A PrehistoricEgyptian Pottery Experiment

Andrew Jamieson and Ashten Warfe report from theWestern Desert in Egypt on an experiment reproducing

Neolithic pottery traditions discovered in the Dakhleh Oasis

DUKINC, THE 2lHl2-2l)tl3 M U : H A E O U X ; I C A L FItLI) SEASON o f t h e

I )akhieh Oasis Project an experiment was nndertaken reproducingAbove: hired pots -note the remains ot the

Neolithic pottery. Since the beginning of the Dakhleh Oasis Pro- fire pit andjeet the stndy ot the ancient pottery has formed a major research objective discolourationunder the direction ot Dr C~olin Hope. It has been reported by Hope that in surrounditn> the pit.the Oasis ceramic production appears to have been relatively continuous since

f>r,iniK-TECHNICAL No. 20 2005 53

Tlie primary aim of the

experiment u<as to prodnce

pottery that bears the

hallmarks of the Neolithic

potters, using traditional

manuJiicturiHg lechniques

and local resources.

Secondary' objectives

inclnded recording the

labour and time costs

involved in each stage of

manufacture, the

availability of resources

and the impact of

environmental conditions

on the pottery producing

process. One important

objective was to replicate

each stage of production in

keeping with the materials

used hy and the

constraints imposed on the

early poiters. Tlie

available evidence suggests

that the present

geomorphological and

emnronmental setting of

the Dakhleh Oasis is

similar to that of

cintiquity, including the

prehistoric period.

Replicated bou-l (left) and Neolithic bowl (right)

the Early Holocene times. He notes that the e.irliest pottery dates trom theEpipaleolithic Period (c. 9(H)0-S()()0 I3CE) and finds have been made whichenable a sequence to be developed, albeit an interrupted one, extending intothe Islamic Period. Itisstiggestedby Hope, theretbre, that it is possible to studythe manutacture of pottery and its use in the C asis for a time span ofat leastsoon years.

While the Pharaonic and later periods are relatively well know n, the potterytrom the prehistoric periods is less widely understood. The experimentdescribed in this report is intended to complement a detailed stuciy of thearchaeological ceramic material trom the Neolithic Period which is currentlybeing under taken by Ashten Warfe as part of a doctoral dissertation at MonashUniversit)' in the Centre tbr Archaeolog\' and Ancient History. Throughexperimental studies it is hoped to gain insights into the processes of the pastthat are not clearly visible in the archaeological record.

El'lL'ALAtOLITHIC AND NEOLITHIC POTTERY IN THE OASISThe majorit\^ ot the prehistoric pottery thus tar discovered in the Dakhleh

Oasis is essentially crude, characterised by coarse fabrics, uneven and thickwalls and rough, heavily textured surtaces. However, some tnier examples doexist. These later items are made ot tiner materials, thin and even walls andwell finished smooth surtaces. This variation is not a retlection of ditferentpotters' capabilities and it is more likely that the cruder material was designedand manutactured to serve partictilar utilitarian needs and a ditferent set of pri-orities explains the production of the tiner wares.

The Epipalaeolithic Period in the D.ikhleh Oasis has been termed theMasara C'ultural Unit; the Netiiithic Period has been divided into two units:the Bashendi Cultural Unit and the Sheikh Muftah Cultural Unit, the formerprobably being the earlier ot the two. The range of fabrics from the BashendiCultural Unit is quite di\erse, but generally the sherds contain ditfering pro-portions ot sand and fine to coarse shale. Some sherds are also straw tempered.Most vessels appear to have been coil-built. The pottery of the Sheikh MutiahUnit represents a clear tradition of handmade wares, either modelled from asingle piece ot clay in the case ot the smaller vessels, or coil-huilt in tlie case ofthe larger torms.

The local clay deposits that were exploited during prehistoric times pro-\ ided the potters with .ui iron-rich body containing a high percentage of tem-pering material, predominantly in the torm ot shale and sand. While some of

54 c>r.iii,K-TECHNICAt No. 20 2005

these clays may have been treated to remove the larger inclusions, all ot thetabrics distingtiished among the Sheikh Muttah Unit material, for example,appear to contain deliberately abundant non-plastic temper.

CLAY COLLECTION AND PREPARATION

Clay occurs readily in many parts ot the Oasis and has been exploited bypotters in the area tor thousands (5t years. The most commonly occurring claysare iron-bearing silts. Marl clays are also represented but are used less fre-quently. Recent building activities in the immediate vicinity ot the expeditionhouse resulted in the disturbance ot below ground levels exposing abundantclay deposits. This clay was a grey to green colour, a dry consistency andoccurred in a semi-hard to hard state. It appeared to be tree ot extraneousmaterial. A sample was collected (approximately 22 kg) and approximatelyhalfwas prepared for tests in reproducing Neolithic pottery vessels.

The dry clay was broken down into smaller pieces and then crushed usingstones and large chunks ot dry clay. This process required considerable staminaand took approximately one hour and 30 minutes. Once crushed, the groundclay was pushed through a sieve and the larger pieces were pulverised againuntil small enough to pass through the sieve. Water w'as then mixed with thecrushed clay Enough liquid was added to form a plastic paste (about 550 ml).The dry material quickly absorbed the water and became pliable. It was alsoobserved that once the water was added the mixture emitted a distinct damp,clay-like odour.

The clay was then placed in the shade to mature overnight. While it isunknown how long clay was rested in antiquit\', local potters working at Qasr,located at the southern end of the Dakhleh Oasis, prepare and throw claywithin relatively short periods of time (see Ceramics Techtiical 10, 2000, 54-59).

While the clay was resting, a series of ditTerent tempering agents were col-lected to add to the newly prepared clay It was decided to use a variety ot tem-pering agents in order to assess the etlects ot these materials on the texture andappearance ot the clay.

The clay was divided into six portions, each weighing approximately I .X kgand assigned a roman numeral to which the ditlerent kinds .uid quantities oftemper were added. The additives were mixed by hand, kneading the clayuntil evenly distributed. Six ditferent tabrics (1-VI) or wares were preparedbased on the Neolithic examples. Fabric I was tempered with fine to mediumshale. Fabric II was tempered with cciarse sliale. Fabric III was tempered with

Left to right: (^ay deposits al

Ain Giindi in the Dakhleh

Oasis. Sieving crnshed clay.

Mixing clay.

The experiment reproducing

Neolithic pottery was conducted

over a period of nine days (in

December 2002} during which

time clay was collected,

prepared, manufactured and

fired. The experiment was

executed at the expedition dig

house based at Ain Gundi,

near Mut, the capital of the

Dakhleh Oasis. For

convenience, the experiment is

divided here into three main

stages: clay collection and

preparation, the construction

and dr)'ing of objects, and the

firing process.

Cfr.inii, JECHNICAt No. 20 2005 55

Top: Diyi".'^ luvidmadv

pinch and coil-hiiHi pofs.

C'ciilrf: Finishedpo!s

before pit t^rinj^.

Abore: Amiiii'iii'^ pois in

pit ii'iih tiin\^s

coarse straw. To Fabric IV w.is .iddcd tine to niL'tiiuiii sand. Fabric V was tem-pered with crushed tired clay or grog. No temper was added to Fabric VI. Allthe tempering agents were easily available and Leathered trom the immediatearea. For example: sand is readily occtn'ring and abundant in the oasis; straw isused widely in many of the villages of Hakhleh and commonly added to mudbricks as a strengthening agent; shale characterised by deposits of greenishbands m the landscape and often associated with limestone caps, is also widelyavailable; and grog or crushed fired clay, coarse in texture, was obtainable fi"omfragments of broken and discarded pottery.

It was obser\'ed that these diflerent materials altered the texture and worka-bility of the clay. However, as it turned out after the inclusions were added, theclay was not as malleable as was hoped and around 3(l()gm of sand was added toeach of the six samples, including Fabric VI, to enhance the texture and work-ability of the clay.

CONSTRUCTION AND DRYING

All the Neolithic pottery recovered from Dakhleh is handmade. MostNeolithic pots are characterised by simple conical shapes. They commonlyhave rounded bases. The rims tend to be plain and may be rounded or taperslightly. Decoration is not common but rnany vessels have wel! finished sur-faces (wet-smoothed or self-slipped) or with slightly rouletted finishes.

Finch and coil methods are thought to have been the main techniques usedtor production dtiring the prehistoric period. Therefore, simple handmademethods of manufacture were used to replicate the vessels in this experiment.No tools were used and shapes were kept deliberately simple.

The construction of the vessels took place over two days and the drying pro-cess also took a few days. Two test bars ot each fabric were made and incisedwith 10 cm scales to measure shrinkage of the different fabrics. In all, 12 rect-angular-shaped test bars or 'briqtiettes*, 19 vessels and one perforated discwere made.

It was interesting to note that some fabrics performed better than others inthe vessel manufacturing process. For instance, the grog-tempered fabric (V)was the easiest to form, thotigh it tended not to hold its shape, dried rapidlyand the clay appeared to become fatigued as it dried at a faster rate than theother fabrics. Fabric III, tempered with straw, was also particularly hard on thehands, though this may not have been a concern tor the prehistoric potters. Itwas observed that the sand and tine shale tempered fabrics (I and IV) appearedto provide the best balance of workability and strength.

An additional insight gained from this exercise was how the ein'ironmentcan influence the construction process. For instance, it was observed that theideal time to construct vessels, in the open at least, was fVoni mid-morning toearly afternoon. In the early morning and the evening the clay was cold andless malleable and therefore more difficult to work. In the late atternoon thesun and wind would dry the clay at a faster rate than desired making produc-tion difTicult. In some instances exposure to the sun of an unfinished pot wasuseful, accelerating the drying of the vessel increasing its strength so that addi-tional coils could he added without causing a distortion of tiie vessel shape. Inshort, environmental conditions could be exploited or avoided depending onthe potter's interests and priorities during the manufacturing process.

The finished pots were placed initially in the shade and once leatherhardkept in the sun to dry completely. It was noted that after a 24 hour period thetest bars had shrunk by five percent, 48 hours later a further five percentshrinkaj^e was noted, after which point there was no or minimal shrinkage.

56 , i . JECHNICAt No. 20 2005

PROCESS

The tliel tor the firing was collected within 15(1 ni radius of the expeditionhouse. This included whatever natural sources were available and com-bustible, such as donkey dung, straw, palm matting, and twigs and branchesfrom Tamarix and Acacia trees. All these materials, except tor the donke\dung and straw are known to have existed \u the Oasis during the prehistoricperiod. While there is no evidence for grasses in tiie archaeobotanical recordthere could be several alternatives that the Neolithic potters used. Likewise, ifdonkeys were not yet introduced into the Oasis, there were a range ot otheranimals that produced dung. Overall, an estimated 5i} kg ot tuel was collected.

It is generally assumed that most Neolithic pottery was open or pit tired. Asyet no kiln structures have been recorded on prehistoric sites in the Oasis. It issuggested however, that 'proto-kilns' may have been used, which have sincedisappeared from the archaeological record. A 'proto kiln' is conceived as a pitin which both vessels and fuel are stacked together and on top of which dungis heaped to retain sutficiendy high temperatures.

Some of the dry pots and test bars were placed on a layer of straw and palmfibre matting at the base of the pit. Generally, larger and heavier vessels wereplaced at the bottom ot the pit with smaller vessels placed between and abovethe other pots. In most instances the pots were inverted so that the vesselsrested on their rims. Straw, small twigs and palm matting were stacked over theunbaked pots with larger sticks and branches then heaped on top of the pit.

The firing began in the late atternoon (around 16.0(1 hr) to coincide with theonset of the daily northerly winds. In the initial stages ot the firing a consider-able amount of fuel was used, mosdy branches of Tamarix and Acacia, becauseof the dry state of this material. It was estimated about 20 kg or so ot fuel wasplaced on top ot the fire during the first stages ot tiring. In an attempt to reducethe rate of combustion, the opening ot the pit was reduced w'ith surroundingsand so that the wind was prevented from channelling or tanning the tlames otthe fire excessively. After 40 minutes, most of the fast burning fuel was smoul-dering and had formed a canopy ot ash over the fire pit. At this point, the wallsot the pit were pushed inwards and the slow burning tuel (dung) was heaped ontop of the fire to completely smother any remaining fianies. The pit was left tosmoulder slowly overnight and for the next day and a half While there were notlaniL's and minimal snn:)ke emanating trom the pit it was possible to detect theglow ot hot coals in some places and teel an intense heat radiating from the pit.No gauges were used to measure the precise firing temperature but it is possibleto estimate the firing range based upon the texture and hardness ot the bakedvessels. It was noted that the baked pots have a 'bricky' resonance when strucksuggesting a firing temperature in the range of55(l-750°C.

OliSEIWAnONSThere was a low tailure rate with only two vessels that were extensively pit-

ted and one that revealed nnnor tractures along coil joins. More importantly,at least tor the purposes ot this study, the experiment was considered successtulas the vessels were tound to resemble the Neolithic pots in a number of ways.The reddish-brown fired surface colours ot the experimental vessels closelyresemble the colour range of the Neolithic pots. Examination ofthe texture otthe iron-rich fired clay used in the experiments appeared similar, it not thesame, as that used predominantly by the early potters. .Scratch tests conductedon the experiment.il vessels to determnie hardness post firing also fit the pro-file ot the ancient Neolithic examples. The section or core ot the experimen-tal pottery resembles most of the ancient pottery. The same firing patterns and

Top: Initial stages of firitig.

Centre: Later stages of

firing following ihe addition

of animal dnng.

Above: Last stages of pit

firing indicating

smouldering ash and

chanoal.

C.T.II I IKJECHNICAL NO. 20 2005 57

Top: Unpackitii; the pit.

Aho\'c: \lap o

colouration occur. The only ditlerence that might be noted is the inclusions.In the case ot Fabric V the grog particles look nothing like the temper found inthe Neolithic pottery. Likewise the shaie particles in the shale-tempered Fab-rics I and II do not resemble the ancient pottery.

An unexpected but important finding resulting form this experiment re-lated to the firing pit. Archaeologists freqnently encounter ash lenses andremains of firing pits. Frec]uently these are interpreted as heartlis tised fromcooking. It is quite likely that some of these features from antiquity may in factrepresent the remains of pottery firing pits.

Perhaps the most important finding resulting from this experiment is that itdeiuonstrates that it is possible to produce pottery in one location without theneed to travel U>ng distances to collect the necessary and difTerent resonrces forsuccessttil ceramic manufacture.

REFER KNt.:hHope, C.A., "PottL-iy Manufacture in the Dakhleh Oasis", in C.S. Chiirchcr &• A.|.Mills (eds.). Reports from the Survey of the Dakhlch Oasis Western Desert of Egypt1977-1987. Dakhleh Oasis Project: Monograph 2, CXvhnw Monograph 99. Exeter,l999.pp.21S-243.

Andrewjanuesun is the R. E. Ross Curator of the Classics and Archafolog\- (.oUei-tion andlecturer in the Ccntrt- tor Clas.sic!i and Archaeology' at the University of Melbourne.Ashten Warfe is completing a PhO in the C'entrt- for Archaeolog\' jnd Ancient History .uMonasli Univfi-siry. The authors would like to thaiik I )r C!olin A. Hope, co-mvestigatorot tiie Dakhleh Oasis Project and DircLtorofthc t Centre for Archaeology and Aiuient His-tor\' .It Monash University', tor supporting this expeninent and pro\iding guidance on thestudy and rese;nvh on tht- .uuicnt pottery of the Dakhleh Oasis.

58 c\r.,ni]csTECHNiCAL No. 20 2005