Gordon 1992; Gordon andMcAndrews 1992). One benefit of inland researchwas encountering many twentieth-centurytraditional use sites.

To gain a better understanding of thetraditional use sites, boat surveys in 1994 withBill Twain focused on known pictographs andpreviously registered nineteenth- and twentieth-century sites. A new vein quartz quarry site(CfHa-50) and nine pictograph sites (CfHa-3,

CfHa-6, -7, -8, CgHa-8, -9, -10, -21, -28) werere-examined (see Zawadzka, this volume). We alsovisited two historic cemetery sites (CfHa-16 andCfHa-30), the 1834 Hudson’s Bay CompanyOutpost (CfHa-1) on Temagami Island, a Kattfamily camp with precontact lithics (CgHa-24),and the large nineteenth-century village at AustinBay (CfHa-28). Bill’s stories and historicalinformation about other Teme-Augama Anishnabaisites were also recorded (Gordon 1995a).

Figure 35. Lake Temagami Central Hub archaeological site locations. Inset shows South Arm. Hatched lines indicate1985-1987 survey areas.


West-Central Hub SitesOn the west mainland, 13 km of shoreline wassurveyed by boat from the forest fire–denudedNorthwest Arm, north of the Three Pines site, tothe entrance to the Southwest Arm (Figure 35).The Sand Point site (CgHa-1) lies on a low pointwith a wide sand and gravel beach, east of theThree Pines site. This landform is a recurved sandspit that is building out into the deep channel atthe entrance to the Northwest Arm. The 40 by 30m clearing is used as a canoers’ campsite. A total of14.5 square metres was excavated by Conway(1986), resulting in the recovery of 57 lithic tools,818 lithic debitage, 1 copper knife, and sherdsfrom 3 pottery vessels. Two 3.0 m long by 0.50 mdeep hearth features were found, containing fire-cracked rock and fire-reddened sand. Calcinedfaunal bone was adjacent to but not in thesefeatures. Raw material usage is predominantlyHBL chert, other cherts, and quartz. Based on thepottery designs, a cord-wrapped stick vessel, ahorizontal push-pull vessel, and low-collared vesselwith trailed horizontals over obliques, Conwayassigns most of this collection to the LateWoodland period. Two large, rectilinear greywacketools found at the exposed tip of the point areassigned to the Archaic period (Conway 1986).This site has been subject to the same longshoredrift and lake transgression as the Three Pines site500 m west, resulting in potentially inundatedArchaic period components.

Farther south, the Daily site (CfHa-33) is a 3m high rocky promontory overlooking a verysmall shoreline indentation. Modern camping useis indicated by a stone fireplace and lumber left bycottagers for canoe campers to burn. One small12 by 14 m area is flat enough for pitching a tent.Ten test pits yielded only two HBL chert retouchflakes.

Bedrock cliffs rise steeply from the shore southof the Daily site until a small bay is reached. TheArgillite site (CfHa-31) lies on a small point ofland at the south end of this bay. The point isnotable for a natural, canoe-sized slot in itssmooth bedrock margin. Modern camping has leftthe shallow soil denuded of vegetation. Test pits(11) produced three flakes and modern refuse. Inthe northwest corner of the bay is an old logging

road. A logging boom can be seen in a 1977 aerialphotograph.

At the entrance to the Southwest Arm, fourquartz veins were examined (Simony 1964). Threewere vertical, difficult to reach, and lesshomogenous than the Crystal site (CfHa-34).This quarry site is a 20 m long vein of milky whitequartz extending horizontally into a V-shapedembayment. The exposed portion is 1.4 m abovethe water and 0.75 m wide, while the deepunderwater portion is wider. Hammering scars areevident above water, in contrast to the smoothsurface below. With lower lake levels in the past,much more of this vein would have been exposed.Quartz shatter is evident all around and on a smallisland opposite.

The Blueberry site (CfHa-32) is an associatedlithic workshop 200 m south. Large chunks offine-grained vein quartz debitage, flakes, and chertflakes were found on the surface and in test pits(14). This modern campsite on a south-facingpoint gets blasted by the prevailing southwestwinds funnelled up the long, open SouthwestArm. Angular bedrock encircles the point, makingaccess difficult from the water.

The Cattle site (CfHa-36) is the only islandsite with deep sand deposits that was test pitted. Itis a large, grassy expanse strewn with boulders ona promontory with steep rocky shorelines. Onesmall cobble beach on the south side offers accessto the site. Test pits (10) revealed only one plainpottery sherd and one flake. Local informants saidthat cattle had once been pastured here.

East-Central Hub SitesAlong the east mainland (Figure 35), four km ofshoreline was examined, from 1 km north of theWitch Point site to Mule Bay (Gordon 1986,1987). Sealrock Point at Granny Bay and theKokoko Bay entrance were also examined(Gordon 1995a). A stretch of 11 km fromMatagama Point into Cross Bay and down toPelican Point was also briefly examined (Gordon1989, 1995).

Sealrock Point (CgHa-24) is a previouslyregistered site. Here a 100 m x 50 m point of landjuts into Granny Bay. It has a rocky shoreline andthin deposits of soil, with jack pine near the point


and white pine farther inland. Vein quartz flakeswere found in the north campers’ clearing, with achert scraper in the south clearing. According toBill Twain, this point was formerly called SquirrelPoint and is a known stopover camp for the Kattfamily (Gordon 1995a). The Katt family lived onDiamond Lake, to the north (Pollock 1992), andwould gather here when they came down to LakeTemagami.

According to Bill Twain, people wouldfrequently stop at Sealrock Point, which lies at theintersection of the open west channel (Granny Bayand Devil’s Bay) and the North Arm of LakeTemagami. A traveller could easily get caught bywinds from any direction. He noted that peoplewould live on this point especially during flyseason, as it gets a lot of wind. In his terminologySealrock was not a “long time camp,” not a“permanent camp,” just a stopover for a few days(Gordon 1995a).

TheWitch Bay Sugarbush site (CgHa-18) wasregistered by Conway as a maple sugarbush site,based on information from Bear Island residents(Ontario Ministry of Tourism, Culture and Sport,Site Record Form). When the site was visited in1986, probable remnants of a square woodenstructure consisting of lumber and tar paper werenoted. Near the shore were metal and glassartifacts and unburned moose bones. Thestructure may have served as a temporarycookhouse, but could also be a winter trap cabin.Revisited in 1994, the cabin remains were eitherhidden under new growth or washed out andredeposited by storms.

The Kokoko Bay quarry site (CfHa-50) is alarge, horizontal vein of milky quartz. It extendsfor 15 m along the east shore of an island at theentrance to Kokoko Bay. The exposed portion is 2m high, with another 1 m extending below thewater. Cedar and jack pine grow in shallow soilatop the vein. Chunks and chips of all sizes arefound in the water (Gordon 1995a). It is a good-quality, homogenous material, unlike the narrowveins in the pink Archean granite on the oppositemainland shore (Simony 1964). This site may bethe primary source for the milky vein quartz flakesand core fragments found at the Witch Point site.

The Island 245 site (CfHa-35) is on a very

small, exposed bedrock island with a few pocketsof soil and trees. It lies in the channel betweenTemagami Island and the east mainland. Threetest pits yielded two flakes. A side-notched greychert projectile point lay on the bedrock surfaceat the north end.

The Cross Bay site (CfHa-37) occupies arelatively flat clearing beside an expanse of smoothbedrock on the north side of a narrow channelconnecting the Central Hub to Cross Bay. The siteis frequented by campers and lacks both groundcover and a humus horizon. Most of the artifactrecoveries were surface finds. The eight tools aretwo chert biface fragments, four chert scraperfragments, a vein quartz side-notched point base,and a vein quartz scraper fragment. Of the 293lithic debitage, 60 percent was vein quartz,followed by chert, mudrock, and HBL chert. Ahigh proportion of vein quartz debitage is due inpart to a large rock with a quartz inclusionuncovered in one of two test pits. One ball claypipe stem was also recovered. This channel wouldoffer a strategic place for fishing. The site also hasa commanding view of both Cross Bay and LakeTemagami.

The Boulder site (CfHa-38) is a small pointof smooth bedrock forming the west side of thenarrow channel into northern Cross Bay. It is asmall modern campers’ clearing, 26 m by 22 m.Two of five test pits yielded 4 broken tools, 17debitage of mostly vein quartz, and 5 calcinedfaunal bone. Vein quartz tools include two bifacefragments and a scraper fragment, along with anHBL chert scarper fragment.

Another small lithic site in Cross Bay is theHigh Rock Island site (CfHa-9), at the base of thelookout path. This modern campers’ site is easilyaccessible because a smooth rock ledge slopes intothe lake. It is well protected by a small islanddirectly opposite, occupying the narrow southernchannel between Cross Bay and the open water ofLake Temagami. Surface collecting produced avein quartz biface fragment, a piece of vein quartzshatter, and a chert flake.

The lookout path climbs a steep NipissingDiabase dyke at the northeast tip of High RockIsland. It offers a commanding view of the CentralHub of Lake Temagami. Along the path and at the


top are numerous depressions and rocks, identifiedby Conway based on informant information ashuman-made structures for defensive purposes(Ontario Ministry of Tourism, Culture and Sport,Archaeological Site Record Form). Bill Twain saidthe lookout point served as a watchtower. Thefollowing paraphrases his words:

Before the HBC outpost was built in1834, Temagami people used to camp atWabikon on the southeast corner ofTemagami Island. High Rock Island isdirectly southeast across the channel.When Iroquois raiders came up toTemagami, the Temagami People usedsmoke signals to tell if the Iroquois werecoming from the south or north.Ottawa and Montreal River Peoplewould say if the Iroquois were comingand warn the Temagami People. So theTemagami People posted lookouts onHigh Rock Island [Gordon 1995a:17].

Temagami Island is the largest island in theCentral Hub. The Wabikon site (CfHa-13) lieswithin the privately-owned wilderness camp of thesame name, where Conway observed severalprecontact hearths and a few slate flakes. Based onan oral history from Bill Twain, collected byConway in July 1984, Wabikon is described as a“summer Indian village pre-1850, an historic andprecontact Temagami Indian village (OntarioMinistry of Tourism, Culture and Sport, SiteRecord Form).

The Temagami Island Post (CfHa-1) lies justsoutheast of Wabikon. At the head of a small bayare two rectilinear depressions lined with cobbles,but no timbers. One measures 2 by 3 m. Fartherinland are the caved-in remains of walk-in coldstorage cellar dug into the sandy hill. The exteriordimension is 6 m by 4 m. The interior walls arelined with vertical squared timbers, 8 inches (20.3cm) wide. The structure is about 2 m high fromthe sand floor to the roof, made of 2 inch by 6inch (5.1 x 15.2 cm) planks. For insulation, a deeplayer of sand covered the roof, atop which a livewhite pine is now growing (Gordon 1995a).

These structural remains may be the 1834outpost or part of its later expansion in 1857. The

Hudson’s Bay Company opened an outpost onLake Temagami in September 1834. “The‘Temagamingue’ post was thus established on thesouth side to Katay-Teme-augama, ‘Old TemagamiIsland’ just beyond the ancient Wabikon, thesummer ‘Flower’ settlement of the Temagamis”(Hodgins and Benidickson 1989:30). It wasoperated intermittently between 1834 and 1857.

[T]he Temagami outpost was suppliedfrom Timiskaming and open only fromautumn to spring, with three or fourhands hired to transport the outfit, totend the potato garden and to make thefall fishery […] Indians from Temagamicontinued to trade at neighbouringHBC posts including FortTimiskaming, Nipissing andMatawagamingue. They also dealt withopposition traders [Hodgins andBenidickson 1989:30].

In 1876 a new HBC post (CfHa-2) was built atBear Island (Hodgins and Benidickson 1989;Mitchell 1977).

The large Temagami Island Indian BurialGround (CfHa-30) is located atop an 8 to 9 mhigh sandy hill, southeast of the HBC outpost.Posts and rails from a split rail fence and a largefallen white cross marks its boundaries. These wereerected by a former owner of Wabikon whomaintained this cemetery. Two weatheredheadstones with embossed inscriptions for LouisEgwina, 1891, and Peter Keewagano, 1892, facesouth–southwest across the bay. The east-facingheadstone for a child, George Paul, 1889, iserected within a wooden picket fence enclosure.Other graves have fallen enclosures, have blue-painted crosses, or are unmarked. Bill Twain saidthat he had records and genealogies for all themarked graves (Gordon 1995a).

High Island Cemetery (CfHa-16) wasidentified by Bill Twain as several unmarkedgraves on the east central side of a sand and gravelisland, south of Bear Island. Only one possibledepression was noted in this flat area of denseunderbrush in a grove of tall poplars. It is locatedabout 12 m above the lake. Several paths leaduphill from a lakeside clearing on the north end of


the island toward the cemetery. To paraphrase BillTwain:

People from the Southwest Arm of LakeTemagami came up to Bear Island andburied their dead on the east side of thisisland. They chose a place with a view ofthe rising sun. The graves were just pitsdug in the sandy ground. People wouldcut trees down for a burial area, but theydid not erect grave markers. They didnot maintain the graves but just “leftthem alone” [Gordon 1995a:9].

South End SiteThe Austin Bay site (CfHa-28) includes thestructural and artifactual remains of a largenineteenth- to early twentieth-century village ofthe Teme-Augama Anishnabai (Gordon 1995a;Pollock 1992; Pollock and Koistinen 1993). Itoccupies the mainland peninsula and secludedwestern bay between Austin Bay and the SouthArm (Figure 35). Among its above-ground andsemi-subterranean structures are remnants ofdovetail-jointed log cabins, a barn for a horse andcow, cold storage structures, and root cellars.There are also trails and cleared land for potatogardens. Large artifacts include the remains of aCamp Matagami barge, a more recent cedar stripcanoe, a wooden sled with screw-fastened steel

runners, a collection of old vehicles, heating andcook stoves, metal bed frames, clear glass bottles,glazed plates, and also deer and moose bone. Thevarious cabins are difficult to locate as they weremostly burned down by the Ministry of NaturalResources in the 1970s. With Bill Twain, thearchaeology field crew visited the cabins of AlecPaul, Neass Twain (Bill’s grandfather), MichelTwain, and Donald Mckenzie, as well as“Temagami Ned’s” campsite (Figure 36). Theshrubby undergrowth on these former clearings isquite distinctive from the surrounding groves ofbirch, poplar, and pine that cover the sandy soilof the large Austin Bay site (Gordon 1995a).

Northeast End SitesThe precontact lithic find spot, the Portage site(CgGw-4), has been described earlier. A secondsite, a twentieth-century hunting camp, was alsofound inland from the Northeast Arm of LakeTemagami (Figure 32). The Caribou Cabin site(CgGw-6) is located on a narrow bedrock ridge100 m west of Caribou Lake. This log structure,2.5 m square, has seven courses of unpeeled,unnotched logs of 12 to 14 cm in diameter.Round metal spikes with a length of 6 inches(15.2 cm) are driven into the corners, and severallarge cobbles are at the interior corners.Crumbling clear plastic sheeting covers the walls,which are 56 cm high. One wall continues 2.5 msouth of the main square. A plastic vase andflowers, a two-person bow saw without a blade,and a black rubber floor mat were noted outside.I have seen similar arrangements at the familycamp of Job Halfaday at North Caribou Lake(Gordon 1985, 1988a). Job first built low logcabin walls, wrapped in plastic sheeting to keepout the wind. He then pitched a prospector’scanvas A-frame tent over top, with an entrancewayin which to store firewood and tools, and thesleeping quarters at the far interior of the tent.

North End SitesIn 1991, in conjunction with palaeo-shorelineresearch, selected locales at the north end wereexamined by boat, including Sharp Rock Inlet,Bear Bog, Deer Island, and Whitefish Bay, whileBarnac Lake was accessed by logging road

Figure 36.William (Bill) Twain, former Chief andTeme-Augama Anishnabai Elder, took the 1994archaeology crew to visit many traditional use sites,including these log cabin remains at the Austin Bay siteon the South Arm of Lake Temagami.


(Gordon 1992). Three new lakeside precontactsites were discovered (Figure 29). Mayhue Rock(CgHa-33) is a low, grassy promontory at thenortheast entrance to Sharp Rock Inlet. It is afavoured stopover for canoers accessing the LadyEvelyn-Smoothwater Wilderness Park viaDiamond Lake to the north. Test pits (6) revealeda 7 cm thick layer of fire-reddened sand with twochert flakes and one calcined bone, near surfacefinds of one chert flake and three vein quartzshatter fragments.

The Deer Island pictographs (CgHa-23) andthe South Deer site (CgHa-34) have beendescribed above. At the northeast corner of DeerIsland are the melted and twisted remains of theLady Evelyn Hotel (CgHa-35), a three-and-a-halfstorey, 108-bed tourist hotel built in 1904, whichburnt down in 1912 (Hodgins and Benidickson1989:119-121). Just northeast of the hotel, ascatter of HBL chert and vein quartz flakes werenoted in the shallow soil beside a rocky point.

One possible lithic tool source lies in the farnorthwest corner of Whitefish Bay. It is a verticalrock face of Gowganda Formation, with a stratumof greywacke (conglomerate) over a sub-conchoidalfracture, medium grey siltstone. Asquarrying marks were not evident, it was notregistered as a site. However, this is the onlyflakeable siltstone I have seen at the lake edge, andgrey siltstone debitage does occur at the ThreePines site.

Two inland hunting camps of the Teme-Augama Anishnabai were recorded (Gordon1992). The Stove site (ChHa-3) is on a flat terracebelow Bear Bog, containing a collection of olderitems, such as a rectangular wood-burning stove,an oil lamp base, an old jug, and a rusted steelwash basin. Similar artifacts were found inland ata 1939 winter base camp on North Caribou Lake(Gordon 1985, 1988a) and at older winter travelcamps in James Bay (Chism 1978; Gordon 1980).The Cache site (ChHa-4) is located 40 m east ofBarnac Lake on the same sandy outwash plainseen at Ferguson Bay. Here a semi-subterraneanwood structure was found. It is 1.5 by 2.5 m insize, built of slender vertical poles for the walls,roof, and floor, topped by a 30 cm high moundof dirt. This may have been a winter cache or

small ice house associated with an older huntingcamp (Gordon 1992).

The Lake Temagami site is located at themouth of the Anima Nipissing River, at thenortheast corner of Ferguson Bay. This part of theoutwash plain and sandy beach lies within thegrounds of Camp Wanapitei, a wilderness camp.In the nineteenth century this was used as asummer gathering area by the Teme-AugamaAnishnabai, and it is here that Father Paradis builthis Mission du Sacre Coeur in 1891 (Hodgins1976; Hodgins and Benidickson 1989). Remainsof the log chapel were visible in 1986, east of theriver, while nine surface-collected areas within thecamp area produced vein quartz debitage (Gordon1986, 1987). Subsequent excavations byCarscallen (1994a, 1994b) show a large, multi-component site with four discrete clusters of lithicartifacts of quartz, rhyolite, and greywacke over40,000 m2, as well as pottery and a dog burial.

Ethnographic Studieson the Seasonal Economic RoundTo help interpret these surveyed sites, it is worthsummarizing anthropological information on theseasonal economic round. Ethnographic andethnohistorical studies of Cree and Ojibwapeoples in northern Ontario and Québec (Jenkins1939; Rogers 1962, 1963a, 1963b, 1966, 1969,1973; Rogers and Black 1976; Speck 1915a,1915b; Tanner 1979) describe techniques ofhunting, trapping, and gathering seasonallyavailable food resources. From fall to spring, small,kin-based groups periodically moved to differentcamping locations to exploit the fish, mammal,bird, and plant resources as these became available.In summer, when food was more accessible, smallgroups would assemble into larger gatherings forseveral weeks to socialize, trade, and prepare againfor the long winter. Ethnographic studiesemphasize the importance of “the ecotoneoccurring along the boundary between land andwater” (Rogers and Black 1976:5) in choosingcamp locations, for hunter-gatherer mobility,communication, resource procurement,subsistence, and habitation.

Open water in summer, lake ice in winter, andthe intermediate “no travel” periods affect modes


and methods of transportation, as well as huntingand fishing activities (Jenkins 1939; Rogers 1962,1973; Rogers and Black 1976; Tanner 1979). Aready supply of firewood for heating and cooking;constructional resources, such as tent poles,saplings, and spruce roots; as well as access to fish,a dependable food staple, are additional factors inchoosing camp locations (Irimoto 1980; Rogers1966, 1973; Tanner 1979). These near-campresources are short term and quickly exhausted,resulting in frequent moves (Rogers and Black1976; Tanner 1979).

Previous Studies onBoreal Forest Settlement PatternsHow does this seasonal economic round translateinto site selection criteria? Are there specificlandscape features which influence the choice ofone location over another? Certain criteria haveemerged from ethno-archaeological research at LacWashadimi in northwestern Québec (Chism1978; Gordon 1980; Tanner 1978a, 1978b) andNorth Caribou Lake in northwestern Ontario(Gordon 1985, 1988a, 1988b). At North CaribouLake, a closed Boreal Forest environment, severalkey factors underlie site selection. Thisinformation is based on discussions with JobHalfaday of Weagamow (Round Lake) and acomparison of 13 traditional use and 10precontact sites. Locations protected from coldnorth and northwest winds were chosen in bothsummer and winter, in particular those withsunny, southern direct exposures. Wind exposurewas particularly important in the bug-infestedwarm months, with a preference for islands.Islands also offered greater protection from forestfires and bears and greater access to fish resourcesin warmer, open water seasons. Flat ground withwell-drained soil, particularly sand, was preferredfor both winter and summer sites. Winter sites,from freeze-up to break-up, tended to be on themainland, from 5 to 50 m inland. For shorelineaccess, sand beaches or low slopes were preferred,while deep water was avoided as being toodangerous for children (Gordon 1985, 1988a,1988b).

These criteria mirror many of those noted fornorth-central Québec, based on information from

Sam and Williams Rabbitskin of Mistassini and131 surveyed sites, including traditional use sitesand 19 precontact sites (Hanks 1983). In the LacCaniapiscau open Boreal Forest, campsites wereselected for comfort and survival. Protection fromwest and northwest winds was afforded by densestands of trees, especially in winter, rather thanridges or other landforms. Year-round proximityto major bodies of water was emphasized as netfishing provided the most dependable food sourceeven in winter (Hanks 1983:352). Locales withsoutherly exposure and those mostly free of surfaceboulders and cobbles were favoured. Precontactlithic sites averaged 30 m inland and 3 m abovewater compared with 19 to 22 m inland and 1.2m above water for traditional use sites (Hanks1983).

Precontact Habitation SitesThe Lake Temagami precontact sites found or re-examined in this research are categorized byfunction and size, depending on amount of levelground and artifact recoveries (Gordon 1990a).Categories include vein quartz quarry (2), lithicreduction workshop/habitation site (1), very smallcamp (3); small to medium habitation camp (8),large habitation camp (3), and very largehabitation camp (1). Tables 20 and 21 presentdirect wind exposure, soil type, and shorelineaccess for 18 lakeside sites, of which 12 arehabitation sites in the Central Hub and 4 arehabitation sites at the north end.

Wind Exposure. Direct wind exposure means overopen water (or lake ice) unblocked by islands orlandforms. Twelve of 16 sites have southerly windexposure. Two are islands directly opposite otherbodies of land and have no direct wind exposure.However South Deer Island (CgHa-34) facessouth, while and High Rock Island (CfHa-9) faceseast. Only three sites have direct wind exposure tothe north or northwest. These are Sand Point(CgHa-1), oriented east–west at the entrance tothe Northwest Arm; Sealrock Point (CgHa-24);on a north-facing point in Granny Bay; and thelithic scatter at the Lady Evelyn Hotel site (CgHa-35), on the northeast point of Deer Island. Twoalso have southeastern exposures. Sealrock Point


is the only site with a north-facing exposure. Testpits were dug at two other mainland locales withdirect north wind exposure: the northwest cornerof the Sand Point peninsula and the sandybaymouth bar at the northeast corner of WitchPoint, both without results. Southern exposureswere probably favoured year-round for warmthand dryness, while locales with more wind wouldbe more desired in the bug season. It is unlikely

that locales with little or no direct wind would beearly summer sites.

Soil Type. Five of the 16 sites are located on deepsand, while the rest are shallow silty sand overbedrock. As described previously, glaciofluvialdeposits, such as baymouth bars, outwash plains,or eskers, are relatively rare in this ruggedlandscape. Thin till over bedrock is much more

Table 20. Landscape characteristics of precontact habitation sites, Central Hub, Lake Temagami.

Site Type, Name,and BordenNumber

Sand Point CgHa-1

Three Pines CgHa-6

Witch Point CgHa-7

Argillite CfHa-31

Daily CfHa-33

Island 245 CfHa-35

Sealrock PointCgHa-24

High Rock IslandCfHa-9

Cattle IslandCfHa-36

Cross Bay CfHa-37

Boulder CfHa-38

BlueberryCfHa-32

Crystal CfHa-34

Kokoko BayCfHa-50

Large habitation (excavated, multi-component)

Very small camp (2–11 lithics)

Small to medium camp ( 1–350 lithics)

Lithic workshop/habitation (530 lithics)

Quartz veins /quarry site (50 lithics)

Soil

deep sand

deep sand

deep sand

shallow sand overbedrock

shallow silt overbedrockpromontory

shallow sand overbedrock

silty sand overbedrock

shallow silty sandover bedrock

deep sand

shallow silty sandover bedrock

shallow fine sandover bedrock

shallow silt overbedrock

shallow fine sandover bedrock onadjacent island

lakeside quartzvein

DirectWind

Exposure

NW; S–SE

S–SE

W–SW–S

NE; SE

NE

S

W–NW–N

none(east-facing)

S- SW

SE

SE

SW

S

E

EstimatedArtifactArea (m2)

400

520

1,120

32

1

50

90

–

75

120

3

585

10

–

Area of LevelGround(m2)

1,200

875

1,500

468

163

600

600

400

1,800

1,000

192

1,350

200

–

HeightAbove Lake

(m)

1

2

4

2

3

2

2.5

1

2.5

2.5

1.5

2.5

2.5

2

Shoreline Access

sand and cobble beach



smooth low bedrock;cobble beach

smooth low bedrock

smooth low bedrock

low bedrock

low smooth bedrock

cobble beach

low smooth bedrock;cobble beach

low smooth bedrock;cobble beach

angular bedrock

smooth low bedrock

Shoreline Access

sand beach

low smooth bedrock

low smooth bedrock

low smooth bedrock


common. Sand offers flat expanses of well-drainedsoil in all seasons, with few rocks or cobbles. Thelargest mainland habitation sites are all on sand,have all been chosen for excavation, and haveyielded high artifact densities. As noted above, twoother sandy locales with north-facing exposuresyielded no lithics; neither did the narrow, south-facing Mule Bay esker.

Shoreline Access. Related to site terrain is accessfrom water to land and land to water. Clearly it iseasier to alight from watercraft or pull up sledswhere there is a low, sloping sandy beach or sandand cobble beach. On all the shallow silt overbedrock sites, access is made easier by low-sloping,smooth bedrock. Only one site, the Blueberry site(CgHa-32), had angular bedrock at the shore. Itsuse as a processing site for vein quartz from thenearby Crystal site (CfHa-34) may have been amitigating factor.

Narrows for Fishing. Other local features of thesesites are worth noting. Five sites (Sand Point,Cross Bay, Boulder, South Deer, and High Rock)form one shore of deep, narrow bedrock channels.Such channels would serve to funnel fish, makingthese good fishing locales. Ethnographic sourcesshow that fishing is a dependable year-roundsubsistence activity, even in winter with the use ofnets. Hanks (1988) questions whether winter netfishing is a post-contact technique. Unmodifiedcobble netsinkers were noted at the Three Pinessite in a Woodland context. Shallow water fishing

with seine nets in the Middle Woodland periodand deep water fishing with improved gill nets inthe Late Woodland period are suggested byCleland (1982) for Upper Great Lakes fisheriessites. However, fishing through winter ice remainsuncertain for precontact occupants of the surveyarea.

Points of Land. Five precontact sites are on largerpoints of land jutting into the main body of thelake (Sand Point, Sealrock Point, Witch Point,Mayhue Rock, Lady Evelyn Hotel). These offernatural quays to open water travellers approachingfrom several directions, particularly useful at timeswhen bad winds or stormy weather arise suddenly.Several sites are on small points of land that alsocreate indentations or small bays, such as theArgillite site with its natural bedrock dock, theslight indentation at the Daily site and the V-shaped channel of the Blueberry site. Theselandforms offer travellers some measure ofprotection, localized sections of calmer water, andlee shore access depending on wind direction.

Summary of Site Selection FactorsThese 16 sites suggests that southern exposures,protection from cold winds and storm tracks,sufficient flat and well-drained ground, andconvenient shoreline access are landscape featuresof habitation sites deliberately chosen byprecontact occupants of Lake Temagami. Thesesame selection criteria are noted for lakes muchfarther north in the open and closed Boreal Forest.

Table 21. Landscape characteristics of precontact habitation sites, North end, Lake Temagami.

Site Type, Name,and BordenNumber

Lake TemagamiCgHa-2

Mayhue RockCgHa-33

South Deer siteCgHa-34

Lady Evelyn HotelCgHa-35

Very large habitation site (excavated, multi-component)

Small to medium camp (3–10 lithics)

Soil

deep sandyoutwash plain

fine silt overbedrock



Direct WindExposure

SW

W–SW

none(south-facing)

N–SE

EstimatedArtifactArea (m2)

40,000

100

40

–

Area of LevelGround(m2)

100,000

1,250

450

750

HeightAbove Lake

(m)

1

2

2

1


While deep sand deposits show repeated use overtime, lakeside sites may be found anywhere thereis enough flat ground to pitch a tent (e.g., Daily,Island 245). In predicting or finding sites, usingbroad locational parameters, such as shorelineorientation, may be inadequate. Very localizedfeatures are important to site location. As Hanks(1988) points out, a stand of trees is sufficientwind protection.

It is important to note that this analysis isbased on archaeological sites in the modernlandscape context. Howmight past environmentalchange have altered precontact campsite selection?Climate change may have involved changes inprevailing wind direction, but not southernexposures. A warmer, drier climate would result inlonger periods of open water and possibly changesin fishing patterns. Changes in forest compositionsuggests local and regional faunal resourcedifferences, affecting subsistence patterns but notnecessarily habitation site locational choices. Themost significant effect on archaeological sites alongthe modern shoreline would be lake level change.

Precontact Surveyed Sitesand Lake Level ChangeThe palaeo-hydrology research described aboveshows that lake level change has had a differentialeffect on Lake Temagami shorelines. The lake haspulled back from shorelines at or near the formernorthern and northeastern outlets while floodingshorelines in the central and southern sections ofthe modern lake. Dams in the twentieth centuryhave both raised and lowered lakes levels, but to afar lesser degree than isostatic rebound. Lakeregression would have impacted precontactoccupations at Mayhue Rock, South Deer Island,Lady Evelyn Hotel, and the Lake Temagami site.As noted previously, old beaches were noted at twoof these sites. It is possible that the lithics found intest pits near the lake on these sites represent themost recent precontact occupations, withpotential older ones farther inland. For example,at the very large Lake Temagami site it may bepossible to develop a relative chronology of thediscrete artifact clusters noted by Carscallen(1994a, 1994b), particularly for artifacts at thedeepest stratigraphic levels, as the area of habitable

space expanded southward.In the central and southern sections of Lake

Temagami, lake transgression since the mid-Holocene has been a factor, flooding the sillbetween once-separate lake basins in the CentralHub and encroaching on living space at SandPoint; Three Pines; and the other smaller, low-elevation sites in this part of the lake. SomeArchaic period artifacts at Three Pines and SandPoint lie closer to the current lakeshore than morerecent occupations, a situation that may also betrue for other surveyed habitation sites in theCentral Hub. Lower lake levels would haveexposed more of the lakeside veins of milky whitequartz at Crystal and Kokoko Quarry, makingthese lithic raw materials more accessible in thepast. Underwater examination may beworthwhile. Relative geochronology may alsobecome useful in dating northern versus southernrock painting sites on Lake Temagami andadjacent Obabika Lake.

Historic Period Habitation Sites and LinkagesArtifacts datable to the nineteenth century wererecovered from Three Pines, Witch Point, andCross Bay. Summer gathering sites at Wabikon onTemagami Island and Sandy Inlet on FergusonBay are known from written and oral history. Fourof these five locales are on sandy subsoil. All ofthem have southern exposures and are protectedfrom northern winds. This is also true for thesummer gathering place on the southwest cornerof Bear Island.

For the Historic (pre-1900) period, theHistorical Map of Temagami (Macdonald 1993)depicts hundreds of campsites; 600 geographicplace names used by the Anishnawbeg (Ojibway);and a vast network of 1,300 nastawagan, or travelroutes (Macdonald 1987). The map is based oninterviews conducted by Craig Macdonald with200 elders from Bear Island and other reservesover 20 years, starting in the 1970s, along withprimary source material (Macdonald 1987). Healso has traversed many of these winter trails andportage routes (Jenish 2006). The importance ofthis map is both in its content and its applicabilityto the archaeological record. Archaeological sitesmaps give the impression of separated places,


whereas this ethnographic-based map shows thelinkages, presenting campsites as “beads on astring”: a series of short stops within thecontinuous, integrated movement of hunter-gatherers over the entire landscape.

The map encompasses Lake Temagami andregion, including LakeWanapitei to the southeast,Lake Timiskaming to the east, and Lady EvelynLake to the north (see Figure 1). The nastawaganinclude travel routes along lakes, using canoes inopen water or using toboggans and snowshoes onlake ice in winter. Cross-country trails includeshorter summer portage trails across narrowstretches of land or longer portages along riversand creeks connecting large and small lakes.Summer portage trails were also used in winter.Specific winter-only trails called b o n-ka-nah wereused solely for travel by snowshoe, sleigh, ortoboggan. These winter trails were wellmaintained up until the 1940s, with the longerones showing “a skillful use of swamps andgeological faults to keep the trails direct and tominimize climbing” (Macdonald 1987:187). Themap depicts regional topography before 35 lakesand rivers in the area were flooded by moderndams. The geographical place names describeeither topographical features or food resources thatcould be obtained in specific locales (Macdonald1993; Jenish 2006).

Most precontact sites described above areincluded in the Historical Map of Temagami,demonstrating continuity in land use and siteselection criteria. On the west-Central Hub, thevery small habitation sites of Argillite and Daily, aswell as the Crystal quarry, are not included. Thereare, however, two portage trails marked, leadingwest from the small bay at Argillite: one to a smallinterior lake, the other to Gull Lake, which runsparallel to the Southwest Arm of Lake Temagami(Macdonald 1993). The open water or lake icealong steep cliffs between Argillite and Daily wasnot used as a north–south travel route. It can besubject to strong prevailing southwest winds,funnelled up the Southwest Arm.

For the east-Central Hub, Gordon (1991a)suggested that north–south travellers wouldprobably use Cross Bay and other protected baysfarther south, rather than risk the open water of

Lake Temagami. The map confirms this inference,showing no travel routes on the west, or openwater side, of High Rock Island. At the northeastend, the find spot called Portage site lies on awinter-only trail between the Northeast Arm andCassels Lake. The Caribou Cabin site lies along aportage route between the two lakes. Neither ofthese sites, one precontact, the other one twentiethcentury, are marked on this pre-1900 map.

At the north end all the precontact sites aremarked, but not the siltstone source at WhitefishBay. Mayhue Rock and environs was used toaccess an old portage route between LakeTemagami and Diamond Lake, past the rapids atSharp Rock Inlet (Macdonald 1993). Thetraditional use sites, Stove site and Cache site, bothwith indicators of cold weather usage, are notmarked on the historical map, lending supportthat they are indeed twentieth-century traditionaluse sites. The cold weather structure at the Cachesite bears constructional similarities with a deadfalltrap built of vertical posts illustrated on the map.A portage route south of the Cache site follows astream connecting Barnac Lake with FergusonBay. Numerous campsites are marked around thenorth end of Ferguson Bay, not just the LakeTemagami site near the outlet of the AnimaNipissing River (Macdonald 1993).

Traditional Use Habitation SitesTraditional use sites of the late nineteenth to earlytwentieth century encountered in these surveys arecharacterized by remnants of above-ground andin-ground wooden structures, such as cabins andbarns, and traditional and mass-manufacturedmaterial goods. Subsistence resources indicated byartifacts and informant data include moose anddeer hunting; tapping of occasional stands ofmaple trees, such as the sugarbush at Austin Bayand Witch Bay; potato gardens; and cows.Remains of canoes for open water travel and sledsfor ice and snow travel were noted. Sites rangefrom the large village at the mainland Austin Bay,with semi-permanent log cabin structures; toisolated cold weather mainland camps, such as theStove site and Cache site; to possible trap cabins atWitch Bay Sugarbush and Caribou Lake Cabin;to stopover camps, such as Sealrock Point.


Material remains were found lakeside as well as upto 40 m inland. Similar selection criteria were usedas in the precontact sites, emphasizing level, well-drained substrate, southern exposures, and windypoints in fly season, and more protected locales incolder weather (Table 22).

Cemetery SitesThe two cemetery sites, one with unmarked gravesand one with headstones or picket fenceenclosures, have similar landscape attributes tothose noted for cemeteries at North Caribou Lakein northwestern Ontario and described by JobHalfaday (Gordon 1985, 1988a). These attributesinclude a sandy substrate, a markedly higherelevation, and a good view over the lake facing eastto south, “to face the rising sun.” Grave sites aredistant from habitation areas and well above thelake in elevation (see Table 22).

While the question of dating grave stylesrequires further research, one possible sequencecan be suggested. Graves with headstones onTemagami Island (CfHa-30) date from 1889to1892. The unmarked depressions here and atHigh Island Cemetery site (CfHa-16) are likelyolder. The picket enclosures could be intermediatein age. However, wooden picket fence enclosuresare also found at Bear Island, which is a morerecent settlement. Therefore the picket fenceenclosure may be an alternative to headstones andperhaps the preferred style of the late nineteenthcentury and early twentieth century, dependingon the Bear Island grave dates. According to BillTwain, he has never heard of the use of a tiny logcabin structure (wa-ha-gen), which was used in theearly twentieth century at Pekangekum (Dunning1959: Plate 5) and North Caribou Lake innorthwestern Ontario (Gordon 1985, 1988a).

Table 22. Traditional use sites, Lake Temagami.Location, Name,and BordenNumber

Stove site ChHa-3(near Bear Bog)

Cache site ChHa-4(Barnac Lake)

Caribou CabinCgGw-6(Caribou Lake

Witch BaySugarbush CgHa-18

1834 HBC OutpostCfHa-1(Temagami Island)

High IslandCemetery CfHa-16

Temagami IslandBurial GroundCfHa-30

Austin Bay CfHa-28(19th cent. village)

North Arm

Northeast Arm

Central Hub

South Arm

DirectWind

Exposure

S

W

S

W

S

E

S

E, W

Terrain

flat silty sandterrace

sandy flatoutwash plain

flat low ridge,thin till overbedrock

low sand,shallow bay

low sand,shallow bay

high sand hill

high sand hill

sand, protectedshallow bay

Distanceinland(m)

40

40

100

10

10–20

40

15

20–50

Elevationabove Lake

(m)

4

4

2

1

3 to 4

12

9

0 to 4

Structures/Materials

Winter Camp - oil lamp base, older style metaljug and wash basin, wood-burning stove

Cold season cache - semi-subterraneanstructure, 1.5 by 2.5 m, slender vertical poles

A-frame tent base - log structure, 5 by 2.5 m,7 courses, plastic flowers, 2-person bow saw

wooden cabin structure; tar paper, brownJavex bottle, wash tub, moose vertebrae

2 cobble-lined foundations, 2 by 3 m; coldcellar 8 by 4 m

unmarked graves

1889–1892 headstones, picket fences, severalunmarked graves

log cabins, barns, root cellars, trails, gardens,canoes, sleds, old vehicles, stoves and bedframes


Witch Point Site Preliminary Results:Explaining Variation with the

Three Pines Site

To further develop an archaeological sequence forLake Temagami, a second site was selected forcontrolled excavations in 1993 and 1994 (Gordon1993, 1994a, 1994b, 1995a, 1995b). The palaeo-hydrological reconstruction of Lake Temagamihelped inform the choice of sites. It was reasonedthat the earliest precontact occupations at thenorth end of the lake would be distributed over aseries of inland shorelines, while those toward thesouth end would now be submerged. So ideally, asite in the Central Hub of the lake would be thebest candidate. The Witch Point site fit therequirements. It is 3.6 km directly east of ThreePines, on the eastern mainland shore of LakeTemagami, at the entrance to the North Arm(Figures 35 and 37). The site is located at thesouthwest tip of a 4 m high sand and gravel esker,forming a long point of land around the shallowWitch Bay. What is most interesting, and thefocus of this section, is explaining the strongstructural, temporal, and cultural differences

between these two multi-component mainlandsites.

Previous ResearchIn 1982, a two-week salvage excavation wasconducted at the Witch Point site by Conway(1982; Smith 1983). An excavation block of 45m2 oriented along the west side of the clearing wasopened (Figure 38). This work revealed a “longand intensive use” as “all areas contain virtuallyuninterrupted hearths and ash bands” (Conway1982:2), including calcined faunal bone “confinedto the hearths” (Smith 1983:1). A total of 5880cultural items was recovered, along with quantitiesof red and yellow ochre.

Conway identified five components based onartifact attributes and comparison with other siteassemblages. Slate and greywacke choppers,scrapers, and knives and two stemmed points wereassigned to a Middle Archaic component, basedon similarities with the Montreal River site lowerlevels (Knight 1977). Other slate tools, apparentlymade on local beach cobbles and including a cacheof pecked axes, were assigned to the Late Archaic,based on similarities to finds from theSmoothwater Lake site (Conway 1982:3; Pollock1976). The Middle Woodland was represented bya single dentate stamp vessel. “Flat slate knives,side-notched Temagami style arrowheads,extensive use of red ochre and local pottery whichis loosely based on Iroquoian motifs” wereassigned to the Late Woodland period (Conway1982:3). Representing the postcontact period wereseveral artifacts, including a “Davidson-Glasgowclay pipe (1863-1891)” and “an 1871 nickel”(Conway 1982:2).

Modern Landscape SettingWitch Point is a sand and gravel esker that extendswestward 1.25 km into the lake, with a maximumwidth of 0.25 km (Simony 1964). Its finger-likeshape creates the north shore of the protectedWitch Bay. A smaller point of land marks thesouth end of the bay. Four metres high at its tip,the esker climbs sharply higher to a steep face onthe north side and a sandy erosional face on thesouth side. Post-glacial wave action and erosionhave reworked the esker, as indicated by a small

Figure 37. Location of the Witch Point Site (CgHa-7)atop a sand and gravel esker, east-Central Hub of LakeTemagami.


baymouth bar and peat bog at the northeastcorner, a series of storm beaches on the south side,and a shallow sandbar extending into the lakefrom the tip (Figure 37).

The extreme tip of the esker curves to thesouthwest. Here an MNR-designated campsiteoccupies the upper portion: a flat grass-coveredtriangular clearing from 4 to 19 m wide by 19.5 m

Figure 38.Witch Point site esker morphology and excavation grids.


long. The west side has a steep slope, part of whichhas an active erosional face. To the east, a gentlerforested slope leads to a narrow terrace marked bydeer and moose trails. An eroded pathway leadssouthward to a small lower terrace. Cobble andsand beaches encircle the point (Figure 38).

The site area comprises the small lower terraceat the tip and the triangular clearing, and extendsnorthward into an open mixed forest of smallchoke cherry with some poplar, paper birch, andspruce. Based on positive test pits in the openforest and lower terrace, and on surface finds onerosional faces and pathways, the minimum arealextent of the site is 1,120 m2 (56 m from the lowerterrace up into the open forest by 20 m wide).

Excavation Methods and Artifact RecoveriesExcavation focused on the east half of the clearing.In 1993, a north block of 22.5 one-metre-squareunits was opened, while in 1994, an adjacentsouth block of 30 units extended toward the eskertip. Excavation and recording methods were thesame as those described for the Three Pines site,using one-metre-square units subdivided intoquadrants, with 3 cm vertical levels. At WitchPoint, all unit walls were profiled andphotographed. All soil was screened through a ¼inch (6 mm) mesh, with a 1/16 inch (1.6 mm)mesh screen used occasionally. Units were dug toan average depth of 24 cm, generally slightlydeeper in the north block, becoming slightlyshallower as the land slopes southward toward theesker tip. Several units were taken down to a depthof 54 cm.

Preliminary artifact identifications arepresented in Table 23. Cataloguing of the artifactsis still ongoing, as is a formal Harris Matrixstratigraphic analysis. Information presented hereis based on field observations summarized inpreliminary reports (Gordon 1993, 1994a, 1994b,1995a, 1995b). Artifact associations and phasingare therefore subject to revision.

Characteristics of and ContrastsbetweenWitch Point and Three PinesWhile preliminary, the work at the Witch Pointsite is summarized here because of the strongcontrasts with the Three Pines site. The 1993–

1994 excavations revealed significantly deeperanthropogenic deposits, including a thick darkgrey-brown organic-enriched layer, extensive fire-cracked rock deposits, and a greater overall artifactdensity. In the lower, probably Archaic levels ofthe north and south blocks, there was substantialmudrock and vein quartz primary reductionactivity. Fewer Middle Woodland Laurel potteryvessels were recovered compared to Three Pines,but a similar number of Late Woodland, cord-impressed, pottery vessels were found. Alsorecovered were additional pottery vessels withIroquoian tradition stylistic elements, a style ofpottery absent at Three Pines. These combinedfactors give the impression that Witch Point is alarger site used more intensively or extensively,particularly in the Archaic and Late Woodlandperiods. In addition, several Witch Point findssuggest ritual or ceremonial activities in the LateWoodland period.

Culturally Modified Soils and Cobble StructuresTable 24 shows an off-site soil profile of a humo-ferric podzol, 25 m north of the clearing at WitchPoint. The humus-enriched mineral horizon (Ah)is quite thick, but lacks an Ae eluvial horizon. Theparent materials are lenses of sand, gravel, andsmall cobbles deposited during deglaciation. Theyshow different periods of low- versus high-energymeltwater flow under the glacial ice forming theWitch Point esker.

On-site soils at the Witch Point site showsimilar natural podzolic soil development,including L, F, and H organic horizons, over adark brown Ah horizon, a weakly developed lightgrey eluvial Ae horizon, and a yellow-brown Bfsand horizon. Added to the Ah layer areanthropogenic deposits of burnt wood andcharcoal, creating a thick dark brown to dark greysilty sand layer that was quite compacted. Fromthe middle to the base of this layer are fire-crackedrock and large, well-rounded cobbles, oftenincreasing in size with depth (Figure 39). Theserocks and cobbles are much larger than thenaturally occurring well-rounded gravel and smallcobbles of the esker substrate. In some areas, rocksare densely packed (Figure 40). Below this arediscrete areas of grey ashy sand and patches of fire-


Table 23.Witch Point site preliminary artifact identificationsModern/recreational camping

Modern/older hunting, fishing, prospecting camps

19th–early 20th cent./construction

Historic fur trade

Glass - medicine bottle, clear and brown bottle glass

Metal - aluminum foil, 1980s coins, beer and other bottle caps, bent modern nails

Plastic - bread bag tags; fishing rod parts

Soil Features: Intrusive pits with food containers

Faunal - unburned beaver and moose bone

Metal - .22 long rifle and shotgun cartridges, 44 Magnum cartridge, 1930s linesinkers, 1940s coins, prospectors’ claim tags, pen knife

Glass - older style Pepsi bottles

Soil Features: thick organic deposits, fresh charcoal, small fire-cracked rockhearths, intrusive pits with very dark grey silty sand

Metal - square nails

Soil Features: post moulds, one with an old concrete base

Lithics - gunflint

Glass - white, red, green, blue small beads,

Metal - rolled copper bead, copper crooked knife, tinkling cones

Pottery - Iroquoian-like vessels with castellations, oblique incised collars andshoulder punctates; channelled rim vessels; bright orange-red paste plain pottery;plain pipes, cord-marked vessels

Lithic tools - small side-notched HBL chert and other chert projectile points, endscrapers, flake knives,

Lithic debitage - small flakes of HBL chert, other chert, vein quartz, clear quartz

Faunal bone - calcined

Mineral - red ochre

Soil features: fire-cracked rock concentrations, rock and calcined bone hearths,charcoal and fire-reddened soil deposits

Pottery - dentate stamp and pseudo–scallop shell decorated sherds

Lithics tools - HBL chert and Gordon Lake chert triangular end scrapers, largecobble with parallel grooves

Lithic debitage - vein quartz, clear quartz flakes

Soil features: small fire-cracked rock concentrations

Lithic tools - mudrock excurvate biface, celt, large bifacial flake knives, flakedcobble axe, large crescentic mudrock unifacial flake knife

Lithic debitage - large cores or preforms of vein quartz, vein quartz flakes and shatter

Metal - native copper (?)

Soil features: Concentrations of large mudrock cobbles, many fire-cracked; greyashy hearth remnants

Late Woodland

Middle Woodland (eastern edge of excavation)

Archaic/lower levels


Figure 39.Witch Point site – A unit profile showing the density of fire-cracked rock, all imported by precontactinhabitants onto the sandy esker top, mostly from the cobble beaches below.

Figure 40.Witch Point site – A plan view of a one-metre square showing the top of a densely packed “rock pavement”found in the north excavation block near the dog burial.


reddened sand, some with calcined faunal bone,underlain by yellow-brown sand (Figure 41).Sterile levels are found at 24–27 cm deep.

Witch Point occupants from all time periodsspent considerable time and effort hauling cobblesup from the beaches below. In the upper levels,probably Late Woodland, fire-cracked rocks aredensely packed. In some squares the net effect is arock pavement with more rock than soil (Figure40). Cobbles may have served a variety ofpurposes, such as roasting pits; hide smokinghearths; and sweat baths—a function suggested byseveral Bear Island residents who visited theexcavations. Middle Woodland pottery isassociated with smaller cobbles for hearths. In thelower levels, the type of fire-cracked rock changesto rather massive pieces of laminated GowgandaFormation mudrock. By adding these culturaldeposits to the natural sandy matrix, Witch Pointinhabitants contributed to the site’s thickerdeposits. The rock would have trapped wind-blown sediments and protected underlying layersfrom damage by successive occupations.

In contrast, the Three Pines site is shallower,with sterile levels in the yellow-brown sand at 15–18 cm deep. It has a thin, discontinuous,organic-enriched dark brown deposits, underlainby more extensive areas of grey ashy sand and fire-reddened sand, often containing calcined faunalbone. It has only discontinuous, small

concentrations of fire-cracked rock, even thoughcobbles are numerous at the lake edge. Settlementstructures include a Late Woodland rectangularrock platform, a Late Woodland semi-circularhearth windbreak, and numerous MiddleWoodland cooking and disposal hearths with fire-cracked rock and grey ashy sand.

Depth and Artifact DensitiesAt Witch Point artifact-bearing deposits aredeeper, an average of 24 cm compared with 18 cmfor Three Pines. Artifact densities are also higher.Preliminary counts show an average of 350 itemsper one-metre-square unit in the north block and290 in the south block. The overall site average is299 items per unit compared with 153 at ThreePines. By dividing these counts by the averagedepth of the excavation units, Witch Point showsa 47% higher artifact density than does ThreePines (299/24 = 12.5 versus 153/18 = 8.5).

Archaic Lithic Reduction ActivitiesArtifacts in the lower levels of theWitch Point site,assigned to the Archaic period, include largemudrock tools, such as excurvate bifaces, and largebifacial and unifacial flake knives. One hugemudrock cobble was pecked into an axe. As notedabove, the fire-cracked rock in the lower levels ismostly very large pieces of Gowganda Formationlaminated mudrock, easily obtained as beachcobbles or quarried from the numerous bedrockoutcrops. I suggest that primary lithic reductionof this softer, sub-conchoidal fracture raw materialoccurred on site, with the possibility of thermalalteration being used to produce smaller cores andpreforms.

Large vein quartz cores, preforms, and shatterare also numerous in these lower levels, suggestingprimary lithic reduction of this harder material.Vein quartz is found both in and beside discreteareas of grey ashy sand. One possible source is thelarge horizontal vein of milky coloured quartz, theKokoko Bay quarry site (CfHa-50), 7 kmsouthwest of Witch Point.

In contrast, at the Three Pines site, primarystage lithic reduction is not evident in the Archaicperiod units. While it has several large mudrocktools, large cores or chunks of Gowganda

Table 24.Witch Point site “off-site” soil profile.

Depth (cm)

1–0

0–15

15–32

32–45

45–54

54–103

Description

L horizon; forest litter

Ah horizon; humus-enriched sand;dark brown

Upper Bf horizon; iron-enriched sand;medium yellow-brown

Lower Bf horizon; iron-enriched sand;light yellow-brown

C horizon; esker formation; lenses ofmedium grey coarse sand, some withgravel (low-energy meltwater)

C horizon; esker formation; dark greycoarse sand with well-rounded, verycoarse gravel and cobbles 30–110 mmdiameter (high-energy meltwater)


Formation mudrock are absent. Three PinesArchaic tools are fashioned on a wide range ofother local raw materials, such as siltstone,quartzite, sandstone, and vein quartz, as well asnon-local blue and grey cherts and clear quartz.However, flakes, shatter, and preforms of thesematerials, including vein quartz, are minor (seeTable 7).

Laurel PotteryThe Witch Point site has far fewer MiddleWoodland Laurel pottery vessels than the ThreePines site. Laurel dentate stamp in the north blockand pseudo–scallop shell pottery fragments in thesouth block are confined to the eastern edge of theclearing. Several Laurel-style scrapers are scatteredin other areas, but these are not associated withpottery. The preliminary impression is thatMiddle Woodland usage of Witch Point is minor.

In contrast, Three Pines was extensively usedby Middle Woodland groups, leaving traces ofcooking and faunal bone disposal hearths,extensive grey ashy living floor remnants, and a

pottery-making area. Pseudo–scallop shell, dentatestamp, and linear stamp vessels are associated withsmall, triangular scrapers; retouched flakes; smallside-notched points; and a variety of celts,abraders, and grinding stones.

Cord-marked PotteryLike Three Pines, Witch Point had very few LateWoodland cord-wrapped stick–impressed or cord-marked pottery sherds. One vessel was found inthe N21 test pit, along with HBL chert flakes. Asecond was noted in the south block. Furtheranalysis will determine if these are Blackduckvessels.

Late Woodland Potterywith Iroquoian Tradition TraitsWitch Point contains a strong representation ofLate Woodland, Iroquoian-like pottery vessels.The north and south blocks both producedcollared vessels, some with castellations, decoratedwith shoulder punctates and with oblique andhorizontal incised lines. While several vesselsresemble Sidey Notched and Huron Inciseddesigns, they do not exactly match southernOntario Iroquoian tradition pottery inconstruction, paste, or overall form (Rick Sutton,personal communication 1994; Alicia Hawkins,personal communication 1994). Associated LateWoodland lithic tools include small, side-notchedHBL projectile points; end scrapers; and flakeknives. In contrast, no such vessels were recoveredat Three Pines.

Possible Ritual ItemsSeveral interesting items at the Witch Point sitemay be associated with Late Woodland rituals andceremonies. These include “red ochre paste”pottery, quantities of red ochre nodules, clearquartz crystals, and a burial of a young dog. Boththe dog burial and the bright orange, thin-walledpottery vessels were found in the north block,where the density of cobbles was high. Except fora few red ochre nodules, these types of finds areabsent at Three Pines.

Explaining VariationThe Witch Point and Three Pines sites, both

Figure 41.Witch Point site - 1994 excavation lookingnortheast across south block to 1993 north block.


mainland sites in the same section of LakeTemagami, exhibit significant differences in sitestructure, settlement features, and culturalcomponents. Witch Point revealed deepanthropogenic sediments, including extensivecobble structures, from the Archaic to Woodlandto Historic periods. It has a strong Late Woodlandrepresentation seen in its Iroquoian-derivedpottery designs. Of special interest are other LateWoodland features which, through ethnographicanalogy, suggest ritual activities. Such findingswere absent in the thinner, discontinuoussediments of the Three Pines site, with itspredominantly Middle Woodland occupations.What could account for the marked differencesbetween these two sites? Comparing the landscapecontext of these two sites in light of the previousgeochronological and palaeo-hydrological researchmay offer some insights into their differences.Ecological change and seasonal differences basedon faunal bone comparison and site locationalfactors are also considered.

Geochronology and Lake TransgressionThe Witch Point site may contain oldercomponents than the Three Pines site. Thepotential for a greater time depth is supported bytwo palaeo-environmental factors. First, theWitchPoint esker is an older, more stable land form thanthe Three Pines baymouth bar. Second, the steep-sided esker was less susceptible to laketransgression than the lower-elevation Three Pinessite. The esker formed during a period of rapiddeglaciation by at least 12,000 years ago, whenLake Temagami was ice-free (Veillette 1988).Whether Witch Point was suitable for occupationat such an early date is unknown. It is possible thatWitch Point was part of the central sill separatingtwo of the ancient lake basins. Ancestral LakeTemagami formed in the northeast basin by10,900 years ago and in the north basin by 9,800years ago (Table 18).

As noted earlier, the two earliest radiocarbondates in northeastern Ontario are c. 8,500 and5,800 years ago (Hanks 1988; Knight 1977).Between these dates, Lake Temagami wasgradually regressing from its former northern andnortheastern outlets while encroaching on palaeo-

shorelines in the rest of the basin, including theWitch Point and the Three Pines sites in theCentral Hub.

Post-glacial wave action and erosion reworkedthe Witch Point esker, but these processes actedon the steep margins, not on the esker top, whereprecontact activity was concentrated. In contrast,the Three Pines site is on a less stable, post-glaciallandform. A combination of erosion andlongshore drift formed the baymouth bar in earlypost-glacial times. As lake levels rose in the centrallake basin, the sandbar and peat bog configurationshifted northward, reaching its present-daylocation by 7,500 years ago. Any potential earlyArchaic period occupations of the front edge ofthe Three Pines site have now been inundated bya 4 m lake level rise.

At 7,500 years ago, although the 8 m climbup to the Witch Point site would have beenconsiderable, the site did offer a well-drained flatlocation along an otherwise rocky mainlandshoreline. As lake levels rose, any nearby lower-elevation sites flooded, rendering Witch Pointmore accessible and desirable. Perhaps rising waterlevels also played a factor in the later spatialdistribution of Woodland components at theWitch Point site. For example, Middle WoodlandLaurel and Late Woodland cord-marked vesselswere found somewhat closer to the esker edgesthan were the Iroquoian-like Late Woodlandvessels.

Ecological and Seasonal VariationCould there be ecological or seasonal variationbetween the two sites that might explain thecultural variation? The regional vegetation historypresented earlier (Table 19) shows that most of theArchaic period enjoyed a warmer, drier climate(Zone 3a Hypsithermal), while the Woodlandperiod as a whole is characterized by a coolingtrend (Zone 3b and 4 Neoglacial). In theHypsithermal, warmer temperatures would meanlonger periods of open water, and drier conditionswould lead to a greater incidence of forest fires.The relatively rapid shift to cooling temperaturesfalls in the Late Archaic period, with a dramaticcrash in white pine pollen. In Liu’s (1990)reconstruction, the Boreal Forest–Mixed Forest


ecotone, which had reached the northern edge ofthe Canadian Shield in the Hypsithermal,retreated southward to stabilize at its currentlocation, 50 km north of Lake Temagami, in theNeoglacial period. Slight cooling and increasedprecipitation continues throughout the Woodlandperiod. There is no obvious correlation betweenclimate change and cultural change that wouldhelp explain site variation at Lake Temagami.

Determining seasonal variation through faunalbone identification is difficult on Canadian Shieldsites. Unburned bone usually does not survivelong because of soil acidity, while calcined bone isfragile. Dense bone with distinctive skeletalfeatures, such as common loon, beaver andungulates, tends to be easier to recognize asfragments in faunal assemblages. More delicatebird and fish bone, which include some warmseason indicators may be under-represented. Incomparing the taxa identified from the WitchPoint salvage excavation (Smith 1983) with thosefrom the 1986 Three Pines excavation (see Table3), no major seasonal difference can be seen. The1982 collection contains 2 fish as well as 55elements identified to species, mostly beaver, withsome lynx, muskrat, striped skunk, caribou, andmoose (Smith 1983). From Three Pines, 842elements could be identified below the taxonomiclevel of class. These include beaver, white-taileddeer, moose, black bear, marten, pond turtle,ruffed grouse, and common loon. Only the singleloon bone and three pond turtle bones relate towarm weather usage at Three Pines (Prevec 1987).

Seasonal variation through ethnographicanalogy may be more useful. Witch Point shareswith Three Pines many desirable characteristics fora good camping locale. It has well-drained, flatterrain and ease of access from water to shore. Theeast–west trending esker curves to the southwest atits tip. This offers some protection from coldnorth winds while still providing southernexposure. The shallow water around the marginsand cobble beach provides a good landing placefor watercraft.

Witch Point may have attracted more peoplebecause of its location, particularly in the openwater season. It is a unique topographic feature onthe lake. Jutting out into the lake, it forms a

natural quay, accessible from all sides, creating asheltered bay. In the summers of 1993–94, localvisitors to the site used whichever side of the pointwas in the lee shore when landing their boats.North of Witch Point, the lake narrows to a longchannel with steeply rising cliffs formed byNipissing Diabase dykes along the eastern shore.Compared with this channel and the wide openGranny Bay farther north, Witch Point wouldhave offered a welcomed respite to open watertravellers. Also, Witch Point lies at the beginningof a major travel route to and fromDiamond Lakeand the Montreal River-Lake Timiskamingdrainage. Bill Twain reported that Witch Pointwas used as an assembly point for families going toand from the large summer gatherings at nearbyBear Island (Gordon 1994b). These landscapefeatures alone may account for a higher artifactyield and density of precontact occupations atWitch Point.

The Three Pines site is on a broad peninsulaand therefore has a greater land mass than thenarrow Witch Point esker. It has a larger near-sitehinterland for resources such as firewood andsphagnum moss from the peat bog. It also enjoysgreater north wind protection. This evidencesuggests that the Three Pines site would besuitable for both warm season and cold seasonusage, and for longer term usage. In contrast,Witch Point probably witnessed shorter term,intensive usage, mostly in periods of open water,over a longer time span than the Three Pines site.

Late Woodland Rituals, AllianceFormation, and Models of Social Change

Archaeological evidence suggests that the WitchPoint site was not only a stopover campsite forhunting and gathering, but also a place of socialinteraction and ceremonial expression for itsprecontact inhabitants. As described above, severalunusual items plus the dog burial were uncoveredin the 1993 north block excavation (Gordon1993, 1994a, 1994b, 1995a, 1995b). These findsare tentatively assigned to the Late Woodland,based on field observation of spatial associationwith Iroquoian-influenced pottery designs. Thetemporal assignment awaits verification through aHarris Matrix stratigraphic analysis and absolute


dating of the dog bones. What makes thesefindings of interest, and worthy of this preliminarydiscussion, is that, first, ethnographic analogysuggests they may be associated with rituals and,second, such items are absent at the Three Pinessite, with its predominantly Middle Woodlandoccupations.

This aspect of the Lake Temagami researchbroadens further to consider questions of hunter-gather behaviour in the context of the humanecosystem (Butzer 1982), leading to questions oflong-term cultural change. Focus turns to themore elusive, but interesting, changing socialcontext. Hunter-gatherer populations interact“not only with the physical environment andbiological one, but also with the socialenvironment” (Bailey 1983:187). This sectionexamines material evidence for ritual behaviour atWitch Point in comparison with other regionalsites. Identifying ritual behaviour archaeologicallyhas certain pitfalls, but ethnographic andethnohistorical studies provide some clarification(Brück 1999). Examining seventeenth-centurywritten accounts of hunter-gatherer groups inproximity to Lake Temagami elucidates some ofthe forms and functions of rituals, which may havecontinuity from the Late Woodland period.Finally, the role of ritual in alliance formation isconsidered, both as possible explanations for theWitch Point evidence and as an underlying themein models of long-term social change (Gilman1983; Madden 1983). It is tentatively proposedthat such models may help explain certain MiddleWoodland to Lake Woodland changes seen innortheastern Ontario.

Witch Point Ritual EvidenceIn the north block, the unburned, partial skeletonof a dog was found without artifacts in a small pitof fine sand. This pit lies below the dark grey-brown layer and intrudes into the sterileyellow-brown B horizon. Skeletal elements includeboth mandibles, a maxilla fragment, oneextremity, and several vertebrae. Epiphyses areunfused, indicating a young animal. Cut markswere observed only on one long bone.

The burial lies beside a particularly denseaccumulation of fire-cracked cobbles, which form

a type of rock pavement. As stated above, onvisiting the site in 1993, several Bear Islandresidents suggested that this rock pavement mayhave been for sweat baths. In north-centralQuébec, small cobble structures were found in twomodern Cree winter camps (Hanks 1983). Thesesmall sweat lodges were 0.5 m in diameter. Hanks(1983) classified these as specialized socio-religiousstructures, called atutson, because physical andspiritual healing are integral practices for the Cree.

The bright orange, thin-walled pottery vesselsfrom the 1993 block have not been seen anywhereelse on the lake or in the region. Perhaps red ochrewas mixed into the clay before firing, or a specificiron-rich clay was used. These unique and delicatevessels may have had a special purpose. Unflaked,exotic clear quartz crystals were also found in thenorth block. Clear quartz flakes occur at WitchPoint in Woodland period assemblages, while aclear quartz biface fragment fashioned from a verylarge crystal was assigned to the Archaic period atthe Three Pines site (see Figure 13: #143).

Quantities of red ochre nodules were found inboth the 1993 and 1994Witch Point excavations.Conway (1982) reported 556 g of red and yellowochre. While a rare commodity regionally, it wasmined locally south of Lake Temagami, accordingto native informants (Conway and Conway1989). From Craig Macdonald’s interviews withBear Island elders, we learn that “TheAnishinawbeg named one of their riversWu-num-man Zipi or the red mud river. They would gothere to collect a rouge-coloured material, actuallyiron oxide, from rocks and they would use it aspaint or dye” (Jenish 2006). Red ochre in ahabitation context may have been used for tentmarkings or coat markings (Skinner 1911). A bearskull marked with vermillion paint is illustratedon Macdonald’s (1993) Historical Map ofTemagami. Suspending animal skulls and otherbones in trees is a symbolic ritual of respect bytraditional hunters to the animal spirits (Gordon1980; Tanner 1979).

This extensive occurrence of red ochre suggestsa potential Late Woodland context for some of theLake Temagami rock art. Of the 36 known rockpainting sites in the area (Dewdney and Kidd1967; Conway and Conway 1989), 21 are on


Lake Temagami itself (see Zawadska, this volume)and several are near Witch Point (Gordon 1995a).Lake level research suggests that pictographs at thenorth end of Lake Temagami date between 1,000and 4,000 years ago. Rock paintings may beindividual expressions of narrative events or dreamquests (Rajnovich 1994), but regardless of themeaning assigned by the maker or subsequentinterpretations by later inhabitants, rock paintingsites serve as localized, public expressions ofpresence, on view to all who pass by.

The plausible inference through ethnographicanalogy is that some artifacts and settlementfeatures at theWitch Point site: the densely packedcobbles; the dog burial; red and yellow ochre;unmodified clear quartz crystals; and delicate,bright red pottery vessels, are related to LateWoodland rituals and ceremonial expression.

Evidence from Other SitesOnly a few dog burials have been reported innortheastern Ontario sites. According to BevSmith (see Oberholtzer 2002), this rarity may bemore a problem of recognition than of presence.At the Lake Temagami site Carscallen (1994b,personal communication 2012) excavated a singledog burial with red ochre. More numerous are thedog burials at the Frank Bay site, on LakeNipissing. Three were reported by Ridley(1954:49) and six by Brizinski (1980). These sixdog bundles were found in the middle to upperstrata (Brizinski and Savage 1983). All the dogskeletons were unburned and dismembered, withtwo underlain by charred birch bark. Four of thebundles were associated with red ochre, and oneburial (number 4) was “accompanied by anexquisitely formed quartz crystal” (Brizinski andSavage 1983:35).

Densely packed fire-cracked rock occurs at theWitch Point site adjacent to the north block dogburial. Such “rock pavement” was not seen onother sites excavated or tested by me on LakeTemagami. However, Carscallen (personalcommunication, 1994) reports heavy rockconcentrations at the sandy Lake Temagami sitethat also yielded a dog burial. Brizinski (1980)found a pavement of fire-cracked rock, measuring1.3 by 0.8 m, in the same excavation area as the six

dog bundle burials at the Frank Bay site.Most dog burials on Upper Great Lakes sites

date to younger than 600 years ago (i.e. after A.D.1400) (Bev Smith, personal communication 1993;Oberholtzer 2002) but dates on charcoal beneathtwo of the Frank Bay dog burials (numbers 1 and2) indicate a greater antiquity, namely, 995 ± 50RCYBP (S-1685) and 895 ± 60 RCYBP (S-1686)(Brizinski and Savage 1983:35). The LakeTemagami site dog skeleton is somewhat youngerat 760 ± 60 RCYBP (T0-4334) (700 ± 35 cal. B.P.Liam Kieser, personal communication 2014).Perhaps the practice of dog burial is somewhatolder in the Lake Nipissing- Lake Temagami area.

On the Mattawa River east of Lake Nipissingis the Port de l’Enfer site (CbGr-1). This red ochremine is rather inaccessible place, as it is locatedwithin a deep cave, high up a cliff face (Tyyska andBurns 1973). Red ochre, or haematite, is a chalky,deep red–coloured mineral that can be easilyground into a powder and mixed with a binder offish oil to form a durable pigment (Rajnovich1994). On its own, it served as a powerfulmedicine (Rajnovich 1994). Brizinski (1980:139)identifies the red ochre and quartz crystal as“exotic or ceremonial goods,” with the latter“acknowledged as having ‘power’ amongAlgonkians” (Brizinski and Savage 1983:35).

Thus, archaeological evidence from both LakeTemagami and Lake Nipissing supports asequence of the deliberate killing (cut marks oncervical vertebrae), dismemberment (cut marks onjoints and long bones), and specialized interment(birch bark covering or in a pit of clean fill) ofindividual, young dogs (Brizinski and Savage1983; Carscallen 1994b; Gordon 1993). Redochre, clear quartz crystals, and rock pavementstructures also occur on sites with dog burials,some in association, some not. Elsewhere, dogburials occur in single-household sites, such as IsleRoyale in Lake Superior (Clark 1990), and withinlarger “village” sites with longhouses, such as theProvidence Bay site (BkHn-3) on ManitoulinIsland (Fox 1990; Smith and Prevec 2000). Dogburials occur on large sandy sites connectingmajor Great Lakes travel routes, such as the FrankBay site, which may have functioned as a LateWoodland trading village (Bandow 2012). Upland


from the Great Lakes, dog burials occur on a large,longer term, possibly a summer gathering site, theLake Temagami site, and on a much smaller,short-term travel site, the Witch Point site.

Defining Ritual BehaviourAccording to Brück (1999:314-315), anthro-pologists and archaeologists agree on certaincharacteristics that define rituals, including highlystructured, repetitive, and prescribed modes ofbehaviour, using symbols or having a symbolic orexpressive nature. She cautions that “a recurrentyet unspoken theme underlying manyarchaeological discussions of ritual is the equationof ritual with non-functional action” (Brück1999:317). Drawing from ethnographic examples,including Brightman’s (1993) research on theRock Cree of northwestern Manitoba, Brückreminds archaeologists that in many societies“where people do not draw such a [modernWestern] categorical distinction between thesacred and the profane, ritual action may not bespatially or temporally distinguished from more‘mundane’ or secular activities” (Brück 1999:319).

One contemporary example from a James BayCree winter camp suggests that the distribution offaunal bone on an archaeological site may indeedreflect ritual behaviour (Gordon 1980). Specific,repetitive disposal practices of sacred animal bonesshow respect and gratitude of human hunters forthe animal spirits (Feit 1973; Rogers 1973; Tanner1979). Ethnohistorical studies of the Rock Cree(Brightman 1993), and contemporary studies ofJames Bay and Mistassini Cree hunting practices(Feit 1973, 2004; Tanner 1979) show that,

in the Cree view of the natural world,the hunter is part of a system and keyCree values (such as respect, sharing,reciprocity, generosity, taking care) applyto relationships between animals andpeoples as well as to those among people[…] Customs and rituals help peopleremember the rules and interpret signalsfrom the environment appropriately[Berkes 1998:111, 125].

So it is plausible that these uncommonarchaeological artifacts, dog burials, and structural

features are not the only expressions of ritual activityon Late Woodland sites, but that some, if not all,played a symbolic role in formalized social practices.The next question is, “What meaning andfunctions might such rituals have expressed?”Documentary evidence from the seventeenthcentury, the early contact period, offers some clues.

Early Contact Period AccountsThe arrival of Europeans explorers, Jesuitsmissionaries, and fur traders in northern NorthAmerica resulted in the first documentaryaccounts of northern forest (Canadian Shield)hunter-gatherers. These accounts are, of course,written from and for a European perspective, butrelied on informant information. Like artifacts,they may represent single events or observations;how far can they be generalized is thus a matter ofinterpretation.

Lake Temagami people were likely in directcontact with Lake Nipissing people who attendedformal ceremonies with dog feasts (Thwaites1896-1910:23:221). On the pre-1900s map(Macdonald 1993), two separate routes to LakeNipissing are marked on rivers flowing south fromLake Temagami, part of the Sturgeon Riverdrainage. Lake Nipissing people were in directcontact with early French fur traders, Huronhorticultural groups to the south, and Algonquianhunting-gathering groups to the north(Champlain 1922:3:41; Thwaites 1896-1910:21:230-240).

The Rivière d Estarjon [sic] is marked on a1616 map draughted by Samuel de Champlain(Heindenreich 1971:24). In his 1615 account,Lake Nipissing people go up the Sturgeon Riverto trade.

This lake is some eight leagues wide andtwenty-five long [Lake Nipissing, whosedimensions are overstated] and into itflows a river [the Sturgeon River] whichcomes from the north-west up whichthey go to trade the goods which wegive them in barter and exchange fortheir furs, with those who dwell here,who live by the chase and fishing[Champlain 1922:3:41].


In 1640, Father Jerome Lalemant offers thisdescription of the Nipissings’ seasonal economicround and long-distance trading relations.

Nipissiriniens […] form a Nation of theAlgonquin tongue which contains morewandering than settled people. Theyseem to have as many abodes as the yearhas seasons, —in the Spring a part ofthem remain for fishing, where theyconsider it the best; a part go away totrade with the tribes which gather onthe shore of the North or icy sea [JamesBay], upon which they voyage ten days,after having spent thirty days upon therivers, in order to reach it. In summer,they all gather together, on the road ofthe Hurons to the French, on the borderof a large lake which bears their name[…] About the middle of Autumn, theybegin to approach our Hurons, uponwhose lands they generally spend thewinter; but, before reaching them, theycatch as many fish as possible, whichthey dry. This is the ordinary moneywith which they buy their main stock ofcorn, although they come supplied withall other goods, as they are a rich peopleand live in comfort. They cultivate alittle land near [Page 239] their Summerdwelling; but it is more for pleasure, andthat they may have fresh food to eat,than for their support [Thwaites 1896-1910:21:230-240].

Another 1640 account from the JesuitRelations identifies localized named groups innortheastern Ontario, of which two can be reliablyassociated with the study area (Bishop 1994a).This is the first time the name Outimagamiappears in written documents.

Leaving the River des Prairies [OttawaRiver] when it turns directly to theNorth, that we may go to theSouthwest, we come to Lake Nipisin,where the Nipisiriniens are found.These have upon their North theTimiscimi, the Outimagami, the

Ouachegami, the Mitchitamou, theOuturbi, the Kiristinon, who live on theshores of the North sea whither theNipisiriniens go to trade [Thwaites1896-1910:18:227-229].

Note that trading relations between LakeNipissing and James Bay groups predate the 1668English fur trade on James Bay and the earliestinterior French posts: 1673 at Nighthawk Lakenear Timmins and 1679 at Lake Timiskaming(Bishop 1994a; Mitchell 1977). As Arthur Ray(1974) has advocated for the English fur trade innorthwestern Ontario and the Prairies, the earlyFrench fur trade may have overlain and made useof pre-existing social alliance networks amongAlgonquian-speaking groups (Hickerson 1960).

The Jesuit Relations also provide severalseventeenth-century accounts of formalceremonies conducted to reaffirm existingalliances and initiate new ones among groups inthe upper Great Lakes (Hickerson 1960). FatherJerome Lalemant’s description is the most detailedof these early accounts (Thwaites 1896-1910:23:209-223). In this account, the Nipissingswere the host, inviting the Sault and the Huron,among other groups, to participate in elaboraterituals over several days (Thwaites 1896-1910:23:209-223; Hickerson 1960). Theseceremonies involved reciprocal gift exchange tosolidify social alliances between linguisticallyrelated groups and with unrelated groups. Higherlevels of social-political integration are suggestedby the elections of chiefs, as well as by possible“early totemic organization” (Hickerson 1960:62).

The event took place in September 1641. Aspart of the Huron contingent, Father JeromeLalemant witnessed a deliberately plannedassembly of 2,000 people in a large bay on theshores of northern Georgian Bay (Thwaites 1896-1910:23:209-223). He described the location asbeing 20 leagues (110 km) from the House of Ste.Marie, placing it perhaps in the modern-day ParrySound area. Activities included gift exchanges,theatrical dances with drumming and music,individual sport competitions, the election ofNipissiriniens Chiefs, and the transfer of namesfrom those who had died since the last feast to theliving (Thwaites 1896-1910:23:209-217).


According to Lalemant (Thwaites 1896-1910:23:217-219) the most solemn and sacredwere the ceremonies for the dead. The womenbuilt an arched cabin of 100 paces long, bringingin bones of the deceased in birchbark containerscovered with beaver robes. A feast was held for thewomen only and they chanted throughout thenight. Additional activities were held in this longcabin the next day, after which came a dog feast.

Algonquin Captains, who acted asMasters of Ceremonies, entered ten ortwelve in line, bearing flour, beavers,and some dogs still alive, with whichthey prepared a splendid Feast for theHurons. The Algonquin Nations wereserved apart, as their Language isentirely different from the Huron.

Afterward, two Meetings were held; oneconsisted of the Algonquins who hadbeen invited to this Solemnity, to whomvarious presents were given, according tothe extent of the Alliance that existedbetween the Nipissiriniens and them[…] The second Assembly was that ofthe Huron Nations, at which theNipissiriniens gave us the highest Seat,the first titles of honor, and marks ofaffection above all their Confederates[Thwaites 1896-1910:23:221].

Lalemant (Thwaites 1896-1910:23:209)himself noted that this “feast of the dead” differedgreatly from those of the Huron, described byGabriel Sagard in 1623-24. Hickerson (1960:81,87) suggests these pre-burial practices wereborrowed from the Huron only by thoseAlgonquian groups in direct contact. Rajnovich(1994) disagrees, arguing that this is the earliestaccount of a Ghost Lodge ceremony, one of manyelaborate ceremonies related to the widespreadAlgonquian Midewiwin, or Society of GoodHearted Ones, which persists throughout thehistoric period and into the twentieth century.

Throughout the historic period, ritual dogsacrifices were documented among Algonquian-speakers of the Great Lakes (Brizinski and Savage1983; Oberholtzer 2002). The purpose of theserituals varied. Dog sacrifices were done to appease

or ask favour from spirits, to accomplish successin hunting or war, and to ward off disease(Oberholtzer 2002). A Teme-Augama Anishnabaielder told heritage consultant Judy Gouin(personal communication 1993) that at thenaming ceremony, young children are fed the meatof a puppy to ward off diseases (Gordon 1993).Ritual killing and consumption of dogs was also aspecial part of the rites of passage, such asadoption and initiation into the Midewiwin ormedicine lodge ceremonies (Brizinski and Savage1983; Long 1974; Oberholtzer 2002; Skinner1911), and, in the 1641 example mentionedabove, as part of an alliance-reinforcing feast.

In 1701, a major gathering of Great Lakespeople occurred in Montreal, to make a peaceagreement between, on the one hand, the Frenchand their allies and, on the other hand, the IroquoisConfederacy (Bohaker 2006). Among the 25political entities the French named in the treatywere the “nepissingues [Nipissings], algonquins,Temiskamingues [Lake Temiskaming people]”(Bohaker 2006:24). Native signatories used 38 or39 pictograph images of mammals, reptiles, fish,and flora on the document itself, making it theearliest known example of such symbolic writingon a treaty document (Bohaker 2006:24).6

In the eighteenth and nineteenth centuries, 20of these animal figures are associated with thetotem clan groups, or nindoodemag (kinshipnetworks), of Algonquian speakers: Ojibwa,Ottawa, Potowatomi, and Algonquin.

Nindoodemag shaped marriage andalliance patterns and facilitated long-distance travel; access to communityresources was also negotiated throughthese networks. Sources dating from theseventeenth century suggest that in thisearlier period and likely before contact,nindoodemag operated as an importantcomponent of Anishinaabe collectiveidentities, fulfilling similar social andpolitical functions [Bohaker 2006:26, 29].

6 In the author’s opinion as an artist, there is a re-markable stylistic similarity between these images onthe 1701 Great Peace (Bohaker 2006:Figures II andIII) and the pictographs illustrated in Rajnovich’s1994 book on northern Ontario rock art.


Bishop (1994a, 1994b) notes that clans werepresent by 1770, while Hickerson (1960) seeshints of totemic evidence in the early contactperiod. Teme-Augama Anishnabai familiescurrently are part of 14 clan names and totems(Teme-Augama Anishnabai/Temagami First Nation2007).

Rituals and AlliancesThus, documents from the 1600s and early 1700sreveal events and social, economic, and politicalpractices among Algonquian-speakers inproximity to Lake Temagami. They show theextensive mobility of these hunter-gatherers,including long-distance trading that probablypredates the French fur trade (Hickerson 1960).Direct or indirect contact between coastal JamesBay, the upland Canadian Shield, andhorticultural regions farther south are indicated(Champlain 1922:3:41; Thwaites 1896-1910:21:230-240; Bishop 1994a) leading to tradeexchanges and other interactions betweenlinguistically related and unrelated groups(Hickerson 1960). Localized name groups areevident, more than those represented in the 1701French tally, including the Outimagami (Bishop1994a; Bohaker 2006). Extensive social networkscalled nindodemag; intensified communal activitywithin autonomous groups; and group integrationbetween allied and/or confederated groups alsoexisted (Bohaker 2006; Hickerson 1960:95).

The dog feast is one element in alliance-reaffirming ceremonies among upper Great LakesAlgonquian groups and their Iroquoian-speakingallies to the south (Hickerson 1960; Thwaites1896-1910:23:209-223). Alliances ceremonieswere also conducted among allied groups and theirenemies, at the behest of the French (Bohaker2006). One key reason for formalized socialalliances is to help groups in times of distress,enabling them to gain access to new territory andresources through reciprocal social and economicobligations (Bailey 1983; Bohaker 2006; Gilman1983; Hickerson 1960; Madden 1983). With thestress of epidemics in the mid-seventeenth centuryand Iroquois raiding attacks, these strong kinshipand inter-group alliances clearly served theirpurpose. Iroquois raiding stories are still part of

Teme-Augama Anishnabai oral history (Conwayand Conway 1989; Gordon 1995a).

In the early post-contact period one impetusfor forming new alliances would have been toaccess trade goods, either directly from the Frenchor through middlemen, such as the Nipissing(Hickerson 1960) and Odawa (Fox 1990). So itcould be argued that the social, economic,political, and ritual practices described in theseearly documents are new forms—a direct result ofthe introduction of rare, highly valuedcommodities and the influx of distinct linguisticand cultural groups (e.g., White 1991).

However, others have argued for continuitybetween the precontact and historic periods.Hickerson (1960) argues for a florescence of pre-existing social patterns in the 1600s, rather than anew response to the early European fur trade.Bohaker (2006) suggests that clan- or totem-basedalliance networks extended back into theseventeenth century and perhaps earlier.Oberholtzer (2002) proposes that dog sacrifice inhistoric times may be a continuity of precontact dogburial practices. Rajnovich (1994) contends thatboth precontact and historic period rock paintingscan be related to totems, power dreams of medicinepeople, andMidewiwin ceremonies (e.g., the latticeglyph at Keso Park on the French River).

Thus it can be inferred that one key functionof rituals in the Late Woodland period was tosymbolically express social and economic alliances,between local groups, distant groups, and eventhose speaking different languages. This possibilityoffers one plausible explanation for ritual evidenceat the Witch Point site. Witch Point was afavourable habitation site for precontact groupstravelling north and south past the steep cliffs andwindy open water of the North Arm. As atraditional use site, according to Bill Twain, it wasan assembly point for families on their way to andfrom the large summer gatherings at nearby BearIsland (Figure 42). The rock pavement mayrepresent successive large hearths in ritualsinvolving feasting, dog sacrifice, and dog burial.Cobble structures for sweat baths are anotherpossible ritual function. These rituals may haveserved to reintegrate dispersed families, reinforcingreciprocal obligations.


Groups travelling long distances who werepassing through Lake Temagami may also havemet here to ritually exchange goods andinformation with local groups. It may not bepossible to get to the specific meaning or forms ofthese rituals, or to the specific groups involved (seeBandow 2012 for an alternative viewpoint).However, the same artifacts and structural featuresthat are present on large sandy gathering sites,such as Frank Bay and the Lake Temagami sites,are also repeated on this smaller travel camp atWitch Point. This evidence suggests that certainLate Woodland rituals had a similar, prescribedformat, whether practised externally, in elaborate,large-scale meetings such as described by Lalemant(Thwaites 1896-1910:23:209-223) for thepurpose of reaffirming alliances between locallynamed groups, adjacent neighbours, and alliesfrom different linguistic groups, or internally,between families and local groups to reinforcemutual social, spiritual, and economic support.

Alliances and Models of Social ChangeThe evidence for ritual behaviour found at theWitch Point is absent from the Three Pines site,which contains no Iroquoian-like pottery vesselsbut does have a strong Middle Woodland Laurelpottery component. There are no specificlocational, ecological, or seasonal factors that helpexplain this variation. So, perhaps these differencesrelate to broader social-cultural trajectories fromthe Middle Woodland to the Late Woodlandperiods reflected in this admittedly insufficientsample of two upland Canadian Shield sites.

Nonetheless it is worth examining several modelsof long-term culture which consider the role ofritual in social alliances.

Several archaeologists have emphasized theimportance and application of alliance andreciprocal exchange theories developed inanthropology (e.g., Levi-Strauss 1969; Sahlins1972) to explanations of long term culture changein the prehistory of hunter-gatherers (Bender1978; Gilman 1983; Madden 1983). Tosummarize briefly: the domestic household orlocal group cannot survive in isolation, but needsto form alliances with others in order to practise aseasonal subsistence economy across a widegeographic area (Bender 1978). One way ofcreating alliances is by exchanging spouses, thusforming kinship groups or mating networks(Bailey 1983). Alliances between households orlocal groups involve reciprocal exchange andobligations. One obligation is to produce morethan the household needs in order to createsurplus goods for exchange (Bender 1978), forexample, trade in high-value material goods(Walthall and Koldehoff 1998). Alliances allowaccess to the territory and resources of othergroups, particularly in times of shortage and stress;offer economic assistance; provide security; andhelp to reduce and resolve conflict (Bailey 1983;Gilman 1983). Alliance networks (Gamble 1983)or social network systems (Madden 1983) rely onvarious means of communication, such as spokenlanguage, information exchange aboutenvironmental conditions, the visual media ofmaterial culture, and ritual (Gamble 1983:203).Increased emphasis on social rituals helps toreinforce long-distance exchange and symbolizesgroup affiliation and role differentiation (Bailey1983).

Gilman (1983) uses alliance and reciprocalexchange theory to develop a two-stage model ofsocial change occasioned by technological changefrom the Middle to Upper Palaeolithic in Europe.He postulates that the Middle Palaeolithic, with alower level of technology, had very low populationdensities. Hunting groups would welcome helpfrom any and all others to prevent food shortages.Therefore there would be no social boundaries.Technological improvements in the Upper

Figure 42.Bear Island PowWow July 1994, Lake Temagami.


Palaeolithic led to increased hunting efficiency; awider range of species exploited; higherpopulation densities; increased ritual, particularlyassociated with human burial; and greater artisticand stylistic expression. The local group now hasmore neighbours, but needs less subsistence aid,given increased production. As households mustbalance the costs against the benefits of externalalliances, each group would therefore restrict thescope of its alliances (Gilman 1983:122-123).Alliances become more fragile and thereforerequire rituals to hold them together.

Maintaining the necessary web of socialrelations requires the balancing ofcontradictory interests and it is this thatmakes ritual reinforcement ofreciprocity necessary […] [In the UpperPalaeolithic] the closed connubium offriends-in-need would requireceremonies to symbolize and cementtheir alliance and style to represent it,not because innovations in techniquehad made mutual aid more necessary,but because higher production security[…] had made social co-operation moreunstable [and] […] liable to breakdown[Gilman 1983:122-123].

Similarly, to explain regional variation inNorwegian Mesolithic sites, Madden (1983:192)develops three models of unbounded andbounded “social networks systems”. Key variablesinclude population density, geographic distancebetween groups, and degree of interactionsresulting in anticipated differences in materialculture. The first model is of an unbounded,undifferentiated social network system with a lowpopulation density spread over an extensive area.Interactions would be continuous but unevenacross the whole system, with the expectation ofmaterial culture homogeneity. The second modelis of differentiated networks because the distancesbetween them are too great to maintain socialinteractions. The third model is “differentiateddue to imposed social boundaries […] imposeddivision to preserve exclusivity of the exploitationterritories and/or resources” (Madden 1983:193).Traits include greater internal integration of

groups and corresponding external differentiationbetween groups in a region, as well as “moreformal and structured linkages” (Madden1983:193). This would be reflected in materialculture as distinct “differences in stylisticbehaviour, discrete stylistic zones,” more exchangeinteractions between bounded groups, as well as“increased symbolic/stylistic behaviour” (Madden1983:194).

Evidence of Change inUpper Great Lake FisheriesCleland (1982:772-775) notes similarcharacteristics in comparing variation betweenMiddle Woodland and Late Woodland fishingsites in the Upper Great Lakes. He providesevidence for improved technology, greaterefficiency of production, greater resourceexploitation, increased communal activity withlabour-intensive practices, greater subsistencesecurity, and increased population. MiddleWoodland peoples seasonally inhabited the shoresof the Great Lakes to exploit shallow-water,spring-spawning fish runs, with new technologiesand techniques, including seine nets withnetsinkers and harpoons. Late Woodland groupswere able to increase production and exploitdifferent fish species by developing a variant of theseine net. The gill net using netsinkers and floatswas employed in deep-water fall fisheries. Fallfisheries provided greater returns, both in quantityand nutritional value, than the spring ones.Surplus fish could be dried, stored, and used intothe winter months. Increased size, number, andduration of lakeshore sites suggests a substantialincrease in Late Woodland population. A labour-intensive practice, the fall fisheries requiredcommunal co-operation because they were ofshort duration, a few weeks only, compared withseveral months for the spring fisheries (Cleland1982:772-775). Cleland’s work supports thepossibility that a shift from open, unboundedsocial networks with more stable alliances towardmore closed social networks with imposedboundaries and unstable alliances (Gilman 1983;Madden 1983) may be applicable to northeasternOntario, albeit on a much smaller time scale.


Middle Woodland:Undifferentiated Social NetworksIf the Middle Woodland social network systemcorresponds to an undifferentiated one, with nosocial boundaries, this might account foruniformity in Middle Woodland assemblagesacross northern Ontario. Other researchers seeenough variation to propose a distinct stylisticzone of “Eastern Laurel” in northeastern Ontario(Pollock 1975; Reid and Rajnovich 1991), but theview here is that there are strong, overridingsimilarities in lithic raw material choices, tooltypes and frequency, method of lithic and potterymanufacture, and pottery designs and motifs.These are evident when comparing Laurel sites onNorth Caribou Lake in northwestern Ontario(Gordon 1985) with collections from acrossnorthern Ontario (e.g., Wright 1967a, 1967b,1968) and the Middle Woodland occupations atthe Three Pines site (Gordon 1991a).

Late Woodland:Toward Differentiated Social NetworksOne anticipated material culture aspect resultingfrom a shift from unbounded to bounded socialnetwork systems is increased distinctions instylistic behaviour (Madden 1983). This shift maybe seen in increased regional variation in LateWoodland pottery. Blackduck pottery occurs onmany Middle Woodland Laurel pottery sites, butnew designs and forms also appear. Innorthwestern Ontario these include, for example,Sandy Lake Ware (Arthurs 1978; Taylor-Hollings1999), Selkirk Ware (Koezur and Wright 1976;Rajnovich and Reid 1978), and Bird Lake Ware(Lenius and Olinyk 1990). In northeasternOntario, later pottery forms are predominantlyIroquoian-like motifs. From Lake Abitibi (Ridley1956, 1958, 1963, 1966) to Lake Temagami toLake Nipissing there is a shared pottery tradition,with the greatest diversity of vessels on thesouthern sites, such as Frank Bay (Ridley 1954;Brizinski 1980).

From an examination of Ridley’s Abitibi andFrank Bay collections (Gordon 1991a), it appearsthat some of these vessels may be direct importsfrom southern Ontario, while others, like the

Witch Point vessels, appear to be copies ofIroquoian pottery motifs, such as Sidey Notchedand Huron Incised designs. In the adjacentAbitibi-Témiscamingue area, Côté and Inksetter(2001) define a “Blackduck episode” as a relativelyshort period, dating between A.D. 1000 and1350, completely superseded by “proto-Huronian” vessels, dating between A.D. 1300 and1650. Neutron activation chemical analyses ofLaurel and Blackduck vessels show locally usedclay sources, while those with Iroquoian Traditionstylistic affinities were manufactured from agreater number of sources (Côté and Inksetter2001).

Increased regional variation in pottery stylesmay be associated with increasing socialboundaries across northern Ontario. Innortheastern Ontario, north–south lake and rivertravel routes served to connect historically knowndistinct groups. The occurrence of “exotic” vesselsand the replacement of Blackduck vessels withIroquoian-like vessels on upland Canadian Shieldsites supports more formalized trade, direct orindirect, and other forms of interactions betweengroups of different economic, social, and(potentially) linguistic affiliations (i.e., northernhunter-gatherers with southern horticulturalists)in the later part of the Late Woodland period.Pottery appears infrequently on upland sites in theJames Bay region of northern Québec (e.g., Chism1977, 1978), suggesting that these trade relationshad a finite distance.

What is in the pottery vessels may have beenmore important than the vessels themselves. At theFrank Bay site the presence of burned corn cobsdating to at least A.D. 1000 (Brizinski and Savage1983), indicates contact with maize cultivation orcultivators. By this date, intensive corn, bean, andsquash horticulture is clearly in place in the lowerGreat Lakes (Jamieson 1992), although the usageof maize (Hart and Lovis 2013) and squash (Foxand Garrad 2004) in precontact diets may beconsiderably older. At some point, tobacco mayalso have been a highly valued commodity forlong-distance trade, as suggested by the presenceof Late Woodland period ceramic pipes at theFrank Bay and Witch Point sites. The functionalpracticality of fragile ceramic vessels on the rocky


Canadian Shield has always been questionable inmy view. Perhaps the vessels themselves becamevalued symbols of a household’s alliance networks,long after the contents were consumed.

On the upland Canadian Shield ofnortheastern Ontario in the Late Woodland, thereis evidence of food surplus from Great Lakesfisheries, population increase, increased stylisticvariation in pottery, a new pottery tradition withsouthern traits, long distance trade, andpresumably other interactions with differenteconomic modes (i.e., maize cultivation), as well asdifferent linguistic groups (i.e., Iroquoianspeakers). Southern-tier Algonquian speakers maythemselves have practised horticulture (Fox andGarrad 2004:124). Longhouse villages, such asseen at Providence Bay onManitoulin Island (Fox1990), or even potential Algonquian villageswithin Huronia itself (Fox and Garrad 2004) maybe one result of these newly developing alliancesand interactions. Another result may be increasedceremonialism, as seen in the dog burials afterA.D. 1000, as mobile groups had wider choicesand greater reasons to form alliances using thesymbols and practice of ritual to keep fragile tiesfrom falling apart (Gilman 1983). Trends begunin the Late Woodland period toward more sociallybounded networks may also account for earlypost-contact evidence of higher levels of socio-political organization noted above (Hickerson1960:62; Bohaker 2006). These potential socio-cultural trajectories in the Late Woodland innortheastern Ontario would not have taken placein a vacuum. They may have been influenced bybroader regional—if not continental—developments, such as suggested by Jamieson(1992:70), who relates Iroquoian development insouthern Ontario to a process of“Mississippification,” resulting in networks ofsedentary fortified villages, cultivating maize,linked by lineage and clan affiliation, and engagedin warfare and raiding. Such developments maybe another impetus for increased Late Woodlandceremonialism to symbolize alliance formation onCanadian Shield sites in northeastern Ontario.

Summary

Proglacial Lakes Post-Algonquin and BarlowGlacial ice retreated from Lake Temagami around12,150 cal B.P. (10,400 ± 240 RCYBP) (Veillette1988). Deep proglacial lakes, held up by the icemargin to the northeast, inundated the Temagamibasin, as seen by varved clays in peat coresextracted from Bear Bog, near Sharp Rock Inlet(North Arm), and Baseball Bog (Northeast Arm),near Cassels Lake. The lithology and radiocarbondates from Baseball Bog correlate well withVeillette’s (1988) post-glacial reconstruction. LakePost-Algonquin reached northward from the LakeHuron basin into southern Lake Timiskaming,extending a narrow arm into the Temagami basin.This proglacial lake drained southward around11,500 cal B.P. Early proglacial Lake Barlowfollowed. It reached out of the Timiskaming basinand into the Northeast Arm. Cold, deep watervarves in Baseball Bog indicate a water depth of30 to 50 m (Veillette 1988). Thus proglacial LakeBarlow inundated altitudes from 319 up to 338m asl, creating palaeo-shorelines as high as 46 mabove the current elevation of 293 m asl for LakeTemagami.

Ancestral Lake TemagamiEarly Lake Temagami separates from proglacialLake Barlow, as seen in a shift from proglacialsediments to lake muds in the pollen cores. Thischange dates the establishment of ancestral LakeTemagami to 10,900 cal B.P. at Baseball Bog and9,800 cal B.P. at Bear Bog. Isostatic reboundplayed a significant role in the levels and drainagepatterns of this early lake.

A predictive model for locating ancestralshorelines was developed by applying isostaticrebound curves extrapolated from adjacent LakeTimiskaming (Lewis and Anderson 1989) to the50 km long Temagami basin. It predicted that theTemagami basin was tilted significantly in thedirection of continental ice retreat. A displacementof 30 m at the north end, relative to the south endof the lake basin 11,000 years ago, rebounded toa 15 m displacement at 10,000 years ago. The“tilt” model also predicted that ancestral LakeTemagami comprised three separate lakes: one inthe north half from the Central Hub northward,


a second lake in the Northeast Arm, and a third inthe Central Hub, South and Southwest Arms.

Fieldwork confirmed two predicted palaeo-outlets; a 60 m wide, high-energy spillwaydraining north through Sharp Rock Inlet towardDiamond Lake and a multi-channelled outletdraining northeast from the Northeast Arm intoCassels Lake. The modern southern outletthrough Outlet Bay-Cross Lake was not active atthis time. Water levels in these ancestral lakes werenot deep enough to form varves, but were stillabove the modern level of 293 m asl. Palaeo-shorelines in the form of water-washed bouldersand sand ridges were observed at altitudes of 295m to 305 m asl at Sharp Rock Inlet, Bear Bog,Deer Island, and Ferguson Bay (North Arm) andat Caribou Lake and Jessie Fen (Northeast Arm).These old lakeshores are 2 to 12 m above themodern lake, ranging from 16 to 400 m inlanddepending on slope.

Holocene Lake Regression (North)and Lake Transgression (South)As the earth’s crust continued to reboundthroughout the Holocene, lake levels declined inthe north and northeast basins and began floodingpalaeo-shorelines in the central and south basins ofancestral Lake Temagami. This is seen in atransition from lake mud to pond mud or peat inthe Baseball Bog and Bear Bog peat cores,estimated at 7,600 years ago. Eventually thenortheast palaeo-outlet closed, while the northernoutlet had a much reduced water flow. At somepoint, the sill across the Central Hub wasinundated, creating one large lake and opening amajor southern outlet, which operates today,through Cross Lake.

Evidence for rising water levels in the centralbasin is seen in the peat core extracted from theThree Pines Bog, beside the Three Pines site. Thissphagnum peat bog is separated from the lake bya narrow, permeable sandbar. Its water tablefluctuates with lake levels. The depth and rate ofpeat growth in the core reflects changes in the levelof Lake Temagami itself. Peat began forming atthe Three Pines Bog in the mid-Holocene, datedat 7,500 cal B.P. The accumulation of 4 m of peatin the bog at a rate of 5.3 cm per century,

corresponds to a 4 m rise in lake over 7,500 years:evidence of continuing lake transgression in thecentral basin due to differential isostatic rebound.

Palaeo-hydrology andPotential Palaeo-Indian OccupationsPalaeo-hydrological reconstructions for LakeTemagami accord well with those from LacMégantic, a finger lake in southeastern Québec,which has a 12,000-year record of occupation,from Early Palaeo-Indian times onward (Loewenet al. 2005). The key evidence from both lakes, ofrelevance to finding Palaeo-Indian period sites, isthat early post-glacial palaeo-shorelines are wellinland from modern lakeshores. At LakeTemagami, palaeo-shorelines are found near theold outlets identified in the north and northeast,where lake regression has occurred. In contrast,away and farther south of these palaeo-outlets laketransgression has occured. Early to mid-Holocenesites that once occupied these former lakeshoresmay now be partially or completely inundated,depending on slope and elevation.

There is, as yet, no evidence of Palaeo-Indianoccupations in northeastern Ontario inland fromthe Great Lakes. Late Palaeo-Indian componentshave been identified southwest of Lake Temagami,at Killarney and Manitoulin Island on GeorgianBay (Greenman 1943; Julig 2002; Lee 1957;Storck 1984). In northwestern Ontario, surveys ofproglacial Lake Minong have found late Palaeo-Indian Plano sites (Dawson 1983; Fox 1975; Julig1995). Based on site landscape andgeochronology, Pollock (1984) suggests that theLake Abitibi Jordan site (DeHa-8) could havebeen occupied in the late Palaeo-Indian periodfollowing the drainage of proglacial Lake Ojibway.

Systematic archaeological survey of palaeo-shorelines of proglacial lakes or ancestral LakeTemagami remains to be done. Brief inlandsurveys near both palaeo-outlets produced nospecific evidence of Palaeo-Indian occupations.Inland survey in the town of Temagami, whichstraddles the former northeast outlet, produced asingle bipolar decortication flake struck from acobble of a dark brown metasediment (the Portagesite). One clear archaeological indicator of higherwater levels in the North Arm is the Deer Island


Pictographs site, with glyphs ranging from 1.7 to3.1 m above the modern lake level, compared withthose in the Central Hub, which are about 1 mabove the water.

One intriguing hint of Palaeo-Indian presencecomes from the Gillies Depot I site (CiGw-3), 25km northeast of Lake Temagami (Knight 1977).Paul Racher (1988) identified a lanceolate AgateBasin–style point. It has very fine collateral flakingand is manufactured from a thin blade ofbrownish red chalcedony. One face has a largedetached flake, while the other shows extensivewater and sand abrasion (Knight 1977:186). Thissandy site was excavated when Ontario Hydrolowered the water levels in the Montreal River.Although a possible import by a precontactcollector, the site landscape lends support to thetool’s antiquity. It is identified as “nearshore andbeach sediments, 1–20 m thick, deposited duringregression of glacial lakes in less than 50 m ofwater” (Veillette 1986).

If this region was colonized in early post-glacial times by Palaeo-Indian hunter-gatherers,they faced relatively rapid landscape change. Cold,proglacial lakes formed and drained due tocontinued ice retreat and significant isostaticrebound. Settlement choices, travel routes, andresource usage would have been in a state ofconsiderable flux. Tundra vegetation may haveexisted for a relatively short period (Saarnisto1974), followed by an open, patchy early BorealForest with more hardwoods than today (Hall etal. 1994 ; Liu 1990). Zone 1 in the Bear Bogpollen diagram shows white spruce and jack/redpine with some balsam fir and birch in thecomposition of this early forest. Campsites alongpalaeo-shorelines would have a relatively shorttime depth, producing few artifacts. Theexception would be quarry sites where specificbedrock sources were exploited for tool making,as seen in the use of Lorrain quartzite atSheguiandah (Lee 1953, 1957; Julig 2002), thesame Lorrain Formation which outcrops on LakeTemagami, and of a silicious welded tuff or “greenchert” from the nearby the Mt. Goldsmith quarryat the Jordan site, Lake Abitibi (Pollock 1984,personal communication 2014).

Archaic OccupationsAn early Archaic component at the Foxie Ottersite dated to c. 8500 cal B.P. (Hanks 1988), ornear the beginning of the Hypsithermal warmingperiod. A date of c. 5800 cal B.P. was obtained inthe lowest Shield Archaic stratum at the MontrealRiver site (Knight 1977). Cores from Bear Bogand Three Pines Bog show Pollen Zone 3a, datingbetween 8,200 and 3,800 cal B.P., with high levelsof white pine and much lower levels of jack/redpine. Climatic warming caused a shift from aClosed Boreal Forest to a Mixed Conifer–Hardwood Forest. In Liu’s (1990) reconstruction,a Great Lakes–St. Lawrence mixed forest reachedthe edge of the Hudson Bay Lowlands.Physiographic differences underlay forestcomposition, with a prolific increase of white pineon the Canadian Shield uplands and an expansionof white cedar in the Great Clay Belt region. Meanannual temperatures were 1 to 2 degrees Celsiushigher than today, and conditions were likelydrier.

Archaic period hunter-gatherers may haveenjoyed an increase in biomass, different animalspecies such as deer instead of moose and caribou,and an increase in plant growth and plantresources. Warmer temperatures meant a longerperiod between freeze-up and break-up. Sitesassociated with open water could have beenoccupied for longer periods. Slightly drierconditions would mean an increase in forest firefrequency and possibly lower lake levels. Usingpollen-climate transfer functions (Bartlein andWebb 1985), the Three Pines Bog pollen dataindicate a higher mean annual temperature (18.8,compared with 18.0 degrees C today) with a lowermean annual precipitation (650 mm comparedwith 820 mm today) at 7,500 years ago. Bearingin mind the constraints of the Canadian Shieldtopography on settlement patterns and travelroutes, Archaic period hunter-gatherersnonetheless enjoyed warmer, drier, and generallymore favourable environments than did any earlierPalaeo-Indian groups or later Woodland groups.Previously held views that associated ShieldArchaic hunter-gatherers with marginal, harshBoreal Forest conditions (e.g., Wright 1972a) cannow be rejected for northeastern Ontario.


The Witch Point site, on the east centralmainland, is situated on top of a high sand andgravel esker. It may have older Archaic occupationsthan the Three Pines site, as the landform itself isolder and its physiography made the area lessvulnerable to rising lake levels. From a preliminaryanalysis of Witch Point, Archaic period hearthremnants appear as discrete areas of grey ashysand, slightly intruding into the yellow-brown Bfsoil horizon. Artifacts in the deepest layers,tentatively assigned to the Archaic period, includelarge tools, such as excurvate bifaces and largebifacial and unifacial flake knives on GowgandaFormation mudrock collected locally. One hugecobble was pecked into an axe. Conway’s (1982:3)salvage work collected similar large tools. Veinquartz cores, preforms, and shatter are alsoprevalent in deeper deposits. The Kokoko Bayquartz vein (CfHa-50), 7 km to the southwest,may be one source. It is a horizontal lakeside veinof milky coloured, homogenous quartz.

The Three Pines site is situated on a sandybaymouth bar on the west mainland of centralLake Temagami. It was available for occupationby at least 7,500 years ago, based on the beginningof peat formation in the adjacent Three Pines Bog.Extrapolating from the present slope, the lowerterrace and active beach may have extended 20 mfarther south. Lake transgression and longshoredrift have gradually reworked and inundated thefront edge of the site. Therefore, in the Archaicperiod, the Three Pines site provided hunter-gatherers with a larger area of flat, well-drainedterrain for camping and working. Anyarchaeological remains near its old lake marginswould now be inundated.

These processes did not, however, effect theupper terrace occupations. Settlement tracesinclude deep basins of grey sand with cobbles;dark orange-brown sand, often with calcinedfaunal bone; including one hearth with beaver andmammal bone. Discrete areas of orange-brownsand in which several Archaic tools were located,may be the faded remnants of once brighter fire-reddened hearths. Hearths were used for cooking,warmth, and faunal bone disposal.

Archaic period tools include various largerside-notched projectile points; a large stemmed

projectile point or drill; medium to large excurvatebifaces, including one bi-pointed biface; retouchedflakes; flake knives; and a large snub-nosed scraper.Archaic period toolmakers used locally availableraw materials that have sub-conchoidal fractureproperties. These include a distinctivemetamorphosed olive-grey quartz siltstone orquartzite from the Lorrain, Gordon Lake, or BarRiver Formations (leaf-shaped biface, a stemmedprojectile point or drill); a very fine-grained lightgrey Gowganda Formation sandstone (side-notched projectile point); softer greenish greyGowganda Formation mudshale, mudstone, andclayshale (flake knives and a large leaf-shapedbiface); and vein quartz (preform). These materialswere either quarried directly at bedrock outcropsand/or gathered as rounded cobbles found in theglacial drift. While the Three Pines site containedseveral Archaic period large mudrock tools, largecores or chunks of Gowganda Formation mudrockwere absent.

The tools recovered suggest hunting,butchering, and food and hide preparation. Onlymammals and beaver bone were identified fromthe one calcined bone deposit associated with theArchaic occupations. Tool making and primarylithic reduction were not major activities at ThreePines, judging by the low frequencies ofdecortication flakes, thinning flakes, large cores,or shatter of the above-mentioned lithic rawmaterials. If Archaic toolmakers did quarry rockfrom local outcrops on Lake Temagami, these werenot in proximity to the Three Pines site. The largesize of the bifacial tools suggests use of core, ratherthan flake, tools. The need for large cores andpreforms underlies the use of local sub-conchoidalfracture material. Some of the local raw material isprone to step fractures, giving a visual impressionof crude manufacture. In reality, many of theseArchaic period tools were made with skill andprecise control. Nodules of conchoidallyfracturing chert, either HBL or local chert, mayhave been too small for the needs of Archaicknappers.

Middle Woodland OccupationsNeoglacial cooling is represented by Zone 3b inthe Three Pines Bog core, dating from 3,800 to


1,200 years ago. A rapid decline in white pineoccurs, while cooler boreal species, such as jack/redpine, birch, and some spruce, increase. The BorealForest–Mixed Forest ecotone began to migratesouthward from the present day Timmins area,eventually stabilizing 50 km north of LakeTemagami where it currently remains (Liu 1990;Liu and Lam 1985).

The landscape setting of the Three Pines sitewould have offered Middle Woodland periodhunter-gatherers many features favourable toseasonal settlements. The relatively flat upper sandterrace provided well-drained, level terrain forerecting habitation and work structures. Thegentle slope of the sand beach provided ease ofaccess from land to water and a good landing placefor canoes. It is warmer and sheltered, beinglocated on an L-shaped section of the shorelinefacing south and southeast. It is protected fromthe cold north and northwest winds and from theprevailing southwest winds. A 120-degree vista ofthe west central section of Lake Temagami allowsfor inhabitants to observe mobile game and peopleentering or leaving the Northwest Arm. Fish couldbe obtained beyond the shallows and especially inthe deep northwest channel. The adjacent peatbog and the upland forest would provide plantresources for subsistence, construction purposes,and fuel. Middle Woodland period hunter-gatherers probably selected this locale for the samereasons that earlier Archaic groups did. However,the amount of usable space was much reduced forMiddle Woodland groups due to laketransgression.

Middle Woodland occupations are wellrepresented at Three Pines. Settlement tracesinclude extensive grey ash and cobble deposits,representing traces of high-temperature hearthsand living floor smears. Localized areas of fire-reddened sand with calcined faunal bone indicatecooking or disposal hearths or perhaps their cooleredges. A pottery-making area is suggested by aunique deposit of brown clay, sand, and smallstones. A single post mould or pit was alsoidentified. Ground squirrel burrows, tree rootgrowth, erosion, and subsequent cultural activities(e.g., modern garbage pits) have affected thepreservation, boundaries, and spatial distribution

of these remnant settlement structures.The extensive nature of these settlement traces

suggests considerable time depth and repeatedseasonal occupations by Middle Woodlandhunter-gatherers. Based on the landscape settingand identified faunal taxa, the Three Pines site wasprobably suitable for occupation at any time of theyear. This interpretation is consistent with thefaunal taxa identified from specific MiddleWoodland contexts, including beaver, moose, andwhite-tailed deer—three game species that couldhave been captured during any season of the year.Beaver was better tasting in the colder months andtheir winter pelts thicker (Job Halfaday, personalcommunication 1981). The inhabitants probablyalso exploited bear, wolf, pond turtles, and fish.The latter resource is inferred from two clusters ofunmodified pebbles identified as net sinkers. Ofthese taxa, pond turtles are a warm seasonindicator. Thus, Middle Woodland groups mayhave used the Three Pines site during both thewarmer and the colder seasons.

The Middle Woodland period is marked bythe introduction of ceramic technology. At theThree Pines site, pottery vessels include pseudo–scallop shell stamped, linear stamped, dentatestamped, and combination linear and dentatestamped designs on the lip and/or exterior. Rimvessels show slightly outflaring rims and straightwalls, while one plain body vessel shows a conicalbase and straight walls. The identification ofpottery temper lithology was a useful adjunct todecorative motif as a means of separating potterysherds into individual vessels. For temper, pottersused granitoid rock, or mixed sand and granitoidrock. Only one vessel had the dark grey speckledNipissing diabase temper common to the LateWoodland pottery.

Laurel pottery tended to be in spatially distinctgroups of different pottery designs and temper. Insome cases, tree roots, rodent tunnels, and modernbackfill from pits served to protect the potteryclusters from erosion and subsequent culturalactivities. While functional differences and thepreferences of individual potmakers may bereflected in the different pottery vessels and in thespatial clusters, there is little clear stratigraphicbasis for temporally ordering the different pottery


vessel deposits. The Point Peninsula potteryassemblage at the Frank Bay site included vesselssimilar to certain pseudo–scallop shell and linearstamp Rim Vessels from Three Pines, while vesselsfrom Ridley’s Primary Transitional stratum werecloser to a dentate stamped Body Vessel. As noted,this latter vessel may bear similarities withsouthern Ontario Saugeen ware.

Middle Woodland flintknappers madeextensive use of fine-quality, colourful, non-localHBL chert from farther north obtained by tradeor travel. There is also some use of a poorer quality“local” chert and other unidentified cherts. Allthese hard, conchoidal-fracturing rocks, availableonly in small nodules, were used for smalltriangular and trapezoidal scrapers, retouchedflakes, and small side-notched projectile points.

Local clayshale, mudshale, sandstone, andsiltstone were used for flake knives, while cobblesand pebbles were used for grinding implements,abraders, celts, netsinkers, and utilized cobbles.This material occurs as beach cobbles and asbedrock outcrops.Vein quartz was also used for theoccasional side-notched projectile point andbiface. One source of this material occurs within8 km (the Crystal site).

Late Woodland OccupationsContinued climatic cooling characterizes thisperiod. Zone 4 (c. 1,200–0 cal B.P.) of the ThreePines Bog pollen diagram shows an increase inspruce and fir, with a slight decrease in white pine.A Mixed Conifer and Hardwood Forest occupiedthe Lake Temagami area, while the Boreal Forest–Mixed Forest ecotone remained farther north upto the present day. Lake levels were only slightlylower than today.

Although they have been more subject toerosion and other processes, Late Woodlandoccupations of the Three Pines site are less intenseand of shallower time depth than those of theMiddle Woodland period. Settlement tracesinclude two rock structures. One is a deliberatelyplaced line of rectangular rocks that may havebeen used for a working surface, hearth platform,or windbreak. Another is a semi-circle of cobblearound an ash and sand deposit. This may havebeen a hearth with a stone outline or a hearth

windbreak, as the cobbles are situated on thesoutheast. Ethnographic examples of Cree andOjibway bush camps show that the doorways ofconical tents are oriented to the southeast (Chism1978; Gordon 1980, 1985). The remnants ofseveral fire-reddened sand cooking and calcinedbone disposal hearths also contain Late Woodlandcomponents.

Late Woodland pottery at the Three Pines siteincludes cord-wrapped stick–impressed rimdesigns and a cord-wrapped paddle–textured bodydesign. One distinctive characteristic of the threevessels is the use of a dark, speckled temper, theresult of a high percentage of grey mafic mineralsin the Nipissing Diabase temper. While Blackduckvessels are found at the Frank Bay site (Brizinski1980) and at Lake Abitibi (Ridley 1966), theseThree Pines vessels do not have characteristicBlackduck motifs. However, at least oneBlackduck vessel was identified at Witch Point.

At Three Pines, HBL chert, unidentifiedchert, and “local” chert were used for small side-notched projectile points; scrapers, generallysquare in form; and retouched flakes. The lack ofclearly visible stratigraphic separation of LateWoodland occupations at Three Pines from otherprecontact occupations makes it difficult to assignadditional lithic tools to this period. As LateWoodland occupations units of stratification werespatially discontinuous, there was no stratigraphicbasis for temporal ordering these occupationtraces.

Another type of Late Woodland pottery vessel,linear-trailed collars and castellations, is found atthe nearby Sand Point (Conway 1986) andWitchPoint sites. The linear trailed sherds from ThreePines site are too small to determine whether theyare from similar vessels. The Witch Pointexcavations produced collared vessels, some withcastellations, decorated with shoulder punctatesand oblique and horizontal incised lines. Whileseveral resemble Sidey Notched and HuronIncised designs, they do not exactly matchsouthern Ontario vessels in construction, paste, oroverall form. This ware is generally cross-datedwith the pottery sequences of horticulturalIroquoian sites in southern Ontario (Conway1982, 1986; Ridley 1966). Such vessels are well


represented in the upper strata of the AbitibiNarrows, Ghost River Garden, and Frank Bay sites(Ridley 1954, 1958, 1966; Brizinski 1980).Recent work in the Abitibi-Témiscamingue areaof Québec dates these vessels with affinities toIroquoian Tradition stylistic traits to between A.D.1300 and 1650 (Côté and Inksetter 2001). Directand indirect trade, overwintering of CanadianShield hunter-gatherers with horticultural Hurongroups, and Iroquois raiding are all possible socialinteractions underlying this late precontact sharedpottery tradition (Côté and Inksetter 2001; Fox1990; Fox and Garrad 2004; Heidenreich 1971).

The Witch Point site revealed certain itemsthat suggest ceremonial or ritual activities. Theseinclude red ochre paste, thin-walled potteryvessels, red ochre nodules, and clear quartzcrystals. Of particular interest are the dense cobblestructures, including a rock pavement with anearby burial of the unburned partial skeleton ofa young dog in a sterile small pit of fine sand. TheWitch Point landscape offers a natural quay forgroups travelling up and down the lake and a flat,well-drained living area for short- but not long-term duration. It may well have served as ameeting place for families who practised rituals tosymbolize and reaffirm group alliances.

Early Contact Period: 1600sThe arrival of European explorers, Jesuitmissionaries, and fur traders in northern NorthAmerica brought durable material items as tradegoods. Manufactured items of sixteenth- orseventeeth- century European origin have not yetbeen identified in artifact assemblages on LakeTemagami. The lake is upland from early majorfur trade routes between the St. Lawrence Riverand the upper Great Lakes. Prior to 1615, Huronpeople traded with the French via Algonquianmiddlemen using the Lake Huron-LakeNipissing-Ottawa River trade route (Fox 1990;Heidenreich 1971). Lake Temagami people maynot have directly encountered French explorersand fur traders at this time, but probably had traderelations with the Nipissings, who did (Champlain1922:3:41). A 1640 account from the JesuitRelations identifies localized named groups innortheastern Ontario, of which two, the

Outimagami and the Timiscimi can be reliablyassociated with the study area (Bishop 1994a;Thwaites 1896-1910:18:227-229). The Frenchopened a northern inland post at Nighthawk Lakein 1673 and operated another at LakeTimiskaming from 1679 to 1688 (Bishop 1994a).This latter post was located at the mouth ofMontreal River. It was visited in 1686 byChevalier de Troyes on his overland excursionfrom Montreal to capture English Hudson’s BayCompany posts on James Bay (Mitchell 1977:8).The English had established earlier fur trade postsbetween 1668 and 1673 at the mouths of theRupert, Moose, and Albany rivers (Bishop 1994a).

Historic Period: 1700sWith the objective of ending hostilities betweenthe French and their allies and the Iroquoians,the Great Peace was held in Montreal in 1701.This gathering of 1,300 Natives and 1,200French included Nipissing and Temiskamingpeople, according to the French classification of25 signatories. However, as Bohaker (2006)points out, the pictograph signatures on thetreaty actually number between 38 and 39 Nativegroupings. Perhaps some of these signaturesrepresented groups inhabiting the LakeTemagami area. Certainly stories of hostileIroquoian raids are part of the oral history ofTemagami people and surrounding groups(Conway and Conway 1989). Stories told byBear Island elder Bill Twain are associated withspecific locales, such as High Rock Island andnearby Portage Bay (Gordon 1995a).

The fur trade continued under French controlwith a Timiskaming post operating intermittentlyfrom 1720 until 1760. In 1785, the Hudson’s BayCompany also built a Timiskaming post (Mitchell1977). One artifact from the Three Pines site maydate to this period: an eighteenth- to nineteenth-century black English musket flint (Noël Hume1980).

Historic Period: 1800sAt the Three Pines site, an 1846–1876Henderson-Montreal clay pipe stem (Smith 1986)was found with two punched-out gunspalls, non-calcined beaver, and deer bone in a hunting camp


hearth. Ball clay pipe fragments and two rusteddecorative metal items were nearby. Another pipestem was collected from the Cross Bay site. At theWitch Point site small glass beads, a crooked knife,and two rolled copper tinkling cones wereretrieved in the 1993 excavations, while Conway(1982) reported other mass-made objects of thisage obtained through the fur trade.

Fur trade establishments on Lake Temagamiitself appear quite late (Hodgins and Benidickson1989:29-35). An official HBC outpost operatedsporadically on Temagami Island from 1834 to1857 (Hodgins and Benidickson 1989; Mitchell1977).

In 1833 Samuel Peck and CharlesHarris, two free traders working fromPenetanguishene who had been regularlytrading on Nipissing, moved up to LakeTemagami and wintered over there into1834 “with a large outfit.” […] Peckand Harris were probably the firstEuropeans to winter on LakeTemagami. Chief Trader AngusCameron at Fort Timiskaming movedquickly to have Harris bought off andenlisted into the service of the HBC[…] After the winter of 1834-5, theharassed Peck left Temagami for good[Hodgins and Benidickson 1989:29-30].

Bill Twain told a story which explains somedifficulties in operating this early outpost:

One day some Indians traded at thepost, but they lost what they hadpurchased in the water when their boatsoverturned. So the Indians went backfor more goods at the post. The traderrefused to replace the lost items andthreatened to shoot them. So theIndians did not go back to this post.Instead they went to Ville Marie[Timiskaming] and to Nipissing. Thetrader reported that there were nolonger any Indians at Temagami, andthe post was closed. It was a long timebefore the next post was built [Gordon1995a:13].

An expanded HBC post was established in 1857also on Temagami Island. Further research will beneeded to date the cobble structures and walk-incold storage observed at CfHa-1 and determine ifthis was the location of both the early outpost andthe later enlarged HBC post (see Gordon 1995afor discussion). The HBC post moved to BearIsland (CfHa-2) in 1876 (Hodgins andBenidickson 1989; Mitchell 1977).

Euro-Canadians located their commercial andreligious establishments near Teme-AugamaAnishnabai summer gathering areas characterizedby large expanses of flat, well-drained sand. Onewas at Wabikon on the south shore of TemagamiIsland, another at the southwest corner of BearIsland. In 1891, Father Paradis built his Missiondu Sacre Coeur on the sandy outwash plain ofFerguson Bay, also a summer gathering area,beside the Anima Nipissing River outlet (LakeTemagami site). The headstones from theTemagami Island Burial Ground date from 1889to 1892. One is carved with a rose, a second witha cross and the monogram “IHS,” all Christiansymbols. Father Paradis continued the mission anda farm until 1924, when St. Ursula’s RomanCatholic Church was built on Bear Island(Hodgins 1976).

In Bill Twain’s words, people camped all overthe land, as seen in the hundreds of campsitesidentified by Bear Island residents and other nativeinformants for the pre-1900 period illustrated inMacdonald’s (1993) Historical Map of Temagami.These camps are on large lakes and small interiorlakes, and on islands and the mainland. They werelinked by nastawgan, a vast web of travel routes,including summer portages, winter-only interiortrails, and open water or lake ice trails (Macdonald1987).

The traditional way of life described below isreminiscent of Father Jerome Lalemant’sdescription of the Nipissings in 1640 (Thwaites1896-1910:21:230-240).

Many of them [Craig Macdonald’sinformants] had been born and raised inthe bush and learned to hunt and trapfrom their elders, but by the 1970sthose ageing trappers were the last link


with the ancient world of theAnishinawbeg. For centuries, theAnishinawbeg people would spend thewinters in small camps of two or, atmost, three families and work theirtraplines. In the spring they wouldharvest maple syrup and in the summerfamilies from throughout the territorywould gather on Ka-tay Te-mee-ay-ga-maw Minis, an island in LakeTemagami. They harvested whitefishand cultivated maize, a small, hardyvariety of corn, and in the fall returnedto their winter camps [Jenish 2006].

Modern Period: 1900sTraditional Use Campsites. From a material cultureperspective, change is seen in the increasedfrequency of durable and datable, mass-produceditems at hunting and fishing campsites.Twentieth-century hunting camps at the ThreePines site could be identified by a range of goods,including coins. At the Witch Point site olderhunting camp hearths were indicated byunburned beaver and moose bone and artifacts incharcoal-laden deposits. Surface remains at theStove site include, among other things, a metalwood-burning stove, a steel wash basin, and an oillamp base, suggesting a winter hunting camp(1930-40s?) situated 40 m in from the currentlakeshore. Traditional construction techniques areseen in a small, semi-subterranean structure ofvertical wooden poles with a dirt-covered roof forcold storage at the Cache site, 40 m inland fromBarnac Lake. All of these sites suggest mobilehunting camps, in contrast to thecontemporaneous, but more permanentsettlement at Austin Bay with its log cabins androot cellars. The traditional way of life waschanging, especially by the 1940s when theinterior, winter-only trails ceased to be maintained(Macdonald 1987). Today, although most familiesno longer live “on the land,” the Teme-AugamaAnishnabai maintain a strong relationship withn’Daki Menan and their history.

Euro-Canadian Development.Major changes occurin the twentieth century. Along with theintroduction of new artifacts, there are differenttypes of sites and landscape transformations,occasioned by the arrival of larger groups of Euro-Canadians. These new groups came north usingnew transportation methods, to work in large-scale resource extraction industries such as loggingand mining. The pollen record shows a rise inragweed pollen (Ambrosia) in the top part of theThree Pines pollen core and in the “Lake 306”core (Hall et al. 1994). This reflects the impact onregional vegetation of late nineteenth- andtwentieth-century land clearance for logging andmining at Lake Temagami itself and, farming inthe Little Clay Belt. Lake levels also changed asearthen dams of the late nineteenth century werereplaced with wooden and later concrete dams forhydroelectric power. Accessibility to the area byrail (TNOR) and the islands by paddle wheelsteamer led to a tourist industry (Hodgins andBenidickson 1989). Sites and finds indicative ofall these activities include the Owaissa Sawmill sitenorth of Temagami; prospectors’ metal claim tagsrecovered from the Witch Point site; and themelted and twisted remains of the Lady EvelynHotel, which burned down in 1912.

Wilderness recreational use became a majoractivity on Lake Temagami with the developmentof wilderness camps for Canadian and Americanyouths in the 1930s at Ferguson Bay andWabikon, among others (Hodgins 1987). Shorter-term tourist activities are indicated at the ThreePines site by food containers, cameras parts, picnicrefuse, boating refuse, and lightweight tent parts.Picnickers, boaters, and overnight campers duggarbage pits and tent drainage ditches and built aseries of picnic hearths, eventually resulting in thelarge picnic fire mound. These activities served insome cases to destroy and in other cases to protectthe underlying precontact remains. Canoecampers and others often use the same portages,trails, and campsites as were once traversed andoccupied by families of hunter-gatherersthroughout the long prehistory and history of thearea—one of change and continuity, witnessed bythis beautiful deep water lake.


Approaches to the Archaeological Record:Discussion and Implications

for Future Research

The “Constraints” ofCanadian Shield ArchaeologyIn northeastern Ontario and adjacent areas, thenorthern forested (Canadian Shield) landscapepresents many constraints and challenges toarchaeologists in finding, analyzing, andinterpreting archaeological sites (e.g., Brizinski1980; Hanks 1988; Knight 1977; Kritsch-Armstrong 1982; Marois and Gauthier 1989;Pollock 1976, 1984). Sites tend to be thinlydispersed over vast areas. The rugged terrain,exposed bedrock, and dense forest cover makefinding them difficult. The podzolic soils of theBoreal Forest have weakly developed soil horizons,generally characterized as thin, shallow, andcompressed (Canadian Soil Survey Committee1978). Periodic forest fires destroy the organiccontent and further serve to collapsearchaeological soil layers (Hinshelwood 1997).Over time, soil acidity disintegrates most organicmaterial, leaving only lithics, ceramics, andcalcined faunal bone, often in minute quantities.Finally, archaeological sites are easily disturbed bymodern development, recreational activities, andseasonal lake level changes, both natural andhuman-made (e.g., Knight 1977; Pollock 1984;Ridley 1954, 1966).

In contrast to southern Ontario, at the time ofthis fieldwork, only a dozen sites from LakeNipissing to Lake Abitibi had been fully excavatedand published. They are geographically widespreadand differ in function (e.g., quarry sites vs.habitation sites) and time depth, making it difficultto compare and correlate artifacts. As a result,artifact comparisons reveal few similarities withother site assemblages. Furthermore, it is difficultto obtain organic material from reliable precontactcontexts for radiocarbon dating (Hinshelwood1997); hence, absolute dates are rare. All of thesefactors complicate attempts to sort thearchaeological record into meaningful temporaland cultural components. How then can we get tobigger questions of long-term culture change withsuch a sporadic and discontinuous record?

Using a Harris Matrix on Precontact SitesResearch conducted at Lake Temagami hasattempted to overcome some of the limitations onthe archaeological record noted above.Investigations at the Three Pines site focused onthe natural and cultural formations processes thathave interacted to create the site and its landscapesetting over time. In this approach, thearchaeological record is a time series of physicaldeposits, whose form and nature reflect thecumulative total of all events, activities, andprocesses that have interacted to create thearchaeological site as it exists today (Gordon1991a, 1991b). The archaeological site is apresent-day phenomenon, but its past can only beknown through its stratigraphic record (Butzer1982; Schiffer 1987), similar to other stratigraphicsources of information about the past, such as thelithological column or the pollen core.

At the time of this field research (Gordon1990a, 1991a), a “stratigraphic/contextualapproach” required a complete shift in perception,as northern sites had been viewed in moreconventional terms (e.g., Gordon 1985). Theterm “stratified” had been reserved for specificlandscape contexts in northern Ontario (James V.Wright, personal communication 1991), such asthe Pic River or Frank Bay sites (Ridley 1954; J.Wright 1967a). In these locales, riverine-depositedsediments form lenses of sterile soil betweenburied dark Ah soil horizons and serve to separateartifact-bearing lenses (Ridley 1950-53; Wright1972b). Because some sites were thought to lack“cultural stratigraphy” (Brizinkski 1980),excavated artifacts had been temporally sorted bydepth, varying from 4 cm to 12 cm thick arbitrarylevels (e.g., Brizinski 1980; Knight 1977); bydiscrete artifact clusters, referred to as “horizontalstratigraphy” (Pollock 1976); or by continuous soilhorizons, as is the case for the Mattawan stratumof the Frank Bay site (Ridley 1950-53). However,by viewing the Three Pines site as a time series ofminute physical deposits and culturalmodifications, it became feasible to use a HarrisMatrix stratigraphic analysis (Harris 1979a). It isnot that the Three Pines site is an exceptional sitebecause of its stratigraphy, but rather that asystematic stratigraphic approach has been


explicitly employed to reveal the sequence ofoccupations.

A Harris Matrix analysis, combined withdetailed excavation procedures andgeoarchaeological analyses, provides a powerfultool to sort out artifacts and other archaeologicalsediments into a relative temporal framework.Although constructing the matrix is time-consuming, it has many tangible benefits. It helpsovercome some of the challenges of collapsed orcompressed stratigraphy (Hinselwood 1997;Pollock 1976) and spatial discontinuity (Pollock1976; Brizinski 1980). Modern “disturbances” cannow be included as part of the archaeologicalrecord. Stratigraphically examining occupationswith a known time depth, such as the Modern (atthe time of the field work covering 86 years) andHistoric periods (150 years), gives an appreciationof the minute scale of these physical traces. Thisapproach also assisted in determining whichnatural and cultural site formation processes wereacting uniformly over the site area, versus thosethat were localized. Some served to destroy andothers to protect the older deposits. Despite itsshallowness, the Three Pines site reflects certainartifact patterns evident on sites with much deeperdeposits and conventionally defined stratigraphy,such as the Frank Bay site (Ridley 1954) and theMontreal River site (Knight 1977). At the ThreePines site subtle settlement and disposal patternsthrough time are evident, while lithic tool trendsand changes in procurement strategies can bestratigraphically verified.

Clearly this method alone cannot overcome alllimitations on interpretations imposed by shallowdeposits, seasonal reuse, multilinear sequences, the(in)visibility of strata, and artifact movement in asandy matrix. However, it does offer a systematicand objective approach to these conditions. Oneproblem at Three Pines is that the discontinuousnature of the units of stratification did not allowfor temporal separation of individual MiddleWoodland units or individual Archaic period unitsacross the site. Soil micromorphological analyses(Courty et al. 1989) might help detect moreminute patterns of stratification. Absolute datingwould also be useful, provided samples could beobtained from reliable stratigraphic contexts.

Would different excavation procedures havebeen more compatible with applying a HarrisMatrix analysis in the laboratory? At the ThreePines site systematic recording procedures using 3cm vertical levels and horizontal one-metre squaressubdivided into quadrants were employed. Forconsistency, the same excavation and recordingtechniques, but with more emphasis on recordingsoil changes, more vertical profiles, and morephotographs, were employed at Witch Point.Harris (1979b) later proposed an excavationstrategy which defines single-layer plans coupledwith stratigraphic sequencing at the time ofexcavation, identifying the soil or architecturalmatrix first, then its dimensions. In contrast, theReid method used here, started with the spatialdimension (quadrant/level) and then identifiedsoil matrix second. Which is better is not known.Certainly applying Harris Matrix principlesduring field excavation would be moreadvantageous. But what matters most is flexibilityin systematically recording and collecting a widerange of data in order to reproduce a three-dimensional model of the site, its soil contexts,and its contents in the lab. Many individualobservations in the field only became repetitivepatterns after considerable comparative analysis inthe lab. All field information—vertical depth,horizontal distribution, and soil context—needsto be recorded in a systematic and sensitivefashion, at a scale appropriate to a podzolic soilmatrix being used seasonally, by small groups, fora few days or weeks, over years and centuries.

Certain minor modifications were made whenapplying the Harris Matrix analysis to the ThreePines site. The generic term “unit of stratification”was used for all deposits, and the interpretation ofevents and/or processes was placed inaccompanying charts and text. There was no apriori equation of any one unit of stratificationwith a single deposition event or cultural ortemporal period. Thus the 3 cm vertical recordinglevels were used as depth indicators, separatingartifacts in one 3 cm level from those in another,even within the same unit of stratification. Whatcauses some confusion for Harris Matrix scholarsis that “interfaces,” or the sharp contact betweendifferent deposits, were not separately numbered.


However, the important role of interfaces indetermining stratigraphic succession wasmaintained. For example, in the accompanyingtext and charts, pits are described as two separateevents—their excavation and subsequent infilling.The reason this minor modification was adoptedwhen first presenting this material (see Gordon1991b) was that separately numbering interfaceswould 1) over-represent modern or historic periodunits of stratification where interface boundarieswere more visible and 2) give the incorrectimpression that natural boundaries, such as soilcolour shifts due to weathering in the lower levels,were being treated as culturally created interfaces.Perhaps complexity is not a bad thing, however,and now that more researchers are familiar withthe Harris Matrix, one might recommendadhering more strictly to Harris’s numberingsystem.

One of the key reasons for doing detailedstratigraphic analyses is to allow for a more solidchronological sequencing of artifacts by which totest and revise “diagnostic” and other artifactpatterns. Where historically documented,architectural sites are at an advantage is in theidentification and dating of artifacts. I hope thatin the future, as the Harris Matrix is applied tomore precontact sites, there will be considerableinteraction between stratigraphic and artifactanalyses, to sort out and identify temporallysignificant artifacts.

This method could be applied retroactively toexcavated sites. Ridley’s (1950-1953) field notesindicate that the Frank Bay site would make anexcellent candidate, with its long sequence ofoccupations, buried Ah horizons, and settlementfeatures. For example, the Mattawan stratumcould be subdivided into different depositionalevents, separating shallow from deeply buriedtools, which may help to better distinguish MiddleWoodland from Archaic period tools (e.g., Wright1967b). Stratigraphic analysis could potentiallyhelp with the dog burials, first assigned to theHistoric period, based on spatial proximity to furtrade items, but later radiocarbon dated to 600years earlier (Brizinski and Savage 1983). One keylesson from the Three Pines site analysis is thatone soil horizon or vertical level is not necessarily

equivalent to one depositional phase or timeperiod, and that spatial association used by itself asa temporal indicator can be misleading.

Lake Level Change, Archaeological SurveyMethods, and Site InterpretationLake level change has long been recognized asimportant to northeastern Ontario archaeologicalsites. For example, Ridley (1954) noted that theMattawan Stratum (mostly Archaic period) of theFrank Bay site on Lake Nipissing had evidence ofseasonal or intermittent inundation, which hepostulated resulted from northeast storms. It isplausible that the water-smoothed, white quartzitetools with iron staining in the Mattawancollection actually relate more to sustained highwater levels in early post-glacial or mid-Holocenetimes. Ridley also suggested that “the history oflake levels may have significance in the location ofearly habitation sites on the [Lake Abitibi] shores”(Ridley 1966:18). In modern times, fluctuatingwater levels have had destructive effects on LakeAbitibi sites, particularly the Abitibi Narrows site(Ridley 1966), the Jessup site (Kritsh-Armstrong1982), and the Jordan site (Pollock 1984).Artificially lowered levels in the Montreal Riverled to the discovery of a lanceolate point at GillesDepot of Late Palaeo-Indian origin (Knight 1977;Racher 1988).

At the Three Pines site some Archaic periodcomponents were just a few metres closer to thelakeshore than Historic period ones. A pollen corefrom the adjacent Three Pines Bog showed a 4 mlake level rise steadily inundating the site over7,500 years. However, as detailed herein,predictive palaeo-hydrological modelling andresulting pollen core analyses have shown thatisostatic rebound has been a major cause of lakelevel change in the Temagami basin. Other factorsdo of course play a role in lake level variability.Loewen et al. (2005) point out the complexinterplay of factors, such as isostatic rebound,hydro-climatic changes in seasonal precipitationand evapo-transpiration, local vegetationdifferences, and aquifer characteristics of eachwatershed. They advocate a basin-by-basinapproach to reconstructing lake levels over time.

The predictive “tilt model” employed here


offers archaeologists a relatively simple, easilyadapted approach for reconstructing ancientshorelines. It concentrates inland survey to areaswith a higher probability of yielding Palaeo-Indianor early Archaic sites—that is, emergentshorelines. Modern lakes with only a southwarddrainage would make the best candidates forapplying this tilt model, as such lakes are likely toshow the greatest isostatic rebound–relatedchanges in shorelines over time. Lake Temagamidrains both to the north and south from a centralsill, adding complexity to the model. While digitalterrain maps (DEM) are now available toresearchers, bathymetric values still need to betaken from base maps for most Ontario lakes. Wedid note several errors in the topographicmapping, which are best corrected by field survey.Geologists and archaeologists work on quitedifferent scales, meaning a lake level rise of 17 mmay be considered slight by the former, while itcan be substantial for the latter. For heritageresource planning or GIS predictive modelling(e.g., Dalla Bona 2000; Hamilton 2000), using astandard inland distance from modern lakeshoreswould not be appropriate for site prediction.Palaeo-shorelines at Lake Temagami were noted atwidely varying distances inland depending onlocal topography and location with respect topalaeo-outlets.

Palaeo-hydrological research on LakeTemagami reveals emerging shorelines near thenorthern palaeo-outlets while the rest of the basinis submerging. This has significance for howfuture archaeological surveys will be designed andhow known precontact sites on Lake Temagamicould be interpreted (e.g., Dalla Bona 2000).While surveys along modern shorelines aresuccessful in locating very small to very largehabitation sites, gaps in the cultural sequences willresult if only this method is used. Potentialhabitation sites of early colonizers of LakeTemagami are likely to have low artifact counts(except at bedrock quarry sites), be more deeplyburied, and be found inland near the palaeo-outlets. Perhaps the site selection criteria identifiedin this research for lakeside sites (well-drained levelground, north wind protection, ease of access towater) may extend to these hypothetical early sites.

With modern logging roads, mining, andpipelines, inland areas are increasingly accessible.Not all interior locales have been impacted bythese activities.

One benefit of inland survey at LakeTemagami has been recording numeroustraditional use sites of the Teme-AugamaAnishnabai. These twentieth-century sites, tracesof traditional land use and settlement patterns, arerapidly disappearing. Containing both above-ground and in-ground wooden structures, they aresubject to the destructive effects of weathering,forest fires, and modern development. In my view,one important obligation for archaeologists is todocument all material traces of humanoccupation, both Native and non-Native, throughall time periods—precontact, historic, andmodern.

Examining traditional use sites, especially withthe help of local informants, also suggestsalternative places to look for precontact sitesbeyond the modern lakeshore. Inland traditionaluse sites are not easily spotted from open water byboat survey. In summer, dense vegetation at theshore often conceals inland flat areas. When thelake, bogs, and ground are frozen, alternativechoices are available for hunter-gatherers, whichcan influence winter trail and camp site locations(Macdonald 1987, 1993). Repeated human usagedirectly impacts forest regeneration, creating areasof open vegetation or secondary regrowth: anotherclue to finding archaeological sites.

Researchers will need to be very imaginativein designing survey strategies. It is easier to seelandscape features when there are no deciduousleaves but also no snow cover. Fall or spring pre-survey, even after freeze-up, may be useful inconjunction with testing in open water periods.In designing surveys, the importance of laketransgression since the mid-Holocene is alsocritical. Lake cores taken near bogs would showlake sediments over peat bog deposits insubmerging parts of the lake basin. Surveys atnatural low water levels in April or in co-ordination with hydro-electric companies mayprove helpful in finding lost edges of lakeside sites.One useful approach might be to compare mapssuch as Macdonald’s (1993) Historical Map of


Temagami, which shows water levels before hydro-electric development, with modern shoreline data.This might reveal differences in the nastawganplacement and traditional use locales (Macdonald1987, 1993) compared with modern recreationalmaps of canoe routes, portages, and canoe campsites. Underwater work is also an option (Loewenet al. 2005).

Understanding the local effects of lake levelchange is important for site interpretation. Lakelevel change may provide a method of relativelydating rock art, such as the higher-elevation DeerIsland pictographs at the north end comparedwith the lower-elevation pictographs in theCentral Hub. Receding lakeshores may helprelatively date deeply deposited artifact clusters onlow-elevation emerging shorelines, such as theLake Temagami site. Lake level change has adifferential impact on artifact distributiondepending on landscape context and slope. WitchPoint is an esker, an older and higher landform,less affected by rising water levels than the lower-elevation Three Pines and Sand Point baymouthbars. On these two sites, lake transgression mayprovide a method of relatively dating specificarchaeological deposits. Lithic raw materialsources, such as the grey siltstone at Whitefish Bayand the lakeside quartz veins at the Kokoko Bayand Crystal sites, are also differentially affected bylake level change, depending on their locationrelative to palaeo-outlets. In the absence of specificpalaeo-hydrological reconstructions, one needs tothoroughly examine any site landscape as if it wasboth emerging and submerging, to cover allpossibilities.

The “Strengths” ofCanadian Shield ArchaeologyMany of the “constraints” on the archaeologicalrecord of the northern forest (Canadian Shield)described above are actually clear reflections ofhow groups utilized the precontact landscape.From an examination of Ojibwa and Cree hunter-gatherer adaptive strategies (Feit 1973, 1993,2004; Irimoto 1980; Jenkins 1939; Rogers 1962,1963a, 1963b, 1966, 1969, 1973; Rogers andBlack 1976; Speck 1915a, 1915b; Tanner 1979),certain features emerge that are of critical

importance to conceptualizing the specific natureof this archaeological record. These include the

1) extreme mobility of northern hunter-gatherers;

2) small size of habitation/hunting groups;3) seasonal nature of the economic round;4) importance of the land–water ecotone;5) desirability of certain types of locales; and6) short-term usefulness of any one locale.

Any one site could only be occupied briefly,measured in days or a few weeks. A site’s usefulnesswould be dependent not only on local, seasonallyavailable resources, but also on the short-term,near-camp resources such as fish, snowshoe hare,and firewood. The number of occupants at anyone locale would be quite low. While summergathering locations may support a larger numberof inhabitants, the family or kin groups wouldoccupy spatially discrete areas. In all cases, onlyminor, discontinuous traces of each occupationwould be evident at any one site.

From archaeological surveys on LakeTemagami, it appears that precontact humanactivity is concentrated at or near lake or rivershores, both ancient and modern. The choice ofcamp sites is seriously constrained by topography.While a small travel camp can occur whereverthere is enough ground to pitch one tent, such asthe Daily site or Island 245, only certain places oneach lake are likely to have repeat occupations.Gaps in usage would be anticipated to allow forregeneration of required resources. Highlydesirable places for habitation are those that offera reasonable expanse of well-drained, level ground;north wind protection; and ease of access. Owingto surficial geology, these are rare, occurring onsand points, eskers, and outwash plains (e.g.,Three Pines, Sand Point, Witch Point, and LakeTemagami sites).

These sites contain remnants of occupationsover different seasons, years, and centuries. It is forthese reasons that such sites are selected forexcavation. They are easier to excavate, yieldhigher artifact counts, offer a multi-componentsequence of occupations, and provide informationon long-term culture change. However,archaeological survey, ethnohistorical sources, and


local informant information show a much widerrange of site type, site usage, and site location. Thepoint is that if only highly desireable habitationlocales are chosen for excavation, our understandingof the minute scale of deposition, density ofartifacts, temporal scale, diversity of sites, and wide-ranging use of the landscape and settlementpatterns will be constrained, if not skewed.

Any archaeological site is but one stop in awhole series of habitation and special purposesites. Basic daily activities may be the same, butsite-specific or seasonally derived activities wouldvary. Differential preservation of organic versusinorganic materials eliminates evidence for entiretypes of activities. Therefore each archaeologicalsite contains only minor traces of material cultureand other evidence of site usage, representing buta fraction of all the human activity that occurredon the site, which in turn is but a small part of allhuman activity on the lake or in the wider region.Each site therefore offers a tiny but uniquesnapshot of the complex dynamics of hunter-gatherer economic, social, and spiritual behaviour.

Although the challenges are many, we canappreciate the distinctive qualities of northeasternOntario archaeological sites. These are as follows:

1) The landscape concentrates potentialhabitation locations to shorelines, bothancient and modern.

2) The most desirable locations couldpotentially hold thousands of years ofhuman history, from soon after deglaciationto modern times.

3) The archaeological site reveals a time seriesof subtle traces of past human behaviour,both individual and collective.

4) In these shallow deposits, “modern,”“historic,” and “precontact” tend to form acontiguous record, giving us theopportunity to document all culturaloccupations of a site.

5) Each archaeological site will offer its ownunique temporal and stratigraphic record.

Thus each site as a whole can be viewed as aunique record and given an equal role in thedeveloping mosaic of Canadian Shield prehistory,history, and cultural transformation.

Acknowledgements. My appreciation goes toconference organizers and editors Alicia Hawkinsand Bill Fox for giving me this opportunity tojourney back to archaeology; and to JockMcAndrews and Ian Campbell for theirinvaluable, ongoing contributions to this project.My thanks to the many reviewers, who includeAndrew Stewart and John Triggs, for theirinsightful comments. Love and gratitude to myfriends in light, past and present, and to mywonderful family, John, Emily, and ConnorSwitzer, for their patient support, graphics andeditorial help.

Major research funding was provided by theOntario Heritage Foundation with additionalfunds from the Presidential Committee onNorthern Science Training and Research,McMaster University. Support came also from theDoctoral Fellowship Program of the SocialSciences and Humanities Research Council, andthrough doctoral scholarships and bursaries fromMcMaster University. I would like to thank mycommittee members, the chairpersons, faculty,staff, and fellow students at the Department ofAnthropology, McMaster University, from 1980–1991, and the chair and faculty of the Departmentof Anthropology, University of Waterloo, from1991–1994, for sharing their time and guidancealong the way.

Thanks to my four great field crews, includingmy husband John Switzer, David Johnston,Deirdre Mackenzie-Ward, James Bandow, JamesMolnar, Alicia Hawkins, Cory Andrews, and fromBear Island, Ardie Moore, Jason Ball, and JodyBecker. Archaeology lab assistance was ablyprovided by Philip Woodley, Dorinda Sussex,David Johnston (draftsman of many originalfigures), Penny Young, and many other studentsfrom bothMcMaster and Laurentian Universities.Thanks go also to Rosemary Prevec for faunalbone identification, and Geof Burbidge for soilsand lithic raw material identification.

For the pollen core research, we would like tothank field crew members Brad, Stuart, Liz, andZicheng Yu, all botany students of McAndrews atthe University of Toronto. Our appreciation alsoto Darakshan Siddiqi for macrofossilidentifications and Kam-Biu Liu for pollen zone


correlation, done at the Royal Ontario Museum.This work has greatly benefited from many

researchers graciously sharing their time andinformation including Jean-Luc Pilon, James V.Wright, and Roger Marois (Archaeological Surveyof Canada), Helen Devereaux (LaurentianUniversity), Dean Knight (Wilfrid LaurierUniversity), Thor Conway, Jean Veillette(Geological Survey of Canada), Edward Harris(Bermuda Maritime Museum), Ann Balmer(Algonquin Associates), the staff at the TemagamiOntario Ministry of Natural Resources, JohnPollock (Woodland Heritage Services), LiamKieser (Isotrace Laboratory) and Karl Hele(Concordia University).

On Lake Temagami, the Plumsteads ofBoatline Bay Marina and the Moskwa family ofLoon Lodge all helped make the field seasonssuccessful. My special thanks to the late William(Bill) Twain, former Chief and Teme-AugamaAnishnabai Elder and to his son Don Twain whosefriendship (and dock) were invaluable. Manythanks also to former Teme-Augama AnishnabaiChief Gary Potts, former Temagami First NationChief Holly Charyna and Second Chief BarryMcKenzie for their support, and to the manypeople of Bear Island who visited the excavationsand shared their knowledge.Miigwech.

Andrews, J.T.1989

Arthurs, D.1978

Bailey, G.N.1983

Bandow, J.B.2012

Bartlein, P.J., and T. Webb III1985

Bell, P., and D. Wright1987

Bender, B.1978

Berkes, F.1998

Birkeland, P.W.1984

Bishop, C.A.1994a

Postglacial emergence and submergence. InQuaternary Geology of Canada andGreenland, edited by R.J. Fulton, pp. 546-562. Geology of Canada 1. GeologicalSurvey of Canada, Ottawa.

Sandy Lake Ware in Northwestern Ontario:A Distributional Study. ManitobaArchaeological Quarterly 2(1-2):57-64.

Editorial for Chapter 16: Social interactionand economic change: Large scaleperspectives. In Hunter-Gatherer Economy inPrehistory: A European Perspective, edited byG.N. Bailey, pp. 187-190. CambridgeUniversity Press, Cambridge.

White Dogs, Black Bears and GhostGamblers: Two Late Woodland MidewiwinAspects from Ontario. Papers of the FortiethAlgonquian Conference, 2008, edited by K.S.Hele and J.R. Valentine, pp. 40-58. StateUniversity of New York Press, Albany.

Mean July Temperature Estimates for 6,000yr B.P. in Eastern North America: RegressionEquations for Estimates from Fossil-PollenData. Syllogeus 55:301-342.

Rocks and Minerals: MacMillan Field Guides.Collier Books, New York.

Gatherer-Hunter to Farmer: A SocialPerspective.World Archaeology 10(2):204-222.

Indigenous knowledge and resourcemanagement systems. In Linking Social AndEconomic Systems: Management Practices andSocial Mechanisms for Building Resilience,edited by F. Berkes and C. Folke, pp. 98-128.Cambridge University Press, Cambridge.

Soils and Geomorphology. Oxford UniversityPress, New York.

Northern Algonquians 1550–1760. InAboriginal Ontario: Historical Perspectives onthe First Nations, edited by E.S. Rogers andD.B. Smith, pp. 275-288. Dundurn Press,Toronto.

References Cited


1994b

Black, D.W.1991

Bohaker, B.2006

Boissonneau, A.N.1968

Boudreau, R.E.A, J.M. Galloway, R.T. Patterson,A. Kumar, and F.A. Michel2005

Boyle, D.1900

1904

1905

Brightman, R.1993

Brizinski, M.A.1980

Brizinski, M.A., and H. Savage1983

Northern Algonquians 1760–1821. InAboriginal Ontario: Historical Perspectives onthe First Nations, edited by E.S. Rogers andD.B. Smith, pp. 289-306. Dundurn Press,Toronto.

Stratigraphic Integrity in Northeastern ShellMiddens: An Example from the InsularQuoddy Region. In Prehistoric Archaeology inthe Maritime Provinces, edited by M. Dealand S. Blair, pp. 201-215. Committee onArchaeological Cooperation, Reports inArchaeology 8. Council of MaritimePremiers, Fredericton.

“Nindoodemag”: The Significance ofAlgonquian Kinship Networks in the EasternGreat Lakes Region, 1600–1701. TheWilliam and Mary Quarterly, Third Series63(1):23-52.

Glacial History of Northeastern Ontario II:The Timiskaming-Algoma Area. CanadianJournal of Earth Sciences 5:97-109.

A paleolimnological Record of HoloceneClimate and Environmental Change in theTemagami Region, Northeastern Ontario.Journal of Paleolimnology 33:445-461.

Archaeological Report 1899. Being Part ofAppendix to the Report of the Minister ofEducation Ontario. L.K. Cameron, Toronto.Archaeological Report 1903. Being Part ofAppendix to the Report of the Minister ofEducation Ontario. L.K. Cameron, Toronto.Archaeological Report 1904. Being Part ofAppendix to the Report of the Minister ofEducation Ontario. L.K. Cameron, Toronto.

Grateful Prey: Rock-Cree Human-AnimalRelationships. University of California Press,Berkeley.

Where Eagles Fly: An Archaeological Survey ofLake Nipissing. Unpublished MA thesis,Department of Anthropology, McMasterUniversity, Hamilton, Ontario.

Dog Sacrifice among the Algonkian Indians:An Example from the Frank Bay Site.Ontario Archaeology 39:41-56.

Brück, J.1999

Burbidge, G.H.1988

Butzer, K.W.1982

CalPal2014

Canadian Soil Survey Committee1978

Card, K.D., W.H. McIlwaine, and H.D. Meyn1973

Carscallen, C.1994a

1994b

Champlain, S. de1922

Chism, J.V.1977

1978

Clark, C.P.1990

Ritual and Rationality: Some Problems ofInterpretation in European Archaeology.European Journal of Archaeology 2(3):313-344.

Lake Temagami Archaeology: GeologicalReport. Ms. on file, Ontario HeritageFoundation, Toronto.

Archaeology as Human Ecology: Method andTheory for a Contextual Approach. CambridgeUniversity Press, Cambridge.

Online CalPal: Cologne RadiocarbonCalibration and Paleoclimate ResearchPackage. <http://www.calpal-online.de>.

The Canadian System of Soil Classification.Research Branch Publication 1646. CanadaDepartment of Agriculture, Ottawa.

Geology of the Maple Mountain Area,Timiskaming, Nipissing and Sudbury Districts.Geological Report 106. Ontario Division ofMines, Ottawa.

The Lake Temagami Site (CgHa-2): Intra-siteLithic Resource Patterning and RegionalProcurement Strategies. Ms. on file, Ministryof Culture, Tourism and Recreation,Toronto.Excavations at the Lake Temagami Site,Temagami, Ontario. Annual ArchaeologicalReport Ontario 5 (New Series), edited by P.L. Storck, pp. 116-121. Ontario HeritageFoundation, Toronto.

The works of Samuel de Champlain, vol. 3.Translated and edited by H.H. Langton andW.F. Ganong. Champlain Society, Toronto.

Archaeology at Washadimi: The 1977 FieldSummary. Ms. on file Départementd’Archéologie et Ethnographie, Ministère desAffaires Culturelles, Québec.Archaeology at Washadimi: The 1978 ChismField Summary. Ms. on file Départementd’Archéologie et Ethnographie, Ministère desAffaires Culturelles, Québec.

A Dog Burial from Isle Royale Lake Superior:An Example of Household Ritual Sacrifice in


Cleland, C.E.1982

Conway, T.1982

1984

1986

Conway, T., and J. Conway1989

Côté, M., and L. Inksetter2001

Court Judgments1984

1991

1999

Courty, M. A., P. Goldberg, and R. MacPhail1989

the Terminal Woodland Period.Midcontinental Journal of Archaeology15(2):265-278.

The Inshore Fishery of the Northern GreatLakes: Its Development and Importance inPrehistory. American Antiquity 47(4):761-784.

A Preliminary Report on the Witch Point Site,Lake Temagami. Ms. on file, HeritageBranch, Northeastern Region, Ministry ofCitizenship and Culture, Toronto.An Archaeological Assessment of Obabika Lake.Ms. on file, Heritage Branch, NortheasternRegion, Ministry of Citizenship and Culture,Toronto.The Sand Point Site, Lake Temagami. Ms. onfile, Heritage Branch, Northeastern Region,Ministry of Citizenship and Culture,Toronto.

An Ethno-Archaeological Study of AlgonkianRock Art in Northeastern Ontario, Canada.Ontario Archaeology 49:34-59.

Ceramics and Chronology of the LatePrehistoric Period: The Abitibi-Témiscaminque Case. A Collection of PapersPresented at the 33rd annul meeting of theCanadian Archaeological Association, editedby J-L. Pilon, M.W. Kirby, and C. Thériaultpp. 111-127. The Ontario ArchaeologySociety, Toronto.

Ontario Supreme Court, Attorney Generalfor Ontario v. Bear Island Foundation et al.,CanLII 2136(ON SC).<http://temagami.nativeweb.org/archive/ongov-vs-bearisland.pdf>.Supreme Court of Canada, Ontario(Attorney General) v. Bear IslandFoundation, [1991] 2 SCR 570 <http://scc-csc.lexum.com/scc-csc/scc-csc/en/item/782/index.do>.Court of Appeal for Ontario, Bear IslandFoundation, et al. v. Ontario CanLII 9307(ON CA) < http://canlii.ca/t/1f9tw>.

Soils and Micromorphology in Archaeology.Cambridge University Press, Cambridge.

Dalla Bona, L.2000

Dawson, K.C.A.1983

Dewdney, S., and K. Kidd1967

Dunning, R.W.1959

Dyke, A.S., and V.K. Prest1987

Faegri, K., and J. Iversen1975

Feit, H.A.1973

1993

2004

Fenwick, I.M.1968

Fox, W.A.1975

Protecting Cultural Resources through ForestManagement Planning Using ArchaeologicalPredictive Modeling. In Practical Applicationsof GIS for Archaeologists: A PredictiveModelling Toolkit, edited by K.L. Wescott andR.J. Brandon, pp. 78-106. Taylor andFrancis, Philadelphia.

Cummins Site: A Late Paleo-Indian (Plano)site at Thunder Bay, Ontario. OntarioArchaeology 39:3-31.

Indian Rock Paintings of the Great Lakes.University of Toronto Press, Toronto.

Social and Economic Change among theNorthern Ojibwa. University of TorontoPress, Toronto.

Paleogeography of Northern North America,18,000–5,000 years ago. Geological Surveyof Canada, Map 1703A.

Textbook of Pollen Analysis. Munksgaard,Copenhagen.

The Ethno-Ecology of the Waswanipi Creeor How Hunters Can Manage TheirResources. In Cultural Ecology: Readings onthe Canadian Indians and Eskimos, edited byB. Cox. pp. 115-125. McClelland andStewart, Toronto.The Construction of Algonquian HuntingTerritories: Private Property as Moral Lesson,Policy Advocacy, and Ethnographic Error. InColonial Situations: Essays on theContextualization of Ethnographic Knowledge,edited by G.W. Stocking, Jr., pp. 109-134.University of Wisconsin Press, Madison.Hunting and the Quest for Power: The JamesBay Cree and Whitemen in the 20thCentury. In Native Peoples: The CanadianExperience, edited by R.B. Morrison andC.R. Wilson, pp. 101-128. OxfordUniversity Press, Toronto.

Pedology as a Tool in ArchaeologicalInvestigations.Ontario Archaeology 11:27-38.

The Paleo-Indian Lakehead Complex. InPapers contributed to the CanadianArchaeological Association Annual Meeting,


1990

Fox, W.A., and C. Garrad2004

Gamble, C.1983

Gilman, A.1983

Gordon, D.L.1980

1985

1986

1987

1988a

1988b

1989

1990a

March 1975, pp. 28-29. Historical Planningand Research, Ontario Ministry of Cultureand Recreation, Toronto.The Odawa. In The Archaeology of SouthernOntario to A.D. 1650, edited by C.J. Ellisand N. Ferris, pp. 457-473. OccasionalPublication 5. London Chapter, OntarioArchaeological Society, London.

Hurons in Algonquian Land. OntarioArchaeology 77/78:121-134.

Culture and Society in the Upper Palaeolithicof Europe. In Hunter-Gatherer Economy inPrehistory: A European Perspective, edited byG.N. Bailey, pp. 201-211. CambridgeUniversity Press, Cambridge.

Explaining the Upper Paleolithic Revolution.In Hunter-Gatherer Economy in Prehistory: AEuropean Perspective, edited by G.N. Bailey,pp. 115-126. Cambridge University Press,Cambridge.

Reflections on Refuse: A ContemporaryExample from James Bay, Québec. CanadianJournal of Archaeology 4:83-97.North Caribou Lake Archaeology.ConservationArchaeology Report 9. Northwestern Region,Ontario Ministry of Citizenship and Culture,Toronto.LakeTemagami Archaeology, 1986: PreliminaryField Report. Ms. on file, Ontario HeritageFoundation, Toronto.Archaeology at Lake Temagami, 1986: FinalField Report. Ms. on file, Ontario HeritageFoundation, Toronto.Discovering Patterns of Change at NorthCaribou Lake in the Post-Contact Period:Implications for Finding and Interpreting theArchaeological Record. In Boreal Forest andSub-Arctic Archaeology, edited by C.S. Reid,pp. 3-36. Occasional Publication 6. LondonChapter, Ontario Archaeological Society,London.Lake Temagami and the NorthernExperience. Arch Notes 88(2):25-30.Lake Temagami Archaeological Research:1987–1988, Final Field Report. Ms. on file,Ontario Heritage Foundation, Toronto.Prehistoric Occupations at Lake Temagami.In Temagami: A Debate onWilderness, editedby M. Bray and A. Thomson, pp. 153-183.Dundurn Press, Toronto.

1990b

1990c

1991a

1991b

1992

1993

1994a

1994b

1995a

1995b

Fieldwork at Lake Temagami. AnnualArchaeological Report Ontario 1 (New Series),edited by P. L. Storck, pp. 39-40. OntarioHeritage Foundation, Toronto.Analysis of the Three Pines Site, LakeTemagami. Annual Archaeological ReportOntario 1 (New Series), edited by P. L.Storck, pp. 80-82. Ontario HeritageFoundation, Toronto.A Contextual Approach to the Archaeology ofLake Temagami. Ms. on file, Department ofAnthropology, McMaster University,Hamilton.In Context: Approaches to theArchaeological Record of the Three Pines Site(CgHa-6), Lake Temagami, Ontario. Paperpresented at the 24th Annual Meeting of theCanadian Archaeological Association, St.John’s, Newfoundland. Ms. on file, OntarioHeritage Foundation, Toronto.Research Report on Palaeo-hydrology andPrehistoric Occupations of Lake Temagami,Northeastern Ontario. Ms. on file, OntarioHeritage Foundation, Toronto.Preliminary Report on 1993 ArchaeologicalField Work at the Witch Point Site (CgHa-7),LakeTemagami, Ontario. Ms. on file, OntarioHeritage Foundation, Ministry of Culture,Tourism and Recreation, Toronto.Rocks, Water and a Dog: Structural Variationbetween theWitch Point Site (CgHa-7) and theThree Pines Site (CgHa-6), Lake Temagami.Paper presented at the 27th Annual Meetingof the Canadian Archaeological Association,Edmonton, Alberta.Excavations at the Witch Point Site (CgHa-7), Lake Temagami, Northeastern Ontario.Annual Archaeological Report Ontario 5 (NewSeries), edited by P.L. Storck, pp. 129-132.Ontario Heritage Foundation, Toronto.Report on the 1994 Field Season: FurtherExcavation at the Witch Point Site CgHa-7and Survey of Prehistoric and TraditionalNative Sites on Lake Temagami, NortheasternOntario: Final Field Report. Ms. on file,Ontario Heritage Foundation, Ministry ofCulture, Tourism and Recreation, Toronto.Continued Excavations at the Witch PointSite (CgHa-7) and Survey of Prehistoric andNative Sites on Lake Temagami,Northeastern Ontario. Annual ArchaeologicalReport Ontario 6 (New Series), edited by P.L.Storck, pp. 103-107. Ontario HeritageFoundation, Toronto.


Gordon, D.L., and J.H. McAndrews1992

Gordon, D.L., J.H. McAndrews, and I.D. Campbell1992

Grant, J.A.1964

Greenman, E.1943

Hall, A.1990

Hall, R.I., K.E. Duff, and P.A. Quinby1994

Hamilton, S.1981

2000

Hanks, C.C.1983

1988

Field Testing a Model of Palaeo-hydrologyfor Prehistoric Site Prediction at LakeTemagami, Northeastern Ontario. AnnualArchaeological Report Ontario 3 (New Series),edited by P.L. Storck, pp. 80-86. OntarioHeritage Foundation, Toronto.

Palaeo-hydrology and Prehistoric Occupationsat Lake Temagami: Preliminary Research.Paper presented at the 25th Annual Meetingof the Canadian Archaeological Association,London, Ontario. Ms. on file, OntarioHeritage Foundation and Ministry ofCulture, Tourism and Recreation, Toronto.

Geology of the Vogt-Hobbs Area, District ofNipissing. Geological Report 22. OntarioDepartment of Mines, Ottawa.

An Early Industry on a Raised Beach nearKillarney, Ontario. American Antiquity8:260-265.

Where Justice Lies: Aboriginal Rights andWrongs in Temagami. In Temagami: ADebate onWilderness, edited by M. Bray andA. Thomson, pp. 223-253. Dundurn Press,Toronto.

A 10,000 Year Vegetation History of theTemagami Region of Ontario with SpecialEmphasis on Eastern White Pine. AncientForest Exploration and Research Report 4.<http://www.ancientforest.org/rr4.html>

The Archaeology of the Wenesaga Rapids.Archaeological Research Report 17.Archaeology and Heritage Planning Branch,Ministry of Culture and Communication.Archaeological Predictive Modelling in theBoreal Forest: No Easy Answers. CanadianJournal of Archaeology 24(1-2):41-76.

An Ethnoarchaeological Approach to theSeasonality of Historic Cree Sites in CentralQuébec. Arctic 36(4):350-355.The Foxie Otter Site: A Multi-componentOccupation North of Lake Huron.Anthropological Papers 79. Museum ofAnthropology, University of Michigan, AnnArbor.

Harris, E.C.1975

1977

1979a

1979b

1989

Harris, E.C.,M.R. Brown III andG.J. Brown (editors)1993

Hart, J.P. and W.A. Lovis2013

Heidenreich, C.1971

Hewitt, D.F., and E.B. Freeman1978

Hickerson, H.1960

Hinshelwood, A.1997

Hodgins, B.W.1976

1987

Hodgins, B.W., and J. Benidickson1989

The Stratigraphic Sequence: A question oftime.World Archaeology 7:109-121.Units of Archaeological Stratification.Norwegian Archaeological Review 10:84-94.Principles of Archaeological Stratigraphy.Academic Press, New York.The Laws of Archaeological Stratigraphy.World Archaeology 11:111-117.Principles of Archaeological Stratigraphy.2nd ed. Academic Press, New York.

Practices of Archaeological Stratigraphy, editedby E.C. Harris, M.R. Brown III, and G.J.Brown. Academic Press, London.

Reevaluating What We Know About theHistories of Maize in Northeastern NorthAmerica: A Review of Current Evidence.Journal of Archaeological Research 21:175-216.

Huronia: A History and Geography of theHuron Indians 1600–1650.McClelland andStewart, Toronto.

Rocks and Minerals of Ontario. GeologicalCircular 13. Ontario Ministry of NaturalResources, Toronto.

The Feast of the Dead among theSeventeenth Century Algonkians of theUpper Great Lakes. American Anthropologist62(1):81-107.

Boreal Forest Fire Ecology andArchaeological Site Formation: An Examplefrom Northern Ontario.Ontario Archaeology62:63-92.

Paradis of Temagami. Highway Book Shop,Cobalt, Ontario.The Lure of the Temagami-based CanoeTrip. In Nastawgan: The Canadian North byCanoe and Snowshoe, edited by B.W. Hodginsand M. Hobbs, pp. 189-202. DundurnPress, Toronto.

The Temagami Experience: Recreation,Resources and Aboriginal Rights in theNorthern Ontario Wilderness. University ofToronto Press, Toronto.


Irimoto, T.1980

Jamieson, S.M.1992

Jenish, D.Y.2006

Jenkins, W.H.1939

Julig, P.1989

1995

2002

Karrow, P.F., and R.S. Geddes1987

Knight, D.H.1977

Koezur, P., and J. V. Wright1976

Kritsch-Armstrong, I.D.1982

Chipewyan Ecology: Group Structure andCaribou Hunting System. Senri EthnologicalStudies 8. National Museum of Ethnology,Osaka.

Regional Interaction and Ontario IroquoisEvolution. Canadian Journal of Archaeology16:70-88.

Mapman of Temagami. Legion Magazine,May 1. Canvet Publications.<http://www.legionmagazine.com/en/index.php/2006/05/mapman-of-temagami/>.

Notes on the Hunting Economy of the AbitibiIndians. Anthropological Series 9. TheCatholic University of America, Washington.

An evaluation of the Contextual Approach inPredicting Site Locations in the Boreal Forest.Paper presented at the 16th Annual Meetingof the Ontario Archaeological Society.The Cummins Complex and PaleoindianOccupations in the Northwestern LakeSuperior Region. Ontario ArchaeologicalReport 2. Ontario Heritage Foundation,Toronto.The Sheguiandah Site: Archaeological,Geological and Paleobotanical Studies at aPaleoindian Site on Manitoulin Island,Ontario. Mercury Series Paper 161.Archaeological Survey of Canada, CanadianMuseum of Civilization, Hull, Québec.

Drift Carbonate on the Canadian Shield.Canadian Journal of Earth Sciences 24:365-369.

The Montreal River and the Shield Archaic.Unpublished PhD dissertation, Departmentof Anthropology, University of Toronto,Toronto.

The Potato Island Site, District of Kenora,Ontario. Mercury Series Paper 51.Archaeological Survey of Canada, NationalMuseum of Man, Ottawa.

Queries near the Quarry: A TechnologicalAnalysis of the Jessup Lithic Workshop Site.Unpublished MA thesis, Department ofAnthropology, McMaster University,Hamilton, Ontario.

Lee, T.E.1953

1957

Lenius, B.J., and D.M. Olinyk1990

Lévi-Strauss, C.1969

Lewis, C.F.M., and T.W. Anderson1989

Liu, K-B.1990

Liu, K-B., and N. Lam1985

Loewen, B., C. Chapdelaine, and P. Richard2005

Long, J.1974

Lovell, H.L.1977

Lovell, H.L., and T.W. Caine1970

A Preliminary Report on the SheguiandahSite, Manitoulin Island. In Annual Report ofthe National Museum of Canada for the FiscalYear 1950–51.National Museum of CanadaBulletin 128:58-67.The Antiquity of the Sheguiandah Site: TheCanadian Field-Naturalist 71(3):117-137.

The Rainy River Composite: Revisions toLate Woodland Taxonomy. In TheWoodlandTradition in the Western Great Lakes: PapersPresented to Elden Johnson, edited by G.Gibbon, pp. 77-112. Publications inAnthropology 4. University of Minnesota,Minneapolis.

The Elementary Structures of Kinship. BeaconPress, Boston.

Oscillations of Levels and Cool Phases of theLaurentian Great Lakes Caused by Inflowsfrom Glacial Lakes Agassiz and Barlow-Ojibway. Journal of Paleoliminology 2:99-146.

Holocene Paleoecology of the Boreal Forestand Great Lakes–St. Lawrence Forest inNorthern Ontario. Ecological Monographs60(2):179-212.

Paleovegetational Reconstruction Based onModern and Fossil Pollen Data: AnApplication of Discriminant Analysis. Annalsof the Association of American Geographers75(1):115-130.

Holocene Shoreline Occupations and WaterLevel Change at Lac Mégantic, Québec.Canadian Journal of Archaeology 29(2):267-280.

Voyages and Travels of an Indian Interpreterand Trader. Facsimile ed. Originallypublished 1791, Coles Publishing, Toronto.

Geology of the Englehart-Earlton Area, Districtof Timiskaming. Miscellaneous Paper 69.Ontario Geological Survey, Toronto.

Lake Timiskaming Rift Valley. MiscellaneousPaper 39. Ontario Department of Mines,Toronto.


Macdonald, C.K.1987

1993

Madden, M.1983

Marois, R., and P. Gauthier1989

McNab, D.T.1999

2009

McNeil, K.1990

Mitchell, E.A.1977

Noble, W.C.1982

Noël Hume, I.1980

Oberholtzer, C.2002

Ontario Geological Survey1973

The Nastawgan: Traditional Routes of Travelin the Temagami District. In Nastawgan: TheCanadian North by Canoe and Snowshoe,edited by B.W. Hodgins and M. Hobbs, pp.183-188. Dundurn Press, Toronto.Historical Map of Temagami. OntarioGeographic Names Board.

Social Network Systems amongst Hunter-Gatherers Considered in Southern Norway.In Hunter-Gatherer Economy in Prehistory: AEuropean Perspective, edited by G.N. Bailey,pp. 191-200. Cambridge University Press,Cambridge.

Les Abitibis. Archaeological Survey ofCanada, Mercury Series Paper 140. CanadianMuseum of Civilization, Hull, Québec.

Circles of Time: Aboriginal Land Rights andResistance in Ontario. Wilfrid LaurierUniversity Press, Waterloo, Ontario.No Place for Fairness: Indigenous Land Rightsand Policy in the Bear Island Case and Beyond.McGill-Queen’s University Press, Montreal.

The Temagami Indian Land Claim:Loosening the Judicial Strait-Jacket. InTemagami: A Debate onWilderness, edited byM. Bray and A. Thomson, pp. 185-222.Dundurn Press, Toronto.

Fort Timiskaming and the Fur Trade.University of Toronto Press, Toronto.

Algonquian Archaeology in NortheasternOntario. In Approaches to AlgonquianArchaeology, edited by M. Hanna and B.Kooyman, pp. 35-55. Proceedings of the13th Annual Chacmool Conference.Department of Archaeology, University ofCalgary, Calgary.

A Guide to Artifacts of Colonial America.Alfred. A. Knopf, New York.

Fleshing Out the Evidence: From ArchaicDog Burials to Historic Dog Feasts. OntarioArchaeology 73:3-14.

Map 2259, Diamond Lake. Scale 1:63,360.Ontario Geological Survey.

Ontario Ministry of Natural Resources1969

1973

1978

1982

Pelleck, J.A.1983

Phillips, W.H.C.1907

Pollock, J.W.1975

1976

1984

1992

Pollock, J.W., and R. Koistinen1993

Fisheries Management Maps: LakeTemagami, Temagami Region. Scale1:64,000. Ontario Ministry of NaturalResources, Toronto.Lake Temagami Plan for Land Use andRecreation Development. Ontario Ministryof Natural Resources, Toronto.Background Information and Approach toPolicy: Northeastern Ontario: Strategic LandUse and Planning Process. Ontario Ministryof Natural Resources, Toronto.Ontario Base Maps. Scale 1:20,000 OntarioMinistry of Natural Resources, Toronto.

Forestry Point (EgKl-1): A Woodland Site onRed Lake, Northwestern Ontario.Conservation Archaeology Report,Northwestern Region Report 3. OntarioMinistry of Citizenship and Culture, Kenora,Ontario.

Rock Paintings at Temagami District.Archaeological Report 1906, pp 41-47. BeingPart of Appendix to the Report of theMinister of Education Ontario. L.K.Cameron, Toronto.

Algonquian Culture Development andArchaeological Sequences in NortheasternOntario. Canadian Archaeological AssociationBulletin 7:1-53.The Culture History of Kirkland Lake District,Northeastern Ontario. Mercury Series Paper54. Archaeological Survey of Canada,National Museum of Man, Ottawa.A Technological Analysis of Lake AbitibiBifaces. Unpublished PhD Dissertation,Department of Anthropology, University ofAlberta, Edmonton.Consulting Activities of Settlement SurveysLimited in Northern Ontario. AnnualArchaeological Report Ontario 3 (New Series),edited by P.L. Storck, pp. 46-54. OntarioHeritage Foundation, Toronto.

Teme-Augama Anishnabai Cultural HeritageInventory Planning and Assessment on n’Daki-Menan. Paper presented at the 26th AnnualMeeting of the Canadian ArchaeologicalAssociation, Montreal, Québec.


Potts, G.1989

Prevec, R.1987

Racher, P.1988

Rajnovich, G.1994

Rajnovich, M.G.N., and C.S. Reid1978

Ray, A.J.1974

Reid, C.S., and G. Rajnovich1991

Ridley, F.1950-1953

1954

1956

1957

1958

1963

1966

Teme-Augama Anishnabai: Last-DitchDefence of a Priceless Homeland. In DrumBeat: Anger and Renewal in Indian Country,edited by B. Richardson, pp. 201-228.University of Toronto Press, Toronto.

The Three Pines Site CgHa-6: FaunalReport. Ms. on file, Ontario HeritageFoundation, Toronto.

Analysis of Gillies Depot I and Marsh Bay I.Ms. on file, Department of Sociology andAnthropology, Wilfrid Laurier University,Waterloo.

Reading Rock Art: Interpreting the Indian RockPaintings of the Canadian Shield. NaturalHeritage Books, Dundurn Press, Toronto.

Selkirk and Clearwater Lake Ceramics onLake of the Woods: An Overview. Archae-facts: Journal of the Archaeological Society ofSouth-Western Manitoba 5(2,3):43-48.

Indians in the Fur Trade: Their Roles asTrappers, Hunters and Middlemen in theLands Southwest of Hudson Bay, 1660–1870.University of Toronto Press, Toronto.

Laurel: a Re-evaluation of the Spatial, Socialand Temporal Paradigms. Canadian Journalof Archaeology 15:193-234.

Field Notes, Frank Bay Site (CbGw-1).Ms. on file, Archaeological Survey of Canada,Canadian Museum of Civilization, Hull.Québec.The Frank Bay Site, Lake Nipissing, Ontario.American Antiquity 20:40-50.An Archaeological Reconnaissance of LakeAbitibi, Ontario. Ontario History 43(1):18-23.A Preliminary Comment on ArcticRegionalism. Pennsylvania Archaeologist 27(3-4):145-148.Sites on Ghost River, Lake Abitibi.Pennsylvania Archaeologist 28(1):3-20.The Red Pine Point Site. AnthropologicalJournal of Canada 2(3):7-10.Archaeology of Lake Abitibi Ontario-Québec. Anthropological Journal of Canada4(2):2-50.

Roed, M.A., and D.R. Hallett1979

Rogers, E.S.1962

1963a

1963b

1966

1969

1973

Rogers, E.S., and M. Black1976

Saarnisto, M.1974

Sahlins, M.1972

1987

Simony, P.S.1964

Skinner, A.1911

Northern Ontario Engineering GeologyTerrain Study, Data Base Map, MapleMountain. Ontario Geological Survey Map5023.

The Round Lake Ojibwa. Art andArchaeology Occasional Papers 5. RoyalOntario Museum, Toronto.Changing Settlement Patterns of the Cree-Ojibwa of Northern Ontario. SouthwesternJournal of Anthropology 19(1):64-88.The Hunting Group-Hunting TerritoryComplex among the Mistassini Indians.Bulletin 195. National Museum of Canada,Ottawa.Subsistence Areas of the Cree-Ojibwa of theEastern Subarctic: A Preliminary Study. InContributions to Anthropology 1963-64, PartII, pp. 87-118. National Museum of CanadaBulletin 204.Band Organization among the Indians ofEastern Subarctic Canada. In Contributionsto Anthropology: Band Societies, edited by D.Damas, pp. 21-50. Bulletin 218. NationalMuseum of Canada, Ottawa.The Quest for Food and Furs: The MistassiniCree, 1953-1954. Publications in Ethnology5. National Museum of Man, Ottawa.

Subsistence Strategy in the Fish and HarePeriod, Northern Ontario: The WeagamowOjibwa 1880–1920. Journal ofAnthropological Research 32(1):1-43.

The Deglaciation History of the LakeSuperior Region and Its ClimaticImplications.Quaternary Research 4:316-339.

Stone Age Economics. Aldine-Atherton,Chicago.Schiffer, M.B.Formation Processes of the ArchaeologicalRecord. University of New Mexico Press,Albuquerque.

Geology of Northwestern Timagami Region,District of Nipissing. Geological Report 28.Ontario Department of Mines, Toronto.

Notes on the Eastern Cree and NorthernSaulteaux. Anthropological Papers of theAmerican Museum of Natural History 9,Part 1.


Smith, B.A.1983

Smith, B.A., and R. Prevec2000

Smith, R.H.1986

Speck, F.G.1915a

1915b

Storck, P.L.1984

Stucki, B.R.1993

Tanner, A.1978a

1978b

1979

Faunal Analysis of the Witch Point Site(1982 excavation). Ms. on file, HeritageBranch, Northeastern Region, OntarioMinistry of Citizenship and Culture,Toronto.

Economic Strategies and CommunityPatterning at the Providence Bay Site,Manitoulin Island. Ontario Archaeology69:76-91.

Analysis of the Clay Tobacco PipeAssemblage from the Front Street Site (AjGu-15), Toronto. Ontario Archaeology 46:55-61.

Family Hunting Territories and Social Life ofVarious Algonkian Bands of the OttawaValley.Geological Survey Memoir 70.Anthropological Series 8. CanadaDepartment of Mines, Ottawa.Myths and Folk-Lore of the TimiskamingAlgonquin andTimagami Ojibwa.GeologicalSurvey Memoir 71. Anthropological Series 9.Canada Department of Mines, Ottawa.

Research into the Paleo-Indian Occupationsof Ontario: A Review. Ontario Archaeology41:3-28.

Three-Dimensional Assessment of ActivityAreas in a Shell Midden: an Example for theHoko River RockShelter, State ofWashington. In Practices of ArchaeologicalStratigraphy, edited by E.C. Harris, M.R.Brown III, and G.J. Brown, pp. 122-138.Academic Press, London.

Ethno-Archaeology in the Region of theJames Bay Project. Ms. on file, Départementd’Archéologie et Ethnographie, Ministère desAffaires Culturelles, Québec.Field Notes: Washadimi, Baie James,Summer 1978. Ms. on file, Départementd’Archéologie et Ethnographie. Ministère desAffaires Culturelles, Québec.Bringing Home Animals: Religious Ideologyand Mode of Production of the Mistassini CreeHunters. Institute of Social and EconomicResearch, Memorial University ofNewfoundland, St. John’s, Newfoundland.

Taylor-Hollings, J.S.1999

Teme-Augama Anishnabai/Temagami First Nation2007

Thwaites, R.G. (editor)1896-1901

Tyyska, A., and J. Burns1973

Veillette, J.1986

1988

1989

Walthall, J.A., and B. Koldehoff1998

White, R.1991

Wright, H.E., Jr.1967

Wright, J.V.1967a

The Northwestern Extent of Sandy LakeWare:A Canadian Perspective. Unpublished MAthesis, Department of Anthropology andArchaeology, University of Saskatchewan,Saskatoon.

Family Names and Totems.<http://www.ndakimenan.com/CLANS.htm>.

The Jesuit Relations and Allied Documents.73 vols. Burrows Brothers, Cleveland.<http://puffin.creighton.edu/jesuit/relations>.

Archaeology from North Bay to Mattawa.Heritage Branch Report 19. OntarioMinistry of Culture and Recreation, Toronto.

Map1624 A Surficial geology, Haileybury,Ontario-Québec. Scale 1:100,000.Geological Survey of Canada.Déglaciation et Evolution des LacsProglaciaires Post-Algonquin et Barlow auTémiscamingue, Québec et Ontario.Géographie Physique et Quaternaire 42(1):7-31.Ice Movements, Till Sheets and GlacialTransport in Abitibi-Timiskaming, Québecand Ontario. In Drift Prospecting, edited byR.N. Dilabio and W.B. Coker, pp. 139-154.Geological Survey of Canada Paper 89-20.

Hunter-Gatherer Interaction and AllianceFormation: Dalton and the Cult of the LongBlade. Plains Anthropologist 43:257-273.

The Middle Ground: Indians, Empires andRepublics in the Great Lakes Region, 1650-1815. Cambridge University Press,Cambridge.

A Square-Rod Piston Sampler for LakeSediments: Notes. Journal of SedimentaryPetrology 37(3):975-976.

Pic River Site: A Stratified Late WoodlandSite on the North Shore of Lake Superior. pp.54-99. Anthropological Series 72. NationalMuseum of Canada Bulletin 206.


1967b

1968

The Laurel Tradition and The MiddleWoodland Period. Anthropological Series79,National Museum of Canada Bulletin 217.The Michipicoten Site, Ontario.Contributions to Anthropology VI:Archaeology and Physical Anthropology, pp.1-85. Anthropological Series 82, NationalMuseum of Canada Bulletin 224.

1972a

1972b

1979

The Shield Archaic. Publications inArchaeology 3. National Museum of Man,Ottawa.Ontario Prehistory. Van Nostrand andReinhold, Toronto.The Shield Archaic: A Critique of a Critique.Manitoba Archaeological Quarterly 3(3-4):30-35.

Les enquêtes de sites préhistoriques et historiques, l’excavation des sites à composants multiples de Three Pines (CgHa-6) etdeWitch Point (CgHa-7) et le carottage de pollen de trois tourbières fournissent des renseignements étendus quant à l’histoireculturelle et naturelle en évolution du lac Temagami. Une meilleure connaissance de l’évolution des paléorivages modifiel’interprétation des sites connus et la méthodologie d’enquête. Des enquêtes des rivages modernes, premièrement ceux le longdes rivages et deuxièmement ceux internes, ont engendré des sites de campement préhistoriques, des ateliers lithiques, descarrières de veine de quartz et des sites d’utilisation traditionnelle de la Première nation Teme-Augama Anishnabai datantdes XIXe et XXe siècles. Compte tenu des contraintes topographiques du lac Temagami, les chasseurs-cueilleurs favorisaientces endroits limités ayant un terrain plat et bien drainé, une protection contre les vents froids et une facilité d’accès aurivage.

Le site Three Pines possède un sol mince et comprimé qui est typique des sites préhistoriques de la région du Boucliercanadien, mais une analyse modifiée du diagramme stratigraphique (de la version originale Harris matrix) permet unaperçu de la séquence stratigraphique des occupations de la période Archaïque à celle moderne. Des changements significatifsau niveau du paysage sur le site semblent avoir affecté la configuration spatiale de l’utilisation du site et potentiellementavoir conduit à la perte de composants primitifs. Contrairement au site Three Pines, les résultats préliminaires du siteWitch Point (CgHa-7) montrent des gisements plus profonds, une plus grande densité d’artéfacts, de nombreux vestiges degalet et des activités de façonnage. Ces différences structurelles sont liées aux variations relatifs en lien avec le paysage dusite, l’altitude, la saisonnalité et l’utilisation culturelle. Un comportement rituel du Sylvicole supérieur, suggéré par unenterrement de chien et par de l’ocre rouge au siteWitch Point, est examiné dans un contexte de dires primitifs postcontactsoù les rituels renforçaient les alliances de groupe.

Diana L. GordonDigital Albatross Design and Consulting Inc.58 Canada St.Hamilton, OntarioL8P [email protected]

John H. McAndrewsSenior Curator EmeritusNatural HistoryRoyal Ontario MuseumToronto, [email protected]

Ian D. CampbellDirectorScience Coordination DivisionAgriculture and Agri-FoodCanadaOttawa, [email protected]