Petrie's Sequence Dates : Amratian = SD 31-37; Gerzean = SD 38-62; Semainian = SD 63-76 (link to University College)

Petrie's goal was to arrange the columns so that the largest frequencies of the different forms concentrated along the diagonal of his matrix of 900 graves and 9 types; Kendall (1971) called this technique "the Petrie Concentration Principle".   The result was a series which progressed across the strips from earliest to latest, anchored at the late end by W-Ware.    Having obtained what he felt was the best ordering of the graves based on his knowledge of C-Ware and W-Ware, Petrie divided the 900 graves into 51 "Sequence Dates" (S.D.) each containing 18 graves.    To these he assigned the numbers 30 through 80.   Wisely, Petrie left sequence dates unassigned at the beginning in case an earlier culture should be discovered, which it subsequently was at Badari by Brunton and Caton-Thompson (1928).

Later Petrie divided the whole range into three groupings which he termed Amratian, Gerzean, and Semainean, names derived from "type" sites where particular ceramic forms had first been identified.   The first two terms were widely adopted by other scholars, and continue in use by some (e.g. Friedman, 1994; Kantor, 1992), though his Semainean period was not, owing to its rather ambiguous definition at the type site of Semaineh and the inability of others to discern such a period at other places (see Kantor, 1944).   Some researchers (e.g. Mace, 1909; Hoffman, 1982) have used the term "Protodynastic" to refer to the period between the end of the Gerzean and the beginning of the Dynastic Age.

It was not long before other scholars began noticing problems with Petrie's shortcuts.   For example, Scharff (1926:73) noticed that the large, globular, wavy-handled jars (Petrie's form W1) co-occurred with one of the supposedly more degenerate forms (Petrie's form 24) at Abusir El Meleq in Lower Egypt.   The globular forms were found in numerous other contexts much later than those assigned to them by Petrie.   As Friedman notes, though, "nevertheless, owing to the geographical distance between Abusir El Meleq in Lower Egypt and the Upper Egyptian Nagada culture, Scharff was unwilling to reject the S.D. system as a whole, but simply stated that Petrie's system did not work well in the cemeteries of the north " (Friedman, 1981:2).
   Later, Baumgartel re-examined the material that Petrie used to create his system and concluded that the wavy-handled jars were not well dated.   The earliest, Petrie's W1 form, had been purchased rather than excavated.   Baumgartel believed that all of the wavy-handled forms were contemporary except for the First Dynasty cylindrical shapes (1955:42; also c.f. Kantor, 1947:77).   Kaiser also found the chronological differences between the globular and upright forms to be extremely small, though it must be stressed here that Petrie never implied that each of his fifty sequence dates was of equal length.   In fact, Friedman points out that since there are more ceramics in later Predynastic graves (see Castillos, 1982, 1983), the later sequence dates probably represent shorter time spans, while the earlier Sequence Dates, dominated by B- and C-Wares, probably represent longer periods (Friedman, 1981:6-7).   Essentially, then, Petrie's system works in its broad outline (Amratian, Gerzean, and Semainan or Protodynastic) but is not very reliable in its details.

 Kaiser's Stufe Dating Method:  As early as 1928, Guy Brunton, using Petrie's Sequence Dating, noticed that certain cemeteries in Middle Egypt exhibited temporal clustering in space.   For example, Badari Cemetery 3800 has the earliest graves in the center and later graves placed to the east and west (Brunton and Caton-Thompson, 1928:51).
   Werner Kaiser (1957) attempted to overcome some of the difficulties with Petrie's system by developing a dating method that takes advantage of the horizontal stratification observed by Brunton as well as typological differences in Predynastic ceramics.   This was an attempt to date the contents of tombs in part by the position of the grave, rather than dating the graves according to their contents.
   His system divides the Predynastic into three main periods, Nagada I, II, and III, and several sub-periods, referred to as stufen (German for "stages").    Using the maps of Armant, Kaiser divided the cemetery into three main periods based on relative percentages of B-Ware, R-Ware, and L-Ware.   Eleven, and later fifteen sub-periods (Kaiser, 1990) were created based on ceramic type clustering within the main periods : Stufe ("Stage") Ia-Ic, Stufe IIa-IId2, Stufe IIIa1-IIIb2.   These sub-divisions relied heavily on the spatial distribution of ceramic types because only 115 forms occurred more than once, and many were found in only two graves (Kaiser, 1957:69; also c.f. Patch, 1991:157-161).   W-Ware, C-Ware, and D-Ware were fitted into the system rather than used to determine it because they were relatively rare forms at Armant.   Kaiser's system in some respects validates Petrie's work, in that his broad divisions, Nagada I through III, generally parallel Petrie's.   Nagada I is roughly synonymous with Amratian, Nagada II with Gerzean, and Nagada III with Semainean or Protodynastic (Kantor, 1992:7).

Many scholars have adopted Kaiser's Stufe dating system since it validates and expands on Petrie's system of three main periods in the Predynastic, while eliminating some of the problems.
   There are, however, several difficulties in the Stufe system.   First, it appears somewhat tautological.   Brunton used Petrie's Sequence Dates to infer temporal drift in Predynastic cemeteries (Brunton and Caton-Thompson, 1928).   Then Kaiser used this spatial patterning to develop his Stufe dates and assign ceramic forms to his subdivisions.

The "space-as-time " assumption that underlies Kaiser's system is based on Brunton's time-as-space observation at Badari, and reaches its greatest extreme when graves that contain no artifacts are dated, presumably on the basis of dated graves nearby (extant data files for the Naga-ed-Dêr cemetery contain such dates for empty graves).   If the assigned Sequence Date of a given grave can vary by as much as twenty sequence dates on either side, as Kaiser (1957:69) noted, then the underlying spatial pattern that Brunton observed may be a result of incorrect grave dates, meaning that any subsequent dating based on the spatial pattern that he observed is also questionable.   And the problem cannot be corrected by re-dating the cemeteries with the Stufe system and then checking their spatial layout, since that would create an even more vicious tautology.
   When Friedman applied Kaiser's system at Naga-ed-Dêr, she found that the relationship between space and grave dates did not appear in Cemetery N7000 (1981:70).   Rather than early graves being confined to only one area, and later graves to another, in a form of "drift," Friedman found several areas with graves from a variety of Kaiser's Stufen.   Thus even Kaiser's system discerns no clear association between space and time in Cemetery N7000.

A second problem associated with the Stufe dating system is the way it has been adopted throughout the rest of Egypt.   The forms that Kaiser assigned to his various stufen have been used like "index fossils" by other archaeologists to re-date graves in other cemeteries and to date the deposits on settlement sites "from Hierakonpolis in the far south (Adams, 1987) to Buto in the north-western Delta (von der Way, 1991)" (Wilkinson, 1996:10).

The implicit assumption of uniform ceramic development that underlies such applications of the Stufe system (or Petrie's, for that matter) is beginning to be replaced by an understanding of the regional character of Predynastic ceramic production and development.   Naville stated emphatically that "for pottery, the only true classification is not chronological : it is geographical, or rather, local" (1914:xi).   Scharff (1928) suggested that there were regional variations in Petrie's C-Ware class, and Finkenstaedt distinguishes three principle regions, each of which "produced a type of C ware peculiar to it."   She infers that "in some cases, individual sites evolved a distinctive local variation on the regional style" (1981:7).   Kaiser himself warned that differences between sequences at Armant and Nagada might reflect differences in development at the two sites (1957:73).   More recently Patch notes that "it would be surprising if all Predynastic sites followed the same development " (1992:192).
   Finally, Friedman's analysis of ceramics from the settlements at Hierakonpolis, Nagada, and Hemamiah found that "previously suspected, but poorly defined, regional differences within the Amratian settlement ceramic assemblages in each of the geographical regions were clearly apparent from an examination of the utilitarian pottery or kitchen wares at each site " (1994:865).   Tempering exhibits the largest regional differences, but Friedman also notes variation in manufacturing technique, surface treatment, and shape.   She also identifies "minor, but possibly regionally significant morphological differences" in the B-Wares and P-Wares (1994:871), sufficient to rule out the notion that the ceramics came from a single production center, and further (she) expands on Finkenstaedt's observed regional differentiation in the C-Ware class.   The results of Friedman's exhaustive study parallel those of Holmes' (1989), who documents differences in the lithic assemblages in the same areas.   These distinct regional differences make it difficult to justify a normative, "index fossil" approach to dating the Predynastic.

A third problem with the Stufe system is internal to it.
   A number of researchers have noted difficulties in the transitions between Kaiser's Stufen Ic and IIa.   Based on radiocarbon dates, Hassan (1988:138) suggests that Amratian/Nagada I encompasses Stufen I-IIb.   Kemp (1982) reached similar conclusions, based on his computer seriation (discussed below).   And Hendrickx (1993:5) proposes that the division between Nagada I and Nagada II should be placed between Stufen IIb and IIc.   Friedman surmises that 

...it remains to be seen if all the divisions of the Nagada I phase proposed by Kaiser are valid.   The distinction between Nagada Ia and Ib is particularly questionable, given the small number and restricted geographic range of the graves from which these phases were distinguished.

Kaiser (1957) divided the Nagada I phase of the mortuary complex into three subphases (Nagada Ia-c) on the basis of supposed stylistic developments in the pottery.   Only the division between Nagada Ic and the transitional Nagada IIa phase was demonstrated by the spatial distribution of the graves within the cemetery at Armant [1994:925, footnote 6].

Problems also are noted in the transition between Kaiser's Nagada II and III periods.   Mortenson (1991:15) pointed out that the archaeological evidence for the differences between Stufen II and III are quite unclear, and Hendrickx agrees (1993:18).   Wilkinson's work indicates that at Matmar, Mostagedda, Mahasna, and Hierakonpolis "the transition from Naqada II to Naqada III seems to occur between Stufen IIIa1 and IIIa2.   Furthermore, at Matmar, Mostagedda, and Mahasna there is no division within phase 2 which could correspond to a separate Stufe IIIa1" (1996:63).   The same is true of Hierakonpolis 2.   Wilkinson concludes that these results call Kaiser's method into question on a number of points, and support the approach of formulating separate site-based local chronologies which can then subsequently be correlated, rather than trying to devise an over-arching chronological scheme for the whole of Egypt.

 Multivariate Analysis:  The tautological nature of the assumptions that underlie traditional approaches to Predynastic chronology, the regional nature of the ceramic assemblages, and the internal difficulties in the Stufe dating system suggest that a fresh approach is needed.   Barry Kemp (1982) used a multi-dimensional scaling (MDS) program, HORSHU, to test the sequence of graves in Cemeteries A and B at el-'Amrah, Armant 1400-1500, and el-Mahasna.
   MDS, like correspondence analysis, is a form of "dual scaling," that can arrange both cases and types in chronological order (see Nishisato, 1980).   HORSHU, developed by Kendall (1971) specifically to automate the process Petrie used to create his original seriation, can handle no more than one hundred graves because of the computationally intensive nature of non-metric MDS (see Shennan, 1990:281-283 for a brief discussion of MDS).
   Kemp condensed the Petrie corpus into 43 types for his analysis, and distinguished three clusters of graves based on their ceramic types.   The clusters are interpreted to confirm Petrie's basic division of the Predynastic into Amratian, Gerzean, and Semainean (and the major divisions in Kaiser's Stufe dating method), although Kemp's division between Amratian/Nagada I and Gerzean/Nagada II appears somewhat later than Petrie's.

Since Kemp's initial computer seriation, others have used similar methods :
1) Seidlmayr (1990) employed seriation to identify local chronologies at individual sites, and then correlated the results.
2) Wilkinson (1996) used the Bonn Seriation Program (a form of correspondence analysis, see Scollar, 1993) to develop independent chronologies at 8 different Predynastic and early dynastic sites.   He subsequently correlated them by connecting the various phases from the eight sites to the Stufe system.   However, his results were hampered by lack of external dating controls, such as those provided by radiocarbon dating.
3) Savage (1995, 1997) used correspondence analysis to date 143 graves from Cemetery N7000 at Naga-ed-Dêr, and concluded that they should be grouped into four use phases that span parts of the three Predynastic periods (essentially late Nagada I, all of Nagada II, and early Nagada III).   Like Friedman's earlier work, it appeared that there were graves from all four use phases in most parts of the cemetery.   The results suggested that the spatial clusters observed in the cemetery were the burial grounds of distinct descent groups.   Twelve radiocarbon dates from graves on Savage's seriation curve helped establish an absolute time range for the N7000 cemetery (Savage, 1998; see below).

To summarize, there are clear problems with traditional chronological methods.   While the three methods used in the past all agree on a tripartite division of the Predynastic, they disagree on further subdivisions, and attempts to rectify problems in any method by resorting to another and then re-dating graves in the first method only (to) create tautological nightmares.
   Furthermore, the regional variation in Predynastic ceramic assemblages is generally not considered (but see Wilkinson, 1996 for a notable exception), meaning that chronological schemes worked out in one place tend to be applied uncritically in other places.   The whole question of "space as time" rests on an assumption that needs to be verified by an external, independent dating method.   O'Shea stresses the importance of independent reference points : "precise dating, particularly dating that is independent of the material culture and behavior being examined (as in radiometric dating), provides a critical underpinning for any serious anthropological research into the past" (1996:16).   The only way out of the current muddle is to develop a radiocarbon based chronology.

 IV. Dating the Predynastic with Radiocarbon Methods:  Radiocarbon dates have been obtained from many Predynastic and dynastic sites (for some recent compilations see Close, 1980, 1984, 1988; Derricourt, 1971; Hassan, 1984a, 1985; Kantor, 1992).   Hassan (1988:138) tentatively suggests a Predynastic radiocarbon chronology as follows : Early Predynastic (EP), 4000-3900 BC; Middle Predynastic, 3900-3650 BC; Late Predynastic, 3650-3300 BC; Terminal Predynastic, 3300-3050 BC.

However, the radiocarbon dates from Predynastic contexts are not without problems; most absolute dates are from settlement sites, not cemeteries.   Because of the history of Egyptian archaeological research there are two parallel chronological schemes for the Predynastic period.   The first, based on relative methods, is exemplified by Petrie's (1901), Kaiser's (1957), and Kemp's (1982) work with ceramics from cemetery contexts (mostly whole vessels).   The second is based more on sherds for its ceramic typology, and on radiocarbon dates obtained mostly from settlement sites (and tombs from the later dynastic period).

D-Ware (Naga-ed-Dêr)

   However, the ceramic assemblages in settlement sites are different than those in cemeteries (e.g., decorated, marl clay pots, termed "D-Ware" by Petrie, are very rare in settlements [see Hoffman, 1987] but fairly common in cemeteries; low-fired, coarse clay pans, called "Bread Molds," are hardly ever found in cemeteries but are abundant in settlements).
   The two systems are not well connected.   Predynastic cemetery sites remain largely undated by radiocarbon methods, and the possibility exists that cemetery chronologies may not coincide with those from settlements.

Cemetery contexts contain by far the largest body of excavated material from the Predynastic in Upper Egypt.   They are virtually the only source of whole vessels and the best source of organic remains that are specifically associated with individual, dateable events (the interment event).   Furthermore, other Near Eastern chronological schemes tie into the Predynastic and Dynastic Egyptian dates primarily through the presence of ceramics recovered from cemeteries.   Thus, an ideal solution to the chronological problems in the Predynastic and the Bronze Age would be to develop a radiocarbon chronology based on cemetery data.   Such a goal is clearly in line with Bruins and Mook's (1989:1024) assertion that, "many more samples from ancient Egypt ought to be investigated, as urged in the 1969 [Olsson, 1970] Twelfth Nobel Symposium."
   At present, there are very few dates from Predynastic cemeteries, less than 100 radiocarbon dates have been published from Upper Egypt; fewer than 20 are from mortuary contexts.   Five of these are based on older methods, and they were calculated on materials gathered from a number of contexts.   For example, Libby's sample C-810 consisted of about 3 ounces of human hair, derived from four different graves at Nagada with different sequence dates in Petrie's scheme !   In addition, Libby's dates have wide standard deviations (about 300 years), so the 95% confidence intervals extend over 1,200 years !
   As a result, many of the extant dates from Predynastic cemeteries are not as precise as we would like.   Three dates from graves at Hierakonpolis have been secured.   Sample BM-1127A from Tomb 100 at Hierakonpolis is clearly problematic at 12,900 ± 120 b.p. (Burleigh, 1983:364).   Another Tomb 100 date (BM-1127B) has a calibrated one-sigma range of 235 years at 3900-3665 BC (Burleigh, 1983:364), but its wide error estimate is not useful for a more precise radiocarbon chronology or for a comparison to the tomb's ceramic inventory.   At Hierakonpolis Locality 6, Tomb 1 was nearly empty when excavated; many of the sherds were recovered from an adjacent spoil pile presumed to have been left by plunderers (Hoffman, 1982:41).   Comparing the radiocarbon date to the original tomb contents is not possible.

Hassan and Robinson (1987) note that "the existing corpus of radiocarbon measurements for ancient Egypt is, with a few exceptions, not fully satisfactory."   Since most of what we know about the Predynastic period in Upper Egypt comes from cemetery excavations, a radiocarbon based chronology of the cemeteries is needed.   Such a chronology could be used to help establish a concordance between cemetery and settlement contexts.   In addition, a cemetery-based radiocarbon series would allow us to test hypotheses related to ceramic chronology and cemetery use and development.

It is not easy to do this, though.   Most of the known Predynastic cemeteries were excavated in the late 19th or early 20th centuries, long before radiocarbon techniques were available.   Hence, it is often necessary to rely on materials collected in some cases more than a century ago.   Many excavators did not collect samples of materials suitable for radiocarbon analysis, or the extant materials are not well provenienced.   Some otherwise well-documented materials, which would have been suitable for dating, were treated with petrochemically based preservatives in the days before radiocarbon dating, thus ruining them for assays.   As a result, there has been considerable doubt about whether a radiocarbon chronology could be recovered from Predynastic cemeteries.

What is needed is a Predynastic cemetery dug in a well-controlled manner, where there is an abundance of uncontaminated materials dateable by radiocarbon methods, and thorough documentation.   Not many Predynastic cemeteries fulfill these requirements, but, happily, Cemetery N7000 at Naga-ed-Dêr does.
   The excavators, Albert Lythgoe (1965) and G.E. Smith, took careful field notes, made drawings of the entire cemetery and virtually every grave, provided provenience information on artifacts, and took over 1,500 large-format photographs.   In addition, they recovered many organic objects.   These are curated at the Hearst Museum of Anthropology, U.C. Berkeley.

 Results to Date in Cemetery N7000:  As a result of Friedman's (1981) and Savage's (1995, 1997, 1998) efforts, considerable progress has been made in establishing a firm chronological foundation in Cemetery N7000.
   Friedman (1981:Appendix III) estimated Sequence Date ranges and Stufe dates for graves in the cemetery with ceramics.   Podzorski (1995, personal communication) estimated dates for graves that did not contain ceramics, using their proximity to those that did, based on the questionable assumption that "space equals time."   Correspondence analysis (CA) was used to analyze 143 graves, each with a minimum of three ceramic vessels, and at least two ceramic ware "types," to establish four use phases in the cemetery (Savage, 1995:98-104).
   CA reduces variability in the data under analysis by producing a set of scores along "Eigenvectors."    A CA result captures temporal variation when a scatterplot of points on the first two Eigenvectors assumes a linear to horseshoe shape (Bech, 1988; Bolviken, et al. 1982; Hojlund 1988; Holm-Olsen 1988; Madsen 1988).    The CA seriation plot from Cemetery N7000 (Figure 1) matches such a shape, capturing temporal variation well.

Figure 1 : Correspondence analysis (in red, C-14 graves)

  A number of graves on the seriation curve contained organic materials (Figure 1, numbers); of these, sixteen samples from fifteen graves were submitted for radiocarbon analysis to the NSF Arizona AMS Facility (Figure 1, red numbers).   These materials were donated by the Hearst Museum of Anthropology to test whether the organic items curated there from Cemetery N7000 were still viable for radiocarbon dating, and to help fix Savage's seriation curve in absolute time.

Twelve radiocarbon dates were obtained (Table 1); four of the samples were not of sufficient size after cleaning to be dated, but larger samples are dateable.   Using Bayesian methods to combine dates calibrated with the OxCal program (Bayes, 1763; Iverson, 1984; see Bronk-Ramsey, 1995), Savage concluded that there were four radiocarbon-based use phases in Cemetery N7000 (Table 2), ...

'sixteen samples from fifteen graves submitted for radiocarbon analysis'

... indicating that the cemetery was most likely in use between about 3800 and 3090 BC.   These phases correspond remarkably well with the four use phases developed independently through seriation (Savage 1998:242-243).
   Then, by recalibrating a series of 58 published dates from Upper Egypt and using the same methods to combine the results, Savage (1998: Table 4, Figure 5) suggested that each of Hassan's radiocarbon based periods in the Predynastic could be divided into two smaller phases (EP I and II, MP I and II etc.).   A comparison of the four use phases from Cemetery N7000 to the recalibrated phasing from Upper Egypt showed remarkable correspondence (Savage 1998:243-247) : Naga-ed-Dêr Phase 1 corresponds to the Middle Predynastic II, ...

Extended Tentative Chronology (Hassan)

... Phase 2 to the Late Predynastic I, Phase 3 to the Late Predynastic II, and Naga-ed-Dêr Phase 4 dated to the Terminal Predynastic 1 (thus helping fill a/(the) gap in Hassan's radiocarbon chronology from Upper Egypt).

"2 Signas" are used for chronology purpose

This initial test of the viability of organic materials from the Predynastic cemetery at Naga-ed-Dêr showed that the objects curated at the Hearst Museum are capable of producing reliable, and vital AMS dates.   Preservatives and contamination do not appear to be a significant problem; initial pretreatment is able to remove spurious wood wool or cotton fibers from the original packing materials used in 1904, thus allowing the samples to be dated.   These results contribute to a badly needed, calibrated radiocarbon chronology for Predynastic cemetery remains from Upper Egypt.   They constitute one of the first series of AMS dates to be run on Predynastic cemetery remains, and the results fit well with the recalibrated chronology of the Upper Egyptian sequence derived mostly from settlements.   This preliminary work sets the stage for an expanded program of radiocarbon dating with Cemetery N7000 materials.

 V. Proposed Work: Artifact inventories from the various graves in the phases discerned at Naga-ed-Dêr can be compared with those from settlement sites, judging contemporaneity on the basis of their absolute chronologies rather than their relative dates; thus allowing us to explore more fully questions related to different ceramic inventories in cemetery and settlement sites from the same periods.
   Further tests on additional samples will allow the dates published in Savage (1998) to be refined, and allow questions relating to different methods of relative ceramic dating to be addressed.   For example, specific ceramic types common in the cemetery can be assigned absolute date ranges rather than seriation-based stages.   In addition, a complete radiocarbon-based chronology will allow other questions about social organization and the use of mortuary space at Naga-ed-Dêr to be addressed.   Savage (1995, 1997) suggests that burial clusters in the cemetery represent specific social groups in Predynastic society rather that clusters of burials dating to similar time periods, and additional dates will clarify these relationships.   Accordingly, this proposal seeks support to obtain 100 samples from the hatched graves shown in Figure 2; they will be submitted to the NSF Arizona AMS Facility.

Figure 2 - In red, graves still containing Black-topped Red wares/sherds (BkT).

Since Predynastic B-Ware contains a high concentration of atmospheric carbon (because the black top is produced by inverting the vessel in organic material during firing) these sherds are directly dateable, as are many other organic samples in the cemetery.   The Hearst Museum preserves some 210 dateable samples from 135 different graves, not counting human bone (Savage 1995:Tables A6 and A12; Figure 2).   Results already obtained from the ceramic chronologies and initial radiocarbon dates in Cemetery N7000 suggest a number of important hypotheses which can be tested by dating these materials :

1) The use life of Cemetery N7000 can be divided into four phases. The twelve dates obtained so far suggest, but are not sufficient to conclusively demonstrate, the four-phase division developed through seriation.   Through the date combining routines in the OxCal program (Bronk-Ramsey, 1995) the new dates can be grouped, and compared to the groups already established.   Four radiocarbon-dated use phases will tend to support Savage's (1995, 1997) seriation, but more sub-phases will force a re-interpretation of temporal trends in the cemetery.

2) The Cemetery's temporal development supports the notion that "Space equals Time" in Predynastic cemeteries.   Once the dates are grouped, to test Hypothesis 2, they can be plotted on the Cemetery N7000 map.   By so doing, it will be visually clear whether graves of similar date cluster together in space.   In addition, G-square tests for significant association between temporal groups and spatial clusters discerned by Savage (1995:81-85, 104-107) will test Hypothesis 2.   If G-square tests fail to reject the null hypothesis (that "space does not equal time") by failing to show a significant association between burial clusters (see Savage 1995, 1997; Figure 3) and temporal groups, then it suggests that interpretations of other Predynastic cemeteries that rely on this previously untested assumption must be re-evaluated.   On the other hand, if the null hypothesis is rejected, then it will be necessary to re-evaluate the cultural significance of the spatial clusters visible in Cemetery N7000.   In addition, radiocarbon dates run on samples from adjacent graves (see Figure 2) will allow this hypothesis to be tested directly; since adjacent graves, or nearly adjacent graves, should date to the same period.

3) The Predynastic radiocarbon chronology from Upper Egypt is divisible into four main periods, each with two phases.   As suggested by Hassan (1988) and modified by Savage (1998), the current radiocarbon chronology assumes this configuration.   Adding 100 additional dates to the available data will more than double the number of radiocarbon dates from the Upper Egyptian Predynastic; after calibrating and grouping with the OxCal program, the expanded data will allow a revised periodization to be developed.   If the grouped results indicate fewer, larger periods, then Predynastic ceramic chronologies that suggest shorter developmental phases will not be supported by the radiocarbon dates.   However, it is conceivable that additional sub-periods will be discernable with 100 new dates, which will tend to support current ceramics-based chronological frameworks.

In addition to these formal hypotheses, there are several ideas which may be examined through a dating program conducted on organic material and B-Ware pottery: 

1) Dates run on B-Ware sherds and organic items such as seeds from the same graves will address the issue of whether ceramics were made specifically for the mortuary ritual.   Vessels with no use wear would appear to have been made specifically for funerals, but if their dates appear earlier than dates from seeds or matting (which probably date within one harvest cycle to the time of burial), then it will be possible to suggest that certain vessels were curated before burial.

2) One hundred radiocarbon dates will allow us to establish manufacturing date ranges for many different types of Predynastic pottery from the Petrie Corpus (Petrie, 1928) and other sources [e.g. Armant (Mond and Meyers, 1937), Badari (Brunton and Caton-Thompson, 1928), Harageh (Engelbach and Gunn, 1923), Mahasna (Ayrton and Loat, 1911), Matmar (Brunton, 1948), Mostagedda (Brunton, 1937), Petrie's corpus of Protohistoric pottery (Petrie, 1953), and The Archaeological Survey of Nubia (Reisner, 1910)] by combining radiocarbon dates from a series of graves with the same ceramic types.
   This should provide significant information for cross-dating graves that do not contain enough ceramics for seriation or organic materials for radiocarbon dating, and will provide a series of temporal markers for comparison with other Predynastic cemeteries and settlements.

3) Hassan (1988) estimated the end of the Predynastic era by filling in a gap between his Late Predynastic and radiocarbon dates obtained from the 1st Dynasty tomb of Aha (Hassan and Robinson, 1987).   Three of the twelve dates already obtained from Cemetery N7000 have helped fill the gap (Savage 1998: Table 3); additional dates should clarify the situation, and provide needed documentation for the critical transition to the Dynastic Period in Egypt.

4) The development of a radiocarbon-based chronology at Naga-ed-Dêr will provide vital data for dating other sites in the region, and important information to compare with chronologies that can be developed in other regions.   By using this comparative method, assumptions based on the uniformity of Predynastic ceramic chronologies across multiple regions may be explored for the first time.

5) Developing a detailed radiocarbon chronology for Naga-ed-Dêr allows many important social questions to be addressed in a more informative way.   Savage (1995, 1997) suggests that the spatial clusters observed in Cemetery N7000 (Figure 3) represent the burial grounds of separate descent groups, which competed with each other in various ways.   Tomb robbing, architectural elaboration, control of various exotic trade items, and developments in the mortuary ritual appear to be cluster-specific in the cemetery.   These results indicate that a substantial degree of inter-group competition existed at the local level during the Predynastic, and help frame an understanding of the development of the Egyptian state based on circumscription (Carneiro, 1970) and individual actions taken by competing groups or individuals (see Marcus and Flannery, 1996).   By refining and expanding the number of dated graves from Cemetery N7000 many of these questions can be addressed with renewed vigor.

The N7000 cemetary at Naga-ed-Dêr

 VI. Methods and Procedures: 

 Selection of Samples: Much of the material available for dating consists of vegetable matter contained in pots included as grave offerings.   These comprised mainly seeds or chaff, sometimes charred.   Grain or seed offerings included in graves are thought to have originated as either the remains of a funeral meal consumed at the grave site, or as food offerings intended to accompany the deceased on their journey to the afterlife (see Adams, 1988).   It is likely that the age of the materials themselves corresponds very closely to that of the interment event (probably within one harvest cycle).

Other objects available for dating include human hair, matting, and leather or cloth from articles of clothing (Table 3).

Table 3 : Potential new dateable items, including BkT wares.

   Of these, the human hair is perhaps most problematic in terms of its connection to the burial event.   The excavators observed numerous burials which included balls of human hair as grave offerings, such as from graves 7055, 7130, 7491 and 7596 (samples of these are present in the Hearst Museum collections).   Grave 7491, for example, included about 50 small coils and balls of fine brown hair, along with straight hair, mostly corn-yellow but with traces of the same brown in it.   The straight hair corresponded to that of the hair found on the cranium.   Lythgoe remarked that this provided "the solution of the balls and coils of hair found occasionally in other graves of this cemetery.   It was the hair of the individual at an earlier age which had been preserved and was finally buried with the individual" (Lythgoe and Dunham, 1965:310).   Thus, the samples of human hair do not necessarily correspond to the burial episode itself, but to an earlier event.   Radiocarbon age determinations based on these hair balls would be expected to read somewhat earlier than the interment.   One standard deviation of the radiocarbon date likely encompasses any discrepancy in age between the hair and the interment, but I will not use hair for dating unless it is absolutely necessary.

Matting and clothing from the graves also are likely to correspond closely enough to the date of the interment; they are clearly somewhat earlier than the burial event itself, but probably not so much as to overly bias the radiocarbon results.

Thus, there is a clear "order of preferability" in sample selection.   The items most directly related to the funeral event are seeds or grain stored in pots deposited in the grave, or stomach contents of the bodies; every effort will be taken to sample these items first, obtaining as wide a distribution in the cemetery as possible.   Matting that wrapped the bodies is also closely related to the interment, and graves with matting samples will be chosen as second priority.   Clothing probably dates closely to the time of interment, and ceramics may have been manufactured specifically for the mortuary ritual (see Hoffman, 1989).   Ceramics will be sampled specifically to judge whether they date closely to the time of other dated items in specific graves.

    The 210 dateable items from 135 graves are many more than this proposal seeks to date.   The final selection process will be performed in collaboration with the Hearst Museum staff, and the actual samples will be taken by the Museum staff.   Every effort will also be made to sample items whose specific location in the grave was shown by field drawings (see Savage, 1998: Figures 2 and 3; Figure 4).

Because there are a number of different hypotheses and other questions that can be addressed through the dating process, care must be taken to select samples whose dating will accomplish as many goals as possible.   Since one goal is to test the dates of graves in close proximity, clearly some groups of samples must be collected from adjacent, or nearly adjacent graves.   However, another goal of the research is to date graves in every part of the cemetery, so samples must be collected from widely dispersed graves as well (see Figure 2).   The final determinations will be made based on the condition and quantity of dateable items.   All samples will be taken in accordance with standard archaeological practices (e.g. Bowman, 1990; Dancey 1981:163-164), wrapped in aluminum foil, and delivered to the NSF Arizona AMS Facility.   Only items with secure proveniences will be sampled; that is, only items described in Lythgoe's field notes, and which have an existing catalog card at the Museum.   Sample sizes will be in accordance with requirements published by the NSF Arizona AMS Facility (NSF-UA, 1994).

 Analysis Procedures:  Once the raw dates are obtained from the University of Arizona AMS facility, they will be calibrated and grouped using Bronk-Ramsey's OxCal program.   "Calibrated ¹⁴C dating ...enable linkages, if not possible otherwise, of historical chronologies with archaeological strata, artifacts, and environmental history..." (Bruins and Mook, 1989:1026).
   In 1986 a high-precision calibration using Irish Oak was published (Stuiver and Kra, 1986).   The Irish Oak calibration curve has an estimated error of less than twenty years, but only extends back to about 4000 BC, so dates from the early part of the Predynastic period are often outside its range (Shaw, 1985).   A more recent calibration curve (Stuiver et al., 1993) will be used here; it is as precise as the Irish Oak curve, but extends back considerably further.

In recent years it has been recognized that Bayesian methods of radiocarbon calibration are preferable to earlier methods (Bowman, 1994:841; also see Bowman and Leese, 1995).   Earlier calibration routines (e.g. Stuiver and Reimer, 1986) used an intercept method, but more recent programs (Bronk Ramsey, 1995; Niklaus et al., 1992; Stuiver and Reimer, 1993; Van der Plicht, 1993) have adopted a Bayesian routine which uses "the eminently reasonable a priori assumption that, in the absence of any other information to the contrary, all calendar ages for the event being dated are equally likely" (Bowman, 1994:840).   Furthermore, because Bayesian date combination techniques narrow the probable range of a group of dates rather than expanding the range as the Long and Rippeteau (1974) date averaging method did, it is possible to provide two-sigma date ranges that are frequently shorter than the one-sigma ranges calculated with earlier methods.   The specific techniques used will follow those reported in Savage (1998).

 VII. Broader Impacts of the Study: Because of its importance to Near Eastern archaeology as a whole, the chronology of Ancient Egypt has assumed critical importance.   But, because early chronological systems were based on flawed methods or incomplete data, their resulting dates are subject to question.   Particularly questionable are dates related to the all-important beginning of the 1st Dynasty in Egypt.   Earlier dates for the Gerzean/Nagada II and Protodynastic/Nagada III periods are vital for establishing the beginning and duration of the Early Bronze Age, while the accession of Egypt's 1st Dynasty helps determine the break between Early Bronze II and III in the southern Levant.   A revised and expanded radiocarbon chronology from Egypt will allow more precise estimates to be made for these crucial developments in the larger Near East.

In addition, an effective cemetery-based radiocarbon chronology from Upper Egypt will help bridge the unfortunate gap that currently exists between cemeteries inadequately dated by earlier methods and settlement sites dated by radiocarbon techniques.   By filling this gap, it will be possible for the first time to directly compare artifact assemblages from cemeteries and settlements, and help determine the meaning of the differences we can now dimly discern.   Furthermore, dating Cemetery N7000 at Naga-ed-Dêr will provide an important springboard for exploring important questions about social organization in the Predynastic, and lead to revised insights about the origins of the Egyptian state.

Beyond its clear importance for establishing a badly-needed radiocarbon-based chronology for the Predynastic that can be cross-referenced to other culture areas in the ancient Near East, the results of this study can contribute to the education of archaeological field and laboratory students.   The raw dates can be used to instruct students in methods of calibration, date combination, and statistical analysis through a course in Archaeometry.   Using these data, students may experiment with different date combinations and add additional dates to the database as they become available.

It is an unfortunate fact that many Near Eastern excavations fail to recover dateable samples, mainly because many archaeologists working in the region have relied on relative ceramic chronologies exclusively.   By conducting this research, it is hoped that the utility of radiocarbon-based chronologies will become more apparent in the field, and hence, encourage other archaeologists to collect more samples.   As local archaeologists are trained, a greater appreciation for the need for absolute chronologies that are comparable across regions will lead to more samples being taken, and more published dates by scholars who are currently under-represented in the field.   Since there is no AMS facility in the Near East, developing a radiocarbon-based chronology will necessitate increased cooperation between local, Near Eastern archaeologists and laboratories in Europe and the United States, thus helping to forge closer ties within the scholarly community.

 VIII. Dissemination of Results:   The initial test of sixteen samples from the Hearst Museum was published by the Journal of Archaeological Science  ( Savage 1998).   The results of this larger study will also be submitted to the JAS, and more popular versions of the study will be developed and sent to other magazines, such as Archaeology.
   A longer monograph is anticipated that will present the complete publication of the dates, including a careful comparison of these dates with others from the Predynastic, and the impact the results have on dates of specific ceramic forms and their cross-dating to other cemeteries.   The raw data will be made available on the World-Wide Web, along with a summary of the results; these data will accompany data from Cemetery N7000 that I have already placed on my Web site at archaeology.asu.edu/Jordan/download.html.   Finally, it is anticipated that an exhibit will be developed for display at the Hearst Museum, demonstrating how the dating was accomplished, and explaining the results to the public.

 top

Adams, Barbara

Albright, William F.

Ayrton, Edward R. and W. L. S. Loat

Bard, Kathryn,

" (26°00'N, 32°22'E; Abadiyeh) - Predynastic settlements have not been well preserved - an absolute chronology of the Nagada culture relative phases has been collated by Hendrickx (1996:64) based on Hassan's radiocarbon dates - a long gap in dates between the Nagada Ic and Nagada IId phases (Görsdorf et al. 1998). - Hendrickx.s Absolute Chronology : Nagada Ia-IIb approximately 3900-3650 BC - calibrated using OxCal v3.8 - Nagada culture potsherds and lithics, ranging in date from about 3700 BC to 3200/3100 BC. "

Baumgartel, Elise J.

Bayes, T.R. 1763 An Essay Towards Solving a Problem in the Doctrine of Chances. Philosophical Transactions of the Royal Society 53, 370-418.

Bech, Jens-Henrik

Bolviken, Erik, Ericka Helskog, Knut Helskog, Inger Marie Holms-Olsen, Leiv Solheim and Reider Bertelsen

Bourriau, Janine; Paul Nicholson. Pamela Rose.

Bowman, S. G. E.

Bowman, S. G. E. and M. N. Leese

Braun, Eliot

" EB I could be lengthened by up to 450 years - The end of EB I placed early in the reign of Dynasty I's third king Horus Djer.. - The Nahal Mishmar Cave seems dated by its mats, in which the hoard of Chalcolithic objects was wrapped. - BM 229, Saqqara tomb 3503 (Djer) 4520 ± 65 3351-3306 (Hendrickx, 1999) "

Breasted, James Henry

Bronk Ramsey, Christopher

Bruins, H.H. and W.G. Mook

Brunton, Guy

Brunton, Guy and Gertrude Caton-Thompson

"The Badari area is located on the Nile east bank, roughly 30km to the south of Assyut, where Badarian (4800 - 4000 BC) settlements and graves extend 33 km southwards in the Mostagedda and Matmar region. There are 41 cemeteries and 40 settlement locations in the low desert that overlooks the floodplain beneath the high desert limestone plateau cliffs. ... The best known pottery of this period is the black-topped and burnished wares which was carried on into the Naqada Periods." (touregypt.net)

Burleigh, R.

Carneiro, Robert

Castillos, Juan

Close, Angela

Crichton, Michael

Dancey, William S.

Derricourt, Robin M.

Emery, Walter B.

Engelbach, R. and Gunn, B.

Finkenstaedt, Elizabeth

Friedman, Renée

Görsdorf, Jochen; Günter Dreyer and Ulrich Hartung

" Sixteen newly ¹⁴C-dated samples were mainly taken from remains of wooden roofs and coffins, or in the case of the earliest tombs from mattings. - The dating results are but 100-150 yr earlier than the so far established historical chronology. - tomb U -j (Naqada IIIa2) contained the earliest evidence of writing. - A total of 22 dates are presented : tomb B19/ Dyn.I.1 (3350-3300 cal BC); tomb U-503/ Nagada Ib (3710-3510). - The overlapping range of two dated samples from the tomb of Qa'a give an age between 2920 and 2910 cal BC. - a more accurate form of the general chronology (Hassan 1988) - We can date Naqada IIIa2 to the middle of the 34th century BC."

Hassan, Fekri A.

" Hassan (1984a: 683) concludes that dates from three early Nagada sites provide a midpoint estimate of 3760 +/- 40 B.C., and dates of the Nagada II zone of south Town provide a range of 3600-3300 B.C. A calibrated radiocarbon date from a Nagada III tomb at Hierakonpolis of 3025 +/- 80 B.C. (WIS-1180) has been published by Hoffman (1982: 42) ."

Hassan, Fekri and S. Robinson

Hassan, Tassie & van Wetering

Hayes, William C.

Hendrickx, Stan.

Hendrickx, Stan ; Renée Friedman, Krzysztof M. Cialowicz

Hendrickx, S.: R. Friedman and F. Loyens.

" How the black top was produced remains a major question ... black-topping was a secondary, post-kiln, process."

Hoffman, Michael Allen

Hojlund, Flemming

Holmes, Diane

Holm-Olsen, Inger Marie

Iversen, Gudmund R.

Kaiser, Werner

Kantor, Helene J.

Keita, S.O.Y. (National Human Genome Center at Howard University)

Kemp, Barry J.

Kendall, David G.

Kitchen, K.A.

Lerner, Gerda

Libby, Walter

Long, A. and Rippeteau, R. B.

Lythgoe, Albert M. (founding curator of the Metropolitan - 1868-1934)

Lythgoe, Albert M. and Dows Dunham (ed.)

Mace, Arthur C.

Madsen, Thorsten

Marcus, Joyce and Kent V.Flannery

Masahiro Baba & Masanori Saito

Millard, A. & Wilkinson, T.

Mond, Sir Oliver and Robert H. Meyers

Mortenson, Bodil

Naville, Edward

Newton, Steven,

Nicholls NSFUA

Niklaus, T. R., G. Bonani, M. Simonius, M. Suter, and W. Wölfli

Nishisato, Shizuhiko

Olsson, I. U. ed.

Patch, Diana Craig

Payne, Joan Crowfoot

Petrie, W.M. Flinders

Petrie, W.M. Flinders and J.E. Quibell

Podzorski, Patricia

Reisner, George A.

Reisner, George A. [Hearst Egyptian Expedition, University of California, Berkeley]. and Mace, A.C., and Lythgoe, Albert M., and Dunham, Dows.

" George Reisner excavated in four Predynastic cemeteries: El-Ahaiwah, Ballas, Naga ed-Deir, and Shurafa."

Savage, Stephen H.

" In the earliest phase of the cemetery, trade was directed toward the south. In the second phase evidence of outside trade vanishes at about the time of the Chalcolithic collapse in the Southern Levant. In the third phase, trade is oriented toward Syria and Mesopotamia. ... The political unification of Egypt may be the result of the efforts of Upper Egyptian chieftains to control the lucrative trade routes with Southwest Asia."

" Reliable radiocarbon dates from these contexts are especially valuable because of their ability to tie relative ceramic chronologies to more recent radiocarbon chronologies from Predynastic settlement sites. ... A comparison of these results with other radiocarbon dates from Upper Egypt (Hassan) establishes synchronisms between the N7000 cemetery and settlement sites, while helping to fill in gaps in the known radiocarbon chronology."

" This paper reports results from 30 samples in the current batch and 12 samples from an earlier set of submissions. ... The initial results are encouraging: about 65 ceramic types now have date ranges assigned to them, and the whole suite of dates will allow nearly 500 types to be dated. "

Scharff, Alexander

Savage, Stephen H. & Harrison, T.P.

Stephen H. Savage, Donald R. Keller, and Kurt Zamora

Scollar, Irwin

Seidlmayr, S. J.

Shaw, Ian M. E.

Shennan, Stephen

Stager, Lawrence

Stuiver, M. and R. S. Kra eds.

Stuiver, M., A. Long, and R. S. Kra eds.

Stuiver, M. and P.J. Reimer

Trigger, Bruce G.

Van der Plicht, J.

von der Way, Thomas

Wilkinson, Toby A.H.

Ward, Cheryl

" Grave enclosures in the Predynastic Naga-ed-Dˆer cemetery (Lythgoe & Dunham 1965; phase dates in Savage 1998) demonstrate an increased standardisation and complexity of woodworking technology. "

 top

Home | Vita | Projects | Abstracts | Lab Book | Cemetery Data| E-mail| Breakout| NSF Proposal