Archaeologists know from experience that Demeter sanctuaries can be counted on to warehouse large quantities of objects. The worship of the mother and daughter goddesses demanded mare than anything else the placation of the forces of Nature and the Afterlife through the dedication of gifts on a regular calendrical basis. Their precincts, from Sicily to the coast of Asia Minor, have been found choked with votive accumulations not unlike our own at Cyrene. Less immediately predictable are the elements of their architectural design which, depending on purpose, time and place, can consist of everything from rooms in converted houses to isolated shrine buildings to full-blown compound sanctuaries supplied with protective surrounding walls (periboloi), altars, storage chambers, underground votive pits, fountains and other lustration facilities, initiation halls, monumental entranceways, independent sacred houses or, less commonly, temples, and finally even the occasional cave to remind the cult’s initiates of Persephone’s reluctant entry into the Underworld. As we shall see, the Cyrene Sanctuary falls into something like the latter category (compound sanctuary).
The City of Cyrene
Cyrene was the leading city of the Libyan Pentapolis or ‘Region of the Five Cities.’ It was settled by Greek colonists from Thera and other island and mainland cities toward the end of the 7th century B.C. (see Uhlenbrock’s box on the founding of Cyrene, and Schaus, this issue), and its formal existence as a center for organized Hellenic life persisted until the Arab invasions (A., 643). The ancient town was established about 13 km (ca. 8 miles) inland from the Mediterranean on the 600 meter-high crest of a limestone plateau known today as the Gebel Akhdar or “Green Mountain” (Fig. 2). The plateau or gebel forms an important physical barrier separating coastal eastern Libya from the Saharan region to its south. During the winter and early spring the mountain heights block the southerly drift of water-laden clouds, squeezing from them enough rain to support a fertile, temperate northern Mediterranean environment a mere 64 km north of one of the most incandescently arid and sterile regions on the globe.
Cyrene agriculturally based economy thrived on the export of wheat, legumes, fruit, sheep and goat-derived products, horses, and a highly sought-after herbal plant known as silphium (Fig. 3; see box), which grew exclusively on the Libyan gebel. The city itself lay inside a protective circuit of stone defensive walls, erected during the Ptolemaic era, that measure overall just under 1,600 m northwest by southeast and ca. 1,100 m northeast by southwest (Fig. 4). It rose on two massive hills. The southwest hill (on which lie the acropolis, agora, and forum) is totally free of modern building; the northeast is largely covered with the old Arab village of Shahat, stands of reforested evergreens, and cultivated ploughlands, and remains largely unexplored.
The ancient urban center was divided by three main roads. The Valley Road follows the sloping valley between the two hills to the Sanctuary of Apollo with its standard Greek entrance gateway, temples, altars, fountains, theater, and, later, Roman-period baths. The second road, named after the city’s first king Battus, connects the still unexcavated acropolis zone with the city gymnasium and Roman-period forum. The third road crosses the main axis of the city east of the forum. At its intersection with the Valley Road were more temples, a basilica, and a series of important Roman-period urban villas. Off in the northeast corner of the walled city rose a massive temple dedicated to Zeus (currently undergoing reconstruction) and the city’s still unexcavated circus or hippodrome.
A century-long experiment with a republican form of government followed the collapse of Cyrene’s hereditary monarchy around 440 B.C. The region then fell under the control of the Egyptian Ptolemies after Alexander the Great’s death in 323. By the early 1st century B.C. it was swallowed up by the Romans, but continued to retain its distinctively Greek cultural bent until overrun by the Arab forces of Amr Ibn el-Aasi in the 7th century A.D.
The Sanctuary’s Foundation and Appearance
The Extramural Sanctuary of Demeter and Persephone was laid out about a generation after the initial foundation of Cyrene and lasted in use until badly damaged by an earthquake in A.D. 282 (see Fig. 3 in Preface; Figs. 8, 7) and eventually totally destroyed by an even more severe earthquake in A.D. 365. In its heyday, which when measured in terms of architectural expansion seems to have coincided with the reigns of Trajan through Antoninus Pius (A.D. 98 through 181), ?the Sanctuary covered more than 9,000 sq m. Its structures were distributed over 20 m of abruptly rising ground, broken into three major divisions prosaically designated the Lower, Middle, and Upper Sanctuary levels.
Entry to the lower northwest corner from the nearby walled city was gained in antiquity by means of a bridge (S27/S28) across the wadi drain (Fig. 5: nos. 6,7). Narrow steps cut in the steep opposite face of the wadi above the bridge permitted access to the city’s agora (Fig. 5: nos. 10,11) through some still undisclosed opening in the walled ramparts. A monumental staircase (Fig. 5: no. 4) connected the Sanctuary’s upper grounds to an unidentified walled complex (Fig. 5: A,B) installed at a higher level on the great hill rising to the south. The principal entrance to the Upper Sanctuary during the Roman period was provided by a four-columned propylaeum or gateway (520; Fig. 11), strategically positioned in front of the junction of the monumental hillside staircase and the ancient road leading back to the southeast suburban quarter of the city along the rim of the wadi. In this way the Sanctuary grounds were architecturally linked with both the city and the countryside that fay to its south.
To judge from its emblematic depictions on the pre-Hellenistic Cyrenean coins, the silphium plant was composed of a strongly ribbed stalk crowned by a terminal umbel (i.e., a flower whose pedicels or stems spring from the same point to form a rounded cluster) and two opposing pairs of protuberant leaves, as well as axillary umbels that sprang obliquely from the stalk (Fig. 3). From an early time it was harvested as an uncultivated wildflower by the native Libyans, who transmitted their knowledge of its food and medicinal properties to the first Greek colonists. The juice was used as a condiment and drug, the stalk eaten as a vegetable, and the residue fed to cattle. A recent article has even suggested that it was used as an ancient oral contraceptive (Riddle and Estes 1992:226, 230). Since it grew wild, its survival depended on two conditions. First, that it not be overharvested, and second that the Libyan herdsmen keep their sheep and goats from eating it.
Apparently as early as the 3rd century B.C. the Libyans began a deliberate sabotage of silphium out of protest against excessive taxation and the restriction of their traditional grazing lands. By the 1st century A.D. its rarity made silphium an extremely expensive commodity to acquire on the foreign market, and by the beginning of the 5th century the plant was virtually extinct, although, interestingly enough, Bishop Synesi us (b. Cyrene ca. A.D. 370) seems to have known of its existence as a cultivated plant. Reports of its rediscovery on the Gebel Akhdar by an Italian botanist during the summer of 1991 remain to be confirmed.
The individual components making up the Sanctuary in its final pre-earthquake development (Preface, Fig. 3; Figs. 6, 7) are mainly local variations of the same stock features previously detailed for other full-blown Demeter sanctuaries: a system of terraces (T11-T13, T15, T20-T22), interior retaining walls (W1, W2), and stairways (G2, RI-R4) surrounded by an outer lateral peribolos (T14, T17, T18). These in turn support, contain, and give access to a variety of independent shrine houses (S1, 55, SO, S7, S8), storage rooms (59, 511, S12), water installations (Fl, F2), a votive dump (S18), cult rooms variously distinguished by columnar-fronted entrances and/or mosaic floors (S10, S16), an open-air access corridor with mosaic floor (S22), as well as the already mentioned colonnaded propylaeum (520) and Southwest Building (S17). A number of units still elude positive identification (52-S4, 514, S23, S24, 525, S26). While no permanent altars have come to light, the relatively large number of portable stone altars found, ranging in height from a few centimeters (for burning incense?) to over a meter and a half, undoubtedly helped to accommodate most of the sacrificial activity taking place within the Sanctuary’s central core. The major permanent altars remain to be discovered at some future date on the level of the Upper Sanctuary.
In the absence of detailed information about the still largely unexcavated colonnaded Southwest Building (S17), which, on analogy with Eleusis (the best-known of all of Demeter’s mainland sanctuaries), could be an initiation or mystery hall (where sacred rites were conducted in secrecy), the most memorable features of the Cyrene installation are its hillside terracing, propylaeum, and independent shrine houses.
The post-A.D. 115 monumental terrace wall (T20) separating the Lower from Middle Sanctuary is the latest addition to a series of similar terracing devices that go back to the foundation of the Sanctuary at the end of the 7th century B.C. (Fig. 8). Its impressive thickness covers vestiges of an Archaic rubble wall (P8/P9), succeeded by a late 6th, early 5th century pseudoisodomic (low alternating with high courses) cut-stone terrace (T3/T4), which in turn was replaced by a more conventional ashlar or squared-block facing (T10) of Augustan date around the time of Christ. In a literal way the terrace physically incorporated every major developmental phase experienced by the Sanctuary over a 750-year period.
The most flamboyant aspect of this terrace had nothing to do with its latest and most massive T20 faring but instead was linked with the preceding T10 phase, when a facing of such exceptionally regular block units was thrown across the eastern two-thirds of the Middle Sanctuary that its excavators immediately nicknamed it the ‘Thirty-Two Centimeter Wall’ (Figs. 7, 8). When this was installed as a reinforcement to the pseudoisodomic wall around the beginning of the present era, its builders backfilled the ca. 3-5 m deep V-shaped gap (Fig. 9) between its inner face and the sloping bedrock scarp with earth mixed with massed concentrations of artifacts. This created in effect a gigantic storage bin filled with discarded votive paraphernalia.
The deposit behind the wall seems best explained as fill carried from dumps of votives buried at earlier times in other parts of the Sanctuary grounds, combined with the debris of dismantled buildings. From what was learned from several controlled tests, the man-made objects fall into four broad categories: (1) pottery and lamps; (2) stone and terracotta statues, statuettes, and figurines; (3) personal ornaments such as gems, seals, jewelry, and pendants; and (4) miscellaneous objects ranging from glass (cups and bowls) to faience (figurines), iron (utensils, blades), bronze (fragments of sheeting, strainers, rings), alabaster (bowls, jars), and terracotta (loom weights).
Sherds and terracotta figurines, dating mainly from the 6th and 5th centuries B.C., were recovered in unusually heavy concentrations. A total of 52 stone statue and statuette body fragments, 10 head fragments, 7 relief fragments, and a nearly intact two-thirds life-size female statue were also found. With the exception of the lower half of a 6th century kore (female figure), the statues appear to be mainly post-Classical in date. More will he said about individual objects in the other articles.
The Museum’s Excavation
Of the Sanctuary’s three internal divisions, only the Middle Sanctuary has been archaeologically investigated in something approximating its entirety. When The University Museum took over the project in 1973 (see box in Preface), a decision was made to limit activity on the level of the Lower Sanctuary to clearing its overburden of scrub brush and to recording whatever wall features already stuck up above the ground without benefit of excavation. As work progressed through the mid-1970s it became increasingly plain that the grounds which we had initially assumed made up the site’s climactic element constituted instead a middle zone and that the western half of what must have been originally the actual Upper Sanctuary was dominated by a large colonnaded hall-like building (see Preface, Fig. 3: 517). This led to a further decision to limit the exploration of the Upper Sanctuary to a series of test trenches while attempting to complete the excavation of all of the Middle Sanctuary features that came to light south of the expedition’s railway line. With the Museum’s presence in Libya halted in 1981, the practical result was that the entire southern upper edge of the Sanctuary, including the back of the large Southwest Building (517), had to be left buried.
The colonnaded propylaeum (S20), erected following the Jewish rebellion that rocked the province of Cyrenaica in A.D. 115, still awaits restoration on paper as part of the final volume of the site’s publication. In the meantime, based on the close similarity of its plan (Fig. 11), decorative order (hybrid Corinthian), and overall dimensions with the Roman propylaeum restored by the Italian Archaeological Mission in Cyrene’s nearby Sanctuary of Apollo (Fig. 10), one has little difficulty in imagining its general appearance.
This leaves the sacred houses, which is what we call the five independently sited shrine buildings distributed over the Middle Sanctuary slope (see Preface, Fig. 3; Figs. 6, 7). The earliest (S1, S5, 56) seem to belong to the first 125 years of Sanctuary development, while the remaining two are additions of the Hellenistic (57) and early Roman period (58). Although no two build ings are exact copies of each other, the group shares in common a simple columnless plan, modest dimensions, and single entrance. Four of the houses opened north, down the slope toward the city, but the fifth (56) faced east. 55 and S6 opened at their short ends; SI, the earliest in the series, and S7 were virtually squares, while S8, the latest, was entered through one of its long sides. No doubt coincidentally, the interior length was nearly precisely double the interior width of S8. At least two of the buildings (55, 57) had stone benches across their back walls as well as painted plaster interiors.
A closer look at the S7 Sacred House will serve to illustrate the type. It was a simple 4.2 m square single room structure, opening north with, as already said, a low stone bench stretched across its rear wall (Fig. 13). Panels with linear geometric designs covered the inner surfaces of its back and side walls. The outer walls were built from ashlar blocks of an excellent grade of local limestone, free of obvious shell inclusions and capable of taking a crisp finish.
Both plan and elevation appear to have been based on modular increments of a local variation of the “Alexandrian foot” (a measure of length associated with Ptolemaic Egypt), here measuring 35 cm. To restore the building, staff architect Jim Thorn applied a simple grid to it plan and four sides (Figs. 13, 14) Each grid unit is made up of three local ancient feet (1.05 m). Thus them width of the north facade abovo stylobate (pavement) level equals units, the total height 6 units, them pediment width 5 units, and them pediment height 1 unit. The con stituent parts of the decorative entab lature and door are restored iron actually surviving fragments, whosl dimensions also appear to be frac Lions of the same grid unit broker down into a half-foot or single foot The pan tiles covering its roof mea sure 1 by 1.5 ancient feet.
This then is the model for the typ of building that dominates the Middle Sanctuary. Its presence, Omuta tics mutants, can he readily traced 11 other Demeter sanctuaries thorough out the Greek world, although per haps nowhere else is it known to appear with such deliberate insist tence as here. These buildings are still mysterious, their functions kept deli berately obscure by the rules for secrecy governing the administration of Demeter’s cult throughout anti quity. However, they may have fu] filled such roles as the facilitating c sacrifice and other set forms of ritua the display and storage of votive the reenactment of sacred myth, th commemoration of the dead (whose temple-tombs, reliefs with pedimented tops, and funereal heroons or shrines painted on vases echo the same basic house form), and perhaps even the supply of fixed stations along the processional routes prescribed for the conduct of festivals.
Demeter and Persephone
The last two syllables of the name of Demeter (in Latin, Ceres) mean “mother.” At some early stage she appears to have cast off a younger double of herself, the ‘Virgin’ or ‘Korea,’ whose father may have been Zeus. The Korea’s proper name is Persephone (in Latin, Proserpina). The older goddess governs the fruits of the earth, especially the bread grains, the “staff of life.” The younger goddess, through her mythic union with Hades, lord of the Underworld, with whom she is forced to live some part of the year, symbolizes the power of the grain. Like the goddess the grain is annually renewed only to disappear beneath the ground, just as do the generations of mankind whose inescapable fate is death. Thus both Demeter and the Kore are intertwined with the natural cycle of life and death.
Both goddesses are normally depicted in ancient art as majestic, beautiful adult women equipped with scepters and ears of wheat or torches. Demeter is usually shown seated Fig.12). Her hair, which is said to be blond, frequently hangs long behind her neck, while the Korea’s is more apt to be upswept, at least in later depictions, but it is often difficult to distinguish the two. Persephone is frequently shown in relief sculpture, painting, and mosaic in the act of being forcibly abducted by Hades driving a four-horse chariot. She is more often invoked in ancient literature and on inscriptions by her euphemistic title of Kore than by her true name of Persephone, which carries associations of death and retribution.
Both goddesses may be accompanied in painting and sculpture by piglets, which carry a complex range of symbolic associations ranging from purification to the fecundity of the earth. Piglets are either eaten or sacrificed in the course of various Demeter rites and festivals, in particular during the celebration of the Thesmophoria (see Kane’s article and her Fig. 7, and Uhlenbrock, this issue).
Evolving Nature of the Cult
At the outset the cult drew its support from the colonial agriculturalists who found themselves transplanted from a relatively stable natural setting to an unfamiliar environment in an alien continent, protected from the desert by only a narrow sliver of plateau and surrounded by an unpredictable native Libyan population. lt is little wonder that Cyrene’s people felt almost preternaturally dependent on Demeter, the goddess of vegetation and fertility, for the survival of their community (see box on Demeter and Persephone). This sense of dependency was physically expressed in the endless stream of anonymous gifts brought to the Sanctuary during the 6th and early 5th century. As time went on, people’s interests turned inward, and their traditional concern for communal prosperity blended with more private considerations of death and the immortality of each individual’s soul (see Warden, this issue). This is most eloquently reflected in the personal inscriptions, cultic reliefs, and stone statues of private individuals that tend to dominate the votive offerings of the Sanctuary’s Hellenistic and Roman phases (323 B.C.-A.D. 262).
What became of the Sanctuary following its final destruction? More than 15 centuries elapsed between the A.D. 365 earthquake and Italy’s military occupation of Libya shortly before WWI, leaving scarcely a trace in the archaeological record. Broken wine bottles, spent rifle cartridges, and the odd regimental graffito cut into a fallen block of marble were found mingled with the tops of fallen statues and scattered Greek and Roman potsherds in the surface stratum when the site was first opened. We know that Cyrene’s Christianized Greek population sought protection inside the reduced circuit of the old city walls until the Arab invasions of A.D. 643. Some form of urbanized existence may have lingered on within the shell of the old city until the arrival of true nomadism toward the middle of the 11th century. The discovery of two 7th century Byzantine coins and a single 8th century Arab coin (see Buttrey, this issue) does little to shed light on the veil of otherwise total darkness that descends on the site after the mid 4th century.