No title

Bi

blical Arca

Perspectiveson the AncientWorldfromMesopotamiato the Mediterranean Vol.56 No.2

PAMPERED OR

The

PLAIN

POOCHES. PARIAHS?

AshkelonDog

Burials

June1993

Biblical Archaeolog ist

to theMediterranean on theAncientWorldfromMesopotamia Perspectives A Publication of theAmericanSchoolsof OrientalResearch

Volume56 Number2

June1993

54 About the Authors

55

Pampered Pooches or Plain Pariahs? The Ashkelon Dog Burials Paula Wapnish and Brian Hesse Did the cosmopolitan Persian period port city of Ashkelon have a "pet cemetery?" Dog bones are common at ancient sites, but the Ashkelon finds of thousands of bones and dozens of articulated skeletons, including many puppies, are altogether unique and extraordinary. Ashkelon's buried dogs are a first-class mystery whose solution demands all the forensic capabilities of zooarchaeological science. Were the dogs a special breed? Did they die of natural causes or were they sacrificed? Were the dogs pampered pets lovingly interred or urban hounds merely discarded?

page 55

81 Economics with an EntrepreneurialSpirit:

Early Bronze Tradewith Late Predynastic Egypt

Timothy P Harrison Economics link the late twentieth-century "global village" in ways we readily appreciate. Can the same be said about Egypt and southern Palestine during their earliest historical periods? Advances in archaeological study of exchange items and anthropological modeling of exchange now enable the flow of goods between the two regions to be traced. The network that emerges suggests the presence of "entrepreneurs" moving goods back and forth across the ancient landscape.

94 page 81

The Samaria Ivories, Marzeah, and Biblical Text Eleanor Ferris Beach How important was the visual dimension of ancient texts to their signification? A comparative study of the renowned Samaria ivories helps to reconstruct the lost visual context of several Hebrew biblical texts. As the symbolic backdrop of the marzeah,these carvings inform us about the meaning of this elite institution, and direct our attention to its powerful iconographic significance for the literature of the Hebrew Bible. Among other examples, the power of the visual image preserved in the ivory carving of the "woman at the window" facilitated and ironically undermined the 2 Kings account of the usurper Jehu's encounter with Queen Jezebel.

105 Arti-facts 107 Book Reviews

page 94

On the cover: the plaster-jacketeddog skeleton from the Ashkelon excavationsis flanked by an Egyptiandog portraitfrom the BeniHasan tombs (12th dynasty)and a drawingof the modernPharaohhound. Neitherrepresentationaccuratelymodelsthe Ashkelondog type that more nearlyresembledthe naturallyoccurringpariahdog.

:i:Ra -: --

i__ -

__ I

- P-::::-:-i:----~i ---:: I~:l-:-i-:':-l --:-i

i i_::;;:

i/r:~ iil :.

--:--

--

-1'

E

::

I -.-..---

:-

iii: ;;

--i

:.

i-i

the

Authors Eleanor Ferris Beach

i: i,

i--iii-~

About

---~~-:-

Ellie Beach is Assistant Professor of Religion at Gustavus Adolphus College (Minnesota), where she teaches Bible and Women's Studies. She received her Ph.D. in Old Testament from Claremont Graduate School and has excavated at Zeror (1966), Gezer (1973), and Lahav (1986-87). Her research devotes special attention to relating archaeology and iconography to textual studies.

Timothy P. Harrison

Sub-

Tim Harrison is completing his Ph.D. program at the University of Chicago. He has participated in fieldwork at Jebel Abu Thawwab, Zeiraqon, Tell Halif, and Tell ePUmeiri where he serves as field supervisor of Area D (Early Bronze Age). Presently, he is collecting settlement data on the Madaba Plains region of Central Jordan in order to study the social and economic development of the region during the Early Bronze Age.

Paula Wapnish and Brain Hesse

To place your subscription to BiblicalArchaeologist,complete

ittoScholars thisform andreturn Press,P.O.Box15399,Atlanta,GA 30333-0399. Individual orders must be prepaid by check or money-orderdrawnon a United

Statesbankor by VISA or

Forfasterservicewith MasterCard. or VISA, call(404)727MasterCard 2345. Non-USsubscribersadd $5 for postage. F $35 USindividuals F $45 USinstitutions El $40 non-USindividuals i non-USinstitutions $50 l Checkor money-orderenclosed R MasterCard VISA CARD NUMBER

EXPIRATIONDATE

SIGNATURE . NAME

"

: :

ADDRESS _.

.

COUNTRY

_

i _...

.

_

_

.,

54

BiblicalArchaeologist56:2 (1993)

Drs. Wapnish and Hesse are responsible, jointly and individually, for hundreds of publications on pastoral systems, animal domestication, and faunal analysis: reports, reviews, major papers, academic and popular presentations, photographs, and a field-defining book (Animal Bone Archeology:FromObjectivesto Analysis. Taraxacum:Washington DC, 1985). They have excavated around the world, including Turkey, Iran, Lebanon, Israel, Chile, and Alabama, and provided faunal analysis at such sites as Tel Miqne-Ekron, Ashkelon, Tel Batash-Timna, and Qasile. Both received their Ph.D's at Columbia University. Brian holds the position of Professor in the Department of Anthropology at the University of Alabama at Birmingham, is Director of the University's International Studies Program, and serves as Research Associate with the Smithsonian Museum. Paula lectures at UAB in the Department of History and is Research Associate with the UAB Department of Anthropology, the Smithsonian Institution, as well as Harvard's Semitic Museum. Their daughter, Arielle, will soon turn five.

From

the

Editor

It demanded great skill, patience, and creativity to read Ellie Beach's study on the Samaria ivories in last September's issue of BiblicalArchaeologist.An error in BA's former production system jumbled the pagination of the article and turned it into a papyriologist's dream (or nightmare). Professor Beach's treatment of these famous carvings and their relation to biblical texts is happily re-presented in this issue, with apologies to both the author and the readership. The task of re-constructing Beach's text presented a rough parallel to her own challenge to re-construct the lost iconographic or visual context of ancient texts. Beach supplies the rhetoric of the Samaria ivories through a comparative analysis of a constellation of motifs shared by various caches of Near Eastern ivory carvings. This rhetoric focuses on life-death transitions among the elite. It provides commentary on the biblical marzeah,where ivory studded couches were employed, thereby making sense of certain textual allusions. Moreover, the iconography of the carved motifs furnishes the visual backdrop of another group of texts, affording uncommon access into the imagination of the ancient audience. Entering into the enigmatic imagination of the ancient world also constitutes the heart of Wapnish and Hesse's zooarchaeological research. They address the remarkable and baffling preponderance of dog burials excavated in Persian Period strata of Ashkelon. A commanding assemblage of archaeological, historical, literary, and, above all, animal-bone data is meticulously marshaled. The canine skeletons are re-constructed, and the dogs themselves are once again given visual embodiment with the help of comparative analysis. Osteological data portray the animals' lives and deaths. The re-construction points tellingly to the act of interment itself as the interpretive key. It is precisely at this point, where behavioral motivation comes to the fore, that archaeologists confront what culture historian Robert Darnton refers to as "the unfathomable strangeness of life among the dead" (The Kiss of Lamourette:Reflectionsin CulturalHistory. New York:W. W. Norton & Co., 1990, xiv). What did the urbanites of Ashkelon hope to gain from burying their city's dead dogs? Gain and loss are prominent in Harrison's thinking as he searches for a meaningful re-construction of the patterns of exchange between Egypt and southern Palestine in the Early Bronze Age. The quantities and distributions of traceable articles chart an exchange network and suggest the motivations of Early Bronze Age actors. The "traders" may have belonged to an ancient variety of entrepreneur. Again, the archaeological task constitutes the re-construction of intangible patterns and notions that have long since vanished. The archaeological task-whether it be determining the nature of regional interaction or dog-related behavior or the iconographic function of ivory carvings-always demands more than restoring the correct order of a mis-paginated text. It requires an enlightened and enlightening entry into the minds of our ancient forebears.

David C. Hopkins Editor

Biblical

Archaeologist

on theAncientWorldfrom Perspectives Mesopotamia to theMediterranean EditorDavid C. Hopkins ArtDirectorLyle Rosbotham Book Review Editor James C. Moyer

EditorialAssistant Timothy L. Adamson EditorialCommittee Gerald L. Mattingly JefferyA. Blakely Ernest S. Frerichs Carol L. Meyers Gaetano Palumbo Seymour Gitin Kenneth G. Hoglund Neil A. Silberman Thomas E. Levy Mark S. Smith Gloria London Paul Zimansky Subscriptions Annual subscription rates are $35 for individuals and $45 for institutions. There is a special annual rate of $28 for those over 65, physically challenged, or unemployed. BiblicalArchaeologistis also available as part of the benefits of some ASOR membership categories. Postage for Canadian and other international addresses is an additional $5. Payments should be sent to ASOR Member-

ship/SubscriberServices,P.O.Box15399, Atlanta,GA 30333-0399(ph:404-727-2345; Bitnet:SCHOLARS@EMORYUI). VISA/Mastercardorderscan be phoned in. Backissues Backissues can be obtainedby callingSPCustomerServicesat 800-437-6692 or writingSP CustomerServices,P.O.Box 6996,Alpharetta,GA 30239-6996. PostmasterSend addresschangesto Biblical ASORMembership/Subscriber Archaeologist, Services,P.O.Box 15399,Atlanta,GA 303330399.Second-classpostagepaid at Atlanta,GA and additionaloffices. Copyright? 1993by the AmericanSchools of OrientalResearch. CorrespondenceAll editorialcorrespondence should be addressedto BiblicalArchaeologist, 4500MassachusettsAvenueNW,Washington, DC 20016-5690(ph:202-885-8699; fax:202-8858605).Booksfor review should be sent to Dr. JamesC. Moyer,Departmentof Religious Studies,SouthwestMissouriStateUniversity, 901 SouthNational,Box 167,Springfield,MO 65804-0095. AdvertisingCorrespondenceshould be addressedto SarahFoster,ScholarsPress,P.O. Box 15399,Atlanta,GA 30333-0399(ph:404727-2325;fax:404-727-2348).Ads for the sale of antiquitieswill not be accepted. BiblicalArchaeologist(ISSN 0006-0895) is published quarterly (March, June, September, December) by Scholars Press, 819 Houston Mill Road NE, Atlanta, GA 30329, for the American Schools of Oriental Research (ASOR), 3301 North Charles Street, Baltimore, MD 21218.

L4OFlg

o'

rny/'3m Ca

Pooches Pampered Pariahs? Plain Ashkelon Dog The or

Burials

By Paula Wapnish and Brian Hesse n 1985whentheLeonLevyExpe- bones were found in other parts of

dition renewed excavations at Ashkelon, a large port on the southern coastal plain of modern Israel, the huge quantity of faunal material recovered included many remains of small mammals, animals smaller than sheep/goats but larger than rats or mice. Among this sample were 21 bones of immature canids (for unfamiliar terminology check the glossary on page 76), including articulations (2 to 4 bones each) from 4 individual pups, all excavated in the same area of the site, the soon to be infamous Grid 50. Because the bones were immature, it was not clear whether they were from domestic dogs (Canisfamiliaris), the most likely possibility, or from one or more of the three wild canids of the southern Levant, the wolf (Canis lupus), the fox (Vulpessp.), or the jackal (Canis aureus). Immature morphologies, where the size and shape of bone elements are growing and changing into the adult form, make it difficult to distinguish the remains of closely related species. While the sheer size of an immature bone sometimes can be used to distinguish medium sized dogs from the much larger wolf or the much smaller fox (in terms of size, the jackal is too close to call), some bones were from such young individuals-so small and unformed-that it was impossible to do more than assign them to the canid family. During that first season's excavations, 11 bones of adult dog also were recovered from Grid 50. The partial skeletons of three adult dogs and another 147 individual adult dog

skeletons. Animals consumed on site rarely remained intact after slaughter. They were butchered into halves or quarters and then reduced to smaller cuts depending on consumer preferences.2 In this way the carcass was dispersed throughout the site. After the meat was eaten, bones were often discarded around households or other places of consumption. If not scavenged by roaming animals or otherwise disturbed, these bones then were incorporated into the deposit near the site of disposal. Larger settlements frequently had neighborhood or city-wide dumps that served as the accumulators of refuse. Bones of a once live animal might end up in the same dump area

the site, an impressive representation of canines for one year's efforts. Several questions were apparent. The first issue was the abundance. In our experience, wild canid bones are rare at large historic sites. At Ashkelon, for 5 51 7 62 example, seven years of excavation have produced in 12 65 6 6 146 excess of a million animal bones, but 15l/ less than 30 specimens of fox, a lone wolf possibility, 25 26 27 2 24 S23and not a single jackal. Domestic 2V 30 35 32 dogs are much 33/ more common at ancient cities than -\ their wild cousins, yet their numbers are but a fraction of the typical bone collection. The pro-NJ 51 52 53 54 56 1155 portion of canid/dog bones in the first season's faunal sample simply was much higher than expected. The articulations of immature canids from Grid 50 were also a loom intriguing. It is not unusual to encounter the bones of non-food Topographic plan of Ashkelon shows the three grids which animalsI as multihave produced dog burials.These tinted areas (38, 50, 57) have produced hundreds of dog burials deposited over a relatively ple bone articulathin slice of time. tions or partial

19 17

20 2)

0 ;00

BiblicalArchaeologist56:2 (1993)

55

through many routes of disposal but they would no longer have any anatomical relationship to each other. On the other hand, non-food animals, because they are not butchered, often remain intact. Microfauna and smaller mammals, such as dogs and cats, may be left where they die, or a malodorous carcass may be moved to a street or alley, empty field or dump. If the carcass is covered with dirt there is less likelihood of it being scavenged and a greater possibility of the skeleton remaining intact. Otherwise, accumulations of dirt and debris will eventually bury the carcass but the slower pace of this process will expose the bones to greater disturbance. This pattern holds for large mammals as well, except that carcasses are more likely to be moved because of the stench. After burial, complete skeletons undergo postdepositional processes (including excavation) that often result in the destruction/disappearance or repositioning of some body parts (Hesse and Wapnish 1985). Therefore, partial skeletons or less extensive articulations are all that remain. While not common, there is nothing unique in uncovering bone articulations from an individual donkey or calf in excavation. It is very common, almost expected, to find a few bones or even a limb of a dog or a cat. So what was puzzling about the clusters of canid puppy bones from Grid 50 was not their occurrence as partial articulations, but their agemost were but a few weeks old. Under normal circumstances, immature bones of domestic canids are rare at archaeological sites, those of wild canids almost unknown. The paucity of immature canid bones is explained by the animal's behavior and the attritional processes affecting the carcass. Zoologist Thomas J. Daniels, who has studied free ranging urban dogs, notes that he seldom found dog carcasses "if the animals have initiated movement away from the den site" (pers. comm., May 1988). Any number of bird and mammal scavengers eat exposed dog car56

BiblicalArchaeologist56:2 (1993)

X:

.............

X:

Ii~iii

M,

.. . . .....

:.0

The city of Ashkelon lies next to the sea, bounded by a semi-circularfortification system.

The 150-acresite possessesan occupationalhistorystretchingfromthe fourthmillenium BCE throughthe thirteenthcenturyCE. Photo courtesyof RichardCleave.

casses. If the dead Ashkelon puppies were exposed for even a brief time, it is very unlikely that any part of the skeleton, let alone articulations, would have survived to burial. When a bitch whelps she chooses a concealed location for her litter. It is inconceivable that a wild canid mother unaccustomed to and wary of people would choose a city rather than familiar surroundings in which to whelp. It is equally difficult to see the puppies as naturally occurring deposits of domestic dog bones. Once born, the young remain at the den until they are weaned, which begins about week 5, and attain independence by 3.5 to 4 months of age. This is well past the age of the puppy bones in question. So how did several bones from (not less than) four very young pups who were not litter mates (based on the bones and their locations) end up buried in what was then an open expanse in what we now call Grid 50? Clearly, some unusual human/animal behavior, depositional process, or both were responsible for this phenomenon.

We didn't have to wait long for some answers. In 1986, excavators in Grid 50 recovered many partially articulated and complete skeletons of unquestionable domestic dogs.3 The excitement of finding dog burials was contagious and for the next few summers volunteers and staff competed in inventing ingenious or outrageous explanations for their presence. Had we but known, we might have settled for the uncertainties of that first field season. By the end of the 1992 excavation, 1238 dog finds were logged (more on counting the finds below) and our earlier questions had mutated into intractable problems that eight seasons of digging have yet to resolve.

The Dog Burials Location and Date The renewed excavations at Ashkelon are concentrated on the northern and southern mounds that comprise much of the western part of the site (on the excavation project, see Stager 1991a,b,c). The dog burials discovered were all on the southern mound.

:i~::-: ::: d

: 1

~IP'::--i~s~eP~

~

i:

saP~-_i_,: _:-:?~R~-~'i~6_~~:_::-~:~i:-1: .-.---:::

.-iC-:::::: o: :~::::: i -:::::::::i::_ :_;-::: :::r:~:::::: -:;::::

..

::::::::_1::::

imp%::

::::::-::: :::::: :~"i~:: i'i-i":-:-:i:~:-::::-~:i~i::::::,::::::: *::?::i:-:i~i-:::: :-;::: ":::::?:.;:?::::-::::::

;::::::::::li :::i:

6~-~a

::: ':~:: :-: c:;.. ~E- I ~:: ----::~i- ': ~':x-'~~;iii-is-i-;:ii-:l__l:iiiii:i ~:::::: I::-::

::: ...::

:.=:::-: :: :::~-:-:: :~::: :-:::a:::::: :-:

Partially excavated dog burials are exposed in Grid 50.

The excavation of over 1200 burials has demanded a huge

Thedogs were interredin fillsoverlayinga largewarehouse.The burialswere distributedrandomlythroughoutthe area.

expenditureof energy,leavingbehindthe excitementof the discoveryand creatingseeminglyintractableinterpretiveproblems.

Photos by the author unless otherwise attributed.

Photoby CarlAndrews,courtesyof the LeonLevyExcavations at Ashkelon.

Many were removed in plaster jackets. Grid 50 is situated there, at the southwestern side of the city along the coast. Of the 1238 dog finds noted to date, 970 of them were excavated here from Persian period (538-322 BCE)fills overlaying a large warehouse. The original warehouse of six 30 by 60 foot storerooms was stepped down to the sea on its western side. After an initial phase of use, the terracing was filled with a series of debris-rich layers to level it (approximately) with the eastern half of the building. Dog burials were concentrated in the thick matrix of the western side of the warehouse, although some dogs were recovered from a thinner series of fill sediments to the

east. Because of the seaward erosion on the southwestern portion of the mound, we will never know how far to the west dogs were buried. Excavations during the summer of 1992 in the northwest part of Grid 50 produced fragmentary elements of large scale architecture. Exactly how, or if, these remains relate to the dog inhumations is under study. Directly south of Grid 50 lies Grid 57. About 500 BCE,the first phase of a large building complex was begun here, but before the next phase of building could be initiated, the area was leveled and briefly used for dog burials. Fifty-eight dog finds were excavated from this phase of Grid 57, all corresponding to the dog-burying

phase of Grid 50. Grid 38 in the northeastern sector of the southern mound was also a locus of dog burials, with 181 recovered to date. Numerous subphases comprised the nearly 10 feet of Persian period deposits that contained a complex series of large buildings. Most of the dogs in Grid 38 were buried in streets or thoroughfares between the buildings. In Grid 50 the burials were dug into deposits well dated by ceramic associations to the fifth century BCE.The deposits were capped by architectural remains that establish a terminal date possibly as late as the first part of the fourth century. A similar time frame applies to Grid 57. In Grid 38 dog burial began BiblicalArchaeologist56:2 (1993)

57

later in the Persian period and continued throughout the occupation into the first years of the Hellenistic period (332 BCE). Although the intervening areas between Grid 38 and the two coastal grids have not been excavated, it is reasonable to offer several hypotheses. No clear demarcations of the limits of any of the dog concentrations have been found. While Stager (1991b) refers to the dog burials in Grids 50 and 57 as a cemetery, evidence that a sacred area was set aside for dog inhumation is not compelling. The matrix into which pits were dug accumulated throughout the period of interment. Dogs were buried where there was space, rather than a space being prepared to receive dogs. If we accept this view, then the burials in Grid 38, while somewhat later in date, are part of the same phenomenon. The difference between the two concentrations may be more apparent than real, being conditioned by the architectural nature of the spaces available for burial.

Pattern of Burial In general, each dog burial seems to have been a discrete event. This conclusion is warranted by several observations. The more complete skeletons were found singly, each in its own unlined, shallow pit. The top surfaces of the pits were of varying heights within the sediment matrix, an indication that dogs were interred sporadically. No pattern was discernible in the orientation of the pits, nor in the placement of the corpse within its pit, with respect to the compass or other interments. Each corpse was carefully placed in its grave. There were no skewed limbs, arched back heads, or other skeletal distortions that characterize animals that were just pitched into a convenient hole. This was made especially clear by comparison to a dog recovered from a drain of Hellenistic date in a deposit just above the burial layer in Grid 50. It was both twisted and missing most of its smaller skele58

BiblicalArchaeologist56:2 (1993)

i::iz~i~i::::i:lJ

i~iiliii',ii~iii~liiiiiiii0

OEM

.

~

Plaster jackets enabled the removal and preservation of about 75 skeletons. Jacketing

is expensivebut advantageouswhen time is at a premimumor the skeletonis especially fragile.Completestudyrequiresthe dogs to be re-excavatedfromtheir moderncocoons.

tal elements. The attitude in death was totally different that of the earlier animals. The dogs were buried on their sides with tails carefully arranged to curl toward the feet, sometimes reaching between the lower hindlimbs. In a few cases the feet were so entwined that they may have been bound at the ankles before burial. The limbs were flexed to different degrees. In Grids 50 and 57 the legs were sufficiently extended from the torso to suggest a natural repose. But in Grid 38 the intact burials appeared cramped, with legs drawn up close to the body and the skeleton straining against the edges of a pit just large enough to contain it. This was undoubtedly because the burials were confined to narrow streets. Overall, the careful interment of adult and immature dogs appears just a little less careful in Grid 38. No act of burial was accompanied associations. The pit material by matrices, which were often softer and darker than the surrounding sediment, contained small amounts of pottery, animal bones, flints and

debris from metal working (Grid 50), but none were distinguished enough to deserve the label 'grave good'. This fact is consistent with two additional salient features: 1) No markers were found that would have signalled the location of a pit; 2) Burials were frequently dug into or superim-

.4

Ok. ii.

W. 02& Twopuppies belongingto the categoryof Several Bone Finds.These bones~vere dis-

articulatedbut found in close proximity. Slightlydifferentsizes permittedthem to be segregated.

..

:

oil

In

:$ : -: -:

:

:::

- : -?:

:-

q

-:i -: ;

:.

: - :_ :- j: AN

S I

:i-: ;:

:

iioli two

X

s:

:

cannot prove this theory at present, it should be kept in mind as the search to explain all the features of the burials continues.

:

x:

OUTeef -: ::i:::i:-':an : ~: : :

: : : _,:i~: : : : :

_~_ii-i-ii~:--i--i--i~i i-iiii:i. W:":: ::::::::j::::::'~:: :me:j~: :-?:

:-

_;:

W:7: _: 1

-a

-

IMFME

A :::::::::::t 51. .. .. . . ::-::::::::i~~~:a . . . .. .. . . ::-:: ::::I ..::::j:::-

gii[i'i::::'::::~: :;:::: ::b di

.....

..................... xx:::n:::

... ?i ....

A puppy (top) and adult skelton displayed after removal,

cleaning,and consolidation.Thisprocesssometimesenables disassociatedfindsto be recombined.

posed on top of each other. It is tempting to see these data in a cause and effect relationship, but which way around? Were burials dug into each other because the graves were unmarked, or may we surmise that the graves were unmarked because burials were dug into each other? Since such a chicken-and-egg puzzle will lead nowhere, a different theory is called for. Might the absence of grave goods, markers, and burial disturbances be the result of a more comprehensive motivation? If we take the view that only the act of burial mattered, the position of each grave would not have been important, hence no markers. Moreover, there would no longer be any significance attached to the corpse, rendering the disturbance of previous burials unimportant. This would also account for the lack of any material deposits with the burials. While we

The Ashkelon Dogs Demography Numbers Generating an estimate of the total number of dogs buried is complicated by the fact that so many of the pits were disturbed by subsequent burials. To keep a handle on the numbers involved we defined three types of "dog find": Complete

Dogs,PartialSkeletons,and Several

BoneFinds, this last when only a few disarticulated bones or teeth are found in close proximity.4 Only a dog find from a complete or partial burial can be equated with an individual animal. The many Several Bone Finds may have come from a disturbed partial skeleton, or an animal whose remains were so scattered that it was recorded as more than one dog find. Therefore, the actual number of (excavated) individual dogs will be smaller than the total number of recorded dog finds. The table on page 61 shows the distribution by grid of the raw counts of the three types of dog finds in each of three age categories. Information about the three main concentrations of dogs is supplemented with a tabulation of the additional canid material recovered from the site. The graph on page 61 presents these abundances on a logarithmic scale to illustrate visually the general similarities of the distributions, from the three grids. By

adding the complete and partial skeletons together we come up with a minimum of at least 436 individual dogs, though we suspect that about 600-700 animals are actually represented. This larger figure we believe will emerge when location, relative body size, and stage of osteological development are factored in, thus permitting us to combine some of the dog finds. Sex Dogs of both sexes were buried. Males are identified by the presence of a baculum (os penis), females by its absence in complete burials. While many of the burials were disturbed and so indeterminate, there is no reason to suspect selection for one sex or another.

Age at Death We divided the dogs into three age groups: puppies (0-6 mos.), subadults (6-18 mos.), and adults, based on tooth eruption and wear and the degree of long bone ossification (see discussion in sidebar on skeletal ageing). The most striking characteristic of the burials is the very large number of puppies. The bar charts on page 63 illustrate the percentage distribution of the three age groups by dog find in each of the three grids. A difference emerges between the burials from the seaside grids and those in Grid 38. The percentage of puppies comprising the total number of dog finds in Grids 50 and 57 is 58% and 55%, respectively. By contrast, puppies made up 82% of the sample in Grid 38. This substantive difference may be partly explained by the distribution of finds in the area. Most of the puppies in Grid 38 were recovered as Several Bone Finds, which implies that considerably more disturbance affected the remains, a process that elevated the count of puppies artificially. This may have occurred in antiquity but, more likely, is a result of the failure to recognize puppies during excavation. Counting only the complete and partial skeletons reduces the differences between the grids. The Chi-Square statistic computed for this comparison (1.091, BiblicalArchaeologist56:2 (1993)

59

2df) indicated that there was no reason to suspect that the variability arose from more than chance. Puppies constitute almost two thirds of the sample from the three concentrations (62%).Adults make up 33%, and subadults 5%.5Several problems complicate the easy assumption that this mortality pattern was representative of all the dogs at Ashkelon. For example, if some dogs were buried and some dogs were not (for whatever reason), our data would be skewed. Further distortions would result if certain burial areas had been selected for particular types or ages of dogs. Since we already know that a large portion of the burial area was lost to erosion, how confident are we that the areas excavated accurately reflect burials in all the areas used in antiquity? Fortunately, we are not totally without some controls. We will never be absolutely sure that the buried dogs reflect population mortality, but every time we find Persian occupational debris, we find dog burials rather than the usual scattered dog remains. This at least gives us the impression that an awful lot of the dogs at the site in that time period did get buried. How do we know that dogs were not differentially selected? We can't be sure of that either. But we do know that all ages are represented and that the distribution of mortality is consistent with that experienced by a population of unmanaged urban dogs. Thus, there is no reason to suggest that dogs were selected by age for burial. Finally, the distribution of the ages of the dog finds are similar between the grids (the one anomaly explained by collection bias). Coupled with the large number of burials that have been uncovered and the size of the extent of the exposures in the three grids, we believe that what we have collected fairly represents the dog-burial activities of the Persian period. Cause of Death No skeleton shows evidence of a trauma extensive enough to have killed the animal. There is no evi-

60

BiblicalArchaeologist56:2 (1993)

The Characterof the Accumulation Paleontologists and zooarchaeologists make a distinction between two kinds of faunal accumulation -catastrophic and attritional. Better understood as poles on a continuum, the two categories distinguish two patterns of mortality. In

of the affectedpopulationare equally susceptibleto the cause of death. Therefore,the demographic patternof a catastrophicaccumulation mirrorsthe relativeabundance of the variousage and sex categoriesin the living population that was affected. On the otherhand, attritional

deposit are thought to have died in a brief episode. Death is presumed to have resulted from a single cause or a closely allied set of causes. As an example, the bison kill sites that dot the American

accumulations result from multiple causes operating over longer

a catastrophicaccumulation,all the animalsfound togetherin a

GreatPlainsrepresentthe resultof single very successful hunts. Since bison run in age and sex defined herds at various times of the year, in some cases, entire nursery herds (bison cows and their calves) or whole bachelor herds were exterminated, while in others the entire population was slaughtered. Exceptionally virulent diseases can also produce a catastrophic accumulation. One suggestion that has been made for the Ashkelon dogs is that they were the victims of an epidemic (Smith 1991). This proposition can be evaluated quite precisely since a key characteristic of catastrophic accumulations is that all members

dence that carcasses were cut up or skinned prior to burial. One radius from a non-Persian context is cut, and it is the solitary cut dog bone in the collection. Pathologies and diseased bones are present in perhaps 5% of the adult and subadult specimens (see photographs on page 64). These include damaged paws, knitted breaks, dislocations, parasite infections, and dental anomalies.6 None of these is severe enough to cause death. This does not mean that injuries were never severe. One dog found in Grid 50 in 1992 had a broken and healed first cervical vertebra

periods of time. All membersof the affectedpopulationare not equally susceptibleto the various causes of death. The best example is the kitchen midden of an agro-

pastoralcommunity.Here the animals which are representedare the specificselections/culls of pastoralistsreactingto the needs of pastoralmanagementand the opportunitiesof the marketplace. Usually this would mean that young males and old females predominate in the accumulation, since prime adults would have been spared to increase the productivity of the herd. Stated more generally, the demographic pattern of an attritional sample reflects the differential vulnerability of each age and sex class in the affected population to the mortality processes that kill their members.

(atlas). The damage to the bone is consistent with the dog having been struck sharply just behind the skull. Some teeth show extreme wear, but this is doubtless because in a beach environment like Ashkelon's, sand was mixed into most of what the dogs ate (see photograph on page 65). Late Egyptian cat mummies show evidence of strangulation in broken cervical vertebrae. This has not been noted on any Ashkelon skeleton. Poisoning would, of course, leave no skeletal traces. Death by strangulation or drowning might leave microscopic traces of blood on

Dog find counts presented graphically. The Y-axisis scaled logarithmicallyto show the similarityin the distributions in the three areas.

produce samples with demographic parameters matching those of living populations, attritional samples do not. The Ashkelon age distribution with its abundance of puppies and absence of subadult dogs is clearly attritional. This conclusion is reinforced by the character of the archaeological burials discussed above, which clearly accumulated intermittently. In a study of free-ranging urban dogs, Beck established that approximately 50% of the animals died in their first year. He emphasized that this was a very conservative estimate of mortality because "dead dogs disappear in a few days if not collected, either destroyed by traffic or by natural decay (microorganisms and insects)" (Beck 1973:37). Young adults are least vulnerable, while adult animals die at nearly twice the rate of young adults, but not nearly as high as the rate for puppies. This is a close match to the mortality pattern in the Ashkelon collection, particularly considering the complete dogs. However, the archaeological sample has a slightly higher percentage of puppies. We believe that this difference reflects the difficulty in observing dead puppies in a free-ranging environment, because of predator scavenging. What does the abundance of puppies mean? Ashkelon's residents either must have been managing the dogs in some way, or they were exceptionally alert for dead dogs. The approximate match to urban freeranging dog mortality further suggests that the causes of death for the Ashkelon dogs were many. It is not unreasonable to suggest that in this pre-veterinary period the abundance of pups is related to the host of diseases which afflicted young canids.

the teeth but this possibility has not yet been pursued. However, the distribution of mortality and the available skeletal evidence give us no reason now to believe that the dogs were killed.

What Did the Dogs Look Like? Bones allow us to describe several aspects of appearance: height, weight, limb proportions, and the general shape of the skull. These descriptions can be combined to give a picture of the animal: heavy or

Distribution of Dog Finds Grid 50 Puppy Several Partial Complete Total Subadult Several Partial Complete Total Adult Several Partial Complete

Grid 57

Other

Grid 38U

%

No.

%

No.

%

No.

%

317 146 102 565

51 75 66 58

17 8 7 32

50 73 54 55

100 35 14 149

85 85 61 82

5 0 0 5

17 0 0 17

39 7 8 54

6 4 5 6

0 1 0 1

0 9 0 2

1 1 2 4

1 2 9 2

1 0 0 1

3 0 0 3

266 40

43 21

17 2

50 18

16 5

14 12

23 0

80 0

45

29

6

46

7

30

0

0

No.

23 80 43 16 25 28 351 36 Total 29 181 58 970 Grand Total (1238) Rawcounts are provided (No.) as well as the percentage (%) within an age category (puppy,subadult, adult) for each find category (several, partial, complete). Information is provided for the three main areas (Grids50. 57, and 38U) of burialconcentration as well as for other areas and periods of the site. Determiningthe total number of dogs representedby the finds demands carefulstudy of the location, size, and age of the bone finds.

he-R

IM. L&

210 a,

E z

Several Partial Complete Several Partial Complete Several Partial Complete

Puppy

Subadult

Adult

Because the Ashkelon dog accumulation can be dated to a more or less thin slice of time, an important question is whether the dogs died from attritional or catastrophic causes (see Sidebar). Catastrophic agents

BiblicalArchaeologist56:2 (1993)

61

of

half

right

intercentrum

body

1

of

atlas

The atlas or topmost vertebra of the neck

is composedof three bonyparts.These fuse duringpuppyhoodand permitan estimationof age. Thetop drawingshows a beagle pupvertebra80 daysfromconception.By106 days,the elementsare fusing(bottom). Illustrationis based on Evans,Watson,and de Lahunta(1986).

light, cursorial (adapted for running) or non-cursorial, round or elongate head. However, important features like coat color, texture and markings, ear shape and carriage, tail length and carriage, true muzzle length, etc., are unknowable from skeletal materials. It is these characteristics, together with disposition, that figure prominently in breed definitions. Even more tenuous are links that many authors would like to make between morphology and use as hunting, herding, war, or lap dogs. Therefore our picture of the Ashkelon dogs is very incomplete both visually and functionally. Height and Weight The height of an adult dog can be estimated from the length of the limb bones (Harcourt 1974; Koudelka 1885 provides an older standard). The weight can be calculated from a measurement on the lower mandible (jaw) (Wing 1978; see also Hamblin 1984). Reliable height and weight estimates cannot be made on immature specimens. Only 15 completely 62

BiblicalArchaeologist56:2 (1993)

studied adult dogs are well enough 80 to make preserved these calculations. This small number reported here results not only 840 from the elimination of immature E 0 20 Z 40 but specimens because many adult jaws were crushed beyond reconstruction and a large number of 80 adult skeletons were removed in 60 -o plaster jackets and await study (see c table and graph on pages 66-7). The distribution of the Ashkelon information is presented together with some comparative material. One Persian Period dog E from Tel Batash is included. A number of dogs and one possible wolf o 40 in the collections of Hebrew University z 20 were measured. While most are male individuals (Buck, Fang, Stripe, and Bernard), one sev. Par.com. Total Sev.Par.Com.Total Sev.Par.Com.Total female (Batsheba) Subadult Adult Puppy Subadult Adult Puppy was available. These animals are Age distributions of the dogs found in Grids 50, 57, and 38. related and derive Puppiesare lessthan 6 monthsold, subadults6-18 months,and from Suez. In addiadults18 monthsand older.Moststrikingis the largenumber of puppies,mostof whichdied as veryyoung animals.Quantition, a dog deties are shown as percentagesof each "dogfind"category. scribed as a "saluki" could be measured. It came than would be found in the breed from a Bedouin group in the Sinai. standards of modem dog types. The Ashkelon dogs range in weight from fact that the Persian Period dogs are about 11 to 20 kgs., and vary from 48 to 61 cms. in withers height (shoulsomewhat heavier in build than the der). This range exceeds that of the comparative specimens relates to the for material save the latter's origins in a more arid envicomparative wolf?? and saluki. This is a signifironment (see below). It was possible to compare the cant amount of variation, far more

Ageingthe DogSkeletons Mammalian skeletons are aged by an eruption sequence for the deciduous and permanent dentition (see Arnall 1961 and Williams and Evans 1978 specifically for the dog) and by an ossification and fusion sequence for the bony skeleton. Ossification refers to the progressive replacement of soft tissue by hard in the skeleton. It begins in utero in the dog. Extensive radiographic and dissection studies have identified the ossification centers for each bone element and the pattern of bony development (Evans and Christensen 1979; Hare 1959, 1961). For example, the degree of ossification in the vertebrae can be used (Evans 1958, 1962, 1974; Evans and Christensen 1979; Watson, Evans and de Lahunta 1986). The point at which the arch is fused to the centrum is an easy stage to recognize. With long bones, the second phase of maturation is the stage at which the diaphysis or shaft portion of the bone fuses permanently to the epiphysis, the articular end. Extensive anatomical research has established the ages at which fusions are complete (Chapman 1965; Evans and Christensen 1979; Seoudi 1948; Silver 1969; SumnerSmith 1966). It remains difficult to assign

ages to individual specimens or even to complete skeletons, because of the discrepancies between the numerous fusion sequences. There can be anywhere from a few weeks to a six month difference cited among various authors for the age of fusion of a particular element. Various factors are responsible for the discrepancies, such as which diagnostic technique was used, the type of dog, and the level of nutrition. In addition, all the puppy studies note developmental differences inter- and intralitter in bony growth, so there are no absolutes correlating age and stage of development. In establishing our three age categories we tried to keep the various sequences in perspective to come up with approximate divisions. Our puppy category varies in age from just born to about six months. Based on the ossification of the vertebrae none seem to have been fetuses, but some may have been still-born term puppies. Subadult refers to an animal older than about six months up to a year and a half, whose longbones are still incompletely fused. Finally, adult refers to animals older than a year and a half whose permanent teeth are in wear, and whose longbones are all fused.

Early Iron Age materials from Isin together with an Early Bronze age dog from Tell Brak. The Syria-Anatolia-Iran group includes a variety of archaeological sites in the highland areas to the north of Mesopotamia together with a modem specimen from "Constantinople" described by Lortet and Gaillard (1903). Also tabulated is a comparative series of dogs identified as "saluki" or "greyhound" as well as one possible wolf. These larger canids together with the Suez dogs provide a useful cline of development in both size and robustness. Batsheba can be taken to represent what Clutton-Brock (1979) would refer to as a "founder population," a naturally occurring population which serves as the basis for further selective breeding. The two salukis can be used to represent two historical points in the development of a characteristic of the breed desired by dog fanciers in this century, namely, increased height. As can be seen, there is considerable overlap between the geographic scatters, though there is a slight tendency for dogs at more northerly sites to be taller than the material from the eastern Mediterranean. Several of the comparative archaeological specimens fall outside the Ashkelon range-Korucutepe, Takht-i Suleiman, and Pergamum. This is not surprising given the general biological principles relating environment to morphology. Both salukis, the greyhound, and the wolf?? also are taller than the Ashkelon dogs.

Skeletal Robustness

average height measurements to a number of other archaeological and modem dogs.7 This comparison is presented graphically on page 66. On these charts the arithmetic means calculated from the heights of dogs from sites and comparative samples are arranged in four geographic groupings-Egypt, Eastern Mediterranean, Mesopotamia, and Syria-AnatoliaIran. The Egyptian sample includes a series of modem dogs from Khartoum together with the skeletons of mummified dogs. The authors of

these descriptions (Gaillard and Daressy 1905, Lortet and Gaillard 1903) recognized four types of dog from Egyptian mummies-a pariah type, a greyhound, the Egyptian dog, and a spitz type. While, as we discuss below, these distinctions are unfounded, they are retained here to better represent the variability of dogs in this region. The Mediterranean group includes sites in Israel together with the Hebrew University collection. The Mesopotamian sample includes Old Babylonian and

A further indication of how the animals looked is gained from estimates of skeletal robustness. We calculated this statistic i0r all the limbs based on the ratio of greatest length to least shaft diameter. The ratios for each limb were normalized and then combined into an estimate of overall robustness. The results are illustrated by the scatter graphs on page 68. The sites are arranged as before, though the Egyptian material was not published in a manner that makes this analysis possible. A general trend is BiblicalArchaeologist56:2 (1993)

63

again present. The dogs in the Eastern Mediterranean and Mesopotamian regions tend to be more robust than their cousins to the north. The Hebrew .... University saluki is by far the most . ......... lightly built individual. Rearranging the site and X~ gg comparative sam....... ples by height reveals a strong Pathologies in the Ashkelon dogs. Bones displayed above and pattern: short top rightshow paw and limbpathologies.Thevertebral stature correlates pathologymanifestedby the bones on the middlerightwas with robust build. broughton by Spirocercalupi(diagnosisby RobertD. Powers, What is most Thisconditionusuallyafflictsdogs DVM,AuburnUniversity). that are in poorcondition.Noneof these caseswould have important in the been severeenough to havedirectlyresultedin the death of comparisons is that the dog. The bottom photo detailsa moresevereinjury:a brothe collection from ken and healedcervicalvertebra. Ashkelon is in no its for way atypical region and that the build of the dogs is also consistent specialized for running (Miller 1976). That is, in the forelimb the humerus with the overall morphological patbecomes long with respect to the terning of Near eastern domestic that the There is no scapula and the radius/ulna with dogs. suggestion Ashkelon dogs diverge in basic body respect to the humerus. In the hind what is from limb, the tibia lengthens with repect shape expected given to the femur. The limb ratios of the their environmental location. Ashkelon dogs could be calculated in It is also instructive to compare 12 cases (see graph on page 69). Only the Ashkelon sample to another one collection from dog falls in the gazehound group extremely large southeastern Europe. B6k6nyi (1984), (dogs that hunt by sight rather than to smell such as greyhounds and was able following Hornberger, Roman of several salukis), the others lying in the most types distinguish cursorial of based on estimates part of the range for shepdogs partly herd dogs. The Ashkelon average robustness. The ancient Near Eastern falls in the middle of two weakly material overlaps almost completely defined clusters in the archaeological with B6k6nyi's type 4, the most comand comparative material. The mon form, which he calls "dogs of medium size used for hunting and of Egyptian material has relatively elonabout 50 to 58 cm withers height" gated forelimbs compared to the hindlimbs. None of the Ashkelon (1984:72). By contrast the Mesopodogs and tamian seen the extremes of Syrian-Anatolian-Iranian approaches any in the variability of this Roman Tic samples show the reverse pattern. While we have only one archaeologiGorsium sample. cal sample (Ashkelon) for the Eastern LimbProportions Relative limb proportions can also be Mediterranean, comparative material for the region from Hebrew Univerused to estimate the degree of cursorial adaptation. The relative length of sity overlaps both groups, reinforcing the impression of an intermediate the limb bones increases from trunk character for our collection. to toes as the animal becomes more

al

...

-............:~: ~iii

64

BiblicalArchaeologist56:2 (1993)

ii

.........

Head Shape The shape of the dog skull is traditionally classified into three broad types, brachy-, dolicho-, and mesaticephalic (Evans and Christensen 1979): a relatively short, broad head such as the rounded skull of a golden retriever, a long, narrow head like that of the gazehounds, and a shape between the two. Unfortunately, most of the skulls in the Ashkelon

collection were crushed by grave pressure and could not be reconstructed well enough for accurate measurement. Of the few that were measurable, only one skull even approaches the width to length ratio associated with the gazehound group. Instead, the Ashkelon morphology fits within the narrow end of the middle group which includes dogs such as the shepherd. This was reinforced by the examination of skulls in situ where we had the best approximation of what they looked like. None showed an elongated nasal region or extreme narrowness of the crown. Summarizing all this osteometric information, the Ashkelon dogs appear to be a variable group similar to most other ancient Near Eastern dogs and modern dogs of no particular ancestry. They do differ from those animals identified as salukis and greyhounds. No evidence of selective breeding in the direction of cursorial adaptation is present. This is reinforced by an examination of the hip joint which shows no strong adaptation to a gazehound pelvic form.8

Representations of Dogs in Ancient Sources To what may we compare the Ashkelon dog population? This question is not as ingenuous as it seems, because from the minute it was known that there was something special about the dog burials, suggestions have swirled around the issue fueled by gleanings from ancient texts and artistic representations. How do they aid our understanding of the phenomenon at hand? Open any modern book on dog breeds and it will tell you that the Saluki is reputedly one of the oldest breeds in the world. The Saluki, variously known as the Persian Greyhound, Persian royal hunting dog, gazelle hound and other folk names which give purported clues to its ancestry, is thought to have originated as a dog specialized in the hunting of hares and gazelles. Depic-

tions of Saluki and greyhound type dogs are common in ancient Egyptian reliefs and tomb paintings of the second millennium

........:

i-

BCE.More impor-

tant for our purposes is the dog i!Li!!iili i•• !ii,! •il •!•!! i•il :••ii••i •ii•: .......... iii : ... .... ... ....... from Tell Brak in Iraq, dated even earlier than many Egyptian sources 2580-2455 BCE(calibrated). CluttonMandible with well-worn teeth. The sandy environment of Brock notes that Ashkelondoubtlessresultedin especiallygrittyforage for the the alliance of this dogs in the city. dog "...with the greyhound type not only provides osteological evikept pure.... It is significant that definite types of dogs are depicted over dence for one of the foundation and over again in the Tombs not only breeds of domestic dogs but also for of one period, but in those separated its distribution outside ancient by great numbers of years. These we Egypt..." (1989:219). So here were can, therefore, consider native tantalizing links-an animal of noble breeds" (Ash 1927, reissue 1972:49). ancestry, somehow associated with the Persians, and well known from Egyptian art is reputedly some of the ancient Egypt-a good place to start. best in the ancient world with respect to the accuracy of animal representaThis explains why we paid partictions. But do these images represent ular attention to greyhound mor"breeds"? phology in generating our skeletal Modem breeds are defined as picture of the Ashkelon dogs. much by outward characteristics that Saluki and greyhound type dogs are not registered skeletally, such as are not the only kinds represented in ear shape and coat color and texture, Egyptian art. A recent book on dog as by variations in height, weight, breeds names four similar modemrn and skull shape. Furthermore, the breeds-the Greyhound, Saluki, skeletal distinctions presented as eviPharaoh hound, and Ibizan hounddence for breeds are not in themthat it says not only appear in Egyptselves definitive. Constance Miller, ian reliefs and painting, but that origfor many years a dog show judge inated there as well (Caras 1985). In and well known gaze- or greyhound fact what is truly remarkable is the authority, notes that these modern great diversity of types to be found. In the tombs at Beni Hasan (XIIth dogs have long jaws with no surplus skin at the oral area, moderate to no dynasty, 2000-1800 BCE),dogs of 'stop' of skull, long and low held many sizes, with various head tails, and slender but chase-worthy shapes, coat color, tail length and carframe (pers. comm.). These features, riage, limb proportions, chest depth, and neck length, are depicted. Beni she points out, are extremely similar to the wolf forms of Asia Minor and Hasan is only the best known of any in fact are adaptive characters suited number of examples of this diversity. to the arid, open steppes of the One early authority expressed a widely held belief when he wrote: region. So greyhound type features "In Egypt dogs were very much as represented in Egyptian art are as likely to be natural morphological they are to-day and doubtlessly were BiblicalArchaeologist56:2 (1993)

65

adaptations as they are the result of human selection. Which brings us to the over-arching point concerning breeds, namely, that they are as much human social phenomena as they are biological expressions. Without a sustained human network to establish standards, cull deviants, and so forth, a breed cannot exist. Artistic evidence of different dog types may be evaluated on a number of grounds short of assigning the animals breed status. With no evidence of the artist's intent, there is no a priori reason to assume that the images are meant to represent populationsof similar appearing dogs. They could just as well stand for a single animal. If that's the case, then the images don't represent some early dog breed manual, but a random selection from the whole range of dogs present in the region. Further, there is a whole suite of problems (which cannot be dealt with here) involved in how accurately any artist renders a living animal. A combination of features might portray an animal type generally, but breeds are not general types, having a host of very specific characteristics and proportions,9 which we would do well to doubt were repeatedly captured by ancient artists. In any case, the key point is that breeds deal with populations, not individuals. Not until classical times is there some textual evidence for the maintenance of dog breeds. Hull (1974) claims that the Greeks carefully bred hounds for hunting based on the

20

12

48

50

52

54

56 58 60 62 Harcourt Height (cm)

64

66

68

Estimates of the heights and weights of the Ashkelon adult dogs. The animals were

quitevariable.Dataare found in the table at right.

65

E

5

AUCons

45

Geographical Groupings

probably represents no more than "down-the-line" breeding, or breeding like to like, which ancient peoples had been practicing for thousands of years (with varying intensity). There is no evidence for established standards, culling to standard, studbooks, pedigree records or any of the other essentials of breed maintenance. Roman authors also stressed pure bloodlines, but again, all the particulars necessary for true breeds 66

BiblicalArchaeologist56:2 (1993)

Comparison of the shoulder height of the Ashkelon dogs with other Near Eastern

and comparativespecimens.Thedogs are arrangedin four geographicalgroupings Mesopotamia,Syria/Anabeginning(on the left)with Egypt,the EasternMediterranean, tolia/Iran,and a comparativegroup.The heightsof the dogs show considerableoverlap amongthe geographicalgroupingswith some tendencyfor tallerdogs at morenortherly sites.Dataand referencesarefound in the table on page 77. are not unequivocally present in the written material.10 Boessneck (1988:Abb. 90b) has

illustrated bones from Elephantine dated to the Vth Dynasty that came from a relatively short dog. However,

Height and Weight of Ashkelon and Other Dogs Dog Ashkelon 86-2

Harcourt Height(cm)

Wing Weight(kg)

50.3

14.9

Sex/ Comment

10.5 male 49.0 86-3 54.5 13.0 86-5 54.6 18.2 86-6 17.0 57.0 male 86-7 11.6 50.9 87-1 11.6 87-2 48.3 55.7 15.4 male 87-3 54.1 12.9 87-26 48.9 male 10.3 88-5 19.4 56.9 88-28 draindog, early Hellenistic? 11.6 Islamic 57584 54.5 91-17 51.9 19.3 14.1 49.8 91-18 19.6 male 92-222 60.8 15.7 48.5 Persian Period Batash 51.4 12.7 female Batsheba M4053 16.9 male 68.3 Saluki M3973 57.8 13.2 male Stripe M3975 58.2 15.7 male Fang M3977 16.2 Buck M3974 59.2 male male 54.0 14.5 Bernard M3972 21.2 Wolf?? M4066 62.5 Heights and weights of Ashkelon and other dogs based on estimators of Harcourtand Wing. A contemporary PersianPeriod dog from Batash is induded as well as a series of modern specimens in the collections of Hebrew University.All except the Salukiand the Wolf?? are 'pariah'dogs.

the calculated stature for that animal is 40 cm, not far below the range in the Ashkelon sample, and certainly not as extreme as the dachshund type dog from a Beni Hasan Tomb (XIIth Dynasty) which Boessneck illustrates with the bones (Abb. 91). The best evidence for distinct dog types based on skeletal remains comes from the Roman period. For example, B6k6nyi's (1984) study of Roman dog remains from Tac Gorsium in Pannonia shows enormous differences in height and limb shape, indicating that the dog population was comprised of individuals whose aggregate variations fell outside an expected adaptive range. Even if the types he discerns were not recognized by the Romans who raised the dogs, they were forms with particular characteristics which could have been manipulated in breeding practice. This perspective legitimizes our search of the skeletal remains for morphological trends in the develop-

ment of physical types in a manner that is not based on breed standards, strict definitions that make methodological demands which we are unable to approximate with archaeological evidence. The truth is that the only modem breeds "that can be told from skeletons are the really aberrantones. Normal dog-breeds are characterized by surfacephenomena that make animals lookto be far more different than their skeletons confirm" (Constance Miller, pers. comm.). However, skeletal demonstration of distinctive types either far removed from or outside of natural adaptive ranges, coupled with written references to human behavior suggesting selection, can reasonably be construed as the introduction of different stock or human manipulation of indigenous domesticates. Artistic evidence can reinforce these data but cannot by itself establish the validity of ancient physical animal types.

Another approach to understanding the variability of the Ashkelon dogs is to consider modem dogs in the region. Fortunately, we have the resource of two dedicated scholars, Rudolph and Rudolphina Menzel, who produced descriptive studies of unmanaged dogs in Mandate Palestine (1948, 1960). They applied the term "pariah" to these forms, using a label that had a long history of application to canids in the broad belt from Morocco to India (c.f., Studer 1901). They recognized four forms of naturally occurring types of medium sized animals varying in body and head from a sheep dog to a greyhound type. These forms show a considerable degree of variation in head form, muzzle shape, and ear carriage, ranging from Type 1, shepherd-like dogs, Type 2, dingo-like dogs, Type 3, collie-like dogs, to Type 4, gazehound-like dogs. Variability of coat color and quality, tail carriage, and stature, can be seen in the figures that accompany their article (some of which are reproduced on page 74). This foundation stock has been selectively bred in recent decades to create a breed, the Canaan dog, which is recognized internationally by some kennel societies. Considering the variability of the pariah dog skeleton implied by the Menzels' study, we believe that the Ashkelon dogs would fit comfortably within this group. The Ashkelon sample is best viewed as a naturally occurring canid population with physical characteristics adapted to the hot and semi-arid conditions of the southern Levant. This is reinforced by the slight cline in stature and robustness seen in archaeological dogs throughout the Middle East. More Ancient Dogs A further source of comparison is to consider burials of dogs at other archaeological sites in the Middle East. There is considerable evidence for dog burial throughout the region, but we hasten to add that our review should in no way be considered exhaustive. With the exception of BiblicalArchaeologist56:2 (1993)

67

Egypt, at no site were as many dogs buried as at Ashkelon. The most efficient way to evaluate the comparative burials is to organize the data by time, space, and context, and then to consider the various features. Closest in these particulars are the burials at Ashdod. The site of Ashdod, located a few miles north of Ashkelon just off the coast, was also one of five major Philistine cities. The city declined precipitously after conquest by the Neo-Babylonians at the beginning of the 6th century. Five burials of mature dogs are dated to Persian levels (strata 5-4) in an area given over to industrial activities, mostly metalworking, after the Babylonian conquest. The animals are lying on their sides, usually one dog to a pit (Dothan and Porath 1982). No measurements or anatomical details are given, and the published photograph does not permit any morphological assessment. In an earlier Ashdod report (Dothan 1971), N. Haas reported that a refuse dump of Hellenistic Stratum III yielded many animal bones mostly in anatomical relation. These included the remains of one wild boar, many large cattle, two felids, and nine adult and three immature canids; no details of the skeletons are given. During the summer of 1992, excavations by the Department of Antiquities of the State of Israel uncovered a series of dog burials in a Persian Period site located near Ben Gurion airport. Details of interment appear to be similar to the Ashkelon examples. While no osteological specifics about these finds are yet available, they do enlarge the regional pattern. Unfortunately, our information about dog burials from other sites in Israel is not more revealing. In the Persian Stratum (V) at Tell el-Hesi, a collection of dog remain found in a single pit could have come from a single juvenile individual (Bennett, Jr. and Schwartz 1989). No cranial elements were found. An articulated dog skeleton was uncovered in a mud-brick silo that had been con-

68

BiblicalArchaeologist56:2 (1993)

shoBraU So

8r

0.6

0.4

02

HKELONO J wU.AO

Isin-G

&A 0o

stam k

inWl

-00.2OBahig o -0.8 t

r

se

l#

HbeH