You are on page 1of 14

71

Ambiguous Images: The Problems and Possibilities of Analysing Rock-art


Images in the Egyptian Western Desert

Daniel James

Figure1MapofEgyptshowingsitesmentionedinthe
text(afterMcDonald2009).

INTRODUCTION

For just over seventy years, analysis and interpretation of
rock-art in the Dakhleh Oasis region of the Egyptian
Western Desert has been influenced by the work of the
German anthropologist and specialist in the study of
religious science, Dr Hans Alexander Winkler. Winklers
observations, published in the preliminary volumes Rock-
Drawings of Southern Upper Egypt I (Winkler 1938a)
and Rock-Drawings of Southern Upper Egypt II (Winkler
1939), were based upon two seasons of rock-art
recording, analysis and interpretation carried out in the
winters of 1936-7 and 1937-8. These two works were
ground-breaking: Winkler (1938a, 1) saw an incredible
opportunity for rock-art images to play a valuable role in
adding to our knowledge of the pre-historic forebears of
Dynastic Egypt, with each image acting out the role of a
written text which could be translated and analysed for
important information.

The influence of Winklers work on current rock-art
studies in the Dakhleh region is most apparent when
examining a group of anthropomorphic images referred to
as female figures (Colour Plates 1 6f (Page 161)), with
varying additional descriptions such as pregnant, fat,
and goddesses (e.g. Berger 2006, 196; 2008; Ikram
2009, 75; Judd 2009, 52-53; Krzyzaniak 1987, 185; 2004,
183; Kuciewicz et. al. 2007, 7; Riemer 2006, 499-500;
2009, 40).
1
These rock-art images were the most
significant find Winkler made during his survey of the
Western Desert and appeared distinct from the material
he had recorded from his earlier work in the Nile Valley
and Eastern Desert (Winkler 1939, 27).

The basic shape of the Winkler goddess is described as
female in profile, with a very simple upper body
contrasted with an exaggerated lower body (Winkler
1939, 28-30; 36). Descriptions of further characteristics
vary slightly from researcher to researcher, but typically
include one or more of the following: a pregnant or
exaggerated belly, decorated skirts or decorated lower
body, exaggerated buttocks, steatopygia, exaggerated
skirts, or exaggerated lower body, and breasts (see Table
1).

Although most examples are considered to be in profile
(Plate 3 and 5; Krzyzaniak 1987, 185; Winkler 1939, 28),
there are also examples interpreted as frontal views (Plate
2). Similarly, while some examples have a featureless or
residual upper body (Plate 2 and 3; Berger 2008, 137-
140; Kuciewicz et. al. 2007, 7-8; Winkler 1939, 28),
there are also examples with this portion of the body
decorated (Plate 2 and 3). Winkler goddesses appear to
range in appearance from figurative forms, with clear
anthropomorphic traits, to non-figurative ones that are,
following Winkler (1939, 29), typically interpreted as
schematised or simplified versions of the more elaborate
figures (Plate 4; Berger 2008, 137-140; Krzyzaniak 1991,
63; 2001, 250; Kuciewicz et. al. 2007, 7-8).

What is striking about this series of rock-art images is
that all current attempts to interpret them are based
exclusively upon the conclusions made by Winkler in
1939. His concluding interpretations were that the images
represent females, typically pregnant, that are
representations of deities associated with fertility
(Winkler 1939, 28-30). There is one variation between
Winkler and those that have followed, however, with
Winkler (1939, 29) ultimately proposing that these
images were representations of figurines or statuettes.

Amongst the literature promoting the Winkler-like
interpretations, there are few observations that question
the divine nature or sex of the images. These range from
minor observations that Winkler goddesses are found in
pairings or positions that may actually suggest some of
the females are in fact males (Berger 2008 140;

1
For further discussions see also: Krzyzaniak (1990, 96; 1991, 62;
2001, 250), Krzyzaniak and Kroeper (1991, 59), McDonald (1993, 44-
45). For the purposes of this paper and simplicity these shall be
henceforth referred to here as Winkler goddesses.
DANIEL JAMES
72
Bergmann 2003), to an acknowledgement that the sex of
the figures is generally unclear (Judd 2009, 52). In both
cases the non-female explanations are swiftly dismissed:
Berger (2008, 139-140) argues that any pairings that may
imply a male-female dichotomy are in fact two different
types of women; Judd (2009, 52) despite his admittance
of the ambiguity of sex, appears to contradict it by stating
that the figures have clear feminine features. Thus the
general view presented in the body of published literature
is that the nature of the Winkler goddesses is not up for
further discussion.

This paper goes beyond this current trend of acceptance
to explore what a modern analysis of the rock-art can
show us. I analysed a collection of published and
unpublished rock-art images through a presence/absence
analysis to suggest how the divine nature and sex of the
Winkler goddesses could be objectively tested. The
necessity of formal and replicable tests is paramount,
given the prevailing trend of images interpreted as
Winkler goddesses to lack any sort of explicit analysis or
explanation as to why they are assumed to be
representations of females, or divine beings. The
continued esoteric context of the interpretation of these
images is detrimental to the much needed and burgeoning
rock-art research of the Dakhleh region.


pregnant/
pronounced
belly
exaggerated
buttocks/
skirt/
lower body
decorated
clothing/
skirt
breasts
Winkler
1939
Yes Yes Yes
Cervicek
1986; 1993
Yes Yes
Krzyzaniak
et. al. 1989
2004
Yes Yes
McDonald
1993
Yes Yes Yes
Bergmann
2003
Yes
Huyge
2003
Yes
Kuciewicz
et. al. 2005
ongoing.
Yes Yes Yes
Berger
2006; 2008
Yes Yes Yes Yes
Riemer
2006; 2009
Yes Yes Yes Yes
Ikram 2009 Yes
Judd 2009 Yes Yes Yes
Table1Integralattributeshighlightedbyauthors
discussingWinklergoddesses.

BACKGROUND: WINKLER AT DAKHLEH

In 1936, the President of the Egypt Exploration Society in
London, Sir Robert Mond, tasked Winkler with leading
an expedition to find, survey, catalogue and analyse rock-
art in the Egyptian Eastern and Western Deserts. Winkler,
with a strong background in contemporary Egyptian
religious practices and folklore, saw the untapped
potential of this material and was of the opinion that its
analysis would reveal much about the religious beliefs
and social institutions of the Egyptian prehistoric periods
(Winkler 1938a, 1).

In the summer of 1938, during a survey along the
Ghubari road from Kharga Oasis to Dakhleh Oasis,
roughly 600 kilometres southwest of Cairo and 250
kilometres due west of Luxor, Winkler first encountered
his goddesses. Looking at his field notes (Winkler
1938b) we see that his initial impressions were that he
had found stylised women, which is an interesting
interpretation given that the first figure he encountered
(Plate 1) displayed no overt sexual characteristics (Ucko
and Rosenfeld 1972, 177-182; Verpoorte 2001, 92).

Winkler (1939, 36) noted that his interpretations were
heavily influenced by prehistoric figurines from Crete
and the Near East, and presumably Europe, which were
being interpreted as evidence for early worship of a
Mother Goddess (James 1959; Ucko 1962; 1968). This
interpretation would expand as Winkler progressed
further along the Ghubari road and into Dakhleh Oasis,
culminating in the preliminary publication from his
survey which detailed his case for pregnant females at
Dakhleh (Winkler 1939).

As part of his analysis, Winkler attempted to place the
rock-art he recorded into a chronological framework
based upon the style or manner in which the rock-art
image had been made. The framework for the Western
Desert was based upon the sequence he had established in
the survey of the Eastern Desert (Winkler 1938a). This
sequence consisted of:

Autochthonous Mountain Dwellers cattle
breeders whose art is characterised by figures
wearing Libyan penis sheaths and images of cattle
(Winkler 1938a, 18-20).
Early Nile-Valley Dwellers art characterised by
Nilotic boats (Winkler 1938a, 24-25). For the
Western Desert, the Early Nile-Valley group was
replaced with an Early Oasis Dwellers group and
constituted the inhabitants of the Oases prior to
Dynastic control (Winkler 1939, 27-30).
Eastern Invaders art characterised by boats of a
type foreign to Egypt (Winkler 1938a, 26).
Earliest Hunters characterised purely on
aesthetics, with any primitive imagery, defined
by Winkler (1938a, 28-30) as geometrical or spiral
designs, or hunting scenes instantly ascribed to
this group.

Winkler (1939, 27-31) attributed his goddesses to his
Early Oasis Dwellers, a group of prehistoric indigenous
inhabitants of Dakhleh Oasis who were sedentary and
practiced plant-cultivation (Winkler 1939, 28). Examples
of the goddesses which appeared to Winkler to be
coarser or were found in association with rock-art
depicting primitive imagery, were assumed to be poor
copies of the goddesses made by the Earliest Hunters
(Winkler 1939, 28; 32). Winkler (1939, 33) concluded
Ambiguous Images: Problems and Possibilities of Analysing Rock-art Images in the Egyptian Western Desert

73
that these two groups lived contemporaneously within the
Oasis, at the same time as the Gerzean/Naqada II period
in the Nile Valley.

DAKHLEH POST-WINKLER

In the summer of 1938, Mond and Winkler (1938c)
discussed the potential of further surveys and sondages at
Dakhleh to gain a better understanding of the Early
Oasis Dwellers group Winkler had discovered. Mond,
however, passed away in October 1938 not long after the
proposal for the additional seasons in Egypt (Newberry
1938). As a result Winkler was left without any sort of
financial backing or income and was ultimately forced to
return to Germany. With the outbreak of the Second
World War, Winklers knowledge of Middle Eastern
language and custom provided an opportunity for work
with the German Foreign Service (Junginger 1995a, 156-
157). This diplomatic role was short lived, with Winkler
eventually seeing combat in North Africa and in Poland
(Junginger 1995b, 105-109). Winkler was shot and killed
on the Eastern Front in 1944, ensuring that his 1938 and
1939 preliminary publications would be his last words on
the rock-art of Dakhleh Oasis.

It was not until nearly fifty years after the publication of
Winklers preliminary results that the next survey of
Dakhleh rock-art was undertaken. This work began in
1984 (Krzyzaniak and Kroeper 1985), with a Polish team
from Poznan headed by the late Professor Lech
Krzyzaniak, working as part of the petroglyph unit of the
Dakhleh Oasis Project (DOP). Krzyzaniaks rock-art
research at Dakhleh was focused on data collection and
answering specific research questions that unfortunately
did not involve re-assessing Winklers work, the result of
this being that the publications that ensued presented
brief discussions of the activities of each survey season,
without major synthesis or discussion (Krzyzaniak and
Kroeper 1985; Krzyzaniak 1987; 1990; 1991; Krzyzaniak
and Kroeper 1991; 1993; Krzyzaniak 1993; Krzyzaniak
2001; 2004). As with Winkler, Krzyzaniaks passing in
2004 has meant that a vast wealth of unpublished material
and research has been lost to us.

SPATIAL CONSIDERATIONS

Winkler (1939, 8; 27) originally spatially restricted the
goddesses to a large depression (Winkler sites 62 68 in
Figure 2), situated in the eastern part of Dakhleh Oasis,
which has since been named the Southeast Basin, a
small depression several kilometres east of the village of
Teneida, and north of the road from Dakhleh to Kharga
oasis (McDonald 1993, 44). Krzyzaniak and his team
focused much of their work in this area, relocating all of
the Winkler sites within the basin in addition to surveying
and recording a number of new rock-art sites at which
Winkler goddesses were found (Krzyzaniak 1987; 1990;
1991; 1992; Krzyzaniak and Kroeper 1985; 1993). These
activities initially confirmed Winklers prescribed spatial
limitations for the pregnant women motif, with no
further examples discovered beyond the limits of the
Southeast Basin.

Figure2WinklersitesatDakhleh(afterWinkler1939).

Work in the past decade, however, has revealed that
Winkler goddesses can be observed at other locations
outside of Dakhleh. Examples have since been found at
the nearby areas of Kharga Oasis (Ikram 2009),
Djedefres Water Mountain (DWM) (Berger 2006; 2008;
Bergmann 2003; Riemer 2006; 2009) and Meri (Berger
2006; 2008; Riemer 2006; 2009), as well as at other
locations within Dakhleh Oasis (Krzyzaniak 2004;
Kuciewicz et. al. 2005; 2008) (see Figure 1).
2


IN WINKLERS SHADOW

The reason as to why Winklers interpretations have
persisted over the years can likely be attributed to a
number of interrelated points. Although rock-art research
was heavily undertaken in the Sahara from the 19th
century onwards (e.g. Barth 1857-58; Rhotert 1952; Van
Noten 1978; see also list of major works in Davis 1979)
no further work was done at Dakhleh until Krzyzaniaks
work on the rock-art material. This resulted in Winklers
(1939) text remaining the only publication that dealt with
the rock-art of Dakhleh in any sort of systematic detail
(Andrs Zboray 2009, pers. comm.).

As with the general archaeological record, the rock-art of
Northern Africa has been traditionally interpreted within
a cultural-history framework, where sequences, cultural
units and chronology have been the primary research
focus (Brentjes 1969, 41-68; Davis 1979, 59-74; Holl
2004, 1-5; Muzzolini 1996, 59-69; 2003, 44-58; Smith
1968, 1-39; 2006, 5-8; Wilcox 1984, 29-49). There is,
however, a recent and ongoing trend in world-wide rock-
art studies to push beyond the question of chronology and
cultural sequence in order to ask questions relating to the
social information that is encoded within the material
(e.g. Barta 2010; Hendrickx et. al. 2010; Holl 2004; Le
Quellec 2008a; May and Sanz 2009). This is not to say
that questions of chronology are unimportant, but they are
not the only important questions to ask of rock-art.

Despite and in opposition to the old chronological trends,
those studying the rock-art within the Dakhleh Oasis
region have long attempted to push beyond the
boundaries of chronological assessment (Krzyzaniak

2
For the purposes of this paper the Dakhleh region is considered to
include Dakhleh Oasis proper and the above listed locations.

DANIEL JAMES
74
1987; Krzyzaniak and Kroeper 1985; Winkler 1939). We
find then that, Krzyzaniak (1987, 183; Krzyzaniak and
Kroeper 1993, 78) with his attention focused on
examining the rock-art in order to understand the pre-
historic attitudes of people towards animals, elected to
adopt Winklers stylistic groups due to the sheer amount
of time it would take to establish a brand new system of
classification. This is despite the concern of
archaeologists such as Hoffmann (1991, 233-239),
Kromer (1970, 385-386) and Trigger (1965, 63-64) who
recognised quite early the inherent flaws of Winklers
system.

The unfortunate but unavoidable consequence of
adopting Winklers classifications was the adoption of the
interpretations and inherent biases that shaped them,
resulting in Winkler-styled interpretations and
chronological/stylistic positioning persisting despite their
inherent flaws. While a decision was later made to
improve upon, but not redo, Winklers typology
(Krzyzaniak 1991, 60; Krzyzaniak and Kroeper 1993,
78), it remains unclear what this entailed and there is no
evidence in the publications that followed that any
significant change had taken place.

This touches on one of the predominant issues regarding
the presumed pre-historic rock-art in and around
Dakhleh: the lack of secure dating. Winkler and those
following him have employed style as a primary
chronological tool, and although it has traditionally been
used as a dating method in rock-art studies (Lorblanchet
and Bahn 1993; Muzzolini 2006), its use in this regard
has been challenged (e.g. Conkey and Hastorf 1990;
Lorblanchet and Bahn 1993). Alternative stylistic theory
suggests that style should be regarded as differentiating
methods and means of communication, which may exist
outside the bounds of chronology (Conkey 1990). Any
pigeon-holing of the Dakhleh rock-art thus disregards
not only the possibility that the so-called schematic
forms of the female image are of a different
chronological period to the elaborate ones, but also
disregards the possibility of entirely different embedded
meanings.

We know, through the work of the DOP, the North
Kharga Oasis Survey (NKOS), and the Arid Climate,
Adaption and Cultural Innovation in Africa project
(ACACIA) that the units Winkler identified through the
rock-art imagery he recorded do not equate with what is
known from the archaeological record. At Dakhleh
specifically, the cultural units identified as belonging to
the early to mid Holocene or Late pre-historic period,
which would correspond to the chronological period of
Winklers Earliest Hunters and Early Oasis
Dwellers/Early Nile Valley Dwellers, are the Masara or
Epipalaeolithic (ca. 9300 6500 cal BCE), Bashendi
Neolithic (subdivided into Bashendi A: ca. 6400 5650
cal BCE, and Bashendi B: ca. 5400 3800 cal BCE) and
Sheikh Muftah (which existed into the Early Dynastic
period, ca. 3800 2200 cal BCE) (most recently
McDonald 2007; 2008; 2009). These units have been
extensively defined over the life of the DOP, based upon
artefact assemblages, site features, site locations, and
organic remains. These groups cannot be broken down
into two simple categories of mobile hunter-gatherers and
sedentary plant cultivators as Winklers stylistic groups
suggest. The subsistence strategies and sedentary nature
of these cultural units is far more complex (McDonald
2009).

What is clear is that the adoption of Winklers now
questionable interpretations and theoretical frameworks
has ensured that rock-art research at Dakhleh is
unintentionally steeped in Winklers bias. These
processes have continued following Krzyzaniaks passing
in 2004, with his work at Dakhleh continued by his
students and colleagues, who hold a similar if not
identical stance (e.g. Kuciewicz et. al. 2007). More
concerning is the adoption of these frameworks by those
working outside Dakhleh: at the previously mentioned
sites of Meri, DWM, and Kharga Oasis, and even by
those working in the Eastern Desert (Berger 2006; 2008;
Bergmann 2003; Ikram 2009; Judd 2009; Riemer 2006;
2009). The result of this is that Winkler-styled
interpretations of females, along with their associated
divine connotations, are reported with little hesitation.

THE PRESENCE AND ABSENCE ANALYSIS

It is the major contention of this paper that any
interpretation of Winkler goddesses should be based upon
objective and replicable analysis of the data, not initial
intuition. In order to show what information this sort of
testing regime can provide, a presence and absence
analysis is performed upon a corpus of Winkler
goddesses. The presence and absence test is based upon
the identification of anthropomorphic traits used to
establish humanity and sexual character (Maynard 1977,
398-399; Rosenfeld 1977; Ucko and Rosenfeld 1972). In
terms of its adoption here, it entails a simple examination
of the presence, or absence, of a number of key
characteristics that readily distinguish a human from an
animal (Ucko and Rosenfeld 1972, 149). These attributes,
referred to as anthropomorphic attributes are:
1. Arms.
2. Hands.
3. Legs.
4. Feet.
5. Head.

The first five attributes test the legitimacy of the
supposed anthropomorphic nature of these figures, that is
to say are they human or human-like? The sex of the
images is tested by the presence or absence of primary
and secondary sexual characteristics (Ucko and
Rosenfeld 1972, 157):
6. Breasts (primary sexual characteristic).
7. Penis/Penis Sheath (primary sexual characteristic).
8. Vulva (primary sexual characteristic).
9. Pubic Triangle (secondary sexual characteristic).
10. Beard (secondary sexual characteristic).

Definition of these characteristics is fairly self-
explanatory, although the arms and breasts categories
Ambiguous Images: Problems and Possibilities of Analysing Rock-art Images in the Egyptian Western Desert

75
need some further justification. Authors discussing the
Dakhleh images are vague with their interpretation of an
area of the figure below the head that is typically referred
to as breasts (Berger 2006, 196; 2008, 138; Riemer 2006,
499; 2009, 40). This attribute is typically represented by a
circular engraving, a cupule (round/pecked engraving), or
curvilinear shape (e.g. examples on Plate 2; 4 and 5).
The interpretation of this attribute has the potential to
critically impact on the results of this analysis if
interpreted incorrectly, making it necessary to test
whether it must be classified as an ambiguous
characteristic or make a definitive statement for it as
either breasts and/or arms.

The sexual characteristics detailed above remain the most
reliable way to determine the sex of a figure when
lacking informed material relating to the pre-historic
periods of the Dakhleh region and the cultural groups that
inhabited it (Taon and Chippindale 1998, 6-7). If sexless
figures are shown to exist amongst the corpus, it raises
questions over the definitive statements researchers
currently make regarding the sex of these images. The
corollary to this is that interpretations based upon this
assumption of sex are also called into question.

In addition to these ten attributes, four others have been
included to determine their impact upon the sex and/or
divinity of the images, referred to here as Goddess
attributes. These characteristics, along with the already
listed breasts, are readily cited by authors discussing the
Dakhleh material and appear to have been integral to their
interpretations (Table 1). As such, they are worth testing
to determine whether their use as markers of divinity is
appropriate. These attributes are:
11. Pronounced belly (often considered an indicator of
pregnancy).
12. Exaggerated Buttocks /Exaggerated Skirts
/Exaggerated Lower Body.
13. Decoration of Head (includes hair).
14. Decoration of Lower Body/Skirt/Buttocks.

The analysis of these attributes not only examines their
individual presence or absence, but any correlation or
patterning of attributes.

THE ISSUE OF ROUNDED ARMS/BREASTS AND
BELTS/PENIS SHEATHS.

Before presence/absence testing of attributes can
commence, the rounded arms/breasts require further
consideration. As discussed earlier, identification of this
attribute could directly impact on the results of any
analysis of anthropomorphic characteristics, making this
a necessary and crucial first step.

As it stands, interpretation of the rounded arms/breast
attribute (e.g. Berger 2008) is based upon personal
opinion or simple observation of the art and not a testable
process that elucidates patterns amongst the corpus
(Maynard 1977, 387). If we look at the corpus of
published Winkler goddesses, what becomes readily
apparent is the fact that all examples with definite breasts
also display arms (e.g. Plate 6.1; see also Krzyzaniak
1987, Fig. 4; 2004, Fig. 2). In addition, no examples of
figures with the rounded arms/breast attribute display
additional arms/breasts (Plate 1, 2 and 5). In other words,
if the rounded shape represents breasts, then there are no
examples of rounded breasts with arms. This rounded
arms/breasts component is also found on figures with a
penis/sheath (Plate 4 and 5). Examples also exist (e.g.
Plate 6.2) where the rounded arms/breast attribute is
shown as an incomplete semi-circular engraving, again
implying an arm interpretation.

Another factor which may impact upon this discussion is
the interpretation of the perspective of many of these
images. They are typically referred to as being in profile
with front-facing/frontal figures also occurring, although
somewhat rarer (Plate 6.1). It should be noted that all
figures with definite breasts are only shown from a front-
facing perspective. If a vague sense of anatomical
accuracy is desired, the rounded arms/breasts
characteristic is not well suited as breasts on a true profile
figure. I would suggest however, that the figures
identified as profiles are not true profile figures and are
instead a composite of features from differing
perspectives. Specifically, we find that the lower body
occurs in profile with an upper body and head from a
front-facing perspective. The combination of perspectives
is certainly a Dynastic artistic convention (Robins 1997,
21) that appears to have been ignored or unnoticed by
those interpreting the Winkler goddesses.

These factors suggest that this rounded arms/breast
characteristic represents arms that have gone through a
system of schematisation and not breasts (Maynard 1977;
Rosenfeld 1977). It may suggest a move to a
standardisation of arm representation given the lack of
uniformity of the attribute throughout the corpus in
general. Thus for the purposes of the analysis to follow,
this attribute will be referred to as and considered to be
arms.

Similar issues arise with regard to an attribute described
as a belt (e.g. Berger 2008) on the Winkler goddesses. I
suggest that this component of the image is, in fact, a
depiction of a penis or penis-sheath, based upon its
positioning at the front of the figure when it occurs (e.g.
Plate 3.4 and 4.2). It is at no time depicted at a different
position on the body, which may be expected of a belt. I
would argue, therefore, that the interpretation of this
attribute as a belt is based upon the preconception that
Winkler goddesses are in-fact female.

CONDITIONS OF THE ANALYSIS

A total corpus of 129 individual figures were subjected to
the formal analysis; 67 from the Winkler archive of
published pictures and unpublished plates (Winkler
1939), 21 from the published works of Krzyzaniak
(Krzyzaniak 1987; 1990; 1991; 1993; 2001; 2004
Krzyzaniak and Kroeper 1985; 1991; 1993), 31 from the
published works of Berger (2008), 7 from Riemer (2007),
DANIEL JAMES
76
2 unpublished examples kindly provided by S. Ikram in
2009, and 1 unpublished example kindly provided by C.
Hope in 2007.

Each image was assessed on paper: as colour, photo-
quality prints (e.g. Plates 1 to 5) for the material of
Winkler, Hope and Ikram, and black and white
photocopies in the case of Krzyzaniak, Berger and
Riemer. The reasons for this difference in image
representation are due to the issues of publication: the
relevant works of Berger, Krzyzaniak and Riemer have
all been published in black and white, making colour
reproductions impossible. The quality of the original
photos also varied considerably, ranging from high-
quality print-outs of digital scans made from original
negatives (in the case of Winkler) to blurred digital
photographs (in the case of Ikram).

These issues impacted upon the analysis, resulting in
attributes that were more difficult to spot on some
examples compared to others. This is an unfortunate issue
relating to the material under study and needs to be kept
in mind when reading the results of this analysis. The use
of a hand-lens with a 10x magnification was employed to
make judgments regarding difficult attributes and to
prevent image quality from critically affecting results.

Of further consideration is the fact that the images under
study are petroglyphs incised or engraved into sandstone
that is quite soft and open to damage from the elements or
human agency. Attributes that were partially damaged but
identifiable were recorded for the purposes of the
presence/absence analysis, but only measured when
complete for the proportional quantitative analysis.
Where damage removed the possibility for attribute
identification, it was recorded as such so as to not impact
on the analysis.

In an attempt to acknowledge and work with these issues,
each image representation not reproduced from a Winkler
plate negative was enlarged, where possible, to allow for
better examination of image attributes. In addition, all
images were examined under a 10x hand-lens as already
mentioned, and also measured with a digital vernier
calliper.

RESULTS

The results of the presence and absence test are provided
below in Table 2 and Figure 3:

arms hands legs feet head
Nr. Present 104 6 28 3 103
Table2Presenceofindividualanthropomorphic
attributes.

Table 2 and Figure 3 show the number of individual
anthropomorphic attributes present amongst the corpus of
images. Where an attribute is present, arms and head are
the most common, making up 43% (n=104) and 42%
(n=103) respectively, followed by legs at 12% (n=28),
hands at 2% (n=6) and feet at 1% (n=3). Of the total
corpus (n=129) only 7.8% (n=10) display no
anthropomorphic attributes at all.

Figure3Breakdownofanthropomorphicattributesby
percentage.

Figure 4 presents a tree diagram that shows the possible
combinations of anthropomorphic attributes, with a total
of 26, ranging from single examples such as Head to
Arms, Hands, Legs, Feet and Head. The actual presence
of these in the corpus is presented in Figure 5 below.

The range and number of combinations of
anthropomorphic attributes present varies, with Arms
and Head by far the most common combination of traits,
making up over 50% of the total corpus (n=70). Arms,
Legs and Head are significantly behind at just over 10%
(n=13), followed by both Arms and images lacking any
attribute at 7.8% each (n=10). These figures suggest that
the artists were far more likely to add arms and a head in
combination to the petroglyph than all other attributes
and that a lack of features is relatively infrequent (7.8%
n=10). The limited ranges of attribute combinations,
when compared to the larger potential range (Figure 4),
are interesting. It is rare for an image to display arms and
hands in any sort of combination with only 4.8% in total
displaying a combination involving these two attributes
(n=6). Legs and feet are slightly less common, making up
only 2.4% (n=3) of images.

Moving to the presence or sexual characteristics we find
that only a small proportion of the corpus includes such
attributes (Table 3). This is in stark contrast to the
anthropomorphic attributes where 89.9% of the images
(Anthropomorphic attributes in Figure 7 less the
Without and Damage categories) display
anthropomorphic traits.


breasts
penis/
sheath
vulva
pubic
triangle
beard
Nr.
Present
3 15 0 2 0
Table3Presenceofindividualsexualcharacteristics.

Only a single example depicts a combination of sexual
traits (see Krzyzaniak 1987, Figure 4), exhibiting both
breasts and a pubic triangle. This leaves a total of 19
individual figures (14.7%) that display a primary or
secondary sexual characteristic. The sexual
characteristics are broken down into their relevant
frequencies in Figure 5.
Ambiguous Images: Problems and Possibilities of Analysing Rock-art Images in the Egyptian Western Desert

77

Figure4Possiblerangeofcombinationsofanthropomorphicattributes.

It is instantly noticeable that the penis/sheath is the


predominant sexual characteristic amongst the images
that can be sexed, making up 75% (n=15). Breasts and
pubic triangle make up the remaining numbers with 15%
(n=3) and 10% (n=2) respectively.

Figure5Breakdownoffigureswithasexual
characteristicbypercentage.
The final component of the presence/absence analysis
turns to the attributes typically cited by authors as the
defining characteristics of the motif (Table 1). Table 4
lists the frequency of occurrence of these attributes (note
that breasts, also cited by these authors, has already been
included in the sexual characteristics analysis and is
therefore not repeated here). An additional decorative
element noted by myself (head/hair) has been added to
this analysis to test its potential importance. The
breakdown of these attributes is presented in Figure 6.

Where these integral attributes are present, they most
often consist of exaggerated skirt/buttocks (100%),
followed by the pronounced belly (42.6%) and
skirt/buttock decoration (34.1%). The additional category
of head decoration comes in last at 13.2%, and the
previously tested breasts category comes in at 15%
(Figure 6). This suggests that the exaggeration of the
skirt/buttocks is of considerably higher significance than
all other integral attributes. The potential combinations
of these attributes and their frequency are points worth
considering (Figure 8 and 9).


pronounced
belly
exaggerated
skirt/
buttocks/
lower body
head
dec.
/hair
dec. of
lower
body/skirt/
buttocks
Nr.
Present
55 129 17 44
Table4Presenceofintegralattributesidentifiedby
authors.

Figure6Breakdownofintegralattributesbypercentage
ofcorpusdisplayingtheseattributes
The potential combinations for the integral attribute, of
which there are 15 in total, differs considerably from the
actual range and frequency of combinations present in the
corpus (Figure 9). The exaggerated skirt/buttocks where
lacking any combination of other attributes is by far the
DANIEL JAMES
78
most commonly occurring, making up 40.3% of the total
assemblage. Figures with a greater number of
combinations are the least occurring, ranging from 5.4%
to 3.9%.

The results pertaining to the pronounced belly indicate
that it never occurs on its own, instead typically
appearing in combination with exaggerated buttocks
(21.7%) or with exaggerated buttocks and skirt
decoration (13.2%). It is rarely shown in a combination
that involves head decoration (7.8%) (Figure 9).

Relating these results to sexual characteristics (Table 3
and Figure 5), it is notable that no image contains a
pronounced belly in combination with any sexual
characteristic, be it of male or female designation. The
difficulty is then in determining whether the belly
attribute represents pregnancy, thereby making it a sexual
characteristic determining its female sex, or whether it
represents something else (obesity, for example, or
simply a stylised normal belly). It is possible that
patterns not elucidated by the attribute analysis may be
noted by a closer examination of the decorative elements.

With regard to links between the belly characteristic and
decoration: of the 55 individual examples of the belly
(Table 4), 27 of them (49.1%) have some form of
decoration. 17 (30.9%) display skirt decoration only, 5
(9.1%) display head decoration only, and 5 (9.1%) also
display both head and skirt decoration. When we consider
that there are only a total 44 examples with skirt/buttock
decoration, 50% (n=22) of these occur on the figures with
a belly. Head decoration, on the other hand, occurs 17
times in total, with 58.8% (n=10) occurring on figures
with a belly. But how does this compare to figures that
may be sexed and is there any pattern?

If we turn to figures that can be sexed as females, we
find that all examples of breasts (n=3) occur with
skirt/buttock decoration (100%) while 66.7% (n=2)
display head decoration. All examples with a pubic
triangle (n=2) display both head and skirt decoration.
What can be noted is that skirt/buttock decoration on
definite females makes up only 11.4% of the total
number of skirt decorations, with head decoration on
females making up slightly more at 23.5% of total head
decorations. As stated, no figures displaying explicit
female sexual characteristics also display a pronounced
belly.

Turning to male figures, decoration on images with a
penis is uncommon with only 20% (n=3) displaying any
form of decoration. 66.7% of these are skirt decoration
(n=2), while 33.3% (n=1) are of head and skirt decoration
combined. Skirt decoration on figures that may be
classified as male therefore make up only 6.8% of the
total number of skirt decorations, while head decorations
on males only make up 5.9%.

This leaves a considerable proportion of figures with skirt
decorations unsexed (36.4%), with a smaller number of
figures with head decoration also remaining unsexed
(17.6%). While the results suggest that definite female
figures are more likely to be decorated than males, this
may be skewed by the smaller number of females in the
data set. The resulting numbers, therefore, are potentially
statistically irrelevant or anomalous, and highlight the
previously mentioned issues relating to the corpus of
figures available.

Figure7Anthropomorphicattributespresentaspercentage(%)ofcorpus.
Ambiguous Images: Problems and Possibilities of Analysing Rock-art Images in the Egyptian Western Desert

79

Figure8Treediagramofpossibleintegralattributecombinations.
DISCUSSION

The results presented above clearly show how an
objective, replicable and transparent analysis can add to
the existing discussion regarding the Winkler goddesses
of the Dakhleh Oasis region. The application of this sort
of testing regime should lend significant weight to
determining whether such figures can be objectively
referred to as either females and/or goddesses, or if
indeed a new form of designation is necessary for some
of these images. If the sex of these images is
questionable, then also questionable is the tendency for
them to be lumped together as a single category or type-
class of rock-art image.

The results of the presence/absence test were unexpected
when considering the prevailing interpretative trend for
these images. Looking at the anthropomorphic nature of
the figures first, it may be stated that only 10.1% of
figures cannot be identified as being anthropomorphic in
nature, based on a lack of any attribute determining
anthropomorphism. These figures, often identified by
authors as h or b shapes, or simply as schematics
have previously been interpreted as schematised or
simplified versions of the female image that have been
subjected to a process of simplification whilst retaining
the original meaning and intent of the more elaborate
images. This view was first put forward by Winkler
(1939, 28) and subsequently came to be adopted in turn
by Krzyzaniak (e.g. Krzyzaniak 1990, 96; 2001, 250) and
others working in the region (e.g. Berger 2008, 138;
Kuciewicz et. al. 2007). Similar interpretations have been
made of supposed schematic female figures shown in
profile from the Upper Palaeolithic in Europe (Rosenfeld
1977) and more recently of Neolithic material from
Turkey (Khan 1990; 1996) and Saudi Arabia (Khan 1990:
1996).

Figure9Integralattributecombinationspresentaspercentages(%)oftotalcorpus.
DANIEL JAMES
80
Such interpretations, at least with regard to the Dakhleh
material, attempt to pigeon-hole images that, at first
glance appear to be vastly different in style if not subject
matter, into a single all-encompassing type category:
female goddesses. This is not to say they are necessarily
incorrect, but the assignation of type-classes to this
material must be based on analysis, not intuition, and
allow for variance between each type (Muzzolini 2006).

Regarding the results pertaining to the presence/absence
of sexual characteristics, interesting questions have been
raised regarding the current trend to sex these images as
female. Only 14.7% of the corpus of 129 images can be
sexed based upon sexual characteristics. Of these 14.7%,
75% (n=15) were determined in my analysis to be male
based upon the identification of a penis or penis sheath.
Of the definite female figures, all are depicted from a
frontal perspective. Male figures, on the other hand, occur
only on figures that exhibit a combination of profile
(lower body) and frontal perspectives (upper body). Also
striking is the apparent lack of combinations of sexual
characteristics, with only one image displaying both
breasts and a pubic triangle. This suggests that varied
methods of sexing representations were either infrequent
or inapplicable to rock-artistry. Similarly interesting is
the lack of any additional sexes: hermaphrodites or
transgendered figures. These figures are not uncommon
in prominent rock-art areas to the southwest of Dakhleh,
in the Gilf Kebir and Gebel Uweinat regions (Andrs
Zboray 2009, pers. comm.).

This leaves a significant 85.3% of the rock-art corpus that
must be considered of undetermined sex. This may have
been intentional by the artists, or alternatively other
methods of sexing existed, known to both the creator and
intended audience. These sexless figures, which may
represent immature individuals or spiritual beings, require
specific explanation and interpretation. That they exist as
a distinct and separate category of image type needs to be
considered.

Given the large portion of the corpus that remains
unsexed, it is worth examining the possibility that another
attribute may be a possible sexual characteristic.
Pregnancy is typically cited (Table 1) as one of the main
descriptive elements of these images and obviously
factors heavily into the pregnant goddess interpretations
typically offered as an explanatory framework. If the
belly attribute does indeed represent pregnancy then it
likely reflects the female sex, though the possibility for
transgendered spiritual beings should also be considered.
As the results of my analysis show, 55 individual
examples of the belly (Table 4) exist in the corpus,
making up 42.6% of the total assemblage.

Recently Berger (2008, 140) claimed that the presence of
pronounced bellies indicated pregnancy due to their
position on the body. This conclusion was based upon the
work of Guthrie (2005, 149) who discussed pregnancy
position and how the swelling of a womans abdomen
generally conforms to a specific shape and direction.
What Berger did not consider is the fact that any modern
obstetrician (Dr Barnett 2009, pers. comm.) will point out
that every woman carries their child in a different way
and position, or that Guthrie (2005, 349) also went on to
note that the swollen abdomens are disproportionately
large compared with the rest of the bulges of the body.
This does not appear to be the case with this corpus of
images.

The results show that the belly does not overlap with any
of the other sexual characteristics, making conclusive
determinations somewhat difficult. As the existence of a
belly alone is not enough to make assertions about the sex
of a figure (Verpoorte 2001, 92), we turn to decoration as
a potential means of assistance. Decoration, at least of the
skirts/buttocks area is another attribute typically cited as
characterising these images (Table 1). This is a little
misleading, as only 44 examples display such decoration,
while another 17 display head decoration, although this
decorative element is not mentioned by researchers
discussing these figures.

When we consider the images with a belly, nearly half of
them are decorated (either head or skirt) and exactly half
of all figures with skirt/buttock decoration also have a
belly. But is decoration characteristic of the figures that
can be sexed? On definite female images we find that all
examples with breasts (n=3) occur with skirt/buttock
decoration (100%), while 66.7% (n=2) display head
decoration. Similarly all figures with a pubic triangle
(n=2) display both head and skirt decoration. Decoration
on images that have been classified as being male is less
common with only 20% (n=3) of images with a
penis/sheath displaying any form of decorative element.
Two thirds of these are skirt decoration with the
remaining third of head and skirt decoration combined.

With regard to the frequency of decoration on sexed
figures we see that only 11.4% of the total number of
skirt decorations are found on females, with nearly one
quarter of total head decorations found on females. Skirt
decoration on male figures make up only 6.8% of the
total number of skirt decorations, while head decorations
on males only make up 5.9%. We find then that a
significant portion of skirt decorations are on either
unsexed figures at 31.8% or figures with a belly at 50%.
The percentages for head decoration are slightly smaller
on unsexed figures at 17.6%, but higher on figures with a
belly at 58.8%.

Again, the low number of identifiable females is
obviously a problem with regards to the analysis of
decoration in particular, resulting in small numbers that
are statistically insignificant. This is unfortunately a
problem of the data that can only be solved through
additional surveying and data publication, highlighting
the need for additional work in the Dakhleh region or for
those working there already to publish more detailed
accounts to facilitate greater analysis. Despite these
problems, my results suggest that neither the belly nor
decoration can be used to make unequivocal statements
regarding the sex of a figure.

Ambiguous Images: Problems and Possibilities of Analysing Rock-art Images in the Egyptian Western Desert

81
It is obvious that there is a need for care when discussing
the sex and divine nature of the supposed females at
Dakhleh. As already shown by Ucko (1968) for the
interpretation of prehistoric figurines from Egypt, Crete
and the Near East: male and sexless figures are often
ignored or dismissed by proponents of Mother Goddess
styled explanations. These sorts of explanatory
frameworks are long outdated in the field of figurine
studies; it would be ill of us to adopt them without
question for Winkler goddesses.

The call for greater objectivity in interpreting these
figures has more at stake than a simple academic
distinction. At the heart of this is what archaeology and
rock-art studies ultimately have as one of their goals: to
make statements about the history and social processes of
people and cultural groups. At present, the ideological
and social interpretations of the early inhabitants of the
Dakhleh Oasis region cannot help but be influenced by
the presentation of this corpus as a group of female
goddesses. The implication is that a Goddess-
worshipping group of people inhabited the area, with a
pre-occupation for engraving a pregnant deity onto the
rocks for worship or ritual purposes. There is a tendency
to then link the worship of this goddess with fertility rites
and associations. This might very well be the case, but as
I have shown, the interpretations rely on links that are
tenuous at best.

The current push for new and greater emphasis on well-
thought out interpretations and analyses, over traditional
chronological assessments (e.g. Barta 2010; Hendrickx et.
al. 2010; Holl 2004; Huyge 2002; Le Quellec 2008a) is
an important one. These advances are being made despite
the problems of some interpretations made of material
from the Saharan region, with their needlessly reckless
assumptions and poor analytical methodology
(Chippindale in Keenan 2005, 484; Le Quellec 2008b,
52). It is time for research in the Dakhleh region to follow
suit.

ACKNOWLEDGEMENTS

Plates 1 to 6 are digital scans made by Michael Morrow
from Winkler's original negatives. These are reproduced
courtesy of the Egypt Exploration Society.

I would like to thank Associate Professor Colin Hope and
Dr Bruno David of Monash University for their
supervision of my MA thesis, on which this paper is
based. I am also indebted to Dr Patricia Spencer and Dr
Christopher Naunton of the Egypt Exploration Society
London, and Mr Mike Morrow of the Rock Art
Topographical Survey in Egypts Eastern Desert for their
assistance with the Winkler archive material held at the
EES, and for their permission to use the scans of
Winklers original negatives. Professor Salima Ikram and
Associate Professor Colin Hope are also thanked for their
permission to use unpublished images from Dakhleh and
Kharga Oasis. Without these images and prints my
analyses and this paper would not have been possible.
Sincere thanks must also go to Susanne Arenhvel for
translating Winklers unpublished field notes. I would
also like to thank Associate Professor Colin Hope, Jess
Cox, Lucia Clayton, Johanna Petkov and Maj. Michael
Lee for commenting on an early draft of this paper.
Finally, research undertaken for this paper would not
have been possible without the financial assistance of
Monash University. Any errors contained within the
above paper are the sole responsibility of the author.

BIBLIOGRAPHY

Barta, M. 2010. Swimmers in the Sand. On the Neolithic
Origins of Ancient Egyptian Mythology and Symbolism.
Prague: Dryada Publishers.

Barth, H. 1857-58. Travel and Discoveries in North and
Central Africa. London: Longman, Brown, Green and
Roberts.

Berger, F. 2006. Relative Chronology of Rock Art at
Djedefres Water Mountain, SW Egypt. In
Archaeology of Early Northeastern Africa: In memory of
Lech Krzyzaniak, edited by Kroeper, K; Chlodnicki, M;
Kobusiewicz, M. Studies in African Archaeology 9.
Poznan: Poznan Archaeological Museum, 137-146.

Berger, F. 2008. Rock Art West of Dakhla: The
Women from Dakhla. Rock Art Research 25(2): 137-
145.

Bergmann, C. 2003. Discoveries in the Western Desert
of Egypt. Carlo Bergmanns Discoveries. Accessed 17
th

July 2009. Available from: http://www.carlo-
bergmann.de/Discoveries/discovery.htm.

Brentjes, B. 1969. African rock art. Translated by
Anthony Dent. London: J M Dent and Sons.

Cervicek, P. 1986. Rock Pictures of Upper Egypt and
Nubia . Naples: Instituto Universitario Orientale.

Cervicek, P. 1993. Chorology and Chronology of Upper
Egyptian and Nubian Rock Art Up to 1400 B.C. Sahara
5: 41-48.

Conkey, M. W; Hastorf, C. A. 1990. The Uses of Style in
Archaeology. Cambridge: Cambridge University Press.

Conkey, M. W. 1990. Experimenting with style in
archaeology: some historical and theoretical issues. In
The Uses of Style in Archaeology, edited by Conkey, M.
W; Hastorf, C. A. Cambridge: Cambridge University
Press, 5-17.

Davis, W. M. 1979. Sources for the Study of Rock Art in
the Nile Valley. Gttinger Miszellen 32: 59-74.

Guthrie, R. D. 2005. The Nature of Paleolithic Art.
Chicaco: University of Chicago Press.

DANIEL JAMES
82
Hendrickx, S; Huyge, D; Wendrich, W. 2010. Worship
without Writing. In Egyptian Archaeology, edited by
Wendrich, W. London: Wiley-Blackwell, 15-35.

Hoffmann, M. A. 1991. Egypt before the Pharaohs: The
Prehistoric Foundations of Egyptian Civilization. Austin:
University of Texas Press.

Holl, A. 2004. Saharan Rock Art: Archaeology of
Tassilian Pastoralist Iconography. Walnut Creek: Alta
Mira Press.

Huyge, D. 2002 Cosmology, ideology, and personal
religious practice in ancient Egyptian rock art. In Egypt
and Nubia: Gifts of the Desert, edited by Friedman, R.
London: The British Museum Press, 192-206.

Huyge, D. 2003. Grandeur in Confined Spaces: Current
Rock Art Research in Egypt. In Rock Art Studies: News
of the World 2. Developments in Rock Art Research
1995-1999, edited by Bahn, P. G; Fossatis, A. Oxford,
Oxbow Books, 59-73.

Ikram, S. 2009. Drawing the World: Petroglyphs from
Kharga Oasis. Archo-Nil 19: 67-82.

James, E. O. 1959. The Cult of the Mother Goddess.
London: Thames and Hudson.

Judd, T. 2009. Rock Art of the Eastern Desert of Egypt:
Content, comparisons, dating and significance. BAR
S2008. Oxford: Archaeopress.

Junginger, H. 1995a. Ein Kapitel Religionswissenschaft
whrend der NS-Zeit: Hans Alexander Winkler (1900-
1945). Zeitschrift fr Relgionswissenschaft 2: 137-161.

Junginger, H. 1995b. Das tragische Leben von Hans
Alexander Winkler (1900-1945) und seiner armenischen
Frau Hayastan (1901-1937). Bausteine zur Tbinger
Universittsgeschichte 7: 83-110.

Keenan, J. 2005. Looting the Sahara: The Material,
Intellectual and Social Implications of the Destruction of
Cultural Heritage (Briefing). The Journal of North
African Studies 10 (3-4): 471-489.

Khan, M. 1990. The Problem of Inter Regional
Cultural/Iconographic Contacts in Prehistory. ATLAL:
The Journal of Saudi Arabian Archaeology 13: 35-57.

Khan, M. 1996. Rock Art Research in the Arabian
Peninsula, Levant and Anatolia. In Rock Art Studies:
News of the World I, edited by Bahn, P. G; Fossati, A.
Oxford: Oxbow Books, 95-104.

Kromer, K. 1970. Comments. In Symposium
international dart prhistorique, Valcamonica 1968,
edited by Anati, E. Brescia: Edizione del Centro, 384-
387.

Krzyzaniak, L; Kroeper, K. 1985. Dakhleh Oasis
Project: Report on the Reconnaissance Season of the
Recording of Petroglyphs, December 1985. The Journal
of the Society for the Study of Egyptian Antiquities 15 (4):
138-139.

Krzyzaniak, L; Kroeper, K. 1991. A Face-mask in the
Prehistoric Rock Art of the Dakhleh Oasis?. Archeo-Nil
1 (May): 59-61.

Krzyzaniak, L; Kroeper, K. 1993. The Dakhleh Oasis
Project: Interim Report on the Second (1990) and Third
(1992) Seasons of the Recording of Petroglyphs. The
Journal of the Society for the Study of Egyptian
Antiquities 20: 77-87.

Krzyzaniak, L. 1987. Dakhleh Oasis Project: Interim
Report on the First Season of the Recording of
Petroglyphs January/February 1988. The Journal of the
Society for the Study of Egyptian Antiquities 17: 182-191.

Krzyzaniak, L. 1990. Petroglyphs and the research on
the development of the cultural attitude towards animals
in the Dakhleh Oasis (Egypt). Sahara 3: 95-97.

Krzyzaniak, L. 1991. Dakhleh Oasis Project: Research
on the Petroglyphs 1990. Polish Archaeology in the
Mediterranean II, Reports 1989-1991: 60-64.

Krzyzaniak, L. 1993. Dakhleh Oasis: Research on the
Rock Art 1992. Polish Archaeology in the
Mediterranean IV, Reports 1992: 80-82.

Krzyzaniak, L. 2001. Dakhleh Oasis: Research on
Petroglyphs 2000. Polish Archaeology in the
Mediterranean XII, Reports 2000: 249-257.

Krzyzaniak, L. 2004. Dakhleh Oasis. Research on
petroglyphs. Polish Archaeology in the Mediterranean
XV, Reports 2003: 181-189.

Kuciewicz, E. 2008. Petroglyphs. In Dakhleh Oasis
Project Prehistory Group: 2008 Final Report. Accessed
20
th
September 2009. Available from:
http://arts.monash.edu.au/archaeology/excavations/dakhle
h/assets/documents/doppg-report-2008.pdf.

Kuciewicz, E; Jaroni, E; Kobusciewicz, M. 2005.
Dakhleh Oasis Petroglyph Unit: New Rock Art Sites,
Season 2005. Polish Archaeology in the Mediterranean
XVII, Reports 2005: 279-284

Kuciewicz, E; Jaroni, E; Kobusciewicz, M. 2006.
Petroglyph Unit. In Dakhleh Oasis Project Report to the
Supreme Council of Antiquities 2005-2006 Season.
Accessed 20
th
September 2009. Available from:
http://arts.monash.edu.au/archaeology/excavations/dakhle
h/assets/documents/dakhleh-report-2005-2006.pdf.

Kuciewicz, E; Jaroni, E; Kobusciewicz, M. 2007. The
Petroglyphs Code. Academia 1 (13): 4-8.

Ambiguous Images: Problems and Possibilities of Analysing Rock-art Images in the Egyptian Western Desert

83
Le Quellec, J.-L. 2008a. Can One Read Rock Art? An
Egyptian Example. In Iconography without Texts, edited
by Taylor, P. London: The Warburg Institute, 25-42.

Le Quellec. J.-L. 2008b. Whats New in the Saharah,
2000-2004?. In Rock Art Studies: News of the World 3,
edited by Bahn, P.G; Franklin, N; Strecker, M. Oxford:
Oxbow Books, 52-88.

Lorblanchet, M; Bahn, P. G. 1993. Introduction. In
Rock Art Studies: The Post-Stylistic Era or Where Do We
Go from Here, edited by Lorblanchet, M; Bahn, P. G.
Oxford: Oxbow Books, v-vii.

May, S. K; Sanz, I. D. 2009. Making Sense of Scenes.
Rock Art Research 27 (1): 35-42.

Maynard, L. 1977. Classification and Terminology in
Australian Rock Art. In Form in Indigenous Art:
schematisation in the art of Aboriginal Australia and
prehistoric Europe, edited by Ucko, P. J. Canberra:
Australian Institute of Aboriginal Studies, Canberra, 387-
402.

McDonald, M. M. A. 1993. Dakhleh Oasis Project:
Holocene Prehistory: Interim Report on the 1988 and
1989 Seasons. Journal of the Society for the Study of
Egyptian Antiquities 20: 43-76.

McDonald, M. M. A. 2007. Twilight of the microlith in
the Late Epipalaeolithic of North East Africa? Locality
268 in Dakhleh Oasis, Egypt. Journal of African
Archaeology 5: 79-90.

McDonald, M. M. A. 2008. Emerging social complexity
in the mid-Holocene Egyptian Western Desert: Site 270
and its neighbours in Southeastern Dakhleh Oasis. In
The Oasis Papers 2, Proceedings of the Second
International Conference of the Dakhleh Oasis Project,
edited by Wiseman, M. F. Oxford: Oxbow Books, 54-87.

McDonald, M. M. A. 2009. Increased Sedentism in the
Central Oases of the Egyptian Western Desert in the
Early to Mid-Holocene: Evidence from the Peripheries.
African Archaeological Review 26: 3-43.

Muzzolini, A. 1996. Northern Africa: Some Advances in
Rock Art Studies. In Rock Art Studies: News of the
World 1, edited by Bahn, P. G; Fossati, A. Oxford:
Oxbow Books, 59-69.

Muzzolini, A. 2003. New Data in Saharan Rock Art
1995-1999. In Rock Art Studies: News of the World 2,
edited by Bahn, P. G; Fossati, A. Oxford: Oxbow Books,
44-58.

Muzzolini, A. 2006. Classifying a set of rock art: how to
choose the criteria. Rock Art Research 23 (2): 171-178.

Newberry, P. 1938. Sir Robert Mond, LL.D., F.R.S.
1867-1938. The Journal of Egyptian Archaeology 24 (2):
208-210.
Rhotert, H. 1952. Libysche Felsbilder; Ergebnisse der XI.
und XII. Deutschen Inner-afrikanischen Forschungs-
Expedition (DIAFE) 1933/1934/1935. Darmstadt: L.C.
Wittich Verlag.

Riemer, H. 2006. Out of Dakhla: Cultural diversity and
mobility between the Egyptian Oases and the Great Sand
Seas during the Holocene humid phase. In Archaeology
of Early Northeastern Africa: In memory of Lech
Krzyzaniak, edited by Kroeper, K; Chlodnicki, M;
Kobusiewicz, M. Studies in African Archaeology 9.
Poznan: Poznan Archaeological Museum, 493-526.

Robins, G. 1997. The Art of Ancient Egypt. London:
British Museum Press for the Trustees of the British
Museum.

Rosenfeld, A. 1977. Profile Figures: Schematisation of
the Human Figure in Magdalenian Culture of Europe. In
Form in Indigenous Art: schematisation in the art of
Aboriginal Australia and prehistoric Europe, edited by
Ucko, P. J. Canberra: Australian Institute of Aboriginal
Studies, 90-109.

Rosenfeld, A. 1997. Archaeological Signatures of the
Social Context of Rock Art Production. In Beyond Art:
Pleistocene Image and Symbol, edited by Conkey, M;
Soffer, O; Stratmann, D; Jablonski, N. G. California:
California Academy of Sciences, 27-51.

Riemer, H. 2009. Prehistoric Rock Art Research in the
Western Desert of Egypt. Archeo-Nil 19: 31-46.

Smith, A. B. 2006. Rock Art of the Sahara. The Digging
Stick 23 (1) (April): 5-8.

Smith, P. E. L. 1968. Problems and Possibilities of the
Prehistoric Rock Art of Northern Africa. African
Historical Studies 1 (1): 1-39.

Taon, P. S. C; Chippindale, C. 1998. Introduction: an
archaeology of rock-art through informed methods and
formal methods. In The archaeology of rock-art, edited
by Chippindale, C; Taon, P. S. C. Cambridge:
Cambridge University Press, 1-10.

Trigger, B. G. 1965. History and Settlement in Lower
Nubia. New York: Yale University Press.

Ucko, P. J. 1962. The Interpretation of Prehistoric
Anthropomorphic Figurines. The Journal of the Royal
Anthropological Institute of Great Britain and Ireland 92
(1): 38-54.

Ucko, P. J. 1968. Anthropomorphic Figurines of
Predynastic Egypt and Neolithic Crete with Comparative
Material from the Prehistoric Near East and Mainland
Greece. London: Andrew Szmidla.

Ucko, P; Rosenfeld, A. 1972. Anthropomorphic
Representations in Palaeolithic Art. In Santander
Symopsium: Actas del Symposium Internacional de Arte
DANIEL JAMES
84
Prehistorico Santander 1972. Madrid: Union
Internationale des Sciences Prehistoriques et
Protohistoriques, 149-211.

Van Noten, F. 1978. Rock Art of Jebel Uweinat. Graz:
Akademische Druck- u. Verlagsanstalt.

Verpoorte, A. 2001. Places of Art, Traces of Fire: A
contextual approach to anthropomorphic figurines in the
Pavlovian (Central Europe, 29-24 kyr BP). Leiden:
Archaeological Studies Leiden University.

Willcox, A. R. 1984. The Rock Art of Africa. London:
Croom Helm.

Winkler, H. A. 1938a. Rock Drawings of Southern Upper
Egypt I: Sir Robert Mond Desert Expedition. Season
1936-1937 Preliminary Report. London: The Egypt
Exploration Society.

Winkler, H. A. 1938b. Winklers Field Diary and Notes
(unpublished). Accessed 4 February 2009. Held in Egypt
Exploration Society, London, Archives.

Winkler, H. A. 1938c. Winkler correspondence
(unpublished). Accessed 4 February 2009. Held in Egypt
Exploration Society, London, Archives.

Winkler, H. A. 1939. Rock Drawings of Southern Upper
Egypt II (Including Uwenat): Sir Robert Mond Desert
Expedition. Season 1937-1938 Preliminary Report.
London: The Egypt Exploration Society.

You might also like