Article body

Introduction

In 1980, Charles Martijn and Norman Clermont (1980a) published a volume of Études/Inuit/Studies summarizing the state of knowledge about Inuit in southern Labrador. Historical documents suggested an Inuit presence on the southern Labrador coast by the late 16th to early 17th century (Barkham 1980; Martijn 1980a, 1980b; Sturtevant 1980; Trudel 1980), but the single archaeological site (Martijn and Clermont 1980b) and inconclusive linguistic data (Mailhot et al. 1980) were less convincing. None of the papers were able to demonstrate that the Inuit settled the south in a permanent fashion. Hamilton Inlet, in central Labrador, was offered as the southern terminus of occupation, from which Inuit made seasonal forays south to trade with European fishers and whalers and to pilfer from their camps (Jordan and Kaplan 1980: 39). The problem with this conclusion was that vast areas of southern Labrador and the Quebec North Shore had yet to be archeologically surveyed.

In the years since, Auger (1991, 1993, 1994), Brewster (2006), Dumas and Poirier (1994), Fitzhugh (1989), Rankin (2004, 2006, 2014a), and Stopp (2002) have located Inuit settlements spanning the southern coastline. Yet the number of Inuit winter sod houses, the dwelling form best associated with permanent Inuit settlement in central and northern Labrador, remains limited. This is not due to a deficit of sod-walled dwellings. Stopp (2002: 85) reported 210 sod features between Hamilton Inlet and the Strait of Belle Isle. The real problem is establishing which of these sod features may have been occupied by Inuit. In southern Labrador, sod-covered dwellings occupied by Inuit-Métis families during the 18th and 19th centuries, and by seasonal 19th-century European, American, and Newfoundland “stationer” fishers, leave archaeological traces similar to those of Inuit sod houses, and the occupants’ ethnicity is not easily differentiated by archaeological survey or test excavation.

Because ethnicity, or culturally ascribed identity, has material manifestations that allow group members to recognize shared cultural characteristics, it should be recognizable in the archaeological record. Thus, Stopp (2002: 85) has proposed using cultural features such as mounded stone burials, tent rings, stone fox traps, visible house entrance passages, and material culture made from bone, soapstone, and ground slate to recognize Inuit occupation based on their similarity to features at known Inuit sites elsewhere in Labrador. Unfortunately, these features are not always obvious or present at every Inuit site, particularly when survey is the primary mode of investigation. This paper draws on community interviews and archaeological surveys and excavations undertaken at known or probable Inuit, Inuit-Métis, and stationer fishery sites in and around Sandwich Bay, Labrador, between 2001 and 2013, to expand Stopp’s (2002) site criteria and help resolve the issue of site ethnicity for the area immediately south of Hamilton Inlet. Site location, as well as house and site features, are used to increase confidence in Inuit site classification and to provide strategies for targeted test-excavations elsewhere in southern Labrador. The results of the research also allow for a better understanding of the nature and extent of Inuit occupation in Sandwich Bay.

The Sandwich Bay region

The region of investigation includes Sandwich Bay (the second largest bay on the Labrador coast), the mainland flanking the bay to the north and south, and the adjacent islands (Figure 1). Nearly 90 km separates Sandy Cove, in the north, from the Cape North Peninsula, in the south. Approximately 60 km separates the head of Sandwich Bay from the outer islands. The region has several diverse ecological landscapes, including the sand dune beaches of the Porcupine Strand, the forested shores of the inner bay, and the rocky, exposed tundra of the islands and headlands. The entire zone is rich in marine resources throughout the year, and there are seasonally abundant avian, mammalian, and plant resources to be found in specific locales (see Rankin et al. 2012: 65-66). Inuit traveling to this region would have encountered an ecosystem similar to that found in other parts of Labrador, particularly Hamilton Inlet, where well-documented Inuit sites are located near the present-day community of Rigolet (Jordan 1977; Jordan and Kaplan 1980). Sandwich Bay would have been the next region colonized if Inuit settlement extended further south.

Relevant archaeological research in Sandwich Bay

In 1986, Fitzhugh conducted a two-day survey in Sandwich Bay, identifying three probable Inuit sites, including a series of cache pits and a fox trap at Baird Cove, as well as a tent camp and sod structures at Snack Cove (1989: 168-171). A second trip in 1987 verified the Inuit affiliation of both Snack Cove sites through test excavation and led to the discovery of a second Inuit tent site at nearby Snack Cove Island (Fitzhugh 2009: 136). During a 1992 survey of Sandwich Bay, Stopp (1997) recorded a substantial number of historically-occupied sites, and unclassified rock features. In 2002, she revisited these sites, looking for indicators of Inuit activity, such as fox traps, and added a further six Inuit sites to the regional database (Stopp 2002: 86-87), while suggesting that as many as 12 more sites might also be Inuit based on the presence of sod houses or low-elevation rock caches. My survey of the region undertaken between 2001 and 2006 used Stopp’s (2002) criteria to identify a further seven probable Inuit sites and archival evidence to locate the first definitive Inuit-Métis settlement in Sandwich Bay (Rankin 2002, 2003, 2004, 2007, 2010). In 2007, Pace (2008) conducted interviews and an archaeological survey with local informants and identified two more Inuit sites, three probable Inuit-Métis settlements, and one site that locals associated with the Newfoundland fishery. Kelvin (2011) conducted further interviews and a survey in 2010. Eleven Inuit-Métis sites were identified by community members, including FkBd-05, FkBd-08, and FkBd-09, which Stopp (2002: 82) lists as possible Inuit sites. Three sites were identified as places frequented by Newfoundland fishers, including FkBd-09, which Stopp (2002: 82) listed as a possible Inuit site and which was identified through interviews as an Inuit-Métis settlement area (Kelvin 2011: 82-89, 118). The total number of identified historic period sites in Sandwich Bay now includes 31 probable Inuit sites, 15 probable Inuit-Métis sites, and 4 sites that are likely a result of the Newfoundland stationer fishery (Table 1).

Figure 1

Map of Sandwich Bay area showing sites mentioned in text

Map of Sandwich Bay area showing sites mentioned in text

-> See the list of figures

Excavation was used to explore several of these sites further. Between 2003 and 2005, my students and I excavated one Inuit tent feature at Snack Cove 1 (FkBe-1) and the interiors of three sod-walled Inuit houses at Snack Cove 3 (FkBe-3). This work was followed in 2007 by excavation of the known Inuit-Métis sod house settlement (FkBg-24) (Beaudoin 2008). Between 2009 and 2013, my team returned to Sandwich Bay under the auspices of the SSHRC-funded CURA (Community-University Research Alliance) project “Understanding the Past to Build the Future.” We excavated four Inuit sod houses and two tent rings at Huntingdon Island 5 (FkBg-3), one Inuit sod house at Pigeon Cove (FlBf-06), and one sod house at Norman’s Island (FlBg-7) believed to be associated with the Newfoundland stationer fishery (Dobrota 2014; Murphy 2012; Rankin 2013, 2014a, 2014b).

Table 1

Recorded Inuit, Inuit-Métis, and stationer fishery sites in Sandwich Bay

Recorded Inuit, Inuit-Métis, and stationer fishery sites in Sandwich Bay
*

Identified independently as an Inuit, Inuit-Métis, and stationer fishery site.

**

Stopp (2002) identified this site as a “possible” Inuit settlement. This has now been confirmed.

-> See the list of tables

Inuit sites

Snack Cove 1 (FkBe-1) and Snack Cove 3 (FkBe-3) are located less than 250 m apart on the eastern edge of Huntingdon Island in an outer-coast, tundra environment (Figure 1). These were the only sites in Sandwich Bay positively identified as Inuit settlements in 2003, and we hoped to use data gathered there to identify more settlements. Snack Cove 1, situated at 12 m above sea level (asl), contained three tent rings, while Snack Cove 3, at 2 m asl, contained a minimum of three semi-subterranean sod-walled dwellings (Figure 2). The area between the two sites included two kayak supports, several rock-walled caches, and two boulder burials, ranging in elevation from 1 to 15 m asl. Three sod houses were excavated, each composed of a single room approximately 6 m by 3 m in size with paved stone floors (Figure 3). Timber supported the cut sod walls and roof, with boulders used sporadically to hold sods in place. Five-metre entrance passages with cold traps were oriented south towards the beach. Houses 1 and 2 contained a single rear platform, while House 3 had both a rear platform and a side platform. A lampstand was present on the east side of each house just inside the raised cold trap. External middens abutting the entrance passages were ephemeral, but wall middens were located behind the platforms in each house.

Figure 2

Map of Snack Cove 1 (FkBe-1) and Snack Cove 3 (FkBe-3)

Map of Snack Cove 1 (FkBe-1) and Snack Cove 3 (FkBe-3)

-> See the list of figures

Figure 3

House 3 after excavation, Snack Cove 3 (FkBe-3), 2005. Looking south; scale bar is 1m

House 3 after excavation, Snack Cove 3 (FkBe-3), 2005. Looking south; scale bar is 1m

Photo: Lisa Rankin.

-> See the list of figures

Artifact assemblages were small (H1=29, H2=90, H3=210), and were dominated by European material culture (H1=98%, H2=87%, H3=85%). Caches of nails with heads removed and ends flattened were recovered from each house (Brewster 2006: 25), along with sherds of French Normandy and Martincamp ceramics, and Dutch case bottle glass, all of which could have been scavenged from seasonally abandoned European fishing and whaling stations to the south (Figure 4). No formal trade goods were present. Traditional artifacts, including ground slate end-blades, a toy bow, a float plug, sled shoes, and ulus (with metal blades), were randomly located on the house floors (Figure 5). Faunal assemblages (H1 NISP=1009, H2 NISP=1019, H3 NISP=1510) indicate that seal, caribou, sea birds, and fish were important foods (Brewster 2006). Radiocarbon dates on caribou bone sampled from each house provided dates between the mid-17th and mid-18th centuries, but architectural style and European artifact ranges suggested an early to mid-17th century date (Brewster 2006: 27; Ramsden and Rankin 2013). Small, single-family, outer-coast sod houses of similar architectural style have been found elsewhere on the Labrador coast during the early colonization period (AD 1450-1700), when Inuit were first exploring new landscapes (Jordan and Kaplan 1980; Kaplan 2012; Schledermann 1971).

The tent ring excavated at Snack Cove 1 was approximately 12.5 m by 6 m in size and bisected into two compartments by a central row of stones. A hearth feature was present in the eastern section alongside a slab-stone bench, at floor height, which may have been used as a work station (Brewster 2006: 18). A lead drop pendant and a portion of a harpoon foreshaft were the only artifacts recovered. Forty-three faunal fragments from sea birds, ptarmigan, and seal were recovered. Wood carbon dates and architectural style suggest occupation during the 17th century (Fitzhugh 1989; Kaplan 1983). Because of the similar date range and the continuous distribution of site debris between the sod houses and the tent rings, it has been suggested that the two sites represent summer and winter components of the same settlement (Brewster 2006; Fitzhugh 2009).

Figure 4

Selection of European artifacts: A, basal fragment of a Martincamp stoneware jug; B, iron stein handle; C, hilts from a pair of French swords; D, large blue glass bead; E, coin perforated for suspension. A: Snack Cove 3 (FkBe-3); B, C, D, E: Huntingdon Island 5 (FkBg-3)

Selection of European artifacts: A, basal fragment of a Martincamp stoneware jug; B, iron stein handle; C, hilts from a pair of French swords; D, large blue glass bead; E, coin perforated for suspension. A: Snack Cove 3 (FkBe-3); B, C, D, E: Huntingdon Island 5 (FkBg-3)

Photo: Lisa Rankin.

-> See the list of figures

Figure 5

A selection of traditional Inuit artifacts: A, ground slate end-blade; B, miniature soapstone lamp; C, antler foreshaft; D, whalebone harpoon head with iron blade; E, ulu with iron blade and whalebone handle. A, E: Snack Cove 3 (FkBe-3); B, C: Huntingdon Island 5 (FkBg-3); D: Pigeon Cove (FlBf-6)

A selection of traditional Inuit artifacts: A, ground slate end-blade; B, miniature soapstone lamp; C, antler foreshaft; D, whalebone harpoon head with iron blade; E, ulu with iron blade and whalebone handle. A, E: Snack Cove 3 (FkBe-3); B, C: Huntingdon Island 5 (FkBg-3); D: Pigeon Cove (FlBf-6)

Photo: Lisa Rankin.

-> See the list of figures

Huntingdon Island 5 (FkBg-3) is located on the landward side of the island in an inner-bay environment, next to a polynya (Figure 1). The site is identified in Stopp (2002) as a possible Inuit settlement based on the presence of sod houses. I confirmed this classification in 2006 when test units revealed slab-stone floors. The site contained five semi-subterranean sod houses and six tent rings situated at 2-3 m asl (Figure 6). Two rock-walled caches were located within the settlement. A fox trap, a boulder burial, and two rock-walled caches were located approximately 100 m northwest of the living area ranging in elevation from 2-6 m asl. The excavated interiors of Houses 1 and 2 measured approximately 8 m by 7 m, and 7 m by 6 m from wall to wall, and had 5 m long entrance passages opening south towards a small pond (Rankin 2014a). Both houses had slab-stone floors, and were roofed with sod supported by timbers. Central support-beams were present. Slab-fronted, gravel sleeping platforms lined three walls of House 1.

Figure 6

Map of Huntingdon Island 5 (FkBg-3)

Map of Huntingdon Island 5 (FkBg-3)

-> See the list of figures

House 2 had three rock-covered platforms. Multiple lampstands (H1=4, H2=3) were located in front of the platforms, an indication that several families shared each house. The House 1 artifact assemblage included 322 objects, while House 2 had 595. European goods were prominent (H1=51%, H2=46%). Easily scavenged items such as iron scrap, nails, ceramic sherds, and Basque tile were present in both houses. House 2 contained two glass beads, an indication of formal trade with Europeans. Traditional Inuit items were present (H1=68, H2=107), including a soapstone bowl, a miniature soapstone lamp, sled shoes, and dog trace buckles. Several European items were modified for Inuit use, such as roofing tile used as whetstones and oil lamps. Much of the European material was unevenly distributed (Rankin 2013; 2014a). Clusters of nails and roof tile were found separately along edges of different sleeping platforms, perhaps stored for trade.

Faunal remains, recovered from ephemeral exterior middens located on either side of the entrance passages, included seal (76% of the 618 NISP recovered from House 1, 50% of the 1686 NISP recovered from House 2) followed, in order of abundance, by dog, caribou, and whale. Caribou bone, used to date the houses, indicated they were most likely occupied during the mid-17th century, a date consistent with the architectural style, although only one other 17th-century multi-family house has been recorded; House 2 at Eskimo Island 3 (GaBp-3) in Hamilton Inlet, looks remarkably similar to House 1 at FkBg-3 (Kaplan 1983: 420). The radiocarbon dates do not indicate whether both houses were occupied at the same time, but stones were scavenged from the entrance passage of House 1 to construct the entrance of House 2, an indication that House 1 had been abandoned first. Houses 1 and 2 at FkBg-3 were occupied at approximately the same time as Snack Cove, but the houses at FkBg-3 were occupied by multiple families, a style of living that would become the norm in Labrador Inuit winter dwellings by the 18th century.

Houses 3 and 4 at Huntingdon Island 5 are examples of communal houses occupied during the early to mid-18th century (Murphy 2012; Rankin 2014a). These semi-subterranean houses were roughly trapezoidal, measuring approximately 7 m by 7 m along the longest walls, and being entered through 5 m long entrance passages (Figure 7). Unlike other houses at FkBg-3, the entrance to House 4 opened north toward the beach. Exterior middens flanked the entranceways. The houses were constructed of sod, with timber structural elements, including central beams. Interior features included raised gravel sleeping platforms along three walls, storage alcoves (H3=5, H4=4), and lampstands fronting on the platforms (H3=3, H4=2).

House 3 contained 753 artifacts, while House 4 had 1,195. The percentage of European goods was significant (H3=96%, H4=83%). Approximately 10% of each assemblage had been modified for traditional Inuit use as men’s knives, ulus, and harpoon heads. Other objects, such as metal sewing needles, had replaced Inuit versions. Objects of personal adornment, such as beads, pierced coins, and pierced cutlery, were present, as were two French sword hilts, each tied with a thong. Smoking pipes and French ceramics were scattered throughout the house interiors. Traditional items, such as sled shoes, wound pins, soapstone pots, and toys, were still being manufactured. The distribution of material was random, suggesting that cohabitants shared all but a few decorative items, which may have been used to indicate personal status (Rankin 2014a). Faunal samples (H3 NISP=630, H4 NISP=1910) continued to be dominated by seal, although dog and caribou were significant. Caribou bone, in conjunction with architectural style and artifacts such as coins, French ceramics, padlocks, and swords, was used to date the houses to the mid-18th century (Murphy 2012: 35; Rankin 2012).

Communal houses like those excavated at FkBg-3 have been found throughout central and northern Labrador. They were generally located in protected inner bays, where a broader range of food resources was available, where weather conditions were less extreme, and where communities were hidden from strangers (Kaplan 2012: 23). They are believed to be the homes of extended families occupied by a prestigious male, his sons, and their wives (Kaplan 2012). Living communally offered economic advantages by providing sufficient labour for hunting and for preparing furs and sea mammal oils for trade (Kaplan 2012: 23).

Figure 7

House 3 after excavation, Huntingdon Island 5 (FkBg-3), 2010. Looking south; scale bar is 1 m

House 3 after excavation, Huntingdon Island 5 (FkBg-3), 2010. Looking south; scale bar is 1 m

Photo: Lisa Rankin.

-> See the list of figures

Tent Rings 1 and 4 were also excavated at Huntingdon Island 5 (Dobrota 2014). Tent Ring 1 measured 12 m by 5 m and was bisected centrally by a line of rocks. A hearth was present in the western half, and a stone bench, at floor height, was located in the eastern section. No diagnostically Inuit artifacts were recovered, but 113 objects were recovered from the hearth area. Lead shot, gun flints, nails, and French ceramics were abundant. If not for the characteristic shape of the structure, common at 17th-century Inuit sites throughout Labrador (Kaplan 1983), there would be little to connect it to Inuit settlement. No faunal remains were recovered and no radiocarbon dates submitted. The shape of the tent and French ceramics suggest the tent was occupied in either the 17th or 18th century. Tent Ring 4 was oval, measuring 5 m by 6 m. The only feature was a floor-level rock bench along the north wall. The 642 artifacts recovered included two sled shoes, a fragment of soapstone, and a cache of several hundred glass beads. No faunal remains were recovered and no radiocarbon dates possible. The style of tent was commonly used by Inuit during the 18th and 19th centuries, and can only date the occupation generally (Kaplan 1983: 246-247).

A further communal house was discovered at Pigeon Cove when test excavation uncovered a slab-stone floor (Rankin 2011). FlBf-6, located on a sheltered cove of Newfoundland Island (Figure 1) at 2 m asl, contained a single rectangular sod-walled Inuit house measuring 11 m by 9 m, with a 7 m entrance passage opening south. No other Inuit features were present. The house walls and roof were made from sod with timber structural elements including two interior support-beams. Gravel sleeping platforms flanked three walls, each associated with a lampstand. Of the 4,532 artifacts recovered, 83% were of European origin. Four percent of the European assemblage had been transformed into Inuit tools, including ulus and harpoon heads. Other European items, like the 160 beads, were decorative items. French and English ceramic sherds were prominent (1,119), but so were soapstone pots and lamps, along with other traditional items such as sled shoes, a wooden doll, and bow drill pieces. The artifacts were randomly distributed throughout the house. An exterior midden abutting the entrance passage included 4,281 faunal specimens and was dominated by seal, although caribou were also significant. Ten pig bones were present. Caribou bone dates range between 1725 and 1775, and the presence of English ceramics, likely more abundant following the Treaty of Paris in 1763, might indicate an occupation in the late 18th century.

Based on these excavations, Inuit settlement in Sandwich Bay followed trends seen elsewhere in Labrador. Early occupation began with small, single-family houses located on outer islands and developed towards large, communal houses located in inner-bay environments. Sandwich Bay was occupied by Inuit from the early 17th century, an indication that Inuit colonization of the south had been underway earlier than previously demonstrated archaeologically (Jordan and Kaplan 1980), but in keeping with documentary records (Martijn 1980a, 1980b, 2009); by the late 18th century some settlements were substantial. Excavation of these chronologically and structurally varied Inuit sites in Sandwich Bay has led to increased confidence in the classification of the other 27 Inuit sites recoded in Table 1 and provides core data against which to evaluate them and other potential Inuit sites in southern Labrador. However, a comprehensive comparative sample also requires data from Inuit-Métis and stationer fishery sites; this evidence will be described in the following sections and evaluated in the discussion.

Inuit-Métis

In 1775, British merchant George Cartwright established premises in Sandwich Bay, and soon after British settlers began to arrive, often associated with fishing and trading stations. Given that very few British women came to the Labrador coast at this time (Kennedy 1995; Thornton 1977), the settlers frequently chose Inuit wives. These blended families maintained close ties to merchants for whom they would harvest local resources in exchange for credit and access to Western commodities. The first generation of people of ethnically-mixed heritage preferred to marry those of similar background, thus reinforcing a new ethnic consciousness and lifeway (Kennedy 1997: 8). In 1848, Bishop Edward Feild visited the region, commenting that “many of the occupants of Sandwich Bay are pure Eskimaux, but the majority are Anglo-Eskimaux” (Feild 1849: 19). It is therefore likely that the majority of 19th-century sod dwellings in Sandwich Bay were occupied by Inuit-Métis families, and that this mixed heritage became the 19th-century culture pattern of most southern Inuit.

In 2001, I test-excavated a 19th-century sod house, Porcupine Strand 18 (FkBg-24), situated at 2 m asl near the mouth of North River (Figure 1) (Rankin 2002). Since we were unable to identify the occupants’ ethnicity, the site was left for future research. It was later determined that this house was the residence of Charles Williams of Plymouth, England, whose family was listed as the only residents of North River on Levin T. Reichel’s 1872 map of the region (Beaudoin 2008: 32). A marriage licence located in the Anglican Archives at Memorial University records his marriage to his wife Mary McPherson, a local woman of Scottish and Inuit ancestry. Charles and Mary had several children and resided at North River until Charles’ death in 1879 (Beaudoin et al. 2010). With the mixed ethnicity of the Williams household known, the site was chosen for further excavation in the summer of 2007 (Figure 8).

Figure 8

Williams house after excavation, Porcupine Strand 18 (FkBg-24), 2007. Looking west; scale bar is 1 m

Williams house after excavation, Porcupine Strand 18 (FkBg-24), 2007. Looking west; scale bar is 1 m

Photo: Matthew Beaudoin.

-> See the list of figures

Although the surface form of the house was similar to those of Inuit structures, excavation revealed that the house had been built in a European tradition (Beaudoin 2008). The single-room, semi-subterranean house was approximately 10 m by 4 m in size. It had a timber frame and a split lumber floor held together by wrought nails. Sods covered the exterior for insulation. A south-facing, ground-level door was flanked on the outside by a large midden, and a 5 m long saw pit, filled with wood chips and dust, was located to the rear of the structure. Internal features included a stone platform for an iron stove located in the centre of the northern wall, and a large cellar pit located beneath the floor in the northwest corner of the house.

Beaudoin’s (2008) detailed analysis of the 6,158 artifacts and 976 faunal specimens recovered was used to date the occupation to the mid-to-late 19th century, and to interpret activity areas within the house. Barrel hoops, iron strapping, and faunal remains in the eastern side of the house were interpreted as a storage area, while the west end of the site contained an open work space littered with glass beads (Rankin et al. 2012: 77). Animal traps, gun parts, fishing gear, and whalebone sled runners, all likely used to procure subsistence resources and trade commodities, were interpreted as men’s tools. These objects, stored near the door, were limited to 7% of the assemblage, presumably because male householders worked away from home much of the year hunting, fishing, and trapping, and stored their gear closer to locations of use (Rankin et al. 2012: 77). Domestic activities, thought to be associated with women, were represented by 50% of the assemblage and included English ceramic hollowware (several items repaired with mending holes), cutlery, and cooking pots, often with caribou bone handles. These items, as well as 711 glass beads, 54 buttons, and other items of personal adornment, such as perforated coins and hair combs, were distributed throughout the open work space in the western side of the house (Beaudoin 2008:108). The faunal assemblage, mostly recovered from the midden, consisted entirely of wild foods and was dominated by seal and caribou—a wintertime diet. The lack of flatware suggests that most food was prepared as soups and stews (Beaudoin 2008: 136).

The Inuit-Métis settlement at North River represents the occupants’ mixed ethnicities. Constructed by an English man, the house reflected his preference for a wooden floor and walls, a ground-level doorway, and a wood-burning heat source. The saw-pit would have been used to assist with house construction. However, the internal space was organized in keeping with the Inuit values of his wife and children, who were the primary occupants. The artifact distribution was similar to that found at Inuit sites, with the storage and sleeping areas located along walls fronted by an open activity area (Rankin et al. 2012: 77). The Inuit-Métis likely had greater access to European commodities than earlier Inuit families because they worked for merchants, but the whalebone sled shoes and sled runners, the bone handles, and the ceramics repaired with mending holes indicate that Inuit technologies were not abandoned, and some were adopted by European men. The large number of beads, and perforated coins, suggest that decorative clothing and other goods continued to be made by the Inuit-descended women. Faunal remains indicate that the diet was based on traditional wild foods, and prepared in hollowware vessels, which had replaced soapstone.

In 2007 and 2010, we conducted a series of ethnographic interviews and archaeological surveys for the purpose of identifying additional Inuit-Métis settlements for comparison (Kelvin and Rankin 2014; Pace 2008). We learned that the Sandwich Bay Inuit-Métis practised seasonal transhumance that took them to the outer headlands and islands between July and October to fish for cod, and deep into the sheltered bay during winter months, where they had access to sufficient wood fuel and game, and could reside closer to their trap lines (Kelvin and Rankin 2014: 123). Gardens were planted at the winter residence before families departed for their summer homes so that vegetables could be harvested on their return in the autumn. Cod fishing was a communal activity, and generally several families would maintain homes at the same location, but winter residences were more isolated in order to ensure that everyone had access to the resources they would require for the winter. Sometimes, winter residences, like the one occupied by the Williams family, were also established near salmon-fishing locales, allowing the household to remain in the winter residence until the end of the salmon season in July. Other families made a third move to salmon stations between the trapping and cod-fishing seasons (Kelvin 2011: 54). Construction of all seasonal dwellings was similar, although most interviewees indicated that their winter homes were their real homes and that more effort was put into upkeep (Kelvin 2011). Interviewees reported that their ancestors had occupied rectangular, wooden houses made of split timbers and insulated with sods, much like the Williams residence in North River (ibid.: 96-97). Every family would pack and move their belongings when going from one home to the other, so their possessions at each location would be similar, although the presence of seasonal items, such as trapping and gardening tools, might vary between residences.

Many of these 19th-century locales are still used by the same families, having been passed down from one generation to the next (Kelvin 2011: 96). Interviews and archaeological surveys revealed the locations of 14 more Inuit-Métis settlements (Kelvin 2011; Pace 2008). Archaeological testing was carried out at eight of the locations, including one warm season and seven cold season sites. All sites were located between 1 m and 3 m asl. Test excavations at summer and winter habitation sites revealed semi-subterranean, single-room houses ranging in size from 7 m by 4 m to 14 m by 10 m. Each had ground-level entranceways, split wood floors and walls, cut sod insulation, 19th-century ceramics, and, in some cases, a saw pit and/or garden behind the house (Kelvin 2011: 80-91), an indication that Inuit-Métis architecture and site activity areas followed an ethnically-derived pattern that was distinct from that of earlier Inuit settlement in Sandwich Bay.

European fishers

By the early 19th century, the migratory Newfoundland fishery had become firmly established. Each year, fishers from Newfoundland would travel north to Labrador for the summer fishing season (Kennedy 1995). Many fishers lived aboard schooners during the fishing season, but the “stationers” carried out fishing operations from onshore rooms, which they built or rented. Some stationer crews were all male, but others involved entire families, transported by steamer with necessary personal belongings to the Labrador coast for the summer. The surface remains of their summer homes, or tilts, resemble those of Inuit and Inuit-Métis dwellings. Four known sites in Sandwich Bay, all located on outer islands, or headlands, are likely associated with the stationer fishery. Round Island 1 (FkBd-9), originally recorded by Stopp in 1992, was shared by an English merchant and 10 outfits from Newfoundland around the mid-19th century (Anderson 1988: 17), and contains 11 sod structures, 15 wooden structures, and three bawns used to dry fish. Three other locations—Norman’s Island (FlBg-7), Handy Islands 2 (FlBg-11), and the Creek (FkBd-18)—were identified by local informants as places occupied by the stationer fishery (Kelvin 2011: 118; Pace 2008). I had test-excavated sod houses at two of the locations during a survey in 2001 (FlBg-7 and FlBg-11), but had been unable to determine the ethnicity of site occupants (Rankin 2002). As a result of the information provided by community interviews, we returned to Norman’s Island in 2013 to excavate the single sod house at that location (Figure 1) (Rankin 2014b).

The Norman’s Island site is located on the southern shore of the exposed, outer island. It is 6 m asl. The single room, semi-subterranean house was no more than 7 m long and 6 m wide, with a paved ground-level entranceway (Figure 9). The house was t-shaped, but this shape is likely due to a large rock outcrop that truncated the eastern wall. The floor had been prepared with crushed shell and sand from the beach below. The roof and walls were supported by uncut timber, which had been nailed together and then covered with cut sods. A small earthen wall-bench, located to the west of the entranceway, terminated with four floor-level paving stones used as a hearth. No other features were present within the structure, but a flat, rectangular feature measuring 8 m by 10 m located to the west of the house may have been used as a bawn for drying fish.

Figure 9

Fishery house after excavation, Norman’s Island (FlBf-7), 2013. Looking southeast; scale bar is 1 m

Fishery house after excavation, Norman’s Island (FlBf-7), 2013. Looking southeast; scale bar is 1 m

Photo: Lisa Rankin.

-> See the list of figures

The 646 artifacts recovered were entirely European in origin. Fifty-eight percent of the assemblage was composed of iron nails, most likely relics of wall and roof construction. Domestic activities, such as cooking, sewing, sleeping, and smoking, were well represented by 41% of the assemblage. One piece of lead shot was the only artifact that might be attributed to hunting. McDougall pipes and Cochran and Co. ironstone, both manufactured in Glasgow after 1846, were both present, but nothing in the assemblage suggests that the house predated the mid-19th century. There was both a male and female presence as indicated by hair pins, a decorative women’s brooch, and two metal heel plates from a pair of men’s shoes. A leather luggage tag, bearing the initials E.P.M., was recovered near the doorway. The initials are not attributable to any of the Inuit-Métis known to have occupied the region (Understanding the Past… n.d.). The artifacts were distributed throughout the house, but were most prominent alongside the bench. The small faunal assemblage (NISP=221) included pig and cow remains (15%), as well as various local species, such as duck, shellfish, cod, and salmon, all recovered inside the house, which were available during the summer and early fall.

The site is best interpreted as a briefly occupied, 19th-century summer camp. The expedient nature of house construction was indicated by the lack of significant flooring and limited internal features. The house shape, style of entrance, floor, internal and external features, and artifact and faunal assemblages were distinct from characteristics of earlier Inuit occupations, and also from those of the houses occupied by Inuit-Métis in Sandwich Bay. Given that local informants associated this location with the stationer fishery, it is likely that these differences are meaningful and represent a different manner of inhabiting the outer coast of Sandwich Bay. Some of the characteristics present at Norman’s Island 1 were also recognized at the stationer camp at Handy Islands 2.

The Handy Islands 2 site (FlBg-11), located on an outer island at 2 m asl, contains as many as 12 sod structures; all semi-subterranean dwellings no more than 3 m by 4 m in size, and roughly rectangular in shape. None of these houses have been excavated but test excavations in 2001 at eight of the structures revealed shell and sand floors, ground level entranceways, and a few fragments of 19th-century wheat-pattern ceramics. No other features were observed inside or outside these structures, but more work is required. All four probable stationer sites in Sandwich Bay contain different numbers of sod houses, as well as houses of different sizes, perhaps because the size of the crew using each locale was different. The sites range in elevation above sea level from 2 m at Handy Islands 2 to 11 m asl at the Creek. However, there are similarities between the four sites. Each is located in a good harbour on the outer reaches of the bay. The archaeologically tested sites indicate that single-room, semi-subterranean houses with ground-level entrances and crushed-shell floors were used. Fish bawns were present at two of the sites. The household and site-based features indicate that the stationer sites were used and occupied differently from those occupied by Inuit during the contact period, or by 19th-century Inuit-Métis in Sandwich Bay.

Discussion and conclusion

Upon excavation, the sod-walled houses occupied by Inuit, Inuit-Métis, and stationer fishers in Sandwich Bay have been shown to be distinct. Attributes corresponding to the sites examined are outlined in Table 2. Inuit and Inuit-Métis settlements were constructed with more effort than those occupied by fishers and included prepared floors, as well as substantial internal and external features, such as sleeping platforms, lampstands, entrance passages, caches, storage cellars, and saw pits. Inuit and Inuit-Métis homes were prepared for winter occupancy, and warmth was prioritized. The internal use of space was similar, reflecting the shared culture of the occupants, but methods of construction were unique to each ethnic background, thus making the homes readily identifiable.

In contrast, the house on Norman’s Island had a more expedient nature. The floor was prepared with crushed beach shell, and few features were present. Similar to Inuit tent rings, the fisher’s houses were for summer use. Occupants of both likely spent much of their time outside engaged in other pursuits. Inuit-Métis and Newfoundland fishers both pursued cod during the summer months and were therefore attracted to similar site locations. Round Island 1 (FkBd-9) is a perfect example (Table 1). Stopp (2002) suggested it may be of Inuit origin based on the presence of sod houses, while Anderson (1988) identifies it with particular fishing crews, and local informants recognize it as a “Métis place.” It is unlikely that this site is Inuit, but the presence of both sod and wooden structures suggests both fishers and Inuit-Métis may have lived here. Inuit-Métis summer homes were constructed of wood like their winter homes, but sod insulation may have been used more sparingly for summer dwelling.

Site location is not enough to determine ethnic occupancy. Nor is elevation, although fishery-related sites are more likely to be found at higher elevations, perhaps because, as locals, Inuit-Métis had greater access to the most accessible land. However, Stopp’s (2002) suggestions that low-elevation rock features were likely to be Inuit is strengthened by the fact that fishers and Inuit-Métis did not appear to use them. Nevertheless, Inuit features, including tent rings, burials, and caches, were also recorded at elevations of 15 m asl.

Material culture assemblages indicate that fishers relied on European manufactured items, while Inuit and Inuit-Métis used both European and traditional objects. Inuit sites contained French material reflecting the earlier period of occupation, but if 19th-century Inuit sites are located in Sandwich Bay this will change. Faunal assemblages at Inuit and Inuit-Métis sites were similar. Seal and caribou were important foods for both. Ten pig bones were recovered from Pigeon Cove, making up less than 1% of the assemblage. Use of traditional foods may have reflected both access (food from domesticated sources may have been expensive and in limited supply) and cultural preference. Inuit-Métis women may have favoured preparing meals that were most familiar. In contrast, fishers ate food from both domesticated and local sources.

Excavation, survey, and community interviews undertaken in Sandwich Bay demonstrate that it is possible to distinguish between Inuit, Inuit-Métis, and stationer fishery settlements. Sites were constructed and inhabited differently by each ethnic group. The research has identified site attributes that may help to determine site ethnicity in Sandwich Bay and elsewhere in southern Labrador. Inuit settlements were shown to be the most unique ones, with very distinctive house floors, and internal and external features. Some attributes are common to all settlement types. Site location cannot be considered a reliable indicator of ethnicity, as all ethnic groups inhabited similar locales, and material culture and faunal assemblages should only be used to interpret site ethnicity if substantial assemblages are available.

Table 2

Attributes of Inuit, Inuit-Métis, and stationer fishery settlements in Sandwich Bay

Attributes of Inuit, Inuit-Métis, and stationer fishery settlements in Sandwich Bay

-> See the list of tables