The macroscopic traces observed on vessel surfaces conform to the range discussed by Roux (2019). These individual traces were never found to be correlated with one specific forming technique when found in isolation; instead – as before – the combination of specific traces appearing together, and the configuration of those traces seemed to have the best indication of forming technique. The following traces relating to both forming and finishing were observed in the experimental type set.
Traces observed
The vast majority of traces observed are indicative of what Roux has identified as ‘irregular topography’, that is “a relief with hollows and protrusions” (2019, 144). It is important to note also that these hollows and protrusions can be further described in terms of their “dimensions; morphology; the line, rectilinear and curvilinear (sinuous); orientation, vertical, horizontal, oblique, and concentric; and localization” (Roux 2019, 144).
Hollows
Hollows observed included fissures, crevices, depressions, and imprints. As such, they represent features which plunge into the plane of the vessel surface.
Fissures
Fissures, the “deep incisions situated at the limit of juxtaposed elements” (Roux 2019, 145), principally appear on interior surfaces and range significantly in length. Their orientation includes horizontal, oblique, and concentric examples, often on the same vessel, and the line can be characterized as curvilinear. The presence of this trace did not appear to be influenced by the choice of clay; fissures were present in tempered and untempered examples alike, though it should be stated that untempered examples present the traces somewhat more clearly. Figure 12 shows two such examples.
Crevices
A lesser-observed type of hollow is the crevice, identifiable as “tears in the paste […] either left by tools and formed after superficial pull-outs of a paste with too little hygrometry […] or on the body of wheel-fashioned pastes stretched too fast” (Roux 2019, 145). In the experimental type set described here, it was for the latter reason that these crevices were formed. They are often visible at the widest point of the vessel, whether on interior or exterior surface. They are often quite short and oblique to vertical in orientation. They also tend to occur in small clusters; 2-3 or more will often be visible as a small group in one area of the vessel wall (see Figure 13). Crevices occur on tempered and untempered vessels alike, though it can be noted that the most severe (i.e. greatest number clustered together and longest crevices) were observed on a tempered example.
Depressions
Depressions feature with somewhat low frequency within the experimental type set, and are almost exclusively observed on exterior surfaces. They are defined as “hollows with diffuse contours” (Roux 2019, 144). The morphology of these contours can be indicative of the mechanics of their creation. Within the type set, the most common cause of depressions relates to the presence of areas where walls are thinner compared to adjacent areas. During the application of RKE these thinner areas are weak points and the wall collapses inward slightly. Figure 14 illustrates this, and the thinner area of the wall is due in this case to the location of a coil end.
Imprints
Imprints are identifiable as “negatives left by artifacts or fingers during the course of the different fashioning and finishing operations” (Roux 2019, 145), and are present on virtually every tempered example within the experimental type set. As the vast majority related to the movement of tempering material across the vessel surface while the vessel was being subjected to RKE, these traces are, for the most part, short and rectilinear with a horizontal orientation. The temper itself was 2mm, and as such the imprints are more or less the same size. Figure 15 shows examples of imprints from the experimental material.
Protrusions
In addition to the hollows described above, protrusions were also present. These are divisible into continuous protrusions (including bands and undulations) and intermittent protrusions (including crests, compression folds, and bumps).
Concentric protrusions
Both categories of concentric protrusion were observed in the experimental material: undulations and bands. Bands can be thought of as a type of undulation; in both cases, the morphology of the vessel section presents alternation between convex and concave in a sinuous line. While ‘undulations’ refers to the more smoothly sinuous examples, ‘bands’ are “undulations with a rectangular cross section” (Roux 2019, 145).
Undulations
The undulations present in the experimental material range from small to large, occurring more or less over the full height of the vessel wall. Undulations can appear on interior and/or exterior surfaces. They occur equally on tempered and untempered examples, and an illustration of this trace is visible in Figure 16.
Bands
The experimental material presented relatively few examples with bands, though there was little difference between tempered and untempered examples in terms of morphology or frequency (see Figure 17). As with undulations, the bands range from small to large, more or less occurring over the full height of the vessel, and appearing on the interior and/or the exterior.
Intermittent protrusions
Intermittent protrusions, unlike concentric protrusions, are discontinuous and discrete in nature, with a variety of specific variations possible. A number of these are present on the experimental type set.
Crests
Crests, “the result of accumulation of clay slurry […], present as filiform elevations” (Roux 2019, 145) are observed on tempered and untempered examples alike. These relate almost entirely to the procedure of removing the vessel from the wheel surface, and often portions of the palmar surface of fingers and finger tips are recognizable. Crests are more readily identified on untempered vessels, however, as the impact of the 2mm temper on the surface topography in the tempered examples seemed to both reduce the potential for crest formation as well as obscure the visibility of those crests (see Figure 18).
Compression folds
Compression folds are one of the traces which present the greatest variability within the experimental type set. They are “obtained by the compression of the wall and are often located on narrowing zones (base and neck)” (Roux 2019, 145), and most often appeared on the lower third of vessels in this type set. The folds range from short, small, and thickly distributed over the affected area to long, large, and moderately thickly distributed over the affected area. In any case the orientation of the compression folds is determined by the direction of rotation of the wheel during RKE shaping and thinning – in this case the counterclockwise rotation resulted in folds which are oblique from upper left to lower right on exterior surfaces and lower left to upper right on interior surfaces. Figure 19 shows the range of variation of this trace within the type set, with an untempered small vessel on the left exhibiting many small, fine compression folds as well as another untempered small vessel in the middle exhibiting longer, larger, and more obvious compression folds. The rightmost small tempered vessel in Figure 19 also exhibits small, fine compression folds, illustrating the potential difficulty in recognizing those in conjunction with the topography of a tempered paste’s surfaces.
Bumps
The final type of intermittent protrusion observed within the experimental type set was bumps, “the expression of unequal pressures on the walls” (Roux 2019, 145). This trace was not especially frequently observed, but nevertheless appeared in both tempered and untempered examples. The manifestation of this trace did tend to differ, however, with the presence of temper as the bumps themselves seemed to force temper closer to the surface giving greater texture to the surface of the bump overall. Figure 20 shows the manifestation of bumps on both tempered and untempered examples from the experimental type set.
The ‘wedged coil’ wheel coiling – Diagnostic Features
Fully diagnostic features of ‘wedged coil’ wheel coiling are infrequent, though there are recurring trace combinations within this experimental type set. Fissures in general were only observed in the ‘wedged coil’ wheel coiled material, which confirms observations from previous similar experiments by the author (Jeffra 2015). These cannot be taken as diagnostic for the specific variation (wedged coil) being reported here, but the presence of this evidence for “juxtaposed elements” (Roux 2019, 145) in conjunction with evidence of RKE such as undulations and compression folds (such as that seen in Figure 21) is a strong indication of a combination technique. The distribution of horizontal to oblique linear or sinuous fissures at regular intervals for the height of the vessel (in conjunction with other RKE features) does seem to be an especially strong indication of the use of coils and thus wheel coiling.
Caution in applying this is advisable, for not all examples of known wheel coiling manufacture exhibit the type of fissures described above and it may be the case that fissures manifest as strongly oblique and discontinuous without identifiable regular intervals (see Figure 21 a, c).
The single most notable diagnostic feature is the relationship of traces observed on interior vs exterior surfaces, the result of the forming method itself. As noted above, each coil is applied to the upper, inner surface of the existing wall and the mass of that coil is distributed upward. Even prior to rotation, this creates a volume with fissures from coil joining solely on interior surfaces, which is maintained throughout the potting process. Figure 22 shows an example of this, where the exterior surface shows no evidence of fissures while the interior surface (right) presents a number of long, sinuous, horizontally- to obliquely-oriented fissures at regular intervals for the height of the vessel.
The use of this observation as diagnostic for ‘wedged coil’ wheel coiling should be applied with caution and skepticism, however. This combination of traces could very well be observed in other types of wheel coiled vessels, particularly those which were subjected to further finishing operations with and without RKE. It is perhaps more reliable to consider this combination diagnostic among the plainest and least finely finished archaeological examples, and to take the perspective that a single vessel showing this feature is less reliable than an assemblage where this feature is observed among a large proportion of vessels.
Fissures form one part of another relationship which was only observed within the ‘wedged coil’ experimental material: severe fissures on a vessel interior with corresponding depressions on the vessel exterior. Figure 23 shows one such example of this. The two traces are inextricably linked in this case; the fissures visible on the interior surface are located where two ends of one coil met as well as the horizontal boundary between that and the coil lower on the vessel wall. This intersection and the extension of those fissures beyond the intersecting point show on the interior surface that these juxtaposed elements were insufficiently joined together with RKE pressures. The fact that the ‘wedged coil’ method does not produce fissures on exterior surfaces does not preclude the existence of thinner areas of the wall and the depression visible on the exterior view demonstrates that the poor integration of juxtaposed elements manifests there as hollows – in this case a depression – where the vessel wall is thinner.
The observation of the relationship between these two discrete traces tidily underlines the need to consider traces in context rather than in isolation.
Compression folds as a category of trace were observed in both wheel coiled and wheel thrown vessels, but the dimensions of these compression folds do vary somewhat based on forming technique. In the case of wheel coiled vessels exhibiting this trace, it most often manifested as longer, wider, and deeper folds (see Figure 24).
It may be the case that subsequent experimental work or different experimental protocols produce less of a marked difference between wheel coiled and wheel thrown manifestations of the compression folds seen here. This experiment, however, confirms a relationship between more significant compression folds and wheel coiling which was seen in the earlier experiment (though in that instance compression folds were referred to as torsional rippling, torsional strain, and torsional tearing).
Beyond just the presence of significant compression folds, there is an observed relationship between those folds and undulations and/or bands which was not observed within the wheel thrown material. Two variations of the interaction between compression folds and undulations/bands were observed (see Figure 25). The first variation is characterized by folds which are largely constrained to the troughs of a vessel surface’s undulations, where the peak of one band of undulation acts as a terminus for the folds. The folds themselves may differ morphologically, but in any case can be considered significant in terms of scale (see Figure 25 a-b, d-e).
The second variation on compression folds and undulation/band interaction unique to wheel coiling is identified by the way that the compression folds bisect the peaks and troughs of the undulations (Figure 25 c, f). The compression folds themselves are not the most significant in terms of width and depth, but given that they are not restricted to the troughs of the undulations, they are among the greatest compression folds in terms of length.
The difference between these two variations on compression fold and undulation/band interaction may be simply that the torsional effects during potting (i.e., the force of rotation in opposition to the braking effect of shaping pressures on the vessel wall) were either stronger or weaker relative to the overall strength of the clay, which resulted in compression folds which either extended beyond the potential terminus of an undulation/band peak, or did not. In any case, the presence of this relationship was a feature restricted to wheel coiled vessels in this type set.
The wheel throwing technique – Diagnostic Features
One of the most striking aspect of the wheel thrown vessels within the experimental type set is their overall regularity. Though very few are identified as having regular topography, or “a uniform relief with a continuous plane” (Roux 2019, 143), the irregularities of the topography tend to be very minor and/or low relief.
A first feature observed within the wheel thrown material is the strong tendency toward traces with very low relief, that is, protrusions and hollows which do not deviate significantly from the wider plane of the vessel surface. This low relief does not preclude the presence or variety of observed traces, rather it is a characterization of those traces (see Figure 26).
One particular trace morphology which was exclusive to wheel thrown examples within the type set was the wide distribution of minor compression folds across the majority of a vessel’s surfaces, both interior and exterior. These compression folds follow the same orientation determined by the counterclockwise direction of wheel rotation, but are quite short, narrow, and densely distributed across the vessel. This morphology can result in quite subtle traces, particularly in tempered vessels, but Figure 27 presents the interior and exterior surfaces on one such example.
Wheel thrown vessels with temper presented a specific trace relating to the imprinting and dragging of a temper particle across the vessel surface. The resulting trace might be mistaken for a fissure, though in this case the dragged imprint trace has more of a squared or U-shaped cross section to correspond with the way the temper particle digs a furrow across the surface. In contrast, a fissure is more often V-shaped in cross section as the two convex surfaces of the juxtaposed elements meet. What makes the wheel thrown dragged imprint traces distinct is their straight line, their horizontal orientation, and their occasional great length (see Figure 28).
Comparing Common Features
Few significant features were identified in common between wheel coiling and wheel throwing in this experiment, which should be clear from the lack of strong overlap in previous passages above. One type of trace is, however, observed in both wheel coiled and wheel thrown examples: compression folds. What sets this particular trace apart from others held in common is the fact that the dimensions of those traces have some correspondence with forming technique. Figure 29 shows a side by side comparison of two typical vessels, of the same morphology, with compressions folds which differ in dimension.
It was found that compression folds for wheel coiled examples often manifest as wider, larger, and more sinuous folds than in wheel thrown vessels. For wheel thrown vessels, compression folds most often appeared as small, narrow, straight, and quite fine. In both cases, the pattern of oblique orientation was consistent, and reflected the direction of wheel rotation.
CONCLUSIONS
The experiment described above provides greater detail of and a glimpse into the nuance necessary when studying and evaluating forming traces. The pluralistic nature of individual traces is already well established (Roux 2019). This work can therefore be situated as an investigation of the relationship between a limited range of variables (vessel shape, forming technique, clay type) which can be immediately employed as comparative material against archaeological examples, to better aid with interpreting those pluralistic traces and understanding them in their context. It is the particular goal of the author that this work assist those who do not have the resources to complete their own experimental programs, and it is for this reason that the products of the experiment have been made available open access for consultation as reference material. This ethos of transparency and open access shaped the way that this experiment has been reported upon above, just as it led to the creation of an open access database of the experimental material (available at https://tracingthewheel.eu). This report is complementary to the content of that database, where photographs, forming videos, and 3D models are available for each vessel in the experimental type set described here.
Looking to the broader context in which this experiment is situation, the fact that transitional techniques such as wheel coiling have been identified has created a renewed opportunity to identify precisely when potters made the leap from those transitional techniques to wheel throwing, a technological boundary which was taken for granted not so long ago. That the presence of such combination techniques previously went unobserved is a strong indication that identifying the boundary between those techniques and wheel throwing should reinforce a need for meticulous study of material and caution in assessments. Even so, as more work is done on the topic of identifying the forming techniques used by potters of the past who were using the wheel, there are more opportunities to successfully identify the transition point from wheel coiling to the wheel throwing technique. This experiment can stand as a reference point for making some of these identifications. What should be very clear, however, is that single archaeological examples within a wider assemblage are not strong evidence alone; indeed, within this very type set of vessels produced with known techniques there is still considerable overlap in clusters of traces. It is therefore the hope of the author that interpretations of techniques observed be based on clearly-described descriptions of individual traces as employed above first, and that future studies of archaeological material focus on seeking patterns within assemblages, rather than solitary examples.
BIBLIOGRAPHY
JEFFRA, C. 2011. The Archaeological Study of Innovation: An Experimental Approach to the Pottery Wheel in Bronze Age Crete and Cyprus. PhD, Exeter: University of Exeter.
JEFFRA, C. 2013. “A Re-Examination of Early Wheel Potting in Crete,” The Annual of the British School at Athens 108, 31–49.
JEFFRA, C. 2015. “Experimenting wheel-coiling methods,” The Arkeotek Journal 2.
JEFFRA, C. 2021. “Generalised type sets in experimental ceramics: widening applicability and maximizing cross-cultural assessments,” Interdisciplinaria Archaeologica 12(2): Online First.
ROUX, V. 2019. Ceramics and Society: A Technological Approach to Archaeological Assemblages. Springer.
ROUX, V. & M.A. COURTY. 1998. “Identification of wheel-fashioning methods: technological analysis of 4th-3rd millennium BC oriental ceramics,” Journal of Archaeological Science 25: 747–63.