Introduction

1It is a well-known fact to say that a prevalent share of Egyptian rock art, similar to the rest of the Sahara, comprises images of zoomorphic figures. This also applies to the case of Dakhleh, the Oasis situated in the Western Desert of Egypt, where petroglyphs have been produced since at least the Mid-Holocene (fig. 1). Animal depictions, comprising a large part of the corpus of Dakhleh rock art, are both fascinating and problematic for researchers because these images were executed over a long period of time and appear in various cultural contexts. It is fascinating, because, apparently, animals must have had a certain ‘universal’ value to be depicted in petroglyphs across millennia. It is problematic, because very often one is not able to establish a precise chronology of such images. At times, any dating seems impossible to attain.

2Most of Dakhleh animal petroglyphs seem to have been created in the prehistoric times. Often, we are not able to date them precisely, but minimally we can associate them with a broader cultural horizon, such as with ‘pastro-foraging’ communities of the Mid-Holocene (for Dakhleh we distinguish three prehistoric cultural units: Masara, Bashendi, and Sheikh Muftah; McDonald 2016). This concerns mostly the savannah species, particularly giraffes and antelopes (Riemer 2009), although these animals continued to be depicted during historical periods. A recently published chronological sequence of rock art in the Chufu region (Kuper 2014/2015: 296, Abb. 23), which provides the closest parallel for Dakhleh petroglyphs, seems to confirm some of the observations regarding Dakhleh petroglyphs. The sequence suggests, at least indirectly, that one of the main differences between ‘older’ prehistoric rock art (7^th-5^th millennium BC) and ‘younger’ pre- and dynastic specimens (4^th-3^rd millennium BC) lies in the applied techniques of execution where the former is usually fully pecked or smoothed, while the latter is mostly limited to outlines and stick-like figures. This surely cannot be treated as a strict rule, but to a large extent it is applicable to zoomorphic rock art in Dakhleh.

3A chronological association between depictions of savannah animals and Bashendi culture was suggested by Mary McDonald (1990) and Lech Krzyżaniak (1990) on the basis of proximity in spatial distribution of rock art and other archaeological sites. Similar dating of archaeological context was also the main argument in the age estimation of the Djara petroglyphs (Classen et al. 2009). Mid-Holocene archaeological sites provide context for some of the comparable findings in the Meri region (Riemer 2011), as well as in Chufu (Kuper 2014/2015). Chronology of this older horizon of rock art is, thus, largely established upon spatial proximity of other remains, but also corroborated by evaluating of the state of preservation. However, no attempt of direct dating of rock art has been ever undertaken for the rock art of the Dakhleh region.

4During later historical periods, animals still constituted a vital topic and we encounter again some of the species depicted in prehistory, although in many instances it is not necessarily a ‘style’ of images, but rather their state of preservation, which differentiates them from prehistoric examples. Only in the case of horses and camels can one be sure that they are dealing with fairly late rock art, but even these petroglyphs pose some chronological problems.

5It would be ideal to provide an overview of depicted animal species separately for each chronological horizon. However, this would require determining very clear stylistic distinctions between images, as well as other unambiguous criteria of dating. Because such rigid criteria are often difficult to establish, a general overview of all animal pictographs documented thus far in Dakhleh rock art, is discussed. This means that when discussing a particular species, that species referred to may belong to one or more chronological stages and whenever possible the chronological differences are discussed in the ways certain animals were depicted in different time periods.

Sources

6Altogether, 1429 recognised zoomorphic petroglyphs have been subjected to analysis. The images were documented in various areas across the Oasis as a result of the extended research of the Petroglyph Unit of the Dakhleh Oasis Project (Polkowski et al. 2013). The first group of sources constitutes documentation collected by Lech Krzyżaniak between 1985 and 2002. These rock art sites are mainly located to the south and south-east of Teneida village, as well as to the east of Balat. For the sake of convenience this huge area will be called the Eastern Oasis (henceforth: EO).

7The second large complex of rock art sites is known as the Central Oasis (CO), where it is situated upon a sandstone ridge between Ismant and Balat. The work in this area was conducted since 2005, only to be hindered by a political turmoil of the 2011 revolution, and was eventually stopped due to further political consequences (Polkowski 2018a).

8The analysed material consists mainly of photographic documentation. Unfortunately, some of the archival black & white pictures are not of the best quality, which may in some cases, influence the perception of the studied petroglyphs. The last and smallest group of reviewed images were taken from several publications. This group includes those rock art panels published by Hans Winkler (1939) and Pavel Červíček (1986), not previously covered by L. Krzyżaniak, and a few additional panels with zoomorphs provided in publications by Lisa Giddy (1987) and Michel Baud et al. (1999).

Identified species

9After much scrutiny, 1429 zoomorphic elements have been likely identified as animals. More such animal figures could be potentially recognised among Dakhleh rock art if not for the bad state of preservation. In those instances where there was no certainty whether a petroglyph may be identified as zoomorphic, that image was excluded from analysis. Likewise, all depictions of fantastic beasts, such as Seth animals or gryphons, were not taken into consideration. The overview also omits a category of meandering lines, often interpreted as snakes. The subject is still controversial and is discussed elsewhere (Polkowski 2015a).

10Three levels of identification for analysis have been differentiated as follows:

1. ‘certain’ recognition of a species/genus/family;

2. ‘probable’ identification;

3. classification as an ‘undetermined quadruped’.

11Although the first category consists of ‘certain’ identifications, I am aware that some recognitions may still prove to be disputable. All the identified species, along with numbers of recorded petroglyphs, are presented in figure 2.

12There are more than a hundred petroglyphs classified as ‘probably’ depicting a particular species. They are not considered in figure 2, but will be discussed separately for each species. Finally, nearly a quarter of all images (n=318) are recognised as ‘undetermined quadrupeds’. Both badly preserved depictions and those devoid of diagnostic features, are included in this group. In most cases, identification is plausible, although there are some which allow for a careful association with specific taxa. Because level 2 and 3 identifications are not reliable enough, they will not be taken into consideration in much discussion. It is, however, important to note that such vague groups of zoomorphic imagery do exist and that the analysed set of data is thus unavoidably biased. A positive aspect of the collected material is the fact that it constitutes a large representative number of petroglyphs, recorded across a vast area, allowing to drawing conclusions on a regional scale. As such, this material may prove to be sufficient for conducting comparative analyses with rock art assemblages in other regions of Egypt.

13An important aspect of this study is to compare species as depicted in rock art with those recognised in faunal remains from archaeological sites across the Oasis. The latter have been discussed by C.S. Churcher et al. (2008), based on findings from localities associated with Masara, Bashendi, and Sheikh Muftah units, as well as selected dynastic and Graeco-Roman sites. These records constitute an important reference when interpreting rock art from various periods. Because large parts of faunal fossils are interpreted as consumption remains, a comparison of what was eaten and what was depicted might be fruitfully conducted. All zoomorphic categories of Dakhleh rock art is first reviewed, followed by a brief discussion of issues of a more general character regarding animals in rock art of the Oasis.

Overview of zoomorphic imagery

14As indicated in Figure 2, the two animal species clearly prevailing in the corpus of zoomorphic rock art are giraffes and oryxes. They are followed by cattle and ostriches. These four species constitute more than a half of all zoomorphic depictions, and along with other antelopes and gazelles they usually dominate rock art corpora in other areas of the Western Desert (Ikram 2009a; Riemer 2009).

15As many as 11 species recognised in Dakhleh rock art are represented by less than 10 ‘certain’ specimens, and four of them are associated with singular attestations (plus a few ‘probable’ identifications). No matter which period we analyse, it seems that at any point given artists were choosing from a rather limited selection of species. This picture could be verified if we had tools to identify all these specimens labelled as ‘undetermined quadrupeds’. From 318 petroglyphs of that kind some seem to have features which potentially allow us to suggest recognition of particular species, although with considerable reservations about each identification: – 57 resemble antelopes or gazelles; – 11 potentially depict giraffes; – 23 can represent cattle (but they lack distinctive features, e.g. horns); – 4 can be interpreted as felines/canids; – finally, 2 share features of donkeys and 1 resembles a camel or a horse. The remaining 220 undetermined quadrupeds pose an even more difficult riddle to solve, because they are so generic and often featureless, that they can be variously interpreted. The following overview will not consider them, except to bear in mind that our inability to identify them heavily influences all implications derived from this study.

16Each category of zoomorphs will be discussed in terms of variability of formal features, chronology, analogies, and distribution. However, questions concerning their significance and meaning will not be included here because these issues require a separate study. My intention here is to provide some ‘raw’ data, which, despite a growing number of rock art publications, are still rarely presented. Nonetheless, I will conclude my overview with certain remarks and observations about the collected material.

Giraffe

17‘Certain’ depictions of giraffes number 311, in addition to seven ‘probable’ specimens. It is one of the easiest zoomorphs to identify, as in majority of cases these creatures have very long necks. We are likely dealing here with depictions of Giraffa camelopardalis, but other species, such as G. reticulata, may also be represented (Ikram 2009a: 268-9). Thus far, no bone records which could be assigned to Holocene contexts have been found in the Oasis (Churcher et al. 2008: 15). The nearest bones come from Mudpans where depictions of giraffes were also discovered (Riemer 2009: 37).

18Giraffe is an animal which in Dakhleh rock art was depicted in varying positions, orientation or techniques of execution. We can distinguish three positions of the neck observed in the recorded specimens: more than a half of giraffes (164) are depicted with oblique necks (fig. 3b-d), whereas horizontal (fig. 3a, e) and vertical (fig. 3g) positions characterise 83 and 64 specimens, respectively. Neck seems to be the key feature, as it is usually exaggerated in size, sometimes being four times longer than the main body. Not only length, but also its width is often much too large, which suggests that such petroglyphs were not meant to reflect a ‘real’ look of animals, but were rather intended to be easily recognisable.

Photos: P. L. Polkowski. © The Petroglyph Unit of the Dakhleh Oasis Project and the Poznań Archaeological Museum

19Not only necks but also tails and legs are often exaggerated in length. The tails often appear bushy, but this is not a rule. In the majority of cases four legs are depicted, and in only a couple of examples do we observe a more schematic way of drawing limbs, where both front legs and hindlegs are jointed, and along with a lower body they take a shape of an arch (compare with Le Quellec et al. 2005: 35, fig. 27). Other frequently encountered features are ossicones and ears. These are usually indicated schematically as two, three, or four short lines on top of the head.

20What differentiates all these giraffes from each other are mainly proportions of body elements, their overall shapes and the applied techniques of execution. There is a significant amount of idiosyncrasy involved in the first two characteristics, so description of all variants would be impossible to discuss here. Except for a few images, the rest cannot pretend to be termed as naturalistic in style. In Dakhleh giraffe depictions are rather schematic, although the range of schematisation is very wide. The already mentioned exaggeration of body features is not very naturalistic, and this heavily influences proportions of the body. One also gets the impression that nearly all figures are static, as we do not observe any attempt to depict movement in these petroglyphs.

21What we observe, on the other hand, is the exist­ence of several ‘styles’, understood here in terms of applied techniques and treatment of surface rather than other formal features which, as I mentioned above, may heavily differ between the specimens. Among 319 giraffe depictions (8 ‘probable’ included), 128 figures are fully pecked (fig. 3a-b). Of these 11 can be termed as ‘rectangular’ (fig. 3a). It means that their bodies are geometrically shaped, in contrast to the rest of the pecked figures whose bodies are more realistically depicted, being oval or irregular shaped. Among these fully pecked zoomorphs, a group of at least 13 are extremely long-legged (+ 4 long-legged smoothed specimens; fig. 3d). Such legs are usually very narrow and not straight, but rather slightly erratic. To this we can add another 22 pecked images, whose limbs (and often tails as well) are engraved. Except for this engraved feature they share the traits of the fully pecked giraffes. Altogether 150 pecked giraffes constitute the largest group in terms of surface treatment (for analogies from the Western Desert, see: e.g. Ikram 2009b: 75, fig. 10; Riemer 2009: 40, fig. 12; Negro 2009: 127, fig. 4; Kuper 2014/2015: 295, Abb. 20).

22This set of images is followed by another category of giraffe depictions which have been fully or partially smoothed. In the most elaborate images the applied technique may even resemble sunken relief. There are only 11 fully smoothed figures, whereas the further 83 zoomorphs have a depressed main body with engraved (sometimes pecked) limbs (and often a neck). This group is generally analogous to the pecked one regarding other stylistic traits. Among 83 smoothed giraffes (see similar in Le Quellec et al. 2005: 47, fig. 67), 25 have rectangular bodies (fig. 3c; comparable example: Riemer 2009: 37, fig. 7a). Not infrequently, they are carefully executed and, perhaps significantly, these zoomorphs are often depicted with horizontally positioned necks (14 out of 25).

23The next two groups are technically simpler, with both containing only engraved petroglyphs. The first type refers to stick-like images, because such giraffes consist of five to six lines, representing limbs and a body with a neck (and sometimes a tail). Usually the only feature allowing for identification is a long line of neck. Only 15 such depictions have been recognised so far (fig. 3f).

24The second engraved category of giraffes comprises outline figures (fig. 3g). These images are fairly standardised, and although still schematic, can be considered slightly more realistic. The proportions between various features of body, seem to be more accurate. The most characteristic trait of this type of execution is the way the lower body is usually depicted. It is often a gently curved line forming an arch consisting of front legs, hind legs and abdomen. Of the recorded examples, 21 appear to place this group on the margin, but the characteristics seem enough to suggest that they should be treated as a separate visual tradition among the giraffe figures.

25Finally, there are specimens whose bodies were marked in order to imitate patterns on the giraffe’s coat. We can distinguish two patterns of filling, namely parallel lines or lattice. I counted 17 figures treated in this way, out of which 11 are filled in with lines (fig. 3e), and six are checkered. Most of these giraffes are simply outlined, although three specimens had been first smoothed, and then filled in. I classify all of them within one group, but it must be underlined that each has its own stylistic character.

26The remaining petroglyphs of presumed giraffes do not constitute a coherent stylistic group, therefore they will not be discussed here. Instead, I would like to briefly describe three more features which characterise Dakhleh giraffes, no matter to which type they belong. The first phenomenon concerns the so-called ‘sitting’ giraffes (Deregowski & Berger 1997; Van Hoek 2005). These are animals turned 90^o, thus appearing as if sitting on their hindlegs and buttocks. It is a well-known motif (Ikram 2009a: 268, fig. 2), although its understanding is still debated. In Dakhleh only four examples have been noted so far, one of which is turned 90^o, but its head is pointed down. The remaining giraffes are placed in a classic way. Perhaps they depict dead animals (Polkowski 2018b).

27Another trait is a rare body feature, observed in at least seven figures. It is a clear protrusion on a belly, most probably indicating pregnancy (see e.g. Krzyżaniak 1990: 95, fig. 2). A particularly interesting example is known from the Central Oasis where two such pregnant giraffes are juxtaposed with the Dakhleh archetypal motif of a female figure, apparently also enceinte (Polkowski et al. 2013: 110, fig. 11). This may indicate a meaningful link between two types of images, centred on pregnancy and life in general (see summary in Polkowski 2016: 396).

28Yet another archetypical motif, known also from other regions in Africa, is a giraffe on a leash or tether (Van Hoek 2003). So far, 20 such figures have been registered, exactly half in the CO area, and the other 10 in EO. Usually we deal with giraffes which have long lines attached to their necks or heads. In six cases, however, these ‘ropes’ are held by anthropomorphic figures (fig. 3c; Krzyżaniak 2004: 184, fig. 4). The absence of human figures in the majority of such scenes may mean that their presence could have been implicit, and drawing anthropomorphs was not required every-time. But we cannot also exclude a possibility that these two variants referred to slightly different phenomena.

29There is little doubt that most of the Dakhleh giraffe depictions should be attributed a prehistoric chronology. It was Mary McDonald (1990), and subsequently Lech Krzyżaniak (1990), who suggested that rock art in the Southeast Basin and Rock Art Basin (two areas in EO), where giraffes often co-occur with other zoomorphs (mainly oryxes) and ‘female figures’, are usually found in direct contexts of Bashendi archaeological sites. As these sites have been associated with the Bashendi unit phase B (ca. 5600-3800 BC), we can think of 6^th and 5^th millennia as the main period for producing this kind of giraffe petroglyph portrayal.

30Such chronology seems to be confirmed by other researches. For instance, findings in the Meri region (e.g. site Meri 06/12 to the south of Dakhleh) have been tentatively dated by Heiko Riemer (2011: 248-51) to the same time span. The already mentioned chronological sequence for the Chufu region situates giraffe pictures mainly in the 6^th millennium BC (Kuper 2014/2015: 296, Abb. 23). Giraffe depictions were also registered at sites such as Mudpans (south of Dakhleh), associated mostly with Mid-Holocene remains, therefore forming a strong link between the archaeology and rock art (Riemer 2009: 37, fig. 5). The specimens found in Kharga were connected with chronology based on the Nile Valley rock art, and thus dated to the very late 5^th and 4^th millennia BC (Ikram 2009a: 266).

Photos: P. L. Polkowski: a, c; E. Kuciewicz: b.
© The Petroglyph Unit of the Dakhleh Oasis Project and the Poznań Archaeological Museum

31In light of a fairly wide stylistic range of depictions, as presented above, it is reasonable to suggest that some groups of Dakhleh giraffe depictions may have been executed in late prehistory (4^th millennium BC) and during historical periods. This likely concerns most specimens which were produced in outlines. Their overall diagnostic similarity to other ‘late’ animals (such as cattle) may become a favourable argument. Another observation is that some depictions could have been created as imitations, being inspired by clearly visible and abundant prehistoric rock art (Polkowski 2015b: 53). It is also possible that some of the examples labelled here as stick-like giraffes, as well as some of the zoomorphs with patterned interiors, should be also linked with periods after the 5^th millennium BC.

Oryx

32With 221 ‘certain’ and 25 ‘probable’ figures, oryx antelopes are second regarding the number of depictions. Although according to Churcher et al. (2008: 19) no bones of oryx have been found in Dakhleh (but see Pantalacci & Lesur-Gebremariam 2009: 248, Tab. 1, where oryxes are listed in dynastic bone records from Balat), research from other places in the Western Desert confirm that these animals once lived in the area (e.g. in Djara and Mudpans, see Pöllath 2009: 90, tab. 3).

33The most characteristic feature of these images are long horns, very often exaggerated in length. The rest of the body is fairly proportional, although examples with extremely long necks also occur. Two shapes of horns have been noted. The most prevalent are the scimitar shaped horns, which allow for identification of the Oryx dammah species. Within 221 specimens, 33 have clearly straight horns that can potentially depict Oryx beisa (fig. 4e). Beisa is mentioned by C.S. Churcher et al. (2008: 19) as a species which could have inhabited Dakhleh in the past, and D.J. Osborn and J. Osbornová (1998: 165-8) take into account the possibility that some of the rock art depictions in Egypt, including those in the Western Desert, represent Oryx beisa. As this determination is still uncertain, I will stick to a rank of genus, labelling all petroglyphs under the study as oryxes.

34In comparison with giraffe figures (typically between 50 and 100 cm high), oryxes are usually much smaller (often 10-20 cm). Similar to giraffe representations, they also seem to be depicted as rather static. No traits such as pregnancy have been noted, and their associations with anthropomorphs are rare and indirect (with the exception of a scene of an anthropomorph and dogs attacking oryxes and dibatags; unpublished).

35The Dakhleh oryxes were executed in similar techniques to the giraffes. However, in the case of the oryx, 140 were created by smoothing, not pecking. From this group, almost all (n=138) follow a pattern of a mixed technique of smoothing and engraving (fig. 4a). It is the main body which is ground and often smoothed, whereas the limbs, tail, and horns are carved. There is a notable formal variability between the specimens, from more schematic to fairly realistic figures, but one gets the impression that this variation is lower than in the case of giraffes. In the oryxes showing relief, the rectangular ones (n=11) are the most noticeable. This type, appears more rectangular and mechanical in execution. They look more stylized consisting primarily of markedly fine straight lines (fig. 4b).

36Further, 37 oryxes are pecked: 21 fully (fig. 4c), and 16 with engraved limbs and horns (fig. 4d). In many cases these zoomorphs are very similar to the smoothed ones, particularly the latter group. Specimens produced with other techniques are less frequent. Among them, seven are distinguished as stick-like figures (four with curved horns, three straight; fig. 4e) and 27 outlined (17 curved, 10 straight; fig. 4f). The outlined oryxes can be further subdivided into several types, as they vary from more geometric to curved shapes.

37It is likely that the majority of oryxes belong to the same chronological horizon as giraffes. Analogies come, for instance, from the Djara cave (Classen et al. 2009) where similar images (mostly pecked with engraved additions) are tentatively dated to 6^th/5^th millennium BC. The Oryx is the most abundant zoomorphic figure in Kharga (Ikram 2009a: 267-8), and is also widely represented in other areas such as Abu Tartur (Riemer 2009: 36-8, figs. 3, 8) and Chufu. In Chufu, according to Rudolph Kuper, they could be executed as early as during the 7^th millennium BC (Kuper 2014/2015: 296, Abb. 23).

Gazelle

38Generally, the only reliable feature which helps discern gazelles from other quadrupeds are characteristically shaped horns. These should be incurved twice, and usually, although not always, directed backwards (for similar identifications see examples in Osborn & Osbornová 1998: 175-80). There is also another type of zoomorph with incurved horns (pointed forward) which potentially could depict gazelles, but in this inventory they are classified as dibatags.

39Gazelle figures belong to a group of very problematic depictions regarding identification. The fundamental question is whether one can trust the realism of the horn shapes. For example, we saw that giraffe depictions are in most cases exaggerated and far from being anatomically proportional. The same applies to horns of oryxes, which are usually unrealistically long. For this reason we cannot be sure whether the shape and size of potentially recognised gazelles allow us to identify particular species. The same reservations concern other body features such as the length of tails and necks, and overall body size.

40Gazelles constitute an important element of the Holocene fauna in Dakhleh. Food remains from various contexts confirm the presence of Dorcas gazelle (Gazella dorcas) and Loder’s gazelle (Gazella leptoceros). Churcher et al. (2008: 19) suggest that the Oasis could have also been inhabited by Dama gazelle (Nanger dama dama) and Soemmerring’s gazelle (Nanger soemmerringi), and that the bones of the former were collected, for instance, in Mudpans, Chufu, and Djara (Pöllath 2009: 90, tab. 3). Gazelles still inhabit the Western Desert, including the oases (Berke 2009).

41There are only 11 ‘certain’ and nine ‘probable’ depictions of gazelles in our assemblage, however, it is highly possible that many more remain unrecognised within the large group of undetermined quadrupeds. As with oryxes, the general term ‘gazelle’ is used, because precise identification of species would be a very difficult task. It may also be noted, that these zoomorphs which have unrealistically long necks may possibly depict larger animals, such as the Dama, or a smaller leptoceros. The latter may be recognised especially in examples having relatively long horns pointing upwards. Long-necked specimens with horns directed backwards are reminiscent of Dama gazelle, although one cannot exclude the possibility that they were intended to depict Gerenuk of the Litocranius genus (Churcher et al. [2008: 19] list Litocranius walleri among fauna whose remains might be found in Dakhleh).

42Gazelle images usually co-occur with giraffes and oryxes, although they may also be encountered as isolated petroglyphs. In one case a gazelle is juxtaposed with a snake, and in another scene, it seems to be followed by a feline. The very limited number of depictions does not provide a suitable sample to allow for a more detailed discussion on preferred techniques of execution except to note that six specimens were engraved. These outlined figures, along with two other engraved schematic depictions, are probably not older than 4^th/3^rd millennium BC (fig. 5c). Even four pecked images cannot definitely be ascribed to prehistory (fig. 5b). It is only the smoothed figures with engraved additions (n=4) that can be attributed to prehistory (fig. 5a). Gazelles identified in Djara rock art are associated with 6^th/5^th millennium BC (Classen et al. 2009) and provide a good analogy for some Dakhleh specimens. However, as gazelles inhabited the Oasis since the early Holocene, they could have been depicted in different styles in various periods. For instance, gazelles are the most represented wild animals in the recorded bone material at the Old Kingdom settlement site of Ain Asil (Pantalacci & Lesur-Gebremariam 2009: 248, tab. 1). In this case, a dynastic chronology would not be surprising for some of the gazelle pictographs in Dakhleh.

Dibatag

43This medium-sized antelope, called Ammodorcas clarkei, has a slender body which resembles the Gerenuk, except for the horns. Dibatag’s horns are curved only once and face forward. This feature, observable in at least 18 zoomorphic images, is convincing enough to classify them as dibatags, although no remain of this animal have been noted in the Western Desert (it is, however, on the list of species which could potentially inhabit Dakhleh, according to Churcher et al. 2008: 19). Recognition of dibatag in Egyptian rock art would not be a precedent, because the identification was also proposed by D.J. Osborn & J. Osbornová (1998: 174; based on this recognition, Kingdon’s and Hoffmann’s Mammals of Africa provides information of a possible southwards retreat of this species in the Predynastic period, see Wilhelmi 2013: 389). However, one should be aware of the hypothetical character of this identification. Although other scholars notice that the horns pointed forwards are reminiscent of dibatag, they will classify such specimens as gazelles (Riemer 2011: 250-1, fig. 257; see also Ikram 2009a: 275, fig. 6). On the other hand, although no bones of giraffe have been found in Dakhleh, there is no doubt that rock art depicts these animals. This information leads us to touch upon a well-known question, whether zoomorphic rock art necessarily reflects only the local fauna (another example applies to elephant depictions; Ikram 2009a: 276) or could some of the animals typically encountered elsewhere, be depicted, in the rock art of the oases?

44In light of the above, should these 18 ‘certain’ and one ‘probable’ identifications be questioned, bearing in mind that these figures may be just another visual representation of gazelles. However, as a group, they are much more coherent in terms of a ‘style’. Most of the figures have very similar body shapes, as well as body proportions. Nine specimens are characterised by having a rectangular trunk (fig. 6). These, and another further three with more irregular torsos, are partially smoothed, while limbs and horns are engraved. It is a modus which we have already observed in giraffe, oryx, and gazelle depictions. Six figures are fully pecked, but remain very similar to the rest of the dibatag petroglyphs. All 18 figures have been recorded in EO, whereas the only ‘probable’ depiction comes from CO. Judging by the state of preservation, contexts, and applied techniques of execution, I believe that these petroglyphs should be dated to times before the 4^th millennium BC, most probably to the Bashendi period.

Elephant

45Elephants depicted in Dakhleh rock art pose another puzzle. It seems that such extraordinary animals would have gotten considerable attention in the past, which would be then reflected in the petroglyphic record. However, we observe a completely opposite situation, where only four ‘certain’ specimens have been identified. Including two more potential depictions, elephants constitute a very small number of the depicted zoomorphs in Dakhleh.

46C.S. Churcher et al. (2008: 12) report elephant remains (Loxodonta africana) from several Holocene contexts in Dakhleh, but such records are generally rare in the Western Desert. These animals have unique features which allow for a quick identification, such as trunk, ears, and body posture.

47This said, the four identifications are certain and only the two ‘probable’ images must be left doubtful. Examining the examples, we are finding two very different traditions. Two elephants were produced in prehistory, and two seem to belong to dynastic times. The prehistoric ones differ from each other stylistically, but both occur among early types of rock art occupying the same panel. The first one was published by H.A. Winkler (1939: pl. lvi.1) and it is a fully pecked petroglyph having a tail and a trunk solidly engraved. Its tail is slightly too long, but generally we may say that the figure is exaggerated in only some ways. The second elephant petroglyph was found in CO (Kuciewicz & Kobusiewicz 2011: 240-1, fig. 7). It is outlined with a gentle curvature of the body. It has a bushy tail, small oval ears, and two tusks below a long, bent trunk. Even the oval feet are indicated. Only the narrow legs slightly decrease the level of realism in this example (fig. 7).

48These two specimens should be associated with the same rock art horizon as the majority of giraffes and oryxes. This is implied by the pre­sence of the figures in a direct proximity on both panels, but also by analogies from other regions (e.g. Ikram 2009a: 268, fig. 2; Kuper 2014/2015: 293).

49Two dynastic depictions of elephants provide a different case. They are both executed in a very similar way, resembling each other to a great extent. Moreover, the very concept of these images is unique, perhaps allowing us to see the same person as responsible for the execution of both petroglyphs. The first image was reported from Halfat el-Bir site (Giddy 1987: 277, Graffito n^o 5), while the other is located in CO. Both elephants are engraved. They each have a large body with the head separated from it by a single stroke. A straight trunk is fairly short and hangs down. Also, the tail in both cases is short, and the legs are stick-like. The only trait which differs these two images are the tusks which are very long in the CO example, and of typical length in the case of the other. Both elephants have been found in contexts rich in dynastic rock art. Their state of preservation and technique of execution point to Pharaonic times as well.

Ibex

50The only ten ‘certain’ engravings of Capra ibex nubiana were found on the panel where a unique scene of god Amun Nakht spearing a Libyan warrior is shown (fig. 13b; Krzyżaniak 1999: 132-3, fig. 3). The animals have long curved horns and clearly indicated beards considered to be a feature attributed to male ibexes. The body posture, short legs and tails, and overall body proportions, all fit well to ibex characteristics. Their cloven hooves are also marked by two horizontal strokes. The main feature indicating an ibex identification are the very long and curved horns. To my best knowledge, ibexes are not known from the Western Desert, but this can be just another case when non-local fauna become a subject of rock art.

Ostrich

51Ostriches are easier to identify than most of the quadrupeds (Classen et al. 2009: 50). In Dakhleh the majority of ostrich depictions were executed in a very similar way, facilitating their recognition. However, comparable figures known from Kharga have been tentatively identified as storks or flamingos (Ikram 2009b: 71, fig. 4).

52Struthio camelus belongs to species whose bones are not often found in zooarchaeological materials from the Western Desert (Pöllath 2009: 90, tab. 3). However, on most Holocene humid phase sites, the presence of ostrich eggshells have been widely identified (Churcher et al. 2008: 9; Riemer 2009: 37). Eggshells and one bone fragment were also excavated in the Old Kingdom palace at Ain Asil, even though, as discoverers suggest, ostriches were probably extinct in the area before the 6^th dynasty (Pantalacci & Lesur-Gebremariam 2009: 253).

53Ostrich figures are the 4^th most numerous group of zoomorphs in the Dakhleh record, numbering 99 ‘certain’ depictions and 9 ‘probable’. 60 specimens have been found in CO, while 39 identified in EO. They often appear as single figures, but the most impressive are panels on which they have been composed as ‘flocks’. The largest group so far recorded consists of more than 30 birds.

54I distinguished several modes of execution applied to ostrich images, but there is one particular approach which clearly dominates in this assemblage. It employs a smoothing technique, sometimes used so elegantly that it resembles relief en creux. The additions, such as long necks and legs, are usually engraved (fig. 8a). A total of 78 depictions are executed in such a way, although they can differ in terms of selected details. For instance, although the main body of ostrich can be elongated, it can also be smaller and more oval in shape (Riemer 2009: 36, fig. 4). Some ostriches also have toes indicated as three digits, which is interesting, because feet of Struthio camelus are equipped only with two toes each. Some examples have also a schematically depicted tail.

Photos: P. L. Polkowski © The Petroglyph Unit of the Dakhleh Oasis Project and the Poznań Archaeological Museum

55There are 10 pecked figures (including four ‘probable’) with engraved additions (Classen et al. 2009: 50, fig. 6). Two such ostriches have wings – a feature depicted very rarely. Four specimens are fully pecked. Two of them are very simplified, making their identification uncertain. These pecked images seem cruder than the majority of ostrich depictions.

56Finally, another 12 ostrich engravings can be identified, out of which eight are identified as ‘certain’. Three of them are stick-like figures, composed of four extremely simplified lines. Nine depictions are of the outline type, although they differ in terms of other features. For example, some may have necks drawn with a double line, others with a single line. Three specimens have wings depicted in very idiosyncratic ways.

57The group of outline figures seem to belong to historical periods with some to dynastic times (one example juxtaposed with a clearly Pharaonic anthropomorph), and some to a much later era (most probably Islamic Period). It is also recognised that some depictions made in ‘relief’ may be post-prehistoric. Most of them, however, should be attributed to the times of the Holocene humid phase, or a bit later period of the 5^th-4^th millennium BC. Almost all of the ostriches are represented in profile (as most of the animals are), but two depictions found in CO deserve a separate mention because they are shown en face and with outspread wings (fig. 8b). Moreover, their very long necks are bent midway, making these specimens unique and enigmatic.

Dog

58There are several features indicating that some quadrupeds can be identified as dogs (Canis familiaris). Among the characteristics are prick ears, curved tail, and a slender body. Besides formal traits, some of the canid zoomorphs are depicted chasing other animals. This is another indicator of dog behaviour. A few dynastic specimens characterised by a dose of realism can be seen as depicting hounds.

59Bones of dogs are not commonly found in the Western Desert but they were discovered in some numbers in Dakhleh (Churcher et al. 2008: 11; Pantalacci & Lesur-Gebremariam 2009: 255) and Kharga (Lesur et al. 2011: 68). As expected, in both oases, as well as on other desert sites, bones of other canids were also collected (e.g. Canis aureus or Vulpes sp.; Pöllath 2009: 90, tab. 3), thus the possibility exists that some of the dog depictions could actually represent other canid species. However, all these figures which would constitute very questionable identifications of dogs are treated here as ‘undetermined canids’ (n=8). The remaining canid-like depictions are considered as ‘certain’ figures of dogs (except for one jackal image).

60Among 36 such images, 11 were executed with the use of the mixed technique of smoothing and engraving. Typically, it is the main body which is ground, while legs and tail are marked by lines. Only two pictures were entirely pecked, and another two have pecked bodies with engraved additions. It seems that these 15 depictions can be associated with prehistoric rock art horizons, although more precise chronology is difficult to establish at this time.

61Outlined dog elements are seen in 11 cases. They can be further subdivided based on the level of schematic and geometric shapes, but even the most schematic ones have features typical for dogs. Although it is very difficult to date them, it appears that based on style the eleven examples were produced throughout the 4^th and 3^rd millennia. Another five stick-like dog petroglyphs can be added to this number. The canid figures usually have curled tails, and in two cases very conspicuous paws. These dogs were most likely produced in dynastic times. The same applies to four (one of which is smoothed) very well executed figures which can be identified as hounds. Most of them were found in the context of the Old Kingdom watch-post huts, being thus securely dated (Krzyżaniak & Kroeper 1990: 78, fig. 8). The best preserved scene shows two dogs apparently chasing a warthog (fig. 9).

62Dogs hunting wild animals is a recurring theme in Dakhleh rock art. We observe it on at least four additional panels. Two dogs accompany hunters ambushing a giraffe (Krzyżaniak 1990: 95, fig. 2), three others attack an antelope (hartebeest?; Kuciewicz et al. 2014: 240, fig. 11), and finally four chase after a gazelle-like quadruped. The largest group of dogs has been depicted on a flat boulder found in EO, on which oryxes and dibatags(?) are pursued by 11 dogs. All these figures have ground main bodies and engraved other features.

63In a few scenes dogs are seen accompanying cattle and/or humans, but in one unique case a dog figure is shown juxtaposed with a boat (Kuciewicz et al. 2015: 284, fig. 10), and in another case a stick-like canis is enclosed within a rectangle or enclosure shape. It seems that these two panels do not belong to themes widely depicted in Dakhleh Oasis rock art, but rather that they follow a similar pattern recognised in Kharga (Ikram 2009a: 279).

Jackal

64Canis aureus, known also as Golden Jackal, inhabits the entire North Africa. Its bones were recovered from various sites in the Western Desert (Pöllath 2009: 90, tab. 3), for this reason its potential recognition in rock art should not be surprising. However, just one petroglyph allows for such an identification, and is not without some reservations (fig. 10a). The figure has a slender body, an elongated neck, an elongated muzzle, and prick ears. Moreover, it has a long tail, clearly wider than its limbs. This zoomorph is part of a dynastic composition, where it is teamed with an anthropomorph.

Lion

65Churcher et al. (2008: 18-9) list lions (and other feline carnivores) as animals whose remains “might be recovered in later investigations”. However, no Holocene bones of lions have been recovered thus far in Dakhleh and its broad vicinities (e.g. Ikram 2009a: 279).

66All depictions of lions (Panthera leo) found in the Oasis seem to date from dynastic times. There is only one certain figure from CO, which shows a male lion partly executed in a style resembling some of the cattle depictions, with triangular thighs (fig. 10b). Despite some stylizing of the lion, the pose of the animal remains fairly naturalistic. It has a slender and elegant posture, a long swinging tail, which gives the impression that it is slowly moving. Its mane is indicated, as are the facial traits. The shape of the head is unmistakable. A feature which again finds a parallel in some cattle depictions is a baseline beneath the lion’s paws.

67Three more lions are known from EO. The first one, rather simplified, forms a part of a dynastic composition revolving around an Egyptian archer (Winkler 1939: pl. xi.1). Another extraordinary petroglyph panel is documented from one of the Old Kingdom lookout sites (Winkler site 68), where a flat boulder bears depictions of sandals, cattle, a crocodile (?) and a lion. The latter is an outline figure drawn with much precision and an artistic touch. Particularly impressive are the animal’s legs, shaped realistically from thighs to paws. A long tail ended with a tuft hangs down freely. Its dorsum is slightly concave, with what seems to be a mane indicated over the neck. Only the head is slightly flawed appearing too narrow, while the muzzle seems to be a bit too long. There is also an enigmatic feature, namely the horns which protrude from the head. Is it meaningful? Or, perhaps does it constitute a part of an earlier picture?

68The last lion figure is a female. It was executed in a similar style as the above one, although with less care for details. This resemblance leads one to consider it as a lioness (compare with Ikram 2009a: 278-9, fig. 8), although interpretation as another kind of a big cat is also possible. The remains of a cheetah have been identified from Kharga (Lesur et al. 2011: 62, tab. 2). This suggests the alleged lioness may be in fact Acinonyx jubatus.

69One ‘probable’ lion image executed in a very schematic style with a pecked mane and geometric body, as well as four other specimens have been categorized as undetermined felines. From Dakhleh and other Western Desert sites some smaller cats are also known, such as caracal and the wild cat (Felis silvestris), so the “undetermined” images may represent any of these species. However, they may also refer to animals not inhabiting this region, as it probably is in the case of lions.

Warthog

70The only representation of a warthog in Dakhleh is identified in the scene with two hunting hounds (see above; Krzyżaniak & Kroeper 1990: 78, fig. 8). This figure is both exceptional, and controversial. Churcher et al. (2008: 15) prefer to see it as a depiction of a wild boar (Sus scrofa). Nicolas Manlius (2005) also favours such an identification, but does not exclude the possibility that it is in fact a warthog (Phacochoerus aethio­picus).

71This zoomorph has a slender body and a head with a long muzzle (fig. 9). Its legs are depicted in a simple way. It does not seem to have tusks, which likens it to a boar (because boars have smaller tusks), but the way the tail is depicted may favour a warthog interpretation. It is long and raised, and both its shape and the way it is lifted up resemble tails of running warthogs (Harris 2013: 24, fig. unnumbered). Hence I have identified this figure as a warthog (Phacochoerus aethiopicus).

Hare

72Hare, Lepus capensis, features among bone records often and widely, and is identified especially in the prehistoric sites of all periods (Pöllath 2009: 90, tab. 3). However, no prehistoric images of hare have been noted in Dakhleh (the same for Kharga in Ikram 2009a: 285). Instead, only two images and one ‘probable’ of a clearly dynastic type were recorded (fig. 10c). Both have a hieroglyphic look, although the cruder example is explicitly part of a larger composition (Winkler 1939: pl. xi.1).

Lizard

73On one of the panels in EO, juxtaposed with a dynastic image of cattle, is an engraved depiction of a small lizard resembling a gecko. It is shown from above as if sitting on a boulder, with widely spread limbs. A dynastic dating of this image is suggested because it resembles very much the hieroglyph I1 (Gardiner 1957[2007]: 475). Even the tail is similarly bent. Whether it should be viewed as a zoomorphic depiction, or a hieroglyph, is difficult to establish. As a hieroglyph, this sign was commonly used to denote ‘many’, so its co-occurrence with cattle could be meaningful.

Crocodile

74Crocodiles have not been confirmed as part of the Dakhleh fauna, but a few images leave little doubt as to their identification. There are four ‘certain’ and one ‘probable’ crocodile specimens. Perhaps the only prehistoric ones, which possibly may also depict lizards, are the two figures found in CO and published by Ewa Kuciewicz and Michał Kobusiewicz (2012: 269, fig. 6). They are a typical shape seen at many prehistoric rock art sites in Egypt and the Sudan (e.g. Storemyr 2009: 124-5, 130, figs. 5b, 6, 16). Depicted as if seen from above, the crocodiles are rather simplified and consist of the ground slim main body, a head, outstretched limbs, and a long tail (fig. 10d). They co-occur on the panel with prehistoric depictions of large fauna, including ostriches, giraffes, and oryxes.

75The remaining crocodile figures were executed in dynastic or later times. The one published by H.A. Winkler (1939: IV.2) was assigned by him to the Arab period. It also shows a bird’s eye view of the reptile with small limbs ending with dots, and a long tail composed of a line having on both sides a row of short parallel strokes. This crocodile appears to be devouring a human figure.

76There is also an interesting composition in which an anthropomorph (shown partially in profile; dynastic?) is connected by a line with a crocodile depicted as if seen from above. The animal has very short limbs. The line is held by the human figure, so perhaps it is a hunting scene.

Snake

77This is a problematic category of images, because many of the so-called meandering lines can be interpreted as snakes. It was decided, however, to mention only four petroglyphs depicting snakes, which differ from the typical meanders in terms of size. These four specimens are much smaller. All of them are entirely pecked and have a sinuous shape. All were found in CO on two separate sites. The first one is just ca. 8 cm long, and is juxtaposed with a gazelle. The remaining examples are bigger, but still not longer than 40 cm (fig. 10e). They occur alongside an anthropomorph and oryxes. No other details are indicated, which would allow for species identification. It may be added though that among larger meandering lines, not analysed here, there are examples which display possible ‘snake-like’ features, such as a head (see Polkowski 2015a: fig. 5). However, most of the sinusoidal petroglyphs seem to be devoid of them, making their understanding difficult. In fact, it cannot be ruled out that what we take as a head is actually just an illusion.

Bird

78Birds constitute another group of zoomorphs causing difficulties regarding species identification. A large part of the 44 identified bird images were created in dynastic times, but some may predate them. One example is the only pecked petroglyph showing a fairly big bird with a long and bent neck. It bears some resemblance to a flamingo or stork, but this is at best an educated guess (similar in Kharga, see Ikram 2009b: 71, fig. 4).

79Many depictions were executed in a somewhat hieroglyphic style. There is one depiction of a duck or a goose (sign G38 or G39; see Gardiner 1957[2007]: 471), and at least one swallow (G36; as identified in Giddy 1987: 257). There are six tiny engraved birds which possibly depict falcons, again, produced in a hieroglyphic manner. In case of two large images there is no doubt that a falcon was intended (Polkowski et al. 2013: 111, fig. 13d). They were executed by a skilled hand, as they are finely engraved and detailed. It seems that there is also a depiction of an owl which resembles a G17 sign (Gardiner 1957[2007]: 469).

80An extraordinary piece of rock art was found in the early works of the DOP (Mills 1983). It is a scene involving two anthropomorphic figures, two quadrupeds, and 24 birds. Of the bird group 22 are shown inside an enclosure, indicating either a domestic character of these animals, or an attempt at their domestication. According to L. Krzyżaniak (1990: 96), they could depict young ostriches, but I think that other identifications are equally possible. The birds are small in comparison to human figures. They are executed in a very simplified way and almost no detail is shown. We observe only elongated bodies, small heads with beaks, and fairly long legs. Whatever the species, this composition is absolutely unique in the Western Desert rock art.

81The remaining figures, even if produced in similar manners as the ones described above, pose even more difficulties in identifying particular species. Only two stick-like birds can be safely connected with camel depictions, indicating their relatively young age. Perhaps, they represent cranes, or most possibly, egrets.

Fish

82An unmistakable depiction of a fish was found on a panel containing an image of a sickle-shaped boat (site described in Polkowski & Kobusiewicz 2012). The fish had not been observed during the fieldwork, but only subsequently when photographs were analysed. It is a finely smoothed, very shallow, not easily visible petroglyph. It has a triangular tail, and a well indicated dorsal fin (fig. 10f). It depicts either a Nile perch (Lates niloticus), or Nile tilapia (Oreochromis niloticus), the latter being more probable because spines (a dorsal fin) of the engraved fish are even and regular. Both perch and tilapia petroglyphs are known also from Kharga (Ikram 2009b: 73-4, fig. 9).

Cattle

83Cattle depictions constitute a large group of zoomorphs, and are much different from most of the already described petroglyphs. They appear to be depicted at least 107 times, and there are still 11 more ‘probable’ bovine images. Because they are often drawn schematically, it is possible that many undetermined quadrupeds, similar in style but devoid of key traits, could also depict cattle. Nevertheless, the recognised group provides a substantial sample, which enables us to make observations and draw conclusions.

84It seems that a broad majority of cattle depictions are not older than 4^th millennium BC. There are only singular cases when cattle co-occur with savannah-type fauna, and usually stylistic differences between both groups of animals are clearly noticeable. The smoothing technique has been used to create only two cattle depictions: one image interpreted by H.A. Winkler (1939: 29, pl. xxxix.2) as a sacrificial bull led toward female deities; and the other found by L. Krzyżaniak (1990: 96, fig. 5) showing a bull held on a line by a human figure. Only seven (including three questionable specimens) images were pecked into the rock face. Just two examples have a pecked outline (fig. 11b), and three have fully pecked bodies. Except for this, they still resemble the most common cattle depictions recorded in Dakhleh. The remaining two pecked bovines are attributed to a dynastic origin.

85The rest of the cattle depictions were engraved. Excluding six stick-like petroglyphs, the remaining images are drawn in outline, with a high level of variability of stylistic features. The first clear type of bovine zoomorphs includes animals with a distinctly drawn angular head. Such images are often largely schematised. R. Kuper (2014/2015: 296-7, Abb. 23) dates similar depictions to the 4^th millennium BC, and indeed a late prehistoric chronology seems to fit many of these figures which form an intermediary tradition between the earlier hunter-gatherer rock art and the clearly dynastic petroglyphs. It is, however, highly possible that this mode of depiction continued throughout the 3^rd millennium BC, what can be inferred from possible dynastic rock art contexts for some of the figures in Dakhleh, as well as from analogies from other regions. (Huyge 2002: 195, fig. 2).

86Several subtypes identified in this group are distinguished with respect to body shapes. There are 16 outline figures which are marked with an angular body, so their dorsum and underbelly are drawn as straight lines (Riemer 2011: 234-5, figs. 227, 228). The legs are pictured as more or less parallel lines which do not meet (fig. 11c). Further, 11 similar cattle drawings are seen with legs shaped as elongated triangles (fig. 11d). The dorsum remains straight, but the underbelly may vary from straight to curved. Such an arched lower body characterises yet another group of nine specimens, whose legs are not triangular (fig. 11a). Of all sub-types, these seem to be the most comparable to archetypal cattle depictions known from Nubia, apart from the fact that the latter are usually pecked (Kleinitz 2012). To these, there are six additional cattle figures sharing the way the upper body and head is depicted, but differing in the creation of the legs. Each of them have all four limbs incised as straight lines, but their thighs have a triangular shape. The legs of the remaining four bovines end with hoofs consisting of two short lines (fig. 11e).

87In comparison to older rock art, cattle depictions are characterised by a much higher level of stylistic idiosyncrasy. For instance, we can point to another small group of engravings that shares many of the same features with the ones described above but differ regarding the way their heads are depicted. In five figures the head is not well marked, and one gets the impression that the horns protrude directly from the neck (Negro 2009: 133, fig. 46). In another example the situation is opposite, as the neck is indicated by a narrow line connecting the body with the head. This appears to be a unique feature among the studied cattle petroglyphs. Eight specimens are characterised by having rectangular bodies to which straight parallel legs are attached. The abdomen line is continuous, discriminating these figures from the main type described earlier. These examples of cattle depictions may have been produced either during late prehistory (the Sheikh Muftah period), or during the dynastic era.

88There is more certainty in dating when referring to another group of cattle images from Dakhleh. Again, the stylistic spectrum is broad, but most of these depictions occur in dynastic contexts, leaving little doubt regarding chronology. Moreover, some of the bovines refer to more typical Pharaonic iconography. These petroglyphs always have all four legs depicted which is not a rule in case of other cattle depictions. At least 12 cattle zoomorphs have legs incised as elongated triangles, four have elongated necks, in addition to udders and hoofs, whereas two others half-naturalistic figures have thin legs protruding from triangular thighs (fig. 11f). Fairly certain dynastic images are recognised in six very rectangular specimens, including two with hoofs depicted as if they were feet and four depictions which, despite their still rather angular shapes, tend toward a more naturalistic appearance. Finally, at least 11 cattle images can be termed as realistic appearing images, with each of them having their own artistic quality (fig. 11g).

89Only six zoomorphs have body markings (excluding the fully pecked ones). Two are stippled, two striped, and two are infilled with an irregular grid pattern (fig. 11e). Three cows are shown with a short baseline, perhaps depicting not the living animals but statues (?; see Kuciewicz et al. 2015: 282-3, fig. 8). Three bovines are clearly held on a line by a human figure, clearly identified as a dynastic example (Giddy 1987: 284, Graffiti n^os 13, 16). There are also several instances of cattle juxtaposed with signs which possibly can be interpreted as identity or ownership marks (fig. 11f). At least one animal is depicted as being tethered to a stone (Krzyżaniak 1990: 96, fig. 4). I also counted three pairs of cows and calves. Indication of sex is, however, a rare feature among the cattle petroglyphs in Dakhleh. Only four females and four males seem to be marked with udders and penises, respectively. The latter, being much smaller, are inserted between the legs of adult animals. Finally, there is one instance of a cow in a boat, perhaps depicting goddess Hathor (similar image in Darnell 2013: 49-50).

Horse

90There is a clear difference between CO and EO regarding the quantity of horse depictions. Almost all identified horses (n=38) were found in the EO area, and only two engravings were noted in the central parts. Although few petroglyphs could be executed in the late dynastic times, it seems that the majority of images were created later, most probably in the Arab period, and some perhaps in the modern times.

91Except for four smoothed images (fig. 12b), the largest horse group can be labelled as stick-like petroglyphs, comprising 32 specimens. In general, they display a schematic, stylizing of most of the horse elements. This means that they are fairly simple in design, but are marked with some characteristic features, rarely encountered in other zoomorphic images.

92Very simplified examples (n=6) are seen here, in which a minimal amount of straight lines is employed. It is only a specific shape of the head, which allows us to recognise a horse. There are further 10 images only a slightly more obvious. They are also stick-like in form, but the limbs are depicted in a more naturalistic way. Finally, 16 pictures of stick-like horses stand out in terms of applied graphic solutions. Many of such horses have legs that end with small oval holes indicating hoofs (Winkler 1939: pl. ii.1). Hairy horse tails are not uncommon, and the same applies to the manes indicated by a row of vertical lines. In case of several horse representations, elements of harness were also shown, e.g. reins (fig. 12a). Such depictions find very good parallels in the rock art of the Nile Valley, both Egyptian and Nubian (Huyge 1998: 1380, fig. 2b).

93The remaining horse depictions can be divided into two groups. The first comprises four ground specimens, carefully smoothed into rock face. They are drawn in a semi-naturalistic manner. The front parts of animals especially, were elegantly curved. Although the legs are the only engraved features, some of them are portrayed with a dose of realism (Winkler 1939: pl. V.1).

94The last group consists of engraved outlined depictions which are the most naturalistic among the recorded horses. The shapes are gentle, and the body proportions accurate, although some elements appear too schematic or angular. Nevertheless, these depictions differ greatly from other zoomorphs, particularly from the prehistoric images.

95One more common feature characterising horses is the frequent presence of mounted riders. At least 33 anthropomorphs are riding the horses, being very often equipped with weapons, for instance swords, but usually long lances (Winkler 1939: pl. iii.1). The war-like character of these images seems to be indisputable. What is, however, characteristic is a static appearance of almost all of the horse depictions. They look as if standing, even if they feature larger compositions having a narrative nature.

Donkey

96The small number of recognised donkey figures is surprising, considering their large economic importance since at least Old Kingdom times (Förster 2013). I counted 12 ‘certain’ and two ‘probable’ depictions in Dakhleh, which is more than in Kharga, where only one dynastic image of a donkey has been noted so far (Ikram 2009a: 277-8, fig. 7).

97With such a small sample it is difficult to discuss the technical aspects of their creation. However, two groups can be distinguished. One group comprises seven schematic images, of which only the head with long ears serves as an indicator of a species. The second group consists of seven petroglyphs tending toward either a naturalistic, or a dynastic style (fig. 13a). The latter is very characteristic, as it tries to imitate the official Pharaonic iconography (Krzyżaniak 2001: 252, fig. 3). Four of them were found at site 30/450-E4-6 which is one of the watch-posts dated from the Old Kingdom. It might be then an argument for fairly early dating of such images. However, no elements similar to the ones found in Gebel Uweinat, showing a caravan of donkeys (Förster 2013: 317, figs. 43, 44), have been found in Dakhleh.

Camel

98This group of zoomorphs shares many characteristics with horses, and no doubt, many camel depictions were created during the same time period as horses. We can observe similar visual traits, as well as co-occurrences of both animals within the same compositions. Judging by the state of preservation and types of petroglyphs juxtaposed with camel figures, it appears that the majority were executed since the Arab colonisation of the Oasis.

99The pattern of distribution is also similar, as only two camel petroglyphs are known from CO, whereas the remaining 70 have been found in EO. Of course, the most recognisable feature is a hump which is often the only identifiable trait. There is just one camel element produced in smoothing technique, while the rest are engraved.

100The vast majority of drawings are linear and heavily schematic. At least 36 images are stick-like, and would not differ from horses (or would be undetermined quadrupeds), if not for a half-circle indicating a hump (Winkler 1939: pl. iii.1). Sometimes hoofs are visualised as oval holes, and the neck is slightly longer than in the case of horses (fig. 12d). Another eight examples are comparable, except for the fact that instead of a single line indicating body and neck, these features are drawn with two parallel lines.

101There are also 19 outline figures that appear angular and schematic. None of them suggest movement. The only ones which do are stylized, and because they were produced in a very unique manner they cannot be considered naturalistic. As they display a more dynamic posture, these camel depictions are certainly of a different quality (fig. 12c).

102Only 13 out of 72 camels are depicted with riders, which is the feature differentiating them from horse figures. Among the mounted camels, four are shown with palanquins. Although some specimens are involved in larger compositions, even of a bellicose character, the animals themselves are depicted rather generically as if not moving. In almost all cases the legs are straight and perpendicular to the line of underbelly. It is not uncommon that camels depicted as a group or caravan are extremely similar to each other in terms of their static behaviour. We are dealing, thus, with some attempts at narrating different stories but more by juxtaposing certain generic components than by a realistic depiction of such scenes.

Goat

103Even though goat (Capra hircus) was domesticated in the Western Desert as early as the 6^th millennium BC, the only goat depictions I recognised belong to much later periods. There are three ‘probable’ images, which resemble goats, but they lack decisive features. They occur only in dynastic rock art contexts.

Dakhleh Oasis zoomorphs in light of the Western desert rock art

104If we compare the set of animals identified here (and their numbers) with petroglyphs registered in Kharga (Ikram 2009a; 2009b), Chufu and Meri (Riemer 2011; Kuper 2014/2015), Djara (Classen et al. 2009), and more broadly from the Western Desert (Riemer 2009), we will observe more similarities than differences.

105A prime similarity is that the clear domination of giraffes and oryxes is a common phenomenon in that part of Egypt. The only site, from those mentioned above, which lacks giraffes is Djara cave. Elsewhere giraffes dominate not only in terms of numbers, but also in the scale of visibility, where they are the largest petroglyphs of all. In Kharga they are second regarding quantity, being outnumbered by oryxes. There is, however, no doubt that these two animals were the most frequently depicted zoomorphs in the prehistory of this region.

106This case is similar for gazelles, ostriches, and cattle. Although not so numerous, they are among the most frequent seen petroglyphs at other sites. It is only in Djara, that we find a slightly different situation, where no cattle depictions were recorded. However, in the regions surrounding Dakhleh we also find images of species which have not been recognised in Dakhleh rock art, or at best are among unclear images.

107One such an example is Addax antelope (Addax nasomaculatus). Its most characteristic feature are the corkscrew-shaped horns, which make it difficult to mistake it for any other bovid. Many Addax depictions were found in Djara (Classen et al. 2009: 50, fig. 6), and some are known from Abu Tartur sites, situated close to Dakhleh (Riemer 2009: 36, fig. 3). But in the Oasis itself (as well as in Kharga; see Ikram 2009a: 285) no such images have been found, except for one possible example of the late prehistoric or dynastic date. According to Churcher et al. (2008: 19) Addax bones were not recorded from Dakhleh, but its remains are known from other areas, such as those from Mudpans (Pöllath 2009: 90, tab. 3). This means that these antelopes formerly inhabited the region and their absence from the large corpus of Dakhleh rock art may be significant.

108Another missing animal is barbary sheep (Ammotragus lervia). Known also as arui, it has a long mane extending down under the throat, and characteristically shaped horns curved outwards and backwards. Barbary sheep depictions are known from Meri sites (Riemer 2011: 248-54, figs. 255, 258, 261, 263; Hendrickx et al. 2009), Kharga (Ikram 2009a: 274-5, fig. 6), or Sito dei Mufloni (Morelli et al. 2006: 179, fig. 9), but no certain image is identified from Dakhleh. There is one zoomorph which resembles barbary sheep, with similarly curved horns but body features prompt me to treat it as a depiction of cattle. Pantalacci and Lesur-Gebremariam (2009: 250-1) report bone remains of barbary sheep from Ain Asil. It is also mentioned in the inscribed materials found at this site. Once again it means that this species could have been encountered in the area, but for some reasons it was chosen to be not depicted.

109Among other animals depicted elsewhere, but not in Dakhleh, is the hippopotamus (Hippopotamus amphibius). It is a very specific beast, so it is rather difficult to believe that its depiction may be featured among the ‘undetermined quadrupeds’ category. It is much more likely, however, that individual figures of addax and barbary sheep, devoid of key traits, could be found in this group. Our inability to identify all recorded zoomorphs is a bias which will be extremely difficult to overcome.

Zoomorphs versus faunal remains

110It was already noticed by H. Riemer (2009: 38) and S. Ikram (2009a: 285) that there is an evident discrepancy between the animals depicted on rocks, and the faunal remains obtained from prehistoric sites. Dakhleh fits this scenario very well.

111According to C.S. Churcher et al. (2008), prehistoric localities provided no bones of giraffes and oryxes, even though there are more than 500 images of these two animals. This cannot be a purely taphonomic issue, especially considering that bones of giraffes are sturdy and resistible. We should not therefore consider that rock art reflects only a local environment. Rock art themes were chosen intentionally, and perhaps giraffes and oryxes were depicted precisely because these animals were so rare, or even absent in the Oasis landscape. Such an interpretation may be valid here, until we realise that this situation may be also the result of cultural taboos. If high number of depictions of giraffes reflects in any way their cultural significance, then perhaps it was prohibited to hunt these animals, which would effect in no faunal remains of this kind in archaeological record.

112We can also discuss this subject from an opposite perspective, by comparing those animals which are well represented in faunal remains with their petroglyphic counterparts. If we assume, that the majority of giraffe and oryx images were created in the Holocene humid phase, particularly in times of the Bashendi culture (roughly 7^th-4^th millennium BC), we will see that in that period the most commonly consumed animals were: gazelle, hartebeest, fox, hare, and birds, including ostrich (Bashendi A: McDonald 1998: 132), with a considerable contribution of goats and cattle during the Bashendi B phase (McDonald 1998: 134).

113We have already seen that gazelles are recognised in only a small number of total zoomorph representations, while the rest of the consumed animals are virtually non-existent in rock art. I have not identified any foxes or hares that could be associated with the prehistoric time horizon, and there are only two ‘probable’ hartebeest (Alcelaphus buselaphus) depictions, with one being of late prehistoric or early dynastic origin. Of course, the ‘undetermined quadruped’ category may contain bovids which were meant to depict hartebeests. But even if that is the case, such unidentified specimens cannot be numerous. It suggests that although gazelles and hartebeests were a frequently hunted fauna they were very rarely depicted due to reasons not yet understood. It is, of course, possible that those animals which had a certain symbolic value – to some extent reflected by rock art – might have been considered taboo, and therefore not suitable for consumption. Other animals, such as gazelles or hartebeests, would be then considered free of taboo but also not very important within a symbolic realm of prehistoric people. In light of this it is interesting to note that only ostriches are substantially represented in both faunal remains and rock art findings.

114Another issue concerns depictions in rock art of domesticated animals. For example, we could assume that such a huge economic change as the introduction of cattle and goats, which probably occurred sometime during the late Bashendi A phase (McDonald 2016: 186), should be reflected in the rock art. But apparently it is not. Providing that the proposed dating of the majority of cattle imagery to the 4^th millennium BC and the later is correct, the Holocene humid phase rock art is actually devoid of cattle depictions. To my knowledge, there is not a single rock art panel in Dakhleh, on which ‘Bashendi’ giraffes, oryxes, or ostriches, are accompanied by cattle executed in a similar manner. Is this a result of our biased understanding of rock art and our problems with establishing chronology or do we tend to overrate the significance of economic changes? In other words, should we expect rock art reflect such changes at all? This may also indicate that the economic changes were actually gradual and slow (and not associated with incoming populations) and the change from wild fauna to cattle as a rock art theme constituted a long-term process.

115What we are missing in our prehistoric assemblage from Dakhleh are depictions of carnivores. For example, previously mentioned foxes are absent in rock art. The same applies to other species we know existed in the Dakhleh. For instance, C.S. Churcher et al. (2008: 11-2) mention the evidence of hyaenas, either striped (Hyaena hyaena dubbah), or spotted hyaena (Crocuta crocuta), but except for one ambiguous specimen it is hard to identify any hyaena depictions in Dakhleh. We know that the Western Desert was also inhabited by wild cats, caracals (Pöllath 2009: 90, tab. 3) and perhaps even roamed by cheetahs or lions (Churcher et al. 2008: 18-9). But when we identify such carnivores in rock art, it appears that we are dealing with dynastic petroglyphs. The prehistoric corpus of zoomorphs was therefore centred upon particular species whose nature must have had a specific reference and significance within the conceptual realm of the Oasis dwellers. For some unknown reasons carnivores were excluded from the visual narratives on rocks.

116It is an altogether different situation, when we consider historical zoomorphic imagery, where depictions of cattle and donkeys, and later camels and horses, have a more representational character and their economic significance was important for the culture. Of course, it does not mean that rock art depictions were only about representation, as they also referred to various specific phenomena. What is also explicit, is that domesticates outnumber wild fauna in such ‘late’ rock art. For example, even though animals such as gazelle still inhabited Dakhleh in Graeco-Roman or Arab times, they were rarely depicted, if at all.

Prehistoric petroglyphs in Dakhleh

117Briefly speaking, we can interpret the prehistoric zoomorphic rock art figures in terms not only of their quality, but also quantity. By quality I refer here to both what a petroglyph depicts, and what specific features it has. The former involves considering the nature of particular species and attempting to associate this nature with the potential significance of rock art to their makers. In other words, we ask: what is so specific in the giraffe that it became such an important animal to visually document? At some point, however, we have to realise that it is not just a generic image of a giraffe which we are dealing with, but the variety of giraffes. They all depict the same animal, but details can be considerable. Was a depiction of a giraffe with an unnaturally long neck semantically identical to a more naturalistic picture? What was the purpose of drawing very long legs? Was it just another ‘style’? Or rather an intentional indication of a very specific idea?

118The same can be asked regarding the differences in techniques applied to create the figures. As I mentioned before, various modes of production can be broadly and carefully linked to different time horizons, where pecked and ground petroglyphs would mark an earlier phase, and engraved figures the later one. But how should we interpret differences between giraffes and oryxes, if we assume that most of them are contemporaneous? Out of 319 giraffe depictions, 150 were pecked, and if we further exclude all the specimens which could be attributed to the late or post-prehistoric times, we find that the pecked giraffes constitute well over a half of images. Within the same time frame, two-thirds of oryxes have ground bodies, outnumbering the minority of the pecked bodies. How should we evaluate this emerging pattern? Does it have only a chronological value?

119There are also questions concerning the quantity of individual elements. For instance, is the quantitative hierarchy of giraffes (n=311), oryxes (n=221), and ostriches (n=99), reflecting any hierarchy of their cultural importance when we cannot ascertain how people in the past approached rock art? Could the quantity of these figures play any role in their comprehension? Perhaps the only numbers which were significant were associated with a quantity of petroglyphs within a panel? Out of 148 panels containing giraffes, 69 have more than one giraffe depiction. It means that 53.3% of panels contain just one such zoomorph (it can be the only petroglyph on a panel, but also can be accompanied by figures other than giraffes). The almost identical percentage (50/93=53,7%) characterises a share of panels with single oryxes, with or without additional figures. Could a perception of zoomorphic figures be different depending on the quantity of images portrayed? Were the compositions containing different species random? Or, on the contrary, could different arrangements be variously understood?

120The way giraffes, oryxes, and ostriches co-occur on panels may suggest that indeed such arrangements could be meaningful. All three animals appear together on just six panels – each containing a large composition of petroglyphs. Giraffes and ostriches co-occur on another seven surfaces. Finally, a combination of giraffes and ostriches is found on another 28 panels, whereas oryx co-occurs with ostrich only twice. These estimations are surely influenced by the exclusion of the undetermined quadrupeds, and the fact that a definition of a panel and composition is not a strict one, but such a clear prevalence of giraffe/oryx ensemble is certainly thought-provoking. Is it the state of research, a random pattern, or a meaningful design?

Epilogue

121Over 30 years of rock art research conducted by the Petroglyph Unit have resulted in acquisition of a large body of data. However, we are still far from understanding these petroglyphs, even if their chronology slowly but progressively is being improved and defined. All the questions asked above, and many more left unstated, must be addressed in future studies if we want to more fully comprehend the rock art of the Dakhleh Oasis and the communities that produced the rock art.

122One future topic, omitted in this paper, concerns the micro-regional differences in the patterns of rock art distribution. We clearly see that certain depictions, such as those of horses and camels, can be associated almost exclusively with the eastern parts of Dakhleh. This kind of analysis may prove to be rewarding, especially if petroglyphic sites are considered in a broader archaeological context.

123Another topic which will be addressed in a future study concerns human-animal relationships. Anthropomorphic figures co-occur with zoomorphs frequently enough in the rock art of the region, not to be ignored. For this reason, a full discussion of animal petroglyphs co-occurring with human depictions will be the natural next step to accompany the research topics covered here on the animals portrayed in Dakhleh Oasis rock art.

I would like to express my gratitude to Carolynne Merrell for improving the English in the manuscript, and all her valuable comments. My sincere thanks go to both anonymous reviewers whose remarks and suggestions certainly helped me to improve the paper. This work is part of the author’s research project number 2016/23/D/HS3/00805, financed by the Polish National Science Centre.