Content uploaded by Daniela Holst
Author content
All content in this area was uploaded by Daniela Holst on Aug 30, 2019
Content may be subject to copyright.
This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
Author's personal copy
Hazelnut economy of early Holocene hunteregatherers: a case study from
Mesolithic Duvensee, northern Germany
Daniela Holst
*
Palaeolithic Research Unit, Römisch-Germanisches Zentralmuseum, Schloss Monrepos, D-56567 Neuwied, Germany
article info
Article history:
Received 18 May 2010
Received in revised form
23 June 2010
Accepted 24 June 2010
Keywords:
Mesolithic
Northern Germany
Hazelnut
Subsistence
Land use
High return-harvesting
abstract
Throughout the greater part of human evolution in Europe, use of plant foods is invisible and thus might
have played a secondary role in nutrition. Ecological changes at the beginning of the early Holocene
provoked innovations in early Mesolithic subsistence, focusing on the rich plant resources of the
increasingly forested environment. High-resolution analyses of the excellently preserved and well-dated
special task camps documented in detail at Duvensee, Northern Germany, offer an outstanding oppor-
tunity for case studies on Mesolithic subsistence and land use strategies. Quantification of the nut uti-
lisation demonstrates the great importance of hazelnuts. These studies revealed very high return rates
and allow for absolute assessments of the development of early Holocene economy. Stockpiling of the
energy rich resource and an increased logistical capacity are innovations characterising an intensified
early Mesolithic land use, which is reflected in the stable tradition of uniform seasonal settlement
patterns at early Mesolithic Duvensee. The case study reveals characteristics in early Mesolithic
subsistence and land use that anticipate attributes of the Neolithic economy.
Ó2010 Elsevier Ltd. All rights reserved.
1. Introduction
The evolution of subsistence ein particular the emergence of
hunting and the invention of agriculture eforms one of the key
issues of hominin evolution. There is ample archaeozoological and
biogeochemical evidence for the importance of hunting and
increasing carnivory in hominin evolution, emphasising animal
products as a dominant staple food in past and present hunter-
egatherersocieties(Hublin and Richards,20 09; Cordainet al., 200 0).
The role of plant foods, however, remains obscure for large parts
of prehistory, despite their crucial physiological importance in
supplying carbohydrates and moderating high nitrogenloads of lean
meat (Jones, 2009; Speth, 1989). Earliest evidence for a considerable
and regular plant economy derives from the Near Eastern
Epipalaeolithic, extending back to 20000 years at Ohalo II (Israel)
(Kislev et al.,1992; Weiss et al., 2004). As part of the “Broad Spectrum
Revolution”, the Near Eastern Epipalaeolithic is characterised by the
intensive plant utilisation of a wide range of plants and equipment
for processing such as ground stone tools and hearths (Barlow and
Heck, 2002; Hillman, 2000; Wright, 1994). This implies far-
reaching impacts on land use and mobility as well as associated
social changes that precede the Neolithic in its area of origin.
In contrast, in Europe there is an apparently sudden shift from
a meat-focused to an “imported”Neolithic way of subsistence,
mainly based on cereals. Assuming at least partial population
continuity at the advent of the Neolithic (Haak et al., 2005; but see
Bramanti et al., 2009), the quick change must have been incompat-
ible with human physiology (Cordain et al., 2005; Lindeberg, 2009).
However,earlyHolocenereforestationofcentralEuropemusthave
already provoked innovations in human subsistence strategies in the
preceding Mesolithic. Forested environments are considered as diffi-
cult to exploit due to the dispersed nature of their plant and animal
resources, requiring complex subsistence strategies(Gamble, 1986).
Various lines of evidence point to increased plant utilisation in
the Mesolithic (e.g., Clarke, 1976; Zvelebil, 1994). The importance of
hazelnuts is indicated by the frequent presence of hazelnut shells at
Mesolithic sites, as well as by constructions associated with their
processing. Hazelnuts (untreated, shelled) have a very high ener-
getic value, containing more than 60% fat, 15% proteins and nearly
17% carbohydrate, in addition to high amounts of unsaturated fatty
acids, minerals and vitamins (US Department of Agriculture
(USDA), 2009).
However, the absence of detailed quantitative surveys in inter-
relation with spatial analyses at archaeological sites prevents
verifiable reconstructions of their economic relevance.
*Tel.: þ49 2631 977222; fax: þ49 2631 76357.
E-mail address: holst@rgzm.de
Contents lists available at ScienceDirect
Journal of Archaeological Science
journal homepage: http://www.elsevier.com/locate/jas
0305-4403/$ esee front matter Ó2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jas.2010.06.028
Journal of Archaeological Science 37 (2010) 2871e2880
Author's personal copy
Because of their good preservation the sites at Duvensee
(Northern Germany) form an ideal archive for the study of Mesolithic
nut utilisation. Analyses of these exhaustively documented, chrono-
logicallyand spatially clearly defined special task camps now provide
the opportunity for high-resolution studies on plant utilisation and
the developmentof land use throughout the early Mesolithic.
2. Duvensee
The Duvensee bog is situated in northern Germany, 35 km
northeast of Hamburg (Fig. 1). The former lake formed behind
a Weichselian till. Originally the lake covered an area of more than
4km
2
. Peat formation reducing the surface of the lake began in the
late Preboreal. Numerous sites were discovered during the last
century, of which 12 have been excavated until 2001 (Bokelmann,
1971,1981,1991; Bokelmann et al., 1981). The sites are named
"Wohnplätze" ("living sites", inthe following abbreviated by "Wp"),
after a description by the first excavator, Gustav Schwantes
(Schwantes et al., 1925). From the beginning, investigations at
Duvensee were representative for Mesolithic research in Germany
at a general scale. After discovery and initial excavations in 1924
(Schwantes et al., 1925) the Duvensee area subsequently became
the type location for the Early Mesolithic of Northern Germany,
which came to be known as the "Duvensee-group" or "Duvensee-
culture", designated amongst others by characteristic "Duvensee
bone points" (Schwantes, 1958).
The sites, dating from the late Preboreal (w8900 cal. BC) until
the beginning of the Atlantic (w6500 cal. BC), are situated in close
vicinity to each other on two peninsulas of the north-western
former lakeshore.
Hearth structures with layers of carbonised hazelnut shells
within clearly defined scatters of flint artefacts form the character-
istic feature of the Duvensee settlements. In spite of various organic
implements such as a wooden paddle, arrow shafts or bone points
(Holst, 2007; Schwantes et al., 1925) and favourable preservation
conditions in the lake sediments (gyttja), faunal remains are scarce.
Systematic surveys and excavations beyond the hearth areas could
not reveal further Mesolithic activities. This phenomenon, therefore,
may not be ascribed to diagenetic reasons but seems to point to
a restricted seasonal function of the specialised occupations, also
indicated by the occupation history and duration.
All excavated sites are
14
C dated (Fig. 2;Table 1) in accordance
with palynological correlations to vegetation phases of the early
Holocene landscape development. Occupation started in the late
Preboreal (Wp 8 and 9) and extended to the early Atlantic period
(Wp 19), coveringin total a time span of nearly 2400 years. The onset
and end of settlement at the Duvensee lakeshore coincided notice-
ably with the appearance of Corylus in the pollen spectra and its
significant drop respectively, possibly associated with the so called
“8.2-kyr-BP”cooling phase (e.g., Alley and Ágústsdóttir, 2005;
Tinner and Lotter, 2001). The availability of hazelnuts obviously
determined the recurrent Mesolithic occupations at the lakeshore.
Hazel decline possibly triggered the establishment of alternative
subsistence and land use strategies, providing a rare example of the
concrete consequences of climate changes on land use patterns.
3. Nut processing at Duvensee
3.1. Case studies Wp 8 and 6
The current paper focuses on Wp 8 (late Preboreal) and Wp 6
(Boreal) analyses (Holst, 2007). Refits as well as vertical and hori-
zontal find distributions reflect short periods of accumulation for
both sites, pointing to single ephemeral occupations.
Wp 8 was excavated over a total area of 90 m
2
(Bokelmann et al.,
1981)(Fig. 3). Apart from 14,326 lithic artefacts (5280>2 cm),
a hearth structure (4.7 m
2
) was documented in the centre of the
excavated area, directly at the former shore. Two similar features
(0.2 m
2
and 0.6 m
2
) were located in the western and northern
immediate vicinity.
Wp 6 is located some 150 m apart from Wp 8 on the eastern
peninsula (Fig. 1). An area of 117 m
2
was excavated (Bokelmann,
1981)(Fig. 4). The 11,225 flint artefacts (6740 >2 cm), a central
hearth structure (3.2 m
2
) and two additional, but less well-defined
features to the south (each 1.6 m
2
) were uncovered in the peat layer
(Holst, 2007).
3.2. Hazelnut exploitation at Duvensee
Wp 8 and 6 are characterised by distinct hearth structures with
thick nutshell and sand layers (Figs. 3 and 4). These hearths form
the most apparent evidence for nut exploitation. The features have
Fig. 1. Duvensee. Location of the sites at the early Holocene Duvensee Lake (“Wp”¼“Wohnplatz”; sources: RGZM and TK 25 sheets 2229, 2329).
D. Holst / Journal of Archaeological Science 37 (2010) 2871e28802872
Author's personal copy
been interpreted as roasting facilities for hazelnuts on the basis of
ethnographic examples and experiments (Bokelmann, 1981).
The roasting hearth at Wp 8 consists of a rectangular pine and
birch bark mat that originally measured at least 6 m
2
. A layer of
white sand (with a central lens of brown sand), some 2e14 cm thick
over an area of ca. 4.7 m
2
, mixed with charcoal, hazelnut shells and
lithic artefacts was stratified on top of the mat.
The hearth structures of Wp 6 also consist mainly of white sand
(central structure 3e20 cm thick over an area of ca. 3.2 m
2
, with
a central lens of brown sand) with charcoal and large amounts of
hazelnut shells. For the central hearth the sand had been deposited
in a shallow depression.
For Wp 8 and 6 calculations based on the extension and depth of
the sand layers revealed the costly investment in the roasting
process. The layer of sand on Wp 8 approximates 286 l or 414 kg; at
Wp 6 it adds up to 324 l or 466 kg. Taking into account the
admixture with lithic artefacts and nutshells the sand amount was
carefully reduced for 50%, still resulting in a considerable weight of
sand. The huge amounts of fine white sand were probably brought
in from the western lakeshore, where it was also quarried in the
18th century (Funck, 1963). Water transport may have facilitated
the logistical complexity of the roasting events and constitutes one
of the advantages of location at Duvensee.
The bark mat (Wp 8) or, respectively, the shallow depression,
additionally delimited by a construction of several pine planks
(Wp 6) that were produced on site (wood analyses by U. Tegtmeier),
prevented the valuable sand from dispersing (Figs. 3 and 4).
Nut processing equipment further underlines the importance of
hazelnut exploitation. The frequency of grinding and pounding
tools on Mesolithic sites is considered as an indicator of increased
plant utilisation. Analyses of sandstone and granite artefacts at Wp
8(n¼29) and Wp 6 (n¼6) revealed macroscopic traces of cutting,
percussion or abrasion. Typical massive nut cracking anvil stones
possess a characteristic central depression surrounded by abrasions
(Holst, 2007)(Fig. 5). Starch residues of Corylus and Quercus on
comparable artefacts associated with nut remains from the Spanish
Fig. 2. Duvensee. Calibrated
14
C-ages, calibrated with CalPal 2007 (Weninger et al., 2010). References for uncalibrated
14
C-data in Holst (2007).
D. Holst / Journal of Archaeological Science 37 (2010) 2871e2880 2873
Author's personal copy
site of Font del Ros (Martínez-Moreno and Mora Torcal, in press)
strengthen this interpretation.
More or less rectangular shaped sandstone slabs (80e85 cm
long, 60e75 cm wide, 1.5e2.5 cm thick) at the Duvensee sites
display traces of grinding or cut marks that may attest to further
processing such as the preparation of nut oil or flour.
The significance of hazelnut exploitation is furthermore reflec-
ted by intrasite settlement patterns (Holst, in press, 2007). Pro-
cessing of hazelnuts, the site’s prime function was accompanied by
further settlement activities at Duvensee, as shown by the presence
of thousands of flint artefacts. Refits and attribute analyses revealed
entire sequences of production whereby retooling of arrows was
most important (Holst, 2008).
Geostatistical analyses of density distributions (Kriging) and
their quantitative assessments (Fig. 6), point plots and refits of
stone artefacts formed the basis for spatial analyses of settlement
organisation (Holst, 2007, in press). They display concordantly that
all activities eeven those located more distant from the hearth
(vgl. Fig. 6)ewere focussed on the roasting hearths, which
exclusively determined settlement organisation.
Roasting experiments were undertaken (Bokelmann, 1981;
Score and Mithen, 2000) but so far resulted in high loss rates,
indicating that the roasting of nuts must have been conducted
without direct contact to open fire, but with glowing charcoals as
could be verified by the author’s experiments. Sand, the basis of
roasting features, served as a heat-conductor, similar to the
hearths used for roasting Mongongo nuts by the southern
African !Kung (Yellen, 1977). In principal, the roasting hearths at
Duvensee worked in the same way: a fire was lit on top of a sand
layer. After the fire had burned down, the glowing charcoal was
mixed into the sand. Hazelnuts were then buried and roasted in
the hot sand. Roasting takes just a few minutes in sand heated to
nearly 300
C(Lage, 2004); the short roasting times at moderate
temperatures maintain the nutritional value of nuts (Kirbas¸lar
and Erkmen, 2003). Raking the sand for picking out the roas-
ted nuts then led to the admixture with surrounding stone
artefacts.
There are good reasons for the energy-expensive roasting of
hazelnuts: heating destroys contaminants, induces the typical
nutty flavour, improves colour and structure and extends shelf life
(e.g., Saklar et al., 2003; Wandsnider, 1997). Green or unshelled
nuts are very susceptible to carcinogenic mould fungi. Considering
the high loss rates caused by squirrels and the asynchronous
ripening of nuts, harvesting of green fruits in the Mesolithic is
highly probable and also known from ethnographic contexts (e.g.,
Kuhnlein and Turner, 1991).
In addition, roasting makes nuts easier to crack and grind (Score
and Mithen, 2000, p. 511; Talalay et al., 1984, p. 351). Processing
facilitated stockpiling in concentrating foods by reducing their
volume and transport weight by 50%, an important logistical
advantage for mobile hunteregatherers.
4. Economic significance of hazelnuts at Duvensee
4.1. Botanical investigations
Hazelnut shells were detected on every Wp (except on Wp 9,
Bokelmann, 1991) in large amounts forming compact layers at and
around the roasting hearths.
Besides hazelnuts knotweed (Polygonum convolvulus)was
discovered at Wp 2 in considerable amounts (Schwantes, 1958).
Botanical investigation of a small sample (w50 ml) of hearth
Table 1
Duvensee, calibrated
14
C-ages of all excavated sites. Calibrated with CalPal 2007 (Weninger et al., 2010). References for uncalibrated
14
C-data in Holst, 2007.
Wp Lab-no. Material
14
C age (BP) Cal. BC
d
13
C
1 KI-1883.01 Charcoal 9200 160 8468 183 23.8
KI-1883.02 Hazelnut shell 9170 120 8428 128 23.9
H-431/379 Hazelnut shell 9095170 8277 253
2 KI-1884.01 Charcoal 9420 130 8792 240 24.4
KI-1884.02 Hazelnut shell 9280 100 8516 136 25.5
KIA 326362 Paddle 8477 49 7540 30
KIA 36363 Paddle 826138 7300 90
6 KI-1111 Hazelnut shells 9100 130 8303 194 22.3
KI-1112 Hazelnut shells 8840 110 7967 191 23.2
KI-1113 Hazelnut shells 9090 130 8284 199 23.5
8 KI-1818 Birchbark mat 9640 100 9029 164 25.4
KI-1819 Birchbark mat 9410 110 8774 212 26.7
KI-1885.01 Hazelnut shells 9420 130 8792 240 24.8
KI-1885.03 Charcoal 9440 130 8819 234 23.9
9 KI-3041 Charcoal 9590 90 8992 156
KI-3042 Charcoal 9380 80 8657 101
KI-3043 Charcoal 9600 90 9001 156
KI-3044 Charcoal 9440 80 8824 179
13 (a) KI-2125 Charcoal 8630 160 7774 160 25.85
KI-2126 Charcoal 8700 80 7773 129 25.97
KI-2378 Charcoal 8740 85 7824 149 25.4
13 (b) KI-2128 Charred bark (fireplace apart) 7950 75 6864 131 27.48
19 (a) KI-2731 Charcoal 8040 80 6950 129
KI-2734 Charcoal 7970 90 6876 137
KI-2737 Wood (fireplace) 7840 70 6732 115
KI-2738 Wood (fireplace) 7840 120 6732 182
KI-2739 Charcoal 7970 100 6876 145
KI-2743 Charcoal 7950 120 6862 161
KI-2744 Charcoal 7900 100 6824 154
19 (b) KI-2747 Bark mat 7680 100 6542 83
KI-2736 Charcoal 7600 140 6452 144
D. Holst / Journal of Archaeological Science 37 (2010) 2871e28802874
Author's personal copy
sediment of Wp 6 (analyses by Felix Bittmann; Holst, 2007)
furthermore uncovered carbonised seeds of reed mace (Typha sp.)
(conjoined seeds from one roasted bulb) and a carbonised seed of
yellow water lily (Nuphar lutea).
Harvesting seasons of these ethnographically and archaeologi-
cally documented food resources indicate the end of August/early
September as the time of site occupation. Indicators for other
seasons of occupation such as faunal remains are missing, but in
Fig. 3. Duvensee, Wp 8. Excavation area with roasting structures. Black line: section of the bark mat, shown on photographs on the right (photos: Archaeological State Museum
Schleswig).
Fig. 4. Duvensee, Wp 6. Detail of excavation area with roasting hearths, pine planks, charcoal and hazelnut concentrations (blank outlines: pine stumps of younger age). Below:
section AeB through central hearth.
D. Holst / Journal of Archaeological Science 37 (2010) 2871e2880 2875
Author's personal copy
view of the overwhelming evidence for hazelnut exploitation can
only have played a minor role. The harvesting season for hazelnuts
lasts about only 14 days (Howes, 1948), during which the Duvensee
shore was repeatedly occupied, indicating a stable tradition of land
use throughout the early Mesolithic.
4.2. Quantifications of macro-remains
Macro-remains of hazelnuts and reconstructions of potential
crop yields in comparison to assessments of labour result in high
return rates.
The nutshells could not be recovered or quantified in total during
excavations in Duvensee. Quantifications of the nut amounts pro-
cessed are, therefore, based on extrapolations of shell numbers
counted duringexcavationof a sample of the nutshell layer of Wp 5. A
5l-turf(2520 10 cm) contained 8 whole nuts, 132 shell halves,
139 large and numerous small fragments (Schwantes, 1928). This
corresponds to about 140 hazelnuts or 28 nuts per litre. This volume
was employed for extrapolation to the entire layer of densely packed
nutshells. The surface accounted for at least 12 m
2
(after Schwantes,
1928); for its depth 10 cm are calculated (extent of the sample turf in
accordance with documentation of profiles). This corresponds to
1200 l, accounting for at least33,600 nuts which represent morethan
30 kg of nutmeat (0.9 g per nut kernel: Holst, 2007) containing
180,000 kcal (6000 kcal per 1 kg).
These values compare well with estimations for pit F24 in
Staosnaig, Scotland, (Carruthers, 2000) and extrapolations for
nutshell deposits in Holmegård I, IV, Denmark (Holst, 2007).
The ecological potential of early Holocene hazelnut groves
verifies that these enormous amounts reflect just minimum
numbers of nuts for one season.
4.3. Potential crop yields
High pollen frequencies indicate that early Holocene hazel
populations must have differed enormously from today’s under-
storey. The shrubs or trees possibly formed pure, perhaps even
cultivated hazel woods (Firbas, 1949), dispersed in tessellated plots
of considerable densities (Tinner and Lotter, 2001). The formation
and absolute density of Mesolithic hazel stands are unknown and
cannot be absolutely deduced from pollen frequencies, but can be
approximately estimated from traditional hazel plantations
(Howes, 1948) and ecological studies (Dalke, 1953). The nutrient-
rich loamy substrate of the Baltic young moraine area is among the
most fertile habitats for hazel (Howes, 1948; Firbas, 1949). As hazel
avoids constantly wet soils such as peat (Tallantire, 2002), the area
between the peat margin of the former lakeshore and the Weich-
selian tills east and west of Duvensee are considered as possible
hazel habitats. For the following estimations two hypothetical early
Holocene formations of hazel are assumed, taking into consider-
ation formations of pure hazel woods (here treated as wood-like
plots of plantation size) or a shrub belt along the lakeshore
according to hazel stands at forest margins or as hedgerows today.
Assuming only four plantation-like patchy thickets in the
Duvensee environment each 0.4 ha in size (Table 2:Howes, 1948),
representing only a very small surface of the entire lake area,
a count of 135 reproductive hazelnut shrubs would make a very
conservative estimation adjusted after a figure of 540 trees from
plantations of the same size (Table 2), by taking into account lower
density of wild stands (50%) and a capability of reproduction for
a single shrub just every second year (50%).
The western lakeshore measures 3.5 km. Considering a breadth
of 3 m per tree and a distance of 3 m between two hazel shrubs
(conservative estimation after own observations on free standing
shrubs) a closed hazel margin would consist of 580 plants. Because
of possible gaps or a cover of just half of the lakeshore this is
reduced by 50% (290 trees). Takinginto account unsustainable trees
or other limitations such as restricted exploitation rights another
50% were subtracted and 145 trees remain. So both estimations
(thicket or hedgerow) result in about the same numbers of trees.
Unfortunately data on hazel yields are scarce, especially for wild
occurrences. Their wide ecological amplitude entails highly
differing yields, dependent on ground, light and temperature.
According to comparative data on average 2000 nuts per tree
were assessed (Table 2).
Scaleddown againby 30% to accountforpossiblelosses(e.g.,sterile
nuts), calculations could be based on 1400 nuts (¼1.3 kg) per tree.
Thus 135e
145 hazelnut shrubs would yield about 189,000e203,000
nuts (¼170 e183 kg nutmeat with 0.9 g per nut kernel). According to
numbers given for Kentish cobnut trees (Corylus maxima) which are
doublethe sizeof wildforms,one shrubcould even have yielded 25 kg
nuts or 22.5 kg pure nutmeat, that is 3037e3626 kg for the estimated
shrub number.
Alternative calculations based on quite different reference
figures per ha result in 3.2 kg (Dalke, 1953)or179e358 kg
(Talalay et al., 1984) nuts for the plots, corresponding to an
effective energy gain of 17,280 kcal or 966,600e19,33,200 kcal,
respectively.
Fig. 5. Duvensee. Nut cracking stones. Wp 6 (above): sandstone with abraded surface
areas from different moving directions around central depression (4 mm deep) inter-
fering each other, percussion marks in upper part; below Wp 8, coarse-grained granite
with smooth central depression (different scales).
D. Holst / Journal of Archaeological Science 37 (2010) 2871e28802876
Author's personal copy
The above conservative estimations document affluence and
suggest that the productivity of the natural nut sources were not
a limiting factor in early Mesolithic nut economy!
4.4. Assessments of labour input
Estimates of labour and return rates based upon ethnographic,
historical and experimental data (e.g., Barlow and Mithen, 2000;
Ebersbach, 2003; Howes, 1948; Talalay et al., 1984) emphasize
manpower as the critical value limiting the plant harvest. Not
taking into account the comparatively low expenditure of time for
roasting and transport, at least 1400 nuts can be harvested
(own experiments; Talalay et al., 1984) and 500 nuts cracked per
hour (minimum number after Barlow and Mithen, 2000; Howes,
1948). Processing time outbalances acquisition by far, a good
indicator of the value of this resource. On average 950 storable nuts
Fig. 6. Duvensee, Wp 8 and Wp 6. Interpolation of artefact densities by kriging including all artefacts. White outline: significant isoline areas of maximum increase in artefact
density (0e3 cm); black outline: roasting hearths; black lines: refittings.
D. Holst / Journal of Archaeological Science 37 (2010) 2871e2880 2877
Author's personal copy
(0.9 kg nutmeat, 5130 kcal) could be obtained per person per
hour, equivalent to about 7600 nuts (6.8 kg nutmeat, 40,800 kcal)
per day. According to these results, the return rates of hazelnuts
exceed by far those for wild grasses and acorns, e.g., collected
and processed in the Near Eastern Epipalaeolithic (grasses:
w900e1200 kcal/h; acorns: w850e3000 kcal/h) (Barlow and
Heck, 2002; Wright, 1994)! This is due to the higher energy
content of nuts and their less costly processing. Even considering
possible losses (e.g., burnt nuts), that may have reduced the
already very conservatively calculated output per day of 30% to
5320 nuts or 4.8 kg nutmeat (28,800 kcal) per person, the 11-fold
average adult daily energy requirement (2500 kcal, considering
moderate to high physical activity) could be processed into a stor-
able concentrate. Within the only 14 days of nut harvesting season
one person could have furnished 44% of his annual
energy requirements, storing a considerable surplus as a stockpile
(cf. McComb, 2009).
According to this one person can harvest and process the nuts of
3e4 hazel shrubs per day. Mesolithic crops must have been
determined by group size and duration of stay at the nut groves.
Assumptions of small, “special task groups”, perhaps exclusively
of women (to whom the task of collecting is traditionally ascribed)
have to be rejected for economic reasons since they would
have constituted an insufficient workforce. Processing the above
estimated number of 135e145 trees required about 40 working
days (e.g., 8 persons on 5 days).
Even after very conservative modelling the ecological potential
of hazelnuts in the early Holocene must have exceeded available
manpower. Quantifications of labour and ecological potential as
well as extrapolations of nut remains document concordantly the
high potential of hazelnuts in early Mesolithic subsistence. Hazel-
nuts met a substantial portion of nutrition requirements, providing
a concentrated, easy to shelve and transport fat storage throughout
the year.
5. Discussion
The special task sites Wp 8 and 6 at Duvensee, Northern
Germany, demonstrate essential strategies in Mesolithic subsis-
tence, developed to cope with the difficult living conditions in the
early Holocene forested environments.
Wp 8 and Wp 6, representative for the numerous similar sites in
Duvensee, emphasize the importance of hazelnut exploitation
throughout the early Mesolithic, beginning with the first emer-
gence of hazel in the late Preboreal pollen record. The establish-
ment of a continued seasonal settlement tradition at the special
task camps location over a period of 2400 years is due to many
logistical advantages for occupation that distinguish this ecological
niche, comprising the availability of raw material sources such as
rich hazel stands, flint, sand, wood, but in particular the direct
access to waterways, illustrated by the find of a wooden paddle. An
advanced capacity for transport formed the prerequisite for
intensified seasonal resource exploitation. Quantitative
assessments of the transport weights of nuts, stones and sand
(approx. 260 kg during each occupation) reveal the high efforts for
nut processing and by this confirm the value of this resource.
Modelling of the extent of hazelnut economy documents a large-
scale exploitation of hazelnuts that provided a substantial amount of
nutrition. The high-quality, easy to store and digest resource could
have covered an estimated 44% of human energy demands and was
rich in all essential nutrients. Even a temporarily exclusively hazel-
nut based nutrition would on this evidence not have caused any
deficiency symptoms. The supplementary consumption of the above
mentioned carbohydrate rich fruits as knotweed could further have
balanced the comparatively fat rich hazelnut diet; other fruits must
have provided vitamin C. Meat and fish (supply with vitamin A, zinc)
in seasonally varying proportions of course formed part of the
predominantly plant-based diet, but meat apparently (at least in
certain periods and regions) in contrast to the Palaeolithic could
have played only a subsidiary role in the nutrition.
The intensive utilisation of plant foods, especially hazelnuts,
forms a “hallmark”of the early Mesolithic. For the first time in
European prehistory, subsistence was verifiably based on
a substantial degree on plant food. The achievements of this
innovative strategy of high return-harvestings are evident in the
establishment of a continued settlement tradition at Duvensee. The
protection of the valuable hazelnut thickets that may even have
included their management (Bos and Urz, 2003; Firbas, 1949)
provokes the idea of a right of exploitation or ownership. Storage of
the nutrient-rich fruits could have favoured a reduction of the often
implied high mobility as signs of regionalisation regarding tech-
nical traditions may carefully suggest (Holst, 2007).
5.1. Conclusion
An intensified land use by increased logistical capacity,
concentrated and highly efficient resource exploitation on speci-
alised procurement sites, storage and a plant-based diet are char-
acteristics of early Mesolithic subsistence that distinguish this
period. In this respect it shares common traits with the (albeit
much older) Near Eastern Epipalaeolithic: the innovations of the
early Mesolithic already foreshadow common traits with the
Neolithic way of life. Epipalaeolithic plant utilisation, as part of the
so called “Broad Spectrum Revolution”in comprising a huge variety
of wild plants, obviously forced people to increase insufficient
natural harvests by cultivation (Weiss et al., 2006). The multitude of
plant remains in the archaeological records and the temporarily
coexistent ways of utilisation (gathering, cultivation, domestica-
tion) hamper a clear identification of the use of some taxa and
(quantitative) reconstructions of the absolute importance of gath-
ering versus cultivation (Willcox et al., 2008). In opposition to this,
the early Mesolithic hazelnut economy stands out due to its natural
abundance and a specialised stable gathering system with high
return rates that did not require cultivation. In this, the early
Holocene European environmental conditions offered an oppor-
tunity for high return-harvesting that, instead, could even have
Table 2
References for comparative numbers of potential hazel yields.
Reference nNuts per tree kg Nuts per tree kg/ha Source
Dalke, 1953 ø 2 Ecological study, North-America
Talalay et al., 1984 112e224 Assessment of nut coppices, North-America
Howes, 1948 1500e3000 ø 2e4 Plantation (135 trees per 0.4 ha), England
bis 15
Meg Game (pers. comm. 2006) 50 Plantation of Kentish cobnuts, England
McComb, 2009 970 Wild hazel bush, Northern Ireland
Author’s own observations 2000 Freestanding wild hazel bush, Germany,
Rhineland
D. Holst / Journal of Archaeological Science 37 (2010) 2871e28802878
Author's personal copy
narrowed the foraging spectrum. Basic innovations in nutrition
have been made far before the advent of agriculture, which only
had to be developed in times of scarcity.
In terms of return rates, plant utilisation in the evolution of
hominin nutrition reached its peak in the early Mesolithic! In
contrast to Near Eastern Epipalaeolithic plant exploitation, Meso-
lithic high return-harvesting was not obviously connected with an
increase in population density and social differentiation. This may
be due to the fact that cultivation of hazel plots that would have
forced change to a sedentary way of life were limited or even
unnecessary. On the other hand, the time of early Holocene
hazelnut abundance was comparatively too short for such profound
social developments; Corylus rates declined at the end of the early
Mesolithic and new land use strategies developed.
Acknowledgments
Author would like to thank the Palaeolithic Research Unit
(RGZM) for support, especially S. Gaudzinski-Windheuser and
L. Kindler for valuable comments on earlier drafts of this article. M.
Street and E. Turner kindly provided help with the English
translation.
Author would like to thank U. Tegtmeier, University of Cologne
and F. Bittmann, Lower Saxony Institute for Historical Coastal
Research, for botanical analyses.
The Cusanuswerk, Bonn with a PhD grant provided financial
support for the conduct of the research.
Author would like to thank the two anonymous reviewers for
their valuable comments.
References
Alley, R.B., Ágústsdóttir, A.M., 2005. The 8 k event: cause and consequences of
a major Holocene abrupt climate change. Quat. Sci. Rev. 24,1123e114 9.
Barlow, K.R., Heck, M., 2002. More on acorn eating during the Natufian: expected
patterning in diet and the archaeological record of subsistence. In:
Mason, S.L.R., Hather, J.G. (Eds.), HuntereGatherer Archaeobotany. Perspectives
from the Northern Temperate Zone. Institute of Archaeology UCL, London, pp.
128e145.
Barlow, C., Mithen, S., 2000. The experimental use of elongated pebble tools. In:
Mithen, S. (Ed.), HuntereGatherer Landscape Archaeology. The Southern Heb-
rids Mesolithic Project 1988e98. Oxbow Books, Oxford, pp. 513e521.
Bokelmann, K., 1971. Duvensee, ein Wohnplatz des Mesolithikums in Schleswig-
Holstein und die Duvenseegruppe. OFFA 28, 5e26.
Bokelmann, K., 1981. Eine neue borealzeitliche Fundstelle in Schleswig-Holstein.
Kölner Jahrbuch für Vor-und Frühgeschichte KölnJb, vol. 15, 181e188.
Bokelmann, K., 1991. Duvensee, Wohnplatz 9. Ein präborealzeitlicher Lagerplatz in
Schleswig-Holstein. OFFA 48, 75e114.
Bokelmann, K., Averdieck, F.-R., Willkomm, H.,1981. Duvensee, Wohnplatz 8. Neue
Aspekte zur Sammelwirtschaft im frühen Mesolithikum. OFFA 38, 21e40.
Bos, A.A., Urz, R., 2003. Late Glacial and early Holocene environment in the middle
Lahn River Valley (Hessen, central-west Germany) and the local impact of early
Mesolithic peopledpollen and macrofossil evidence. Veget. Hist. Archaeobot.
12, 19e36.
Bramanti, B., Thomas, M.G., Haak, W., Unterlaender, M., Jores, P., Tambets, K.,
Antanaitis-Jacobs, I., Haidle, M.N., Jankauskas, R., Kind, C.-J., Lueth, F.,
Terberger, T., Hiller, J., Matsumura, S., Forster, P., Burger, J., 2009. Genetic
discontinuity between local hunteregatherers and central Europe’sfirst
farmers. Science 326, 137e140.
Carruthers, W.J., 2000. The charred hazelnut shells and other plant remains. In:
Mithen, S. (Ed.), HuntereGatherer Landscape Archaeology. The Southern Heb-
rids Mesolithic Project 1988e98. Oxbow, Oxford, pp. 407e415.
Clarke, D., 1976. Mesolithic Europe: the economic basis. In: Sieveking de, G.,
Longworth, H., Wilson, K.E. (Eds.), Problems in Economic and Social Archae-
ology. Duckworth, London, pp. 449e481.
Cordain, L., Miller, J.B., Eaton, S.B., Mann, N., Holt, S.H., Speth, J.D., 2000.
Planteanimal subsistence ratios and macronutrient energy estimations in
worldwide hunteregatherer diets. Am. J. Clin. Nutr. 71, 682e692.
Cordain, L., Eaton, S.B., Sebastian, A., Mann, N., Lindeberg, S., Watkins, B.A.,
O’Keefe, J.H., Brand-Miller, J., 2005. Origins and evolution of the western diet:
health implications for the 21st century. Am. J. Clin. Nutr. 81, 341e354.
Dalke, P.D., 1953. Yields of seeds and mast in second growth hardwood forest,
South-central Missouri. J. Wildl. Manage. 17 (3), 378e380.
Ebersbach, R., 2003. Palaeoecological reconstruction and calculation of calorie
requirements at Lake Zurich. Forschungen zur Archäologie im Land Branden-
burg 8, 69e88.
Firbas, F., 1949. Spät- Und Nacheiszeitliche Waldgeschichte Mitteleuropas Nördlich
der Alpen. Band 1: Allgemeine Waldgeschichte. Verlag Gustav Fischer, Jena.
Funck H.,1963. Die Entwässerung des Duvensees. Schriftenreihe desHeimatbund und
Geschichtsvereins HerzogtumLauenburg 10. Kommissionsverlag E. Werner, Mölln.
Gamble, C.S., 1986. The Palaeolithic Settlement of Europe. Cambridge University
Press, Cambridge.
Haak, W., Forster, P., Bramanti, B., Matsumura, S., Brandt, G., Tänzer, M.,
Villems, F., Renfrew, C., Gronenborn, D., Alt, K.W., Burger, J., 2005. Ancient
DNA from the first European farmers in 7500-year-old Neolithic sites.
Science 310, 1016e1018.
Hillman, G.C., 2000. The plant economy of Abu Hureyra 1 and 2 (with contributions
by D. de Moulins). In: Moore, A.M.T., Hillman, G.C., Legge, A.J. (Eds.), Village on
the Euphrates. From Foraging to Farming at Abu Hureyra. Oxford University
Press, New York, pp. 327e422.
Holst, D., 2007. Subsistenz und Landschaftsnutzung im Frühmesolithikum:
Nußröstplätze am Duvensee. PhD Dissertation, University Mainz.
Holst, D., 2008. Zur Entwicklung frühmesolithischer Artefaktproduktion: hand-
werkliche Tradition und Landschaftsnutzung am Duvensee (Schleswig
Holstein). AKorrBl. Archäologisches Korrespondenzblatt 4, 457e476.
Holst, D., in press. Duvenseedsettlement behaviour in the early Mesolithic. In:
Gaudzinski-Windheuser, S., Jöris, O., Sensburg, M., Street, M., Turner, E. (Eds.),
Site-internal spatial organization of hunteregatherer societies: case studies
from the European Palaeolithic and Mesolithic papers submitted at the session
(C58) "Come in.and find out: opening a new door into the analysis of hunt-
eregatherer social organisation and behaviour", held at the 15th U.I.S.P.P.
conference in Lisbon, September 2006. Verlag des Römisch-Germanischen
Zentralmuseums Mainz, Mainz
Howes, F.N., 1948. Nuts. Their Production and Everyday Uses. Faber and Faber,
London.
Hublin, J.-J., Richards, M.P. (Eds.), 2009. The Evolution of Hominin Diets. Integrating
Approaches to the Study of Palaeolithic Subsistence. Springer, Dordrecht.
Jones, M., 2009. In: Hublin, J.-J., Richards, M.P. (Eds.), Moving North: archae-
obotanical evidence for plant diet in Middle and Upper Paleolithic Europe, pp.
171 e180.
Kirbas¸ lar, F.G., Erkmen, G., 2003. Investigation of the effect of roasting temperature
on the nutritive value of hazelnuts. Plant. Foods. Hum. Nutr. 58, 1e10.
Kislev, M.E., Nadel, D., Carmi, I.,1992. Epipalaeolithic (19,000) cereal and fruit diet at
Ohalo II, Sea of Galilee, Israel. Rev. Palaeobot. Palynol. 73, 161e166.
Kuhnlein, H.V., Turner, N.J., 1991. Traditional Plant Foods of Canadian Indigenous
Peoples. Nutrition, Botany and Use. Gordon and Breach Science Publishers,
Paris.
Lage, W., 2004. Zur Interpretation der Lehmstraten in den Feuerstelle des Duven-
seer Moores. Lehmplatten als Gar- und Röstvorrichtungen während des Mes-
olithikums in Schleswig-Holstein. Archäologisches Korrespondenzblatt 34,
293e302.
Lindeberg,S., 2009. Modern human physiology with respect to evolutionary adap-
tations that relate to diet in the past. In: J.-J., Hublin, M.P. Richards.(Eds), The
evolution of hominin diets. Integrating approaches to the study of Palaeolithic
subsistence. Springer, Dordrecht, pp. 43-57.
McComb, A.M.G., 2009. The ecology of hazel (Corylus avellana) nuts in Mesolithic
Ireland. In: McCartan, S.B., Schulting, R., Warren, G., Woodman, P. (Eds.),
Mesolithic Horizons. Papers Presented at the Seventh International Conference
on the Mesolithic in Europe, Belfast 2005. Oxbow Books Oxford and Oakville,
pp. 225e231.
Martínez-Moreno, J., Mora Torcal, R., in press. Spatial patterning and lifestyle at Font
del Ros, a Mesolithic settlement in the south-eastern Pyrenees. In: Gaudzinski-
Windheuser, S., Jöris, O., Sensburg, M., Street, M., Turner, E. (Eds.), Site-internal
Spatial Organization Of HuntereGatherer Societies: Case Studies from the
European Palaeolithic and Mesolithic. Papers submitted at the session (C58)
"Come in.and find out: opening a new door into the analysis of hunter-
egatherer social organisation and behaviour", held at the 15th U.I.S.P.P.
conference in Lisbon, September 2006. Verlag des Römisch-Germanischen
Zentralmuseums Mainz, Mainz
Saklar, S., Ungan, S., Katnas, S., 2003. Microstructural changes in hazelnuts during
roasting. Food. Res. Int. 36, 19e23.
Schwantes, G., 1928. Nordisches Paläolithikum und Mesolithikum. Mitteilungen aus
dem Museum für Völkerkunde Hamburg, vol. 13, 159e252.
Schwantes, G., 1958. Die Urgeschichte. Erster Teil. Wachholtz Verlag, Neumünster.
Schwantes, G., Gripp, K., Beyle, M., 1925. Der frühmesolithische Wohnplatz von
Duvensee. PZ. Prähistorische Zeitschrift 16, 173e177.
Score, D., Mithen, S., 2000. The experimental roasting of hazelnuts. In: Mithen, S.
(Ed.), HuntereGatherer Landscape Archaeology. The Southern Hebrids Meso-
lithic Project 1988e98. Oxbow Books, Oxford, pp. 507e512.
Speth, J.D., 1989. Early hominid hunting and scavenging: the role of meat as an
energy source. J. Hum. Evol. 18, 329e343.
Talalay, L., Keller, D.R., Munson, P.J., 1984. Hickory nuts, walnuts, butternuts, and
hazelnuts: observations and experiments relevant to their aboriginal exploi-
tation in Eastern North America. In: Munson, P.J. (Ed.), Experiments and
Observations on Aboriginal Wild Plant Food Utilization in Eastern North
America. Prehistoy Research Series, vol. VI, 2. Indiana Historical Society, Indi-
anapolis, pp. 338e359.
D. Holst / Journal of Archaeological Science 37 (2010) 2871e2880 2879
Author's personal copy
Tallantire, P.A., 2002. The early-Holocene spread of hazel (Corylus avellana L.) in
Europe north and west of the Alps: an ecological hypothesis. Holocene 12,81e96.
Tinner, W., Lotter, A.F., 2001. Central European vegetation response to abrupt
climate change at 8.2 ka. Geology 29, 551e554.
US Department of Agriculture (USDA), 2009. National Nutrition Database for
Standard Reference, Release 22. National Technical Information Service, USDA,
Springfield.
Wandsnider, L., 1997. The roasted and the boiled: food composition and heat
treatment with special emphasis on pit-hearth cooking. J. Anthropol. Archaeol.
16, 1e48.
Weiss, E., Wetterstrom, W., Nadel, D., Bar-Yosef, O., 2004. The broad spectrum
revisited: evidence from plant remains. Proc. Natl. Acad. Sci. U S A 101,
9551e9555.
Weiss, E., Kislev, M.E., Hartmann, A., 2006. Autonomous cultivation before
domestication. Science 312, 1608e1610.
Weninger, B., Jöris, O., Danzeglocke, U., 2010. CalPal-2007. Cologne Radiocarbon
Calibration & Palaeoclimate Research Package. <http://www.calpal.de/>
(accessed 01.06.10).
Willcox, G., Fornite, S., Herveux, L., 2008. Early Holocene cultivation before
domestication in northern Syria. Veget. Hist. Archaeobot. 17, 313e325.
Wright, K.I., 1994. Ground-stone tools and hunteregatherer subsistence in South-
west Asia: implications for the transition to farming. Am. Antiq. 59, 238e263.
Yellen, J.E., 1977. Archaeological Approaches to the Present. Academic Press, New
York.
Zvelebil, M., 1994. Plant use in the Mesolithic and its role in the transition to
farming. Proc. Prehist. Soc. 60, 35e74.
D. Holst / Journal of Archaeological Science 37 (2010) 2871e28802880
