Acknowledgements

We wish to acknowledge the help and support of the Department of Antiquities of Jordan. Funding for excavations (HM) was generously provided by Mu’tah University. Funding for zooarchaeological analyses (CM) was provided by the Stanford Archaeology Center. We also thank Jérôme Poulalier for his help with the French translation.

2The diverse forms of animal management strategies used in LPPNB communities east of the Jordan Valley are slowly being assembled with the continuing publication of zooarchaeological and stable isotopic analyses of faunal assemblages recovered from the region (Becker 1991, 2004; Wasse 2002; Driesch et al. 2004; Twiss 2007; Makarewicz 2009, 2013b, 2016). Here, we further investigate animal use during the LPPNB through zooarchaeological analysis of faunal remains from the site of es-Sifiya located in the Wadi Mujib, southern Jordan. So far, our understanding of LPPNB animal use is largely derived from faunal assemblages recovered from settlements located in either the Jordanian highlands, such as ‘Ain Ghazal, ‘Ain Jammam, and Basta, or semi-arid mountainous landscapes including Ba’ja and Tel Tif’dan. The es-Sifiya dataset provides much-needed additional information on animal exploitation and the shape of caprine husbandry systems in wadi environments where many LPPNB sites were situated during this crucial period of increased food production and settlement expansion.

Animal husbandry east of the Jordan Valley

3Domesticated sheep and goat herding was the primary means of animal-based subsistence supporting large LPPNB settlements located east of the Jordan Valley (Makarewicz 2013b). Small-bodied sheep appeared suddenly in the middle Euphrates region in the mid-8^th millennium BC, indicating their importation from the north where intensive exploitation of sheep had been already underway for over a millennium (Legge and Rowley-Conwy 2000; Saña Segui 2000), and were present in the Damascus Basin by the mid-8^th millennium BC (Helmer and Gourichon 2008). Sheep were likely first incorporated into LPPNB subsistence systems in the Jordanian highlands by ca. 7300 cal. BC (Makarewicz 2013b), although they may have been introduced to the Balqa highlands somewhat earlier during the mid-8^th millennium BC, inferred by the presence of a few sheep bone specimens identified in the late Middle PPNB layers of ‘Ain Ghazal (Driesch and Wodtke 1997; Wasse 2002). It remains to be seen, however, if these trace specimens represent the initial translocation of sheep to the region or are younger intrusive specimens from overlying LPPNB deposits (Horwitz and Ducos 1998). Direct dating of the bone specimens in question and, in the case of sheep and goats which are difficult to distinguish in some skeletal elements on the basis of morphology, collagen peptide fingerprinting are needed to securely identify the temporal and species assignments of each bone specimen.

4For goats, small herds were loosely managed in the Upper Tigris by the late 9^th millennium BC (Hongo et al. 2005), increasingly exploited at roughly contemporaneous sites located in the upper Euphrates, and first appeared in the middle Euphrates basin during the early 8^th millennium BC (Saña and Tornero 2008). Morphologically wild goats appear to have spread by the mid-late 9^th millennium BC to the Damascus Basin and, later by the late 9^th-early 8^th millennium BC, were intensively managed, indicated by young male culling and horncore torsion associated with domestication (Helmer and Gourichon 2008). Further to the south in regions adjacent to the Jordan Valley, morphologically wild goats were under management during the mid-9^th millennium BC, indicated by an intensification in the exploitation of goats at the expense of gazelles (Horwitz 2003; Sapir-Hen et al. 2016), the harvest of juvenile animals (Sapir-Hen et al. 2016), and the augmentation of goat diets by human caretakers (Makarewicz and Tuross 2012). It remains to be seen, however, if these managed animals were the progenitors of domesticated goats that were exploited at LPPNB settlements east of the Jordan Valley during the late 8^th millennium BC.

The late Pre-Pottery Neolithic B site of es-Sifiya

5Es-Sifiya is a large LPPNB settlement situated on the broad flood plain of the lower Wadi Mujib (ca. 200 m a.s.l.; fig. 1). Perennial water flow in the wadi is supported by an expansive catchment zone that includes numerous smaller wadi systems, fed by precipitation run-off, which flows into the Wadi Mujib (El-Naqa 1993). Today, es-Sifiya sits within an Irano-Turanian phytogeographic zone and receives ca. 200 mm of rainfall per annum between November and March. Warmer and wetter conditions prevailed across the Southern Levant during the Early Holocene (Bar-Matthews et al. 2003), including the lower reaches of the Dead Sea drainage where es-Sifiya is located (Rambeau 2010), with rainfall levels sufficient for dry farming.

6Excavations at es-Sifiya were undertaken in 1994-2000 after the discovery of the site during construction of the Mujib dam (Mahasneh 1997). Surface survey and excavations indicate that the settlement spreads over ca. 10 ha and shares architectural features with other LPPNB settlements located in southern Jordan including Basta, Ba’ja, Khirbet Hammam, Hamarash, and el-Hemmeh. Buildings at es-Sifiya are comprised of substantial, agglutinative, two-story stone architecture that likely served as domestic dwellings and are characteristic for LPPNB settlements in central and southern Jordan. Interior spaces consisted of a large room connected to small, ca. 1-2m² rooms by small openings framed by thresholds and lintels (Mahasneh 2004). These smaller rooms likely functioned as storage rooms. Many room interior floors and walls of the rooms were lime plastered and, occasionally, this plaster was painted red. Numerous large stone querns, stone bowls, pestles, and unbaked clay vessels found in situ on the plaster floor of a large room (Square 11, Area B) indicate intensive food processing was regularly carried out in larger interior spaces. Six radiocarbon determinations obtained from unidentified wood charcoal recovered during excavations of Area C yielded a summed distribution of 7060-6800 cal. BC (Görsdorf 2000) and suggest occupation at es-Sifiya took place during the latter half of the LPPNB. Areas A and B may encompass contemporaneous or earlier LPPNB occupation, but additional radiocarbon dates are needed to better establish the intra-site chronology.

Methods

7Faunal remains were recovered through handpicking and also occasional sieving of sediments through a 5-mm mesh. Taxonomic frequencies of macrofauna recovered by hand-picking are comparable to frequencies generated from sieved assemblages (Sapir-Hen et al. 2017), although particular skeletal elements belonging to medium-sized mammals may be underrepresented, especially under conditions of high fragmentation, in the absence of sieving (Payne 1972). Bone specimens belonging to small mammals, micro-fauna, and young and infant-aged medium-sized animals are also not well recovered in un-sieved assemblages and are underrepresented in abundance measures (Payne 1972). Recent work at LPPNB el-Hemmeh, where all excavated deposits were systematically sieved through a 2-mm sieve mesh, indicates the regular recovery of small, fragile bones of infant-aged ungulates, small mammals, and other microfauna (Makarewicz 2013b). This is in sharp contrast to other LPPNB assemblages recovered through handpicking (e.g., Ba’ja and Basta) where bone and tooth specimens representing these taxonomic groups were notably absent. These findings strongly suggest that young, infant-aged animals and small mammals are underrepresented in the es-Sifiya assemblage. Bone specimens were identified with the assistance of the modern comparative collection located in the Council for British Research in the Levant in Amman. Sheep and goat bone specimens were identified according to morphological criteria defined by Zeder and Lapham (2010) and also by Becker (2004) as defined during analyses of sheep and goats from LPPNB Basta. Decisions regarding the timing of livestock slaughter are influenced by short-and long-term production goals for ante-mortem products of milk, fibre, and traction as well as post-mortem products of meat, fat, and skins (Payne 1973; Redding 1981; Vigne and Helmer 2007). Fodder availability and pasture access also structure the timing of animal harvest. Here, separate survivorship curves indicating the relative frequency of animals surviving increasingly older age categories were generated from the state of epiphyseal fusion and also from the eruption and wear of mandibular molars. Sheep and goat survivorship curves calculated according to the state of epiphyseal fusion were based on animal age at the fusion of the epiphyses of different appendicular elements according to Zeder (2006).

8Mortality profiles computed using mandibular tooth wear provide higher resolution information on animal age at death, particularly for animals older than 3 years when epiphyseal fusion in the caprine appendicular skeleton is largely completed. Sheep and goat tooth wear and estimated ages for each mandibular wear stage (MWS) were recorded according to Payne (1973). The general paucity of complete hemi-mandibles in the es-Sifiya assemblage necessitated the use of loose mandibular molar teeth to generate mortality profiles. While expanding the sample size, this approach either underestimates or over-estimates assignments to a specific mandibular wear stage group. Single teeth such as the dp/4 and third molar (M/3) will provide more specific ageing information than individual wear stages of M/1 and M/2’s which cover a wider range of wear stages. In cases where a single molar tooth could be assigned to multiple wear stages, that specimen was equally divided between those wear stages (“straight allocation”). For cattle, mandibular wear stages for isolated teeth were recorded according to Grant (1982), and dental age classes were defined according to Legge (1992).

9The relative proportion of male and female animals in livestock herds also influences harvesting decisions. For example, the intensive slaughter of juvenile males reflects subsistence decisions concerned with optimal production of meat, while extended survivorship of females is strongly associated with dairy production and progeny (Payne 1973; Vigne and Helmer 2007). For sexually dimorphic taxa like caprines and cattle, biometrical data effectively distinguish bone specimens likely belonging to male and female animals (Zeder 2001). Measurements were taken from appendicular skeletal elements according to A. von den Driesch (1976). Metrical values from any one skeletal part were relatively few in number and, consequently, biometrical data were transformed into Log Size Index (LSI) values calculated by comparing log-transformed measurements from archaeological specimens against those from a wild standard animal (Meadow 1999). Sus metrical data were compared to a female wild boar from Turkey (Hongo and Meadow 2000). For goats, the standard animal is represented by the averaged measurements of one male and one female Capra aegagrus from the Taurus mountains and, for sheep, a female Ovis orientalis from Iran (Uerpmann and Uerpmann 1994). LSI distributions computed for sheep and goat are based on breadth and depth measurements, which are influenced by animal weight (Uerpmann and Uerpmann 1994), and based on the proximal and distal humerus, radius, femur, tibia, and metapodials. First phalanges were excluded from LSI distribution calculations because larger posterior phalanges are difficult to distinguish from smaller anterior phalanges. In addition, greatest breadth (GB) measurements of the calcaneum were not included due to difficulty in obtaining replicable measurements in archaeological specimens. Finally, the GLP measurement of the scapula was also excluded from the body size distribution due to an unusual skew toward very large LSI values seen in the es-Sifiya goat bone specimens. Mixture analysis was applied to the fused distal humerus in order to estimate the contribution of different populations to the overall distribution (Monchot et al. 2005). Mixture analysis was computed using PAST software (version 4.02; Hammer et al. 2001).

Results: Relative abundance of taxa

10A total of 3,736 faunal specimens were recovered from excavated deposits at es-Sifiya (table 1). Sheep and goats were the most commonly encountered taxa comprising ca. 82% of all mammalian bone and tooth specimens identified to genus. Goats were the most intensively exploited animal, representing ca. 30% of specimens identified to genus-level, while sheep were less important, their bones and teeth representing only 11% of the assemblage. Gazelle bone specimens were also present, representing ca. 5% of the assemblage, as well as specimens belonging to Bos sp. (3%) and Sus scrofa (2%).

11A variety of other herbivorous and also carnivorous taxa were also found in the es-Sifiya assemblage. Equids, represented by scapulae fragments, one ulna, one astragalus, and numerous broken molar teeth, were occasionally hunted. The wild ass (Equus africanus) and the onager (E. hemionus) were exploited at low frequencies at other LPPNB settlements including ‘Ain Ghazal, Ba’ja, and Basta (Becker 1991; Driesch and Wodtke 1997; Driesch et al. 2004), but it was not possible to determine the species hunted at es-Sifiya in the absence of metrical data from appendicular elements and morphological information from poorly preserved molars. The metacarpal of a roe deer (Capreolus capreolus) was also identified. The glenoid cavity of a scapula, tuber calcaneum, distal humerus, and a radius were tentatively identified as belonging to fallow deer (Damadama), but this remains a very cautious determination in the absence of a modern comparative specimen available at the CBRL. Fallow deer have been identified at Middle PPNB Yiftahel and Kfar HaHoresh, both located in the Mediterranean woodlands of the Galilee region, as well as at MPPNB Abu Ghosh located in the Judean Hills. East of the Jordan Valley, bone specimens attributed to fallow deer have been recovered from LPPNB Basta (Becker 1991) and a tentatively identified specimen from LPPNB Wadi Shu’eib (Makarewicz 2016). Fox (Vulpes sp.) bone specimens including mandibles, pelvis, and fore- and hindlimbs were moderately abundant (ca. 5%) in the es-Sifiya assemblage. Ribs and vertebrae identified as “small mammal” likely belong to foxes. A femur belonging to a badger (Meles sp.) is also represented in the es-Sifiya assemblage, as are three fused right distal humerii, a fused distal tibia and a fused proximal tibia belonging to Canis sp., and a third phalanx belonging to a small felid. A single tortoise carapace was recovered and some unidentifiable bird shaft fragments. Chelipeds of freshwater crabs were also identified.

Capra sp. and Ovis sp.

12Caprines are the most abundant taxa represented at es-Sifiya, representing ca. 82% of LPPNB bone specimens identified as Ovis/Capra, Capra sp., or Ovis sp. The predominance of caprines at es-Sifiya is similar to the high frequencies of sheep and goats observed for other contemporaneous LPPNB settlements located east of the Jordan Valley (Makarewicz 2013b). Goat bone specimens are more abundant compared to sheep at a ratio of 2.8 to 1.

13Goats exploited at es-Sifiya included domestic goat (C. hircus), bezoar (C. aegagrus), and ibex (C. nubiana). Horn cores exhibiting torsion typical for domesticates (n=22) and scimitar-shaped horn cores (n=6) lacking an anterior keel but displaying a convex surface, morphological features typical for male ibex, were also identified. A large number (n=26) of scimitar-shaped horn cores displaying an anterior keel and quadrilateral cross-section distinctive for male bezoar goats were also present in the faunal assemblage, as were four horncores belonging to female bezoar. Bezoar are also more abundant than ibex at MPPNB Beidha, located ca. 150 km to the south in similarly rugged mountainous terrain, where only ca. 15% of horn cores were attributed to ibex and the remainder to bezoar goat (Hecker 1982). Distinguishing between C. hircus, C. aegagrus, and C. nubiana in the appendicular skeleton on the basis of morphological characteristics is difficult. The appendicular skeleton of the Nubian ibex displays wide variation in numerous morphological characteristics that range from strongly goat-like to strongly sheep-like, even within the same skeletal element (Tchernov 1993: 24; Driesch and Wodtke 1997).

14The distribution of LSI values for fused goat specimens indicates wide variation in body size, reflecting both the exploitation of female domesticates, male domesticates, and wild goat species (fig. 2). Mixture analysis of breadth measurements taken from the fused distal humerus (Bd) produced the best model (based on minimum Akaike Information Criterion score) set to two groups (population 1, 31.5±2.2, 56.4%; population 2, 38.2±2.1, 46.4%). Mixture analysis results for three groups are also shown here (population 1, mean = 31.7±57.8%; population 2, mean = 37.5±0.8, 25.3%; population 3, mean = 40.1±1.7, 16.8%; fig. 3). Male bezoar exhibits larger body size than female bezoar, while male ibex falls within the same size range as female bezoar (Harrison 1968), and female bezoar within the range of male domesticates (Zeder 2001). The distributions generated for both LSI values and distal humerus measurements display similar patterns, with the rightmost peak likely representing large male C. aegagrus and the middle group possibly representing a mix of female bezoar, ibex, and male domesticates. The group encompassing the smallest specimens most reasonably represents mostly small-bodied domesticated females and likely some domesticated males. The distribution of LSI values for sheep skeletal parts is unimodal (fig. 4) and indicates either the predominance of female animals, reduced sexual dimorphism in sheep compared to goats, or possibly, the inclusion of metrical values from ibex bone specimens misidentified as Ovis sp. Animal age may also reduce sexual dimorphism in particular skeletal elements (Popkin et al. 2012). At es-Sifiya, survivorship curves computed according to the timing of epiphyseal fusion for goats, sheep, and a combined sheep/goat category indicate the use of different harvesting strategies for various production goals (table 2, fig. 5). For a combined sheep/goat category, very young animals younger than 6 months in age were occasionally slaughtered, indicated by the presence of unfused proximal radii (n=7). Yearlings were rarely slaughtered, with 93% of caprines surviving beyond 1 year in age. A sharp increase in kill-off then took place with approximately 66% of young juveniles surviving beyond 12-18 months. Survivorship of young adult animals remained high with ca. 60% of surviving beyond 18-30 months in age. The pronounced kill-off of animals between 30-48 months in age with ca. 34% of animals surviving this age class is consistent with Type B milk exploitation where females with decreased reproductive potential and milk production are slaughtered. Older surviving animals may represent older domesticates retained for meat “on the hoof” or hunted bezoar or ibex. Species-specific survivorship curves based on epiphyseal fusion data indicate broadly similar harvesting strategies were applied to younger sheep and goats but may have diverged somewhat for older adult animals. A slightly higher proportion of goats survive beyond 18-30 months in age compared to sheep in same age cohort, but there is pronounced survivorship of sheep older than 30-48 months compared to older goats which are intensively culled.

15Mortality profiles generated for caprines calculated according to mandibular wear stage (fig. 5) are broadly consistent with survivorship curves computed using epiphyseal fusion data while providing more fine-grained information on animal age at slaughter, particularly for animals older than 4 years in age. The mortality profile generated using only teeth that could be confidently assigned to a single mandibular wear stage (n=45; fig. 6a) indicates younger kill-off compared to the mortality profile computed according to straight allocation (n=119; fig. 6b). Both methods indicate lambs and kids were rarely slaughtered and caprines between 6-12 months in age (age class C) harvested. The single mandibular wear stage profile indicates a sharp decline in animal kill-off with moderate kill-off of animals aged between 1-2 years, while the straight allocation profile indicates moderate harvesting of animals between 1-2 and 2-3 years in age reflecting the high number of M/1’s exhibiting wear stage 8 associated with MWS stage D/E according to Payne (1973). The slaughter of very old animals aged between 6-8 years is observed in both profiles, albeit with a more intensive harvest of this age cohort as well as animals aged between 4-6 years in the straight allocation profile. Referring to the single wear stage mortality profile, the inhabitants of es-Sifiya engaged in meat production indicated by the pronounced slaughter of older lambs between 6-12 months (meat type A meat [tender]; Helmer and Vigne 2004), along with some slaughter of yearlings. The straight allocation profile describes older kill-off including the slaughter of yearlings and young adults between 2-3 years in age, a harvesting strategy more consistent with optimised meat production. The slaughter of animals older than 6 years in age visible in both profiles suggests longer-term storage of sheep and goats live on the hoof or possibly hunting of older goats.

16Evidence for milk exploitation at es-Sifiya from mortality profiles generated from tooth wear is ambiguous. The type B milk model developed by Helmer and Vigne (2004) calls for the slaughter of some young lambs between 2-6 months (stage B), intensive kill-off of older lambs between 6-12 months in age (stage C), and the slaughter of older females (stage EF) when their milk yields or lamb production decreases (Helmer and Vigne 2004). The mortality profile generated for es-Sifiya only partially fits these criteria. The general absence of evidence for the harvesting of young lambs at es-Sifiya may reflect decisions to retain young animals, or the pattern is an analytical artefact of recovery methods that selected against the recovery of young animal teeth. The pronounced slaughter of older lambs at es-Sifiya may be associated with dairy production or, as discussed above, intentional slaughter with the aim to produce tender meat. Older females between 2-4 years were consistently culled at es-Sifiya in line with the type B milk model, but it is unclear if the removal of dams was due to decreased milk yields or lamb production hypothesised for this age classes (Helmer and Vigne 2004).

Bos sp.

17At es-Sifiya, Bos sp. comprise ca. 3% of bone specimens identified to genus. The low abundance of Bos is comparable to other LPPNB sites in the region including Basta, Ba’ja, and Khirbet Hammam (Becker 1991; Driesch and Wodtke 1997; Peterson 2004). The few epiphyseal fusion and tooth wear data available for Bos from es-Sifiya indicate that juvenile and adult animals were slaughtered. The presence of a fused distal humerus, a fused calcaneum, a fused distal metatarsal, and two first and two second fused phalanges indicates at least some animals survived beyond 3 years in age. Slaughter of younger animals is indicated by an unfused first phalanx belonging to an animal younger than 1.5 years in age and an unfused calcaneum, which fuses at ca. 3 years in age. Mandibular teeth exhibiting recordable tooth wear stages included a first molar erupting from the mandible which takes place at ca. 6 months in age. Two first molars each exhibiting wear stage “k” and a second molar at wear stage “g” also indicate the slaughter of young adults between 26-36 months in age for meat.

18Only a small number of metrical determinations are available for bovines and, consequently, it remains unclear if Bos was hunted or managed at es-Sifiya (table 3). Both smaller and larger-bodied animals, represented by a small sample of metrical data, were exploited at ‘Ain Ghazal and Wadi Shu’eib during the LPPNB (Makarewicz 2016). Exploitation of female animals at Basta has been argued to indicate early cattle management in southern Jordan, either as part of local developments in animal husbandry or as an established practice introduced from the north (Becker 1991), but their domesticated status remains unclear in the absence of comparative metrical data derived from aurochsen that inhabited southern Jordan. Exploitation of small-bodied Bos and the slaughter of juvenile and young adult animals, patterns consistent with husbandry, are clearly evident in the southern Levant by the 6^th millennium BC during the Pottery Neolithic (Marom and Bar-Oz 2013; Makarewicz 2016).

Gazella sp.

19Gazelle were only occasionally hunted at es-Sifiya, their remains comprising ca. 5% of bone specimens identified to genus. Gazelle were commonly exploited at low to moderate intensities at several LPPNB sites in southern Jordan. At es-Sifiya, numerous horn cores (n=22) exhibiting striated, regularly spaced parallel grooves typical from mountain gazelle (Gazella gazella) were identified representing two female and eighteen male animals. The overwhelming predominance of appendicular elements displaying fused epiphyseal ends indicates hunting of older juvenile or adult animals while mandibular tooth wear for individual teeth indicates that young gazelle between 2-7 months in age (n=5), older animals between 18-48 months (n=2), and also very old animals over 96 months in age (n=1) were hunted. The predominance of male gazelles in the es-Sifiya assemblage may reflect intentional decisions by hunters to target bachelor herds, while the presence of young animals may indicate spring-summer hunting.

Sus scrofa

20The low abundance of swine and bovines at es-Sifiya is somewhat unexpected given its location just above the floodplain of the Wadi Mujib and suggests that, unlike Wadi Shu’eib further to the north, the Wadi Mujib did not support a lush riverine environment. The very low frequency (2%) of pigs at es-Sifiya stands in sharp contrast to the high abundance of suid remains recovered from LPPNB settlements located in northern Jordan including Wadi Shu’eib (20%), Abu Suwwan (25%), and ‘Ain Ghazal (7%; Driesch and Wodtke 1997; Abuhelaleh 2011; Makarewicz 2016). Instead, the minimal exploitation of swine at es-Sifiya is similar to that seen at LPPNB settlements located in drier environments further to the south including Basta (>0.5%), Ba’ja (>0.5%), and Tel Tif’dan (>0.5%; Twiss 2007). The very limited metrical data available for es-Sifiya pigs indicate that exploited animals were, in general, larger than a female wild boar from Turkey (es-Sifiya Sus: distal humerus Bd = 45.6 mm; metatarsal IV Bp = 17.8 mm; metatarsal V BFd = 13.6 mm; Ulna BPC = 23.4 mm, 24.0 mm). A very small sample of epiphyseal fusion data (n=14) indicates 75% of animals survived beyond 18-24 months and 50% beyond 24-36 months in age (table 4), a slaughtering pattern that suggests opportunistic hunting of wild boar.

Vulpes sp.

21Foxes were regularly trapped at es-Sifya, their remains comprising 3% of bone specimens identified to genus. Similar abundances of foxes found at ‘Ain Ghazal (Driesch and Wodtke 1997), Wadi-Shu’eib (Makarewicz 2016), and Ba’ja (ca. 2%; Driesch et al. 2004) suggest that production of fur pelts was an important activity at LPPNB settlements. Several species of fox inhabit Jordan including the red fox (Vulpes vulpes), the small sand fox (Vulpes ruepellii), Blandford’s fox (Vulpes cana), and fennec fox (Vulpes zerda). These species all exhibit similar morphologies in the appendicular skeleton but are somewhat different in body size with the red fox being the largest. Fox measurements from es-Sifiya compared with values from LPPNB Ba’ja measured from specimens identified as belonging to the red fox and fennec fox suggest the red fox was the most commonly exploited species at es-Sifiya and smaller-sized foxes, likely belonging to another species, were also occasionally trapped (fig. 7). Bone specimens from entire axial and appendicular skeleton are represented, indicating carcass processing took place on site. The general absence of fox phalanges in the es-Sifiya assemblage may be caused by retrieval bias introduced by hand-picking, although it may be due to skinning practices that left foot parts attached to furs that were transported elsewhere. Fox bone was also used to make beads. A single radius exhibits a single deep incision, perpendicular to the diaphysis, around the entire shaft circumference that, if completed, would have produced a bead.

Discussion

23The high proportion of goats relative to sheep at es-Sifiya mirrors caprine husbandry systems used at other LPPNB settlements located on the highlands and in the wadi systems of southern Jordan (Makarewicz 2013b). Sheep are as important as goats only at ‘Ain Ghazal where they are represented at a ratio of 1.2 to 1 during the LPPNB (Wasse 2002). The pattern of high survivorship of sheep well into adulthood seen at es-Sifiya is also similar to that observed at other LPPNB settlements east of the Jordan Valley (Makarewicz 2009, 2016). The delayed harvest of sheep may indicate a subsistence choice designed to store meat live on the hoof or wealth accumulation. Domesticated sheep, as a novel subsistence technology introduced in large numbers only at the beginning of the LPPNB and presumably much less abundant than goats, may have been a particularly charismatic and valuable animal that help foster nascent wealth generation (Makarewicz 2013a).

24The low abundance of gazelle (ca. 5%) at es-Sifiya contrasts somewhat with other LPPNB settlements east of the Jordan Valley where gazelle hunting was a more important facet of animal-based subsistence systems. Gazelle were hunted at moderate intensities at ‘Ain Ghazal (12.8%), Wadi Shu’eib (7.4%), Basta (9%), and ‘Ain Jammam (8.7%). However, gazelle were less important at other LPPNB settlements including el-Hemmeh (5.9%) and Ba’ja (3.6%; Driesch et al. 2004; Makarewicz 2013b) and rarely hunted at Tel Tif’dan (>1%; Twiss 2009). At present, it is difficult to determine if differences in the relative importance of gazelle in LPPNB subsistence systems reflect prey choice decisions or variation in prey density associated with spatial variation in the distribution of habitats that supported gazelle populations. ‘Ain Ghazal, Basta, and ‘Ain Jammam are all located in the moister environments of the Jordanian highlands, which hold a higher carrying capacity than the drier and more sparsely vegetated landscapes in which other LPPNB sites are located.

25Despite these environmental differences, the considerable decline in the importance of gazelle over time in Pre-Pottery Neolithic hunting systems at sub-regional scales suggests anthropogenic activities disrupted gazelle herds. At ‘Ain Ghazal, gazelle declined sharply in importance from ca. 25% during the MPPNB to 13% during the LPPNB (Wasse 2002). Further to the south in the Shara’a mountains, gazelle were regularly hunted at MPPNB Beidha (Hecker 1982) but rarely targeted at nearby MPPNB Shkarat Mazied (4.1%) and LPPNB Ba’ja (3.5%; Bangsgaard Jensen 2003; Driesch et al. 2004). In the nearby foothills of the Wadi Araba, gazelles were regularly hunted at PPNA WF16 but are almost completely absent at LPPNB Tel Tif’dan (Twiss 2007). While it remains to be seen if the reduction in gazelle numbers was directly caused by intensive hunting pressure over the long term or indirectly by increased competition for graze and browse with domestic sheep and goat herds, both factors would have effectively decreased gazelle population level and encounter rates over time. Differences in gazelle frequencies between LPPNB settlements may also reflect shifts in predation pressure over time within the approximately 500-year duration LPPNB. The proliferation and expansion of these settlements along with a presumed increase in human populations would have progressively exacerbated overhunting. Additional radiocarbon determinations would help better resolve the occupation chronology of LPPNB settlements, currently established by relatively few dates measured from unidentified wood charcoal, and help provide more refined insights into the temporal dynamics of food production and hunting activities.

Conclusion

26The faunal assemblage from es-Sifiya further demonstrates the central role of intensive sheep and goat husbandry in supporting large LPPNB communities located east of the Jordan Valley. The caprine management systems at es-Sifiya focused on the production of tender meat via the slaughter of older lambs and more optimal meat production involving the moderate slaughter of yearlings accompanied by the retention of female domesticates, at least in the case of goats. This harvesting strategy involved considerable forward-planning and effectively extracted meat from sheep and goat at sustainable levels over the long term. The delayed kill-off of sheep at es-Sifiya reflects a broader pattern visible at other LPPNB settlements east of the Jordan Valley and may reflect meat storage live on the hoof, preservation of valuable and still relatively rare animals newly introduced to the region, or wealth generation. The near absence of suids and extremely low-level exploitation of bovines at es-Sifiya reflect local habitats unsuitable for these taxa (i.e., a dearth of a well-watered floodplain despite perpetual water flow in the Wadi Mujib), but the focus at es-Sifiya on sheep and goat herding along with low-level exploitation of wild goats may reflect tight cultural connections with other LPPNB settlements in southern Jordan also deeply involved with caprine husbandry and the use of wild goats.