Towards a Historical Ecology of Intertidal Foraging in the Mafia Archipelago

Home , Paleozoology, Phalium glaucum

Towards a Historical Ecology of Intertidal Foraging in the Mafia Archipelago: Archaeomalacology and Implications for Marine Resource Management Authors: Patrick Faulkner, Matthew Harris, Othman Haji, Alison Crowther, Mark C. Horton, et. al. Source: Journal of Ethnobiology, 39(2) : 182-203 Published By: Society of Ethnobiology URL: https://doi.org/10.2993/0278-0771-39.2.182

BioOne Complete (complete.BioOne.org) is a full-text database of 200 subscribed and open-access titles in the biological, ecological, and environmental sciences published by nonprofit societies, associations, museums, institutions, and presses.

Your use of this PDF, the BioOne Complete website, and all posted and associated content indicates your acceptance of BioOne’s Terms of Use, available at www.bioone.org/terms-of-use.

Usage of BioOne Complete content is strictly limited to personal, educational, and non-commercial use. Commercial inquiries or rights and permissions requests should be directed to the individual publisher as copyright holder.

BioOne sees sustainable scholarly publishing as an inherently collaborative enterprise connecting authors, nonprofit publishers, academic institutions, research libraries, and research funders in the common goal of maximizing access to critical research.

Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library Journal of Ethnobiology 2019 39(2): 182–203

Towards a Historical Ecology of Intertidal Foraging in the Mafia Archipelago: Archaeomalacology and Implications for Marine Resource Management

Patrick Faulkner1,2*, Matthew Harris3, Othman Haji4, Alison Crowther2,3, Mark C. Horton5, and Nicole L. Boivin2

Abstract. Understanding the timing and nature of human influence on coastal and island ecosystems is becoming a central concern in archaeological research, particularly when investigated within a historical ecology framework. Unfortunately, the coast and islands of eastern Africa have not figured significantly within this growing body of literature, but are important given their historically contingent environmental, social, and political contexts, as well as the considerable threats now posed to marine ecosystems. Here, we begin developing a longer-term understanding of past marine resource use in the Mafia Archipelago (eastern Africa), an area of high ecological importance containing the Mafia Island Marine Park. Focusing on the comparatively less researched marine invertebrates provides a means for initiating discussion on potential past marine ecosystem structure, human foraging and environmental shifts, and the implications for contemporary marine resource management. The available evidence suggests that human-environment interactions over the last 2000 years were complex and dynamic; however, these data raise more questions than answers regarding the specific drivers of changes observed in the archaeomalacological record. This is encouraging as a baseline investigation and emphasizes the need for further engagement with historical ecology by a range of cognate disciplines to enhance our understanding of these complex issues. Keywords: eastern Africa, Juani Island, Iron Age, archaeomalacology, marine resource use

Introduction archaeological inquiry, particularly when Human impacts on oceans and coastal embedded within historical ecology and/or margins from over-exploitation, urban- applied zooarchaeology. Historical ecology ization, industrialization, and resulting is an interdisciplinary approach combining pollution have caused habitat loss, reduced long-term historical perspectives with ecol- species diversity, and accelerated oceanic ogy to understand ecosystem evolution, and atmospheric warming on a global natural and anthropogenic environmental scale (as summarized by Jackson et al. modifications, and human-environmental 2001). While these issues have garnered interactions (Balée and Erickson 2006; worldwide attention, the contemporary Crumley 1994). As such, historical ecol- situation can be viewed as part of a much ogy is particularly effective for framing longer-term history of humans exploiting research agendas focused on the types of coasts and islands (Braje et al. 2017; Jack- issues noted above in a range of terrestrial son et al. 2001). The timing and nature of and aquatic environments (Boivin et al. human interactions with coastal and island 2016; Braje and Rick 2013; Fitzpatrick and ecosystems has become a focal point of Keegan 2007).

1 Department of Archaeology, School of Philosophical and Historical Inquiry, Brennan MacCallum Building A18, The University of Sydney, NSW, 2006 Australia. 2 Max Planck Institute for the Science of Human History, Jena, Germany. 3 School of Social Science, The University of Queensland, Brisbane, Australia. 4 Department of Museums and Antiquities, Zanzibar, Tanzania. 5 Royal Agricultural University, Cirencester, England. *Corresponding author ([email protected])

Possible multidirectional changes in fifteenth centuries CE), with similar trends past socio-economic organization may apparent in the fish data from all four be recognized within a historical ecolog- trenches. Marine and terrestrial tetrapods ical approach, broadening the scope and are relatively scarce through the EIA and significance of zooarchaeological research, MIA-LIA, with small game (e.g., duiker) feeding long-term data into the devel- and marine turtle dominant. Archaeobotan- opment of more effective conservation, ical evidence indicates crops in very small management, and environmental policies quantities, all of which are attributed to the (Balée 2006; Braje et al. 2017; Rick and MIA-LIA, indicating a lack of convincing Lockwood 2013). Schwerdtner Máñez agricultural evidence for the EIA. Based on et al. (2014:1–2), for example, note five these data, Crowther et al. (2016) interpret key areas of inquiry within the historical the economic structure as representing an ecology of marine environments, includ- initial EIA maritime adaptation during the ing assessing the relative importance of phase of island colonization, followed by key drivers of environmental change, a mixed herding/farming economy in the understanding the significance of marine MIA-LIA that continued to include signifi- resources for human societies, and consid- cant marine resources. ering how historical information can be Here, we expand on the previ- used to improve future management. In ous archaeomalacological analyses to reviewing the range of potential threats to focus on trends in human foraging and modern marine environments, Crain et al. human-environment interactions from the (2009:57) also suggest that understanding EIA to the present and, in doing so, consider the cumulative effects of multiple threats the implications of these data for resource to marine ecosystems must be a research management. Following Contreras (2016), priority. Zooarchaeology and historical rather than concentrating on whether envi- ecology can make a significant contribu- ronments influenced humans or vice versa, tion in these areas, providing insights into we explore how humans and their environ- ecosystem change, human ecodynamics, ments interacted. As such, we emphasize and the assessment of ecological bench- the interaction of humans and their envi- marks for a range of marine taxa (e.g., ronment within a specific context, taking a Douglass et al. 2018, 2019; McKechnie et temporal approach to investigate the nature al. 2014; Wolverton et al. 2016). of intertidal foraging and to isolate potential The aim of this paper is to develop variability or stability in these behaviors. an understanding of past marine mollusk Archaeomalacological research in eastern use at the Juani Primary School site in the Africa is in an early stage of development Mafia Archipelago (eastern Africa), an area (e.g., Douglass 2017; Faulkner et al. 2018a, contained within the Mafia Island Marine 2018b; Msemwa 1994), meaning that the Park. Detailed analyses and interpretations past use of these resources remains relatively of the Sealinks Project investigation at this poorly understood. In contrast, archaeolog- site are presented elsewhere (Crowther et ical research in other coastal regions has al. 2016). However, in summary, the site demonstrated the influence that human contained rich ceramic assemblages along- foraging can have on mollusk abundance, side abundant faunal remains dominated richness, and diversity, and how these effects by fish and marine mollusks. Based on contribute to broader ecosystem alterations preliminary analyses from a single trench, (e.g., Giovas et al. 2010; Jerardino 2016; the mollusks reduce in density and rich- Jerardino and Marean 2010; Kyriacou et al. ness from the Early Iron Age (EIA, circa 2015). As marine invertebrates fill a crucial first–sixth centuries CE) to the Middle Iron role within marine ecosystems (e.g., habi- Age-Late Iron Age (MIA-LIA, circa seventh– tat complexity, trophic structure), baseline

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library 184 Faulkner, Harris, Haji, Crowther, Horton, and Boivin

data on long-term molluskan richness and The high productivity of the Mafia Island diversity has implications for contemporary ecosystem is partly due to the nutritional resource use, marine habitat structure, and supply flowing from the Rufiji Delta (TCMP ecosystem/resource management (Peacock 2003:53), with the MIMP containing coral et al. 2012). reefs, mangroves, sea grasses, and lagoons (Andersson and Ngazi 1995:476). Habi- Background tat and species diversity are higher in the eastern part of the Archipelago and the Mafia Archipelago: Environmental Chole Bay area, in particular, comprises a Context mosaic of marine ecosystems. Hard reefs The Mafia Archipelago is located are located to the southwest of Mafia Island 20 km off the southern coast of Tanza- and a 33-km fringing reef extends down nia (Figure 1), 120 km southeast of Dar the length of the Mafia, Juani, and Jibondo es Salaam, and separated from the south- Island coastlines (TCMP 2001:13). Seagrass ern extent of the Rufiji Delta on the beds around Mafia Island contain a high mainland by the shallow Mafia Channel diversity of marine organisms, includ- (Andersson and Ngazi 1995:476; MPRU ing fish, crustacea, and mollusks (TCMP 2011:8; TCMP 2003:53). The Archipelago 2001:25–26). The subtidal areas through- comprises a chain of sandstone and coral out Chole Bay are shallow (less than 20 m rag islands (e.g., Bweeju, Chole, Jibondo, below mean tide levels) and are relatively Juani, and Mafia Islands) and a number of sheltered from extreme environmental smaller reefs (e.g., Mange and Tutia Reefs) conditions, resulting in species-rich coral- (MPRU 2011:8; TCMP 2001:11–12). Each and sponge-dominated reefs, considerable island has different biological and physical seagrass patches, mangroves, and intertidal characteristics, resulting in diverse contem- flats (MPRU 2011:8, 10). Dense mangrove porary socio-economic structures across forests are found in the northeast of Chole the region (Andersson and Ngazi 1995:477, Bay, as well as fringing Chole Island and the 1998:689), and likely influencing the northwest coast of Juani Island (Lewis et al. configuration of past economic activities. 2011:7; MPRU 2011:11; TCMP 2001:18, For example, the current inhabitants of 20). Chole Island primarily focus on agriculture supplemented by marine resources, while, Contemporary Marine Resource Use and on Jibondo Island, the principal economic Mollusk Foraging resources are fish, octopus, sea cucum- Long-term ecological and marine ber, and marine mollusks. On Juani Island, resource use studies have been under- the fertile calcareous soils promote plant taken around the Mafia Archipelago by growth, with the abundance of timber for the Frontier Tanzania Association, the firewood making the island a favored loca- University of Dar es Salaam, and the MIMP, tion for burning coral for lime production focused primarily on coral reefs and asso- (Andersson and Ngazi 1995:477). ciated fauna (e.g., within Chole Bay and A significant development for marine around Utende) (Gutmann Roberts 2014; ecosystem and resource management in Hemsworth 2015; Jones and Fanning 2010). this region was the establishment of the These investigations and the literature Mafia Island Marine Park (MIMP) in 1995 published on marine resource management (Figure 1). Encompassing ~822 km2, this have been concentrated on coral reefs, fish, area contains the highest marine biodiver- marine mammals and reptiles, octopus, sity in Tanzania and potentially the richest and large crustacea (e.g., Scylla serrata) in eastern Africa (Andersson and Ngazi (see Barnes et al. 2002; Garpe and Öhman 1995:476; Moshy and Bryceson 2016:3). 2003; Guard and Mgaya 2002; Jiddawi

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library Archaeomalacology and Implications for Marine Resource Management 185

Figure 1. Map of the Mafia Archipelago showing locations mentioned in the text and the Mafia Island Marine Park (MIMP) boundary. 1: Juani Primary School; 2: Bweeju Island; 3: Chole Island; 4: Jibondo Island; 5: Mange Reef; 6: Tutia Reef; 7: Utende; 8: Kilindoni; 9: Mfuruni; 10: Ras Kichevi; 11: Kanga; 12: Kirongwe; 13: Ukunju Cave; 14: Kua Ruins.

and Öhman 2002; Kamukuru et al. 2005; Roberts 2014:28–29; Hemsworth 2015:17, Moshy and Bryceson 2016; Muhando and 28; Jones and Fanning 2010:20). While it Rumisha 2008; Mwaipopo 2008). Where has been noted that significant numbers invertebrate density and diversity has been of marine invertebrates occur within the recorded, the focus has largely been on MIMP, and along the east African coast, sea urchins and crown-of-thorns starfish more broadly, few detailed analyses of (Acanthaster planci) as indicator species of marine invertebrates have been undertaken reef health, anemones, octopus, sponges, (Fabricius 2005; MPRU 2011; Richmond shrimp, nudibranchs, and sea cucum- 2001). bers. Bivalves and gastropods have been Based on qualitative surveys under- recorded largely to phylum except for a taken across the Archipelago in 1995, small number of commercially important the most important marine resources in taxa (Fosuah and Steer 2009:38; Gutmann terms of income and/or food supply were

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library 186 Faulkner, Harris, Haji, Crowther, Horton, and Boivin

coral collection, finfish, and octopus, collection across the Mafia region for the with sea cucumber harvested for export ornamental or curio market has largely only (Andersson and Ngazi 1995:479– been viewed as being opportunistic and/ 480; MPRU 2011:13). Marine mollusk or sustainable rather than intensive and collecting and sea cucumber harvesting systematic (e.g., Fosuah and Steer 2009:38; involved similar numbers of people, and Lewis et al. 2011:35). However, this may although the consumptive value was low not have always been the case throughout for mollusks, the productive income gener- the Archipelago (Horrill et al. 1996:54). ated was only lower than that reported for For example, C. tigris were a focal taxon finfish and octopus (Andersson and Ngazi harvested for the marine curio market in the 1995:480; Coughanowr et al. 1995:455). early to mid-2000s and, although a ban on The Mafia Archipelago qualitative surveys collection was in place within the MIMP, also indicated a shift away from some 250 kg of illegally collected cowries were marine resources between 1993 and 1998, impounded in 2006-2007 (Hemsworth with increased consumption of fish and 2015:21; Mwaipopo 2008). agricultural products and the virtual disap- Detailed modern socio-economic pearance of intertidal resources linked to and socio-ecological research on marine high exploitation levels (Andersson and mollusks is unfortunately lacking from the Ngazi 1998:690–691; Christie 2011:102; Mafia Archipelago but some insight into MPRU 2011:13). It has been suggested modern mollusk resource use is available. that the extensive intertidal flats within the Carrying out ethnographic observations and MIMP were once abundant in mollusks qualitative surveys of contemporary marine and other invertebrates and, although there resource use, Christie (2011) collected data is a lack of detailed distribution and density within Chole Bay and around Juani Village, data available (Muhando and Rumisha as well as around Kilindoni and Mfuruni on 2008:65), these resources are now viewed the western side of Mafia Island. A similar as having been overexploited (MPRU approach was taken by Brigden and Steer 2011:11, 19). (2009) around Ras Kichevi, also on the west The impacts of commercial harvest- coast of Mafia Island, producing compara- ing of marine mollusks have been of key ble results to those of Christie (2011). It was concern within the MIMP and along the noted that intertidal harvesting was largely eastern African coast, particularly since undertaken by women and children during many of the commercially collected low tide, with some incidental collection by gastropod taxa are large and slow grow- fishermen (who harvested Cypraeidae, C. ing. Examples of these commercial taxa ramosus, and P. trapezium). Terebralia palus- include Cypraea tigris, Monetaria annulus tris was also collected by men, largely for bait (syn. Cypraea annulus), Cypraeacassis rufa, and/or lime production, although women Chicoreus ramosus, Pleuroploca trape- were also involved in harvesting this taxon zium, and various Conus species, many and C. ramosus for subsistence (Brigden of which make a significant contribu- and Steer 2009:19; Christie 2011:126). The tion to the income of coastal inhabitants main mollusks collected from the intertidal (Muhando and Rumisha 2008:10). A study zone included Anadara antiquata, Isog- on the effects of shell collection on the nomon ephippium, Nerita textilis, Volema abundance of gastropods on the Tanzanian pyrum (syn. V. paradisiaca), and Polinices coast found that commercially valuable mammilla. Other species collected included species were less abundant on shores in Architectonica perspectiva as traditional Dar es Salaam and Zanzibar in comparison medicine (Brigden and Steer 2009:21) and with Mafia Island (Newton et al. 1993:242– M. annulus (largely by children), the latter 243; Nordlund et al. 2010:400). Mollusk also having a significant commercial export

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library Archaeomalacology and Implications for Marine Resource Management 187

value (Christie 2011:127). Large piles of M. (Cardiidae), one terrestrial mollusk (Tropi- annulus were recorded by Brigden and Steer dophora spp.), and five marine gastropods (2009:19–20), having been collected during (Cypraeidae, Neritopsidae, Strombidae, periods of poor weather in an opportunistic Terebridae, and Turbinellidae [syn. Vasi- fashion by adult fishermen. A range of taxa dae]). The lack of detail in the reporting of within the Cypraeidae were also collected these assemblages makes their interpretation from the intertidal zone for the curio trade, difficult, but the range of taxa falls in line with Cassis cornuta, Phalium glaucum, (at higher taxonomic levels) with the domi- and C. rufa collected sporadically for that nant species recorded by other researchers. purpose from deeper water (Brigden and Over 15 mollusk species were recorded by Steer 2009:20). Christie (2013:108) from her excavations at Kua Ruins, with the seven most frequently Previous Archaeomalacological Research occurring taxa quantified and presented in Although shell deposits appear to be detail, including V. pyrum, T. palustris, N. an important component of the coastal textilis, C. ramosus/P. trapezium, C. tigris, archaeological record across the Mafia A. antiquata, and M. annulus. Aimed at Archipelago, few detailed archaeomala- investigating status differentiation, Christie cological analyses have been undertaken. (2013:115) recorded higher proportions of Field research on Mafia Island by Chami mollusks from the lower status units in Area (2000:208–210) and by Wynne-Jones 1 (1325–1451 cal CE) compared with Area (2006:7–8) on Mafia and Chole Islands 2 (1520–1950 cal CE). Christie (2013:115) resulted in the identification of significant cited contemporary informants in inter- shell deposits, although details on the taxo- preting these assemblages, who indicated nomic composition and abundance are not mollusk consumption was preferred only available. Further surveys on Mafia Island when access to higher ranked resources was by Christie (2011:163–164, 166) recorded restricted. shell deposits dominated by A. antiquata, T. palustris, and V. pyrum around Mfuruni, Methods with several more largely comprised of T. palustris recorded near Kirongwe. Based Juani Primary School: Site Description, on this evidence, combined with oral histo- Excavation, and Chronological Phasing ries indicating the significant development Located on the north coast of Juani of mangrove forests in the last hundred Island, the Juani Primary School site is years around Mfuruni, Christie (2011:172, positioned on a flat raised terrace front- 181–186, 2013:116) suggests that mollusk ing a narrow beach, sheltered by dense foraging shifted from a focus on A. anti- mangroves and an expansive tidal flat quata, I. ephippium, and V. pyrum (open (Figure 2). Four 2x2 m trenches were exca- intertidal) to T. palustris (mangroves) due to vated in 2012 in the northern area of the coastal ecosystem shifts. site (JS12-03, JS12-04, JS12-05, JS12-06). Comparatively more information on the These trenches were hand excavated in invertebrate fauna is available from excava- controlled stratigraphic sequence using the tions on Juani Island, particularly Chami’s single context method, with thicker contexts (2000, 2004) investigations at Ukunju Cave, excavated in smaller arbitrary units for Kinunda Cave, and Juani Primary School, greater vertical and chronological control. and Christie’s (2011, 2013) analyses of the All excavated materials were dry-sieved Kua Ruins assemblages. Small numbers of (3 mm) on site, with bulk soil sampling molluskan fauna, identified largely to family (30–60 liters per context) implemented for level, are reported by Chami (2000:212, flotation (0.5 mm) and wet-sieving (1 mm) 2004:88, 95), including one marine bivalve to maximize recovery.

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library 188 Faulkner, Harris, Haji, Crowther, Horton, and Boivin

Figure 2. Juani Primary School location and site map.

The stratigraphic and chronologi- Stone Town, Zanzibar, in 2013 and 2015. cal sequence for Juani has been detailed As physical reference collections were not in Crowther et al. (2016) and, as such, is available in Zanzibar, the material was only briefly outlined here. The identifi- identified to the lowest possible taxonomic able stratigraphy across all four trenches level using published material (e.g., Abbott was similar, with the upper topsoil layer and Dance 1998; Richmond 2011; Robin containing mixed modern and archae- 2008, 2011; Rowson et al. 2010; Spry ological materials representing the 1964, 1968). Those specimens that did not LIA-Post-LIA phase. This upper portion of retain diagnostic features for species level the deposit was followed by a sequence of attribution were assigned to the appropri- reddish-brown sandy silts containing much ate higher taxonomic level (e.g., genus or of the archaeological evidence for occupa- family) or one of the indeterminate catego- tion. Underlying the LIA-Post-LIA contexts ries (Harris et al. 2015; Szabó 2009:186). was a MIA-LIA phase (c. 880–1200 cal CE) All nomenclature has been standardized with a small number of Tana Tradition and with reference to the World Register of Kwale ceramics, overlying the EIA deposits Marine Species (WoRMS Editorial Board (385–540 cal CE) characterized by Kwale 2016). ceramics, a dense midden deposit, and very The following analyses are based small quantities of slag from iron working. on Minimum Number of Individuals Based on the recovery of daub from the EIA (MNI) calculation. Following Harris et and MIA-LIA contexts, it is likely that both al. (2015) and Giovas (2009), MNI was phases represent small-scale occupation in recorded using a range of taxon-specific earth-and-thatch structures. Non-Repetitive Elements (NRE). Bivalve NREs (following siding) included the umbo/ Archaeomalacological Data Collection beak, the anterior and posterior portions of The Juani Primary School inverte- the hinge/dentition, and the anterior and brate fauna were analyzed at the House posterior adductor muscle scars. A range of Wonders Museum (Beit-el-Ajaib) in of gastropod NREs were used depend-

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library Archaeomalacology and Implications for Marine Resource Management 189

ing on shell morphology, including the were also summed as a separate dataset. spire, aperture, aperture lip, posterior and A chi-square test was conducted on each anterior canals, and the umbilicus. The dataset to assess change over time. Cypraeidae base and labum, the Neritidae Sample representativeness was as- columellar deck, and the Turbinidae, Nerit- sessed using nestedness analysis and idae, and Pomatiasidae calcified opercula visual inspection of species area curves. were also included. The Polyplacophora These types of analytical approaches can (chiton) MNI calculation was based on the assist in determining whether taxonomic highest counts of the apex of the anterior composition and characteristics of archae- and mucro of the posterior valves (Giovas ological assemblages are representative of 2009:1560). Those specimens that were past ecological communities (e.g., Lyman taxonomically identifiable but did not 2008; Peacock et al. 2012; Wolverton et retain the necessary NREs for MNI calcu- al. 2015). Nestedness assesses whether lation were noted as being present and an samples represent nested subsets, where MNI of 1 assigned per individual context assemblages with lower richness should nest (Giovas 2009:1562; Jerardino and Marean within those with higher richness, thereby 2010:413). As excavation of the Juani indicating those samples were drawn from deposits was undertaken stratigraphically, the same community. Nestedness measures with each context interpreted as a discrete this relationship via temperature, whereby depositional unit within a broader chrono- 0° represents perfectly nested subsets and logical phase, the MNI was calculated for 100° represents no nestedness (Ulrich each context and summed per chrono- et al. 2009; Wolverton et al. 2015:502). logical phase. To ensure independence of Plotting assemblage richness (NTAXA) relative abundance data, MNI calculations against sample size (MNI) should indicate per context and chronological phase within sampling to redundancy: when no new each trench were based on the NRE appro- taxa are added with increasing sample priate to the highest common taxonomic size, the curve should asymptote, indicat- level. ing that the assemblage is taxonomically Diachronic changes in habitat repre- representative (Lyman 2008). Finally, we sentation were analyzed by assigning all use multiple measures of richness and taxa to one of seven habitat categories (see diversity to further assess the characteristics further discussion below). Non-specific of the molluskan assemblages, with these (i.e., genera, family) identifications were indices calculated using Palaeontological assigned multiple habitat categories to Statistics (PAST) version 3.13 (Hammer et account for inter-family variation in life al. 2001). For these measures, the identified habits (following Harris and Weisler 2017). taxonomic categories are grouped to the Several taxa identified to higher levels were highest common level to ensure indepen- also assigned multiple habitats to represent dence in taxonomic classification (Lyman the different locations in which these taxa 2008). Richness is evaluated through may be encountered by foragers. The MNI NTAXA, with the Shannon index (H) and for each habitat category was summed in Simpson’s index (1-D) used to investigate two ways. First, MNI for the seven distinct assemblage heterogeneity and eveness. habitat categories was summed for each Following Harris and Weisler (2017), phase. This leads to an artificial infla- random permutation tests of relative abun- tion of MNI for the total assemblage due dance were carried out using PAST v3.13 to to the inclusion of some taxa in multiple test for significant differences in each diver- habitats. To account for this, the totals for sity measure by chronological phase. This each habitat category (i.e., treating multiple test uses a randomization procedure based habitat assignments as distinct categories) on pooling the original samples to deter-

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library 190 Faulkner, Harris, Haji, Crowther, Horton, and Boivin

mine whether the differences observed in Given the ambiguity in the origin of the the original data could have occurred by incidental categories, and our focus on random sampling (Hammer and Harper economic resources, these incidental taxa 2007:33–36). are excluded from further analyses. The Juani economic assemblage Results comprises 3279 MNI from 128 taxonomic categories. Here, we define economic Assemblage Structure and Shifts in taxa in an inclusive sense, using the term Taxonomic and Habitat Representation to encompass subsistence species and The total invertebrate assemblage those used for medicine, trade, and/or recovered from Juani Primary School decoration (Faulkner et al. 2018a). With comprised 5095 individuals (MNI) and has the need for more detailed archaeological been attributed to a large number of taxo- and socio-ecological research across the nomic categories (Supplementary Table 1). region, further classification of the mollus- Of the 150 invertebrate taxonomic cate- kan assemblages into dietary, medicinal, gories recorded, 65 are to species level ornamental categories, etc., would be (22 marine bivalves, seven terrestrial, and speculative. We have taken a more conser- 36 marine gastropods at 3490 [68.5%] vative approach to the following analyses in MNI), 46 to genus level (one freshwater, defining economic molluskan taxa, but one four terrestrial, and 25 marine gastropods, that still represents nearshore foraging and one freshwater and 15 marine bivalves at the exploitation of certain habitat zones. 776 [15.2%] MNI), and 33 to family or The 20 dominant taxa (Figure 3A) represent subfamily level (one freshwater and eight 82.8% of the economic sample (2716 MNI), marine bivalves, two terrestrial and 22 with the top four taxa consisting of Nerita marine gastropods at 629 [12.4%] MNI). balteata, Pinctada spp., Atactodea striata, The Decapoda (crab), Cirripedia (barna- and Nerita polita. The 20 sub-dominant cle), and Polyplacophora have been taxa contribute 300 MNI (9.2%) to the recorded to class or infraclass level (at 104 assemblage, and the remaining 88 minor [2.0%] MNI), and there are three inde- taxonomic categories contribute 263 MNI terminate categories (bivalve, gastropod, (8.0%). Of note are those taxa recorded in and landsnail at 96 [1.9%] MNI). Juvenile modern socio-ecological surveys and/or individuals and those exhibiting signs of archaeological assemblages, with contribu- predatory drilling or hermitting have been tions to the economic assemblage ranging quantified separately, but are not included from 0.3% (C. ramosus) to 5.2% MNI (N. in the above overview as distinct taxonomic textilis), although none clearly dominates, categories. These taxa likely represent inci- as would be expected based on previous dental collection (i.e., potential foraging discussions of their economic significance. by-products) or natural incorporations into There are important differences in the the deposit. The Cirripedia, Decapoda, and proportional abundance of the dominant indeterminate categories have also been and sub-dominant taxa at Juani by phase included in the list of incidental taxa as (Figure 3A and 3B), illustrating change high levels of fragmentation and weathering through time in the structure of the forag- preclude their potential economic assessing economy. While decreasing in overall ment. The incidental categories comprise a abundance, several taxa increase propor- significant component of the overall assem- tionally from the EIA to the LIA-Post LIA, blage at 35.6% (1816 MNI), highlighting with the most dramatic increases seen in the importance of taphonomic assessment A. striata (4.3–43.3% MNI) and M. annulus of shell assemblages where the aim is to (1.5–9.5% MNI). Several other taxa, includ- investigate relative shifts in economic use. ing the Lucinidae, Cerithiidae, Monetaria

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library Archaeomalacology and Implications for Marine Resource Management 191

Figure 3. Juani chronological phase dominant (A) and sub-dominant mollusc taxa (B) by %MNI. * indicates taxa previously recorded archaeologically or in socio-ecological surveys.

moneta, Quidnipagus palatum, Dosinia in proportional abundance from the EIA spp., and Gafrarium spp., also exhibit a to the LIA-Post LIA occurs in N. balteata proportional increase through the Juani (19.9–0.5% MNI) and Pinctada spp. (13.6– sequence, albeit at a reduced degree (0.5– 0.1% MNI). The variable contribution of 2.7% increase). Several taxa peak in the the dominant/sub-dominant species at MIA, only to decline into the LIA-Post-LIA Juani through time likely reflects a combi- (Polyplacophora, Cypraeidae, C. tigris, A. nation of shifting foraging strategies (for antiquata, and Conus spp.), and there are both dietary and non-subsistence purposes) 13 taxa that occur in the EIA and/or MIA coupled with changes in near shore marine but disappear entirely into the LIA-Post LIA habitat and resource availability. (e.g., C. ramosus, P. trapezium, and Strom- The representation of habitat categories bidae). Finally, the most significant decline (as illustrated for Chole Bay and adjacent to

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library 192 Faulkner, Harris, Haji, Crowther, Horton, and Boivin

the Juani site in Figure 4A) across the three and peak in the LIA-post LIA, accounting phases parallels these changes in taxonomic for 72.9% MNI. Examination of adjusted abundance. Two general and significant residuals for grouped habitats shows a trends exist in the habitat data; an increase significant increase in sand/mudflat taxa in in the representation of intertidal sand and the LIA-Post LIA, accompanied by a signifi- mud flats, and a decrease in exploitation of cant decrease in coral rubble taxa from the coral rubble substrates (Figures 4B and 4C). MIA-LIA onwards. In the MIA-LIA phase, A chi-squared test indicated a significant there is a significant increase in taxa that association between phase and habitat could have been foraged from seagrass/ category (MNI) (2 (6, n  4035)  113.89, algae, coral reefs, and sand habitats. p  0.001, V  0.12). In both grouped and Finally, while adjusted residuals showed ungrouped habitat summations, taxa found that the change was not significant, all in sand and mudflat environments account habitats aside from sand/mud flat decrease for a minimum 53.5% MNI across all phases in the LIA-Post LIA period.

Figure 4. Habitat distribution in Chole Bay (A); Ungrouped (B) and grouped (C) habitat representation through time at Juani.

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library Archaeomalacology and Implications for Marine Resource Management 193

Shell Density, Assemblage Richness and context by chronological phase, deriving Diversity density estimates (MNI/m3) based on the Each of the economic taxonomic cate- total excavated volume per phase. These gories have been combined to the highest data (Figure 5B) show a clear decline common level for analyses of assemblage through time, from 422.6 MNI/m3 in the EIA density, richness, and diversity. Nested- to be almost halved in the MIA-LIA (221.6 ness analyses (Supplementary Tables 2–4) MNI/m3), with a slight decrease into the also indicate highly nested assemblages, LIA-Post LIA (168.9 MNI/m3). This reduc- with low temperature values for the EIA tion in shell density through the sequence (T  13.98°), MIA-LIA (T  18.25°), and is mirrored by assemblage richness (Figure LIA-Post LIA (T  22.89°). Visual inspection 5B), albeit with a less substantial decrease of the species area curves (Figure 5A) indi- from the EIA (42 NTAXA) to the MIA-LIA cates that each of the phase assemblages (37 NTAXA), but a considerable decline have been sampled to redundancy. Thus, into the LIA-Post LIA (21 NTAXA). This following Wolverton et al. (2015:503), the again reinforces the trends in taxonomic invertebrate assemblages for each chrono- representation throughout sequence, with logical phase are seen to be representative. a clear reduction in the range of taxa Economic shell density is calculated represented through time. In terms of assem- by aggregating the MNI data from each blage diversity (Figure 5B), the Shannon

Figure 5. Juani chronological phase species area curves (A) and economic taxa density, richness, and diversity (B).

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library 194 Faulkner, Harris, Haji, Crowther, Horton, and Boivin

(H) and Simpson’s (1-D) indices show very uted across the range of species exploited similar trends through time but deviate (evidenced by increasing diversity and from those seen in density and richness. evenness). The continued reduction in rich- Both measures increase from the EIA (H  ness in the LIA-Post LIA, combined with 2.30; 1-D  0.79) into the MIA-LIA (H  decreasing assemblage diversity, suggests a 2.79; 1-D  0.91), followed by a reduction further switch in the marine mollusk forag- in both diversity measures into the LIA-Post ing strategy, likely influenced by natural LIA (H  2.02; 1-D  0.76). Paired diversity species availability within the dominant permutation tests report significant differ- marine environments. These changes see ences between the EIA and MIA-LIA (H, corresponding shifts in the habitats being p  0.001; 1-D, p  0.001), and between exploited, with an increasing emphasis on the MIA-LIA and LIA-Post LIA (H, p  0.001; intertidal sand/mudflats and a decrease in 1-D, p  0.001), but not between the EIA the coral rubble/reef zones, with reduc- and the LIA-Post LIA (H, p  0.03; 1-D, tions in the abundance of taxa from all p  0.2093). Diversity/evenness is there- other habitat areas between the MIA-LIA fore similar between the earliest and most and LIA-Post LIA. Taxonomic richness natu- recent chronological phases represented rally increases with substratum rugosity at Juani, even though there is a substantial (e.g., Beck 2000; Kovalenko et al. 2012); decrease in both density and assemblage an associated decrease in species rich- richness. Again, these trends likely reflect a ness would be expected with less foraging combination of habitat structure, resource in coral/rock habitats. This does not fully availability, and the type of intertidal forag- explain the combination of trends in ing strategies implemented through time. density, richness, diversity, and taxonomic composition through time. A further expla- Discussion nation for these trends relates to changes in foraging behavior linked to the exploita- Diachronic Variability in Mollusk tion of near vs. distant habitats through Harvesting time, associated with field processing and The Juani Primary School archaeo- transportation of resources (e.g., Bettinger malacological assemblages demonstrate et al. 1997; Bird et al. 2002; Giovas 2016), significant changes through time, empha- with these possibilities also raised in recent sizing variability in the exploitation of analyses of molluskan assemblages from invertebrate resources in this part of the Zanzibar (Faulkner et al. 2018a, 2018b). Mafia Archipelago. From initial EIA occu- Although further regional archaeological pation through to the recent LIA-Post LIA data are required to evaluate this sugges- period, there is a clear reduction in shell tion, these data reinforce the interpretation density, concomitant with a trend of decreas- of the molluskan assemblage differences ing taxonomic richness. Both measures of between the EIA and MIA-LIA presented heterogeneity indicate that the molluskan in our earlier study (Crowther et al. 2016), assemblage becomes more diverse/even in while also demonstrating that variability in the MIA-LIA compared with the EIA, with the observable trends in density, richness, a subsequent decline into the LIA-Post LIA. and diversity continued into the LIA-Post In the context of shifting taxonomic repre- LIA. sentation between these phases, we see a Although comparatively poorly pre- slight reduction in the breadth of inverte- served (and not available for the LIA-Post brate resources exploited in the MIA-LIA LIA), the vertebrate faunal and archae- but, more importantly, where the foraging obotanical evidence for the EIA and strategy is less focused on a small number MIA-LIA at Juani provides further context of taxa, becoming more evenly distrib- for understanding changes in the inverte-

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library Archaeomalacology and Implications for Marine Resource Management 195

brate assemblages. The abundance of fish terized by expanding dietary breadth and remains also decreases from the EIA (249 diversity within the context of increas- NISP) to the MIA-LIA (43 NISP), possibly ing food production and integration into indicating less intensive fish exploitation broader Indian Ocean trade networks through time. There is, however, similarity in (Crowther et al. 2016; Fleisher et al. 2015). the range of fish taxa exploited, particularly While mollusk foraging was focused on the those from coral and nearshore bay habitat nearshore, intertidal sand/mudflats, and zones, as well as the presence of Chon- coral habitats, the increasing abundance drichthyes (shark, ray) in both the EIA (35 of taxa from all other marine habitat zones NISP) and the MIA-LIA (5 NISP) (Crowther suggests a more generalized, indiscriminate et al. 2016). Domesticated animals are foraging strategy in operation (evidenced by virtually absent from the EIA and MIA-LIA decreasing shell density and higher diver- at Juani, other than a possible caprine sity index values for the MIA-LIA). Although tooth and caprine-sized mammal bone in vertebrate faunal and archaeobotanical the EIA. The tetrapod assemblage is domi- data are not available for the LIA-Post LIA nated by wild taxa (primarily bovids, suni phase, a significant shift in emphasis to [Neotragus moschatus], and duiker), with foraging within sand/mudflats and decreas- green sea turtle (Chelonia mydas) recov- ing exploitation of mollusks from all other ered from the EIA contexts (Crowther et al. habitat zones suggests a further transition- 2016). Importantly, the archaeobotanical ary period at Juani whereby the economic evidence indicates a lack of domesticated importance of mollusks further decreased. crops in the EIA, with the introduction of This interpretation needs to be viewed crops of African origin (cowpea [Vigna tentatively, as spatial reconfiguration of unguiculata], sorghum [Sorghum bicolor], economic activity (e.g., field processing, baobab [Adansonia digitata]) during the differential transportation) would also have MIA. Preliminary anthracological data a bearing on these patterns, as would shifts also highlights the presence of mangrove in subsistence versus trade/ornamental use taxa (Rhizophoraceae) in both the EIA and of shell. It appears, however, that, in this MIA-LIA, likely used for building materials last phase, the relative contribution and and/or fuel (Crowther et al. 2016). the broad economic role of mollusks begin Drawing on the key conclusions to approximate that seen in some of the presented by Crowther et al. (2016), and modern socio-economic observations (e.g., highlighted above relative to the broader Brigden and Steer 2009; Christie 2011)— Juani economic evidence, mollusks appear albeit with considerably higher taxonomic to be a significant staple component of an richness at Juani, forming what appears EIA economy focused on foraging and hunt- to be a more minor and supplementary ing. While a range of habitat areas were economic component. exploited, mollusks were primarily gath- Importantly, Andersson and Ngazi ered from two main environmental zones (1998:687, 692) have made the point (sand/mudflats and coral reef/rubble), that one way to cope with higher levels likely reflecting the higher productivity of economic uncertainty and loss of of these areas to the north of Juani Island environmental predictability in coastal envi- and within Chole Bay. The relative impor- ronments is to diversify production between tance of mollusk resources, in line with the increasingly separate ecosystems. Although fish and tetrapod assemblages, appears to based on contemporary observations, decline in the MIA-LIA with domestic crop this is consistent with the interpretations introduction and increasing agricultural presented by Crowther et al. (2016) for activity. During this period, mollusks form EIA populations dropping domesticates one component of an economy charac- from their economy in favor of hunting

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library 196 Faulkner, Harris, Haji, Crowther, Horton, and Boivin

and foraging to facilitate initial coloniza- McKechnie et al. 2014; Rick and Lockwood tion/occupation, followed by the MIA-LIA 2013). Our archaeomalacological data are adoption of a mixed agricultural-foraging suggestive (rather than being definitive) of lifestyle. Positioning these interpretations shifting baselines on the northern coast within the regional archaeomalacologi- of Juani Island, with important changes in cal evidence is difficult given the lack of shell density, taxonomic composition, rich- comparative detail in previously reported ness, and diversity across the three defined assemblages. Differences in the biological occupation phases at Juani through time. and physical structure of each island in the Significantly, there appears to be a consid- Archipelago and the resulting diversity in erable difference when comparing this socio-economic activity, however, suggests evidence with modern data on mollusk that the same trends occurring through time harvesting, a trend related to observations of and space are unlikely. In fact, variability modern economic transitions in the region in economic structure across eastern Africa (Andersson and Ngazi 1998:686). As noted is becoming apparent with increasingly above, species richness decreases from detailed zooarchaeological analyses (e.g., the EIA to the MIA-LIA (NTAXA  42–37), Douglass et al. 2018; Faulkner et al. 2018a, and again into the LIA-Post LIA (NTAXA  2018b; Prendergast et al. 2016, 2017). At 21), which contrasts with lower richness Juani, we see important differences through recorded in the modern socio-economic time in mollusk foraging and economic surveys on Mafia Island by Brigden and behavior, a process that has implications Steer (2009) (NTAXA  12) and Christie not just for our understanding of past human (2011) (NTAXA  9). Many of the modern behavior, but also for modern conservation taxa appear within the Juani archaeologi- and resource management in the MIMP. cal record (Figure 3), although they occur at a low proportion of the total assemblage Implications for Contemporary Resource (0.3–5.2% MNI) and decrease into the Management LIA-Post LIA (except for M. annulus). Four One potential implication of the Juani taxa appear in the modern species list that Primary School archaeomalacological do not occur archaeologically (A. perspec- data for marine resource management in tiva, C. cornuta, C. rufa, and P. glaucum), the MIMP relates to the “shifting baseline the latter three species linked to modern syndrome,” referring to changes through shell curio trade. This is suggestive of a time in how ecosystem benchmarks are further shift in mollusk foraging from the measured (via empirical biological data) LIA-Post LIA to now, with a focus on a very and/or perceived (at a generational level restricted number of predominantly large or by individuals) relative to previously taxa for subsistence to supplement fishing established reference points (Papworth et and agriculture, and increasing (poten- al. 2009:95). The key issue with a shifting tially opportunistic) harvest of non-dietary baseline is that the established reference species within the modern economy. condition represents only specific recent This is particularly relevant to the Mafia historical points in time, which themselves Island Marine Park, where a key research have arisen from earlier significant changes area highlighted in the General Manage- to the ecosystem (e.g., Klein and Thurstan ment Plan relates to understanding “tradi- 2016; Papworth et al. 2009; Pauly 1995). tional and recent” resource use practices As recognized by numerous researchers, (MPRU 2011:28). Acquiring a longer-term archaeological data can therefore contrib- perspective for the assessment of the ute to the recognition and resolution of sustainability of marine resource use within shifting baselines by providing long-term the MIMP is important given that a range of historical evidence (e.g., Dietl et al. 2015; resources, including mollusks, have been

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library Archaeomalacology and Implications for Marine Resource Management 197

seen to be declining due to destructive commercial harvesting but of relevance for resource use and overexploitation (MPRU the LIA-Post LIA, is that while the number 2011:31–32). The Juani evidence provides of mollusks removed may be comparatively an understanding of historical variability low, the cumulative effects can be consid- in resource use, documenting potential erable when harvesting is systematic and shifting baselines, and contributes to under- repeatedly focused on a small number of standing past and present biodiversity (Dietl species (Hockey and Bosman 1986). Apart et al. 2015:89–91; Peacock et al. 2012). Of from direct effects of human exploitation relevance to these last two points, and what on target species, foraging also likely has we cannot do at this stage with the available indirect effects on other non-exploited data, is disentangle human impacts from taxa and, therefore, the entire population natural processes and provide resolution within certain habitat zones (Crowe et al. on anthropogenic and/or natural drivers 2000:284). In fact, the cumulative effects of behind the long-term economic changes in a range of behaviors during the Iron Age in this area. Most marine ecosystem changes this region may have been significant, a fact are caused by multiple drivers that inter- that is well illustrated via modern observa- act in historically contingent, complex, tions of marine resource use in the Mafia and potentially cumulative ways. As such, Archipelago. our ability to distinguish natural variabil- Modern exploitation pressure on a ity through time from human influence range of marine resources in the MIMP or impacts becomes critical (e.g., Braje et have been seen to have resulted in al. 2017; Giovas et al. 2010; Schwerdtner reduced environmental productivity and Máñez et al. 2014:2–3). biodiversity (MPRU 2011:5). This is a From the data available, it is unlikely complex process, combining increasingly that the diachronic variability identified specialized, intensive, direct, and often in the Juani archaeomalacological dataset destructive exploitation of marine species relates to direct human impacts on mollusk to meet market-driven demand (Anders- populations via predation pressure. This son and Ngazi 1998:692). For example, is particularly the case for the EIA and during the 1980s and 1990s, dynamite the MIA-LIA, where generalized forag- fishing and beach-seine nets caused signif- ing strategies incorporated a range of taxa icant subtidal damage to benthic habitats from different habitat zones, suggesting a (MPRU 2011:13). Live coral mining for minimization of impacts by avoiding an local use and commercial sale, although intensive focus on a single taxon, functional now reduced, causes loss of reef habitat, group, or ecological niche (e.g., Braje and a reduction in fish abundance and diver- Erlandson 2013; Giovas 2016; Harris and sity, and, through the reduction of the reef Weisler 2017). Given that mollusk harvest- hydrodynamic barrier, increases beach, ing appears to decrease in intensity through seagrass bed, and mangrove forest erosion time, and particularly within the LIA-Post (Dulvy et al. 1995:361–363). High levels LIA, resource depression alone is unlikely of mangrove exploitation (for boat build- to be driving the taxonomic and habi- ing, house construction, and firewood in tat shifts into this period. As noted above, lime production from mined coral) also given the significantsocio-economic tran- causes erosion to near-shore marine envi- sitions that occurred from the MIA through ronments, disrupting habitat productivity to the Post LIA, it is more likely that multi- and the reproductive cycle of invertebrate ple drivers, both direct and indirect, were taxa (MPRU 2011:23; TCMP 2001:18, 20). influencing the foraging economy struc- Finally, the collection of octopus, mollusks, ture through time. A key point made by and seaweed in intertidal and shallow Crowe et al. (2000:284), relative to modern subtidal reef flats can cause habitat damage

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library 198 Faulkner, Harris, Haji, Crowther, Horton, and Boivin

via trampling (Crowe et al. 2000:285–286; contemporary management in the MIMP, MPRU 2011:22). Taken together as a set we clearly need higher-resolution archae- of interactive and cumulative processes, ological, modern socio-economic, and the potential degree of ecosystem disrup- ecological data for marine invertebrate tion and reduction in biodiversity becomes resource use, density, and distribution apparent. Several of the processes noted from across the region, coupled with the above would also have been in operation integration of environmental and climatic during Iron Age occupation at Juani, partic- data from modern studies and the archae- ularly mangrove and coral exploitation for ological record. These are important construction materials, but the long-term considerations; as noted by Schwerdtner implications of these factors, in combi- Máñez et al. (2014:4), bioregional histories nation with near-shore foraging behavior need to be developed for specific marine and broader socio-economic shifts, are environments to allow comprehensive unknown. historical baselines to be developed.

Conclusion Acknowledgments Following Contreras (2016), what Fieldwork was funded by an award to we can say, based on the Juani Primary Nicole Boivin from the European Research School archaeological data, is that Council (206148). The 2015 mollusk anal- human-environment relationships (and yses were funded by a University of Sydney particularly human foraging in near-shore FASS Research Support grant to Patrick marine environments) were complex and Faulkner. We thank the Sealinks Project dynamic, and likely linked to a range of team members (Annalisa Christie, Ania mutually influential drivers (both envi- Kotarba-Morley, Nik Petek, and Ruth Tibe- ronmental and behavioral). Building on sasa) and the people of Juani who undertook the archaeological data presented here, the excavations at Juani Primary School. there are initial implications for identifying We thank Catherine Lyakurwa (Division potential shifting baselines. These data add of Antiquities, Tanzania), Amina Issa, and to our understanding of the significance Abdallah K. Ali (Department of Museums of the full spectrum of marine resources and Antiquities, Zanzibar) for facilitating through time and space in the Mafia Archi- this research. The authors thank Dr. Tam pelago, but determining the combination Smith, the editors of the Journal of Ethno- of drivers of change in the past needed to biology, and the two anonymous reviewers facilitate the reconstruction of past ecosys- for their comments on the manuscript. tems is required. While we have provided a baseline understanding of the long-term References Cited socio-economic patterns in marine inver- Abbott, R. T., and S. P. Dance. 1998. Compen- tebrate use, the connection of these data dium of Seashells. Odyssey Publishers, El to modern resource management raises Cajon, California. more questions than answers. Given that Andersson, J. E., and Z. Ngazi. 1995. Marine our data have been generated from a Resource Use and the Establishment of single archaeological sequence, we need a Marine Park: Mafia Island, Tanzania. to further explore how representative these AMBIO 24:475–481. data are of broader archaeological trends Andersson, J. E., and Z. Ngazi. 1998. Coastal across the Archipelago, and how represen- Communities’ Production Choices, Risk tative they may be of potential ecosystem Diversification, and Subsistence Behavior: shifts through time. To enhance our under- Responses in Periods of Transition: A Case standing of these long-term records for Study from Tanzania. AMBIO 27:686–693.

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library Archaeomalacology and Implications for Marine Resource Management 199

Balée, W. 2006. The Research Program of pology, Conservation Biology, and Histor- Historical Ecology. Annual Review of ical Ecology. Evolutionary Anthropology Anthropology 35:75–98. 4:116–121. Balée W., and C. L. Erikson, eds. 2006. Time and Brigden, K., and M. D. Steer, eds. 2009. A Survey Complexity in Historical Ecology: Studies of Biodiversity and Anthropogenic Pres- in the Neotropical Lowlands. Columbia sures Existing on a Large Intertidal Habitat University Press, New York. at Ras Kichevi, Mafia Island, Tanzania. Fron- Barnes, D. K. A., N. K. Dulvy, S. H. Priestley, tier Tanzania Marine Research Programme W. R. T. Darwall, V. Choisel, and M. Whit- Report 124. The Society for Environmental tington. 2002. Fishery Characteristics and Exploration, London, United Kingdom. Abundance Estimates of the Mangrove Chami, F. A. 2000. Further Archaeological Crab Scylla serrata in Southern Tanzania Research on Mafia Island. Azania: Archae- and Northern Mozambique. South African ological Research in Africa 35:208–214. Journal of Marine Science 24:19–25. Chami, F. A. 2004. The Archaeology of the Mafia Beck, M. W. 2000. Separating the Elements of Archipelago, Tanzania. In African Archae- Habitat Structure: Independent Effects of ology Network: Reports and Views, edited Habitat Complexity and Structural Compo- by F. A. Chami, G. Pwiti, and C. Radimi- nents on Rocky Intertidal Gastropods. lahy, pp. 73–101. Dar es Salaam University Journal of Experimental Marine Biology and Press, Dar es Salaam, Tanzania. Ecology 249:29–49. Christie, A. C. 2011. Exploring the Social Bettinger, R. L., R. Malhi, and H. McCarthy. Context of Maritime Exploitation in the 1997. Central Place Models of Acorn and Mafia Archipelago, Tanzania: An Archaeo- Mussel Processing. Journal of Archaeolog- logical Perspective. Unpublished Doctoral ical Science 24:887–899. Dissertation, Department of Archaeology, Bird, D. W., J. L. Richardson, P. M. Veth, and A. University of York, York, England. J. Barham. 2002. Explaining Shellfish Vari- Christie, A. C. 2013. Exploring the Social ability in Middens on the Meriam Islands, Context of Maritime Exploitation in Torres Strait, Australia. Journal of Archaeo- Tanzania between the 14th-18th c. AD: logical Science 29:457–469. Recent Research from the Mafia Archi- Boivin, N., M. A. Zeder, D. Q. Fuller, A. Crowther, pelago. In Prehistoric Marine Resource G. Larson, J. M. Erlandson, T. Denham, and Use in the Indo-Pacific Regions, edited by M. D. Petraglia. 2016. Ecological Conse- R. Ono, A. Morrison, and D. Addison, pp. quences of Human Niche Construction: 97–122. Terra Australis 39. ANU Press, Examining Long-Term Anthropogenic Canberra, Australia. Shaping of Global Species Distributions. Contreras, D. 2016. Correlation Is Not Enough: Proceedings of the National Academy of Building Better Arguments in the Archae- Sciences 113:6388–6396. ology of Human-Environment Interactions. Braje, T. J., and J. M. Erlandson. 2013. Human In The Archaeology of Human-Environ- Acceleration of Animal and Plant Extinc- ment Interactions: Strategies for Investi- tions: A Late Pleistocene, Holocene, and gating Anthropogenic Landscapes, edited Anthropocene Continuum. Anthropocene by D. Contreras, pp. 17–35. Routledge, 4:14–23. New York. Braje, T. J., T. P. Leppard, S. M. Fitzpatrick, and Coughanowr, C. A., M. N. Ngoile, and O. J. M. Erlandson. 2017. Archaeology, Histor- Linden. 1995. Coastal Zone Management ical Ecology and Anthropogenic Island in Eastern Africa Including the Island States: Ecosystems. Environmental Conservation A Review of Issues and Initiatives. AMBIO 44:286–297. 24:448–457. Braje, T. J., and T. C. Rick. 2013. From Forest Crain, C. M., B. S. Halpern, M. W. Beck, and Fires to Fisheries Management: Anthro- C. V. Kappel. 2009. Understanding and

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library 200 Faulkner, Harris, Haji, Crowther, Horton, and Boivin

Managing Human Threats to the Coastal Fabricius, K. 2005. Effects of Terrestrial Runoff Marine Environment. Annals of the New on the Ecology of Corals and Coral Reefs: York Academy of Sciences 1162:39–62. Review and Synthesis. Marine Pollution Crowe, T. P., R. C. Thompson, S. Bray, and S. J. Bulletin 50:125–146. Hawkins. 2000. Impacts of Anthropogenic Faulkner, P., M. Harris, A. K. Ali, O. Haji, A. Stress on Rocky Intertidal Communities. Crowther, M. C. Horton, and N. L. Boivin. Journal of Aquatic Ecosystem Stress and 2018a. Characterising Marine Mollusc Recovery 7:273–297. Exploitation in the Eastern African Iron Crowther, A., P. Faulkner, M. E. Prendergast, Age: Archaeomalacological Evidence from E. M. Quintana Morales, M. Horton, E. Unguja Ukuu and Fukuchani, Zanzibar. Wilmsen, A. M. Kotarba-Morley, A. Christie, Quaternary International 471:66–80. N. Petek, R. Tibesasa, and K. Douka. 2016. Faulkner, P., M. Harris, O. Haji, A. K. Ali, A. Coastal Subsistence, Maritime Trade, and Crowther, C. Shipton, M. C. Horton, and the Colonization of Small Offshore Islands N. L. Boivin. 2018b. Long-term Trends in in Eastern African Prehistory. The Journal of Terrestrial and Marine Invertebrate Exploita- Island and Coastal Archaeology 11:211– tion on the Eastern African Coast: Insights 237. from Kuumbi Cave, Zanzibar. Journal of Crumley, C. L., ed. 1994. Historical Ecology: Island and Coastal Archaeology. [online]. Cultural Knowledge and Changing Land- https://doi.org/10.1080/15564894.2018.1 scapes. School of American Research Press, 501442. Santa Fe, New Mexico. Fitzpatrick, S. M., and W. F. Keegan. 2007. Dietl, G. P., S. M. Kidwell, M. Brenner, D. A. Human Impacts and Adaptations in the Burney, K. W. Flessa, S. T. Jackson, and P. Caribbean Islands: An Historical Ecology L. Koch. 2015. Conservation Paleobiology: Approach. Earth and Environmental Leveraging Knowledge of the Past to Inform Science Transactions of the Royal Society of Conservation and Restoration. Annual Edinburgh 98:29–45. Review of Earth and Planetary Sciences Fleisher, J., P. Lane, A. LaViolette, M. Horton, E. 43:79–103. Pollard, E. Quintana Morales, T. Vernet, A. Douglass, K. 2017. The Diversity of Late Holo- Christie, and S. Wynne-Jones. 2015. When cene Shellfish Exploitation in Velondriake, Did the Swahili become Maritime? Amer- Southwest Madagascar. The Journal of ican Anthropologist 117:100–115. Island and Coastal Archaeology 12:333– Fosuah, E. N., and M. D. Steer, eds. 2009. An 359. Assessment of the Coral Reefs on Western Douglass, K., A. R. Antonites, E. M. Quintana Mafia Island, Tanzania. Frontier Tanzania Morales, A. Grealy, M. Bunce, C. Bruwer, Marine Research Programme Report 125. and C. Gough. 2018. Multi-Analytical The Society for Environmental Exploration, Approach to Zooarchaeological Assem- London, United Kingdom. blages Elucidates Late Holocene Coastal Garpe, K. C., and M. C. Öhman. 2003. Coral Lifeways in Southwest Madagascar. Quater- and Fish Distribution Patterns in Mafia nary International 471:111–131. Island Marine Park, Tanzania: Fish–Habitat Douglass, K., J. Walz, E. Quintana Morales, R. Interactions. Hydrobiologia 498:191–211. Marcus, G. Myers, and J. Pollini. 2019. Giovas, C. M. 2009. The Shell Game: Analytic Historical Perspectives on Contemporary Problems in Archaeological Mollusc Quan- Human-Environment Dynamics in South- tification.Journal of Archaeological Science east Africa. Conservation Biology 33:260– 36:1557–1564. 274. DOI:10.1111/cobi.13244. Giovas, C. M. 2016. Though She Be but Little: Dulvy, N. K., D. Stanwell-Smith, W. R. T. Darwall, Resource Resilience, Amerindian Foraging, and C. J. Horrill. 1995. Coral Mining at and Long-Term Adaptive Strategies in the Mafia Island, Tanzania: A Management Grenadines, West Indies. The Journal of Dilemma. AMBIO 24:358–365. Island and Coastal Archaeology 11:238– 263.

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library Archaeomalacology and Implications for Marine Resource Management 201

Giovas, C. M., S. M. Fitzpatrick, M. Clark, and Overfishing and the Recent Collapse of M. Abed. 2010. Evidence for Size Increase Coastal Ecosystems. Science 293:629–637. in an Exploited Mollusc: Humped Conch Jerardino, A. 2016. On the Origins and Signifi- (Strombus gibberulus) at Chelechol ra cance of Pleistocene Coastal Resource use Orrak, Palau from ca. 3000–0 BP. Journal in Southern Africa with Particular Reference of Archaeological Science 37:2788–2798. to Shellfish Gathering. Journal of Anthropo- Guard, M., and Y. D. Mgaya. 2002. The Artis- logical Archaeology 41:213–230. anal Fishery for Octopus cyanea Gray in Jerardino, A., and C. W. Marean. 2010. Shell- Tanzania. AMBIO 31:528–536. fish Gathering, Marine Paleoecology Gutmann Roberts, C. 2014. Frontier Tanzania and Modern Human Behavior: Perspec- Marine Research Programme: Utende, tives from Cave PP13B, Pinnacle Point, Mafia Island, Tanzania. 24th March – 16th South Africa. Journal of Human Evolution June 2014. TZM Phase 142 Science Report. 59:412–424. London, United Kingdom. Jiddawi, N. S., and M. C. Öhman. 2002. Marine Hammer, Ø., and D. A. T. Harper. 2007. Pale- Fisheries in Tanzania. AMBIO 31:518–527. ontological Data Analysis. Blackwell Jones, L., and E. Fanning, eds. 2010. Assessing Publishing, Malden, MA. the Health of Mafia Island’s Marine Ecosys- Hammer, Ø., D. A. T. Harper, and P. D. Ryan. tems. Frontier Tanzania Marine Research 2001. PAST: Paleontological Statistics Soft- Programme Report 129. The Society for ware Package for Education and Data Anal- Environmental Exploration, London, United ysis. Palaeontologia Electronica 4:1–9. Kingdom. Harris, M., and M. Weisler. 2017. Intertidal Kamukuru, A. T., T. Hecht, and Y. D. Mgaya. Foraging on Atolls: Prehistoric Forager 2005. Effects of Exploitation on Age, Decision-Making at Ebon Atoll, Marshall Growth and Mortality of the Blackspot Islands. The Journal of Island and Coastal Snapper, Lutjanus fulviflamma, at Mafia Archaeology 12:200–223. Island, Tanzania. Fisheries Management Harris, M., M. Weisler, and P. Faulkner. 2015. and Ecology 12:45–55. A Refined Protocol for Calculating MNI Klein, E. S., and R. H. Thurstan. 2016. Acknowl- in Archaeological Molluscan Shell Assem- edging Long-Term Ecological Change: blages: A Marshall Islands Case Study. The Problem of Shifting Baselines. In Journal of Archaeological Science 57:168– Perspectives on Oceans Past, edited by K. 179. Schwerdtner Máñez and B. Poulsen, pp. Hemsworth, R. 2015. Frontier Tanzania Marine 11–29. Springer, Dordrecht, The Nether- Research Programme: Utende, Mafia lands. Island, Tanzania. 15th December 2014 – Kovalenko, K. E., S. M. Thomaz, and D. M. Warfe. 16th March 2015. TZM Phase 151 Science 2012. Habitat Complexity: Approaches and Report. London, United Kingdom. Future Directions. Hydrobiologia 685:1–17. Hockey, P. A. R., and A. L. Bosman. 1986. Kyriacou, K., J. E. Parkington, M. Will, A. W. Man as an Intertidal Predator in Transkei: Kandel, and N. J. Conard. 2015. Middle Disturbance, Community Convergence and and Later Stone Age Shellfish Exploitation Management of a Natural Food Resource. Strategies and Coastal Foraging at Hoed- Oikos 46:3–14. jiespunt and Lynch Point, Saldanha Bay, Horrill, J. C., W. R. Darwall, and M. Ngoile. South Africa. Journal of Archaeological 1996. Development of a Marine Protected Science 57:197–206. Area: Mafia Island, Tanzania.AMBIO Lewis, E., B. Margerison, B. Sobkowiak, and E. 25:50–57. Fanning, eds. 2011. A Biophysical Survey Jackson, J. B., M. X. Kirby, W. H. Berger, K. A. of Three New Marine Reserves in Mafia Bjorndal, L. W. Botsford, B. J. Bourque, R. Island, Tanzania. Frontier Tanzania Marine H. Bradbury, R. Cooke, J. Erlandson, J. A. Research Programme Report 128. The Estes, and T. P. Hughes. 2001. Historical Society for Environmental Exploration, London, United Kingdom.

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library 202 Faulkner, Harris, Haji, Crowther, Horton, and Boivin

Lyman, R. L. 2008. Quantitative Paleozoology. Shifting Baseline Syndrome in Conserva- Cambridge University Press, Cambridge, tion. Conservation Letters 2:93–100. England. Pauly, D. 1995. Anecdotes and the Shifting McKechnie, I., D. Lepofsky, M. L. Moss, V. L. Baseline Syndrome of Fisheries. Trends in Butler, T. J. Orchard, G. Coupland, F. Foster, Ecology and Evolution 10:430. M. Caldwell, and K. Lertzman. 2014. Peacock E., C. R. Randklev, S. Wolverton, R. A. Archaeological Data Provide Alternative Palmer, and S. Zaleski. 2012. The “Cultural Hypotheses on Pacific Herring (Clupea Filter,” Human Transport of Mussel Shell, pallasii) Distribution, Abundance, and and the Applied Potential of Zooarchae- Variability. Proceedings of the National ological Data. Ecological Applications Academy of Sciences 111:E807–E816. 22:1446–1459. Moshy, V. H., and I. Bryceson. 2016. Seeing Prendergast, M. E., E. M. Quintana Morales, A. Through Fishers’ Lenses: Exploring Marine Crowther, M. C. Horton, and N. L. Boivin. Ecological Changes within Mafia Island 2017. Dietary Diversity on the Swahili Marine Park, Tanzania. SAGE Open 6:1–18. Coast: The Fauna from Two Zanzibar MPRU (Marine Parks and Reserves Unit). 2011. Trading Locales. International Journal of Mafia Island Marine Park General Manage- Osteoarchaeology 27:621–637. ment Plan. United Republic of Tanzania, Prendergast, M., H. Rouby, P. Punnwong, R. Ministry of Livestock and Fisheries Devel- Marchant, A. Crowther, N. Kourampas, C. opment. Available at: marineparks.go.tz/ Shipton, M. Walsh, K. Lambeck, and N. L. reports/MIMP%20GMP.pdf. Accessed on Boivin. 2016. Continental Island Forma- May 7, 2019. tion and the Archaeology of Defaunation Msemwa, P. J. 1994. An Ethnoarchaeological on Zanzibar, Eastern Africa. PLoS ONE Study on Shellfish Collecting in a Complex 11:e0149565. Urban Setting. Unpublished Doctoral Richmond, M. D. 2001. The Marine Biodiver- Dissertation, Department of Anthropology, sity of the Western Indian Ocean and its Brown University, Providence, RI. Biogeography: How Much do We Know? In Muhando, C. A., and C. K. Rumisha. 2008. Marine Science Development in Tanzania Distribution and Status of Coastal Habitats and Eastern Africa. Proceedings of the 20th and Resources in Tanzania. WWF-Tan- Anniversary Conference on Advances in zania, Dar es Salaam, Tanzania. Marine Science in Tanzania, 28 June – 1 Mwaipopo, R. N. 2008. The Social Dimensions July 1999, edited by M. D. Richmond and of Marine Protected Areas: A Case Study of J. Francis, pp. 241–261. IMS/WIOMSA, the Mafia Island Marine Park in Tanzania. Zanzibar, Tanzania. Samudra Monograph. International Collec- Richmond, M. D. 2011. A Field Guide to the tive in Support of Fishworkers. Chennai, Seashores of Eastern Africa and the Western India. Indian Ocean Islands. SIDA, WIOMSA, Dar Newton, L. C., E. V. H. Parkes, and R. C. es Salaam, Tanzania. Thompson. 1993. The Effects of Shell Rick, T. C., and R. Lockwood. 2013. Integrating Collecting on the Abundance of Gastro- Paleobiology, Archeology, and History to pods on Tanzanian Shores. Biological Inform Biological Conservation. Conserva- Conservation 63:241–245. tion Biology 27:45–54. Nordlund, L., J. Erlandsson, M. de la Torre- Robin, A. 2008. Encyclopedia of Marine Castro, and N. Jiddawi. 2010. Changes in Gastropods. ConchBooks, Hackenheim, an East African Social-Ecological Seagrass Germany. System: Invertebrate Harvesting Affecting Robin, A. 2011. Encyclopedia of Marine Species Composition and Local Livelihood. Bivalves, Including Scaphopods, Polypla- Aquatic Living Resources 23:399–416. cophora and Cephalopods. ConchBooks, Papworth, S. K., J. Rist, L. Coad. and E. J. Hackenheim, Germany. Milner-Gulland. 2009. Evidence for

Journal of Ethnobiology 2019 39(2): 182–203 Downloaded From: https://bioone.org/journals/Journal-of-Ethnobiology on 23 Jun 2019 Terms of Use: https://bioone.org/terms-of-use Access provided by Max Planck Digital Library Archaeomalacology and Implications for Marine Resource Management 203

Rowson, B., B. H. Warren, and C. F. Ngereza. TCMP (Tanzania Coastal Management Partner- 2010. Terrestrial Molluscs of Pemba Island, ship). 2003. Tanzania State of the Coast Zanzibar, Tanzania, and its Status as an Report 2003: The National ICM Strategy and “Oceanic” Island. ZooKeys 70:1–39. Prospects for Poverty Reduction. Coastal Schwerdtner Máñez, K., P. Holm, L. Blight, Management Report # 2002 TCMP, Dar M. Coll, A. MacDiarmid, H. Ojaveer, B. es Salaam, Tanzania. [online] URL: https:// Poulsen, and M. Tull. 2014. The Future of www.oceandocs.org/handle/1834/796. the Oceans Past: Towards a Global Marine Ulrich, W., M. Almeida-Neto, and N. J. Gotelli. Historical Research Initiative. PLoS ONE 2009. A Consumer’s Guide to Nestedness 9:p.e101466. Analysis. Oikos 118:3–17. Spry, J. F. 1964. The Sea Shells of Dar es Salaam: Wolverton, S., L. Nagaoka, and T. C. Rick. 2016. Part II Pelecypoda (Bivalves). Reprinted Applied Zooarchaeology: Five Case Studies. from Tanganyika Notes and Records 63. Eliot Werner Publications, New York. Tanzania Society, Dar es Salaam. Wolverton, S., C. Otaola, G. Neme, M. Giar- Spry, J. F. 1968. The Sea Shells of Dar es Salaam. dina, and A. Gil. 2015. Patch Choice, Land- Part I. Gastropods. Reprinted from Tang- scape Ecology, and Foraging Efficiency: The anyika Notes and Records 56. Tanzania Zooarchaeology of Late Holocene Foragers Society, Dar es Salaam. in Western Argentina. Journal of Ethnobi- Szabó, K. 2009. Molluscan Remains from Fiji. ology 35:499–518. In The Early Prehistory of Fiji, edited by WoRMS Editorial Board. 2016. World Register G. Clark and A. Anderson, pp. 183–211. of Marine Species [website]. Available at: Terra Australis 31. ANU EPress, Canberra, http://www.marinespecies.org at VLIZ. Australia. Wynne-Jones, S. 2006. Reconstructing the TCMP (Tanzania Coastal Management Partner- Long-Term History of the Mafia Archi- ship). 2001. Tanzania State of the Coast pelago. Unpublished report submitted to 2001: People and the Environment. Working COSTECH and the Antiquities Department, Document 5059 TCMP, Dar es Salaam, Ministry of Natural Resources and Tourism, Tanzania. [online] URL: https://www.crc. Tanzania. uri.edu/download/TZ_SOC_2001.pdf.