Quaternary International xxx (2015) 1e13

Contents lists available at ScienceDirect

Quaternary International
journal homepage: www.elsevier.com/locate/quaint

Middle and Late Stone Age of the Niassa region, northern

Mozambique. Preliminary results
Nuno Bicho a, *, Jonathan Haws b, Mussa Raja c, Omar Madime c, Ce
lia Gonalves a,
~o Cascalheira a, Michael Benedetti d, Telmo Pereira a, Vera Aldeias e
Joa
a
ICArEHB, FCHS, Universidade do Algarve, Campus de Gambelas, 8005-139 Faro, Portugal
b
Department of Anthropology, University of Louisville, Louisville, USA
c
Departamento de Arqueologia e Antropologia, Universidade Eduardo Mondlane, Maputo, Mozambique
d
Department of Geography and Geology, University of North Carolina e Wilmington, Wilmington, USA
e
Department of Human Evolution, Max Plank Institute for Evolutionary Anthropology, Leipzig, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Located between modern-day South Africa and Tanzania, both of which have well-known and extensive
Available online xxx Stone Age records, Mozambique's Stone Age sequence remains largely unknown in the broader context
of African Pleistocene prehistory. Such lack of data occurs despite the key geographical location of the
Keywords: country, in southern Africa at the southeastern tip of the Great Rift Valley. As such, Mozambique is an
Middle Stone Age area of interest to evaluate the origins and dispersion of Homo sapiens within Africa, particularly in
Late Stone Age
relation to Middle Stone Age contexts and associated early modern human ecology and cognition.
Mozambique
This paper focuses on preliminary survey results from the Niassa District, near Lake Niassa (also known
Lake Malawi
as Lake Malawi) in northern Mozambique. The results include the discovery and location of more than 80
new surface lithic concentration localities, as well as data from two new sites, the open air surface site of
Ncuala and the rock shelter of Chicaza. For Chicaza we provide a series of new radiocarbon dates for the
Iron Age and Late Stone Age occupations based on preliminary testing carried out at the site.
2015 Elsevier Ltd and INQUA. All rights reserved.

1. Introduction symbolic behavior and expansion of ecological niches by anatom-

The anatomical and behavioral evolution of early modern
humans within Africa is a main topic of research in Paleoanthro-
pology today, likely because the emergence of our species is a key
event in human history. Our evolution has been thought to be a
slow and gradual cultural evolution in Africa, starting perhaps as
early as 200 ka. While the earliest fossil evidence for anatomically
modern humans (AMH) seems to come from east Africa (Clark et al.,
2003; White et al., 2003; McDougall et al., 2005), the emergence of
modern cognition is apparently later and largely based on evidence
from southern Africa as early as 160 ka ago (Marean et al., 2007;
Marean, 2014).
In the last decade, eastern and southern Africa have become the
central region for studying the emergence of modern human
behavior (Wadley, 2001; Henshilwood and Marean, 2003; Marean
and Assefa, 2005; Marean et al., 2007). Many new discoveries of
* Corresponding author.
E-mail address: nbicho@ualg.pt (N. Bicho).
ically modern humans come from caves in the Western Cape region
of South Africa (Henshilwood et al., 2004, 2011; Marean et al.,
2007). In the eastern portion of South Africa, excavations at
Sibudu rockshelter have shed additional light on modern human
behavior (d'Errico et al., 2008; Wadley, 2010). Despite these dis-
coveries, little is known about the origin, timing and processes
through which these new behaviors spread. The major question,
thus, is why the emergence of AMH, both biological and culturally,
took place in specic regions of Africa. There are still large regions
in eastern and southern Africa for which we lack detailed infor-
mation for the Middle Stone Age (MSA), the critical period to un-
derstand the processes of AMH emergence. This patchiness of data
makes it difcult to address broader questions on the origins and
dispersion of Stone Age cultures at a continental scale.
Mozambique's Stone Age sequence remains largely unknown,
despite its key geographical location, between eastern and south-
ern Africa. For instance, the well-known map published by
McBrearty and Brooks (2000: 478, Fig. 1) showed a remarkable
density of sites in eastern and southern Africa but the entire area of
Mozambique was blank, despite punctual previous eldwork

http://dx.doi.org/10.1016/j.quaint.2015.09.059
1040-6182/ 2015 Elsevier Ltd and INQUA. All rights reserved.

Please cite this article in press as: Bicho, N., et al., Middle and Late Stone Age of the Niassa region, northern Mozambique. Preliminary results,
Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.059
2 N. Bicho et al. / Quaternary International xxx (2015) 1e13

Fig. 1. Map with the location of the Stone Age sites known up to 2013 in Mozambique.

developed up to 1980. This apparent lack of relevant archaeological
contexts does not necessarily reect the nonexistence of such re-
cords, but it does highlight an important gap in archaeological
research in the country. For the few sites we do know from
Mozambique, there is often little or no accurate information on the
chronology or cultural characteristics of the assemblages other
than generic attributions to Early Stone Age (ESA), Middle or Late
Stone Age (LSA) phases.
A compilation of the known archaeological sites in Mozambique
was published by Joa ~o Morais some 30 years ago (Morais, 1984).
There were roughly 100 Stone Age sites listed, resulting from
geologic and archaeological survey carried out by Lereno Barradas,
an agronomic engineer who carried out geologic and archaeolog-
ical survey in southern Mozambique (Barradas, 1945, 1947, 1955,
1964, 1965; Adamowicz and Nhatule, 2011), and eld campaigns
by Santos Jnior in the context of the so-called Anthropological
Missions of Mozambique (Missa ~o Antropolo
gica de Moambique).
The latter involved six missions respectively in 1936, 1937, 1945,
1946, 1948, and 1955/56 (Rodrigues, 1998/1999, 2004), with the
identication of close to 90 Stone Age sites incorporated into a
series of archaeological maps of Mozambique (Santos Jnior, 1937,
1938, 1940, 1941, 1944, 1946, 1947, 1950) and reviewed by Roque
(2002) and by Roque and Ferra ~o (2004). No excavations were car-
ried out except for some testing in the rock shelter of Caimane, a

Please cite this article in press as: Bicho, N., et al., Middle and Late Stone Age of the Niassa region, northern Mozambique. Preliminary results,
Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.059
 N. Bicho et al. / Quaternary International xxx (2015) 1e13
site located some 70 km west of the city of Maputo (Morais, 1984).
This site provided a sedimentary sequence with ESA, MSA, and LSA
layers capped by Iron Age occupations, though little additional
information is available (Morais, 1984). Thus, at the time of
McBrearty and Brooks publication, Mozambique had only Stone
Age surface sites and a single tested site with little or no published
information and no absolute dating results.
Since then, however, a few Stone Age sites have been located in
various parts of the country, usually as the result of surveys within
various projects focusing on recent chronologies. The only
exception was the project by J. Mercader in the districts of Lago
and Sanga in the Niassa District in northern Mozambique fully
dedicated to the Stone Age (Mercader et al., 2008, 2009, 2012,
2013). Between 2009 and 2013, this project was redirected to
the Gorongosa (Sofala), Namapa, Montepue ^s, Pemba and Quis-
sanga (Cabo Delgado) regions, but only preliminary work was
carried out in the Gorongosa region. As a result of the work done
by Mercader's team, a few absolute dates are now available for the
cave site of Ngalue, in the Lago district with MSA and LSA occu-
pations (Mercader et al., 2009). In addition, Mercader also exca-
vated and dated two open-air sites, Mikuyu and Mvumu
(Mercader et al., 2008, 2012, 2013), both located on the slopes
above Lake Niassa. The dating results point to a MSA chronology
starting around 150 ka. There are no absolute dates for ESA con-
texts in Mozambique.
In 2011, our team started a new research program to study the
MSA in Mozambique in order to ll the current scientic hiatus for
the country. A preliminary reconnaissance of the southern country
was carried out in 2011 and 2012 funded by Wenner Gren Foun-
dation to one of the authors (JH). In 2014, the team started an
ongoing survey and testing project in various regions of the country
funded by the Portuguese Science Foundation (Fundaa ~o para a
^ncia e Tecnologia), including the regions of Niassa in the north
Cie
and Limpopo in the south.
Our aim was to nd and test new sites and collect artifacts,
fauna and dating samples in order to shed light on the origins of
behavioral traits that are thought to be hallmarks of human
modernity, e.g., use of pigments and beads for personal adorn-
ment and denition of individual and group identity, exploita-
tion of aquatic resources such as shellsh, sh, and use of bone
points. The long-term goal of the present project is to contribute
to the knowledge of human evolution by discovering new
archaeological and paleoanthropological sites in a region that is
largely unknown due to historical circumstances. We chose
Mozambique because it is located between East Africa (where
the earliest AMH have been found some 200 ka) and Southern
Africa (the region where the earliest signs of modern cognition
have been found, dated to c. 165 ka), and thus it is potentially a
fundamental region to investigate the true origin of the AMH as
well as to better understand the migration trajectories of early
AMH.
Here, we present the preliminary results of eld survey in the
Lake Niassa region, in northern Mozambique, which took place
during the month of July in 2014. We focused on three main
geologic contexts:
1) Both margins of the Lunho River Valley with its estuary running
to the Lake Niassa, near the town of Metangula;
2) The inland limestone valley of Chitete, in the region locally
known as Malulu;
3) The inland outcrops of Xiugulu and Unango.
The rst area was targeted for both open air and rock shelters,
while we focused mostly on rock shelters and cave contexts in the
remaining areas.
Please cite this article in press as: Bicho, N., et al., Middle and Late Stone Age of the Niassa region, northern Mozambique. Preliminary results,
Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.059
3
2. The natural setting
Lake Niassa is the southern most extreme of the Rift Valley. The
lake is currently roughly at 460 m above sea level (a.s.l.), while the
adjacent territory rising in certain points up to over 1800 m a.s.l.
From a geologic point of view, the area is marked by the presence of
two major units: the Unango complex and the Karoo Supergroup
(Grantham et al., 2011). The Unango complex dominates the region,
extending from Tanzania to the Malawi border in the southern tip
of the Niassa Lake. This complex is dominated by diverse gneisses
and paragneisses (La chelt, 2004; Grantham et al., 2011) locally
marked by the presence of quartz and quartzites. The Unango
complex is partially overlain by the Karoo sedimentary rocks that
ll the Maniamba Graben (Grantham et al., 2011).
The Karoo Supergroup in the Niassa region is represented by its
Lower and Upper phases. The Upper Karoo is exposed in the
southern margin of the Lunho Valley and is known as the Ecca
Group (Grantham et al., 2011), composed by conglomerates with
pebbles in coarse-grained sandstones, locally cemented by calcite.
The upper part of the sequence is composed of medium-grained
sandstones, siltstones and mudstones, frequently interbedded
with carbonaceous beds. There are also Neoproterozoic intrusions
in the region, including the so-called Monte Neumale Syenite,
forming the Unango outcrop. It comprises a complex of leucocratic,
alkaline, and syenitic rocks.
The Lunho alluvial valley is lled with reddish silty sands of
Pleistocene and Holocene chronologies, while the Niassa lake
margins are marked by light colored medium-grained sands
(Grantham et al., 2011), corresponding to what Mercader described
as Luchamange Beds (Mercader et al., 2012).
Rain patterns in the region are fairly consistent, with a rainy
season during the summer months, and a common drought during
the winter, from June to December. The mean annual rainfall runs
from 700 mm in the low lands around the lake margins while in the
higher elevation the pattern goes up to 1400 mm (Gama, 1990).
Vegetation is fairly open in the lowlands and in the Lunho valley,
and denser forests are located in the highlands, as it is the case of
the Xiugulu and Malulu areas.
3. First mapping results
Based on the early maps from Santos Jnior (1950) and more
recent data acquired through various projects, namely that of
Mercader in northern Mozambique (Mercader et al., 2009), we
were able to produce a series of maps for the Stone Age prehistory
of Mozambique (Gonalves et al., 2014, in press). The maps were
also based on a critical evaluation of the sites and a review of some
of the materials that are presently curated at the IICT (Instituto de
Investigaa~o Cientca e Tropical) in Lisbon, Portugal. We have also
developed an online database with specic information for each of
those sites (accessible at http://icarehb.com/index.php/historical-
background-mz).
These efforts resulted in a total of 258 referenced Stone Age
sites dispersed through all of Mozambique, but with clear con-
centrations mostly in the Maputo, Sofala and Tete regions (Fig. 1).
We were able to list geographic locations for 141 sites, of which
70 are currently precisely relocated. Overall, 53 sites are generi-
cally attributed to Stone Age chronologies, while the remaining
include 79 multicomponent sites (mostly with ESA and MSA ar-
tifacts), 68 ESA, 39 MSA, and 19 LSA. Although these renewed
relocation and mapping efforts greatly increased the known
number of Stone Age sites in Mozambique, several areas, namely
large regions of the Cabo Delgado, Nampula and Inhambane
Districs, are still characterized by a lack of research in Stone Age
archaeology.
4 N. Bicho et al. / Quaternary International xxx (2015) 1e13
4. Survey methods
4.1. Ground-covering methods
A non-systematic survey was carried out by foot, directed to
specic areas where geomorphology and geology increased the
chances to nd Stone Age sites in both uvial terraces and/or caves
and rockshelters. Survey was aided by the use of a DJI Phantom 2
drone equipped with a GoPro 3 camera with live feed. This was, in
many situations, an important tool in our work since it allowed for
better and easier decisions of, for example, where to go or how to
more easily move the team into a specic area.
A set of variables was used to decide on the daily targets. In
limestone areas, we targeted all reachable points with visible
landscape elements that could have either caves or shelters such as
vertical outcrops. These outcrops were located remotely, based on
the geologic and topographic cartography, as well as on the infor-
mation obtained from the drone data. The paths were usually
decided by the local guides, frequently following traditional routes
and, more rarely, open ab initio by the team. Target and path de-
cisions were similar in the case of volcanic outcrops of Xiugulu and
Unango.
In the case of the Lunho Valley, the team tried to cover as much
terrain as possible, starting on the alluvial plain and going up on
either margin to the limit of the Lunho basin. Sometimes, tran-
septs were walked parallel to the river at different altitudes,
usually corresponding to each terrace elevation. The team ele-
ments would walk parallel lines, as close to each other as the
vegetation allowed: while in many cases the savannah like land-
scape permitted short distance between the individual transepts
(i.e., ve meters) and thus visual coverage was close to 100%, in
other cases where vegetation was marked by closed bushes and
trees, distances of a couple hundred meters could separate each of
the team elements. In more radical cases, both in the Lunho Valley
or in the other contexts, the team would walk in a single line to
reach open areas or the target location. In the Mlulu region
tropical forest was present in the entire valley and covered close to
100% of the ground oor; in the other survey areas vegetation was
not close to the surface and the ground was mostly visible and
artifacts were easily located. The results obtained, thus, are highly
diverse in terms of terrain coverage and depended heavily on local
vegetation.
4.2. Survey recording methodology
For site recording the team developed an Android-based
software application (ArcheoSurvey e Cascalheira et al., 2014)
that, although designed and programmed to our specic survey
needs, is fully customizable and it is available freely for download
at http://icarehb.com/index.php/eld-methods-and-techniques-
mz. This app was used in smartphones and tablets and
included a set of variables that allowed the description of the
main characteristics of the sites and their specic surroundings.
Most variables had preset attributes that can be selected from
drop-down menus, such as the cases of site type (Open-air,
Rockshelter, Cave, Other) or visibility (Bad, Reasonable, Good),
while others are simple numeric or text eld where the sur-
veyors can input information on e.g. average area of artifact
dispersion (in square meters) and specic description of the
ndings.
The app also allows to take one general photograph of the site
and to record the absolute coordinates using the GPS chips of
each device. We have tested, in advance, the accuracy of GPS
readings from several brands of smartphones and tablets against
the values obtained from a Trimble Geo XT and the results were
Please cite this article in press as: Bicho, N., et al., Middle and Late Stone Age of the Niassa region, northern Mozambique. Preliminary results,
Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.059
acceptable for our purposes with differences always between 2
and 6 m (Cascalheira et al., 2014). Nonetheless, we have also used
the Trimble Geo XT, with an accuracy of 0.5 m, during survey for
mapping all the main survey paths carried out by the team. The
GPS was set for taking a reading every second, providing an ac-
curate and complete information on each path carried out by the
team during survey. ArcheoSurvey data are stored in each
smartphone/tablet internal memory in a comma separated text
le (CSV) that can be easily imported into most database soft-
wares. All our points (and respective information) and paths were
imported to MS Excel and Google Earth and checked for errors on
a daily basis.
Collection of surface materials was limited to both examples of
the range of raw materials present in each site as well as to diag-
nostic artifacts that could help to identify the chronology or chro-
nologies of the human occupation. There a few exceptions such as
the case of Xiugulu, where we were asked not to collect materials
due to local religious beliefs. The decision of not collecting the
entire ensemble of surface materials was partially based on two
aspects:
a) Since in many instances the surrounding landscape offered no
distinct visual location points, it would be very difcult to go
back, even with the aid of GPS, and relocate the site if there were
no surface artifacts e thus, the collection of all artifacts would
prevent the location of the site in the future;
b) In many cases, due to the lack of roads, the team walked be-
tween 10 and 20 kms and in occasions more than 30 kms every
day. To carry all the surface artifacts found in a single day would
have been too much weight for transportation e still and even
with the limited number of artifacts we collected, we obtained
over 35 kg of lithic artifacts.
5. Results
In our survey, we were able to locate 64 open air localities, ve
caves and 20 rockshelters. Of these, 21 localities can be generally
assigned to MSA, nine to LSA and seven possible multicomponent
ones with MSA and LSA (Fig. 2). We also identied one Iron Age
open-air location, 20 indeterminate Stone Age sites and ve others
with no clear identied chronological elements. No ESA artifacts
were found during survey, although these can potentially be
located in the upstream sections of the valley not targeted during
this eldwork.
At least one case corresponds to a very recent occupation (site
O1) since the lithic assemblage, composed of quartz artifacts
including akes, blades, cores and a crescent, was all contained
within daub square ruin of a recent eroded house. There is no
information on contemporary knapping, but these artifacts seem
very similar, both in technology and raw material, to other LSA
surface scatters we found. The implications are very important for
the future work in the region and for the understanding of Stone
Age lithic technology, since these materials are similar to those
we found in the Lunho valley. We did not collect all the materials
on the surface, but only those that seem technologically diag-
nostic, that is cores and a set of akes. In the future we intend to
go back and collect all the surface materials and possibly test the
location within the walls to gather a sizable assemblage for
detailed study.
5.1. Lunho Valley: sites and location patterns
The number of localities in the Lunho Valley with surface ma-
terials (n 59) seems to present a pattern in terms of altitude and
site distribution (Table 1). While in the northern bank, ndings
 N. Bicho et al. / Quaternary International xxx (2015) 1e13 5

Fig. 2. Map with the location of the new located sites from the 2014 eld season in the Niassa region, northern Mozambique.

Table 1
seem to be more concentrated near the estuary, in the southern
Altitude information on Lunho Valley Stone Age sites.
bank Stone Age materials are essentially located in small patches
along the valley, starting close to the Lake and stretching up to Average Minimum Maximum

60 kms from the estuary. The preliminary chronological attribution All sites 583 487 886
of these localities seems to indicate that the ratio between MSA and All sites in the southern bank 628 524 886
All sites in the northern bank 555 487 615
LSA locations is also different between the two banks. It is also clear
All MSA total 570 495 822
that the southern bank has more MSA sites then LSA, while the All LSA total 562 495 600
northern bank seems to have been extensively used during both MSA sites in the northern bank 545 495 596
Stone Age phases. This pattern may be a consequence of modern LSA sites in the northern bank 567 495 600
surface visibility or even differential erosional rates on the two MSA sites in the southern bank 609 527 822
LSA sites in the southern bank 545 534 566
banks.

Please cite this article in press as: Bicho, N., et al., Middle and Late Stone Age of the Niassa region, northern Mozambique. Preliminary results,
Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.059
6 N. Bicho et al. / Quaternary International xxx (2015) 1e13

It seems clear that, in the southern bank, localities are spread Table 2
across a wider range of altitudes, from 524 up to 886 m a.s.l., while Lithic assemblage from Ncuala.

in the northern bank artifact concentrations are located at lower Class Chert Quartz Other Total
elevations, between 487 and slightly over 600 m a.s.l. The inter- Blade 1 1 2
esting aspect is while the pattern seems to be the same for the Blade Fragments 2 2
northern bank for both MSA and LSA localities, respectively with Bladelet 15 15
545 and 567 m, in the southern bank the pattern is very different. Bladelet Fragments 27 27
Chip 2 127 2 131
There is a clear difference between the two phases, with MSA sites
Chunk 18 531 60 609
using higher altitudes, between 520 and 820 m a.s.l., while LSA is Cobble 14 14
tightly constrained around the 540 m, although sites from both Core 6 136 1 143
phases are spread equally along the southern bank. In one Core Fragments 3 24 27
Flake 19 809 11 839
particular case, the site of Ncuala, we carried out more detailed
Flake Fragments 7 526 2 535
work, described below. Maintenance Products 1 5 6
Total 56 2217 77 2350

5.1.1. The site of Ncuala

Ncuala is an open-air site, located in northern bank of the Lunho
river valley at an altitude of c. 592 m a.s.l. Materials were dispersed
for a relatively at area of c. 350 m2, although a denser concen-
tration area with approximately 80 m2 was identied. We set a
10  5 meters grid and collected every visible artifact using
50  50 cm units. Additionally, a 1  1 meter test pit was excavated,
conrming that the archaeological level is completely exposed on
the surface.
A total of 2350 lithic artifacts were collected from Ncuala
(Table 2, Fig. 3). Most of the artifacts are made from local quartz
materials, both crystalline and milky. Other local materials and at
least two different kinds of chert and chalcedony were also used.
From a technological point of view the assemblage is composed
mostly by akes, very few blades and a small amount of bladelets.
When looking at the cortex presence, the majority of the artifacts
do not present cortical surfaces. In the case of quartz the rare
occurrence of cortex could be due to the fact that quartz present at
the sites in the Lunho valley seem to have eroded and was quarried
from local pegmatites as large chunks that were then broken apart
into core-sized cobbles without allowing time for signicant cortex
to develop on their surfaces (Bennett, 2011).
The assemblage also presents a relatively high percentage of
cores, the most relevant types being simple and bipolar cores.

Fig. 3. Lithic artifacts from Ncuala. A e Backed points in quartz; B e Bipolar cores in quartz.

Please cite this article in press as: Bicho, N., et al., Middle and Late Stone Age of the Niassa region, northern Mozambique. Preliminary results,
Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.059
 N. Bicho et al. / Quaternary International xxx (2015) 1e13 7

Bipolar technology was frequently applied in small quartz cobbles
that were exploited according to their longest natural axis (Fig. 3B).
Chronological attribution in this type of sites, with no absolute
dating or other solid age indicators, is always difcult. In addition to
the high representation of bipolar technology and the extremely
rare presence of prepared (Levallois or discoidal) technologies, the
assemblage from Ncuala has a considerably low presence of
retouched artifacts, being limited to the presence of notches, den-
ticulates, one side-scraper and three abruptly retouched points
made of quartz, two of them fractured in their distal parts (Fig. 3A).
5.2. Chitete Valley and the Chicaza rock shelter
In the Chitete River Valley survey was focused mostly on caves
and rockshelters. This valley is part of a set of other valleys that ow
NE/SW into the Nnodwesi's watershed, all surrounded by carbon-
ate deposits, known as the Malulu deposits, that are characterized
by high variability of limestone types. In this same valley, Mercader
et al. (2009) reported a set of archaeological ndings from the small
cave of Ngalue, attributed to the MSA on techno-typological
grounds and absolute chronology.
The site of Chicaza is located some 1300 m to the NE of Ngalue
and is a small shelter protected by a fairly high limestone scarp,
adjacent to a small permanent water shed that runs to the Chitete
River. The shelter has a at surface area of about 25 m2 where we
found on the surface modern ceramics and a re combustion
remains.
A 1 m2 test was dug down to 1.5 m deep (Fig. 4) following the
natural litho-stratigraphic units and subdivided into articial spits.
Due to the small excavation area and the large blocks present the
articial spits were fairly thick, with 10e15 cm each. We did not
reach either sterile horizons or bedrock due to time constraints as
well as the presence of large blocks at the base of the test pit. Based
on the general topography of the locus, it is likely that the deposit is
more than 3 m thick. The one square meter test revealed the
presence of 3 main layers, all composed of loose very ne sand with
artifacts, mammal bones, large terrestrial gastropod shells, charcoal
and large roots that reach the base of the excavated sequence.
A total of 5 archaeological horizons were identied. The top
layer has mostly ceramics, some decorated, and seems to represent
a recent occupation. Layer 2, about 1 m thick, and based on the
vertical concentration of materials, integrated 3 different archae-
ological horizons (2a, 2b and 2c). The top horizon is dated to the
Iron Age, and includes a series of ceramics, some decorated (Fig. 5)
and the presence of small quartz crescents (Table 3; Fig. 6); the
middle and lowest horizons have LSA occupations characterized by

Fig. 4. North Cut of Chicaza rock shelter.

Please cite this article in press as: Bicho, N., et al., Middle and Late Stone Age of the Niassa region, northern Mozambique. Preliminary results,
Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.059
8 N. Bicho et al. / Quaternary International xxx (2015) 1e13

Fig. 5. Artifacts from Chicaza, from A to D in descending stratigraphical order. All lithic artifacts are quartz akes except for artifact D1 which is a fragment of a bifacial point in
quartz.

the presence of small quartz akes and one quartz backed crescent,
larger than those found in the Iron Age. Layer 3 is also a LSA hori-
zon, similarly characterized by small quartz akes, but it includes a
small fragment of a thin bifacial fragment (Fig. 5D), most likely the
lateral section of a small point. Unfortunately, the number of
artifacts is very small, with just a few akes from each layer making
it virtually impossible to characterize technologically these lithic
assemblages.
We were able to obtain 8 AMS dates on single bone and charcoal
fragments, covering the entire sequence (Table 4). Bone collagen

Table 3
Lithic assemblage from Chicaza.

Class Layer 1 Layer 2a Layer 2b Layer 2c Layer 3 Total

Blade 2 2
Blade Fragments 1 1
Bladelet 6 8 5 19
Bladelet Fragments 2 5 2 2 11
Chip 38 29 44 11 7 129
Chunk 29 28 18 4 8 87
Core 2 1 1 1 5
Flakes 45 42 30 11 3 131
Flake Fragment 29 48 20 7 6 110
Maintenance Products 1 1
Retouched tools 3 2 1 6
Total 151 167 123 36 25 502

Please cite this article in press as: Bicho, N., et al., Middle and Late Stone Age of the Niassa region, northern Mozambique. Preliminary results,
Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.059
 N. Bicho et al. / Quaternary International xxx (2015) 1e13 9

obtained in the case of bones and are reversed in the charcoal. If our
interpretation is correct, then the Iron Age occupation took place in
Chicaza just a few hundred years ago, while two of the previous LSA
occupations occurred in early Holocene times, respectively around
7.5 and 9.2 ka cal BP. The earliest LSA occupation seems to be dated
to the Last Glacial Maximum, some 24 ka ago. This seems to be
conrmed by the observation, during excavation, of a clear strati-
graphical break between Layer 3, where this archaeological horizon
is, and the overlaying Layer 2 dated to the Holocene.
Table 4
AMS dates from Chicaza.

Layer/spit Material Lab code Date BP Cal BPa

2.3 Bone Wk-40712 489 35 621e495

2.6 Bone Wk-40714 6793 30 7676e7589
2.8 Bone Wk-40716 8293 40 9429e9137
Fig. 6. Quartz crescents from Chicaza. A is LSA from the archaeological horizon 2b and 3.10 Bone Wk-40718 19,807 100 24121e23551
B and C are Iron Age (archaeological horizon 2a). 2.3 Charcoal Wk-40713 4048 20 4778e4437
2.6 Charcoal Wk-40715 183 20 <287
2.8 Charcoal Wk-40717 1230 20 1258e1073
samples were ultraltrated. All dates have been calibrated with
3.10 Charcoal Wk-40719 9163 25 10401e10242
OxCal 4.2 (Bronk Ramsey, 2009). Dated materials were selected
a
from those articial spits that seemed to correspond to main Calibration made with Oxcal 4.2 using the IntCal13 calibration curve (Reimer
et al., 2013).
archaeological horizons marked by the high frequency of lithic
materials and bones, separated by levels with very few or no 5.3. Lithic assemblages: an overview
archaeological remains. The results show rstly that: 1) dates from
bone and charcoal are very different and not compatible between Regionally, the predominant raw material is quartz. It appears
the two types of samples; 2) most likely the charcoal results are both in primary and secondary sources spread over the entire
erroneous whereas more trust can be achieved by the bone results. landscape. Quartz primary sources are common in all surveyed
This interpretation results from two considerations: attending to areas either as veins of various thicknesses in the metamorphic and
the fact that the sediment is very loose sand and that there are volcanic outcrops in the Xiugulu and Unango areas and as major
modern roots from top to bottom, being likely that vertical sedimentary beds immediately to the north of the Lunho Valley
misplacement occurred with small light weight charcoal fragments that can reach thicknesses of more than a meter. The latter are
than with heavier complete long bones used for dating; this is exposed at least in the small rivers, parallel to the Lunho that run to
partially conrmed by the fact that correct stratigraphic values are Lake Niassa.

Fig. 7. Lithic artifacts from the Lunho valley. a) Chert levallois ake; b) and c) Quartz levallois cores; d) Quartz retouched ake; e) quartz discoidal core; f) quartz sidescraper;
g) Quartz endscraper; h) Chert discoidal core.

Please cite this article in press as: Bicho, N., et al., Middle and Late Stone Age of the Niassa region, northern Mozambique. Preliminary results,
Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.059
10 N. Bicho et al. / Quaternary International xxx (2015) 1e13

The quartz secondary sources are ubiquitous in the region, with Table 5 (continued )
either small (5e15 cm in length) blocks or pebbles/cobbles. The Site ID Blade Bladelet Core Flake Retouched tools Total
cobbles and pebbles are located in the river valleys as well as on the
A39 1 1
lake margin. A40 2 2
No other knappable raw materials were seen in considerable A42 1 4 2 2 9
quantities during survey, but there are artifacts in chert, quartzite and A43 5 5
silcrete. The location of these raw material sources will need more A44 3 4 7
A47 1 1
detailed geologic survey and likely an expansion of the surveyed area.
A48 12 6 18
We collected 245 diagnostic lithic artifacts during survey from A50 1 1
46 Stone Age sites (Table 5). In addition to these, 10 more artifacts A51 2 2
were collected from site O1, the contemporaneous location A52 1 5 1 1 8
A52.2 2 2
mentioned above. Most artifacts are in quartz, although there are a
A56 2 2
few quartzite (11), chert (8) and silcrete (3) from 15 locations A59 1 2 1 4
(Table 6). There are only three sites with 3 types of raw materials A61 1 8 1 3 13
and none with more, suggesting the use of raw materials was A63 1 1
highly standardized independently of the chronology, based on the A64 1 1
A65 4 7 11
local available quartz and rarely other types of lithic materials, the
A66 11 11
latter possibly coming from outside the immediate landscape. A67 5 5
As stated above, there are no artifacts with ESA diagnostic fea- A68 1 1
tures, such as cleavers or handaxes. Thus, all sites were attributed to Ab2 1 1
Ab2.2 1 3 4
either MSA or LSA as long as there were not any ceramics in asso-
Ab2.3 1 1
ciation. We used a series of technological variables to assign the O1 3 1 4 2 10
lithic assemblages to the MSA. These were the presence of Levallois Total 8 2 108 12 124 254
and discoidal cores, Levallois blanks and multifaceted platforms
(Table 7). Artifacts characterized by the presence of those variables Table 6
are present in 24 sites. Many of these have at least the presence of Non-quartz materials from the Niassa Stone Age sites.
two variables. The presence of Levallois technology is sufcient for Sites Quartzite Chert Silcrete (n)
an automatic attribution to MSA. Since there is no indication of ESA
A11 1 2
in the surveyed area, we are fairly secure that the presence of dis- A17 2 1 7
coidal cores is a good indicative for MSA occupations, more so if it is A21 1 1 19
present with multifacetting. We are, however, far less certain that A31 1 6
the presence of only multifacetted platforms is an accurate diag- A31.2 1 9
A31.3 1 13
nostic of the presence of MSA technology. Based on Table 7 as well as
A31.4 1 4
on the presence of retouched tools (Table 8) we think that there are A31.5 1 2
at least 21 sites that we are fairly certain are dated to the MSA. The A31.6 1 10
collected artifacts samples seem to be characterized by the presence A34 3 11
of small discoidal and, more rarely, Levallois cores. Flakes have A44 1 7
A48 3 18
frequently prepared platforms (Fig. 7), and tend to be small and A63 1 11
wide. Retouched stone tools are mostly notches, with rare dentic- A65 1 1 1
ulates, sidescrapers, endscrapers and retouched akes (Table 8). The A66 1 11
presence of bipolar technology, mostly with the presence of scaled Total 11 8 3
pieces might indicate a late MSA at A21, A26 and A48 sites.
Table 7
Table 5 Presence of diagnostic features for MSA in the Niassa Stone Age sites.
Diagnostic lithic artifact counts for the Niassa sites.
Site ID Levallois cores Discoidal cores Levallois blanks Multifacetting
Site ID Blade Bladelet Core Flake Retouched tools Total A3
A3 1 1 2 A4
A4 1 4 5 A17
A7 4 4 A19
A11 1 1 2 A21
A15 1 1 A26
A16 3 7 10 A31.2
A17 4 1 2 7 A31.3
A19 2 2 A31.5
A21 6 1 12 19 A31.6
A22 1 1 1 3 A31.7
A26 7 3 10 A34
A28 1 1 A36
A31 4 2 6 A37
A31.2 1 8 9 A39
A31.3 3 10 13 A40
A31.4 4 4 A42
A31.5 1 1 2 A47
A31.6 3 1 6 10 A48
A31.7 2 6 8 A52
A32 1 1 A59
A34 5 1 5 11 A61
A36 1 1 A68
A37 1 1 2 Ab2.2

Please cite this article in press as: Bicho, N., et al., Middle and Late Stone Age of the Niassa region, northern Mozambique. Preliminary results,
Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.059
 N. Bicho et al. / Quaternary International xxx (2015) 1e13 11

Table 8
Retouched stone tool artifacts for the Niassa sites.

Sites Awl Denticulate Endscraper Geometric Levallois blank Notch Point Retouched piece Sidescraper Burins Scaled pieces Total

A16 4 2 1 7
A17 1 1 2
A21 1 2 2 3 3 1 12
A22 1 1
A26 1 1 1 3
A28 1 1
A3 1 1
A31 1 1 2
A31.2 1 4 3 8
A31.3 2 1 2 2 2 1 10
A31.4 3 1 4
A31.5 1 1
A31.6 4 1 1 6
A31.7 3 1 1 1 6
A32 1 1
A34 1 1 3 5
A37 1 1
A40 1 1 2
A42 1 1 2
A44 1 2 1 4
A48 3 3 6
A51 1 1 2
A52 1 1
A52.2 2 2
A59 1 1
A61 2 1 3
A63 1 1
A64 1 1
A65 1 3 2 1 7
A66 5 2 3 1 11
A68 1 1
A7 1 1 2 4
Ab2.2 2 1 3
O1 1 1 2
Total 1 17 8 1 11 38 3 24 9 4 8 124

LSA materials are small (Fig. 7), usually no longer than 4 cm.
Cores tend to be relatively simple, frequently prismatic. Bipolar
technology is present, both in the shape of bipolar cores and of
scaled pieces. Radial ake removal from simple unifacial cores is
present in a couple sites (A31.7 and A51) together with the presence
of small backed points similar to those found in Ncuala. Since in one
case there are also Levallois materials, these may indicate either the
mixing of two different chronologies or a transitional phase be-
tween MSA and LSA. Only the analysis of larger collections from
excavations can solve this question.
In general akes are short and bladelets are rare. Retouched
tools are scarce, and include notches, retouched akes, burins,
endscrapers, denticulates, scaled pieces and rare backed points, as
mentioned above. It is also noteworthy that, interestingly, no
typical Levallois or other MSA points were found during survey.
6. Discussion and conclusions
At the beginning of the project and of the 2014 eld season,
it seemed that the Niassa region had a high potential for Stone
Age and particularly MSA studies. This was, in part, based on
the previous work by Mercader's team as well as due to the
geomorphological and geologic characteristics of the area, with
both uvial deposits dated to the Pleistocene and bed rock
appropriate for the formation of caves and rock shelters,
respectively limestone and sandstone. Field work rapidly
conrmed our thoughts with the location of more than 90 lo-
calities in a short period of time (about 20 days of effective eld
work), as well as many rock shelters, albeit some empty or with
little sediment.
While the potential is high for the discovery of both open air and
cave/rock shelter in situ multicomponent Stone Age sites, it is also
clear that there are two latent issues associated with the region. The
main issue is that of a generalized lack of previous systematic work
at a regional scale. Even taking into account Mercader's work,
which resulted in half a dozen known Stone Ages sites, the basic
knowledge on technology and the characteristics of Stone Age lithic
assemblages and faunal remains is, to say the least, scarce. The
exception is the work by Bennett (2011, 2013), who published or
presented at various times an initial technological framework for
the regional MSA, but based mostly on a single site, Mvumu. In
brief, Bennett's conclusions are the high frequency of local quartz
(~95%), simple and prepared core reduction with very scarce bi-
polar technology, absence of blade production, very low frequency
of retouched tools with the complete absence of bifacial technology
that include scrapers and awls, following a low level of formal
standardization. Bennet describes the presence of points, but based
on published photos and drawings, they do not seem, in our
opinion, to have been formally shaped into projectiles. Based on the
very few dates and the lithic analyses from 3 sites (Mvumu, Mikuyu
and Ngalue) Bennet argues for a long lasting stable technology
between 105 and 29 ka ago, corresponding to the duration of the
local MSA. There is no information for the LSA in the region,
although it seems to be present in various sites, including the very
important Ngalue cave, located close to Chicaza. Ultimately, one can
say that, except for the information on the use of plant bers and
starch during the MSA in the region (Mercader et al., 2008;

Please cite this article in press as: Bicho, N., et al., Middle and Late Stone Age of the Niassa region, northern Mozambique. Preliminary results,
Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.059
12 N. Bicho et al. / Quaternary International xxx (2015) 1e13
Mercader, 2009), there was little comparable data on chronology,
subsistence or technology for the Stone Age in the region. Our own
results from the preliminary lithic analyses, conrm those charac-
teristics found by Mercader's team. However, it seems that, while
based on small artifact samples, we were able to nd some addi-
tional characteristics for both MSA and LSA that had not been
referenced before. These include the presence of some Levallois
technology, associated to multifacetting of ake platforms in the
MSA; the presence of Levallois and bipolar technologies in some
sites suggesting that this technological set may correspond to later
MSA of, still, unknown age. Sites such as Ncuala (and A31.7 and
A51), with abundant prismatic and bipolar reduction sequences,
producing short and wide blanks of all types, and the presence of
relatively wide backed bladelet points likely corresponding to the
rst phases of the regional LSA. By the Last Glacial Maximum,
bifacial technology is present in Chicaza following the quartz
technological characteristics of the LSA. While, we have very few
diagnostic materials from the dated occupations from Chicaza, the
lithic artifacts seem to maintain the same general patterns with
both prismatic and bipolar blank production all the way to Holo-
cene times.
Unlike what one would expect, the regional lithic technology
and typology from the Niassa region seems to be fairly different
from known sequences in southern and eastern Africa found in
the neighboring such as Tanzania, Zambia, Malawi or South Af-
rica (e.g., Bisson, 1990; Soriano et al., 2007; Diez-Matin et al.,
2009; Thompson et al., 2012, 2014; Villa et al., 2012; Willoughby,
2012; Wright et al., 2014; Davis et al., 2015; Nightingale et al.,
2015; Schilt et al., 2015). The typical elements of late MSA with
Still Bay, Howiesons Poort, Post-HP and the use of blade tech-
nology (e.g., Soriano et al., 2007; Porraz et al., 2013a) seen in
southern Africa from Sibudu (Wadley, 2007; Conard et al., 2013;
Archer et al., 2014, 2015) to Diepkloof (Porraz et al., 2013a,
2013b) are, presently, not attested in the Niassa. The general
characteristics seen in Tanzania in the Iringa region (Willoughby,
2012) or northern Malawi (Thompson et al., 2012, 2014; Wright
et al., 2014) are only partially present, with the high frequency
of quartz and bipolar technology and the presence of atypical
cores as well as small Levallois and discoidal forms as noted by
Thompson et al. (2012, 2014) in the Caronga District in northern
Malawi. However, based on our current data, Levallois ake and
blade points, geometrics, blade production and large scale bi-
polar production seem to have been essentially absent from the
Niassa region during the MSA.
The same pattern seems to be present in the LSA assemblages
found during the Niassa survey. The LSA materials from the Iringa
region in Tanzania (Willoughby, 2012) and from the Luano Spring
site in Zambia (Bisson, 1990) show only some of the variables found
in the Niassa, where the bladelet production are not observed,
while backing, although present, is clearly rare, even in the form of
crescents. On the other hand, the bifacial projectile technology
during the Last Glacial Maximum found in Chicaza, seems to be, at
this time, unique to the Niassa region.
These differences might be the result of both the lack of
chronological control as well as of small lithic assemblages so far
obtained in Niassa. Also, the extended use of quartz as the main
raw material, instead of chert, quartzite or even silcrete might
have an important impact, at least when compared to more
southern latitudes. Still, we cannot exclude that the apparent
differences between the materials from Niassa and other close by
regions are possibly stylistic and form a regional variant with
specic characteristics that include the absence of blade technol-
ogy and of production of large points during the MSA and the
limited use of backing and microlithic bladelet tools during the
LSA.
Please cite this article in press as: Bicho, N., et al., Middle and Late Stone Age of the Niassa region, northern Mozambique. Preliminary results,
Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.059
The preliminary information on the Lunho valley, resulting on
apparent differential patterns of occupation between the northern
and southern banks of the river as well as between MSA an LSA
occupations, provide an interesting future topic of research, to
which, is added the high numbers of sites and rock shelters to be
tested. The same holds true for the inland highlands with caves and
rock shelters in the Unango and Malulu areas. All three areas can,
theoretically, be the framework for the chronological and techno-
logical regional knowledge on Stone Age archaeology in the Niassa
region.
Acknowledgements
We would like to thank Daniela Matos for her help in analyses of
artifacts at the IICT. We would like to thank to a Funda~ao para a
^ncia e Tecnologia (PTDC/EPH-ARQ/4998/2012) and the Wenner-
Cie
Gren Foundation for Anthropological Research for funding the
research in Mozambique. Fieldwork permits were given to Nuno
Bicho for a four year period starting in 2014 by the Governmental
Authority for Cultural Heritage of Mozambique, Direc~ ao Nacional
do Patrimo
nio Cultural.
References
Adamowicz, L., Nhatule, E., 2011. Environment and social impact assessment for the
proposed exploration in EPC
area A. Archaeological Studies. https://www.
academia.edu/6542957/ARCHAELOGICAL_RESCUE_SURVEY_IN_SASOL_
PROJECT_AREA._MABOTE_DISTRICT (accessed 20.06.14).
Archer, W., Bader, G., Conard, N., 2014. Characterizing the late Pleistocene MSA lithic
technology of Sibudu, KwaZulu-Natal, South Africa. PLoS One. http://dx.doi.org/
10.1371/journal.pone.0098359.
Archer, W., Gunz, Ph, van Niekerk, K., Henshilwood, C., McPherron, S., 2015.
Diachronic change within the still Bay at Blombos cave, South Africa. PLoS One.
http://dx.doi.org/10.1371/journal.pone.0132428.
Barradas, L., 1945. As formao ~es quatern
~es com a
arias do sul do Save e as suas relao
pre
-histo
ria. Boletim da Sociedade de Estudos da Colona de Moambique XIV
(47), 1e33.
Barradas, L., 1947. Origem dos solos quatern
arios do sul de Moambique. Boletim da
Sociedade de Estudos da Colona de Moambique XVI (55), 197e216.
Barradas, L., 1955. Complexos geolo
gicos-arqueolo
gicos do Quaterna
rio no sul de
Moambique. Boletim da Sociedade de Estudos da Colona de Moambique 90,
3e20.
Barradas, L., 1964. Concheiros da antiga baia de Loureno Marques. Estudos sobre
Pre
-histo
ria do Ultramar Portugue ^s 2 (50), 113e163.

rio da beira-mar (Sul de Moambique). In:
Barradas, L., 1965. Rochas do quaterna
Memo
rias do Instituto de Investiga~ ao Cientca de Moambique, vol. 7,
pp. 37e84 (se
rie B).
Bennet, T., 2013. Long-term continuity in Middle Stone Age lithic technology from
Niassa, Mozambique. In: Paper Presented at the Annual Meeting of the Paleo-
anthropological Society.
Bennett, T., 2011. Middle Stone Age Lithic Technology at Mvumu, Niassa,
Mozambique (M.A. Thesis). University of Calgary, Calgary.
Bisson, M., 1990. Lithic reduction sequences as an aid to the analyses f Late Stone
Age quartz assemablges from the Luano Spring, Chingola, Zambia. African
Archaeological Review 8, 103e138.
Bronk Ramsey, C., 2009. Bayesian analysis of radiocarbon dates. Radiocarbon 51,
337e360.
Cascalheira, J., Gonalves, C., Bicho, N., 2014. Smart phones and the use of
customized apps in archaeological projects. The SAA Archaeological Record 14
(5), 20e25.
Clark, J.D., Beyene, Y., WoldeGabriel, G., Hart, W.K., Renne, P.R., Gilbert, H.,
Deeur, A., Suwa, G., Katoh, S., Ludwig, K., Boisserie, J.-R., Asfaw, B., White, T.D.,
2003. Stratigraphic, chronological and behavioural contexts of Pleistocene
Homo sapiens from Middle Awash, Ethiopia. Nature 423, 747e752. http://
dx.doi.org/10.1038/nature01670.
Conard, N., Shmid, V., Zeidi-Kulehparcheh, M., Porraz, G., 2013. Bifacial technology
at Sibudu and its implications for our understanding of the Still Bay. In: 3rd
Annual Meeting of the European Society for the Study of Human Evolution,
Vienna, pp. 19e21.
d'Errico, F., Vanhaeren, M., Wadley, L., 2008. Possible shell beads from the Middle
Stone Age layers of Sibudu cave, South Africa. Journal of Archaeological Science
35, 2675e2685.
Davis, J., Nightingale, S., Thompson, J., Gomani-Chindebvu, E., 2015. Quantifying the
effects of erosional processes on stone artifact concentrations: implications for
site formation at open-air paleolithic sites. In: Paper Presented at the 80th
Annual Meeting of the Society for American Archaeology, San Francisco, USA.
 N. Bicho et al. / Quaternary International xxx (2015) 1e13 13

Diez-Martn, F., Domnguez-Rodrigo, M., Sa
nchez, P., Mabulla, A., Tarrin ~ o, A., Reimer, P., Bard, E., Bayliss, A., Beck, J., Blackwell, P., Bronk Ramsey, C., Buck, C.,
Barba, R., Prendergast, M., de Luque, L., 2009. The Middle to Later Stone Age Cheng, H., Edwards, R., Friedrich, M., Grootes, Pieter M., Guilderson, T.,
technological transition in East Africa. New data from Mumba Rockshelter Bed Haidason, H., Hajdas, I., Hatte
, C., Heaton, T., Hoffmann, D., Hogg, A.,
V (Tanzania) and their implications for the origin of modern human behavior. Hughen, K., Kaiser, K., Kromer, B., Manning, S., Niu, M., Reimer, R., Richards, D.,
Journal of African Archaeology 7, 1e27. Scott, E., Southon, J., Staff, R., Turney, C., van der Plicht, J., 2013. IntCal13 and
Gama, M., 1990. O Povo Yao. Instituto de Investiga~ ao Cientca e Tropical, Lisboa. Marine13 radiocarbon age calibration curves 0e50,000 years cal BP. Radio-
Gonalves, C., Matos, D., Haws, J., Benedetti, M., Riel-Salvatore, J., Raja, M., carbon 55 (4), 1869e1887. Beck.
Madime, O., Bicho, N., 2014. Mapping the early and Middle Stone Age in Rodrigues, M.C., 1998/1999. Os Primo
rdios da Investigaa ~o Arqueolo
gica em
Mozambique: preliminary results. In: Poster Presented at the 14th Congress of Moambique e o Prof. Santos Jnior. Portvgalia XIX-XX, 265e278.
the Pan African Archaeological Association for Prehistory, Johannesburg, South Rodrigues, M.C., 2004. A Arqueologia em Moambique nas Misso ~es Cientcas da
Africa. antiga Junta de Investiga~ ao do Ultramar de 1936e1972 (Ph.D. thesis). Uni-
Gonalves, C., Raja, M., Madime, O., Cascalheira, J., Haws, J., Matos, D., Bicho, N., versidade de Coimbra, Coimbra.
2015. Mapping the Stone Age of Mozambique: the archaeology of the 20th Roque, A.C., 2002. Espo
lio da Missa ~o Antropolo
gica de Moambique. Parte I e
century. African Archaeological Review (in press). Apresenta~ ao do espo
lio e inventa
rio dos materiais arqueolo
gicos do espo
lio.
Grantham, G., Marques, J., Wilson, M., Manhia, V., Hartzer, F., 2011. Explanation of Leba e Estudos de Pre
-Histo
ria e Arqueologia 8, 1e244.
the Geological Map of Mozambique, 1:1000000. Direca ~o Nacional de Geologia, Roque, A.C., Ferra ~o, L., 2004. As Teias da Histo
ria. Import^ancia e contributo dos
Maputo. inventa
rios de materiais para o conhecimento e (re)construa ~o da Histo
ria de
Henshilwood, C., d'Errico, F., van Niekerk, K., Coquinot, Y., Jacobs, Z., Lauritzen, S.-E., Moambique. http://www.ces.uc.pt/lab2004/pdfs/Ana_Roque.pdf. Accessed 20
Menu, M., Garca-Moreno, R., 2011. A 100,000-year-old ochre-processing June 2014.
workshop at Blombos cave, South Africa. Science 334 (6053), 219e222. Santos Jnior, J.R., 1937. Contribuia ~o para o estudo da Idade da Pedra em
http://dx.doi.org/10.1126/science.1211535. Moambique. A esta~ ao ltica de Marissa (Tete). Moambique 12, 95e105.
Henshilwood, C., Marean, C., 2003. The origin of modern human behavior: critique Santos Jnior, J.R., 1938. Relato
rio da Missa ~o Antropolo


gica a Africa do Sul e
of the models and their test implications. Current Anthropology 44 (5), Moambique. Trabalhos da Sociedade Portuguesa de Antropologia e Etnologia
627e651. VIII (IIIeIV), 257e308.
Henshilwood, C.S., D'Errico, F., Vanhaeren, M., van Niekerk, K., Jacobs, Z., 2004. Santos Jnior, J.R., 1940. Pre
-Histo
ria de Moambique. Congresso do Mundo
Middle Stone Age shell beads from South Africa. Science 304 (5669), 404. Portugue ^s 14 (I), 307e356.
chelt, S., 2004. Geology and Mineral Resources of Mozambique. Direca
La ~o Nacional Santos Jnior, J.R., 1941. On the prehistory of Mozambique. Moambique 28, 23e75.
de Geologia, Maputo. Santos Jnior, J.R., 1944. Missa ~o Antropolo
gica de Moambique: alguns resultados
Marean, C., 2014. The origins and signicance of coastal resource use in Africa and de duas campanhas. Separata de Las Ciencias 3, Madrid.
Western Eurasia. Journal of Human Evolution 77, 17e40. Santos Jnior, J.R., 1946. Relato
rio da 4 campanha da Miss~ ao Antropolo
gica de
Marean, C., Assefa, Z., 2005. The Middle and Upper Pleistocene African record for Moambique. In: Anais da Junta das Misso ~es Geogra
cas e de Investigao ~es
the biological and behavioral origins of modern humans. In: Stahl, A. (Ed.), Coloniais 1. Junta das Misso ~es Geogr
acas e de Investigao ~es Coloniais, Lisboa,
African Archaeology: a Critical Introduction. Blackwell Publishing Ltd, pp. 451e463.
pp. 93e129. Santos Jnior, J.R., 1947. Alguns aspectos da 4 campanha da Missa ~o Antropolo
gica
Marean, C.W., Bar-Matthews, M., Bernatchez, J., Fisher, E., Goldberg, P., Herries, A.I., de Moambique. Bulletin de la Socie
te
Portugaise des Sciences Naturelles XV
Jacobs, Z., Jerardino, A., Karkanas, P., Minichillo, T., 2007. Early human use of (23), 128e151.
marine resources and pigment in South Africa during the Middle Pleistocene. Santos Jnior, J.R., 1950. Carta da Pre
-Histo
ria de Moambique. In: Proceedings, XIII
Nature 449, 905e908. Congresso da Associa~ ao Portuguesa para o Progresso das Cie ^ncias (Sep. do
McBrearty, S., Brooks, A.S., 2000. The revolution that wasn't: a new interpretation of Tomo V, 4 seca ~; o). Lisboa, pp. 647e656.
the origin of modern human behavior. Journal of Human Evolution 39 (5), Schilt, F., Mentzer, S., Wright, D., Thompson, J., Gomani-Chindebvu, E., 2015.
453e563. http://dx.doi.org/10.1006/jhev.2000.0435. Micromorphology of Middle to Later Stone Age sites at Mwanganda's village,
McDougall, I., Brown, F., Fleagle, J., 2005. Stratigraphic placement and age of northern Malawi. In: Presented at the 80th Annual Meeting of the Society for
modern humans from Kibish, Ethiopia. Nature 433, 733e736. http://dx.doi.org/ American Archaeology, San Francisco, California.
10.1038/nature 03258. Soriano, S., Villa, P., Wadley, L., 2007. Blade technology and tool forms in the Middle
Mercader, J., 2009. Mozambican grass seed consumption during the Middle Stone Stone Age of South Africa: the Howiesons Poort and post-Howiesons Poort at
Age. Science 326 (5960), 1680e1683. http://dx.doi.org/10.1126/science.1173966. Rose Cottage cave. Journal of Archaeological Science 34 (5), 681e703.
Mercader, J., Asmerom, Y., Bennett, T., Raja, M., Skinner, A., 2009. Initial excavation Thompson, J., Mackay, A., Moor, V., Gomani-Chindebvu, E., 2014. Catchment survey
and dating of Ngalue cave: a Middle Stone Age site along the Niassa rift, in the Karonga district: a landscape-scale analysis of provisioning and core
Mozambique. Journal of Human Evolution 57 (1), 63e74. http://dx.doi.org/ reduction strategies during the Middle Stone Age of northern Malawi. African
10.1016/j.jhevol.2009.03.005. Archaeological Review 31, 447e478.
Mercader, J., Bennett, T., Esselmont, C., Simpson, S., Walde, D., 2013. Phytoliths from Thompson, J., Mackay, A., Wright, D., Welling, M., Greaves, A., Gomani-
Middle Stone Age habitats in the Mozambican rift (105e29 ka). Journal of Chindebvu, E., Simengwa, D., 2012. Renewed investigations into the Middle
Human Evolution 64 (5), 328e336. http://dx.doi.org/10.1016/ Stone Age of northern Malawi. Quaternary International 270, 129e139.
j.jhevol.2012.10.013. Villa, P., Soriano, S., Tsanova, T., Degano, I., Higham, T.F.G., d'Errico, F., Backwell, L.,
Mercader, J., Bennett, T., Raja, M., 2008. Middle Stone Age starch acquisition in the Lucejko, J.J., Colombini, M., Beaumont, P., 2012. Border cave and the beginning
Niassa rift, Mozambique. Quaternary Research 70 (2), 283e300. http:// of the later Stone Age in South Africa. Proceedings of the National Academy of
dx.doi.org/10.1016/j.yqres.2008.04.010. Sciences 109, 13208e13213. http://dx.doi.org/10.1073/pnas.1202629109.
Mercader, J., Gosse, J., Bennett, T., Hidy, A., Rood, D., 2012. Cosmogenic nuclide age Wadley, L., 2001. What is cultural modernity? A general view and a South African
constraints on Middle Stone Age lithics from Niassa, Mozambique. Quaternary perspective from Rose Cottage cave. Cambridge Archaeological Journal 11,
Science Reviews 47, 116e130. http://dx.doi.org/10.1016/j.quascirev.2012.05.018. 201e221.
Morais, J., 1984. Mozambican archaeology: past and present. African Archaeological Wadley, L., 2007. Announcing a still Bay industry at Sibudu cave, South Africa.
Review 2 (1), 113e128. http://dx.doi.org/10.1007/BF01117228. Journal of Human Evolution 52, 681e689.
Nightingale, S., Foster, M., Thompson, J., Jeong-Heon, C., Wright, D., 2015. Spatial Wadley, L., 2010. Were snares and traps used in the Middle Stone Age and does it
and chronological components of Middle Stone Age artifact assemblage vari- matter? A review and a case study from Sibudu, South Africa. Journal of Human
ability in deeply buried alluvial fan contexts. In: Paper Presented at the 80th Evolution 58 (2), 179e192. http://dx.doi.org/10.1016/j.jhevol.2009.10.004.
Annual Meeting of the Society for American Archaeology, San Francisco, USA. White, T., Asfaw, B., DeGust, D., Gilbert, H., Richards, G.D., Suwa, G., Howell, F.C.,
Porraz, G., Parkington, J., Rigaud, J.P., Miller, C., Poggenpoel, C., Tribolo, C., Archer, W., 2003. Pleistocene Homo sapiens from middle awash, Ethiopia. Nature 423,
Cartwwright, C., Charrie
-Duhat, A., Dayet, L., Igreja, M., Mercier, N., Schmidt, P., 742e747. http://dx.doi.org/10.1038/nature01669.
Verna, C., Texier, P.J., 2013a. The MSA sequence of Diepkloof and the history of Willoughby, P., 2012. The Middle and Later Stone Age in the Iringa region of
southern African Late Pleistocene populations. Journal of Archaeological Sci- southern Tanzania. Quaternary International 270, 103e118.
ence 40 (9), 3542e3552. Wright, D., Thompson, J., Mackay, A., Welling, M., Forman, S., Price, G., Zhao, J.,
Porraz, G., Texier, P.J., Archer, W., Piboule, M., Rigaud, J.-P., Tribolo, C., 2013b. Cohen, A., Malijani, O., Gomani-Chindebvu, E., 2014. Renewed Geo-
Technological successions in the Middle Stone Age sequence of Diepkloof rock archaeological investigations of Mwanganda's village (Elephant Butchery site),
shelter, Western Cape, South Africa. Journal of Archaeological Science 40 (9), Karong, Malawi. Geoarchaaeology 29, 98e120.
3376e3400.

Please cite this article in press as: Bicho, N., et al., Middle and Late Stone Age of the Niassa region, northern Mozambique. Preliminary results,
Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.059