6.1.4 Discussion of site assemblages

A number of sites of varying kinds were revealed through the four seasons of fieldwalking. The vast majority of site assemblages consist of undiagnostic Gallo-Roman material, but in a number of cases, particularly the larger groups, the chronologies of the material can be refined somewhat. Most Gallo-Roman sites in the region evidently had access to a range of common coarsewares and semi-finewares. Assemblages from Autun provide what are likely to be relatively close chronological parallels and much of the material, particularly the flagons, is likely to have been made there (Creuzenet 1996a; J. Simon, pers. comm.), with further material probably coming from other production centres in the region identified by Joly (1996; Joly and Mouton 2003).

Table 1: Approximate chronological range of ceramic assemblages

Site nameField no(s)/UFNo. of sherdsLa Tène D to AugustanC 1st-3rd ADC 4th-5th AD
A1: Chantal west27, 6553YY-
A1: Villa de Chantal10, 2341?Y?
A1: La Vente (N)31, 4112YY?
A1: La Vente (NN)852-Y-
A1: N du Jardet451?--
A1: Varveille29, 403Y?-
A2: Beaufort19, 35, 36, 4885YYY
A2: Chevannes2120--Y
A2: L'Haut des Champs33, 34, 64, 75517YY?
A2: Le Grands Champ15, 18165?Y?
A2: Les Bruyères461-?-
A2: Les Chaumes22, 44, 8421-YY
A2: Ravery20, 32, 4994-?Y
A3: Bas Cernat7851?Y-
A3: Champs d'Arroux67, 864-Y-
A3: Champs de la Lie871--?
A3: La Bruyère 100, 8124-Y?
A3: La Grand Pâture762Y--
A3: Le Champ de Maison837?--
A3: Le Champ Réau 8210--Y
A3: Les Grands Quartiers80, 9324YYY
A3: Moulin de Chazeu951?--
A3: Pâture de Rully 997Y-?
Poil : Les Bas de Fontaux796Y?-

The majority of assemblages provide evidence of activity between the 1st-3rd centuries AD, for which a densely settled landscape can be envisaged (Table 1). In some cases, this may be refined to indicate an earlier or later emphasis: for example at Bas Cernat, which may be primarily 1st-2nd century AD in date. The occurrence of Rhenish style fabrics (F2-11) (Table 2) provides a good indication of sites with 2nd-3rd century AD occupation, and overall the groups are notable for their apparent emphasis on that period. However, the small size of most groups and the limited information on the longevity of some pottery forms cautions against placing too much emphasis on such chronological precision.

Table 2: Sites with Rhenish style fabrics (2nd-3rd century AD)

Site nameNo. of sherds
A1: La Vente (N)1
A1: La Vente (NN)1
A1: Villa de Chantal1
A2: Beaufort1
A2: L'Haut des Champs6
A2: Le Grands Champ3
A2: Les Chaumes iii1
A3: Bas Cernat1
A3: La Bruyère1

The small number of diagnostic sherds in most of the assemblages prevents greater precision of dating without more understanding of fabric chronologies. However, the limited research on regional Gallo-Roman coarsewares makes the discussion of site chronology on the basis of fabrics highly problematic. Further study to identify typological forms peculiar to different fabrics is a priority for future research.

Despite the problems, a number of suggestions can be made on the chronology of different fabrics and these have an important bearing on the date of sites identified. The continuation of some fabrics from the late La Tène into the Roman period and of some 'Roman' fabrics into the post-Roman/early medieval period should also be recognised, which evidently limits their chronological use. However, the divergence of assemblages on a number of sites may be significant.

This is most notable with respect to pottery from different locations in the Beaufort/Ravery/Chevannes area. These show a very high proportion of B4 fabrics (Beuvray class PGMIC), fabrics that occur in only small numbers elsewhere, even in the large assemblages from L'Haut des Champs and Le Grands Champ (Table 3). This seems likely to be a chronological indicator, and implies that at least some of the B4 fabrics are not Gallo-Roman. The few B4 sherds diagnostic of form appear to be of very late Roman or possibly early medieval affinity, but no local parallels have yet been recognised. Further afield, similar forms (but in different fabrics) exist at Les Mureaux, Yvelines (Tuffreau-Libre 1992, 109), where they are dated to the end of the 5th century AD, and further afield where they are 5th-6th century AD in date (Raynaud 1990), but further comparison is required. If its late date is accepted, the presence of this highly micaceous and distinctive fabric on other sites may signify very late Roman or early medieval occupation. Theoretically, the strong presence of this fabric at Beaufort, which also has later 1st century BC/earlier 1st century AD material, and in small quantities at L'Haut des Champs, which has probable late La Tène occupation, could suggest this fabric is pre-Roman, but this seems less likely on the basis of the forms so far recovered.

Table 3: Sites with B4A-D fabrics

B4a-d (Beuvray PGMIC)Field no./UFNo. of sherds in assemblage% of assemblage (sherds)
A1 La Vente (N)31, 412 of 1216.7%
A1 Villa de Chantal10, 233 of 417.3%
A2 Beaufort19, 35, 36, 4851 of 8560.0%
A2 Chevannes2118 of 2090.0%
A2 L'Haut des Champs33, 34, 64, 7512 of 5172.3%
A2 Le Grands Champ15, 183 of 1651.8%
A2 Les Chaumes (i)225 of 771.4%
A2 Ravery20, 32, 4982 of 9487.2%
A3 Bas Cernat (Champs d'Arroux S)67, 78, 861 of 551.8%
A3 Champs de la Lie871 of 1100.0%
A3 La Bruyère81, 1002 of 248.3%
A3 Le Champ Réau827 of 1070.0%
A3 Les Grands Quartiers 80, 9312 of 2450.0%
A3 Pâture de Rully994 of 757.1%

The identification of any pre-Roman fabrics has been problematic and distinguishing La Tène D fabrics from Augustan and 1st-century AD coarsewares is likely to be extremely difficult. However, certain coarseware fabrics seem likely to be of late Iron Age or, at latest, early Roman date, as indicated by their occasional occurrence in late La Tène or early Roman vessel forms. These fabrics can be broadly grouped under the classification of PSGROS and PSMIC (including B1, B2a-c, R4, R1a-b). If this dating is correct, their presence on a number of sites may signify late La Tène occupation (Table 4). The sites where these fabrics occur are widespread, although the number of sherds is almost invariably small.

Table 4: Sites with B1, B2A-C, R4 and R1A-B fabrics

B1, B2a-c (Beuvray PSMIC, PSGROS)Field no./UFNo. of sherds in assemblage% of assemblage
A1 Chantal west27, 654 of 537.5%
A1 Villa de Chantal10, 233 of 417.3%
A1 La Vente (N)31, 412 of 1216.7%
A1 Varveille29, 401 of 333.3%
A2 Beaufort19, 35, 36, 482 of 852.4%
A2 L'Haut des Champs33, 34, 64, 7521 of 5174.1%
A2 Le Grands Champ15, 184 of 1652.4%
A3 Le Champs de Maison832 of 728.6%
A3 Les Grands Quartiers80, 931 of 244.2%
A3 Moulin de Chazeu951 of 1100.0%
R4, R1a-b (Beuvray PSGROS/PCGROS)  % of assemblage
A1 Chantal west27, 652 of 533.8%
A1 La Vente (N)31, 411 of 128.3%
A1 Varveille29, 401 of 333.3%
A2 Beaufort19, 35, 36, 484 of 854.7%
A2 L'Haut des Champs33, 34, 64, 7520 of 5173.9%
A2 Le Grands Champ15, 181 of 1650.6%
A3 Bas Cernat (Champs d'Arroux S)67, 78, 868 of 5514.5%
A3 Les Grands Quartiers80, 931 of 244.2%

Comparison of the largest assemblages re-emphasises the probable chronological discrepancies between certain sites (Fig. 95). The large number of PCCRU (flagon fabrics) at both L'Haut des Champs and Le Grands Champ contrasts with the very small proportion at Ravery and, to a lesser extent, Beaufort, emphasising the limited evidence for Gallo-Roman occupation at these latter locations, which instead have large proportions of the undated but probably late B4 fabrics.

The presence of Dressel 1 amphora (found by Niaux) at Chevannes and Pascual 1 at Beaufort suggests 1st-century BC/AD occupation in the area, but if there were continuing Gallo-Roman occupation here, more material of later date would surely be expected, given the size of the assemblage and the proximity of these locations to L'Haut des Champs. It therefore seems likely that there is a genuine chronological divergence between the Beaufort/Ravery/Chevannes sites and those at L'Haut des Champs and Le Grand Champs.

One of the more notable features of the fieldwalking assemblage is the lack of early finewares such as Terra Nigra, Terra Rubra and Campanian pottery. This is somewhat surprising given the presence of early coarsewares and Dressel 1 at a number of sites, and later occurrence of Terra Sigillata and Rhenish wares. Terra Nigra, Terra Rubra and Campanian wares were all reaching the area (e.g. Barral and Joly 2002) and some Gallo-Belgic style wares were produced within Burgundy (Joly and Barral 1992), including a few forms at Autun (Creuzenet 1996a). However, early Roman finewares are also noticeably absent from recent excavations at the Sources de l'Yonne, where Dressel 1 is abundant (Creighton and Moore 2008), and despite their presence on the Temple site itself (Péquinot et al. 1996, 219), so their absence from the survey may well reflect the amount of finewares reaching rural sites in the early Roman period, or the chronology, function, or status of individual sites, rather than collection biases.

The 4th and 5th centuries appear to be represented only at L'Haut des Champs and Le Grands Champ and are absent from other assemblages. However, this may be more a function of the much larger size of these two assemblages, rather than signalling chronological variability.


© Internet Archaeology/Author(s) URL:
Last updated: Wed Oct 29 2008