Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=raza20
Azania: Archaeological Research in Africa
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/raza20
Local animal economies during the nineteenth-
century caravan trade along the Lower Pangani,
northeastern Tanzania: a zooarchaeological
perspective
Thomas J. Biginagwa & Paul J. Lane
To cite this article: Thomas J. Biginagwa & Paul J. Lane (2021): Local animal economies
during the nineteenth-century caravan trade along the Lower Pangani, northeastern
Tanzania: a zooarchaeological perspective, Azania: Archaeological Research in Africa, DOI:
10.1080/0067270X.2021.1925023
To link to this article:  https://doi.org/10.1080/0067270X.2021.1925023
© 2021 The Author(s). Published by Informa
UK Limited, trading as Taylor & Francis
Group
Published online: 08 Jul 2021.
Submit your article to this journal 
View related articles 
View Crossmark data
Local animal economies during the nineteenth-century
caravan trade along the Lower Pangani, northeastern
Tanzania: a zooarchaeological perspective
Thomas J. Biginagwa 1 and Paul J. Lane 2
1Department of Archaeology and Heritage Studies, University of Dar es Salaam, Dar es Salaam, Tanzania;
2Department of Archaeology, University of Cambridge, Cambridge, United Kingdom and School of
Geography, Archaeology and Environmental Studies, University of the Witwatersrand, Johannesburg, South
Africa.
ABSTRACT
The expansion of the caravan trade in eastern Africa during the
nineteenth century is considered to have had significant
ecological, economic and social consequences. While available
historical documentary and oral sources provide valuable
evidence concerning the scale, timing and spatial extent of these,
as well as information about some of the key actors and agents,
there remain significant gaps that have the potential to be filled
by targeted archaeological research. This paper presents one such
study, which aims to establish how influential the expansion of
the caravan trade was on local animal economies, with particular
reference to a sample of known caravan halts on the northern
route on the Pangani River, Tanzania. The results of
zooarchaeological analysis of faunal assemblages recovered from
four sites suggest that the impacts may have been less than has
often been argued by some historians. The study also provides
fresh insight on the continuing importance of wild resources,
especially rodents, in local diets in the late nineteenth century
and on local herd management strategies.
RÉSUMÉ
L’expansion du commerce caravanier en Afrique de l’Est au cours
du dix-neuvième siècle est considérée comme ayant eu des
conséquences écologiques, économiques et sociales importantes.
Les sources historiques documentaires et orales disponibles
fournissent des données précieuses concernant l’échelle, le
moment et l’étendue spatiale de ces changements, ainsi que des
informations sur certains des acteurs et agents clés. Il reste
néanmoins des lacunes importantes qui peuvent être comblées
par des recherches archéologiques ciblées. Cet article présente
une telle étude, qui vise à établir l’influence sur les économies
animales locales de l’expansion du commerce caravanier, faisant
particulièrement référence à certaines étapes caravanières
connues sur la route septentrionale le long du fleuve Pangani, en
Tanzanie. Les résultats de l’analyse zooarchéologique des
ARTICLE HISTORY
Received 3 September 2020
Accepted 7 January 2021
KEYWORDS
Lower Pangani; historical
archaeology; caravan trade;
zooarchaeology; Tanzania
© 2021 The Author(s). Published by Informa UK Limited, trading as Taylor & Francis Group
This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License
(http://creativecommons.org/licenses/by-nc-nd/4.0/), which permits non-commercial re-use, distribution, and reproduction in any
medium, provided the original work is properly cited, and is not altered, transformed, or built upon in any way.
CONTACT Thomas J. Biginagwa biginagwa.thomas@udsm.ac.tz
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA
https://doi.org/10.1080/0067270X.2021.1925023
assemblages fauniques obtenus sur quatre sites suggèrent que les
impacts ont pu être moindres que ce qui a parfois été avancé par
certains historiens. L’étude fournit également un nouvel aperçu
sur l’importance continue des ressources sauvages, en particulier
des rongeurs, dans les régimes alimentaires locaux à la fin du dix-
neuvième siècle, et sur les stratégies locales de gestion des
troupeaux.
Introduction
The consequences of the expansion of the caravan trade in ivory and slaves in East Africa
during the nineteenth century are widely regarded to have been considerable (Beachey
1967; Abir 1968; Rodney 1972; Sheriff 1987; Koponen 1988; Kjekshus 1996). Various
legacies of these consequences have received significant scholarly attention, especially
the transformations in the organisation of labour (Rempel 1998; Rockel 2006; Pallaver
2012), gender relations (Robertson 1997), production strategies (Beidelman 1982;
Håkansson 2008), settlement (Kusimba 2004) and consumption patterns (Prestholdt
2004) triggered by the trade. Several studies have also been undertaken on its impacts
on the region’s elephant populations (Thorbahn 1979; Håkansson 2004; Coutu et al.
2016) and possible changes to local environments (Håkansson et al. 2008; Stump and
Tagseth 2009). Here, we extend this discussion by examining zooarchaeological evidence
from four former caravan halts to assess the effects on local animal economies. We
examine the argument that local and regional food production strategies changed in
response to demands from the passing caravans (Kimambo 1996), especially as the
number and size of caravans increased from c. 1840 onward (Koponen 1988).
Thomas Håkansson (1995: 303–304), in particular, has argued that the production of
surplus food was most likely generated by both agricultural intensification — including
the adoption of new farming strategies notably terracing, irrigation and manuring— and
the expansion of farmland often onto marginal areas with low fertility and greater vul-
nerability to soil erosion (Håkansson 2004: 585–586). Håkansson (2004: 577–578) has
also argued, on the basis of the various references to the exchange of ivory for cattle
in the available historical sources, that specialised pastoralism may have expanded as a
consequence of the stimulus that the ivory trade provided to increase herd sizes. Other
scholars have further suggested that agricultural intensification and/or intensification
of livestock herding during the nineteenth century initiated or possibly exacerbated
soil erosion in certain areas, including along sections of the central caravan route in
Ugogo and around Kondoa (Christianson 1981; Sissons 1984).
While the nineteenth century was undoubtedly a period of significant change, there
are several significant gaps in the written and oral historical records in terms of detail,
chronological precision and the environmental and social consequences of different
events (Lane 2010), and it is unlikely that the impacts were entirely uniform across the
entire region (Kjekshus 1996). In an effort to provide a finer-scaled analysis and better
understanding of local responses, this paper presents the results of a zooarchaeological
study of the animal economies of local communities in the Lower Pangani River
Basin, northeastern Tanzania, prior to and during the expansion of the caravan trade
in the second half of the nineteenth century. Faunal remains retrieved from four
2 T. J. BIGINAGWA AND P. J. LANE
settlements known from written accounts and early cartographic sources to have served
as regular caravan halts were analysed to determine possible changes in the exploitation
patterns of domestic and wild animals as the caravan trade expanded and whether any of
those changes that could be discerned might indicate efforts to increase meat production.
The following four sections describe the geographical and historical contexts of the sites,
the excavation methods used and the approaches adopted for the analysis of the faunal
remains recovered. The results of these analyses are then discussed with reference to
various hypotheses concerning the impacts of the caravan trade on local livelihood strat-
egies. The paper concludes by placing these results in a wider regional setting with some
recommendations for further research.
Geographical and historical context
The four sites studied were Ngombezi (05°10'01"S, 38°25'01"E), Maurui (05°08'095"S, 38°
23'23"E), Old Korogwe (05°09'25"S, 38°28'38"E) and Kwa Sigi (05°08'08"S, 38°23'58"E).
These are all abandoned riparian settlements located along the Pangani River in the
current administrative boundaries of Korogwe District, some 110–118 km inland from
the Indian Ocean (Figure 1a). The area in which the sites are located lies within the
Lower Pangani River Basin (hereafter LPRB), covering the low-lying plain of
the Pangani Basin. The LPRB stretches along the western and southern foothills of the
West Usambara Mountains and drops gently southeastward towards the Indian Ocean
(IUCN 2003). Currently, the semi-arid lowland areas receive between 300 and 800 mm
per annum, while the high altitude and windward portions of the adjacent Usambara
Mountains can receive between 2500 and 3000 mm (IUCN 2003; Kiptala et al. 2013).
The lowland parts of the LPRB experience warm temperatures throughout the year,
reaching 38°C during the dry season (URT 2008).
The climate of the LPRB during the nineteenth century is unlikely to have been sub-
stantially different from that of today, although palaeoecological research has identified
several periods of significantly reduced rainfall leading to extended periods of drought
across eastern Africa during the eighteenth and nineteenth centuries (Bessems et al.
2008) that may have impacted communities within the LPRB. Under such conditions,
settlements located in proximity to the Pangani River, such as those investigated for
this project, may have been better placed to weather climate-induced stress, and may
even have had a strategic advantage over those located farther away from permanent
water sources. On the other hand, as riparian settlements, they would have been more
vulnerable to the effects of extreme flooding, although at none of the excavated sites
were any geological or stratigraphic markers of flood events encountered, suggesting
that this was a low environmental risk during the time of their occupancy.
Historically, the LPRB has been referred to as ‘Zigualand’ or ‘Usegua’ because its main
occupants were and still are the Zigua (also spelled Segua, Zegura and Zegula in early
sources). The Zigua who inhabited the island settlements investigated for this study,
and those currently settled along the Pangani River, called themselves ‘Wazigua wa
Ruvu’1 (the Zigua of the Ruvu River or the Pangani River), referring to their livelihood
along that river (Baumann 1891). In his account of Count Teleki’s expedition to Lake
Turkana, which passed through the Pangani Valley in 1887, Von Höhnel noted, for
example, that the ‘Wasegua on the Pangani are also sometimes spoken of as the
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 3
Waruvu, or dwellers on the river’ and that ‘most of their villages are on islands’, among
them the large village of Kwa Mgumi (Von Höhnel 1894: 61). This contrasts with Zigua
living inland away from the Pangani River, including those in Handeni District.
The Austrian explorer-geographer Oscar Baumann, who also visited the area in the
late nineteenth century, recorded some key features of late nineteenth-century Zigua vil-
lages. The island villages on the Pangani River were of considerable size, with ten to 200
‘huts’ (Figure 1b) and were ‘linked with the outside world by a footbridge known as
Figure 1. a) The Pangani River Basin in its wider geographical contexts, major topographical features,
study sites and other places mentioned in the text (redrawn and modified from Soper 1967: 20); b) a
nineteenth-century view of the island of Old Korogwe (after Hans Meyer cited in Bauman 1890).
4 T. J. BIGINAGWA AND P. J. LANE
uraro’ (Baumann 1891: 176). Away from the river, most villages were fortified by a pali-
sade or a thick euphorbia hedge. In 1857, for example, the explorers Richard Burton and
James Speke stopped at Korogwe, or Kohode as they named it, describing the settlement
as a ‘normal cultivator’s village’ … ‘surrounded and concealed by a stout palisade of tree-
trunks…with low arches formed by inclining the beams’. Inside the palisade were a
cluster of ‘little, thatched, wattle-and-daub-huts’, some round, some square, scattered
over ‘some hundred yards’, with goats, sheep and cattle all stalled nearby (Burton and
Speke 1858: 210). Their accounts and those of other European explorers also suggest
that fortification became increasingly common during the second half of the nineteenth
century when there was an upsurge in inter- and intra-community raiding and heigh-
tened competition partly associated with an increased demand for slaves (Lane 2011:
289–301).
Diverse historical sources indicate that the nineteenth-century Zigua subsisted by cul-
tivating crops, animal husbandry, hunting and fishing, augmenting their livelihoods
through trade with neighbouring communities, coastal towns such as Pangani and
passing caravans (Krapf 1968 [1860]; Farler 1882). They farmed mainly along the
banks of the Pangani River and in the valleys (Krapf 1968 [1860]). Maize and
sorghum were being cultivated at the time of Baumann’s visit, with beans, cassava,
sesame, rice and sweet potatoes among the other crops mentioned by different European
observers. Fishing and livestock-keeping were also subsistence mainstays (Burton 1872).
Baumann (1891) notes that mammalian livestock were taken daily across the Pangani
River to graze outside the islands and returned home in the evening, noting that a mix
of long-horn and short-horn breeds were kept, as well as fairly large herds of goats
but only the occasional sheep.
The accounts by nineteenth-century explorers and missionaries and their accompany-
ing maps (e.g. Burton and Speke 1858; Johnston 1879; Baumann 1891) indicate that the
island settlements served as regular halts during the nineteenth century, provisioning
passing caravans and providing security in exchange for imported commodities. Some
of these, notably cotton cloth, beads and brass wire, were valued as currency (Burton
and Speke 1858: 218; Farler 1882: 741). Customarily, as the caravans approached the vil-
lages, the caravan traders fired guns to announce their arrival and within a short time the
villagers would bring foodstuff and other local products to barter (Burton and Speke
1858; Farler 1882).
Regional, systematic archaeological surveys and targeted excavation have documented
multiple sites in the LPRB, some dating to c. AD 900–1200 (Walz 2010). The evidence
points to the existence of flourishing trade connections between the coast and the interior
well before the nineteenth century, as evidenced by the presence of Indo-Pacific glass
beads and bivalve marine shells on several sites along the Mkomazi Valley, some
110 km north of Korogwe. As Walz (2010) has argued, the nineteenth-century caravan
routes clearly emerged from pre-existing regional trading networks, although the
routes may not have been identical. A case in point is the settlement evidence from
the four sites that are the subject of this study, which were established between the thir-
teenth (Maurui) and seventeenth (Ngombezi) centuries AD and were occupied up to the
mid-twentieth century when they were abandoned (Biginagwa and Ichumbaki 2018).
This contradicts some claims in the historical sources (Erhardt 1853: 30; New 1874:
318; Johnston 1879) that the islands were only first settled during the nineteenth
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 5
century to serve as places of refuge for Zigua fromMaa-speaking cattle raiders. However,
despite their earlier establishment, the stratigraphic distribution of local shell beads and
imported glass beads at these sites (Figure 2), particularly the general lack of Indo-Pacific
glass beads in the lower excavation levels (Biginagwa 2012), suggests that these sites may
have been less connected with the coast than those in the Mkomazi Valley investigated by
Walz (2010).
Moreover, when considered along with radiocarbon dates for Ngombezi and Maurui,
which have the longest stratigraphic sequences, the change in the stratigraphic distri-
bution of different bead types and the abrupt appearance of European glass bead types
provide an effective means of differentiating archaeological deposits formed prior to
the caravan trade from those associated with the caravan trade. Thus, a comparison of
the faunal remains recovered from the layers associated with both phases should, in prin-
ciple, provide a means of testing the hypothesis that expansion of the caravan trade trig-
gered a transformation of local animal economies as inhabitants of these settlements
sought to intensify meat production for exchange with the passing caravans.
Excavation strategy
Fieldwork took place between 2008 and 2013. All excavations targeted cultural mounds,
except at Kwa Sigi, which lacks such features. Deposits were excavated by hand using
trowels, according to context subdivided into arbitrary 10-cm-thick spits to monitor vari-
ation in finds density with depth. All deposits were dry sieved, predominantly through a
5-mm-wide wire mesh, although a 2.5 mm mesh was used when the deposits were par-
ticularly fine to capture smaller finds.
At Ngombezi (the main study site), a 14 × 2 m trench was dug sectioning one of the
largest mounds at the site (Figure 3a), reaching a depth of 3.5 m at the highest point
of the mound before the natural horizon was reached, and 2.1 m at the edge of the
Figure 2. Ngombezi; stratigraphic distribution of glass and non-glass beads.
6 T. J. BIGINAGWA AND P. J. LANE
mound. Thirteen layers were revealed (Figure 3b), representing two distinct phases of
settlement activity indicated by a sequence of wattle-and-daub houses built in the
initial occupation phase that were subsequently demolished and a further phase of
house construction (Figure 3c). A charcoal sample (Wk-25718) from the base of the
anthropogenic deposits at 3.30 m below mound surface provided a radiocarbon date
of 238 ± 30 BP (1641–1810 cal. AD, using the SHCal20 curve), suggesting that the earliest
structures on this island date to between the mid-seventeenth and eighteenth centuries.
At Maurui, a 2 × 2 m trench was dug at the centre of a prominent cultural mound to a
depth of 5.0 m below ground level. Ten distinct layers of deposits were observed.
Although no structural remains were exposed, the large quantity of daub recovered indi-
cates the presence of built structures on the island. A charcoal sample (Poz-59494) from
the base of the anthropogenic deposits (500 cm) yielded a radiocarbon date of 630 ± 30
BP (1304–1416 cal. AD, using the SHCal20 curve), suggesting initial occupation as early
as the fourteenth century. At Old Korogwe, a 2 × 2 m trench was excavated near the
Figure 3. Ngombezi: a) a 2 x 14 m trench in an early stage of excavation (Biginagwa 2008); b) com-
posite section of the south-facing wall of the excavation trench; c) dense layers of house daub (below
dotted red lines) projecting from the walls of the excavation trench indicative of settlement phases
(photographs: Thomas Biginagwa, 2009).
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 7
centre of the main mound, reaching a depth of 3.6 m below ground level. Four distinct
layers were exposed. No structural remains were observed, but, as at Maurui, the recovery
of a limited amount of daub suggests the presence of structures nearby. At Kwa Sigi,
eleven 1 × 1 m test pits were dug across the island to sterile deposits at depths of 10–
150 cm below ground surface; no traces of structural remains were encountered and
the shallowness of the deposits suggests that the site was occupied for a relatively
short period; documentary evidence indicates that it was eventually destroyed by a fire
in November 1891 (Kisbey 1898; Biginagwa and Ichumbaki 2018: 74).
Faunal analysis
All faunal remains were fully analysed, except those recovered from Maurui, where 25
bags containing 9779 faunal specimens (44% of the total recovered) were sampled
across the stratigraphy for analysis. Faunal remains were analysed by TB using reference
collections at the Department of Archaeology, University of York (United Kingdom) and
the School of Veterinary Science, Sokoine University (Tanzania) under the guidance of
Professor Terry O’Connor, and with reference to published identification guides to
regional mammalian fauna (Walker 1985) and fish species (Fischer and Bianchi 1984).
In the absence of appropriate reference collections of freshwater fish from the Pangani
River, knowledgeable local fishermen, who had lived in the study area for many years,
prepared fresh reference materials of different fish taxa to aid and confirm the identifi-
cation of the archaeological specimens. Once the archaeological fish fauna and reference
materials were matched, Swahili and English names were given to the identified fish taxa
and checked against a series of published guides (Bailey et al. 1978; Skelton, 1994; IUCN
2003; Seegers 2008), as well as a standard internet source (http://fish.mongabay.com/
data/Tanzania.htm). Estimating the size of each taxon across the layers would have
informed this study of possible diachronic change in mean body size, as it could
reflect the age of the fish, environmental changes or human choice (Prendergast and
Lane 2010). However, this did not take place because only the large and most durable
elements of different taxa survived archaeologically and reconstructions of any trends
in changes to body size would thus likely have been biased towards larger taxa.
Taphonomic aspects of the terrestrial fauna were recorded for all of the assemblages,
following the six sub-aerial bone-weathering stages proposed by Behrensmeyer (1978). In
Stages 0–5, the bones show progressively greater damage. In Stage 6, the bones may not
have survived in the archaeological record. Generally, a large proportion of the studied
faunal material was in a good state of preservation (i.e. Stages 1 and 2), meaning that
other bone surface markers could be examined straightforwardly with the use of a
normal hand lens.
The relative taxonomic abundance in the assemblages was derived from the Number
of Identifiable Specimens (NISP). This method was chosen over others due to the high
level of bone fragmentation in all the samples. The abundance values were thus obtained
simply by counting all the identifiable cranial and postcranial material that could
definitely be assigned to a particular taxonomic category. To estimate the age of caprines,
the method used by Fiona Marshall (1990), who has worked extensively on domestic
fauna in East Africa, was followed. However, due to the small sample of caprine dental
material, it seemed prudent to limit the number of age classes to five instead of seven
8 T. J. BIGINAGWA AND P. J. LANE
(Table 1). Consequently, this analysis did not separate the age classes of ‘young adult’ and
‘adult’ or those of ‘aged’ and ‘very old’. The age of cattle was estimated by examining long
bone epiphyses, following the procedure described by Silver (1969; see Table 2).
To determine the possible impact of the expansion of the caravan trade on animal pro-
duction and exploitation patterns, we first sought to establish whether the occupants of
the sites were predominantly hunters or stock-keepers. The approach used by Gifford
et al. (1980) and Prendergast and Mutundu (2009) in their investigations into much
earlier transitions from hunting to herding in eastern Africa was applied. This considers
the ratio of domestic stock to wild taxa in faunal assemblages to be a direct measure of the
importance of animal husbandry to hunting. A smaller proportion of domestic species
compared to wild taxa signifies that communities subsisted predominantly by hunting,
while a greater proportion of domestic species signifies a pastoral or agropastoral
economy. The dominance of sheep/goats (henceforth ‘caprines’) in faunal assemblages
is further considered to signify a herding economy because it is easier to start a herd
with caprines as they require less direct tending than cattle (Mutundu 1999).
Estimations of the age at which domestic stock died were used to infer whether herds
were managed in such a way as to produce a surplus that could be traded with the passing
caravans. Reid (1996) notes that if the principal concern of herd management is that it
should grow then individual animals are allowed to live until the end of their reproduc-
tive life. However, most bulls will tend to be slaughtered to reduce competition for food,
thereby lessening the stress on cows and their offspring. Faunal assemblages suggestive of
this strategy are dominated by the remains of older animals, as indicated by patterns of
tooth wear and the fusion of long bone epiphyses. If herds are managed to produce meat,
most animals are slaughtered during their prime meat-bearing age— around 12 months
in the case of caprines, and 48 months for cattle (Dahl and Hjort 1976: 167). Assemblages
revealing this strategy typically have a range of animal age classes indicated by the pres-
ence of fused and unfused long bones, as well as teeth at various stages of wear.
Reid (1996) provides three conditions that determine the stages at which bulls are
eliminated from cattle herds. First, is the importance of meat in the diet of herders. If
the herd is managed primarily to produce meat, most bulls will be allowed to approach
their full meat-bearing potential before they are slaughtered, hence a ‘late off-take’.
Table 1. Caprine age classes based on dental evidence (adopted and slightly modified from Marshall
1990).
Code Age group Descriptors Approximate age in months
1 Neonate Not erupted/ unworn deciduous dentition < 1
2 Young juvenile DP4 in wear, M1 not erupted ∼ 6
3 Old juvenile M1 in wear, M2 not erupted =/> 12
4 Adult M2 in wear, M3 not erupted =/> 27
5 Aged Permanent dentition in full wear, M3 fully worn Over 27
Table 2. Estimated time for cattle long bone epiphyseal closure (after Silver 1969).
Fusing time category Estimated time Element involved
Early fusing time <18 months Distal humerus; proximal radius
Medium fusing time 2–3 years Distal tibia; distal metapodials
Late fusing time 3.5 years and above Proximal humerus; proximal tibia; distal radius; proximal and distal femora
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 9
Second, is the ability of the cattle-keeping settlement to consume or advantageously trade
meat resources. In such a scenario, there is no need to let an animal grow fully if the
settlement is small and cannot utilise all the meat. Third, are the ecological circumstances
of a particular place or time. Good access to food and water reduces the need for selective
culling, for example. Other socio-economic, cultural and environmental factors can
further influence slaughter patterns (Dahl and Hjort 1976; Quinlan et al. 2016).
Results
Table 3 and Figure 4 present the general composition of the analysed faunal material,
showing that just over half were minimally identifiable (identifiable to skeletal element
but not to exact taxonomic unit). These materials were recorded under approximate
mammal size classes, one to four, in accordance with Brain (1981) and Bunn and
Kroll (1986). Table 3 shows the dominance of medium-sized animals (MM3: approxi-
mate body weight 20–80 kg, for example sheep/goats and warthogs). Large animals fol-
lowed (MM4: approximate body weight 80 kg and above, for example cattle and Cape
buffalo). The least represented group was of small animals (MM1 and 2: less than
20 kg, such as rodents and suni/dikdik). However, it should be noted that most of the
Figure 4. General composition of the analysed faunal materials from all the sites studied.
Table 3. General composition of the analysed faunal assemblages from all study sites.
Categories
Sub-
categories Ngombezi %
Old
Korogwe %
Kwa
Sigi % Maurui % TOTAL %
Maximally
Identifiable
Terrestrial 5075 24.68 859 19.9 192 13.83 2660 27.2 8786 36
Fish 1776 8.63 1290 29.88 475 34.22 743 7.59 4284
SUB-TOTAL 13,070
Minimally
Identifiable
MM1 657 3.19 33 0.76 13 0.93 586 5.99 1289 51
MM2 995 4.83 195 4.51 7 0.5 880 8.99 2077
MM3 5912 28.75 1270 29.42 297 21.39 2152 22 9631
MM4 3325 16.16 347 8.03 325 23.41 1301 13.3 5298
SUB-TOTAL 18,295
Non-Identifiable 2823 13.72 322 7.46 79 5.69 1457 14.89 4681 13
TOTAL 20,563 100 4,316 100 1388 100 9779 100 36,046 100
10 T. J. BIGINAGWA AND P. J. LANE
bones of smaller mammals were less fragmented and so more easily identifiable to
specific taxa (Table 3).
The condition of the preserved bones was excellent, with most cortical surfaces intact
and visible. This enabled 36.2% of the material to be identified to specific taxa. The good
preservation was attributable to the presence of ash dominating the main deposits of the
excavated mounds, which may have buffered and preserved the chemistry of the buried
bones. The bone surface markers recorded were cuts, burning, and tooth pitting. The
proportion of bones with cuts ranged between 5 and 7% across the sites. The incidence
of cuts was much higher on skeletal elements of medium-sized (e.g. caprines and gazelle)
and large animals (e.g. cattle and hartebeest) than on smaller ones. Most of the specimens
that had cuts were flat and long bones, as well as vertebrae pieces. A wild cat (Felis silves-
tris) pelvis from Ngombezi had a cut mark, perhaps indicating it had been skinned for its
pelt.
Burn marks occurred on less than 10% of all faunal specimens analysed. They
occurred mostly on the long bones, ribs, and phalanges of medium-sized animals. Inter-
estingly, over half of the elephant shrew material had burn marks. Most of these were on
mandibles and examples of the tibia and fibula uniquely fused together. Such small
mammals were likely roasted whole without being butchered, which probably accounts
for the frequency of burn marks associated with this species
Tooth marks were rare, less than one percent. They occurred on various bone
elements of small, medium-sized, and large animals. A juvenile elephant (Loxodonta afri-
cana) tibia from Ngombezi had tooth marks, possibly made by a carnivore.
Table 4 and Figure 5 present the distribution of the maximally identifiable faunal
specimens (NISP), up to skeletal elements and specific taxonomic units across the
studied sites. From the findings, it can be concluded that the inhabitants of the
studied settlements practised similar animal economies, consuming both domestic and
wild animals, as well as fish. This further suggests that the environments in which exploi-
tation occurred were similar. The following sections offer interpretations of the domestic
and non-domestic components.
Interpretation of the faunal assemblages
Animal husbandry
The inhabitants of Ngombezi, Maurui, Old Korogwe and Kwa Sigi kept domesticated
chicken (Gallus gallus), sheep/goats (Ovis aries/Capra hircus) and cattle (Bos taurus).
Table 4 indicates the overall dominance of chickens, especially at Ngombezi and
Maurui, where the largest quantity of faunal remains was retrieved. Caprines followed
closely, as they were abundant at all the sites, followed by cattle. This consumption
pattern reflects that of African agropastoralists where smaller animals are more fre-
quently consumed than larger ones (Herrero et al. 2013).
The faunal remains of domesticated animals occurred consistently across the strati-
graphy in all the excavated trenches (Figure 6a–c). The only difference was the variation
in abundance in the archaeological layers. At Ngombezi, all species were under-rep-
resented in Layers 4, 5, 9 and 12, which contained lumps of house daub and pottery
but very few faunal remains. At Old Korogwe, there was little evidence of chickens
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 11
i-
n
Ta
bl
e
4.
Ta
xo
no
m
ic
re
pr
es
en
ta
tio
n
by
N
IS
P
fo
r
al
lt
he
si
te
s
st
ud
ie
d.
Ca
te
go
ry
Sc
ie
nt
ifi
c
na
m
e
Co
m
m
on
na
m
e
N
go
m
be
zi
M
au
ru
i
O
ld
Ko
ro
gw
e
Kw
a
Si
gi
To
ta
l
Pr
ef
er
re
d
ha
bi
ta
ts
(a
ft
er
Ki
ng
do
n
19
88
)
D
om
es
ti
c
O
vi
s
ar
ie
s/
Ca
pr
a
hi
rc
us
Sh
ee
p/
go
at
(c
ap
rin
e)
1,
31
7
64
6
10
0
23
2,
08
6
Al
lh
um
an
ha
bi
ta
ts
Bo
s
ta
ur
us
Ca
tt
le
54
8
37
1
39
17
97
5
G
al
lu
s
ga
llu
s
Ch
ic
ke
n
1,
53
1
71
8
74
19
2,
34
2
M
am
m
al
Ra
tt
us
ra
tt
us
Bl
ac
k
ra
t
6
3
7
1
17
Ar
ou
nd
hu
m
an
se
tt
le
m
en
ts
Th
ry
on
om
ys
sw
in
de
ria
nu
s
Ca
ne
ra
t
72
3
26
5
15
7
44
1,
18
9
Re
ed
be
ds
an
d
gr
as
s
ar
ea
s
al
on
g
riv
er
s
an
d
la
ke
s
Th
ry
on
om
ys
gr
eg
or
ia
nu
s
Cr
ic
et
om
ys
ga
m
bi
an
us
G
ia
nt
po
uc
he
d
ra
t
12
3
14
2
6
9
28
0
Al
m
os
t
al
lh
ab
ita
ts
?
U
ni
de
nt
ifi
ed
ro
de
nt
62
96
38
8
44
59
0
?
Pe
tr
od
ro
m
us
te
tr
ad
ac
ty
lu
s
El
ep
ha
nt
sh
re
w
19
8
13
9
22
10
36
9
Ri
ve
rin
e
st
rip
s/
fo
re
st
ed
riv
er
ba
nk
s
El
ep
ha
nt
ul
us
sp
.
D
en
dr
oh
yr
ax
ar
bo
re
us
Tr
ee
hy
ra
x
15
7
1
2
25
G
al
le
ry
fo
re
st
,r
iv
er
in
e
st
rip
s
Sy
lv
ic
ap
ra
gr
im
m
ia
Bu
sh
/c
om
m
on
du
ik
er
15
7
28
24
12
22
1
Sa
va
nn
a
an
d
w
oo
dl
an
ds
N
eo
tr
ag
us
m
os
ch
at
us
/
M
ad
oq
ua
gu
en
th
er
i
Su
ni
/D
ik
di
k
11
32
2
2
47
Fo
re
st
s/
th
ic
ke
ts
/
ar
id
ad
ap
te
d
Ae
py
ce
ro
s
m
el
am
pu
s
Im
pa
la
56
36
8
2
10
2
D
ec
id
uo
us
w
oo
ds
,b
us
hl
an
ds
,‘
ed
ge
’h
ab
ita
ts
G
az
el
la
gr
an
ti
G
ra
nt
’s
ga
ze
lle
30
12
3
0
45
M
ix
ed
sh
ru
b
w
oo
dl
an
ds
an
d
sa
va
nn
a
pl
ai
ns
Re
du
nc
a
re
du
nc
a
Bo
ho
r
re
ed
bu
ck
9
7
0
16
W
at
er
de
pe
nd
en
t;
fl
oo
d
pl
ai
ns
,r
ee
d
be
ds
Po
ta
m
oc
ho
er
us
po
rc
us
Bu
sh
pi
g
32
6
13
2
53
Fo
re
st
s,
riv
er
in
e
ha
bi
ta
ts
w
ith
th
ic
k
co
ve
r,
re
ed
be
ds
,m
ar
sh
es
,
m
an
gr
ov
e
sw
am
p
O
ry
ct
er
op
us
af
er
Aa
rd
va
rk
7
12
0
0
19
D
ep
en
de
nt
on
ab
un
da
nc
e
of
te
rm
ite
s
Al
ce
la
ph
us
bu
se
la
ph
us
Re
d
ha
rt
eb
ee
st
19
8
1
0
28
Sa
va
nn
a,
gr
as
sl
an
d,
w
oo
dl
an
d
H
ip
po
tr
ag
us
eq
ui
nu
s
Ro
an
an
te
lo
pe
0
1
0
1
Sa
va
nn
a
w
oo
dl
an
d,
m
ix
ed
bu
sh
an
d
gr
as
sl
an
ds
Sy
nc
er
us
ca
ff
er
Ca
pe
Bu
ff
al
o
3
0
0
3
6
Fo
re
st
,g
ra
ss
y
gl
ad
es
Lo
xo
da
nt
a
af
ric
an
a
El
ep
ha
nt
1
0
0
0
1
Lo
w
la
nd
fo
re
st
s
Eq
uu
s
qu
ag
ga
Co
m
m
on
ze
br
a
5
6
0
0
11
W
oo
de
d
gr
as
sl
an
d
H
ip
po
po
ta
m
us
am
ph
ib
iu
s
H
ip
po
po
ta
m
us
16
0
0
16
Ri
ve
rs
H
el
og
al
e
pa
rv
ul
a
(d
w
ar
f
m
on
go
os
e)
M
on
go
os
e
19
9
2
1
31
W
oo
de
d
to
fo
re
st
ar
ea
s/
co
m
m
on
w
he
re
th
er
e
is
an
ab
un
da
nc
e
of
aa
rd
va
rk
an
d
pa
ng
ol
in
bu
rr
ow
s
G
al
er
el
la
sa
ng
ui
ne
us
(s
le
nd
er
m
on
go
os
e)
Ci
ve
tt
ic
tis
ci
ve
tt
a
Af
ric
an
ci
ve
t
11
2
13
W
he
re
ve
r
th
er
e
is
co
ve
r
Fe
lis
sil
ve
st
ris
W
ild
ca
t
8
8
?
Ca
ni
s
fa
m
ili
ar
is
D
og
1
1
H
um
an
se
tt
le
m
en
t
12 T. J. BIGINAGWA AND P. J. LANE
?
U
ni
de
nt
ifi
ed
ca
rn
iv
or
e
13
6
3
22
?
Pa
pi
o
sp
.
Ba
bo
on
2
2
W
oo
dl
an
ds
,f
or
es
ts
Ce
rc
op
ith
ec
us
ae
th
io
ps
Ve
rv
et
m
on
ke
y
12
4
-
-
16
W
oo
dl
an
d
al
on
g
st
re
am
s,
riv
er
s
an
d
la
ke
s
Bi
rd
Av
es
Bi
rd
16
3
87
1
25
1
?
Re
pt
ile
Va
ra
nu
s
ni
lo
tic
us
N
ile
m
on
ito
r
liz
ar
d
1
1
Pe
rm
an
en
t
riv
er
ba
nk
s
an
d
da
m
s
St
ig
m
oc
he
ly
s
pa
rd
al
is
Le
op
ar
d
to
rt
oi
se
1
1
2
Co
as
ta
lp
la
in
,b
us
hv
el
d
Fi
sh
Sy
no
do
nt
is
sp
.
Ca
tfi
sh
[N
go
go
go
]
29
8
75
10
5
39
51
7
Fr
es
hw
at
er
Cl
ar
ia
s
sp
.
Ca
tfi
sh
[K
am
ba
le
]
23
4
97
88
46
46
5
Fr
es
hw
at
er
O
re
oc
hr
om
is
sp
.
Ti
la
pi
a
[P
er
eg
e]
78
39
10
16
14
3
Fr
es
hw
at
er
Ba
rb
us
sp
.
Ba
rb
el
s
[K
uy
u]
43
53
25
9
13
0
Fr
es
hw
at
er
An
gu
ill
a
sp
.
Ee
ls
[N
in
gu
]
12
34
7
-
53
Fr
es
hw
at
er
Ta
xo
no
m
ic
al
ly
un
id
en
tifi
ed
11
11
44
5
10
55
36
5
29
76
Fr
es
hw
at
er
TO
TA
L
68
51
34
03
21
49
66
7
13
,0
70
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 13
Layer 1. Only one specimen was recovered from this layer, while cattle and caprines
were both present. All domestic stock taxa and a diverse range of non-domestic
species were present at both Ngombezi and Maurui in each of the successive pre-
caravan, caravan and post-caravan trade phases of settlement. The two settlements
also revealed the longest dated archaeological sequences and both yielded a significant
quantity of faunal remains.
The representation of body parts and skeletal elements of caprines and cattle is con-
gruent with nineteenth-century eyewitness accounts (Burton and Speke 1858: 199; John-
ston 1879: 546), which record that these were herded in the study area (and not acquired
from neighbouring pastoralists). This is attested by the presence of both high utility parts
(such as scapulae and upper limbs) and low utility ones (such as lower limbs and meta-
podials) in the analysed assemblages. Skeletal elements of cattle and caprines from all
sites occurred in a ratio of 1:2 on average. This suggests that the preservation and recov-
ery of bones was not biased against smaller bones, as might have been expected if they
had been subject to different preservation and/or recovery factors.
Figure 5. Distribution of the maximally identifiable faunal materials between study sites.
Figure 6. Examples of domesticated animal species recovered in excavation: a) chicken femur; b)
caprine humerus; c) cattle scapula.
14 T. J. BIGINAGWA AND P. J. LANE
Mortality profiles
The dental and long bone evidence revealed the age to which caprines and cattle
lived before being culled. The examined dental evidence (teeth eruption and
stages of wear) revealed the dominance of adult and aged caprines at Ngombezi,
together constituting 61.5% (Table 5; Figure 7a). Young and old juveniles consti-
tuted 38.4%. No neonates were found. Examination of long bone epiphyseal
fusion revealed a close match with the dental evidence, indicating that most caprines
at Ngombezi had already reached 24 months and above before they were slaugh-
tered. There was a similar pattern at Maurui, where 82% (n = 27) of the examined
jaws revealed that caprines were at least 24 months old before being slaughtered.
For cattle, the examined long bone epiphyses showed that over 50% of them lived
up to three years or more before being slaughtered (Table 6; Figure 7b). In addition,
all five cattle jaws recovered from Maurui showed that the cattle were seven years or
older, which is an advanced age for this species (Meado and White 1979; Marshall
1990; Ryan et al. 2000).
An important observation needs to be made with regard to Figure 7b, which shows
the trend in the ratio of the fused to unfused long bones of cattle, increasing from
younger to older cohorts. In principle, the trend should have been in the opposite
direction. Since the bones of cattle aged 1–2 years are under-represented in the assem-
blages, this suggests that cattle of this age group were killed farther away from the
settlements. This interpretation corroborates the eyewitness accounts of nineteenth-
century travellers (Burton and Speke 1858; Johnston 1879), who observed that live-
stock from the Ruvu/Pangani Valley were sent to market for sale in the coastal
towns of Tanga, Pangani and Saadani; these individuals are likely to have been
younger animals.
Generally, the mortality profiles of the domestic animals imply that herds were
managed for growth and continuity rather than for the consumption of meat. The
culling practices that were uncovered suggest that there was little need to consume dom-
estic stock, something that would have led to the slaughter of young animals. This, in
turn, implies that live animals, in particular cattle, had other socio-economic values to
the inhabitants of these settlements such as for blood, milk, dung, traction power and
symbolic value, rather than for meat. These issues are explored in detail in subsequent
sections.
Table 5. Ngombezi: scores for caprine ages using dental evidence.
LAYERS
AGE GROUP
Neonate Young juvenile Old juvenile Adult Aged Total
1 1 1 2
2 1 1
3 3 2 2 6 13
4 1 1
5 2 2
6 1 3 3 3 10
7 4 1 5
8 1 2 2 5
TOTAL 0 7 8 12 12 39
% 0 17.94 20.51 30.76 30.76 100
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 15
Hunting
The composition of wild animal species in the faunal assemblages at all four sites is
similar (Table 4), suggesting that their inhabitants exploited wild animals in similar
ways. This is unsurprising given the location of these islands and their proximity to
each other.
Figure 7. Ngombezi: mortality profiles of domestic stock: a) caprine age profile reconstructed using
dental evidence; b) cattle age profile reconstructed using long bone evidence.
Table 6. Ngombezi: scores for cattle ages using long bone epiphyses.
Fusion Phase Total elements Total fused Total unfused
Early
(<18 months)
24 15
(62.5%)
9
(37.5%)
Middle
(24–36 months)
63 51
(80.9%)
12
(19%)
Late
(≥42 months)
86 41
(47.6%)
45
(52.3%)
16 T. J. BIGINAGWA AND P. J. LANE
Small mammals
Rodents, elephant shrews and hyraxes together formed 73% of the wild mammalian com-
ponent. Of these, rodents formed the majority of the wild mammalian species consumed
(Table 4). The implication of the consumption of small mammals is explored below.
Three rodent taxa were identified: Thryonomys sp. (cane rat), Cricetomys gambianus
(giant pouched rat) and Rattus rattus (black rat). Cane rats were the most abundant, con-
stituting 57.2% of rodent material (Figure 8a). Two species of cane rat were identified on
the basis of mandible size and mandibular dentition: Thryonomys swinderianus (the
larger species) and Thryonomys gregorianus (the smaller species). The second largest
rodent group was the giant pouched rat, which comprised 13.4% of all the rodent
material. The giant pouched rat was represented by one species, Cricetomys gambianus
(Figure 8b). The least represented rodent was the black rat (Rattus rattus), at less than
1% of the total amount of rodent material; it is unlikely to have been a food species.
The cane rat and giant pouched rat were consistently distributed across the stratigra-
phy at Ngombezi, Old Korogwe and Maurui. Black rats appear only in the middle levels
and higher up. This suggests that they are a later introduction to the settlements. Pre-
ndergast et al. (2017) report that Rattus rattus arrived in Unguja Ukuu, Zanzibar,
Figure 8. Examples of non-domesticated animal species recovered in excavation: a) cane rat mand-
ible; b) giant pouched rat mandible; c) elephant shrew mandible; d) metapodials of antelope species
of various sizes (photographs: Thomas Biginagwa, 2011).
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 17
from Southeast Asia between the seventh and eighth centuries AD. The appearance of
Rattus rattus around the point at which imported European glass beads appear in the
sites suggests that their introduction may have been associated with the expansion of
trade contacts between coastal and interior communities.
Elephant shrews constituted the second largest group (15%) of small-bodied wild
animals consumed at the settlements (Table 4). They were identified through well-pre-
served dental material and uniquely fused tibiae/fibulae. Two species were identified
based on the size of their premolars and canine alveoli (Figure 8c). The smaller taxon
(the majority in the assemblages at all the sites) matched well to the four-toed elephant
shrew Petrodromus tetradactylus. The less abundant species belonged to the larger taxon
Elephantulus sp. At Ngombezi, elephant shrew remains were recovered from all but Layer
13, while at Old Korogwe and Maurui they were present in all the archaeological layers,
although with a slight variation in quantity across them. As discussed below, the environ-
mental conditions of the studied sites and the ecological niches preferred by Petrodromus
tetradactylus and Elephantulus sp. suggest that both were likely procured near the
settlements.
Wild ungulates and other larger mammals
Various species of ungulates were represented by 565 skeletal elements (Figure 8d).
Nearly half the material (n = 268; 47.4%) belonged to common duiker (Sylvicapra
grimmia), suni (Neotragus moschatus) and dikdik (Madoqua guentheri). As listed in
Table 4, various medium-sized animal species (n = 235; 41.5%) were also accessible to
the inhabitants of the settlements. These comprised impala (Aepyceros melampus),
Grant’s gazelle (Gazella granti), bohor reedbuck (Redunca redunca), bushpig (Potamo-
choerus porcus) and aardvark (Orycteropus afer). Furthermore, a few large ungulate
species were also intermittently accessible (Table 4). It is worth noting that all 16 hippo-
potamus (Hippopotamus amphibius) bones were recovered from Maurui. There is a pool
at the northern end of Maurui Island formed by swirling water, which today hosts a thriv-
ing hippopotamus population and may well have been where the occupants of Mauri
hunted these animals in the past.
The stratigraphic distribution of the fauna shows that ungulates of all sizes were
almost completely missing in the two uppermost (post-caravan trade) layers at Old
Korogwe. Although present across the stratigraphy at Ngombezi and Maurui, they simi-
larly declined in the uppermost two layers at both sites. Such a general decline can be
explained by the expansion of human settlements and the associated clearance of
forest in areas around the study sites. This inference is supported by the increased pres-
ence of rodents in the uppermost layers at both sites. Studies show that the abundance of
rodents is accentuated in areas where humans have cleared land to cultivate crops
(Ashton 2000; Shenkut et al. 2006).
Other terrestrial fauna
The material of most of these was recovered from Ngombezi and Maurui (Table 4), con-
stituting avian fauna (n = 251), various carnivores (n = 75) such as dwarf mongoose (Helo-
gale parvula), slender mongoose (Galerella sanguineus), African civet (Civettictis civetta),
wildcat (Felis silvestris) and domestic dog (Canis familiaris). Non-human primates were
also present in the assemblages (n = 18), namely vervet monkey (Cercopithecus aethiops)
18 T. J. BIGINAGWA AND P. J. LANE
and baboon (Papio sp.). Monkeys and baboons are still common around the study sites. It
is uncertain whether any of these were used as food, although there is some documentary
evidence for the consumption of dogs in neighbouring Usambara (Krapf 1860: 225).
Fishing
Taxonomically identified fish fauna constituted 31% (n = 1308) of all the fish bones
recovered from all the sites studied (Figure 9). Siluriformes dominated the assemblages,
with two main families, namely Mochokidae and Clariidae (Table 7). The catfish of the
genus Synodontis was the mochokid present in the assemblages. It was identified by its
Figure 9. Examples of fish faunal specimens recovered in excavation (photographs: Thomas Bigi-
nagwa, 2011).
Table 7. Taxonomically identified fish fauna from all the sites studied by NISP.
Taxa Ngombezi Old Korogwe Kwa Sigi Maurui Total %
Clariidae (cf. Clarias spp.) 234 88 46 97 465 35.5
Mochokidae (cf. Synodontis spp.) 298 105 39 75 517 39.5
Tilapiine (cf. Oreochromis spp.) 78 10 16 39 143 11
Cyprinidae (cf. Barbus spp.) 43 25 9 53 130 10
Anguillidae (cf. Anguilla spp.) 12 7 - 34 53 4
Total Site NISP 665 235 110 298 1308 100
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 19
pectoral fin spines, which survive longer in archaeological deposits. Although present in
all layers across the sites, the taxon was less abundant in deeper basal levels at Ngombezi
and Maurui. This may be due to poor preservation conditions. Skelton (1994) indicates
that the taxon is native to the Pangani River and prefers open water. Tamatamah (2007)
indicates that Synodontis copes well with seasonal changes of habitat because it can eat
many types of food.
Clariidae are the second most abundant catfish in the fish assemblages (Table 7).
Clarias was the only genus identified based on the head bones. It feeds on other small
fish, molluscs, detritus and aquatic weeds (Tamatamah 2007). The taxon is also native
to the Pangani River and prefers stagnant water (Seegers 2008). Cichlidae rank third
in the taxonomically identified fish fauna. The family is represented by the genus Oreo-
chromis (cf. Oreochromis korogwe). The taxon is found in the Pangani River, thriving in
normal riverine habitats.Oreochromis has the ability to adapt to an altered hydrology and
other changes in river habitats (Lowe 1955).
A few dental specimens were definitively attributed to Barbus spp. of the family Cyprini-
dae. Barbus inhabits the deeper stretches of rivers with rocky shores or riparian trees
(Skelton1994). The taxon is naturally distributed and is native to thePangani River (Banister
1973). Finally, a few fish bone specimens were attributed to the family Anguillidae, which is
reported as being common in the Pangani River (Skelton 1994). Although seemingly rare in
the analysed assemblages, this may well be due to the low number of easily identifiable
elements for this taxon,making recognition of its presence in archaeologicalfishbone assem-
blages challenging, rather than reflecting a lack of consumption of the species.
Generally, the quantity of fish bones identified of specific taxa from each excavation
unit and layer was relatively small. Furthermore, most fish taxa were under-represented
in the basal deposits, probably due to poor preservation conditions. As a result, no sig-
nificant trend can be claimed for the spatial-temporal distribution of fish fauna in the
excavated layers at the studied sites.
Discussion
Domestic livestock were an important component of the livelihood strategies of the
inhabitants of Ngombezi, Maurui, Old Korogwe and Kwa Sigi. At the main study site
Figure 10. Ngombezi: dietary contribution of domestic and wild animal species per settlement phase
excluding fish.
20 T. J. BIGINAGWA AND P. J. LANE
of Ngombezi, for example, domestic species comprised approximately two-thirds of the
meat supply while the non-domestic component formed one-third. However, hunting
and fishing significantly supplemented the meat supply. The observed consumption of
both domestic and wild resources was consistent throughout all the studied settlements
(Figure 10).
Mortality profiles suggest that caprines and cattle were managed primarily for sustain-
ing the growth of the herd and continuity rather than for immediate consumption. Cattle
and caprines were allowed to live to maturity before they were slaughtered. This suggests
that there was little pressure to consume domestic animals. The observed herd manage-
ment strategies and culling practices are similar to those documented among modern
agropastoralists in relatively unstressed conditions (Reid 1996), whereby bulls are left
to attain maximum weight before being slaughtered (late off-take). For example, the
East African zebu attains its maximum weight when about 40 months old; in stress-
free conditions cows are left to increase the herd size until they are at least ten years
old, which is considered a less productive age (Dahl and Hjort 1976).
Giblin (1992) depicts unstressed environments in the late pre-colonial LPRB, arguing
that although a large part of Zigualand experienced drought farming was possible along
the banks of the Pangani River and in valleys where moist and fertile land was available
(Krapf 1860: 373). Livestock-keeping was also possible because farmers were able to
control the growth of woody scrub, a favourable habitat for ticks and parasites, by reg-
ularly burning it and clearing trees and could thus regularly allow animals to graze.
Additionally, the hunting of wild animals, particularly bushpigs and Cape buffalo (Syn-
cerus caffer), helped to reduce the spread of infectious diseases to livestock, as most wild
species can carry infectious bovine parasites (Giblin 1992: 19).
Overall, the pressure to consume cattle and caprines was probably insignificant
because chickens, wild animals and fish contributed significantly to the meat supply
for the inhabitants and perhaps also to that of the caravans. In rural regions in Africa,
chickens are kept by almost every household (Kitalyi 1998). They are less expensive to
keep as they need less space. Unlike the mammalian stock, chickens can survive
periods of drought and famine. They are an important instant supply of meat during
an unanticipated shortage of the savoury component of meals, particularly when a
family receives unexpected guests. Chickens can withstand a high off-take rate because
they reproduce quickly. Explorers who journeyed through East Africa during the
height of the caravan trade reported that chickens were among the important items
used to provision caravans in exchange for glass beads (Burton 1859: 184).
Similarly, small mammalian fauna, and rodents in particular, likely contributed sig-
nificantly to the diet at the studied settlements. The assemblages recovered from the
four sites possibly represent the largest ratio of rodents to other mammalian species
reported thus far from archaeological sites in East Africa, and so warrants more discus-
sion. Importantly, it should be noted that the environment around the study sites (Table
4) must have attracted small mammals, including rodents, which supplied the inhabitants
and possibly the caravans with meat. For example, grasscutters, the most abundant
rodent species in the assemblages, prefer areas near rivers and streams or marshes due
to the presence of edible plants (Opara 2010). Rodents can also adapt to newly deforested
areas where food and commercial crops are grown and to secondary savanna resulting
from deforestation (Ashton 2000). Rodents have a higher reproduction rate and their
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 21
populations can recover quickly in situations of high exploitation (Stahl 1982; Walker
1995). Taking these factors into account, their abundance compared with other
members of local fauna in the studied assemblages may thus be due to extractive anthro-
pogenic activities having taken place in the area.
The consumption of rodents by current African populations does not signify subsis-
tence stress (Malaisse 1997) — contemporary Zigua practices in the study area being a
case in point— as farmers typically hunt rodents primarily to protect crops and only sec-
ondarily as an alternative source of meat (Assogbadjo et al. 2005). Nonetheless, Stahl
(1982) indicates that approximately 67–76% of a rodent is edible, so they can contribute
significantly to the meat diet when they are caught in abundance. In Bénin, a study by
Assogbadjo and colleagues (2005) found that more than 53% of the villagers near
Lama Forest prefer rodent meat to other wild meat.
The consumption of wild animals also appears to have been quite common across East
African farming communities for some considerable time and continues to this day
(Driver 2008; Ceppi and Nielsen 2014). Helm (2000), for example, reports a subsistence
economy based on exploiting both domestic and non-domestic fauna resources at five
Kenyan coastal hinterland, Early to Later iron-working and farming settlements. The
inhabitants of three of these settlements namely Mgombani 02 (c. AD 775), Chombo
01 (c. AD 900) and Mteza 01 (c. AD 895) kept a small number of domestic livestock
initially, but these were not their main source of meat. Instead, they relied on hunting,
fishing and gathering terrestrial snails. Even when animal husbandry became a major
occupation at the Middle and Later iron-working and farming settlements of Mtsengo
01 (c. 1415) and Mbuyuni 01 (c. 1640), hunting wild game and gathering land snails
were still practised (Helm 2000: 257; see also Foutch et al. 2009 for a similar observation
concerning pre-colonial animal economies in the northern Rukwa Valley, Tanzania).
Prendergast and colleagues (2017) report rather similar animal exploitation patterns
from two mid-first millennium Iron Age farming communities at Fukuchani and
Unguja Ukuu in Zanzibar. Inhabitants of these settlements exploited both marine and
terrestrial resources throughout; the latter comprising mainly wild fauna, especially
small bovids and rodents. As reported here for the Lower Pangani, Prendergast and col-
leagues also observed an unchanged ratio of wild to domestic animals throughout the
occupation of the two settlements, highlighting the continued importance of wild
resources long after the transition to farming.
Archaeological assemblages with many large prey taxa could suggest a more plentiful
environment than those dominated by small prey taxa. Kusimba (2004), for example,
reports a reduction in hunting large to medium-sized game to a reliance on smaller
mammals, reptiles and birds during the course the occupation of rock shelters on Mt
Kasigau in western Tsavo, southeast Kenya. He interpreted this shift in focus as an indi-
cator of subsistence stress brought about by social and political instability arising from
intensification of the regional slave trade and exacerbated by climatic change in the
area some 300 years ago.
Despite well-documented social disruptions arising from the intensification of the
regional slave trade in the nineteenth century across the Pangani Basin (Giblin 1992),
the animal economies of the sites reported here (which were occupied broadly coevally
with those on Mt Kasigau) do not seem to have been affected in the same way and the
local environment, encompassing open grassland, wooded grassland, savanna woodland,
22 T. J. BIGINAGWA AND P. J. LANE
montane forest, river edges and marshes, appears to have supported a diverse range of
wild animals during the nineteenth century and earlier. As Table 4 shows, several ungu-
lates preferring different habitats were available for human exploitation requiring diverse
hunting strategies, perhaps ranging from individual to group efforts assisted by dogs
(Lynwood 1994; Juwayeyi 2008).
That farmers hunt wild game and also small rodents is not very surprising given the
widely attested importance of hunting in pre-colonial East Africa and the damage that
wild animals from elephants to cane rats can cause to crops. Horticultural activities in
other sub-tropical areas can also lead to an increase in the abundance of small
mammals, simultaneously creating enhanced threats to crop harvests and a convenient
source of high-quality protein (Speth and Scott 1989: 74). Additionally, some species
may have been hunted for other purposes. At the Kibaoni site in the Rukwa Valley, Tan-
zania, for example, Foutch and colleagues (2009) reported a cut-marked femur of a
medium-sized cat (serval (Leptailurus serval) or caracal (Caracal caracal)). They inter-
preted this as the product of witchcraft rituals, which are still practised in the area.
The remains of a wild cat bearing a cut mark recovered from Ngombezi might have
been used for similar purposes, although equally plausible alternative interpretations
could be forwarded.
The large quantity of fish bones recovered from the excavations implies that they were
abundant and available for consumption, as also reported in the historical sources
(Baumann 1891: 177). The occurrence of similar fish taxa at the different sites is not sur-
prising given their proximity to each other along the Pangani River, and the occupants of
all four sites seemingly exploited similar aquatic habitats. It is likely that fish species that
were present in the Pangani River during the nineteenth century have not fluctuated sig-
nificantly since then, despite the construction of dams further upstream. Oral infor-
mation gathered from local residents during the excavations indicated that some fish
species are found in large quantities seasonally and others are common all the year
round. For example, Ningu (Anguilla) and Kuyu (Barbus) are available during the hot
season. Perege (Oreochromis) and Kambale (Clarias) are available throughout the year,
although more so during the rainy season.
The common fishing methods still employed include the use of basket traps (migono)
made of thin wooden rods and cord made from palm leaves (Figure 11). These are placed
in the river with each basket’s opening facing upstream so that the fish flowing with the
water are trapped in them. There are also wicker basket traps (masega), which are similar
to migono, but smaller with a wide opening at one end into which the fish swim. The
other end is narrow, preventing the fish from turning round. Nets (kimia or lwavu)
are also used, but not frequently. Another method is to build reed fence-traps across
the river. Line fishing with hooks is also used, but is less common, and no fishhooks
of any description were recovered from the excavations, unlike at coastal sites where
fishhooks have been present since at least the mid-first millennium AD (Prendergast
et al. 2017).
Conclusions
This zooarchaeological study did not detect significant changes in patterns of animal pro-
duction and exploitation during the height of the nineteenth-century caravan trade at a
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 23
sample of known caravan staging points in the LPRB of northeastern Tanzania. No clear
evidence was recovered that would indicate that either a surplus was intentionally being
produced or that there were significant changes to patterns of meat consumption in
response to the expansion of the caravan trade, although there was possibly a decline
in access to hunted game in the post-caravan trade era. It is unclear whether this was
because wild game had been locally extirpated or was more due to the enforcement of
game ordinances by the German colonial authorities (the first such ordinance was intro-
duced in 1896), restricting those permitted to hunt legally (Gissibl 2016: 73–96) or a com-
bination of these factors.
Nonetheless, based on the zooarchaeological analysis presented here, it would appear
that herds were managed throughout the period of occupation of these riverine sites pri-
marily for growth and continuity rather than for the immediate production of meat. The
taxonomic diversity of wild prey consumed at these settlements indicates a focus on small
to medium-sized animals, with occasional contributions from larger species. In addition
to fish, the supply of wild meat would have supplemented that of domestic animals. Such
a widespread consumption of wild resources would thus suggest that if there had been a
shortage of meat from domestic livestock at the climax of the caravan trade this was easily
met by hunting and fishing.
Figure 11. Kwa Sigi, Zigua fishing basket (photograph: Paul Lane, 2009).
24 T. J. BIGINAGWA AND P. J. LANE
It is also reasonable to conclude that domestic animals were of other socio-economic
and cultural value to the islands’ inhabitants, rather than being just a convenient source
of meat. Giblin (1992: 34) has noted that livestock functioned in politics and patronage
among nineteenth-century Zigua communities. The exchange of livestock strengthened
political loyalties and patron-client relations in Zigua farming communities and between
Zigua farmers and neighbouring Maasai pastoralists. Importantly, the male heads of
Zigua families and clans would win the affection and loyalty of their children by prom-
ising them an inheritance in the form of livestock. This helped ward off the competing
affection of the mothers’ relatives, which was pronounced among the Zigua, among
whom descent is traced matrilineally (Giblin 1992: 35). Additionally, cattle in particular
were used to pay bridewealth for sons about to marry.
Recent analysis of the archaeobotanical remains recovered from Maurui Island has
also revealed continuity in the production and consumption of the main local food
crops, notably sorghum (Sorghum bicolor) and millet (Pennisetum glaucum), throughout
the occupation of this settlement and at the apex of the caravan trade (Kawiche 2016).
Excavations at the site of Kwa Fungo, another nineteenth-century Zigua village that
served as a caravan halt, located roughly 30 km to the southwest of Korogwe, also uncov-
ered evidence for maize (Zea mays) consumption, as well as a distinctive local ceramic
assemblage that differs from that recovered from the sites discussed here (Lane 2011;
Croucher 2015: 220–223). As at Ngombezi, only limited quantities of imported beads
or European ceramics were present.
In neither case, however, is the archaeobotanical information sufficient to assess
whether the inhabitants of these sites sought to intensify agricultural rather than livestock
production as a response to the trading opportunities provided by passing caravans. Such
agricultural intensification certainly happened elsewhere— the best example being at the
southern end of Lake Baringo (Anderson 1989; Petek 2018) — and has also been
suggested for parts of the eastern Pare Mountains bordering the Pangani River Valley
(Håkansson 2008). Further comparative studies of other caravan staging posts along
the Pangani corridor beyond those reported here are clearly needed so as to place the
Waruvu island settlements around Korogwe in a broader regional perspective. Walz
(2010), for example, reports on several such sites to the north of Korogwe, which,
based on the materials recovered, he considered could have been linked directly with
the coast via trade for many centuries before the arrival of the Portuguese. Moreover,
as Håkansson (2004, 2008) notes, the historical sources suggest that Maasai communities,
in particular, were able to increase their herds in response to the economic opportunities
provided by the expanding caravan trade. Further archaeological investigation of these
issues needs to examine pastoralist sites alongside those of farming communities.
The findings of our study do raise questions, however, about the frequency and
volume of caravans that actually passed through this landscape. Far from being as numer-
ous and voracious in their demand for supplies as has sometimes been assumed, it is
possible that the Lower Pangani route was only infrequently used by caravans during
the height of the caravan trade. This may explain, for example, why there was little
pressure on the local inhabitants to intensify food production. Contemporary nine-
teenth-century published sources certainly demonstrate the shifting nature of the
various trade routes. In one, for example, the Reverend J.P. Farler (1882: 740), who
was stationed at the Magila Mission in Usambara, observed that ‘lately caravans to the
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 25
Masai country and the Victoria Nyanza have taken a new route via Magila, instead of
following the valley of the Luvu [Pangani]. This has been done to escape the extortions
of Semboja, chief of the Wakilindi, who resides at Mazindi, and levies tribute on the car-
avans both going and returning’.
It is perhaps also appropriate to re-examine some of the assumptions (including our
own) that commonly underlie estimates of the demands that passing caravans made on
the communities they camped among over the course of their long journeys to and from
the interior. In a description of the arrival of a caravan at Njemps at the southern end of
Lake Baringo on 27 December 1887, which was by then a common terminus for caravans
emanating from Pangani, Von Höhnel (1894: 24) records that the members of this par-
ticular caravan ‘had subsisted for more than three months on a handful each of dhurra or
eleusine and were about to undertake the return journey to Miansini without any store of
provisions with them’. He commented further that the ‘hardships connected with the
bringing of ivory down to the coast are very great’, implying that such shortage of
food supplies may have been quite common and due to a ‘thoughtless want of prep-
aration’ (Von Höhnel 1894: 27).
That latter remark probably needs to be taken with a grain of salt (and a few para-
graphs earlier Von Höhnel describes the diet of his own expedition as being short on
carbohydrates and mostly comprising meat from wild and domestic animals) and more
reflective of the challenges that caravans may have faced trying to provision themselves.
That communities living along the trade routes drove a hard bargain is attested by the
frequency with which they rejected certain trade goods, especially particular types of
glass beads, even when such goods had been in demand a year or so previously (see Pal-
laver 2009 and Prestholdt 2004 for wider discussion and specific examples). This cer-
tainly seems to have been the case with the inhabitants of the two Njemps
settlements at the time of Count Teleki’s visit (Petek 2018). According to Von
Höhnel (1894: 5) they were by now ‘quite spoiled by the constant and long visits
they receive from caravans’ and had become ‘very exacting about what they will take
in payment for their wares’. For example, they would ‘have nothing to do with glass
beads’ and trade goods had to be used to obtain even ‘the very smallest quantity of
grain’, while they traded ivory only for cattle, demanding also ‘a considerable
tribute’ on top of these payments.
In contrast to the evidence from Baringo, Mgombere’s (2017) recent archaeobotanical
study of two caravan halts at Mang’ua and Kikole in southwestern Tanzania found no
evidence for specialisation in the production of food crops at these sites during the era
of the caravan trade. Instead, various crops, including finger millet, pearl millet, maize,
rice, coconuts, legumes and cotton, were cultivated and consumed, as in previous centu-
ries. In her study of Ujiji, the caravan terminus on Lake Tanganyika for the central
caravan route and at Uvinza, a salt production site, also on the central route, Wynne-
Jones (2010) similarly found more evidential support for continuities rather than wide-
spread changes in social practices during the height of the caravan trade. For example,
there was limited evidence for imported trade goods (such as ceramics and glass
beads) providing a significant source of wealth or power. Instead, Wynne-Jones found
that the power of local élites continued to derive from their control over salt production
and that the caravan trade did not have a particularly transformative impact on those
seeking power and authority. Emerging archaeological findings such as these should
26 T. J. BIGINAGWA AND P. J. LANE
serve, perhaps, as a caution against assuming that the nineteenth-century caravan trade
was universally transformative across all East Africa. As argued here, if these narratives
remain untested, they are likely to mask the complexity and dynamism of local practices.
Note
1. Or simply Waruvu; also rendered in some sources as Ruvu, Rufu, Luvu and, rarely, Luffa.
Acknowledgements
This study was undertaken as part of TB’s doctoral research at the University of York within the
Historical Ecologies of East African Landscapes (HEEAL) project funded by a European Union
Marie Curie Excellence Grant (MEXT-CT-2006042704) awarded to PL (2007-2011). TB received
additional support for fieldwork at Maurui and for dating radiocarbon samples from, respectively,
the African Humanities Program (AHP) and the Volkswagen Foundation (VWF). Research
permits were awarded by the Antiquities Department of the Ministry of Natural Resources and
Tourism through the University of Dar es Salaam Archaeology Practical Training Field Schools
in 2008, 2009 and 2013. Particular gratitude is due to all the UDSM students and former students
who assisted with the excavations and finds processing for all their hard work and diligent record-
ing. We also appreciate the guidance and advice provided by Prof. Terry O’Connor to TB on the
analysis of the faunal remains; Benson Kimeu’s help with the topographical surveys; the support
and insights offered by the other HEEAL team members, Daryl Stump, Matthias Heckmann,
Ashley Coutu and Pauline von Hellermann; and the practical and logistical support provided
by the University of Dar es Salaam, the British Institute in Eastern Africa, the University of
York and local residents around the study sites. Finally, we would like to thank two anonymous
reviewers, Daryl Stump, Federica Sulas and Peter Robertshaw for their helpful comments on an
earlier draft. We alone are responsible for any errors.
Notes on contributors
Thomas J. Biginagwa (PhD 2012, University of York) is an archaeozoologist and historical archae-
ologist based in the Department of Archaeology and Heritage Studies at the University of Dar es
Salaam. His doctoral dissertation examined animal economies practised by local communities
against the context of nineteenth-century caravan trade expansion in the Lower Pangani, north-
eastern Tanzania. His current research interests lie in historical ecology and zooarchaeology,
exploring human-environmental interactions over the last two millennia.
Paul Lane (PhD 1986, University of Cambridge) is the Jennifer Ward Oppenheimer Professor of
the Deep History and Archaeology of Africa in the Department of Archaeology, University of
Cambridge, and an Honorary Research Fellow in the School of Geography, Archaeology and
Environmental Studies, University of the Witwatersrand. His interests include post-colonial and
Indigenous archaeology in Africa, landscape historical ecology, maritime archaeology and the
origins and development of farming and herding in eastern Africa.
Contributions
TB co-designed the research, undertook data collection and faunal analyses, drafted the
original manuscript and contributed to its revision and editing. PL conceived and co-
designed the research, obtained funding and contributed to the writing and editing of
the manuscript.
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 27
ORCID
Thomas J. Biginagwa http://orcid.org/0000-0001-5558-3838
Paul J. Lane http://orcid.org/0000-0002-9936-1310
References
Abir, M. 1968. “Caravan trade and history in the northern parts of East Africa.” Paideuma 14: 103–
120.
Anderson, D.M. 1989. “Agriculture and irrigation technology at Lake Baringo in the nineteenth
century.” Azania 24: 84–97.
Ashton, A. 2000. “Ecology of rodent communities in agricultural habitats in eastern Sierra Leone:
cocoa groves as forest refugia.” Tropical Ecology 41: 127–142.
Assogbadjo, A.E., Codjia, J.T.C., Sinsin, B., Ekue, M.R.M. and Mensah, G.A. 2005. “Importance of
rodents as a human food source in Benin.” Belgian Journal of Zoology 135: 11–15.
Bailey, R.G., Churchfield, S., Petr, T. and Pimm, R. 1978. “The ecology of the fishes in Nyumba ya
Mungu reservoir, Tanzania.” Biological Journal of the Linnean Society 10: 109–137.
Baumann, O. 1891. Usambara and Seine Nachbargebiete. Berlin: D. Reimer.
Beachey, R. 1967. “The East African ivory trade in the nineteenth century.” Journal of African
History 8: 269–290.
Behrensmeyer, K. 1978. “Taphonomic and ecologic information from bone weathering.”
Palaeobiology 4: 150–162.
Beidelman, T. 1982. “The organization and maintenance of caravans by the Church Missionary
Society in Tanzania in the nineteenth century.” International Journal of African Historical
Studies 15: 601–623.
Bessems, I., Verschuren, D., Russell, J.M., Hus, J., Mees, F. and Cumming, B.F., 2008.
“Palaeolimnological evidence for widespread late 18th century drought across equatorial East
Africa.” Palaeogeography, Palaeoclimatology, Palaeoecology 259: 107–120.
Biginagwa, T.J 2012. “Historical archaeology of the 19th century caravan trade in north-eastern
Tanzania: a zooarchaeological perspective.” PhD diss., University of York.
Biginagwa, T.J. and Ichumbaki, E.B. 2018. “Settlement history of the islands on the Pangani River,
northeastern Tanzania.” Azania: Archaeological Research in Africa 53: 63–82.
Brain, C.K. 1981. The Hunters or the Hunted: An Introduction to African Cave Taphonomy.
Chicago: University of Chicago Press.
Bunn, H. and Kroll, M. 1986. “Systematic butchery by Plio-Pleistocene hominids at Olduvai
Gorge, Tanzania.” Current Anthropology 29: 135–149.
Burton, R.F. 1859. “The Lake Regions of Central Equatorial Africa.” Journal of the Royal
Geographical Society of London 29: 1–454.
Burton, R.F. 1872. Zanzibar: City, Island, and Coast. London: Tinsley.
Burton, R.F and Speke, J. 1858. “A coasting voyage from Mombasa to the Pangani River; visit to
Sultan Kimwere; and progress of the expedition into the interior.” Journal of the Royal
Geographical Society of London 28: 188–226.
Ceppi, S.L. and Nielsen, M.R. 2014. “A comparative study on bushmeat consumption patterns in
ten tribes in Tanzania.” Tropical Conservation Science 7: 272–287.
Christianson, C. 1981. Soil Erosion and Sedimentation in Semi-Arid Tanzania: Studies of
Environmental Changes and Ecological Imbalance. Uppsala: Scandinavian Institute of African
Studies.
Coutu, A.N., Lee-Thorp, J.A., Collins, M.J. and Lane, P.J. 2016. “Mapping the elephants of the 19th
century East African ivory trade with a multi-isotope approach.” PLoS ONE 11 (10): e0163606b.
Croucher, S.K. 2015 Capitalism and Cloves: An Archaeology of Plantation Life on Nineteenth-
Century Zanzibar. London: Springer.
Dahl, G. and Hjort, A. 1976. Having Herds: Pastoral Herd Growth and Household Economy.
Stockholm: University of Stockholm, Department of Social Anthropology.
28 T. J. BIGINAGWA AND P. J. LANE
Driver, J. 2008. “Models for zooarchaeologists from modern bushmeat studies.” In Animals and
Archaeology: Archaeozoological Papers in Honour of Ina Plug, edited by S. Badenhorst, P.J.
Mitchell and J. Driver, 21–33. Oxford: Archaeopress.
Erhardt, J. 1853. “Journal containing an account of my journey to Usambara and back and of three
months stay with the king of the country, Kimeri, from the 9th of August to December 1853.”
Church Missionary Society Microfilm CA5/09/16,35.
Farler, J.P 1882. “Native routes in East Africa from Pangani to the Masai country and Victoria
Nyanza.” Journal of the Royal Geographical Society of London 4: 730–742.
Fischer, W. and Bianchi, G. 1984. FAO Species Identification Sheets for Fishery Purposes: Western
Indian Ocean (Fishing Area 51). Rome: FAO.
Foutch, A.E., Steele, T.E. and O’ Brien, C.J. 2009. “Faunal analysis from Kibaoni, a late precolonial
Pimbwe village in Rukwa Valley, Tanzania: first reconstructions of cultural and environmental
histories.” Azania: Archaeological Research in Africa 44: 257–267.
Giblin, J. 1992. The Politics of Environmental Control in North-Eastern Tanzania, 1840–1940.
Philadelphia: University of Pennsylvania.
Gifford, D., Isaac, G.L. and Nelson, C. 1980. “Evidence for predation and pastoralism at Prolonged
Drift, a Pastoral Neolithic site in Kenya.” Azania 15: 57–108.
Gissibl, B. 2016. The Nature of German Imperialism: Conservation and the Politics of Wildlife in
Colonial East Africa. Oxford: Berghahn.
Håkansson, T.N. 1995. “Irrigation, population pressure, and exchange in precolonial Pare,
Tanzania.” Research in Economic Anthropology 16: 297–324.
Håkansson, T.N. 2004. “The human ecology of world systems in East Africa: the impact of the
ivory trade.” Human Ecology 32: 561–591.
Håkansson, T.N. 2008. “The decentralized landscape: regional wealth and the expansion of pro-
duction in northern Tanzania before the eve of colonialism.” In Economies and the
Transformation of Landscape, edited by L. Cligget and C. Pool, 239–266. Walnut Creek:
AltaMira Press.
Håkansson, T.N., Widgren, M. and Börjeson, L. 2008. “Introduction: historical and regional per-
spectives on landscape transformation in north-eastern Tanzania, 1850–2000.” International
Journal of African Historical Studies 41: 369–382.
Helm, R.M. 2000. “Conflicting histories: the archaeology of the iron-working, farming commu-
nities in the central and southern coast region of Kenya.” PhD diss., University of Bristol.
Herrero, M., Grace, D., Njuki, J., Johnson, N., Enahoro, D., Silvestri, S. and Rufino, M. 2013. “The
roles of livestock in developing countries.” Animal 7: 3–18.
IUCN. 2003. The Pangani River Basin: A Situation Analysis. Nairobi: IUCN.
Johnston, K. 1879. “Notes of a trip from Zanzibar to Usambara in February and March 1879.”
Journal of the Royal Geographical Society of London 1: 545–558.
Juwayeyi, Y. 2008. “Human and animal interaction on the Shire Highlands, Malawi: the evidence
from Malowa rock shelter.” In Animals and People: Archaeozoological Papers in Honour of Ina
Plug, edited by S. Badenhorst, P.J. Mitchell and J. Driver, 83–93. Oxford: Archaeopress.
Kawiche, K. 2016. “Archaeobotanical evidence of crop production during the nineteenth century
caravan trade in Lower Pangani, north-eastern Tanzania.” MA diss., University of Dar es
Salaam.
Kimambo, I. 1996. “Environmental control and hunger in the mountains and plains of nineteenth
century north-eastern Tanzania.” In Custodians of the Land: Ecology and Culture in the History
of Tanzania, edited by G. Maddox, J. Giblin and I. Kimambo, 71–95. London: James Currey.
Kiptala, J.K., Mohamed, Y., Mul, M.L., Cheema, M.J.M. and Van der Zaag, P. 2013. “Land use and
land cover classification using phenological variability from MODIS vegetation in the Upper
Pangani River Basin, Eastern Africa.” Physics and Chemistry of the Earth, Parts A/B/C 66: 112–122.
Kisbey, W. 1898. “The burning of Kwa Sigi village.” African Tidings 7(109).
Kitalyi, A.J. 1998. Village Chicken Production Systems in Rural Africa: Household Food Security and
Gender Issues. Rome: FAO.
Kjekshus, H. 1996. Ecology Control and Economic Development in East African History: The Case of
Tanganyika 1850–1950. Oxford: James Currey.
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 29
Koponen, J. 1988. People and Production in Late Pre-Colonial Tanzania: History and Structures.
Uppsala: Scandinavian Institute of African Studies.
Krapf, J. 1860 [1968]. Travels, Researches, and Missionary Labours During an Eighteen Years’
Residence in Eastern Africa. Abingdon: Frank Cass and Co.
Kusimba, C.M. 2004. “Archaeology of slavery in East Africa.” African Archaeological Review 21:
59–88.
Lane, P.J. 2010. “Developing landscape historical ecologies in eastern Africa: an outline of current
research and potential future directions.” African Studies 69: 299–322.
Lane, P.J. 2011. “Slavery and slave trading in eastern Africa: exploring the intersections of histori-
cal sources and archaeological evidence. In Slavery in Africa: Archaeology and Memory, edited
by P.J. Lane and K.C. MacDonald, 281–314. Oxford: Oxford University Press
Lowe, R.H. 1955. New Species of Tilapia (Pisces, Cichlidae) from Lake Jipe and the Pangani River,
East Africa. London: British Museum (Natural History).
Lynwood, F. 1994. Rodent Pest Management in Eastern Africa. Rome: FAO Plant Production and
Protection Paper 123.
Malaisse, F. 1997. Se Nourrir en Forêt Claire Africaine: Approche Ecologique et Nutritionnelle.
Wageningen: CTA Spore 75.
Marshall, F. 1990. “Cattle herds and caprine flocks: early pastoral strategies in south-west Kenya.”
In Early Pastoralists of South-Western Kenya, edited by P.T. Robertshaw, 205–260. Nairobi:
British Institute in Eastern Africa.
Meadow, S. and White, J. 1979. Structure of the Herd and Determinants of Offtake Rates in Kajiado
District in Kenya, 1962–1977. London: Overseas Development Institute.
Mgombere, C. 2017. “Archaeological evidence of crop economies practiced along the nineteenth
century caravan route in southern Tanzania.” MA diss., University of Dar es Salaam.
Mutundu, K. 1999. Ethnohistoric Archaeology of the Mukogodo in North-Central Kenya: Hunter-
Gatherer Subsistence and the Transition to Pastoralism in Secondary Settings. Oxford: British
Archaeological Reports.
New, C. 1874. “Journey from the Pangani, via Wadigo, to Mombasa.” Journal of the Royal
Geographical Society of London 19: 317–323.
Opara, M. N. 2010. “Grasscutter: the haematology and major parasites.” Research Journal of
Parasitology 5: 214–223.
Pallaver, K. 2009. “‘A recognized currency in beads.’ Glass beads as money in 19th-century East
Africa: the central caravan road”. In Money in Africa, edited by C. Eagleton, H. Fuller and J.
Perkins, 20–29. London: British Museum.
Pallaver, K. 2012. Muslim communities, long-distance trade and wage labour along the central
caravan road: Tanzania, 19th century. Storicamente 8 (2012): 20. https://storicamente.org/
pallaver_tanzania
Petek, N. 2018. Archaeological Perspectives on Risk and Community Resilience in the Baringo
Lowlands, Kenya. Uppsala: Department of Archaeology and Ancient History, Uppsala
University.
Prendergast, M.E., Buckley, M., Crowther, A., Frantz, L., Eager, H., Lebrasseur, O., Hutterer, R.,
Hulme-Beaman, A., Van Neer, W., Douka, K. and Veall, M.A. 2017. “Reconstructing Asian
faunal introductions to eastern Africa from multi-proxy biomolecular and archaeological data-
sets.” PLoS ONE 12(8): e0182565.
Prendergast, M.E. and Lane, P.J. 2010. “Middle Holocene fishing strategies in East Africa: zooarch-
aeological analysis of Pundo, a Kansyore shell midden in northern Nyanza (Kenya).”
International Journal of Osteoarchaeology 20: 88–112.
Prendergast, M.E. and Mutundu, K. 2009. “Late Holocene zooarchaeology in East Africa: ethno-
graphic analogues and interpretive challenges.” Documenta Archaeobiologiae 7: 203–232.
Prendergast, M.E., Quintana Morales, E.M., Crowther, A., Horton, M.C. and Boivin, N.L. 2017.
“Dietary diversity on the Swahili coast: the fauna from two Zanzibar trading locales.”
International Journal of Osteoarchaeology 27: 621–637.
Prestholdt, J. 2004. “On the global repercussions of East African consumerism.” American
Historical Review 109: 755–781.
30 T. J. BIGINAGWA AND P. J. LANE
Quinlan, J., Rumas, I., Naisikye, G., Quinlan, M. and Yoder, J. 2016. “Searching for symbolic value
of cattle: tropical livestock units, market price, and cultural value of Maasai livestock.”
Ethnobiology Letters 7: 76–86.
Reid, D.A.M. 1996. “Cattle herds and redistribution of cattle resources.” World Archaeology 28:
43–57.
Rempel, R. 1998. “Trade and transformation: participation in the ivory trade in late 19th-century
East and Central Africa.” Canadian Journal of Development Studies 19: 529–552.
Robertson, C.C. 1997. “Gender and trade relations in central Kenya in the late nineteenth century.”
International Journal of African Historical Studies 30: 23–47.
Rockel, S. 2006. Carriers of Culture: Labour on the Road in Nineteenth-Century East Africa. Oxford:
James Currey.
Rodney, W. 1972. How Europe Underdeveloped Africa. London: Bogle-L’Ouverture Publications.
Ryan, K., Munene, K., Kahinju, S.M. and Kunoni, P.N. 2000. “Ethnographic perspectives on cattle
management in semi-arid environments: a case study from Maasailand.” In The Origins and
Development of African Livestock: Archaeology, Genetics, Linguistics, and Ethnography, edited
by R.M. Blench and K.C. MacDonald, 462–477. London: Routledge.
Seegers, L. 2008. “The fishes collected by G.A Fischer in East Africa in 1883 and 1885/86.”
Zoosystematics and Evolution 84: 149–195.
Shenkut, M., Mebrate, A. and Balakrishnan, M. 2006. “Distribution and abundance of rodents in
farmlands: a case study in Alleltu Woreda, Ethiopia.” SINET: Ethiopian Journal of Science 29:
63–70.
Sheriff, A. 1987. Slaves, Spices and Ivory in Zanzibar: Integration of an East African Commercial
Empire into the World Economy 1770–1873. London: James Currey.
Silver, I.A. 1969. “The ageing of the domestic animals.” In Science in Archaeology, edited by D.
Brothwell and E.S. Higgs, 283–302. London: Thames and Hudson.
Sissons, J. 1984. “Economic prosperity in Ugogo, East Africa, 1860–1890.” PhD diss., University of
Toronto.
Skelton, P. H. 1994. “Diversity and distribution of freshwater fishes in East and southern Africa.”
Annales du Musée Royal de l’Afrique Centrale Sciences Zoologiques 275: 95–131.
Speth, J.D. and Scott, S.L. 1989. “Horticulture and large-mammal hunting: the role of resource
depletion and the constraints of time and labor.” In S. Kent (ed.), Farmers as Hunters: The
Implications of Sedentism, edited by S. Kent, 71–79. Cambridge: Cambridge University Press.
Stahl, P. 1982. “On small mammal remains in archaeological context.” American Antiquity 47:
822–829.
Stump, D. and Tagseth, M. 2009. “The history of pre-colonial and early colonial agriculture on
Kilimanjaro: a review.” In Culture, History and Identity: Landscapes of Inhabitation in the
Mount Kilimanjaro Area, Tanzania, edited by T. Clack, 107–124. Oxford: Archaeopress.
Tamatamah, R.A. 2007. “Environmental flow assessment (EFA) Wami River Basin, Tanzania:
aquatic ecology component of the Wami River EFA study. Literature review for BBM work-
shop.” Unpublished report, Wami-Ruvu Basin Water Office, Morogoro, Tanzania.
Thorbahn, P.F. 1979. “The pre-colonial ivory trade of East Africa: reconstruction of a human-ele-
phant ecosystem.” Ph.D. diss., University of Massachusetts.
URT (United Republic of Tanzania). 2008. Korogwe District Investment Profile. Dar es Salaam:
Government Printer.
VonHöhnel 1894Discovery of Lakes Rudolf and StefanieVol. II. London: Longmans, Green andCo.
Walker, N.J. 1995. Late Pleistocene and Holocene Hunter-Gatherers of the Matopos: An
Archaeological Study of Change and Continuity in Zimbabwe. Uppsala: Societas
Archaeologica Upsaliensis.
Walker, R. 1985. A Guide to Postcranial Bones of East African Animals. Norwich: Hylochoerus.
Walz, J.R. 2010. “Route to a regional past: an archaeology of the Lower Pangani (Ruvu) Basin,
Tanzania, 500-1900 C.E.” PhD diss., University of Florida.
Wynne-Jones, S. 2010. “Lines of desire: power and materiality along a Tanzanian caravan route.”
Journal of World Prehistory 23: 219–237.
AZANIA: ARCHAEOLOGICAL RESEARCH IN AFRICA 31