Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 18 Introduction In contrast with those from other provinces in Canada, Nova Scotian archaeofaunal assemblages from prehistoric sites have received comparatively little professional analysis (Murphy and Black 1996). This neglect is ironic, considering the high density of coastal shell midden sites in the province, which tend to contain well-preserved archaeofaunal remains. As a result, relatively little has been published about coastal economies during the Maritime Woodland Period (ca. 3300–350 cal B.P.) in the province, though pioneering work by Stewart (Nash and Stewart 1986, 1990, Stewart 1989), and later Rojo (1986, 1990) have provided critical insights. Furthermore, few comparisons with more comprehensively studied archaeological deposits in Maine and New Brunswick have been made. This is a significant deficit, as defining the economic and cultural links between these regions is critical to understanding regional culture history. The E’se’get1 Archaeology Project is a community- based research endeavor focused on defining the Maritime Woodland prehistory of Nova Scotia’s South Shore, and in particular the relationship between ancient Mi’kmaq and the coastal ecosystem. An integral part of this work is the development of an economic and subsistence history in Port Joli Harbour (Fig. 1), to be used as a baseline for assessing aspects of culture change, human–animal relationships, and identity amongst the region’s ancient Mi’kmaw inhabitants. In this paper, we use the accumulated record of radiocarbon dates, site location, site structure, and faunal remains from 2 excavated sites, AlDf-24 and AlDf-30, to construct an economic history of Port Joli Harbour. In particular, we focus on defining economic changes which occurred over the Middle Maritime Woodland Period (2150 –1300 cal B.P.) to the Protohistoric era (550–350 cal B.P.), including harvest strategies, foraging efficiency and return rates, transport and butchery, processing, and seasonality. Additionally, we consider the taphonomic history of the assemblages and its potential impact on faunal frequencies. Background Port Joli Harbour, on Nova Scotia’s southern shore, is a long, shallow arm of the Scotian Shelf. Located in an area of Devonian–Carboniferous bedrock, the outer harbor is protected on the north and south by granitoid headlands composed of bedrock, boulders, and cobble beaches, while similar outcrops punctuate portions of its interior coastline. Large granitoid glacial erratics are common throughout this region of the South Shore, and dot its coastline areas. Numerous streams and several small rivers drain into the harbor, and it is consequently rich in sediment and nutrients. The harbor’s coastline is characterized by extensive marshes, eelgrass beds, and foreshore beaches and mud flats. Peatlands and wetlands are common in upland areas away from the coast, although dense Acadian-boreal forest, dominated by red spruce (Picea rubens), white spruce (Picea glauca), balsam fir (Abies balasmea), white pine (Pinus strobus), and sugar maple (Acer saccharum), covers much of the landscape. Softshell clams (Mya arenaria) and other shellfish are abundant on the extensive beach systems An Economic History of the Maritime Woodland Period in Port Joli Harbour, Nova Scotia Matthew W. Betts1,*, Meghan Burchell2, and Bernd R. Schöne3 Abstract - Five seasons of survey and excavation in Port Joli Harbour, NS, Canada, have resulted in a high-resolution archaeofaunal sample from 2 contrasting shell-bearing sites: AlDf-24, and AlDf-30 (Jack’s Brook). In this paper, we discuss the evidence for differences in mollusk-, fish-, and mammal-harvesting strategies between contemporaneously occupied sites. Furthermore, we highlight shifts in Mi’kmaw exploitation of coastal resources around the Middle to Late Maritime Woodland transition (ca. 1300 cal B.P.). Finally, we present insights regarding shellfish-harvesting strategies and site seasonality from isotopic analysis of softshell clam (Mya arenaria) shells. In the process, we construct a history of human– animal relationships in Port Joli, and reveal crucial similarities and important differences with Wabanaki economic strategies in adjacent regions. North American East Coast Shell Midden Research Journal of the North Atlantic 1Canadian Museum of History, 100, rue Laurier Street, Gatineau, QC K1A 0M8, Canada. 2Department of Archaeology, Memorial University of Newfoundland, St. John’s, NL A1C 5S7, Canada. 3Institute of Geosciences, Institute of Geosciences, Applied and Analytical Paleontology, University of Mainz, Johann-Joachim-Becher-Weg 21, 55128 Mainz, GERMANY. *Corresponding author - Matthew.Betts@museedelhistoire.ca. 2017 Special Volume 10:18–41 Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 19 in the harbor, and migratory waterfowl, especially Canada geese (Branta canadensis), overwinter here in abundance. Grey seals (Halichoerus grypus) are common on the rocky outcrops and cobble beaches at the head of the harbor, especially in spring and fall. Atlantic eels (Anguilla rostrate) and alewife (Alosa pseudoharengus) run in the small rivers and lakes throughout the region. Recent research by Neil et al. (2014:210) has documented the effect of sea-level rise on the development of Port Joli Harbour. Around 3000 cal B.P., the harbor was entirely above water; the outer portions only became inundated ca. 2000 cal B.P. Extensive foreshore flats, suitable for clam beds, developed ca. 1500 cal B.P., and the present coastline was not established until ca. 500 cal B.P. As we will describe later in the report, the development of this coastline had a substantial impact on settlement and economy in Port Joli. The archaeological potential of Port Joli was first noted by Thomas Head Raddall (n.d.), an author, historian, and artifact collector, who conducted some exploratory excavations in local shell middens in the early and middle 20th century. John Erskine, a naturalist working for the Nova Scotia Museum, tested or completely excavated more than a dozen sites on the east and west shores of Port Joli Harbour in the late 1950s and early 1960s (Erskine 1962, 1986). His excavations, while lacking in methodological rigor by professional standards, resulted in a relatively comprehensive collection of artifacts and faunal remains which suggested an intensive occupation spanning the Maritime Woodland Period, ca. 3300–350 cal B.P. In 1967, Eric Millard documented previously unrecorded archaeological sites on the eastern side of the harbor, one of which, AlDf-1, he excavated over the following 3 summers. This was to be the last archaeology conducted in the region for nearly 30 years. In the early and middle 1990s, Stephen Powell (1990, 1995) revisited the area to re-locate the sites Erskine discovered and to conduct further reconnaissance in preparation for the development of Thomas Raddall Provincial Park. Powell discovered several large shell middens and was the first to describe that many of the coastal sites in the region had been extensively disturbed by Erskine and other collectors. Given the density of shell middens in the harbor (Betts 2009, 2010), Port Joli is believed to contain one of the most comprehensive Maritime Woodland coastal archaeological sequences in southwestern Nova Scotia, if not the entire province. In 2008 and 2009, researchers from the Canadian Museum of History returned to Port Joli to conduct walking surveys and text excavations, primarily focusing on publically and privately owned lands on the southern side of Port Joli Harbour (Betts 2009, 2010). The surveys resulted in the discovery of 2 new shell midden sites, and the mapping and accurate geolocation of shell midden sites discovered by Raddall, Erskine, and Powell. In total, 21 shell middens have been documented in Port Joli (Fig. 1), representing the greatest density of such features anywhere in the province of Nova Scotia (MARI online). Other shell middens are expected to exist, particularly on the northern shore of the harbor, where there has been more extensive private development and comparatively less archaeological investigation. In 2010 and 2012, extensive excavations were carried out on undisturbed deposits, with an emphasis on excavating shell midden deposits and dwelling features (Betts 2011, Betts and Hrynick 2013). In total, 8 sites were excavated or tested between 2008 and 2012, and previous excavations by Erskine and Millard have provided varying degrees of information on an additional 9 sites, providing a relatively comprehensive overview of the deposits in the harbor. Based on the data recovered, Port Joli shellbearing sites can be classified into 4 primary types (Fig. 1). The first are large shell midden mounds, located within 30 m of the high-water mark, and 300 m2 or greater in area. They have limited artifact densities (especially lithics), larger portions of unbroken clam shells, limited soil development, and large amounts of charcoal. These middens can have depths up to 1 m and universally have large shell-bearing “black soil” (Black 2002) deposits located at the landward edge of the sites. These sites are all located at mid-harbor, opposite the most extensive clam flats (Fig. 1). The second type are large shell-bearing black-soil deposits, greater than 300 m2, with a high proportion of organic matrix and limited proportions of shell. These sites are often characterized by high quantities of lithic debitage and charcoal and highly fragmented faunal remains. They can occur adjacent to the coastline or as much as 350 m inland, but are all located close to the head of the harbor, and are always adjacent to streams or rivers (or relict stream beds). The third type consists of small shell-bearing black-soil deposits with small associated shell middens located close to the coast. These are often less than ~150 m2 and less than 50 cm in depth, and are situated within 30 m of the shoreline. The deposits are associated with high quantities of lithic debitage and comminuted charcoal. The fourth and most unique type of site is similar to the third type of site, except that these sites are located between 140–400 Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 20 m inland from the coast. These “near-interior” sites are all located in forested locations at higher elevations, close to small freshwater streams. They are characterized by smaller shell-middens containing broken and unbroken shells, typically less than 50 cm in depth. The middens occur in multiple distinct loci near the edges of the sites, and are often associated with extensive shell-bearing black soil deposits adjacent to the shell middens. AlDf-24 AlDf-24 (Fig. 2) is a very large shell-bearing and shell midden site located within 30 m of the beach, at approximately the midpoint of Port Joli Harbour’s southern shoreline. The dominating feature of AlDf-24 is a large shell-mound, Area A, covering nearly 300 m2. To the south of Area A, away from the beach, several other activity areas were located via a subsurface soil probe (Betts 2009, 2010). Area B is an extensive, shallow, shell-bearing deposit with significant quantities of cultural black soil with abundant finely crushed shell. Area C, located south of a boulder ridge on a small terrace, contains a deep, shell-bearing black-soil deposit as well as a small shell-midden on its eastern border. Further south, and on the highest terrace on the site, Area D appears similar to deposits in Area C (it has not been tested). The Area A shell midden has a large flat surface, tapering abruptly on its margins to the forest floor, and a large glacial erratic forms its northwest margin. Over 1 m deep in sections, the western half of the midden had been substantially impacted by previous collecting activities, but the eastern portion remained undisturbed. The 2010 excavations focused on excavating a transect through this undisturbed portion to assess overall variability in the midden (Fig. 2). The midden did not develop on a uniform surface, but instead was deposited in an area strewn with numerous large boulders, some greater than 3 m3 in volume. As the midden accumulated, it filled in the spaces between the boulders, eventually creating a large flattopped mound. Figure 1. Map of Port Joli Harbour showing location of AlDf-24 and AlDf-30, including distribution of shell-bearing site types. Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 21 Much of the midden is composed of large unbroken clam shells so numerous that in some portions there was very little soil matrix between the shells. Regardless, the midden is clearly stratified, exhibiting a continuous profile across more than 9 m on the north–south axis. Alternating strata manifested primarily as shifts in (a) the degree of shell fragmentation and the presence of unbroken shells, (b) the portion of soil matrix, (c) the presence of trace amounts of mollusc shell other than softshell clam, and (d) the density of artifacts, particularly ceramics. In no instances other than the immediate pre-subsoil layer did any strata manifest as a shell-bearing black-soil layer. Artifacts, especially lithics, were very rare in the midden, although pottery was relatively abundant. While fragmented, animal bone recovered from the midden tended to be less fragmented than in other deposits in Port Joli Harbour. The only features associated with the midden— a dwelling surface, labelled Feature 2, and a human burial (see Betts et al. 2012)—occurred immediately below the sod surface, and not within the midden deposit (both features are not considered further in this paper). The lack of features within the midden proper suggests that it accumulated as a specialpurpose area of activity— something that precluded other activities, particularly dwelling-related activities, from occurring on the mid- Figure 2. Plan map of AlDf-24. den as it developed. Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 22 AlDf-24 Area C (Fig. 2) was a much more complex deposit than the large midden in Area A. The uppermost stratum (Level 2) appears to be a black-soil shell-bearing activity or living surface of relatively recent age. Below this, in Levels 3 and 3b, we encountered a buried dwelling feature (Feature 4), which appeared to be the remains of a dismantled wigwam-style structure, approximately 3 m long and 2.75 m wide (see Hrynick et al. 2012 for a full description). The floor deposit manifested as a layer of compact black soil, containing abundant lithics, comminuted charcoal, and highly fragmented shell. Despite the fragmentation of the shell, preservation of animal bones in the black-soil feature was excellent, and was comparable to the adjacent midden (Betts 2011). Excavation suggests the feature was a shallow depression no more than 15 cm deeper than the surrounding surface when it was occupied. Relevant to the present paper, a large angular rock, interpreted as an anvil stone for breaking ungulate bones, was documented in the southwest corner of the dwelling. Numerous spirally cracked long bone fragments were found adjacent to the stone. Below the Feature 4 floor deposit were alternating lenses of dark black compacted soil with variable amounts of shell which extended to a maximum of ~55 cm below the surface. The deposits appear to be a complex palimpsest of black-soil shell-bearing layers, shell middens, and dwelling-floor surfaces. To the east of these deposits, a relatively deep and complexly stratified midden was deposited to a depth of ~65 cm below the surface. It appears from the stratigraphic profile that Feature 4 and earlier house floors developed on top of the midden surfaces and were occupied as the midden developed around them. A complex articulation occurred between the house feature and midden strata; because the house floors developed on top of and beside the midden, the floor layers often appeared to gradually transition to higher densities of the more coarsely broken shell that characterizes the midden. For this reason, we were able to associate strata in the midden with strata in the floor deposits. Though small, the midden contained a much denser proportion of artifacts and faunal remains than was recovered from the Area A midden. While unbroken and coarsely broken shell did occur in the Area C midden, their density was less than in the Area A midden. AlDf-30 AlDf-30 (Jack’s Brook) is a small multi-component shell midden site located in a densely wooded forest ~250 m inland from a beach just north of Scotch Point (Fig. 3). The site sits on a small, sparsely treed, knoll that is surrounded on the west, south, and east by a fen (Neil et al. 2014), and on the north by a small stream known as Jack’s Brook (Raddall n.d.). Archaeological deposits at the site are composed of at least 3 discrete shell middens, labelled Areas A, B, and C, as well as a small centrally located black-soil deposit (Fig. 3). Excavations at the site in 2008, 2009, 2010, and 2012 focused on Midden A and the black-soil area adjacent to it. Area A is a well-preserved shell midden, with abundant complete and coarsely broken softshell clam valves. The midden is relatively uniform in depth (~15–20 cm thick), but increases slightly in thickness (~30 cm) as it moves north towards the apex of the midden. As with AlDf-24 Area A, lithics were infrequent in the midden, though ceramics and animal bones were abundant. At its southeastern margin, the midden abruptly ends in a ~10 cm “step”, transitioning to a very thin (~1 cm) scatter of shells. Excavation of the black-soil area in 2012 (see detailed description in Hrynick and Betts 2014, 2017 [this volume]) revealed a sequence of undisturbed dwelling floors and a possible sweat lodge feature. The lower house floor and the sweat lodge feature are likely to be chronologically associated with the Area A midden. Unlike AlDf-24 Area C, this area of the site contained little fragmented-shell, and uncalcined animal remains were not preserved in this area of the site. The largely unidentifiable and highly fragmented assemblage of calcined bones from this area is not considered in the present paper. Chronology Charcoal and terrestrial mammal bone samples were carefully collected for radiocarbon dating from all strata/features throughout the 2008–2010 excavations in Port Joli. As described in Table 1, eight AMS radiocarbon assays have been conducted on bone and charcoal samples from the contexts described in this paper. The resulting dates span a period ca. 1630 ± 40 to ca. 380 ± 40 radiocarbon years B.P. (normalized). When calibrated, the 5 dates associated with AlDf-24 Area A and AlDf-30 Area A cluster around ca. 1550–1300 cal B.P., suggesting occupation of these contexts during the end of the Middle Maritime Woodland Period (Black 2002:304), or Ceramic Period 3 (Petersen and Sanger 1993). All dates overlap substantially at 2σ, suggesting that the sites were occupied contemporaneously. Perhaps most intriguing is the range of dates returned from the AlDf-24 Area A midden. These samples were collected from strata spanning the lowest level (Layer 3n) to the surface (Layer 3b) and clearly indicate the near Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 23 in basal cultural strata, while the latest date, 380 ± 40 B.P. (510–310 cal B.P.), is derived from caribou bone recovered from the cultural layer immediately below the sod. A radiocarbon date on a caribou bone found in the Feature 4 hearth (Feature 4a) dates the dwelling floor to ca. 660 ± 40 B.P. (680–550 cal B.P.), or the Later Late Maritime Woodland Period (Black 2002). It should be noted that the apparent hiatus in the Area C deposits likely represents a gap contemporaneous nature of the deposit, suggesting a rapid accumulation of over 1 m of shell sometime in the decades around 1450 cal B.P. Unlike the deposits at AlDf-24 Area A, the Area C deposits span a much larger time range—from the Early Late Maritime Woodland Period to the early Protohistoric Period (following Black’s [2002] periodization). The earliest date, 1260 ± 40 B.P. (1290– 1070 cal B.P.), is derived from caribou bone found Figure 3. Plan map of AlDf-30. Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 24 each natural or arbitrary level from each unit. Within features, especially potential dwelling floors, excavation proceeded by quadrant (with all recovered materials screened and bagged by quadrant). We acknowledge that the different sampling protocols used in middens versus features could potentially hamper the comparison of taxonomic frequencies between deposits, especially for fish, birds, and small mammals. To assess this possibility, we screened three ~5-L bulk midden samples from AlDf-24 Area C through 6-mm and then 3-mm mesh, collected the faunal remains, and identified them using standard zooarchaeological techniques. There was no difference in the number of identified specimens (NISP) between the 6-mm- and 3-mmscreened samples, and there was no difference in the unidentifiable fraction of birds and mammals between the samples. However, 5% of the unidentifiable fish remains were not captured by the 6-mm mesh. This suggests that, overall, contexts from Port Joli screened through 3-mm and 6-mm mesh should be comparable, but caution is warranted when unidentified fish fractions (i.e. unidentified class counts) are compared. Further pertinent to this discussion is the nature of the small vertebrate fraction recovered from the bulk-sample analysis described later in this report. While the bulk-sample analyses (see below) did not include taxonomic identification, visual inspection of the remains reveals that there were no unusually small vertebrate taxa in the bulk samples, including small fish and bird taxa. This finding suggests 2 things: (a) our screening protocols are generally adequate for the average size of taxa harvested in Port Joli, and more importantly, (b) the bulk samples did not reveal any new taxa not already identified in the screened faunal assemblages. Analysis and Results The assemblages discussed in this paper were identified with the aid of physical reference collections housed at the Howard Savage in sampling (samples from intervening layers have not yet been assayed) and not in occupation. We amalgamated the faunal samples based on stratigraphic and radiocarbon date associations to form a sequence of 5 contexts spanning the Middle Woodland to Protohistoric periods. AlDf-24 Area A is considered to represent 1 large assemblage dating to ca. 1530–1300 cal B.P. Dates from AlDf-30 Area A (1520–1330 cal B.P.) overlap substantially with AlDf-24 Area A, and hence the 2 deposits are considered to be roughly contemporaneous. AlDf-24 Area C Level 2 is considered to be 1 assemblage, dating to ca. 510–310 cal B.P., the Protohistoric era of Port Joli. AlDf-24 Area C Levels 3 and 3b represent the entire Feature 4 household and associated midden assemblage, dated to ca. 680–550 cal B.P., the Later Late Maritime Woodland. Finally, AlDf-24 Area C Levels 3f–3k are considered contemporaneous for the purposes of this report, and represent the earlier Late Maritime Woodland Period, dating as early as 1300 cal B.P. Sampling Strategy The E’se’get Project employed a standardized excavation strategy applied to all sites and units. The primary unit of excavation was the 100 cm x 100 cm square; excavation followed natural stratigraphy, with arbitrary 5-cm layers assigned in deep strata for additional vertical control. All deposits were excavated by trowel, and undifferentiated midden deposits were screened through 6-mm (1/4-inch) mesh, while features and “kitchen middens” associated with dwelling features were screened through 3-mm (1/8-inch) mesh to enhance recovery. Water screening was not employed, but bulk midden samples were recovered from each stratigraphic layer in each unit to further enhance recovery in in midden deposit. These consisted of 3–5-L samples of unscreened matrix, typically taken from the center of each unit at the start of each level. Faunal remains, undecorated pottery, and lithics were collected in level bags representing Table 1. Radiocarbon dates for Port Joli samples. 2σ calibration CMC # 13C/12C Conventional B.C./A.D. B.C./A.D. Borden # (lab #) BETA # 0/00 age older younger Older Younger Area, feature, and level Service AlDf-30 1746 255231 -22.9 1630 ± 40 340 540 1610 1410 Area A, Level 3b AMS AlDf-24 A 1745 255230 -23.8 1540 ± 40 420 610 1530 1340 Area A, Level 3g AMS AlDf-24 A 1749 273513 -23.9 1470 ± 40 540 650 1410 1300 Area A, Level 3m AMS AlDf-24 A 1748 256564 -23.5 1430 ± 40 560 660 1390 1290 Area A, Level 3b AMS AlDf-24 A 1769 297241 -23.8 1420 ± 30 590 660 1360 1290 Area A, Level 3N/0 AMS AlDf-24 C 1758 288733 -20.8 1260 ± 40 660 880 1290 1070 Area C, Level 3i AMS AlDf-24 C 1756 286106 -21.7 660 ± 40 1270 1400 680 550 Area C, Feature 4a (hearth) AMS AlDf-24 C 1757 288732 -22.6 380 ± 40 1440 1640 510 310 Area C, Level 2 AMS Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 25 Osteoarchaeological Laboratory at the University of Toronto, with additional Atlantic Canadian reference materials housed at the Canadian Museum of History and the Royal Ontario Museum. All faunal remains from AlDf-24 Area A and C, as well as AlDf-30 Area A, were included in this analysis. The vertebrate remains from the AlDf-30 house features and the small sample from the AlDf-30 Area B midden are not considered in this report. As a methodological note, we acknowledge difficulty in the identification of cervid remains in the assemblage, despite having access to comprehensive reference collections and other reference materials. In particular, it was sometimes difficult to determine if the taxon of some elements was woodland caribou (Rangifer tarandus caribou), or white-tailed deer (Odocoileus virginianus). In some instances, potential caribou remains often seemed to be very gracile in comparison to contemporary woodland caribou reference specimens. This likely reflects the fact that the caribou remains came from a now extinct/extirpated Maritime Peninsula herd. In these cases, a “cervidae” identification (Table 2) was given to the problematic element, which in this paper refers to woodland caribou or white-tailed deer (no potential moose [Alces alces] bones occur in this taxonomic category). Bulk sample analysis Our analysis begins with a consideration of the shell midden matrix from these sites. Bulk midden samples selected for this analysis come from the deepest part of the midden at AlDf-24 Area A (unit N56W50, Levels 3–3m), and the shell midden deposits to the east of AlDf-24 Area C (units N51W51, N52W51, Levels 2–3j). The AlDf-30 bulk samples were selected from the Area A midden, unit N52W55, Levels 3–3E. For the purposes of this paper, we aggregated data from all stratigraphic levels (surface to basal levels), except in AlDf-24 Area C, where we aggregated levels based on chronological and stratigraphic indicators described above. Bulk samples were sorted using graduated sieves followed by hand sorting under magnification. Artifacts were counted and weighed, while faunal remains, soil fraction, non-cultural lithics, and mollusc taxa were weighed to the nearest milligram. Figure 4 describes the proportion of marine shell to combined soil and non-cultural lithic fraction. AlDf-24 Area A clearly stands out, exhibiting a very large proportion of shell to soil/lithics. The contemporaneous Middle Maritime Woodland AlDf- 30 and the Early Late Maritime Woodland contexts are highly similar but have nearly 20% less shell by weight than in AlDf-24 Area A. Shell volume increases in the later portions of the AlDf-24 Area C midden. There is a general increase in the density of artifacts per kilogram in the bulk samples over time (Fig. 5), with AlDf-24 Area A exhibiting far fewer artifacts than contemporaneous and later deposits. Scrutiny of the charcoal fraction (Fig. 6) reveals Figure 4. Percent total weight (kg) of marine shell compared to percent total combined weight of soil and non-cultural lithics in the bulk midden samples. Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 26 Table 2. Number of identified specimens (NISP) and percent, minimum number of individuals (MNI), richness, and unidentified counts for taxa in Port Joli assemblages. AlDf-24 Area C AlDf-24 Area A AlDf-30 Area A Early Late Proto Taxon Common name NISP % MNI NISP % MNI NISP % MNI NISP % MNI NISP % MNI Actinopterygii Squalas acanthias Spiny dogfish 19 0.71 1 78 3.80 1 0 0.00 0 0 0.00 0 0 0.00 0 Anguilla rostrata American eel 692 25.85 6 0 0.00 0 0 0.00 0 0 0.00 0 0 0.00 0 Gadus morhua Atlantic cod 57 2.13 12 75 3.65 1 21 11.41 1 2 0.57 1 0 0.00 0 Migrogadus tomcod Tomcod 37 1.38 2 0 0.00 0 0 0.00 0 0 0.00 0 0 0.00 0 Gadidae Cod/pollock 0 0.00 0 4 0.19 1 0 0.00 0 0 0.00 0 0 0.00 0 Gadidae True cods 593 22.15 11 1089 52.99 8 40 21.74 1 3 0.85 1 0 0.00 0 Total identified Actinopterygii 1398 1246 61 5 0 Unidentified Actinopterygii 1824 850 227 25 0 Total Actinopterygii 3222 2096 288 30 0 Aves Gavia immer Common loon 28 1.05 2 1 0.05 1 0 0.00 0 2 0.57 1 0 0.00 0 Branta canadensis Canada goose 443 16.55 12 26 1.27 3 4 2.17 1 4 1.14 1 0 0.00 0 Anserinae Swan/goose 0 0.00 0 5 0.24 1 0 0.00 0 0 0.00 0 0 0.00 0 Anserini Goose 0 0.00 0 2 0.10 1 0 0.00 0 0 0.00 0 0 0.00 0 Anas rubripes American black duck 60 2.24 6 49 2.38 7 0 0.00 0 1 0.28 1 0 0.00 0 Anatinae Dabbling duck 55 2.05 3 75 3.65 4 0 0.00 0 0 0.00 0 0 0.00 0 Aythaya collaris Ring-necked duck 0 0.00 0 4 0.19 1 0 0.00 0 0 0.00 0 0 0.00 0 Somateria mollissima Common eider 223 8.33 9 29 1.41 6 0 0.00 0 0 0.00 0 0 0.00 0 Anitidae Black duck/eider 0 0.00 0 57 2.77 5 0 0.00 0 0 0.00 0 0 0.00 0 Mergus sp. Merganser 13 0.49 2 9 0.44 3 0 0.00 0 0 0.00 0 0 0.00 0 Anatidae Duck/goose/swan 150 5.60 7 9 0.44 4 3 1.63 1 2 0.57 1 1 0.54 1 Larus smithsonianus American herring gull 3 0.11 1 0 0.00 0 0 0.00 0 0 0.00 0 0 0.00 0 Laridae Gull 0 0.00 0 6 0.29 1 0 0.00 0 0 0.00 0 0 0.00 0 Total identifiedAves 975 272 7 9 1 Unidentified Aves 4966 338 165 650 23 Total Aves 5941 610 172 659 24 Mammalia Lepus americanus Snowshoe hare 136 5.08 7 442 21.51 17 8 4.35 1 20 5.68 1 4 2.17 1 Castor canadensis North Amer. beaver 6 0.22 1 4 0.19 3 1 0.54 1 7 1.99 2 1 0.54 1 Erethizon dorsatum Porcupine 3 0.11 1 0 0.00 0 2 1.09 1 0 0.00 0 0 0.00 0 Muridae Mouse/shrew 4 0.15 1 0 0.00 0 0 0.00 0 0 0.00 0 0 0.00 0 Tamias striatus Eastern chipmunk 1 0.04 1 0 0.00 0 3 1.63 1 0 0.00 0 0 0.00 0 Rodentia Rodent 1 0.04 1 0 0.00 0 0 0.00 0 0 0.00 0 0 0.00 0 Tamiasciurus American red squirrel 1 0.04 1 1 0.05 1 0 0.00 0 0 0.00 0 0 0.00 0 hudsonicus Scurius sp. Squirrel 0 0.00 0 0 0.00 0 0 0.00 0 0 0.00 0 0 0.00 0 Canis lupus Eastern wolf 7 0.26 1 0 0.00 0 1 0.54 1 0 0.00 0 0 0.00 0 Canis lupus familiaris Domestic dog 2 0.07 1 18 0.88 1 1 0.54 1 0 0.00 0 0 0.00 0 Canis sp. Coyote/wolf/dog 0 0.00 0 14 0.68 1 0 0.00 0 0 0.00 0 0 0.00 0 Ursus americanus American black bear 2 0.07 1 3 0.15 1 0 0.00 0 0 0.00 0 0 0.00 0 Mustelinae Mink/marten 2 0.07 1 0 0.00 0 0 0.00 0 0 0.00 0 0 0.00 0 Lontra canadensis River otter 1 0.04 1 0 0.00 0 0 0.00 0 0 0.00 0 0 0.00 0 Lynx canadensis Lynx 6 0.22 1 0 0.00 0 1 0.54 1 1 0.28 1 0 0.00 0 Carnivora Carnivore 7 0.26 1 0 0.00 0 1 0.54 1 2 0.57 1 0 0.00 0 Halichoerus grypus Grey seal 6 0.22 1 4 0.19 1 0 0.00 0 0 0.00 0 0 0.00 0 Pagophilus Harp seal 3 0.11 1 0 0.00 0 0 0.00 0 0 0.00 0 0 0.00 0 groenlandicus Phocidae Small seal 21 0.78 2 0 0.00 0 0 0.00 0 0 0.00 0 0 0.00 0 Alces alces Moose 17 0.64 1 15 0.73 1 4 2.17 1 3 0.85 1 3 1.63 1 Rangifer tarandus Woodland caribou 52 1.94 2 36 1.75 1 1 0.54 1 1 0.28 1 0 0.00 0 caribou Odocoileus virginianus White-tailed deer 12 0.45 1 0 0.00 0 22 11.96 2 98 27.84 3 34 18.48 2 Cervidae Caribou/deer 14 0.52 1 0 0.00 0 71 38.59 3 206 58.52 4 141 76.63 5 Total identified Mammalia 304 537 116 338 183 Unidentified Mammalia 1546 515 2153 5507 1729 Total Mammalia 1850 1052 2269 5845 1912 Spirally fragmented (Mammalia) 41 0 662 1376 289 Unidentified Indeterminate 3185 2281 628 1935 0 Richness (all taxa) 26 16 11 10 5 Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 27 is some variation in the presence and absence of these species, which may be instructive. For example, urchin is only present in any quantity in the AlDf-24 Area A deposits (Fig. 7). Interestingly, our field notes indicate there is a stratigraphic pattern to their presence, occurring as a thin layer roughly every 10–15 cm throughout the depth of the midden. Mussel shell appears in greatest quantity in AlDf-24 Area C, and in extremely low quantities elsewhere. Again, in the large AlDf-24 Area A midden, mussel remains appeared to be stratigraphically correlated, appearing as a trace in deposits every 10–15 cm. that while there is a large proportion of charcoal in AlDf-24 Area A compared to the contemporaneous AlDf-30, there is a general decrease in the proportion of charcoal over time in these middens. The bulk-sample analysis reveals that mollusc richness and diversity in Port Joli middens is extremely low, with only 4 species represented, including, softshell clam, blue mussel (Mytilus edulis), green sea urchin (Lytechinus variegatus), and northern whelk (Buccinum undatum). By weight, softshell clams dominate deposits in all periods and times, and never comprise less than 99.88% of the shell fraction by weight. That said, there Figure 5. Number of artifacts per kilogram in the bulk midden samples. Figure 6. Percent total weight (kg) of charcoal in the bulk midden samples. Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 28 Class analysis (Fig. 9) of the identified animal remains provides a general indication of economic practices at these sites. At both of the Middle Maritime Woodland contexts, fishing was a crucial activity, and fish remains account for greater than 50% of the identified specimens. AlDf-24 Area A, however, shows an increased emphasis on birds, while AlDf-30 displays a greater emphasis on mammals. The AlDf- 24 Area C contexts are very different, indicating a greater than 60% proportion (NISP) of mammals, which increases over time. Fish remains are generally abundant in the early portion of the assemblage, but decline through both the Late and Protohistoric portions of the deposit. In all Late Maritime Woodland contexts, birds are notably infrequent. A more detailed look at the identified fraction provides some critical insights about these differences (Fig. 10, Table 2). In AlDf-24 Area A, fish, especially gadids and American eel, dominate the assemblage. Canada goose and common eider are also present in significant quantities, and snowshoe hare (Lepus americanus) account for more than 5% of the assemblage. Woodland caribou, moose, and whitetailed deer are also present in the assemblage but in comparatively low frequencies. However, moose and caribou remains are relatively more numerous in this context than white-tailed deer. The AlDf-30 assemblage exhibits a significantly different faunal profile. It is clearly dominated by gadids, and more than 20% of the assemblage is composed of snowshoe hare, with lesser inputs from Canada geese and ducks. Caribou and moose are present Faunal remains generally accounted for less than 1% of the bulk samples by weight. However, some differences were visible; vertebrate remains occur in relatively equal proportions in the AlDf-24 Area A deposits and the early and late deposits in AlDf-24 Area C. AlDf-30 bulk samples suggest an increase in the proportion of vertebrates, while the Protohistoric sample from AlDf-24 Area C suggests a substantial relative increase in vertebrate remains. Vertebrate Zooarchaeology Table 2 describes the vertebrate faunal assemblages from AlDf-24 and AlDf-30. As described in Table 2, richness, here measured as the number of non-overlapping taxa, varies considerably between sites. The contemporaneous assemblages from AlDf- 24 Area A and AlDf-30 exhibit large differences in richness, primarily accounted for by the decrease in fish and mammals, notably seals and furbearers, in the latter context. Richness declines further in the AlDf-24 Area C contexts, this time accounted for by a decrease in bird and fish taxa. An examination of the effect of sample size on richness values (Fig. 8) indicates a strong positive correlation. This effect can be overcome somewhat by amalgamating the AlDf-24 Area C contexts. Once amalgamated, overall richness equals 14, which compares well with AlDf-30. However, there appears to be an unmistakeable decline in avian taxa, particularly associated with ducks, in these Late Maritime Woodland contexts (Table 2). Figure 7. Percent total weight (kg) of unique mollusc taxa (calculated as a percentage of the total marine shell fraction) in the bulk midden samples. Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 29 in relatively low quantities, and importantly, no white-tailed deer was identified in this assemblage. The Late Woodland deposits at AlDf-24 Area C exhibit radically different faunal signatures from the earlier deposits. The assemblages are dominated by ungulates, and cervids increase in proportion through time. While identification was sometimes difficult, it appears that white-tailed deer may also increase in proportion through time in these contexts. This pattern stands in contrast to that associated with the Middle Maritime Woodland contexts, which exhibited less identified white-tailed deer (in fact, none were identified at AlDf-30). Gadids, likely Atlantic cod (Gadus morhua), are frequent only in the early portion of the AlDf-24 Area C contexts, and birds and furbearers are underrepresented in all 3 assemblages. Notably, no seal remains were recovered from any of the Late Maritime Woodland contexts. The differences in cervid representation between the contexts merit further consideration (Table 2). Figure 8. Vertebrate richness plotted against sample size (NISP). Figure 9. Percent NISP of birds, mammals, and fish in the Port Joli assemblages. Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 30 In both the Middle Maritime Woodland contexts (AlDf-24 Area A and AlDf-30), moose and caribou are much more abundant than white-tailed deer (the latter context contained no deer). In the Late Maritime Woodland contexts, moose and caribou decline significantly while deer increases to dominate the cervid assemblage. The replacement of moose/ caribou by white-tailed deer is concomitant with a shift towards terrestrial ungulates in general, and therefore appears to represents a critical shift in economies at Port Joli. Cervid element distributions High frequencies of woodland caribou and whitetailed deer remains permit the opportunity to analyze ungulate element distributions. We aggregated all cervid remains except for moose when calculating element distributions in each assemblage; this approach was necessary to increase sample size, though we recognize that transport may have differed between woodland caribou and white-tailed deer due to unique hunting strategies. As described in Table 3, we calculated minimum number of element (MNE), minimum animal units (MAU), and % MAU values for cervid remains recovered from each context considered in this report. To assess transport, processing, and butchery, we compared these against the food utility index (FUI; Metcalfe and Jones 1988) and the meat drying index (MDI; Friesen 2001). We note that these values were developed for use on caribou, but since our “cervid” taxonomic category contains abundant caribou elements, and because known utility and density indices for white-tailed deer correlate strongly with caribou (e.g., Madrigal and Zimmerman- Holt 2002, Lyman 1994), we believe using caribou indices will not introduce significant bias in our assessment of the element distributions. Table 3 displays Spearman’s rank order correlation coefficients for % MAU and the FUI. In all of the AlDf-24 contexts, there is a moderate to strong negative correlation with the FUI, but at AlDf-30 there is a moderate positive (but insignificant) correlation. Negative correlations between % MAU and the FUI are common on archaeological sites because of attrition and identification issues (Marean and Frey 1997), but the positive correlation was unexpected and may, given attrition, suggest that large meat-laden elements are more frequent at AlDf-30. Correlations between % MAU values and the MDI (Table 3) are similarly negative, indicating limited evidence for meat drying at these sites. To test if density attrition was a factor in the element distributions, we compared the % MAU to a density index (Table 3) developed for caribou (Lam et al. 1999, 2003). An interesting pattern develops here, where we see a decline in correlation over time. In the latest 2 assemblages, there are weak or negative correlations with density that are not significant. This is an intriguing metric for the impact of diagenesis on the faunal remains in these sites, and potentially, the correlation with time and density-mediated attrition. Regardless, as is typical of most faunas, the Middle and Early Maritime Woodland Port Joli assemblages appear to be significantly impacted by densitymediated attrition, and this factor must be carefully weighed when interpreting the faunas. Figure 10. Percent NISP of general taxonomic categories in the Port Joli assemblages. Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 31 many of the migratory bird and fish species can be found in Port Joli Harbour year round. However, a few key species can provide some degree of seasonal indication. American eels migrate from lakes and rivers in large numbers between August and December, with peak activity in September/October. In locations like Port Joli, where eelgrass is plentiful in shallow harbors and bays, eels can be found in large schools in the summer. In fact, an intensive estuarine eel fishery was implemented by some Mi’kmaq in the summer months in such locations (Weiler 2011:14). Spiny dogfish (Squalus acanthias) are a migratory species in Nova Scotia, moving onshore between June and November (Leim and Scott 1966). Unfortunately, most of the duck and geese species are resident year round. For example, Canada geese are known to overwinter in Port Joli Harbour in significant numbers, but can also be found in relatively large numbers throughout the year. Of the waterfowl, only ringnecked duck (Aythya collaris) is non-resident in the area and migrates through only in December. The eastern chipmunk (Tamias striatus) is known to enter a torpor state between the months of November and March (Snyder 1982). It is important to note that these are burrowing rodents, but we encountered no evidence of burrowing activities in the form of soil features or complete skeletons in any of the middens we excavated. Grey seals haul out around the headlands of Port Joli Harbour in May through August/September (Breed et al. 2009:3217). Harp Finally, to roughly measure the potential for bone grease and marrow extracted from cervid remains, we assessed the proportion of spirally cracked (green-fractured) remains in the entire assemblage, including unidentifiable elements and identified fragments (Table 2). While the unidentified fraction could not be distinguished to a taxon below Mammalia, we note that all were large-sized remains consistent with ungulates, and that other-large sized mammals in the collection were present only in exceptionally small frequencies. Table 2 displays the frequencies of spirally fractured fragments in comparison to the entire mammalian assemblage. It is noteworthy that no spirally fractured elements were identified in the AlDf-30 assemblage, and that they were present in very low frequencies (0.5%) of the contemporaneous AlDf-24 Area A assemblage. However, frequencies of spirally fractured elements rise sharply in the Late Maritime Woodland contexts, especially in the Late and Proto assemblages. We point out that these assemblages are associated with an inferred “anvil stone”, likely used to break bones for marrow extraction in the dwelling deposits in AlDf-24 Area C (Hrynick et al. 2012:10). Seasonality Indicators Vertebrate remains Seasonality indicators based on the vertebrate assemblages are difficult to assess in this case because Table 3. Minimum number of elements (MNE) and minimal animal units (MAU) values for family Cervidae. AlDf-24 Area C AlDf-24 Area A AlDf-30 Area A Early Late Proto Element (# per skeleton) MNE MAU MNE MAU MNE MAU MNE MAU MNE MAU Skull (1) 2 2.00 0 0.00 3 3.00 2 2.00 2 2.00 Mandible (2) 1 0.50 1 0.50 6 3.00 9 4.50 3 1.50 Atlas (1) 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 Axis (1) 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 Cervical (5) 0 0.00 0 0.00 1 0.20 1 0.20 0 0.00 Thoracic (13) 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 Lumbar (6,7) 1 0.17 1 0.17 0 0.00 0 0.00 0 0.00 Sacrum (4,5) 0 0.00 1 0.25 0 0.00 0 0.00 0 0.00 Rib (26) 1 0.04 0 0.00 0 0.00 0 0.00 0 0.00 Sternum (2) 0 0.00 0 0.00 0 0.00 0 0.00 0 0.00 Scapula (2) 1 0.50 0 0.00 1 0.50 0 0.00 0 0.00 Humerus (2) 0 0.00 1 0.50 3 1.50 3 1.50 2 1.00 Radio-ulna (2) 1 0.50 0 0.00 2 1.00 6 3.00 3 1.50 Metacarpal (2) 5 2.50 0 0.00 1 0.50 6 3.00 6 3.00 Innominate (2) 0 0.00 0 0.00 0 0.00 2 1.00 1 0.50 Femur (2) 0 0.00 2 1.00 1 0.50 2 1.00 1 0.50 Metatarsal (2) 1 0.50 1 0.50 5 2.50 9 4.50 2 1.00 Phalanges (24) 11 0.46 5 0.21 11 0.46 60 2.50 47 1.96 Correlation FUI2 -0.62 (0.071) 0.45 (0.317) -0.52 (0.126) -0.59 (0.071) -0.87 (0.002) Correlation MDI2 -0.08 (0.005) -0.85 (0.015) -0.62 (0.055) -0.50 (0.139) -0.41 (0.269) Correlation density index2 0.08 (0.009) 0.94 (0.005) 0.48 (0.195) 0.20 (0.599) -0.34 (0.417) Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 32 a disturbance such as storms, predation, or by natural lesions (Marin et al. 2012). The action of shell remodeling may influence stable oxygen isotope values, and therefore a remodeled portion of the shell may not correspond to annual or seasonal banding, nor may it contain a continuous record of local environmental conditions. We avoided sampling any shells with any sort of growth disruption or changes in growth-band formation on the external or internal portion of the shell. In preparation for δ18O sampling at the terminal growing edge, shells were coated with LePage metal epoxy along the axis of maximum growth then cut using a Buehler ISOMET precision low-speed saw with a diamond coated wafering blade. The shell sections were mounted on glass slides with the metal epoxy and cut to a thickness of 3 mm. The samples were polished on a Buehler Ecomet variable speed grinder/polisher using 3 polishing steps: (1) grinding with Al2O3 slurry on a 240 grit paper, (2) fine grinding with distilled H2O and 320 grit paper, and (3) final polishing with a felt micro-polish cloth and 0.3 μm Al2O3 slurry. Between each grinding and polishing step, shells were cleaned in an ultrasonic bath and allowed to air dry. To obtain shell carbonate powder for δ18O analysis, we carefully micro-milled the surrounding metal epoxy away from the ventral margin. Micro-milling began at the ventral margin and moved in ~100-μm consecutive steps that contoured the micro-growth lines using a 1-mm diamond-coated cylindrical drill bit (Komet/Gebr. Brasseler GmbH and Co. KG model no. 835 104 010) on a Mimo precision drill mounted onto a Zeiss Stemi-2000 stereo microscope. Continuous micro-milling, as opposed to drilling, allows for an uninterrupted record to be obtained from the shell’s growth history and minimizes the effects of time-averaging, which is the mixing of different time intervals/seasons of shell growth (Goodwin et al. 2004). Samples were milled along growth lines in continuous steps ranging from 0.3 mm to 1.02 mm, with an average distance of 0.59 mm between the center points of each sample. Shell carbonate powder samples were processed in a Thermo Finnigan MAT 253 continuous flow isotope-ratio mass spectrometer coupled to a GasBench II in the Department of Applied and Analytical Paleontology at the University of Mainz, Mainz, Germany. The δ18Oshell values were calibrated against NBS-19 (δ18O = -2.20‰) with a 1σ external reproducibility (= accuracy) of ±0.07‰, and an internal precision of 0.07‰. The δ18Oshell values are expressed relative to the international Vienna Pee Dee Belemnite (VPDB) standard and are given as seals (Pagophilus groenlandicus) are rare along Nova Scotia’s South Shore, but when they are seen, it is typically in March and April. Based on these data, it is difficult to assign a season of occupation to these sites. However, as we show below, they can be used with more definitive seasonality indicators to provide specificity about the seasonal economic practices of the residents of Port Joli. We note that cod otoliths were recovered in bulk midden samples from many contexts, but preservation of these elements was generally quite poor, and analysis of these osseous structures is ongoing. Shell stable isotope analysis In many shell seasonality studies, either the oxygen isotope values or the growth patterns at the terminal growth margin are used to establish the season of shellfish collection. However, with Mya, the chondrophore (or hinge) is more durable than the terminal growth and preserves well in archaeological contexts (however, we were able to recover some well-preserved specimens with the terminal edge intact). Annual growth lines in the chondrophore, or hinge portion, of softshell clam have been used to establish seasonality in archaeological contexts on the eastern seaboard and the coast of Maine (Belcher 1989; Black 1993; Cerrato et al. 1991,1993; Heflich 2002). By producing acetate peels or thin sections of the chondrophore, Cerrato et al. (1991) identified the presence of both annual and intra-annual features that could be used to interpret seasonal patterns of shellfish collection. To assess the season of occupation at the Port Joli shell midden sites, we conducted high-resolution stable isotope analysis on softshell clam. A study of live-collected softshell clam from the immediate region of AlDf-30 showed that stable oxygen isotopes (δ18Oshell) can also be used to reliably determine seasonality (see Burchell et a. 2014 for a detailed overview of the technique). To examine seasonal patterns of shellfish collection, and possible seasonal site occupation, we analyzed 14 archaeological specimens of softshell clam from AlDf-24 (n = 8) and AlDf-30 (n = 6). Samples were derived from AlDf-24 Area A (N56W50, Levels 3J and 3M) and shell midden deposits to the east of AlDf-24 Area C (units N51W51, N52W51, Levels 3f and 3j, respectively). The AlDf-30 shell samples were collected from the Area A midden, unit N52W56, Level 3c. Prior to sampling for stable isotopes, all shells were examined under 40x magnification to ensure that the ventral margin (the last phase of growth) was intact. The ventral margin of this species is subject to remodeling since its relatively thin shell structure can re-generate after Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 33 culate the relative importance of a taxon with the following function: AI = A/(A + B), where A is the frequency of a high-ranked taxon (or taxa), and B is the frequency of a lower-ranked taxon (or taxa) (Bayham 1979, Ugan and Bright 2001). Because the equation is normed, AI values range between 0 and 1; if the frequency of taxon A increases or decreases, then the abundance index increases and decreases concomitantly. Abundance indices are generally resistant to taphonomic and collection biases, as long as these biases are systemic to all the assemblages being compared, as in the Port Joli samples (e.g., Ugan and Bright 2001:1313; although see cautions by Lyman 2003). AI values were originally developed to test shifts in return rate (Bayham 1979); a decline in an AI is typically interpreted a decrease in the encounter rate of the highly ranked taxa (taxon A), and thus overall foraging efficiency (e.g., Lyman 2003:596, Ugan and Bright 200:1312). Such declines are often interpreted as ‘‘resource depressions’’ of large-bodied taxa resulting from human predation or climactic change. Figure 11 (Table 5) describes changes in abundance indices for taxa in the Port Joli assemblages. Here we have selected an aggregate of all taxa in the family Anatidae for taxon B primarily because they are the only lowerranked taxa present in all assemblages (see Betts and Friesen [2005] for discussion of selection criteria for taxon B). The per mil (‰). In general, the more positive the value, the colder the sea temperature when the shell died, and the more negative the value, the warmer. Table 4 reveals that while the sample sizes for each site are relative small compared to other stable isotope studies of shellfish seasonality, overall the results appear to be indicating seasonal trends in the collection. Of the 14 shells samples, 11 were collected in the spring, and 1 in the autumn, with only 1 shell harvested in the winter and 1 shell in the summer months. An Economic History of Port Joli To assess the general taxonomic shifts in the Port Joli faunal assemblages, we used index measures (cf. Bayham 1979; Broughton 1997, 1999) that cal- Table 4. Seasonality determinations from shell isotope analyses. Site Winter Spring Summer Autumn Total AlDf-24 Area A 5 1 6 AlDf-24 Area C 2 2 AlDf-30 1 4 1 6 Total 1 11 1 1 14 Table 5. Abundance index values for the Port Joli assemblages. AlDf-24 AlDf-30 AlDf-24 AlDf-24 AlDf-24 Index Area A Area A Area C (early) Area C (late) Area C (Proto) Cervidae/ Cervidae + Anatidae 0.0914 0.1614 0.9333 0.9778 0.9944 Gadidae/ Gadidae + Anatidae 0.4212 0.8151 0.8971 0.4167 0.0000 Phocidae/ Phocidae + Anatidae 0.0308 0.0149 0.0000 0.0000 0.0000 Anserini / Anserini + duck 0.5579 0.1116 1.0000 0.8000 0.0000 Figure 11. Abundance index values for the Port Joli assemblages. Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 34 indices provide good support for the trends noted above, but also elucidate some key economic shifts. For example, the cervid index is similar between AlDf-24 Area A and AlDf-30, though it is slightly higher in the latter; however, it increases substantially in the early component of AlDf-24 Area C and gradually increases in the rest of the Late Maritime Woodland deposits. This result indicates that return rates for terrestrial prey increased substantially over the sequence, right up to the moment of European contact. This trend is almost precisely the opposite for seals, which decline steadily over the sequence. This result might be interpreted as a resource depression for marine mammals, though we stress that sample sizes are exceedingly low for seals. The gadid index is intriguing, because it suggests that fishing is much more important in the AlDf-30 contexts compared to the AlDf-24 Area A deposits. Interestingly, the importance of gadid increases in the early Late Woodland assemblages, but declines sharply in the later Late Woodland Deposits. Again, this finding suggests substantially declining return rates from gadid fishing in the latter part of the Late Maritime Woodland. Finally, the goose index follows these general trends from the Late Maritime Woodland through to the Protohistoric periods, but indicates a significant difference between the contemporaneous AlDf-24 Area A and AlDf-30 contexts. The data presented above allows us to move towards building up an economic history for Port Joli Harbour over the Late Middle Maritime Woodland to the Protohistoric Period. Seasonality of sites has long been a critical issue in Maritime Peninsula archaeology. Combining the limited vertebrate seasonality data with the shell isotope analysis suggests some interesting seasonal trends. One of the most important revelations from the shell isotope data is that all sites/contexts considered in this paper appear to have been occupied for at least a portion of the spring (Table 4), meaning that contemporaneous occupations in the Middle Maritime Woodland period may have overlapped, perhaps substantially. If AlDf-24 Area A and AlDf-30 were contemporaneous, this might indicate that subsets of the Wabanaki population that used Port Joli (perhaps family groups) operated independently, choosing to come together and disperse in complex ways. The large midden at AlDf-24 Area A appears to have been occupied in the early spring through the summer months, as suggested by the presence of eels, toporic rodents, and the shell isotopic data. There is some possibility, based on the vertebrate taxa, that this occupation extended into the fall months, perhaps as late as November. In contrast, the vertebrate and invertebrate data from AlDf-30 indicates a fall through spring occupation, with a definitive winter signature indicated by both the shell isotopes and the presence of ring-necked duck. The shellfish information from AlDf-24 Area C again shows a relatively clear spring signature, though we stress the sample size is much smaller. The vertebrate species from AlDf-24 Area C are heavily dominated by cervid remains, which may partially suggest an intensive ungulate hunt in the fall months, when they are in prime condition. The percentages of gadid and Canada goose in the Early Late Woodland assemblage are similar to those encountered in AlDf-30, so it is possible that it also was occupied in the cold season. Only continued isotopic analysis on the shells from these sites can confirm this possibility. Economically, one definitive quality is shared by many of the contexts. Clams dominate the shellfish frequencies in all sites, and other molluscs are only present in trace frequencies, typically less than 0.5% of the identified fraction. However, the bulk-sample analysis indicates that AlDf-24 Area A is significantly different from other contexts described in this report. Specifically, it contains less artifacts, soil, and non-cultural lithic fractions but significantly larger proportions of marine shell (softshell clam) and charcoal than any other context. As suggested in Sanger (1996:523), a deposit such as this might be indicative of “a location specifically designed to dry shellfish for storage”. Given the depth of the midden, its lack of cultural features, and its rapid deposition, a viable explanation for the midden is that it represents the results of seasonally intensive softshell clam processing, likely involving smoking (as evidenced by the large fraction of charcoal in the deposit). The seasonality data suggest this intensive activity occurred primarily during the spring and perhaps summer months. Given the size and rapid accumulation of the midden, this could only have been achieved with a population higher than that at other sites. This site probably represented an aggregation of families, specifically for the communal collection and processing of shellfish for later consumption. While no contemporaneous dwellings were found in the AlDf-24 Area A midden itself, the adjacent black-soil midden in Area B may have represented a possible location for multiple dwellings, where the extensive beach, far above the high-water mark, would have served as a very suitable location, especially in summer. Other taxa, especially eels, Atlantic cod, and birds were taken in abundance at AlDf-24 Area A, while cervids, small seals, and other mammals are significantly underrepresented in this context. These findings are notable because the faunal signature emphasizes taxa that were known Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 35 to have been hunted collectively and cooperatively (e.g., shellfish, birds, and fish), rather than mammal hunts that would have generally consisted of a few hunters. Based on our assessment of seasonality, it appears that the contemporaneous AlDf-30 site may have been occupied by a subset of families (likely just one family) who occupied, or at least participated in, the AlDf-24 Area A spring/summer clam processing event. The fact that both exhibit evidence of springtime shellfish collecting indicates a degree of seasonal overlap in their occupations—suggesting variability in the timing and or degree of participation in the communal shellfish processing. Given the close proximity of AlDf-24 Area A and AlDf-30, it is possible that individuals living at AlDf-30 may have remained at the site for some period during the intensive spring clam-processing season. Following the summer months of eel fishing, birding, and clamming at Port Joli, it appears that some families moved a little further inland for the fall and winter months to sites like AlDf-30. This both provided additional shelter against winter storms and cold, but also readier access to the upland areas of Port Joli Harbour that may have been more attractive to caribou and moose. They also had greater access to snowshoe hare in the forest than they did closer to the coast, as is abundantly reflected in the faunal remains. A significant difference between AlDf-24 Area A and AlDf-30 is the lack of whitetailed deer in the former; perhaps AlDf-30 residents participated in a fall caribou hunt at the site, while AlDf-24 Area A residents practiced opportunistic hunts. Regardless, fishing was still practiced in the cold season at AlDf-30, and the faunal remains suggest a significant effort was placed on capturing gadids, especially large Atlantic cod. Additionally, overwintering Canada goose were taken from Port Joli Harbour in abundance. In summary, the first Wabanaki occupation of Port Joli Harbour occurred during the Middle Maritime Woodland Period, ca 1450 cal B.P., and was tied to sea-level rise and the establishment of large clam-flats at mid-harbor, in the vicinity of AlDf-24 (Neil et al. 2014). So productive were these clam beds that they were exploited by Wabanaki families who came together to collectively harvest and process the clams for storage. While this practice resulted in several massive shell-middens, the radiocarbon evidence suggests this productivity did not last long. The sites may have been occupied for only a few generations before either (a) the clam flats could no longer support such intensive exploitation, or (b) groups focused instead on the harvest of other, less labor-intensive, resources. The Late Maritime Woodland assemblages from AlDf-24 Area C must be interpreted with caution because sample sizes are somewhat lower, and evidence from the richness and cervid element analysis suggests attrition is a significant factor in all these shell-bearing contexts. Given this evidence, it is possible that the reduced incidence of fish, and especially bird, remains is related to taphonomic processes in these contexts. However, we note that major portions of the assemblages from both the Early and Late contexts at AlDf-24 Area C included undisturbed shell midden deposits. Also, the cervid element analysis shows that density attrition was lower in these contexts than in AlDf-24 Area A and AlDf-30. Furthermore, a review of the identification catalogue indicates that only 6% of the bird and fish remains in the AlDf-24 Area C deposit were recovered from the midden, as opposed to the shellbearing black-soil deposits. Finally, only ~10% of remains in the unidentified fraction in all deposits came from fish and birds. This finding suggests that preservation is likely not responsible for the low frequency of fish and birds in the Late Maritime Woodland at Port Joli. Instead, we propose that it represents a real and significant shift in the exploitation of marine resources in the harbor. This shift is coeval with the large frequencies of cervid remains recovered from the Late Maritime Woodland deposits at AlDf-24 Area C. Proportions of cervid remains increase steadily over the sequence, with a concomitant decline in fish and bird remains. Interestingly, while the transition is relatively abrupt, the Early Maritime Woodland contexts at AlDf-24 Area C do appear to transition from the Middle Maritime Woodland deposits. For example, gadids are still relatively abundant in the early deposits, as are hares and waterfowl. However, by the Late and Protohistoric levels of the deposit, bird taxa decline significantly (as do hare in the Protohistoric levels). Structurally, the AlDf-24 Area C deposit most closely resembles AlDf-30; they are both composed of a central shell-bearing black-soil midden with evidence of a series of living floors, likely created by a single reoccupied wigwam, with an adjacent “kitchen midden” (e.g., Hrynick and Betts 2014, 2017 [this volume]; Hrynick et al. 2012). The high quantities of cervid remains and hare may suggest a fall occupation, when these taxa were in their prime. However, the shell isotope data clearly indicates a spring season of collection. Though more research is needed, especially on the shell isotopes, we propose that the AlDf-24 Area C deposits represent a coldseason, fall to spring, occupation similar to AlDf- 30. Regardless, the archaeofaunal and structural Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 36 differences between AlDf-24 Area C and AlDf-24 Area A suggest a fundamental difference in the way this portion of Port Joli’s coastline was used in the Late Maritime Woodland Period. In effect, AlDf-24 Area A ceased to support warm-season aggregations of families for the purposes of processing clams for later storage, and became a cold-season site occupied by 1 or perhaps 2 families who specialized in hunting cervids. In some respects, it appears to be largely a continuation of the Middle Maritime Woodland practice of coastal cold-season sites, like AlDf-30, but with radically different economic signatures. Unfortunately, the sites documented in this report do not appear to provide evidence for warm-season occupations in this area of Port Joli. However, as described above, we documented evidence of very large shell-bearing black-soil midden sites located near the head of the harbor at rivers and lake outlets. We hypothesize that these are large Late Maritime Woodland warmseason aggregation sites, similar to those in the Middle Maritime Woodland Period. We tested several of these sites, and while analysis is still ongoing, we observed an extreme reduction in the ratios of clam shell to soil in these deposits. This finding is consistent with the fact that clam flats are not as extensive near the head of the harbor (e.g., Neil et al. 2014). Although we did not encounter this in any of our testing, Erskine (1986:92) claims to have found great deposits of herring scales at one of these sites (AlDf-03), and it is possible that herring fishing was a staple economic pursuit in these contexts. We note that Erskine’s fieldwork observations have largely been substantiated by our archaeological investigations, and while his cultural historical reconstruction is out of date, there is no compelling reason to assume that his field observations about fish scales are in error (e.g., Betts 2011, Betts and Hrynick 2013). Alewife spawn in local rivers and streams in the region between April and July, and all of these large shell-bearing black-soil sites are located on current or former lake outlet streams where alewife would have been available. Unfortunately, this hypothesis can only be confirmed with further analysis, though it is consistent with a focus on warm-season herring fishing at rivers near the heads of marine inlets in the Late Maritime Woodland recently identified by Milner (2014) in St. Croix. Placing Port Joli in a Regional Economic Context We interpret the large Middle Maritime Woodland shell midden at AIDf-24 Area A in Port Joli as a location where multiple families came together in the warm season for the communal purpose of collecting, preserving (perhaps via smoking), and storing shellfish for use in subsequent seasons. Erskine (1962) had reached this conclusion about the large shell-midden at AlDf-25 over 50 years ago, and we suspect the nearby midden at AlDf-26 represents a similar deposit. If so, these 3 deposits represent a unique class of site, which appear to have no analogue in the rest of the Maritime Provinces or the northern Gulf of Maine (e.g., Sanger 1996:563). Evidence from AlDf- 24 Area C indicates that the nature of occupation of these large processing sites changed radically in the Late Maritime Woodland period, ca. 1300 to 1200 cal B.P., with a shift from large summer occupations to limited, but intensive, cold-season occupations. Unfortunately, the sites considered in this paper do not appear to provide evidence for Late Maritime Woodland warm-season occupations in this area of Port Joli, but large shell-bearing black-soil middens located on streams and rivers near the head of the harbour are probable candidates for warm-season aggregation sites during this time. It should be noted that existing models for resource use in Nova Scotia (e.g., Davis 1987,1993; Nash 1984; Nash and Miller 1987; Neitfeld 1981; Stewart 1989) have reached little consensus on prehistoric settlement and subsistence in the province. Davis’ (1993:97) contiguous habitat subsistence model “… sees four habitat zones as contributing to settlement patterns and transhumance in precontact times. The zones include an inshore marine habitat, an intertidal habitat, a riverine/lakes habitat and a forest habitat.” Davis’ model relies heavily on use of interior resources, particularly near lake outlets during a portion of the year, and in many respects is similar to the traditional summer/coastal, winter/ interior “ethnohistoric maritime model” (e.g., Nash and Miller 1987:42). Critical of the ethnohistoric model, Nash (1984:225) proposed a “mosaic model” of settlement and subsistence, whereby “adaptation to resource variability … yields a mosaic of maritime adaptions—each a local expression of a flexible and generalized economy …”. In a review of existing faunal data she had produced in the 1980s, Stewart (1989) concluded that, in Cape Breton Island at least, “… there is some evidence for use of the coast in the winter and inland areas in the summer …”. Neitfeld (1981:5), in her comprehensive compilation of subsistence data for the region, proposed the possibility for permanent coastal settlement. While she had “no direct archaeological indications of seasonality” (Neitfeld 1981:558), she proposed that Middle Maritime Woodland peoples on the South Shore occupied the coast year-round with “relatively permanent fall-through-spring coastal settlements” near large shellfish beds (Neitfeld 1981:209–210), Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 37 and practiced a mobile interior/coastal settlement strategy during the summer months. She further proposed that this system reversed itself in the Late Maritime Woodland, with large semi-permanent summer sites located on the coast and smaller, coldseason sites located in the deep interior. If precontact settlement and subsistence in Nova Scotia was a mosaic, then the Port Joli data provides a new vantage point from which to consider such proposed variability. Here we have direct evidence of year-round settlement on the coast for the Middle Maritime Woodland Period, with indirect evidence for a reorganized, but still year round, settlement in Port Joli during the Late Maritime Woodland Period. These sites suggest, essentially, exploitation of precisely the same resources year round, but in different frequencies based on season. This pattern seems to have been linked to both population aggregations tied to resource extraction and processing (e.g., AlDf-24 Area A) and perhaps a need for lessexposed locations during the winter months at sites where terrestrial resources were more readily available (e.g., AlDf-30). In the Middle Maritime Woodland Period, these seasonal moves were very limited indeed, with small family groups moving camps perhaps several hundred meters a season. This form of seasonal movement on the coast is similar to that identified in Passamaquoddy Bay (e.g., Black 2002, Sanger 1996), where there is a pattern of moreexposed, warm-season coastal sites and protected, or sheltered, cold-season sites. However, we note that in Passamaquoddy Bay these more-exposed locations often were on islands, while warm-season sites were located on the mainland. Despite the presence of onshore and offshore haul-out locations in the vicinity of Port Joli, seal remains in Port Joli middens occur only in trace frequencies, in stark contrast to many Passamaquoddy Bay sites (e.g., Black 2002, 2017 [this volume]). Furthermore, the peak in seal exploitation evident in the Early Late Maritime Woodland in Passamaquoddy Bay (Black 2017 [this volume]) is not apparent in Port Joli at all (in fact, there is no evidence for seal exploitation at all in the Late Maritime Woodland in Port Joli). Ingraham et al. (2016) have suggested that selective discard of phocid remains may be responsible for the general dearth of seal remains at some Maine shell midden sites. Regardless, frequencies of seal remains in Port Joli middens are an order of magnitude smaller than most comparable Maine coastal sites, when sample sizes are taken into account (e.g., Black 2017 [this volume]:table 3, Davis 1987:200–201). This dearth of seal bones, despite the proximity of seal haul-outs, does indicate that seals were not as intensively exploited at Port Joli as in other regions. It is possible that the proximity of the haulout locations to the earliest Port Joli sites resulted in a seal microhabitat depression (e.g., Betts et al. 2009), where seals changed their haul-out locations in response to human proximity. On the relatively linear South Shore coastline, this might have involved a move to haul-outs many kilometers away (e.g., Port Mouton or Port L’Hebert), meaning that seals essentially became unavailable for local hunting forays. The insular nature of Passamaquoddy Bay may have prevented such microhabitat depressions from occurring in much of that region. The limited range of harvested molluscs seems generally comparable to sites in the Northern Gulf of Maine, but shellfish species frequencies are very different from Passamaquoddy Bay deposits. There, mussels and other shellfish are sometimes as great as 10–30% percent by shell weight (e.g., Black 2002:308, 311); in Port Joli, clams comprise more than 99% of the total shell fraction by weight. This variation is likely due to a contrast in predominate substrate; the sandy shores of Port Joli are very different from the rocky intertidal zones, attractive to mussels and urchins, accessible from many Passamaquoddy Bay sites. The most prominent subsistence shift in Port Joli appears to have been a switch from a fishing- and birding-dominated economy to one dominated by cervids. The increasing quantity of cervid remains over the Middle Maritime Woodland to Late Maritime Woodland transition is very similar to that identified by Spiess et al. (2006:150) at the Indian Town Island site in Maine. However, in that case, frequencies of fish and bird remains are substantially higher than in the Port Joli deposits. Similarities and differences between the regions also occur in the stratigraphic and depositional composition of deposits. Substantial Middle Maritime Woodland shell middens have been well documented in the Gulf of Maine, but they are characterized by “alternating sets of living floor [gravel based] and shell midden deposits sandwiched between natural soil layers” (Black 1991:51; see also Black 2002:308). While the deep Middle Maritime Woodland processing middens at the Port Joli are clearly stratified, with alternating layers of broken and unbroken shell, there is no evidence of dwelling floors or formal activity areas within them. In fact, the artifact diversity and density in these middens is significantly lower than in many contemporaneous Gulf of Maine shell middens. These structural differences are intriguing and suggest contrasts in both resource specialization (e.g., clam processing), and the structure and use of sites between the 2 regions. Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 38 The difference in the composition of Middle and Late Maritime Woodland deposits at Port Joli is striking. At Port Joli, that transition appears to occur at ca. 1300 cal B.P., almost simultaneously with that defined in the Quoddy region by Black (2002). In Port Joli, the shift seems to be characterized as one from large summertime processing middens on the coast and smaller, wintertime habitation sites in the near interior to large black-soil middens occupied year-round at the head of the harbor, with smaller winter satellite camps occurring both on the coast and near interior. The seasonality and settlement particulars of this reconstruction are at odds with those proposed by Neitfeld (1981), particularly for the Middle Maritime Woodland, yet the year-round focus on the coast is supported by our research. The appearance of black-soil middens of substantial areal extent in the Late Maritime Woodland is similar to the Quoddy region sequence as defined by Black (2002), although in the Quoddy region these deposits are often located on top of the Middle Maritime Woodland shell middens, not in different locations. At Port Joli, the association of these large black-soil sites with lake outlet streams might suggest that these occupations were focused on the summer exploitation of alewife runs, which probably accounts for the significant decrease of shell in these deposits. In fact, Erskine (1986:92) only noted recovering herring bones and scales at the large black-soil middens associated with lake outlets. Other differences are apparent. Most perplexing of these is the absence of evidence for Early Maritime Woodland deposits in Port Joli. Unlike the Northern Gulf of Maine region, these do not occur as black-soil middens below the Middle Maritime Woodland middens (e.g., Black 2002). In fact, no diagnostically Early Maritime Woodland material has been recovered from any of the Port Joli excavations in the last 60 years (one pre-1500 cal B.P. radiocarbon date reported by Betts [2011] is now considered to be contaminated). Sea-level rise and its impact on the size and distribution of foreshore clam flats is clearly a significant driver of the distribution of sites throughout the Maritime Woodland Period at Port Joli (e.g., Neil et al. 2014). This record appears in stark contrast to recent research at the Boswell site (e.g., Deal 2013), where evidence of occupation can be traced to at least the Terminal Archaic, highlighting the importance of inundation modeling in the exploration of prehistoric coastal settlement. What were the social and environmental mechanisms behind the shifts in Port Joli settlement and subsistence? The shift from the Middle Maritime Woodland to the Late Woodland ca. 1300 cal B.P. was associated with a decline in spruce and birch pollen in Port Joli (e.g., Neil et al. 2014: 205) and an increase in sphagnum and pine pollen. The total pollen accumulation rate also peaked during this period (Neil et al. 2014:206). Furthermore, annual temperature spiked at ca. 1250 cal B.P., as did total annual precipitation (Neil et al. 2014:figure 5). These findings indicate that climate had changed to a warmer and wetter environment during this period. While it is difficult to understand how these changes affected the local faunal populations, perhaps warmer temperatures, particularly in winter, favored ungulate populations, while rising sea levels might have inundated formerly productive clam flats and eel grass beds, resulting in reduced access to mollusc and bird species that relied on them. Climate change ca. 1250 cal B.P. may also explain the Late Maritime Woodland shift to white-tailed deer-dominated assemblages from earlier caribouand moose-dominated assemblages in the Middle Maritime Woodland. With the return of warmer temperatures, white-tailed deer populations may have increased in the region, while caribou and moose populations declined, consistent with known low/high-latitude axis adaptations for deer species in North America (Geist 1998). While these environmental shifts potentially had an impact on economic practices in Port Joli, social and demographic factors may be equally responsible for changes in economic and settlement practices in Port Joli. More research is needed, specifically into the nature of the large shell-bearing black-soil sites closer to the head of the harbor. Investigation of material excavated from these deposits is ongoing, and inclusion of these data will both help complete the economic history of Port Joli and permit more informed inferences on the nature of its development. Acknowledgments The faunas described in this paper were collected as part of a public archaeology project made possible by collaboration with Acadia First Nation, Thomas Raddall Provincial Park, and the Nova Scotia Department of Natural Resources. We especially wish to thank Acadia First Nation for their collaboration, support, and consultation on all aspects of the project. We also appreciate the assistance of University of New Brunswick and Acadia First Nations students who helped to excavate these faunal assemblages. Funding for this research was provided by the Canadian Museum of History and Memorial University of Newfoundland. Other support was provided by the Harrison Lewis Centre, and the University of Connecticut. An early version of this paper was presented at the 2014 Eastern States Archaeological Federation meeting; we are sincerely grateful for comments we received there. Finally, we wish to thank Danielle Desmarais, Vasa Lukich, Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 39 and Lesley Howse, who identified the vertebrate remains under contract. Their analyses were supported in part by the Howard Savage Osteoarchaeological Laboratory at the University of Toronto. Literature Cited Bayham, F., 1979. Factors influencing the archaic pattern of animal utilization. Kiva 44:219–235. Belcher, W.A. 1989. Prehistoric fish exploitation in east Penobscot Bay, Maine: The Knox site and sea-level rise. Archaeology of Eastern North America 17:175–191. Betts, M.W. 2009. Permit report: E’se’get archaeology project 2008 field season. Manuscript on file. Nova Scotia Museum, Halifax, NS, Canada. Betts, M.W. 2010. Permit report: E’se’get archaeology project 2009 field season. Manuscript on file. Nova Scotia Museum, Halifax, NS, Canada. Betts, M.W. 2011. Permit report: E’se’get archaeology project 2010 field season. Manuscript on file. Nova Scotia Museum, Halifax, NS, Canada. Betts, M.W., and T.M. Friesen. 2005. Declining foraging returns from an inexhaustible resource? Abundance indices and beluga whaling in the western Canadian Arctic. Journal of Anthropological Archaeology 25:59–81. Betts, M.W., and M.G, Hrynick. 2013. Permit report: E’se’get archaeology project 2012 field season. Manuscript on file. Nova Scotia Museum, Halifax, NS, Canada. Betts, M.W., S.E. Blair, and D.W. Black. 2012. Perspectivism, mortuary symbolism, and human–shark relationships on the Maritime Peninsula. American Antiquity 77(4):621–45. Black, D.W. 1991. Stratigraphic integrity in northeastern shell middens: An example from the insular Quoddy Region. Pp. 205–220, In M. Deal and S. Blair (Eds.). Prehistoric Archaeology in the Maritime Provinces: Past and Present Research. Maritime Committee on Archaeological Cooperation, Fredericton, NB, Canada. 330 pp. Black D.W. 1993. Living close to the ledge: Prehistoric human ecology of the Bliss Islands, Quoddy Region, New Brunswick, Canada. Occasional Papers in Northeastern Archaeology No. 6. Copetown Press, St. John’s, NL, Canada. Black, D. 2002. Out of the blue and into the black: The Middle–Late Maritime Woodland transition in the Quoddy Region, New Brunswick, Canada. Pp. 301– 320, In J.P. Hart and C.B. Reith (Ed.). Subsistence- Settlement Change, A.D. 700–1300. New York State Education Department, Albany, NY, USA. 360 pp. Black, D. 2017. Archaeological sea mammal remains from the Maritime provinces of Canada. Journal of the North Atlantic Special Volume 10:70–89. Breed, G.A., I.D. Jonsen, R.A. Myers, W.D. Bowen, and M.L. Leonard. 2009. Sex-specific, seasonal foraging tactics of adult grey seals (Halichoerus grypus) revealed by state–space analysis. Ecology 90:3209– 3221. Broughton, J. 1997. Widening diet breadth, declining foraging efficiency, and prehistoric harvest pressure: Ichthyofaunal evidence from the Emeryville Shellmound, California. Antiquity 71:845–862. Broughton, J. 1999. Resource depression and intensification during the late Holocene, San Francisco Bay: Evidence from the Emeryville Shellmound vertebrate fauna. University of California Anthropological Records 32. Berkeley, CA, USA. 158 pp. Burchell, M., M. Betts, and K.A. Patton. 2014. Preliminary analysis of stable oxygen isotopes and shell growth in the softshell clam, Mya arenaria: Implications for interpreting seasonality and shellfish harvesting in Port Joli, Nova Scotia. North Atlantic Archaeology 3:93–108. Cerrato, R.M., H.V.E. Wallace, and K.G. Lightfoot. 1991. Tidal and seasonal patterns in the chondrophore of the softshell clam, Mya arenaria. Biological Bulletin 181:307–311. Cerrato, R.M., K.G. Lightfoot, and H.V.E. Wallace. 1993. Prehistoric shellfish-harvesting strategies: Implications from the growth patterns of softshell clams (Mya arenaria). Antiquity 67:358–369. Davis, S.A. 1987. Man, molluscs and mammals: A study of land use in the later Holocene of the Maritime Provinces of Canada. Unpublished Ph.D. Dissertation. University of Oxford, Oxford, UK. Davis, S.A. 1993. The Ceramic Period in Nova Scotia. Pp. 85–100, In M. Deal and S. Blair (Eds.). Prehistoric Archaeology in the Maritime Provinces: Past and Present Research. Maritime Committee on Archaeological Cooperation, Fredericton, NB, Canada. 330 pp. Deal, M., 2013. Boswell Site Archaeological Project 2012. Archaeology in Nova Scotia: 2012 News 4:24–35. Erskine, J.S. 1962. Micmac Notes 1962. Nova Scotia Museum, Halifax, NS, Canada. Erskine, J.S. 1986. Unpublished Papers on the Archaeology of the Maritime Provinces. Edited by. M Deal. St. Mary’s University, Halifax, NS, Canada. 268 pp. Friesen, T.M. 2001. A zooarchaeological signature for meat storage: Re-thinking the drying utility index. American Antiquity 66(2):315–331. Geist, V. 1998. Deer of the World: Their Evolution, Behaviour, and Ecology. Stackpole Books, Mechanicsville, PA, USA. Goodwin, D.H., K.W. Flessa, M.A. Téllez-Duarte, D.L. Dettmann, B.R. Schöne, and G.A. Avlla-Serrano. 2004. Detecting time-averaging and spatial mixing using oxygen isotope variation: A case study. Palaeogeography, Palaeoclimatology, Palaeoecology 205:1–21. Heflich, A. 2002. Examination of precolonial seasonal habitation at the Coastal Studies Centre through a study of the softshell clam (Mya arenaria) growth patterns. Unpublished manuscript. School of Marine and Atmospheric Sciences, Stony Brook University, Stony Brook, NY, USA. Hrynick, M.G., and M.W. Betts. 2014. Identifying sweathouses in the archaeological record: A Maritime Woodland period case study from Nova Scotia, Canada. Journal of Anthropological Archaeology 35:92–105. Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 40 Hrynick, M.G., and M.W. Betts. 2017. A relational approach to hunter-gatherer architecture and gendered use of space at Port Joli Harbour, Nova Scotia. Journal of the North Atlantic Psecial Volume 10:1–17. Hrynick, M.G., M.W. Betts, and D.W. Black. 2012. A Late Maritime Woodland period dwelling feature from Nova Scotia’s South Shore: Evidence for patterned use of domestic space. Archaeology of Eastern North America 40:1–25. Ingraham, R.C., B.S. Robinson, K.D. Sobolik, and A. S. Heller. In press. Left for the tide to take back: Specialized processing of seals on Machias Bay, Maine. Journal of Island and Coastal Archaeology. Lam, Y. M., X. Chen, and O. M. Pearson. 1999. Intertaxonomic variability in patterns of bone density and the differential representation of bovid, cervid, and equid elements in the archaeological record. American Antiquity 64 (3):343–362. Lam, Y. M., O.M. Pearson, C.W. Marean, and X. Chen. 2003. Bone-density studies in zooarchaeology. Journal of Archaeological Science 30:1701–1708. Leim, A.H., and W.B. Scott. 1966. Fishes of the Atlantic coast of Canada. Fisheries Research Board of Canada #155. Ottawa, ON, Canada. 485 pp. Lyman, L. 1994. Vertebrate Taphonomy. Cambridge University Press, New York, NY. 524 pp. Lyman, R.L. 2003. The influence of time averaging and space averaging on the application of foraging theory in zooarchaeology. Journal of Archeological Science 30:595–610. Madrigal, T.C., and J. Zimmerman-Holt. 2002. Whitetailed deer meat and marrow return rates and their application to eastern Woodlands archaeology. American Antiquity 67(4):745–759. Marean, C.W., and C.J. Frey. 1997. Animal bones from caves to cities: Reverse utility curves as methodological artifacts. American Antiquity 62 (4):698–711. Marin, F., N. Le Roy, and B. Marie. 2012. The formation and mineralization of mollusk shell. Frontiers in Bioscience S4:1099–1125. Maritime Archaeological Resource Inventory (MARI). 2010. MARI On-Line. Available online at https:// mims.ednet.ns.ca/. Accessed 21 January 2010. Website is being revised, but data is available from the Nova Scotia Museum, Halifax, NS, Canada. Metcalfe, D., and K. Jones. 1988. A reconsideration of animal body-part utility indices. American Antiquity 53(3):486–504. Milner, C. 2014. The Precontact Village at St. Croix (BfDa-1), Nova Scotia: Explorations of site size and stratigraphic integrity. Unpublished M.A. Thesis. Department of Archaeology, Memorial University of Newfoundland, St. John’s, NL, Canada. Murphy, B.M., and D.W. Black. 1996. Zooarchaeology in the Canadian Maritimes. Canadian Zooarchaeology 9:2–19. Nash, R.J. 1984. Mi’kmaq: Economics and Evolution. St. Francis Xavier University, Antigonish, NS, Canada. 312 pp. Nash, R.J., and V.P. Miller. 1987. Model building and the case of the Micmac Economy. Man in the Northeast 34:41–56. Nash, R., and F. Stewart. 1986. Mi’kmaq economics and evolution: Faunal remains from the Delorey Island site (BjCj-9) of Nova Scotia. Nova Scotia Curatorial Report #57, Halifax, NS, Canada. 126 pp. Nash, R., and F. Stewart. 1990. Melanson: A large Micmac village in Kings County Nova Scotia. Nova Scotia Curatorial Report #67, Halifax, NS, Canada. 213 pp. Neil, K., K. Gajewski, and M. Betts. 2014. Human–ecosystem interactions in relation to Holocene environmental change in Port Joli Harbour, southwestern Nova Scotia, Canada. Quaternary Research 81(2):203–212. Nietfeld, P. 1981. Determinants of aboriginal Micmac political structure. Ph.D. Dissertation. Department of Anthology, University of New Mexico, Albuquerque, NM, USA Petersen, J.B., and D. Sanger. 1993. An aboriginal ceramic sequence for Maine and the Maritime provinces. Pp. 113–170, In M. Deal and S. Blair (Eds.). Prehistoric Archaeology in the Maritime Provinces: Past and Present Research. Maritime Committee on Archaeological Cooperation, Fredericton, NB, Canada. 330 pp. Powell, S. 1990. Archaeological survey of Sandy Bay Provincial Park, Cole Harbour Day Use Park and Crystal Crescent Beach Provincial Park. Manuscript on file, Nova Scotia Museum, Halifax, NS, Canada. Powell, S. 1995. Archaeological evaluation of AlDf-26, Port Joli. Ms. on file, Nova Scotia Museum, Halifax, NS, Canada. Raddall, T.H. n.d. Indian kitchen-middens in Queens County. Manuscript on file. Dalhousie University Archives, Halifax, NS, Canada. Rojo, A. 1986. Live length and weight of cod (Gadus morhua) estimated from various skeletal elements. North American Archaeologist 7:329–351. Rojo, A. 1990. Faunal analysis of fish remains from Cellar’s Cove, Nova Scotia. Archaeology of Eastern North America 18:89–108. Sanger, D. 1996. Testing the models: Hunter-gatherer use of space in the Gulf of Maine, USA. World Archaeology 27(3): 512–526. Snyder, D.P. 1982. Tamias striatus. Mammalian Species 168:1–8. Spiess, A.E., K.D. 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. Stewart, F. 1989. Seasonal movements of Indians in Acadia as evidence by historical documents and vertebrate faunal remains from archaeological sites. Man in the Northeast 38:55–77. Ugan, A., and J. Bright. 2001. Measuring foraging efficiency with archaeological faunas: The relationship between relative abundance indices and foraging returns. Journal of Archaeological Science 28:1309– 1321. Journal of the North Atlantic M.W. Betts, M. Burchell, and B.R. Schöne 2017 Special Volume 10 41 Weiler, M. 2011. Mi’kmaq and the American eel. Traditional knowledge relating to the American eel by mainland Nova Scotia Mi’kmaq AFSAR Project # 1734. Report to Environment Canada and Fisheries and Oceans Canada March. Mi’kma’ki All Points Services, Shubenacadie, NS, Canada. 39 pp. Endnotes 1E’se’get is a Mi’kmaw word meaning “to dig for clams”. 2Spearman’s rank order correlation (probabilities in brackets).