Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 42 Introduction Since the mid-19th century, shell middens have played a significant role in the development of archaeological methodology. These high-density accumulations of carbonate-rich invertebrate remains create unique environments that preserve the organic remains of human activity usually lost in more-acidic contexts. Such preservation of faunal and floral remains, in addition to ceramics, lithic tools, and other artifacts, makes shell middens ideal features to explore past human activities and the environments that shaped—and were shaped by—these activities. The diverse nature of the remains recovered from shell middens requires archaeologists to engage in interdisciplinary research to comprehend fully the roles that these features played in past human societies. These interdisciplinary roots stretch back to the mid-19th century when the extensive kjoekkenmoeddinger or “kitchen-middens” found along Scandinavian shores were studied not only by archaeologists such as Jens J.A. Worsaae, but also by biologists and geologists (see Morlot 1861:292; Trigger 1986:xii, 1991:82). A report by the Swiss geologist A. Morlot (1861) describing the work of Worsaae and others had an important impact on archaeologists in both Europe and North America ((Trigger 1986: xiii)). Both Trigger (1986:xiii) and Claassen (1998:3–5) note that the early period of shell midden research saw innovations in stratigraphic approaches to excavation and attempts to reconstruct the depositional processes and rates that contributed to the formation of middens. Following this early period, attention shifted away from the shells that comprised the matrix of these middens and towards the artifacts and ecofacts found therein (Claassen 1998:5). Over the last thirty years, greater attention in shell midden archaeology has been paid to the information that lies in the shells themselves. While the first avocational and professional archaeologists were drawn to shell middens for the artifacts and ecofacts they preserved, by the late 20th century researchers had used the most-abundant remains of these middens—the shells—for such diverse purposes as calculating the importance of shellfish in the diets of ancient peoples (Meighan 1972; Quitmyer 1985),1 determining the season in which the shells were harvested (see Andrus and Crowe 2000; Custer and Doms 1990; Deith 1983, 1985; Quitmyer and Jones 1997; Spiess and Lewis 2001:49–54), understanding the economic and political uses of shell (see Ceci 1984, Claassen 1991), and reconstructing the past environments from which these mollusks were harvested (see Ambrose 1967, Claassen 1998, Jerardino 1997, Quitmyer and Reitz 2006). Recent advances in isotope analysis and thin-section microscopy have allowed further refinement of seasonality studies, the results of which have challenged our understanding of the settlement patterns of ancient peoples who dwelt along the east coast of North America (see Betts et al. 2017 [this volume], Culleton et al. 2009, Quitmyer et al. 2005, Villagran et al. 2010). These latest studies are expanding shell midden archaeology beyond interpreting the site itself to understanding how a site relates to the broader ecosystem within which it was created. Some studies of shell middens along the northeastern coasts of the Atlantic Ocean have noted changes in mollusk species over time, although a surprising number of publications favor descriptions of overall weight, species identity, and relative Temporal Changes in Marine Shellfish? A Preliminary Archaeological Perspective from the Northumberland Strait Michelle Lelièvre* Abstract - This paper reports results of test excavations conducted at BjCo-02, a shell midden on the Mi’kmaw island of Maligomish located off the southern coast of the Northumberland Strait in northeastern Nova Scotia. While the site yielded few artifacts, preliminary observations indicate changes in the proportions of the two dominant shellfish species (eastern oyster and soft-shell clam) between 1500 and 500 y.b.p. The appearance of eastern oyster at ca. 1500 y.b.p. and its virtual absence ca. 500 y.b.p. suggest that the Maligomish midden conforms and, at the same time, challenges previously observed patterns at other shell midden sites in the Maritime Provinces of Canada and along the east coast of the United States. Drawing on the preliminary archaeological data from BjCo-02, and previous palynological studies from the region, this paper argues for the recognition of Northumberland Strait as a unique environment within the broader northeastern North American region. The paper also makes recommendations for future research to confirm the apparent species shift, including the proper calibration of radiocarbon dates to account for marine reservoir effects. North American East Coast Shell Midden Research Journal of the North Atlantic *Department of Anthropology and American Studies Program, The College of William and Mary, Williamsburg, VA, USA; malelievre@wm.edu. 2017 Special Volume 10:42–58 Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 43 species proportions rather than reporting on these measurements stratigraphically. This approach may be due to research questions focused on site-level analyses of human activities such as food gathering, processing, and preparation, and on the recovery of chronologically diagnostic artifacts rather than regionally directed questions of environmental reconstruction and change over time. Given the depositional complexity of these sites, and the biases inherent in excavating and interpreting them, some archaeologists are skeptical about the possibility of revealing an accurate understanding of the peoples whose activities created shell middens (see Peacock 2000, Stein 1992:8–10). Stratigraphic excavations at BjCo-02, a shell midden site located on the south side of Maligomish (Indian Island) in Nova Scotia, Canada (see Fig. 1), provided an opportunity for a fine-scale analysis of depositional change over time. Maligomish is a 16-ha island that forms part of the reserved lands of the Mi’kmaw2 First Nation of Pictou Landing. It is located in Merigomish Harbour, along the southern coast of the Northumberland Strait—the body of water that separates the Canadian province of Prince Edward Island from the mainland. The excavations were the culmination of the archaeological phase of a 27-month interdisciplinary and collaborative research project with members of the Pictou Landing First Nation. In discussing the excavation’s results, this article focuses particularly on the change in mollusk species over time observed in a 1 m x 1 m test unit. The presence of eastern oyster (Crassostrea virginica) during time periods when the species was not observed at other shell midden sites in northeastern North America (hereafter, the Northeast) warrants further research to understand the environmental and/or cultural factors that distinguish the formation of the Maligomish midden. Rather than treating the Northeast region monolithically— as many archaeologists do—the appearance of the eastern oyster in a Northumberland Strait midden calls for a recognition of variations within this broad region and the implications of these differences on past human activities. This article begins by discussing briefly the history of shell midden archaeology in the Northumberland Strait, focusing particularly on Merigomish Harbour. It then describes the excavation and collection strategies employed during the 2008 field seasons on Maligomish. The article continues by describing the results of a quantitative comparison of eastern oyster and softshell clam (Mya arenaria) in each of five arbitrary strata in an excavated 1 m x 1 m test unit, placing these results in context with a selection of other shell midden sites in the Northeast for which the presence of eastern oyster has been reported. Possible biases and sampling errors introduced to the study are also discussed. Finally, I make recommendations for future research that could confirm changes in mollusk species over time and refine our understanding of the Northumberland Strait’s unique environment and climate. Research Background - The Northumberland Strait and Merigomish Harbour Few places along the 13,300 km of Nova Scotia’s coastline have had waters warm enough in either the pre- or post-European-contact periods to sustain populations of eastern oyster. One such region is the Northumberland Strait, a body of water covering approximately 12,000 km2 between the Canadian provinces of New Brunswick, Nova Scotia, and Prince Edward Island (see Fig. 1). Kranck (1971:4) notes that during much of the post-glacial period, what is now the Northumberland Strait was two estuaries separated by a ridge of land. This area became a continuous body of water ca. 5000 y.b.p. when the isthmus to Prince Edward Island was breached (Krank 1971:4). Glacial erosion then contributed to the enlargement of these drowned river systems (Davis and Browne 1996:241). With depths as shallow as 10 to 24 m in its western and central areas (Kranck 1971:1), the Northumberland Strait’s waters have an average annual temperature of 20 °C, making them the warmest marine waters north of the Carolinas. These warm temperatures support a native eastern oyster population in the “shallow inshore waters of protected bays and estuaries” (Davis and Browne 1996:242). The Nova Scotia (i.e., southern) shore of the Northumberland Strait is defined by alternating low ridges and valleys (Davis and Browne 1996:108). At its west end, the shore is a coastal plain underlain by fine, red sandstones of the Late Carboniferous period. Moving east, the underlying geology of this shoreline becomes older, with Middle Carboniferous strata that have been cut through by drowned river valleys forming several harbors (Davis and Browne 1996:108). The forests of this area are dominated by black spruce (Picea mariana), jack pine (Pinus banksiana), white spruce (Picea glauca), red spruce (Picea rubens), and red maple (Acer rubrum), with eastern hemlock (Tsuga canadensis) and white pine (Pinus strobus) also occurring (Davis and Browne 1996:109). Tidal marshes are common along the shores of the Northumberland Strait. These diverse ecosystems provide habitat for native and migratory fauna, invertebrates, and various grass species. Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 44 Situated at the eastern extent of the exposed coastal plain, Merigomish Harbour is one of the largest (~25 km2) and deepest along the Northumberland Strait. Located in Pictou County, NS, this long and narrow harbor forms a kind of archipelago, with 11 islands ranging in size of several square kilometers (e.g., the aptly named Big Island of Merigomish) to a few hectares. The mainland south of Merigomish Harbour forms the eastern extent of the Northumberland coastal plain, resulting in an “undulating to rolling” topography (Webb 1990:10). South of the coastal plain are the Antigonish Highlands, which include hills rising as high as 340 m. The bedrock geology surrounding Merigomish Harbour includes two primary geological groups of the Carboniferous Period: “Pictou” to the north and “Mabou” to the south (Kepple 2000). Three major rivers empty into Merigomish Harbour, forming estuary systems. This rich environment has made Merigomish Harbour, and the Northumberland Strait more broadly, an important area for the indigenous Mi’kmaq at least since the Middle Woodland Period (ca. 2400–1000 y.b.p.). To date, Merigomish Harbour has the densest concentration of archaeological sites along the Northumberland Strait, with shell middens dating to the Middle and Late Woodland periods (ca. 1000–500 y.b.p.) located on its southern coast and on most of its 11 islands (see Table 1). The histori- Figure 1. Maps of the study region. Top left: the Maritime Provinces of Canada. Bottom right: Inset of the Northumberland Strait. Top right: Inset of Merigomish Harbour. Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 45 cal record documents Mi’kmaw use of the harbor as early as the 18th century. Mi’kmaq from the Pictou Landing First Nation, as well as other communities from mainland Nova Scotia and Cape Breton Island, continue to use the islands and fish the harbor (see Lelièvre 2012). One of these islands—Maligomish— is located in the western end of Merigomish Harbour ~300 m east of Quarry Island and is accessible only by boat or canoe (UTM: ~5052600N, 540200E). Its bedrock geology indicates that Maligomish is younger than the surrounding area, being part of the Permian Period’s Cumberland Group (Kepple 2000). The vegetation on Maligomish consists of hardwoods, mostly Querqus (oak), Betulus (birch), and Acer (maple) species. A few softwood trees are concentrated at the west of the island. There are coarse beach grasses along the northern and eastern shores of the island. The western and southern shores have rocky beaches. The sea level can vary up to 1 m between high and low tides. The vegetation on the island appears quite different to that encountered on the mainland. The island is relatively open-forested compared to the dense second-growth forests that have dominated the mainland for the past 50 years. Maligomish is one of two islands in the harbor that are part of the Pictou Landing First Nation’s reserved lands. Most maps indicate Maligomish as “Indian Island”. Many people from the Pictou Landing First Nation and other Mi’kmaw communities have built camps along the perimeter of the island. Although most community members only use these camps in the summer, others frequent the island almost year-round. Mi’kmaq from the Pictou Landing First Nation participate in the annual food fishery during the autumn salmon run in the estuary of Sutherland’s River. Some collect shellfish and spear eels. Others hunt deer and rabbits. The archaeological investigation of Merigomish Harbour began with the Reverend George Patterson, who explored Merigomish and Pictou Harbours in the 19th century in search of pre-contact artifacts. He presented his collection to Dalhousie College (now Dalhousie University) in 1888 (see Patterson 1877:28–31], 1889). W.J. Wintemberg, then working at the National Museum of Man (now the Canadian Museum of History), studied this collection when he visited Nova Scotia in 1912 (CMH-L&A 1912). That summer he also traveled to Pictou County where he located three of the shell middens reported by Patterson and discovered an additional three in Merigomish Harbour. Harlan Smith, also at the National Museum of Man, returned to Merigomish Harbour in 1914 and identified 12 additional shell middens, including one located on the south side of Maligomish (Smith and Wintemberg 1929:6).3 The avocational archaeologist, John Erskine, excavated the midden on Maligomish in 1960. In his 1961 report, he bemoaned the context of the midden having been ruined by Harlan Smith’s digging. Erskine (1961:22) described two “sub-sites” within the midden, which he considered to be camps. As will be described in greater detail below, Erskine noted the presence of eastern oyster remains in both sub-sites. Some of Erskine’s collections from Maligomish and other middens in Merigomish Harbour are now stored at the Nova Scotia Museum in Halifax. Since Erskine’s excavations in Merigomish Harbour, little archaeological work has been undertaken in the area. A few local Mi’kmaq report other archaeological activity on the island during the 1960s or 1970s. It is uncertain whether this more recent work was conducted by a qualified archaeologist or by pothunters posing as archaeologists. David Keenlyside, the former Curator of Atlantic Provinces Archaeology with the Archaeological Survey of Canada, surveyed the Merigomish Harbour shell middens in 1978, but did not conduct sub-surface tests on the Maligomish middens (see Keenlyside 1980). In the early 1990s, Ronald Nash directed an undergraduate excavation at Kerr Point—one of the Reverend Patterson’s original collection areas (see Snow 1994). By this time, the Nova Scotia Museum had assigned Borden numbers to the sites identified by Smith and Wintemberg (see Borden and Wilson 1952 for more on Borden numbers). The shell midden on Maligomish was given the designation BjCo-02. While Erskine paid closer attention than Smith to changes in the stratigraphy of the Merigomish Harbour shell middens, he did not excavate systematically or with specific research questions in mind. The archaeological fieldwork conducted on Maligomish in 2007 and 2008 was designed to answer questions regarding the role played by changes in the social and political lives of pre-contact Mi’kmaq. The shell midden on Maligomish was considered ideal for answering these questions given its long association with Mi’kmaw activity and the potential for recovering organic remains that could indicate the seasons during which the site was occupied. Moreover, the incremental growth of invertebrate faunal remains could be observed to determine the season of harvest. Methods Before the fieldwork began, several community meetings were held at the Pictou Landing First Nation, during which some Mi’kmaq expressed Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 46 Table 1. AMS radiocarbon dates from the Northumberland Strait. All dates were calibrated using OxCal’s (v. 4.2.4) IntCal 13 curve. The Marine13 calibration curve was applied to dates obtained from marine shells (see Reimer et al. 2013). The ΔR value for the Pugwash Basin (-60 ± 40) was applied for the Marine13 calibration (see McNeeley et al. 2006). . n.r. = not reported. Median Depth Borden 14C age Corrected 14C age (y.b.p.) age AA Lab # Sample ID (cm) Site name number Material Weight (g) δ13C F (y.b.p.) 68.2% 95.4% (yr BP) AA90571 A2008NS73-5C2 5–10 Maligomish BjCo-02 Mya arenaria 4.383 +2.2 0.9023 ± 0.0059 826 ± 53 506–418 532–326 457 AA94602 A2008NS73-5C3 10–15 Maligomish BjCo-02 Mya arenaria 6.338 1.8 0.8828 ± 0.0038 1001 ± 34 620–550 645–522 584 AA94603 A2008NS73-5C4 15–20 Maligomish BjCo-02 Mya arenaria 4.853 0.9 0.8876 ± 0.0038 958 ± 34 598–516 625–500 554 AA90572 A2008NS73-5C5 20–25 Maligomish BjCo-02 Mya arenaria 2.732 -1.9 0.7864 ± 0.0052 1930 ± 53 1537–1405 1608–1344 1476 AA86506† A2008NS02-TP2-2 10 Maligomish BjCo-02 Mya arenaria 7.338 2.1 0.8943 ± 0.0040 897 ± 35 534–482 593–450 509 AA86507† A2008NS02-TP2-6 20 Maligomish BjCo-02 Mya arenaria 5.932 2.1 0.8991 ± 0.0041 854 ± 35 509–456 532–424 482 AA86508† A2008NS02-TP2-8 26–31 Maligomish BjCo-02 Crassostrea virginica 48.512 0.9 0.7846 ± 0.0036 1949 ± 36 1540–1440 1599–1392 1500 S-973* n.r. n.r. Olding Island BjCo-05 Alces alces n.r. -20 n.r. 280 ± 60 437–155 496–... ‡ 359 Beta-65930/ CAMS-9319§ n.r. ~15 Kerr Point BjCo-15 Charcoal n.r. -26.2 n.r. 2130 ± 60 2297–2004 2311–1952 2118 GSC-3218* n.r. 20–30 Delorey Island BjCj-09 Charcoal n.r. -25.0 n.r. 810 ± 70 788–676 910–660 742 I-11619* n.r. 20–30 Delorey Island BjCj-09 Charcoal n.r. n.r. n.r. 1595 ± 80 1563–1393 1694–1328 1486 S-1602* n.r. 22–24 Cox/Swanson BkCq-10 Charcoal 5.5 -25 n.r. 1420 ± 70 1386–1284 1520–1184 1334 S-1604* n.r. 40–42 Cox/Swanson BkCq-10 Marine shell n.r. -0.0 n.r. 1110 ± 85 755–570 845–524 672 S-1603* n.r. 54 Cox/Swanson BkCq-10 Charcoal 7.5 -25 n.r. 840 ± 60 891–687 909–675 762 † Dates previously published in Mudie and Lelièvre (2013). * Dates previously published in the Canadian Archaeological Radiocarbon Database (CARD). See http://www.canadianarchaeology.ca/ § See Snow 1994:15. ‡ OxCal output suggests the date may extend out of range of 280 ± 60. Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 47 concern that the archaeological work could disturb the spirits whom many believe dwell on the island. In order to mitigate these concerns, the project was designed in two phases. The first was a full surface survey of the island to record any previously undocumented archaeological features. These features were measured, sketched, and photographed only; no sub-surface testing was conducted. The second phase consisted of excavating limited test units in the shell midden. At the suggestion of one of the field assistants, who was a Mi’kmaw woman from the Pictou Landing First Nation, our field crew took further precautions against offending the spirits of Maligomish. Each morning she led us in a simple ceremony to purify ourselves with sweetgrass and make an offering of tobacco. The methods described below focus on those employed during the second phase of archaeological investigation on Maligomish, which occurred in the spring and fall of 2008. BjCo-02 is situated in one of the busiest areas of Maligomish (Fig. 2). A path that leads from the island’s temporary wharf to the church runs through the middle of the midden, and several buildings stand on either side. The northern edge of the midden is at the base of a steep hill that slopes south of the church and cemetery. The southern edge of the midden has been subjected to constant erosion. A high density of shell and very dark brown to black soil is visible in the bank south of the midden. In the spring of 2008, several large, discontinuous areas (~10–15 m x 15–25 m) of midden context were exposed on the surface (see Lelièvre 2008:6). The sub-surface testing phase of the archaeological investigations on Maligomish had several objectives. In the spring of 2008, we made an attempt to discern the boundaries of the midden and to understand its stratigraphy by excavating three 40 cm x 40 cm test pits (TP1-TP3) placed in each of the north, middle, and south areas of the midden (see Fig. 2). These units were excavated following the natural stratigraphy of the midden. All contents were hand-screened using a 0.64-cm (1/4-inch) mesh. Whole samples of eastern oysters and soft-shell clams that had intact chondrophores were collected for future seasonality analyses (see Black 1993:88–92, Spiess et al. 2006). We also collected two separate sets of bulk soil samples—one for macrobotanical analysis and another for pollen analysis. The bulk samples were collected “arbitrarily” from every 10 cm of the profiles (see Mudie and Lelièvre 2013). Of these three test pits, TP2 was the most productive, yielding numerous intact eastern oyster and soft-shell clam samples, along with a few lithic tool fragments and charcoal. In the fall of 2008, our field crew returned to Maligomish to continue our sub-surface testing of the midden. Having heard reports from some members of the Pictou Landing First Nation that the midden extended further west, east, and north than the area tested in the spring, we began the fall field season Figure 2. Site plan of BjCo-02 indicating the locations of Operations 1–5 (conducted under permit A2008NS73) and Test Pits (TP) 1–3 (conducted under permit A2008NS02). ■ = test pit. • = auger test. Note: units 4B and 5C were not excavated, and so are not shown. Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 48 remains, lithics, and ceramics were collected. Most of these materials were recovered from the screen. However, one in situ charcoal sample was collected from sub-operation 4D. To observe the depositional history of the midden, we established three 1 m x1 m units contiguous with test pits 4A and 4B (Fig 2: Operation 5). Our goal was to excavate all three units, which would have provided a profile of ~3.5 m in length. However, the field season allowed only enough time to excavate one 1 m x 1 m unit. In the southwestern corner of each unit, we left ~25 cm2 unexcavated for column sampling (see Spiess et al. 2006). Unlike the test pits that had been opened to this point, the 1 m x 1 m units were not excavated according to their natural stratigraphy. Instead, 5-cm contoured, arbitrary levels were used in order to control observations of potentially confusing midden deposits (see Fladmark 1978:90). Plan views of the bottom of each 5-cm level were drawn and photographed. The volume of matrix removed from each arbitrary level was tracked by counting the number of 9-L buckets filled during excavation (see Table 2). All of the contents of each arbitrary level were screened through a 0.64-cm (1/4-inch) mesh, separating any recovered remains according to material type. In by digging seven 40 cm x 40 cm test pits (Fig. 2: Operation 1) and numerous auger tests (Fig. 2: Operations 2 and 3) to better define its boundaries.4 From the test pits of Operation 1, lithics, floral, and faunal remains, including samples of soft-shell clam were collected. The auger tests of Operations 2 and 3 yielded negligible shell remains. With the boundaries defined, our work then shifted to excavating test pits located in areas of the midden undisturbed by either Smith or Erskine. While no records existed of Smith’s excavations, Erskine had recorded the rough location of his testing grid, along with a schematic of the units that were excavated. With this information, an area of the midden less likely to have been disturbed was identified. 5 Three 40 cm x 40 cm test pits were excavated following the natural stratigraphy, with changes in shell density and matrix, and the occurrence of artifacts and ecofacts, noted as the digging proceeded (Fig. 2: Operation 4). As with the previously excavated test pits, all of the contents of Operation 4’s units were screened through a 0.64-cm (1/4-inch) mesh. Whole soft-shell clam shells, soft-shell clam chondrophores, whole oyster shells, whole shells of other invertebrate species,6 vertebrate fauna, floral Table 2. Invertebrate fauna results from unit 5C. Median calibrated Matrix Absolute Relative radiocarbon volume weight weight Level age (y.b.p.) (L) Species (g) (%) NISP 1 (0–5 cm) N/A 70.5 Mya arenaria 1200 87.3 Crassostrea virginica 175 12.7 Unidentified 5 2 (5–10 cm) 457 73.0 Mya arenaria 5500 85.9 Crassostrea virginica 900 14.1 Mytilus edulis 5 Crepidula fornicata 1 Unidentified 5 3 (10–15 cm) 584 64.0 Mya arenaria 2600 55.3 Crassostrea virginica 2100 44.7 Mytilus edulis 2 Crepidula fornicata 27 Unidentified 4 4 (15–20 cm) 554 56.5 Mya arenaria 1100 35.5 Crassostrea virginica 2000 64.5 Mytilus edulis 2 Crepidula fornicata 26 Mercenaria mercenaria 4 Spisula solidissima 2 Unidentified 12 5 (20–25 cm) 1476 78.0 Mya arenaria 1200 27.9 Crassostrea virginica 3100 72.1 Mytilus edulis 5 Crepidula fornicata 26 Mercenaria mercenaria 5 Spisula solidissima 1 Unidentified 5 Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 49 den in 2008. Most of the invertebrate samples were identified by species or genus. For expediency, the most-abundant taxa (soft-shell clam and eastern oyster) were weighed and the less-abundant taxa (e.g., blue mussel and Atlantic surf clam) were counted.7 In those contexts where the samples were counted, we calculated the number of identified species (NISP), rather than the minimum number of individuals (MNI). There have been no formal zooarchaeological situ charcoal and fauna samples were recorded and collected. As in Operation 4, we collected whole soft-shell clam shells, soft-shell clam chondrophores, whole oyster shells, whole shells of other invertebrate species, vertebrate fauna, floral remains, lithics, and ceramics. To date, only basic species identification and quantification have been conducted for the invertebrate remains collected from the Maligomish mid- Figure 3. Drawing and descriptions of the north profile of unit (operation) 5C. Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 50 analyses on the invertebrate samples.8 The column samples collected from unit 5C were processed using flotation at the Paleoethnobotany Laboratory at the Memorial University of Newfoundland under the supervision of Michael Deal. Sub-samples have been preserved for future analyses. Accelerator mass spectometry (AMS) radiocarbon dates were obtained for four of the five arbitrary levels excavated in unit 5C. Three AMS radiocarbon dates were obtained for TP2. All dates were obtained from marine shell samples (see Table 1). Results The strategy to excavate the 1 m x 1 m test units in arbitrary 5-cm levels was designed to detect Figure 4. Top: Comparison of the proportions soft-shell clam (Mya arenaria) and eastern oyster (Crassostrea virginica) collected from each level of unit 5C (by weight). Bottom: Calib rated radiocarbon dates for levels 2–5 of unit 5C. Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 51 time, have been noted by avocational and professional archaeologists and museum curators through much of the history of shell midden archaeology in the Northeast. For example, in 1934 the Nova Scotia Museum purchased artifacts that a Mi’kmaw man had recovered from a shell midden on an island near Maligomish. The museum curator H. Piers made the following note in the accession ledger: “44 stone implements and other remains. Found by Charles Wilmot in an ancient Indian shell-heap, throughout a stratum of oyster shells from about 1½ to 3 ft. deep, which also contained some pieces of moose antlers” (Province of Nova Scotia 1935:28). The reference to the depth of the eastern oyster stratum is significant not only for being mentioned but also for its magnitude. Describing BjCo-02, John Erskine recognized differences in deposition, but he did not include depth measurements to distinguish these deposits in his reports. He did, however, make the following remark about the different species of shells represented at various depths: “A peculiarity common to both (sub-sites) was that the bottom two inches consisted wholly of oyster and mussel shells, but no artifacts were found in this layer” (Erskine 1986:72). Erskine and others interpreted such changes as signs of cultural shifts. Referring to middens in Nova Scotia with the remains of eastern oyster and quahog (Mercenaria mercenaria), Erskine (1969:3) claimed that they “are a sign of a separate subculture where artifacts are lacking.” A similar view was expressed by Mary Butler in the late 1930s in her discussion of shell midden sites in New York: “a definite difference in culture between the oyster shell-eaters of Westchester County and the musseleaters of northern Dutchess County, with the suggestion of an overlap of the two cultures at Cruger’s Island” (quoted in Brennan 1963:55). Dean Snow (1972) later discussed the interplay between changing sea levels and shifts in pre-contact use of marine resources, observing a transition from primarily quahog, blue mussel, and eastern oyster in lowerlevel strata of sites along Maine’s coast to primarily common clam in the upper strata. He interpreted this cultural transition in historical terms as being driven by changes in the food preferences of pre-contact peoples and developments of new technologies for harvesting (Snow 1972:213). Pretola also explained the species change in terms of access to technology: “First, the easily collected oysters, mussels, and quahogs were exploited. As technological skills increased, less obvious shellfish such as the softshelled clam (Mya arenaria) were gathered” (Pretola 1973:7 quoted in Kirakosian 2014:110). In contrast, other archaeologists, such as Waters (1965) and Braun (1974) have argued that the changes in species were not the result of cultural historical depositional changes over time at a fine scale of resolution. Details not recorded during previous excavations of BjCo-02 were observed. However, what was gained in resolution, was lost in scope. The result is a very detailed view of a very small section of this shell midden. Unit 5C was excavated to 25 cm below surface and offered the most extensive view of the shell midden’s depositional history (see Fig. 3). Definite patterns of shell distribution were visible in levels 3 to 5. A high concentration of shell was observed in the southeastern corner of the unit in levels 4 and 5. The shell species transitioned from a combination of eastern oyster, soft-shell clam, blue mussel, slipper limpet (Crepidula fornicata), and Atlantic surf clam (Spisula solidissima) in level 4 to a predominance of eastern oyster in level 5.9 This shift matches the concentration of eastern oyster observed in test pit 4B, whose southwest corner was contiguous with the northeast corner of 5C (Fig. 2). Like 4B, the eastern oyster shells recovered in 5C were whole.10 The heavy concentration of charcoal and vertebrate faunal remains at ~25 cm below surface in the middle of the west wall and along the south wall of 5C suggests a possible hearth feature. Sterile soil was reached in the middle of 5C and along the east end of its north profile. Figure 4 (top) illustrates the comparative weights of whole shells and chondrophores of the soft-shell clam and whole eastern oyster shells recovered from each of the five arbitrary levels excavated in test unit 5C. Figure 4 (bottom) is a multi-plot graph of the two-sigma age ranges for the radiocarbon dates collected from BjCo-02. All dates were taken from marine shells, were calibrated using OxCal’s IntCal13 and Marine1311 calibration curves, and were corrected for marine reservoir effects by using the ΔR value for the Pugwash Basin, located about 80 km from Maligomish (see Table 1). The dates from 5C cluster around 1500 y.b.p and 500–600 y.b.p.12 These dates also correspond with the peaks in respective representation of eastern oyster and soft-shell clam in unit 5C. The oldest date in 5C was obtained from shell collected in the lowest level (5), where eastern oyster was by far the dominant species. Eastern oyster was still dominant in level 4 (625–500 y.b.p.), but by a much smaller margin. In the younger levels (1–3) soft-shell clam became the dominant species. Thus, the predominance of eastern oyster in BjCo-02 appears to be coincident with a period that preceded a gap in the use of this midden. Discussion The proliferation of shell middens during the Woodland Period, and shifts in mollusk species over Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 52 ingly dominated by soft-shell clam. Finally, at the Turner Farm site, Spiess and Lewis (2001:5–7) documented 3 shell-bearing occupations dating to between approximately 4555 to 2275 y.b.p. with soft-shell clam dominating and no eastern oyster reported. In comparison to these sites in Maine, the Hudson estuary, and regions further south, the midden on Maligomish appears to conform to and, at the same time, challenge previously observed patterns. It conforms to the patterns from the Hudson and the southern sites for the first emergence of eastern oyster in archaeological contexts around 1500 y.b.p. However, BjCo-02 challenges the temporal duration of eastern oyster because the species’ demise appears to have happened later there than at the southern sites. Thus, one of the messages in the Maligomish midden appears to reference a change between 1500 and 500 y.b.p. Determining the nature of that change, however, will require further research. There are limitations on the use of land and freshwater mollusks as climatic and chronological indicators. As early as 1969, J.G. Evans noted that local environmental factors such as changes in moisture and sedimentation rates affect local populations. Consequently, when comparing species types and proportions between archaeological sites, any similarities or differences may be due to local factors and not global processes (Evans 1969). Thus, statements such as Braun’s that the Labrador Current controls “marine temperatures for all northeastern coastal waters north of Cape Cod” (Braun 1974:593) should be taken cautiously by archaeologists who wish to interpret their materials by way of analogy with those from other coastal sites. Variations not only in latitude but also water temperature and salinity, tidal patterns, and the history of glacial retreat combine to make unique local environments. The Northumberland Strait appears to be one such unique local environment, especially in comparison with those areas where other shell midden sites in the Northeast area are found; namely, the Quoddy region of New Brunswick and Maine, the Atlantic Coast of Canada, the Bay of Fundy, and sites further south along the New England coast. In addition, its unique marine geology, which contributes to the Strait’s warm temperatures, and a recent study of the pollen and non-palynological polymorphs from BjCo-02 seems to confirm that the vegetative history of this area is very different than that of sites studied on or near the Atlantic coast (see Mudie and Lelièvre 2013; see also Jetté and Mott 1989). Given these differences, the processes, but instead signaled a change in climate.13 The interest in interpreting changes in shellfish species over time as indicators of climate change has persisted amongst archaeologists and other scientists working in northeastern North America and around the world (see Carbotte et al. 2004, Claassen 1998:126–134). Indeed, Sandweiss and Kelly (2012) demonstrate that archaeological studies of species change and geomorphological processes have not only contributed vastly to archaeology, but also to its sister disciplines. They report that observations of sedimentation at archaeological sites excavated by David Sanger in the Penobscot River Valley in Maine have allowed for the creation of a “generalized sedimentary sequence for the Valley that has not been found in nonanthropogenic deposit” (Sandweiss and Kelly 2012:381). Archaeologists and geologists working in shellbearing contexts in northeastern North America have observed a common pattern regarding the temporal occurrence of eastern oyster and other species. Carbotte et al. (2004) reported on the presence of fossil oyster beds in the Hudson River estuary. They suggest that the fluctuations in oyster representation may be due to warm–cool cycles during the Holocene. Within the Hudson, oysters flourished during the Hypsithermal or mid-Holocene warm period, disappeared with the onset of a cooler climate at 4000–5000 y.b.p., and returned during the warmer conditions of the late Holocene (Carbotte et al. 2004:220). Those authors further suggest that the most recent demise of oysters at 900–500 cal. y.b.p. may have accompanied the Little Ice Age (Carbotte et al. 2004:222). They also compare their geological results to those observed by archaeologists working in shell-bearing contexts along the Atlantic coast, including David Sanger’s (see Sanger and Sanger 1986) work at the Damariscotta River middens where eastern oysters were reported from only 2400–1000 y.b.p. (Carbotte et al. 2004:221). And they refer to Claassen’s (1986) work in the southeastern USA, where eastern oyster only dominates shell-bearing sites after 2500–1500 y.b.p. In Spiess et al.’s (2006:145–147) report on their excavations of the Indiantown Island shell midden on the Gulf of Maine coast, the authors remark that very few eastern oysters were recovered from any of the six cultural units associated with the shell midden that they excavated, which were dated using diagnostic ceramics to between 1650 and 400 y.b.p. At Devil’s Head in Calais, ME, USA, Spiess and Cranmer (2005:46) similarly dated the site to between 2200 and 600 y.b.p. , with a gap of 700 years between 1600 and 900 y.b.p. This site was again overwhelmJournal of the North Atlantic M. Lelièvre 2017 Special Volume 10 53 archaeological sites found along the Northumberland shores may not be readily compared to sites along the Atlantic coast whence the bulk of our coastal data derive. A further factor complicating the potential of Maligomish and other middens to reveal the story of climate change in the Woodland Period is the reliability and comparability of the recovered data. The results from the 2008 excavations of BjCo-02 were limited in scope and included several sampling and processing errors that may bias the results. One of these inaccuracies is the use of a non-local ΔR value to correct for local marine reservoir effects. As Deo et al. (2004:772) explain: “the ‘reservoir effect’ in marine samples arises because the oceans are depleted in 14C compared with the atmosphere and the deficiency is transmitted to marine organisms.” Due to the potential depletion of 14C, marine shell samples tend to date much older than charcoal samples from the same deposit. Proper calibration of dates obtained from marine shells can mitigate this problem, although several researchers have demonstrated the marine reservoir effects vary widely by location, by mollusk species, between individuals of the same species, and over the lifetime of an individual mollusk specimen (see Culleton et al. 2006, Hadden and Cherkinsk 2015, Rick and Henkes 2014). Hadden and Cherkinsky (2015:469) note that “variations in coastal geomorphology, ocean circulation, and upwelling create localized, time-dependent deviations from the global-averaged marine reservoir age.” While upwelling is unlikely to be a significant factor in the relatively shallow waters of the Northumberland Strait and Merigomish Harbour, other factors may contribute to the marine reservoir effect being different from the global average.14 For example, the presence of Carboniferous materials and significant freshwater inputs could cause 14C activity to differ significantly between coastal regions (Hadden and Cherkinsky 2015:469). The bedrock geology of the mainland south of Merigomish Harbour consists of Carboniferous deposits. Additionally, three rivers and several streams empty into the harbor. Although McNeely et al. (2006) have quantified the marine reservoir effect in waters off the Canadian coastlines, the ΔR value closest to Merigomish Harbour is approximately 80km west in Pugwash Basin and was derived from a sample of eastern oyster. This ΔR value may not be an accurate measurement of the marine reservoir effects in Merigomish Harbour and may not be useful for calibrating the radiocarbon dates derived from the soft-shell clam samples from BjCo-02 for several reasons. First, Rick and Henkes (2014) demonstrate that there are variations among the ages of eastern oyster samples from various locations within the Chesapeake Bay, suggesting that variation may also be likely for the eastern oyster found in coastal waters farther north. Second, Hadden and Cherkinsky's (2015) observations of variations in ΔR values between species Figure 5. Comparison of (left) calibrated radiocarbon dates derived from archaeological marine shells collected from levels 2–5 of unit 5C to (right) calibrated radiocarbon dates derived from terrestrial samples from other Northumberland Strait sites (see also Table 1). Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 54 suggest that the ΔR value derived from the eastern oyster at Pugwash Basin may be inaccurate for calibrating soft-shell clam 80 km away in Merigomish Harbour. Finally, Culleton et al. (2006) recommend taking samples for AMS dating from multiple growth increments on marine shell because variations in marine reservoir effects have been observed between seasons and over the lifetime of individual marine shell samples.15 The AMS dates for the Maligomish midden were derived from small subsamples of the individual shells submitted for dating (R. Cruz, NSF-Arizona AMS Laboratory, Tucson, AZ, USA, pers. comm.). Additional research will be required in order to mitigate the potential sources of error in calibrating the radiocarbon ages of the marine shells collected from the Maligomish midden and from other sites along the Northumberland Strait (see Table 1). Ideally, a ΔR value for Merigomish Harbour would be determined using both eastern oyster and soft-shell clam samples. Until such work can be completed, comparison to terrestrial samples collected in Merigomish Harbour and the broader Northumberland Strait may be useful for evaluating the accuracy of the calibrated dates from Maligomish’s marine shells. Figure 5 compares the calibrated dates from Maligomish (left) to those measured on charcoal and moose (Alces alces) collagen (right) from sites along the Northumberland Strait (see also Table 1). The dates from Maligomish cluster around 1500 y.b.p. and 500 y.b.p. A similar pattern is not observed at the sites for which there are multiple radiocarbon dates derived from terrestrial samples—Cox-Swanson (samples S-1602 and S-1603) and Delorey Island (samples GSC-3218 and I-11619)—although Delorey Island returned one date ca. 1500 y.b.p. and the two-sigma range for the oldest date from Cox-Swanson falls within ca. 1500 y.b.p. The remaining dates from the Northumberland Strait, including two from Merigomish Harbour, are either older (BjCo-15) or younger (S-973) than the range observed on Maligomish. Thus, with the limited data available at present, the 1000-year gap observed at Maligomish does not appear to reflect a gap in cultural activity elsewhere in the Northumberland Strait. Possible biases Several sampling and quantification errors have introduced biases to this study. Consequently, the results reported herein should be considered tentative. These biases include the use of a 0.64-cm (1/4-inch) mesh screen, which results in the possibility that invertebrate taxa remains smaller than 0.64 cm (1/4 inch) could have been lost to the back-dirt pile, thus over-representing the soft-shell clam and eastern oyster discussed above. The collection of non-repeating elements, such as the soft-shell clam chondrophores, may have also overrepresented the number of soft-shell clam in the midden. The reported NISP may include, for example, the right valve and chondrophore from the same individual.16 Conversely, the quantities of eastern oyster may be underrepresented in the present study. Only whole eastern oyster shells—not chondrophores—were systematically collected. Many of the eastern oyster shells were extremely friable. The remains of these and other species (e.g., Mytilus edulis [blue mussel]) were so deteriorated in some of the lower levels of the Maligomish test excavations that they had the consistency of silt to coarse sand. These deteriorated shells were not quantified and, therefore, may underrepresent the presence of eastern oyster. A further source of bias may be the work of previous excavators, which include both professional archaeologists and amateur diggers. Some of the strata excavated in 2008 may have been disturbed, while unit 5A showed definite signs of disturbance at ~15 cm below site datum (b.s.d). The results reported above focus on unit 5C, which appeared undisturbed. Finally, the calibrated AMS radiocarbon dates measured on the marine shells may be inaccurate due to the local marine reservoir effects. At present, the only ΔR value used to correct for marine reservoir effects that has been calculated for the Northumberland Strait is from the Pugwash Basin, located approximately 80 km from Merigomish Harbour (Fig. 1). The implications of calibrating an AMS radiocarbon date with a non-local ΔR value are provided in the discussion above. Given the biases in the sampling of the Maligomish midden and in processing its shellfish remains, the current study does not attempt to infer a cause for the apparent shift in species proportions over time.17 The shift may be due to environmental changes, changes in cultural practices, changes in site function over time (see Russo 1988:66), and/or may simply reflect the sampling biases.18 However, the presence of eastern oyster—a rare occurrence for shell midden sites this far north—warrants consideration. The results reported herein may be useful for future shell midden studies in the Northumberland Strait and along other coastal waters of the North Atlantic. Conclusion Several questions emerge from this small sample of the Maligomish midden (BjCo-02). Is the Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 55 Travel Fund from the Division of Social Sciences at the University of Chicago, the Whatcom Museum, and the Mi’kmaq-Nova Scotia-Canada Tripartite Forum. I would also like to acknowledge with deep gratitude the members of the Pictou Landing First Nation who worked with me to develop and execute the archaeological research on Maligomish, including Leonard Cremo, Cheryl Denny, Dominic Denny, former Chief Ann Francis-Muise, Lorraine Francis, Michelle Francis-Denny, Ralph Francis, Sadie Francis, Edie Nicholas, Mary Irene Nicholas, Laura Prosper, Louise Sapier, Martin (Junior) Sapier, and Florence Walsh. Literature Cited Ambrose, W.R. 1967. Archaeology and shell middens. Archaeology and Physical Anthropology in Oceania 2(3):169–187. Andrus, C.F.T., and D.E. Crowe. 2000. Geochemical analysis of Crassostrea virginica as a method to determine season of capture. Journal of Archaeological Science 27:33–42. Betts, M.W., M. Burchell, and B.R. Schöne. 2017. An economic history of the Maitime Woodland Period in Port Joli Harbour, Nova Scotia. Journal of the North Atlantic Special Volume 10:18–41. Black, D.W. 1993. What images return: A Study of the stratigraphy and seasonality of a shell midden in insular Quoddy region, New Brunswick. Manuscripts in Archaeology 27. Archaeological Services, Department of Municipalities, Culture, and Housing, Fredricton, NB, Canada. Borden, C.E., and D. Wilson. 1952. A Uniform Site Designation Scheme for Canada. Anthropology in British Columbia 3:44–48. Braun, D.P. 1974. Explanatory models for the evolution of coastal adaptation in prehistoric eastern New England. American Antiquity 39(4):582–596. Brennan, L.A. 1963. A 6000-year-old midden of Virginia oyster shell at Croton Point, lower mid-Hudson. Pp. 55-59, In The Coastal Archaeology Reader: Selections from the New York State Archaeological Association Bulletin. 1954–1977. Available online at http://nysarchaeology. org/download/nysaa/bulletin/number_29. pdf. Accessed 6 December 2016. Campbell, S. 1981. The Duwamish No. 1 site, A lower Puget Sound shell midden. Office of Public Archaeology, Research Report 1. University of Washington, Seattle, WA. Canadian Museum of History - Library and Archives (CMH-L&A). 1912. Wintemberg, W.J. Field Notes. N.B., N.S. 1912. Manuscript Nos. 196. Vol. 5. p. 28. Ottawa, ON, Canada. Carbotte, S.M., R.E. Bell, W.B.F. Ryan, C. McHugh, A. Slage, F. Nitsche, and J. Rubenstone. 2004. Environmental change and oyster colonization within the Hudson River estuary linked to Holocene climate. Geo-Marine Letters 24:212–224. Ceci, L. 1984. Shell midden deposits as coastal resources. World Archaeology 16(1):62–74. 1000-year gap simply due to a sampling error or is it reflective of an environmental shift and/or changes in cultural practices? What accounts for the apparent persistence of eastern oyster at Maligomish when it is almost completely absent from some coastal sites in Maine and had been phased out from sites further south well before the Late Woodland Period? Claassen (1986) considers several hypotheses to explain the temporal shifts in species, including the die-off of species, overexploitation, innovations in harvesting technology, and environmental change. Of these, she is most persuaded by the environmental hypothesis, citing rising sea levels as the primary cause. Carbotte et al. (2004:220) report that the primary environmental factors that influence oyster growth include salinity, temperature, substrate type, and sedimentation rate. It seems that the way forward for contextualizing the Northumberland Strait among other Atlantic shell midden sites is to first gather more information regarding these factors. All of this work can contribute to developing models for human use and impact on local environments rather than treating the entire coast of the Northeast as one uniform region. Most importantly, the answer to the apparent puzzle presented at Maligomish will be more excavation, better sampling of the midden context (see Claassen 1998:99–104), systematic analysis of the marine shell remains, and studies comparing the radiocarbon ages of archaeological marine shells to marine shells of known age to determine localized marine reservoir effects (see Hadden and Cherinsky 2015, Rick and Henkes 2014, Rick et al. 2012)—objectives made all the more urgent by the effects of contemporary climate change that threaten the survival of these archaeological contexts. Acknowledgments Thank you to the anonymous reviewers for their careful readings of the first draft of this paper and to Philippa Ascough and Keith Goldfarb for their editorial guidance. Thank you to Matthew Betts and Martin (Gabe) Hyrnick for organizing the session on east coast shell middens for the 2014 meeting of the Eastern States Archaeological Federation and for their comments on the revised manuscript. Special thanks also to several participants in that session who helped me at various stages during the fieldwork reported here. These participants include David Black, David Sanger, and Arthur Spiess, who all offered advice on how to excavate shell middens. Matthew Betts was very helpful in granting me access to collections at the Canadian Museum of History. Other archaeologists and museum staff in Nova Scotia offered much support: Scott Buchanan, Tim Bernard, David Christianson, Stephen Davis, Roger Lewis, Heather MacLeod-Leslie, Debra McNabb, Robert Ogilvie, Stephen Powell, and Leah Rosenmeier. The research was funded by the Janco Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 56 Fladmark, K. 1978. A Guide to Basic Archaeological Field Procedures. Department of Archaeology, Simon Fraser University. Publication No. 4. Simon Fraser University, Burnaby, BC, Canada. Glassow, M.A. 2000. Weighing vs. counting shellfish remains: A comment on Mason, Peterson, and Tiffany. American Antiquity 65(2):407–414. Hadden, C., and A. Cherkinsky. 2015. 14C variations in pre-bomb nearshore habitats of the Florida Panhandle, USA. Radiocarbon 57:469–479. Jerardino, A. 1997. Changes in Shellfish Species Composition and Mean Shell Size from a Late-Holocene Record of the West Coast of Southern Africa. Journal of Archaeological Science 24:1031–1044. Jetté, H., and R.J. Mott. 1989. Palynostratigraphie du Tardiglaciaire et de l’Holocène de la région du Lac Chance Harbour, Nouvelle-Écosse. Géographie physique et Quaternaire. 43:27–38. Johnson, M. 2008. Maillard, Pierre. Dictionary of Canadian Biography Online. Available online at http:// www.biographi.ca/EN/ShowBio.asp?BioId=35630. Accessed 2 March 2008. Keenlyside, D.L. 1980. 1978. Pp. 14–16, In R.J.M. Marois (Ed.). Archaeology Survey of Canada Annual Reviews 1977–1979. National Museums of Canada, Ottawa, ON, Canada. Kepple, J.D. (Compiler). 2000. Geological map of the Province of Nova Scotia, Nova Scotia Department of Natural Resources, Minerals and Energy Branch, Halifax, NS, Canada. Map ME 2000-1, scale 1:500,000. Kirakosian, K.V. 2014. Curious monuments of the simplest kind: Shell midden archaeology in Massachusetts. Doctoral Dissertation. University of Massachusetts – Amherst, MA, USA. 504 pp. Kranck, K. 1971. Surficial Geology of Northumberland Strait. Marine Science Paper 5. Geological Survey of Canada Paper 71-53. Marine Sciences Branch: Department of the Environment, Ottawa, ON, Canada. Lelièvre, M.A. 2008. Heritage Research Permit A2008NS02: Report on surface investigations and sub-surface testing of Maligomish (Indian Island), Pictou County, Nova Scotia. Spring 2008. Unpublished manuscript on file with Museum Operations, Nova Scotia Museum, Halifax, NS, Canada. Lelièvre, M.A. 2012. Ajiwisin (you move from one place to another): Mobility, emplacement, and politics in (Post-) Colonial Nova Scotia. Ph.D. Dissertation. University of Chicago, Chicago, IL, USA. Mason, R.D., M.L. Peterson, and J.A. Tiffany. 1998. Weighing vs. counting: Measurement reliability and the California School of midden analysis. American Antiquity 63(2):303–324. McNeely, R., A.S. Dyke, and J.R. Southon. 2006. Canadian marine reservoir ages: Preliminary data assessment. Geological Survey Canada Open File 5049, Ottawa, ON, Canada. Meehan, B. 1982. Shell Bed to Shell Midden. Australian National University, Canberra, Australia. Meighan, C.W. 1972. Midden remains and prehistoric food resources. Archaeological Survey Annual Report 14:3–11. Claassen, C. 1986. Temporal patterns in marine shellfishspecies use along the Atlantic coast in the southeastern United States. Southeastern Archaeology 5:120–137. Claassen, C. 1991. Normative thinking and shell-bearing sites. Pp. 249–298. In M.B. Schiffer (Ed.). Advances in Archaeological Method and Theory. Vol. 3. The University of Arizona Press, Tucson, AZ, USA. Claassen, C. 1998. Shells. Cambridge University Press, Cambridge, UK. Claassen, C. 2000. Quantifying shell: Comments on Mason, Peterson, and Tiffany. American Antiquity 65(2):415–418. Confederacy of Mainland Mi’kmaq. 2007. Kekina’muek: Learning about the Mi’kmaq of Nova Scotia. Eastern Woodland Publishing, Truro, NS, Canada. Culleton, B.J., D.J. Kennett, B.L. Ingram, J.M. Erlandson, and J.R. Southon. 2006. Intrashell radiocarbon variability in Marine mollusks. Radiocarbon. 48:387–400. Culleton, B.J., D.J. Kennett, T.L Jones. 2009. Oxygen isotope seasonality in a temperate estuarine shell midden: A case study from CA-ALA-17 on the San Francisco Bay, California. Journal of Archaeological Science 36:1354–1363. Custer, J.F., and K.R. Doms. 1990. Analysis of microgrowth patterns of the American Oyster (Crassostrea virginica) in the Middle Atlantic Region of Eastern North America: Archaeological application. Journal of Archaeological Science 17:151–160. Davis, D.S., and S. Browne. 1996. The Natural History of Nova Scotia. Volumes I and II. Nimbus Publishing and the Nova Scotia Museum, Halifax, NS, Canada. Deith, M.R. 1983. Seasonality of shell collecting determined by oxygen isotope analysis of marine shells from Asturian sites in Cantabria. Pp. 67–76, In C. Grigson and J. Clutton-Brock (Eds.). Animals and Archaeology. Vol. 2. Shell Middens, Fishes, and Birds, B.A.R. International Series 183, Oxford, UK. Deith, M.R. 1985. Seasonality from shells: An evaluation of two techniques for seasonal dating of marine molluscs. Pp. 119–130, In N.R.J. Fieller, D.D. Gilbertson, and N.G.A. Ralph (Eds.). Palaeobiological Investigations Research Design, Methods, and Data Analysis. B.A.R. International Series 266, Oxford, UK. Deo, J.N., J.O. Stone, and J.K. Stein. 2004. Building confidence in shell: Variations in the marine radiocarbon reservoir correction for the northwest coast over the past 3000 years. American Antiquity 69(4):771–786. Erskine, J. 1961. Micmac Notes, 1960. Nova Scotia Museum, Occasional Paper 11, Archaeological series no. 1. Nova Scotia Museum, Halifax, NS, Canada. Erskine, J. 1969. What can be learned from shell-heaps? Archaeological Society of Nova Scotia, Halifax, NS, Canada. Bulletin. No. 3. Erskine, J. 1986. Unpublished papers on the archaeology of the Maritime provinces. Compiled by M. Deal, Department of Anthropology, St. Mary’s University. Halifax, NS, Canada. Evans, J.G. 1969. Land and freshwater mollusca in archaeology: Chronological aspects. World Archaeology 1(2):170–183. Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 57 Russo, M. 1988. A Comment on Temporal Patterns in Marine Shellfish Use in Florida and Georgia. Southeastern Archaeology 7:61–68. Sandweiss, D.H., and A.R. Kelly. 2012 Archaeological contributions to climate change research: The archaeological record as a paleoclimatic and paleoenvironmental archive. Annual Review of Anthropology 41:371–391. Sanger, D., and M.J. Sanger. 1986. Boom and bust on the river: The story of the Damariscotta oyster shell heaps. Archaeology of Eastern North America 14:65–78. Smith, H. and W.J. Wintemberg. 1929. Some Shell-heaps in Nova Scotia. National Museum of Canada Bulletin, 47. National Museums of Canada, Ottawa, ON, Canada. 193 pp. Snow, D.R. 1972. Rising sea level and prehistoric cultural ecology in northern New England. American Antiquity 37(2):211–221. Snow, H.E. 1994. Archaeological investigations at Kerr Point (BjCo-15) Merigomish Harbour, Nova Scotia. BA Honours Thesis. Department of Sociology and Anthropology, St. Francis Xavier University, Antingonish, NS, Canada Spiess, A.E., and L. Cranmer. 2005. Devils Head, Calais, and Site 97.10: Archaeological Survey for the Land Maine’s Future Board. Maine Historic Preservation Commission, Augusta, ME, USA. Spiess, A.E., and R.A. Lewis. 2001. The Turner Farm Fauna: 5000 Years of Hunting and Fishing in Penobscot Bay, Maine. Occasional Publications in Maine Archaeology. Number Eleven. The Maine State Museum, The Maine Historic Preservation Commission and the Maine Archaeological Society, Augusta, ME, USA. 177 pp. Spiess, A.E., K. Sobolik, D. Crader, J. Mosher and D. Wilson. 2006. Cod, clams, and deer: The food remains from Indiantown Island. Archaeology of Eastern North America 34:141–187. Stein, J.K. (Ed.) 1992. Deciphering a Shell Midden. Academic Press, Inc., San Diego, CA, USA. 375 pp. Trigger, B.G. (Ed.) 1986. Native Shell Mounds of North America: Early Studies. Garland Publishing Inc., New York, NY, USA. Trigger, B.G. 1991. A History of Archaeological Thought. Cambridge University Press, Cambridge, UK. 500 pp. Villagran, X.S., A.L. Balbo, M. Madella, A.Vila, and J. Estevez. 2011. Experimental micromorphology in Tierra del Fuego (Argentina): Building a reference collection for the study of shell middens in cold climates. Journal of Archaeological Science 38(3):588–604. Waters, J.H. 1965. Animal remains from some New England sites. Bulletin of the Archaeological Society of Connecticut 33:5–11. Webb, K.T. 1990. Soils of Pictou County, Nova Scotia. Report No. 18 (with accompanying map sheets: Soils of Pictou County [East and West]). Nova Scotia Soil Survey. Research Branch, Agriculture Canada, Ottawa, ON, Canada. Morlot, A. 1861. General views on archaeology. Pp. 285–343, In Annual Report of the Smithsonian Institution for 1860. Smithsonian Institution, Washington, DC, USA. Mudie, P.J., and M.A. Lelièvre. 2013. Palynological study of a Mi’kmaw shell midden, Northeast Nova Scotia, Canada. Journal of Archaeological Science 40:2161–2175. Patterson, Rev. G. 1877. A History of the County of Pictou, Nova Scotia. Dawson Brothers, Montreal, QC, Canada. Patterson, Rev. G. 1889. The Stone Age in Nova Scotia, as illustrated by a collection of relics presented to Dalhousie College. Transactions of the Nova Scotian Institute of Natural Science 7(3):231–252. Peacock, E. 2000. Assessing bias in archaeological shell assemblages. Journal of Field Archaeology 27(2):183–196. Pretola, J. 1973. The Marshall site: An early contact shell midden from Nantucket Island. Masters Thesis. University of Massachusetts – Amherst, MA, USA. Province of Nova Scotia (H. Piers). 1935. Report on the Provincial Museum and Science Library for the fiscal year 1933–1934. Provincial Secretary, King’s Printer, Halifax, NS, Canada. Quitmyer, I.R. 1985. Aboriginal subsistence activities in the Kings Bay locality. Pp. 73–91, In W.H. Adams (Ed.). Aboriginal Subsistence and Settlement Archaeology of the Kings Bay Locality. Volume 2: Zooarchaeology. Department of Anthropology, University of Florida, Gainsville, FL, USA. Quitmyer, I.R., and D.S. Jones. 1997. The Sclerochronology of hard clams, Mercenaria spp., from the southeastern USA: A method of elucidating the zooarchaeological records of seasonal resource procurement and seasonality in prehistoric shell middens. Journal of Archaeological Science 24:825–840. Quitmyer, I.R., and E.J. Reitz. 2006. Marine trophic levels targeted between AD 300 and 1500 on the Georgia coast, USA. Journal of Archaeological Science 33:806–822. Quitmyer, I.R., D.S. Jones, and C.F.T. Andrus. 2005. Seasonal collection of Coquina Clams (Donax variavilis) During the Archaic and St. Johns Periods in coastal northeast Flordia. Pp. 18–28, In D. Bar-Yosef (Ed.). Archaeomalacology: Molluscs in Former Environments of Human Behavior. Oxbow Press, London, UK. Rand, S. 1902. A Dictionary of the Micmac Language. The Patriot Publishing Company, Charlottetown, PEI, Canada. Reimer, Paula J. et al. 2013. IntCal13 and Marine13 radiocarbon age calibration curves 0–50,000 years cal BP. Radiocarbon 55:1869–1887. Rick, T.C., and G.A. Henkes. 2014. Radiocarbon variability in Crassostrea virginica shells from the Chesapeake Bay, USA. Radiocarbon 56:305–311. Rick, T.C., G.A. Henkes, D.L. Lowrey, S.M. Colman, B.J. Culleton. 2012. Marine radiocarbon reservoir corrections (ΔR) for Chesapeake Bay and the Middle Atlantic Coast of North America. Quaternary Research 77:205–210. Journal of the North Atlantic M. Lelièvre 2017 Special Volume 10 58 Endnotes 1 See also Meehan’s (1982) ethnoarchaeological study of shellfish gathering among the Anbarra of Arnhem Land in northern Australia 2Mi’kmaw is used as a singular noun, as an adjective to modify both singular and plural nouns and when discussing the Mi’kmaw language. Mi’kmaq is a plural noun (see Confederacy of Mainland Mi'kmaq 2007:24). 3In his 1929 report for the National Museum of Man, Smith did not describe having excavated this midden (Shell-heap “D” in the Smith report; see Smith and Wintemberg 1929:6, 7–8). Instead he stated that he had purchased five artifacts likely obtained from the midden from a Mi’kmaw man named Joseph Philip who lived on Maligomish. In 1960, the avocational archaeologist John Erskine reported that he had encountered one of Smith’s companions who had told him that Smith had actually excavated on Maligomish for half a day before the Mi’kmaw caretaker of the island had warned him of f for trespassing on the reserve (Erskine 1961:21). This second-hand report suggests that Smith conducted excavations on Maligomish but did not report them in the official publication. All of the artifacts that Smith and Wintermberg collected from Merigomish Harbour and the surrounding area are now housed at the Canadian Museum of History in Gatineau, QC, Canada. 4During the fall 2008 excavations, we adopted the Parks Canada provenience system (see http://www.pc.gc.ca/eng/ docs/pc/guide/fp-es/4.aspx. Accessed 16 October 2015). 5Despite the best efforts to avoid the ghost of Erskine, he resurfaced in one of the 1 m x 1 m test units: Operation 5A. We abandoned this unit after excavating 15 cm because we were able to observe the outlines of a previous excavation unit, most likely one of Erskine’s. 6These other species included: Mytilus edulis, Spisula solidissima, Mercenaria mercenaria, Crepidula sp., and Littorina sp. 7See Mason et al. 1998, Glassow 2000, and Claassen 2000 for a debate on weighing vs. counting shellfish remains. 8Any future research on the materials collected from the Maligomish midden (BjCo-02) under Nova Scotia Heritage Research Permits A2007NS74, A2008NS02, and A2008NS73 will have to be conducted with permission from the Chief and Council of the Pictou Landing First Nation (PLFN). These collections were temporarily transferred in August 2010 to the Nova Scotia Museum, where they are being held in trust for the PLFN until a suitable storage facility operated by a Mi’kmaw organization becomes available. 9A similar variety of mollusk species was recovered from the Rustico Island shell midden (CcCt-01) off Prince Edward Island. See the unpublished fieldnotes of Birgitta Wallace and her crew from Rustico Island (7F) which are on file with Parks Canada in Dartmouth, NS (Janet Stoddard, Collections Specialist, Parks Canada, Dartmouth, NS, Canada, pers. comm.). 10See Black (1993:60–62) for a discussion of marine shell stratigraphy in shell midden sites from the Quoddy region of New Brunswick. 11Reimer et al. (2013:1871) note that “because Marine13 is based on tropical and subtropical records, its application to 14C ages from samples at higher latitudes must take into consideration additional and possibly large changes in the age of the local surface ocean.” The AMS dates reported here were calibrated using both Marine09 and Marine13. The two-sigma ranges for the samples from 5C2 and 5C3 were exactly the same using each calibration curve. The two-sigma ranges for the samples from 5C4 and 5C5 varied slightly between the calibration using the Marine09 and Marine13 curves. For 5C4, Marine09 = 624–499 y.b.p.; Marine13 = 625–500 y.b.p. A greater variation was observed for 5C5: Marine09 = 1613–1342; Marine 13 = 1608–1344). 12The gap of approximately 1000 years was also observed 15 m to the east of unit 5C in Test Pit 2, which was excavated during the spring 2008 field season (see Table 1). 13My thanks to Katharine Vickers Kirakosian for directing my attention to the debates between Waters, William Ritchie, and John Pretrola, as discussed in her dissertation (Kirakosian 2014). 14Culleton et al. (2006:389) note that “upwelling of deep ocean water is associated with older apparent 14C age and larger ΔR values because of slow mixing that leaves the global marine 14C reservoir depleted relative to the atmosphere.” 15Culleton et al. (2006:398) note that annual to seasonal changes in ΔR on the scale of 100–200 14C yr represent significant sources of random dating error for archaeologists working in aquatic settings. 16One of the anonymous reviewers of the first draft of this article recommended re-analyzing the collection to estimate the minimum number of individuals by sorting the valves and chondrophores side by side, with the more abundant side being interpreted as the MNI. Accessing the collection for this re-analysis will require the permission of the Pictou Landing First Nation. See note 8. 17Claassen (1998:101–104) urges caution when trying to infer environmental or human behavior from poorly sampled remains. She cites Campbell’s (1981:220–223) sampling strategy at Duwamish No. 1 site in Seattle, WA: “No attempt to determine if the differences in shell species composition between proveniences were due to cultural or environmental changes because Campbell could not justify the assumption that the shell data from the four test units were representative of a site-wide temporal trend” (Claassen 1998:103). 18On the ability of archaeologists to infer the aquatic habitats of the mollusk species represented in middens, Stein (1992:9) has argued that such inferences would require “accepting the assumption that the shellfish species found in shell middens are adequate reflections of environmental conditions in adjacent habitats. One must assume that people were selecting shellfish randomly, depositing a random sample of the species inhabiting the bay.”