An important baseline study of ceramics and material culture in southern Ecuador, which establishes the region’s artistic and trade connections with better-known areas of the prehistoric Andes.
Challuabamba (chī-wa-bamba)—now a developing suburb of Cuenca, the principal city in the southern highlands of Ecuador—has been known for a century as an ancient site that produced exceptionally fine pottery in great quantities. Suspecting that Challuabamban ceramics might provide a link between earlier, preceramic culture and later, highly developed Formative period art, Terence Grieder led an archaeological investigation of the site between 1995 and 2001. In this book, he and the team of art historians and archaeologists who excavated at Challuabamba present their findings, which establish the community's importance as a center in a network of trade and artistic influence that extended to the Amazon River basin and the Pacific Coast.
Art and Archaeology of Challuabamba, Ecuador presents an extensive analysis of ceramics dating to 2100-1100 BC, along with descriptions of stamps and seals, stone and shell artifacts, burials and their offerings, human remains, and zooarchaeology. Grieder and his coauthors demonstrate that the pottery of Challuabamba fills a gap between early and late Formative styles and also has a definite connection with later highland styles in Peru. They draw on all the material remains to reconstruct the first clear picture of Challuabamba's prehistory, including agriculture and health, interregional contacts and exchange, red-banded incised ware and ceramic production, and shamanism and cosmology.
Because southern Ecuador has received relatively little archaeological study, Art and Archaeology of Challuabamba, Ecuador offers important baseline data for what promises to be a key sector of the prehistoric Andean region.
- Author's Note
- Chapter 1. Introduction to the Project
- Chapter 2. Pottery Wares and Forms
- Chapter 3. Petrographic Analysis of Selected Ceramics from Two Sites in Ecuador (David V. Hill)
- Chapter 4. Pottery Decoration
- Chapter 5. Pottery Comparisons
- Chapter 6. Effigy Vessels and Figurines
- Chapter 7. Stamps and Seals
- Chapter 8. Stone and Shell
- Chapter 9. The Burials and Their Offerings (James D. Farmer)
- Chapter 10. Human Remains (Douglas H. Ubelaker)
- Chapter 11. Zooarchaeology (Peter W. Stahl)
- Chapter 12. Reconstructing Challuabamba's History
Today Challuabamba (chī-wa-bamba) is a developing suburb of Cuenca, the principal city in the southern highlands of Ecuador. The old farmsteads that once composed the village are being replaced by large suburban houses, and even the Pan-American Highway along the north bank is being superseded by a new four-lane divided highway on the south bank. The bridge over the Tomebamba River, just seven kilometers downstream from the site of our excavations, is a vital link in a wide network of roads leading north to Riobamba and Quito, west to the coastal plain and Guayaquil, and south to Cuenca and beyond to Peru or eastward to the Amazon basin. This is not an invention of modern transportation systems but follows what must have been ancient trails along an array of rivers and streams that join to drive a gorge through the eastern foothills of the Andes on their way to the Amazon. Challuabamba sits near the junction of these ancient trails.
Challuabamba has been known as an archaeological site for nearly a century since Max Uhle's work there in the 1920s. In an interview in 1962 with a Quito newspaper the archaeologist Edward Lanning described the site in these words: "From a professional point of view, Challuabamba is the most fantastic site in the world, since the ceramics one encounters there in unbelievable quantities are extremely fine". My interest in Challuabamba was stimulated by the results of my previous studies in the highlands of northern Peru. At the preceramic site of La Galgada, occupied c. 2800-1800 BC, we had found carved stone bowls, textiles which revealed the progression from looped and twined to heddle loom construction, and the first traces of pottery about 2000 BC (Grieder et al. 1988:185). At Pashash, in the same province, we had unearthed very fine Recuay style pottery and jewelry of gilt arsenic bronze, dated between AD 300 and 600. While both La Galgada and Pashash had large stone buildings in styles unknown in Ecuador, the pottery and textiles showed techniques developed earlier in the Valdivia culture of coastal Ecuador (Marcos 1979). Challuabamba, in the Ecuadorian highlands on the eastern slope of the Andean cordillera, looked like the ideal location to search for links in the chain that led to La Galgada and Pashash. In 1995, with research funds from the University of Texas, I undertook a series of small excavations along the banks of the Tomebamba River on the site of Challuabamba. (Those place-names, given after the Inca conquest of the 1480s, are in the Quechua language: tome or tumi [axe], challua [fish], and bamba [field].)
The area along the river is in the cool, rainy highlands, at about 2,300 m (7,545 ft) above sea level at Challuabamba. Flowing east from lakes near the crest of the Andes, the Tomebamba River joins a group of highland streams to form a major Amazon tributary, and its waters eventually reach the Atlantic Ocean, some 3,400 km (roughly 2,100 miles) to the east.
Inhabitants of the Tomebamba Valley can reach the Pacific Ocean much more easily, crossing a low pass toward the southwest and following the Jubones River about 130 km down the steep Andean escarpment to the shores of the Gulf of Guayaquil. Although only 300 km south of the equator, the surrounding regions provide almost all the varieties of natural environments, from sandy desert to tropical rain forest, from ocean beach to frigid tundra.
A visitor looking for ruins at Challuabamba will be disappointed, for there are few signs on the surface of ancient life, only grassy meadows, old farmsteads, a modern highway, and the beginnings of suburban development. Nevertheless, the valley has been the site of a series of important archaeological studies.
While the archaeology of Ecuador's coastal lowlands has been intensively investigated during the last fifty years, the southern highlands are still in a relatively early stage of exploration. Max Uhle (1922b:108) was the first to write of the archaeological resources of Azuay Province, including "the unexpected discovery of great treasures," followed by "a period of frantic search of the whole region" between 1856 and 1899. The result was the destruction of the ancient tombs and "everything they contained, only the gold being saved" (see also Salazar 1995:143-155). Uhle's work at Challuabamba remains the most important account of the site. Uhle (1922a:206), long familiar with Peruvian sites, was inspired by his reading of H. J. Spinden's study of the Maya and believed that he had found a civilization founded by the Maya in the interior of Ecuador. While the Maya connection has not found acceptance (Collier and Murra 1943:90), Uhle's definition of a "Chaullabamba [sic] civilization" occupying nearly all the Ecuadorian highlands south of 2 degrees south latitude (the Río Chanchan) and parts of the western lowlands has remained influential. In this report we return to the correct spelling (according to the Ecuadorian Instituto Geográfico Militar) of the site name to distinguish it from the archaeological materials gathered under the rubric "Chaullabamba civilization," which are still mostly hard to assign a specific location or period.
The exploratory studies made by Max Uhle are fundamental to the work that followed, though the diffusionist theory that inspired Uhle has been proven wrong. Donald Collier and John Murra (1943), at the end of their studies at Cerro Narrío (about 20 km north of Challuabamba), made a survey of collections and archaeology of the Cuenca area. The close relationships among the ancient peoples of those neighboring areas have always been clear, and Wendell Bennett's (1946) excavations along the Tomebamba River carried on the work of Collier and Murra. The ceramic typologies established by Collier and Murra and Bennett have served as models for later investigators.
Several archaeologists have studied Formative sites in the area in recent years, and studies in other nearby districts offer comparative data (Figure 1.3). Elizabeth Carmichael, Warwick Bray, and John Erickson (1979) found some material closely related to Challuabamba in the Jubones Valley, and Carmichael (1981) obtained radiocarbon dates from Challuabamba as part of that larger study focused on the Jubones River. Mathilde Temme (1999) has made extensive studies of Putushío, near the headwaters of the Jubones, in the highlands south of Challuabamba, while Jorge Arellano (1994, 1997) has studied early settlements with ceramics at Loma Pucara in the Cebadas Valley, to the north between Cuenca and Riobamba. Pirincay, 20 km downriver from Challuabamba, has been excavated by Norman Hammond and Karen Olsen Bruhns (1987), by Bruhns (1989, 2003), and by Bruhns et al. (1990, 1994) (and see Miller and Gill 1990). With a long sequence of building, ceramics, and craft specialization, Pirincay shows important connections both with the Pacific lowlands and with the Upano River Valley in the Oriente (eastern lowlands), mainly during the first and second millennia BC. Dominique Gomis (1989, 1999) has published studies of Challuabamba ceramics, and her work is also presented in Anne-Marie Hocquenghem et al. (1993). Her analysis, based on different criteria, provides a useful check on our work and seems to represent the same early period in the history of the site. Her early period is referred to by Michael Tellenbach (1998:305) as "Apangora [sic]" (Abangora), the creek that enters the river on the Vasquez property about 500 m east of our excavations, where traces of excavation remain. Jaime Idrovo Urigüen (1992, 1999) refers to Gomis's work at the site and proposes a larger picture of Formative developments in the region. References to these studies are made throughout this report. The Challuabamba ceramic collections are particularly closely related to those from Pirincay and Cerro Narrío.
When we map out the locations of previously published studies in the Tomebamba/Paute River system it becomes evident that these valleys have been an important focus of archaeological interest during the twentieth century. It is clear that we are still far from a complete picture of that society and its history. Its region may have extended from Cerro Narrío in the north to at least the Jubones Valley in the south. It may have been united by a common language, as it was by Cañari in Inca times, as well as by ceramic and other artifact styles. In this study we can present only an account of the remains recovered at Challuabamba, but wider comparisons are offered to suggest the range of cultural interaction.
The Present Study
For as far as human records and memories go back, the autochthonous population of the provinces of Azuay and Cañar has been the Cañari people. They dominated most of the territories of the modern provinces of Cañar, Azuay, and Loja and the northern coastal region of El Oro before AD 1400 and only lost the parts of Loja south of the Jubones River to their sometime allies the Jívaro (or Shuar) in the 1400s. In the hierarchy of Cañari chiefs, the chief ruling at Tomebamba was considered paramount; when the Inca Empire conquered the region in the 1480s, it became an important city and received its Quechua name. Emperor Huayna Capac (reigned 1493-1524) was reported (by Betanzos 1996 :121, 182) to have been born at Tomebamba, on the site now occupied by the city of Cuenca, which he endowed with monuments and sacred sites that established its rank as equivalent to that of Cuzco. Fiercely independent, the Cañari welcomed the Spanish invasion and joined Francisco Pizarro's forces against the Inca. After the fall of the Inca Empire in 1533 the region became part of the Spanish colonies. An early Spanish document describes fishing as an important occupation in the Tomebamba Valley (Murra 1963:799), and the inhabitants of Challuabamba (Fish Field) are supposed to have paid their taxes in fish from the river.
After discussing contacts and similarities between Formative Peru and Ecuador, Richard Burger (2003:481) concluded that "despite these similarities and linkages, the world of the Ecuadorian Formative appears to have been profoundly different from that of the Central Andes in terms of its economic organization, its sociopolitical structure, and ideological frameworks that made life comprehensible and meaningful." Different attitudes about their social group, leadership, and the possibilities open to them as individuals are evident in the patterns of habitation and production in Ecuador compared with their contemporaries in Peru. The Challuabamba project was conceived as a way to explore the Ecuadorian version of Formative life and art.
Excavation is the only way to obtain specimens of ancient art in their original cultural context. Too often we are dependent on the excavations of treasure-hunters who have no interest in the artifacts beyond their market value and provide no useful information about the provenience of their finds. The Challuabamba excavations were undertaken to contribute to filling a gap in the art-historical record of the north Andean highlands. The resulting sample consists of over 63,000 potsherds and about a dozen nearly complete ceramic specimens, another dozen stone carvings, traces of building foundations and wall materials, evidence of seven burials, and a very large collection of animal bones. Based on the stratigraphy and other contextual evidence we can put much of the sample into a temporal sequence, assign it approximate dates, and show its technical and design relationships to other sites. While we have a little evidence of social rank, we are not in a position to speculate about social organization in general. But the material provides a good record of technical, design, and expressive achievements during a crucial period in the history of the region.
It was not only research in Ecuador that focused archaeological interest on that country, for Peruvian preceramic and Initial Period sites have long yielded materials with Ecuadorian connections: Spondylus shell in preceramic levels, unquestionably imported from Ecuador, followed by early loom weaving, evidently practiced earlier in Ecuador (Marcos 1979), and ceramics, probably inspired by Valdivia pottery in Ecuador made centuries earlier (Grieder et al. 1988:1, 16, 189-190, 197).
A general picture of cultural development in South America has been forming since about 1950. One of the important advances was the making of ceramics. As in the case of the other major technical achievements of Pre-Columbian South America in textiles and metallurgy, ceramic products were valued for both utilitarian and expressive purposes. With our present knowledge we can make some generalizations about the sequence of those technical advances: looped and twined textiles invented by preceramic societies, ceramics spreading through the tropical lowlands before 4000 BC and down the Peruvian coast and into the Andean highlands by about 2000 BC, and the earliest experiments with metalwork, which seem to have occurred about 1500 BC (Burger 1992:127).
Although Challuabamba's climate is too damp for the survival of textiles and our excavations show no signs of metalwork, introduction of ceramics from sources on the Ecuadorian coast and relationships with other early pottery-making sites are manifest in the excavated materials. On the basis of this general picture we can hypothesize that much of the material in our excavations was produced during the second millennium before the modern era.
The present study emphasizes art and the development of the Formative culture that fostered it. Two radiocarbon assays of wood carbon that we collected at Challuabamba gave dates in the range 2300-1700 BC (calibrated), which encouraged us to continue our search there for early participants in the developmental surge implied by its name, Formative Culture.
In June 1995, having obtained the approval of the Ecuadorian National Institute of Cultural Patrimony in Quito and Cuenca and accompanied by the Ecuadorian archaeologist Antonio Carrillo, our small crew began some preliminary examination of the south bank of the Tomebamba between two small tributary streams, the Salado and the Abangora. The area consisted of pleasant meadows and woods, with some modest houses of local farmers interspersed with a few fine new houses of the emerging suburbs. The site is bordered on the north by the river and on the south by hills, with a huge construction project at the base of the hills, where a highway and water treatment facilities for the city of Cuenca are being built. The only hint of ancient construction was lines of river boulders visible in the upper levels of the riverbanks; in places potsherds were eroding from the bank, testifying to an ancient occupation. In my notes I identified the common pottery type as Cerro Narrío Red-on-Buff, a Formative type named at the site about 20 km north that had been studied by Collier and Murra (1943). This, we concluded, was part of ancient Challuabamba. In the following days, accompanied by Antonio Carrillo as the local representative of the national archaeological administration, I made an agreement with the landowner, Francisco Chimbo, to make small excavations into the river terrace on his land.
The project's first year was designed to obtain a stratigraphic profile to sterile soil and gather some information about the lines of stones eroding from the upper levels. The 1996 field season moved upriver about 100 m to the Jaramillo property, through the courtesy of Juan Pablo Jaramillo. Two excavations were laid out at the edge of the terrace, designated Cuts 2 and 3 (Figure 1.4). In subsequent years the upper levels between Cuts 2 and 3 were expanded to reveal the late walls. Cuts 4 and 5, made in 2000 following magnetometer and ground-penetrating radar (GPR) mapping in 1999, were exposures intended to examine late floor and wall patterns. The excavations reassured us of the integrity of the deposits, which showed no evidence of having been subjected to looting or prolonged flooding. Although erosion had occurred at the edge of the terrace, and perhaps between some stratigraphic levels, the main body of the terrace was intact, with horizontal strata and no interruptions in the sequence of cultural deposits.
Cut 1 is shown in plan and section (Figure 1.5). The location was selected on the basis of a line of river boulders appearing near the surface of the terrace and a fairly deep vertical terrace wall showing numerous potsherds. The eroded bank was cut to reveal the strata in a 3 m wide vertical profile to establish the depth of cultural deposits (1.6 m) and an exposure 1.5 x 5 m was laid out parallel to the edge of the terrace 1 m south of the profile.
Cut 2 was a 3-m square at the north edge of the terrace, excavated in 1996-1997 (Figures 1.6 and 1.7). Five cultural levels are defined (the sixth is sterile base). The upper level contains several fragments of modern glass along with abundant ceramic, lithic, and bone debris. The later of two layers of river cobbles occurs at about 50 cm. A layer of smaller cobbles lies at about 70 cm. The larger upper layer is associated with four shallow pit burials (Burials 1-5) with offerings. While the upper layer seems associated with the burials, neither layer shows a definable building pattern. Level 3 has two layers of ash and charcoal, and in Level 4 we encountered the only good post-mold in our excavations (about 37 cm deep), with no trace of the post but with a large Red-on-Cream potsherd lying in the bottom. Level 5 shows some ancient excavation, with the very light soil of Level 5 resting on top of a layer of dark Level 4 material.
Cut 3 is shown in a plan, a section, and a view. Located about 20 m west of Cut 2, it is the largest area and the deepest of our excavations. Figure 1.8 shows a plan of the excavation and a synthetic section from the north (the river side) into the trench, with the south wall of Sector G in the center, the east wall of the trench on the left, and the west wall on the right. Each sector is 2 x 2 m. The depth of the cultural deposit is 2.1 m in seven natural strata in the trench (Sectors A-J) (Figure 1.9). The remaining area of Cut 3 (Figure 1.4) was excavated only to the base of Level 2 to reveal the pattern of the stones.
Level 1 in Sector L contained a bed of bone debris about 1 m square near the surface, with the remains of a butchered deer. The bones had all been cooked or burned and broken into small fragments, except for a larger piece of antler. The materials suggest a small group feast, with an unusually fine bridge-spout bottle (Figure 4.16) broken in the debris. Levels 3 and 4 of Sectors E and G contained small fragments of bone and fine potsherds that were interpreted as evidence of disturbed burials; and in Sector H at the base of Level 4 and in the surface of Level 5 were two intact burials (Burials 6 and 7) with offerings. In Sector F there are signs of ancient excavation in Level 4 into Levels 5 and 6.
An area about 8.5 x 19 m was exposed to the base of Level 2 to trace the large stone walls. The parallel lines of river boulders are about 2 m apart, running northeast-southwest. The eastern end of the southern wall appears to be a massive revetment for a low platform. These late constructions are discussed below.
Cuts 4 and 5 were excavated in 2000 following magnetometer and ground-penetrating radar mapping of the meadow area in 1999, which suggested stone constructions in those areas (Figure 1.10). The small stone lines in the center of Figure 1.11 may have lined a stockade wall or perhaps a small canal. A later large-stone wall is at the left (Figure 1.12).
Cut 5 was also in response to the magnetometer/GPR study and showed traces of two levels of walls and a series of fragmentary floors, the deepest at 60 cm, with traces of small stone lines. Traces of burning show on the mud floor under the east wall. On top of the small stones are a line of larger river boulders at about 50 cm depth. They appear to be part of a structure of which two pieces of floor remain: one with numerous traces of burning and carbon deposit, the other packed brown adobe (Figures 1.13 and 1.14).
Stone Used in Building
The most obvious traces of ancient occupation on the site are the clusters of waterworn boulders from the river evident in the eroded face of the upper terrace of the riverbank. The largest on average measure about 50 cm in their longest dimension. Their appearance in lines high above the riverbed is an indication that they were carried there by human beings. While boulders were moved by periods of strong current, the patterns of stones seen in the riverbed have nothing in common with those above the riverbank. Excavation reveals that lines of stones are found only in the upper half of the deposit and can be separated into an earlier style using smaller stones averaging 30 to 40 cm in their longest dimension and a later style using larger boulders (Figures 1.15 and 1.16). Both periods of building are apparent in the photograph of Cut 4 (Figures 1.11 and 1.12), with parallel lines of small stones in the deeper area (about 60 cm) and a row of larger boulders stacked across the near corner, representing the latest period.
Max Uhle (1922b:207-208; my translation) refers to "foundations, formed generally of river stones, which still show the outlines of their ancient buildings," which he saw in Challuabamba, Huancarcuchu, and Carmen, all sites that could be considered part of what we might call "greater Challuabamba." Gomis reported (in Idrovo Urigüen 1999:119) house foundations in Challuabamba made of river boulders in circular and rectangular patterns. These alignments had a long tradition behind them. The earliest examples currently known are in the upper Zaña Valley, in northern Peru, where stone-framed earth platforms bearing cane and daub domestic structures date from the period 6000-4000 BC (Dillehay, Netherly, and Rossen 1989).
Although no platforms are known to survive intact, there is good evidence for a sequence of building practices. Earliest to appear are chunks of dried mud bearing the imprints of posts and cords which indicate the erection of walls of bajareque (wattle and daub) (Figure 1.17).
Found in the lower parts of Level 4 of Cut 3, the chunks of mud show that the walls were substantial and solid, with a smooth layer of about 3 cm of mud on the surface and a total wall thickness of more than 10 cm. The smaller stones which appear in Levels 2, 3, and 4 seem to have been used along the base of bajareque walls. They were superseded by the large stone alignments which appear to have supported earthen platforms for buildings. Small areas of packed clay floors survive in all the excavations, presumably interior floors which had some protection from the weather and erosion.
The development from house floors at the same level as the exterior earth surface, separated from the ground outside only by a stockade wall with small stones set along its base, to a floor set on an earth platform about 20 or 30 cm thick supported by a double or triple range of large stones along its margins, with the stockade wall heavily plastered with mud, seems to have occurred during Level 2. Such a raised floor would have been drier than the earlier floors that were level with the ground outside, and the plastered walls would have provided shelter from wind and rain.
The patterns of stone, especially in Cuts 2, 4 (Figure 1.12), and 5, suggest continual building, demolition, reuse, and rebuilding, with stones being thrown aside as walls were taken down then often recovered and reused at the edge of floors or bases of walls or buried in fill as a new platform was leveled. The riverbanks seem to have been favored places for building (more so than even a short distance back from the river), for cultural deposits are deeper nearer the river and the evidence of rebuilding is greater.
The building remains found by Jean Guffroy (1987: Plates 15, 16) at La Vega, Catamayo, south of Azuay on the headwaters of the Chira River system, are the most similar to the Challuabamba remains and particularly to the later lines of large stones, the principal difference being the source of the stones: a quarry at La Vega and the riverbed at Challuabamba. At Cotocollao, in contrast, many postholes were found but no rock alignments (Villalba 1988:41-59). The single good posthole identified at Challuabamba and the numerous pieces of mud plaster with post-molds suggest that the walls there were similar to the quincha (bajareque) walls that Guffroy (1994: Figure 22) shows for Cerro Ñañañique, on the upper Piura River in northern Peru, although those later constructions were not based on stone revetments.
We have tested ten specimens, selected because they seemed likely to provide secure points in the stratigraphic sequence. To reach the closest approximation of a calendar date requires the use of the calibrated calculation, given here in the 68% calibration. They are given in order from earlier to more recent. The first five, TX numbers (University of Texas and University of Washington laboratories), are radiocarbon tests on burnt wood; the second five, AA numbers (University of Arizona),
are accelerated mass spectrometry (AMS) dates on animal bones selected from the storage collection.
The AMS dates were taken from specimens of animal bones selected by Peter Stahl from Cut 3, Sector G, and submitted to the University of Arizona laboratory. They give dates that confirm the integrity of the stratigraphy but run about four hundred years later than the wood carbon series (Tables 1.1 and 1.2).
The AMS dates correlate most closely with four dates obtained by a team from the British Museum led by Elizabeth Carmichael. Although the samples were assigned to the Challuabamba site, the details of the excavations and information on the samples have not been published. The calibrated dates range between about 1100 BC and 950 BC (Carmichael 1981:176). While the TX series suggests a sequence of about 800 years, the AA series suggests about 350 years. This discrepancy is not unusual. For example, among the 55 radiocarbon tests reported by Marcelo Villalba O. (1988: Figure 136, pp. 242-243) for the site of Cotocollao, the thirteen tests on human bone all fall between 805 and 450 BC (uncalibrated), while the wood carbon samples date between 1545 and about 545 BC.
Austin Long and his colleagues (1989:231) of the Arizona laboratory discuss methods for obtaining reliable AMS radiocarbon dates from bone, noting that "bone dates often elicit undisguised skepticism and rebuke." Their analysis of the problem applies particularly to Challuabamba:
Fresh bone containing protein (mostly collagen) and inorganic minerals (primarily apatite), when exposed to moist oxidizing conditions will begin to change in response to its environment. The protein will, with the help of water and micro-organisms, oxidize and degrade to smaller, more soluble molecules. Water flowing through the system will help remove both organic and inorganic components of the bone, leaving a porous, high-surface area structure. This inorganic structure can adsorb organic matter produced in the soil, which would most likely be younger than the bone and have a different 14C content.
This comment comes from the Arizona laboratory which tested the Challuabamba bones. Since the prevailing system of dating to which Challuabamba must be compared derives almost entirely from dates based on wood carbon, it seems most reasonable to make our comparisons on the basis of that system.
A series of excavations by natural levels cannot be expected to show a similar series of levels, since each location will almost certainly have its particular history. The two oldest (and nearly identical) dates both come from Level 4 (the top of that level in Cut 1 and the bottom of the level in Cut 3), which suggests that those deposits represent the same period. Those two dates (calibrated) imply that Level 4 in both locations was deposited around 2000 BC.
The midpoints of the other three calibrated dates (1450, 1425, 1400), all associated with Levels 3 and 2, suggest that those levels were deposited around 1400 BC. All three of those dates are based on wood carbon collected in two excavations about 25 m apart, with some common stratigraphy. All five of our excavations have stone revetments in the upper levels, which can all be dated in the period 1400-1200 BC, judging by these three radiocarbon dates. A diagram of the comparative stratigraphic levels of the excavations with the locations of the carbon samples makes the general chronological picture somewhat clearer (Figure 1.18). Returning to Cut 1, if Level 4 was deposited in the early 1700s BC and the stone revetments were built in the 1200s BC, the 25 cm of Level 3 would represent about 500 years—surely an indication that considerable material had been eroded in ancient times. It is also clear that erosion from the surface has removed all deposits later than about 500 BC, with the black clay of Level 1, containing a mixture of ancient potsherds and modern glass, having been laid down in modern floods. It is significant that there is no evidence of flooding below the surface level; all the earlier levels have friable soil containing little clay but much undisturbed cultural material.