# The analysis of small drinking vessels

### Introduction

This chapter investigates artefact variability of Attic Geometric drinking vessels, and more specifically of kantharoi and skyphoi. Small sized drinking pots comprise the vast majority of the material examined in this thesis. Macroscopic analysis is conducted on 165 ceramic artefacts in total. These are 153 vessels with complete profiles and 12 incomplete vessels or sherds. The majority of this material (149 artefacts or 90.3%) comes from Athens, while 16 pieces (or 9.7%) have been identified as broadly Attic by Coldstream (2003b; 2010).

### The argument

This article argues that according to metrical features, proportions and fabrics, the production of small drinking vessels was broadly standardised all across the Geometric period. Even though there were technological traditions regulating their production, the degree of internal variability among statistical clusters suggests that the conceptualisation of such pots was not strict and potters enjoyed certain freedom in their work. The production of Attic Geometric skyphoi involved a large number of potters, some of which specialised in specific sub-typologies. Their workshops were probably not clustered in a single production site but they were scattered in different locations. The decorative characteristics of kantharoi were highly standardised by contrast to those of skyphoi. The decoration of skyphoi exhibited higher degree of artefact variability and experimentation not only in relation to kantharoi, but also in relation to any other ceramic class examined in previous chapters. This technological variability became stronger after MGII, suggesting that painters of skyphoi enjoyed some freedom in their work similar to that of potters.

### Analysis of metrical features and proportions

*The Athenian Agora (supplemented by the Kynosarges burials and the collections of the British School at Athens)*

The analysis of metrical features and proportions of small drinking vessels is conducted on 47 pots with complete profiles (14 kantharoi and 33 skyphoi) according to the methods described in Chapter 3. From this assemblage, 36 vessels come from the Athenian Agora (11 kantharoi and 25 skyphoi), 7 from the Kynosarges burials (1 kantharos and 6 skyphoi) and 4 from the collections of the British School at Athens (2 kantharoi and 2 skyphoi). Metrical features and proportions of all drinking vessels from Kynosarges and the British School’s collections are analysed in a single section, together with the Agora pottery due to their typological and stylistic similarities. For the total assemblage of 47 pots with complete profiles, 13 vessels come from burial contexts (6 kantharoi and 7 skyphoi), 30 come from mixed non-burial deposits (6 kantharoi and 24 skyphoi), while the archaeological context of 4 vessels is unknown (2 kantharoi and 2 skyphoi). Charts 1 and 3 present metrical features and proportions for the Agora assemblage, and Charts 4 and 5 record the same features for the Kynosarges and British School assemblages. In addition to this study, Chart 2 presents another 11 pieces (4 kantharoi and 7 skyphoi) with some surviving metrical features coming from incomplete or fragmented pottery from the Athenian Agora. These fragments are not used in the analysis of metrical features and proportions; however, they supplement the analysis of fabrics and decorative technology further below.

The first thing to notice in the comparison between skyphoi and kantharoi in Chart 1 is that the rim diameters of the latter are by majority deformed. Rim deformations often exceed 1cm. By contrast, all rim diameters of skyphoi are uniform. Both wares are wheel made and manufactured during a single episode on the potter’s wheel. Footed vessels or pots with ring bases were probably produced during two episodes on the wheel. In general, the chaîne opératoire of both typological classes shows similarities with regard to the initial steps of their forming processes. Rim deformations for kantharoi most likely occurred during the second step of their production, the handle attachment stage, which was different compared to that of skyphoi. According to Figure 1, the handles of kantharoi were stuck vertically on the vessel’s walls on two contact points along a vessel’s external surface: one right at the side of the rim and another one further below it, towards the middle of the vessel’s walls. By contrast, the handles of skyphoi were stuck horizontally on the vessel’s walls, on a contact area that did not touch the rim.

*Figure 1: Different handles attachment techniques for kantharoi and skyphoi.*

The rim deformation of kantharoi was most likely due to excessive pressure along the rim during the effort to attach the handles, resulting to alterations in their rim diameter axis. Furthermore, during the sequence of manufacture from base to rim (bottom to top), the thickest areas of a drinking vessel were located towards its base and the thinnest towards its rim, which was formed at the very end of the first episode on the potter’s wheel. As the rim was the thinnest part of a kantharos, the attachment of its handles resulted to easier deformations due to hand pressure. By contrast, the attachment of skyphoi handles on the sides of the vessel, on areas with thicker walls, resulted to homogenous shapes without deformations along the rim axis.

*Figure 2: Scatter-graph of correlation between net height and base diameter. Athenian/Attic kantharoi from the Agora, Kynosarges, and the British School’s at Athens collections with complete profiles.*

Due to lack of clear rim diameter measurements for kantharoi, their comparison with skyphoi takes place in relation to their net height and base diameter measurements. According to the correlation presented in Figure 2, kantharoi form two distinct clusters. Cluster 1 consists of 6 kantharoi dating between MGII-LGIa and SG, which show low degree of artefact variability. All vessels are of small sizes, including the footed SG kantharos P7196:

Cluster 1 is characterised by net heights below 8cm and base diameters bellow 6cm. By contrast, cluster 2 is more scattered compared to cluster 1, and shows greater artefact variability. It consists of 8 vessels dating between MG and SG times, which belong to two sub-classes: low-handled kantharoi of normal sizes and high-handled kantharoi:

According to the above, it appears likely that the production of small sized kantharoi was more standardised compared to that of normal sized pots. Both types of kantharoi show no distinct chronological patterns and they cluster according to size instead of chronological period. Furthermore, the artefacts from Kynosarges and the British School’s collections blend nicely with the material from the Agora, suggesting that all vessels were produced based on similar conceptualisations.

*Figure 3: Scatter-graph of correlation between net height and base diameter. Athenian/Attic skyphoi from the Agora, Kynosarges, and the British School’s at Athens collections with complete profiles.*

The situation with skyphoi from the same contexts is slightly different. According to Figure 3, skyphoi form two clusters that are not entirely distinct. Cluster 1 is the largest and consists of 29 vessels of mixed dates suggesting no distinct chronological groupings:

Cluster 1 includes all vessels from Kynosarges and the British School’s collections. By contrast to Kantharoi, skyphoi from Kynosarges appear on the margins of this rhomboid cluster, suggesting a production that was probably distinct although not entirely different compared to the Agora vessels. Furthermore, two skyphoi from Kynosarges (K10 and K88) show a distinct conceptualisation that is not noted on any other vessel: their heights and base diameters are equal (see Chart 4). As both vessels date close to MGII, it appears likely that these were the products of the same potter or workshop. Finally, the only gadrooned skyphos (sensu Coldstream 1968; 2003b, 345) stands at the top margin of the rhomboid cluster: A342 has the broadest base (7.8 cm) and the largest height (8.5 cm) in the entire assemblage. This vessel is likely to suggest that drinking cups imitating metallic prototypes were produced larger than other pots, perhaps copying some features of the metallic originals (see Borell 1978, 93-4 and Markoe 1985, 117-27); however, their broader conceptualisation was not entirely distinct.

Cluster 2 consists of 4 vessels, which exhibit net heights above 9 cm. With exception of P21807, the rest are three wide skyphoi dating in MGII-LGIa. All three vessels have been recovered in the same context (Well D12:3) and their clustering is likely to suggest that they were the products of the same workshop:

According to the above, the production of skyphoi is homogeneous with no characteristic variations, neither with regard to specific sub-typologies, nor with regard to specific chronological periods. The majority of vessels form a single and large cluster, in which minor internal variability is likely to suggest differences among workshops. This cluster suggests a low degree of artefact variability connected to standardisation in the production of such vessels.

By contrast to the study on oinochoai, the production of drinking vessels with ring bases was probably not popular during the Geometric period. In this study, it is only 5 out of 33 skyphoi that carry ring bases, dating between MGII and LGIIa:

Kantharoi with ring bases are 4 out of 14 vessels in total, dating between EGI and the early 7th century (SG style):

The small number of drinking vessels with ring bases verifies that the production of skyphoi and kantharoi was standardised even though some exceptions existed through time. By contrast to oinochoai with ring bases, the production of which was popular until the beginning of the Late Geometric era, drinking vessels were mostly produced with flat bases all across two centuries.

The possibility of standardisation in the production of drinking vessels could be attributed to specific technological traditions that regulated the conceptualisation of such shapes. A comparison of mean proportions for kantharoi and skyphoi according to the data recorded in Charts 3 and 5 from all three sites reveals low standard deviations for the proportion of handle attachment height to net height:

The kantharoi assemblage used for the calculation of this proportion consists of 11 out of 14 vessels, and the skyphoi assemblage consists of 30 out of 33 vessels. These reduced numbers are due to pots with no recorded handle attachments, either because their handles are missing or they are reconstructed with plaster and expected to be inaccurate. Furthermore, the low standard deviation in the proportion of base diameter to rim diameter for skyphoi (5.18) is likely to suggest a second pattern that requires further investigation.

According to the correlation of handle attachment height to net height in Figure 4, the regression line for skyphoi follows the equation y = 0.703x – 0.6618 (where y = handle attachment height and x = net height). In other words the handles of skyphoi were attached roughly at 70% (or 7/10) of a vessel’s net height. The difference of 0.6618 cm is too small to be considered. The coefficient of determination of the regression line (R2=0.9073) shows relatively strong statistical correlation at 90.73%. Kantharoi produce a similar regression pattern, parallel to that of skyphoi: it follows the equation y = 0.6947x – 0.9822 (where y = handle attachment height and x = net height). The equation shows that the handles of kantharoi were attached roughly at 69.5% of a vessel’s net height. Again, the difference of 0.9822 cm is small to be considered. The coefficient of determination (R2=0.9475) shows strong statistical correlation (94.75%).

*Figure 4: Scatter-graph of correlation between net height and handle attachment height. Athenian/Attic drinking vessels from the Agora, Kynosarges and the British School’s at Athens collections with complete profiles and recorded handles.*

Based the above, the production of kantharoi and skyphoi followed similar conceptualisations, which indicate connections in the chaînes opératoires of both typological classes. The handles of both wares were attached at roughly 70% (or 7/10) of a vessel’s net height all across the Geometric era, and this was a strong technological tradition in the production of such pottery. The coefficients of determination of the regression lines in Figure 4 suggest that the statistical correlation for kantharoi is slightly higher compared to that of skyphoi. This observation shows that the production of kantharoi was more standardised compared to the latter.

*Figure 5: Scatter-graph of correlation between rim and base diameter. Athenian/Attic skyphoi from the Agora, Kynosarges, and the British School’s at Athens collections with complete profiles, shown all together.*

A high degree of standardisation in the production of skyphoi is also noted with regard to the low standard deviation (5.18) of the mean proportion of base diameter to rim diameter explained earlier. This proportion cannot be studied for kantharoi, as their rims are highly deformed. The correlation of the two metrical features in Figure 5 shows that the vessels are highly scattered. The regression line for 33 skyphoi with complete profiles from three sites follows the equation y = 0.4502x + 0.1109 (where y = base diameter and x = rim diameter). The coefficient of determination (R2=0.7256) shows weak statistical correlation (72.56%).

A second correlation of the same assemblage divided by different sites in Figure 6 shows that the Agora material is scattered closer to the regression line. The material from Kynosarges and the British School’s collections stands on the margins of the cluster comprised of the Agora vessels. The regression line for 25 skyphoi from the Agora follows the equation y = 0.4502x + 0.1109 (where y = base diameter and x = rim diameter). The coefficient of determination (R2=0.8215) shows slightly stronger statistical correlation compared to the regression line in Figure 5, even though the percentage of correlation is not entirely satisfactory (82.15%).

*Figure 6: Scatter-graph of correlation between rim and base diameter. Athenian/Attic skyphoi from the Agora, Kynosarges, and the British School’s at Athens collections with complete profiles, shown separately.*

According to both graphs, the proportional relationship between the two metrical features of skyphoi is likely to follow a specific pattern, where base diameters are roughly 44% or 45% of a vessel’s rim diameter. This relationship could be due to distinct conceptualisations followed in the production of skyphoi, although variations were common. Such variations are likely to explain the low degree of statistical correlation of the regression lines in Figures 5 and 6. The metrical features for the Kynosarges material are more scattered compared to those from the Agora, which is likely to suggest that these pots were produced by different workshops or another group of potters.

*The Kerameikos cemetery*

The analysis of metrical features and proportions of small drinking vessels from the Kerameikos cemetery is conducted on 95 artefacts with complete profiles. The assemblage consists of 69 decorated skyphoi and 26 kantharoi. All vessels have been recovered in burials and their grave contexts are recorded in Charts 6 and 7.

In the original publication of kantharoi from Kerameikos, Karl Kübler (1954) only recorded their net height measurement, which he noted as height without handles (Höhe ohne Henkeln). Unfortunately, for the metrical features of skyphoi he did not show the same degree of consistency: some vessels were recorded by their height and others by their maximum diameter. Here, Chart 6 records real net heights only for the skyphoi published with this measurement in the original Kerameikos volume. A second column records calculated net heights for those skyphoi that were originally published by their maximum diameter. All other metrical features in Charts 6, 7, 8 and 9 were measured in smaller scale through published photographs and were then calculated in real scale based on the original real net height or real maximum diameter measurements. Even though rim diameters could have been calculated for kantharoi, these measurements and their related proportions were not included in any of the charts. Similarly to the Agora assemblage, the rim diameters of kantharoi from Kerameikos are expected to be deformed.

To ensure the accuracy of calculated measurements, 5 skyphoi from the Kerameikos assemblage were chosen for macroscopic analysis and two separate accuracy tests were conducted, which are presented in Charts 10 and 11. During these accuracy tests, real rim and base diameters and real proportions or rim and base diameters to net height were compared to the ones that were calculated though published photographs for the same artefacts. According to Chart 10, the differences between real and calculated base diameters for the Kerameikos skyphoi range between -0.2 cm and +0.9 cm. Differences between real and calculated proportions of base diameter to net height range between -2.4% and +6.1%. Even though differences in base diameters are less than ±1 cm, differences in the percentages of proportions of base diameter to net height occasionally exceed 5%, which raises the possibility of statistical bias. For this reason, a second test was carried out in relation to rim diameters. According to Chart 11, differences between real and calculated rim diameters for the same pots range between -0.9 cm and +0.3 cm. Differences between real and calculated proportions of rim diameter to net height range between -4.6% and +4.2%. Despite the lower degree of statistical bias demonstrated in the second test, calculated metrical features for drinking vessels need to be treated with caution.

Both accuracy tests show that the analysis of drinking vessels from Kerameikos is more problematic compared to that of closed ceramic containers and pouring vessels. Firstly, a height or diameter difference of ±1 cm for large or medium sized vessels is likely to be negligible in relation to the overall size of a vessel. However, in the case of drinking cups, such differences up to ±1 cm could mean significant bias due to the smaller sizes of such pots. Secondly, vessel deformations are not always clear through published photographs and originally recorded measurements are likely to describe maximum metrical features instead of mean metrical features. In this case, bias is likely due to different recording strategies and the more deformed the vessels, the more the bias.

*Figure 7: Scatter-graph of correlation between net height and base diameter. Athenian kantharoi from the Kerameikos cemetery with complete profiles.*

According to the correlation of base diameter and net height for the Kerameikos kantharoi in Figure 7, the assemblage forms two distinct clusters with similar properties compared to the Agora kantharoi. Cluster 1 consists of 11 vessels, which are all small kantharoi (with heights smaller than 8cm), either with high or low handles, two of which are footed. This cluster shows no distinct chronological properties. Vessels range between EGI and SG times:

Cluster 2 consists of 14 vessels, which are all typical kantharoi with high or low handles. One vessel, 1251, is footed. This cluster appears more scattered compare to cluster 1 and includes different sub-typologies. Vessels date between EGII and LGIIa, suggesting no distinct chronological patterns:

Finally, the footed kantharos 930 is a loner. This vessel is different compared to all other footed kantharoi due to its longer foot, which could have functioned as a handle. According to the above, kantharoi from Kerameikos diversify according to size instead of typological or chronological variation.

The correlation of net height and base diameter for the Kerameikos skyphoi in Figure 8 shows that the majority of vessels form a distinct triangular cluster. In this cluster there is a minor degree of artefact variability, which relates to specific sub-typologies. Firstly, all wide skyphoi with stirrup handles (Coldstream 1968, 18) dating between EGII and MGI are gathered at the top corner of the triangular cluster.

*Figure 8: Scatter graph of correlation between net height and base diameter. Athenian skyphoi from the Kerameikos cemetery with complete profiles.*

Secondly, all gadrooned skyphoi dating in LGI are gathered at the bottom right corner of the triangular cluster. These two patterns are likely to suggest that both sub-groups are the products of distinct chaînes opératoires or workshops. They were probably produced in characteristic sizes although their conceptualisation never diverted from the main norm. Figure 8 also shows that 4 pots stand out and need to be treated as loners. These vessels come from different chronological periods and typological sub-groups:

According to the above comparisons from Kerameikos, the production of both skyphoi and kantharoi appears standardised and follows no patterns of significant change across time. The comparison of mean proportions for all vessels based on the information recorded in Charts 8 and 9 shows that the lowest standard deviations are encountered in the proportion of handle attachment height to net height:

The lesser degree of artefact variability is likely to suggest a technological tradition in the production of such drinking vessels, similar to the one observed in the previous section for the Agora material. Furthermore, the proportion of base diameter to rim diameter for skyphoi requires further investigation despite that its standard deviation is relatively high (10.11).

The correlation of handle attachment height to net height in Figure 9 shows that kantharoi and skyphoi from Kerameikos are relatively scattered, even though a central tendency is clear for skyphoi. According to the graph, both wares follow regression patterns that are no different compared to those from the Agora-Kynosarges-British School assemblage. The regression line for kantharoi follows the equation y = 0.7097x - 1.3636 (where y = handle attachment height and x = net height). This means that the handles of these vessels were attached at a height of roughly 71% of a vessel’s net height, with a small difference of 1.36 cm. The coefficient of determination (R² = 0.8638) shows relatively satisfactory statistical correlation (86.38%). The regression line for skyphoi follows the equation y = 0.673x – 0.8208 (where y = handle attachment height and x = net height). This means that the handles of skyphoi were attached at roughly 67.3% of a vessel’s net height, while the difference of 0.8208 cm is small to be considered. This percentage could also be translated as a fraction of 2/3 or 67% of a vessel’s net height. The same fraction was noted with regard to handle attachment heights of neck-handled amphorae. The coefficient of determination of this regression line (R2=0.8585) shows relatively satisfactory statistical correlation (85.85%).

*Figure 9: Scatter-graph of correlation between net height and handle attachment height with regression lines. Athenian drinking vessels from the Kerameikos cemetery with complete profiles and recorded handles.*

Despite the overall variation and the relatively satisfactory degrees of statistical correlation of both regression lines, the material from Kerameikos suggests similar patterns to the ones observed for the Agora-Kynosarges-British School assemblage. The handles of kantharoi were attached at roughly 70% of a vessel’s net height (or 7/10) and those of skyphoi at roughly 67% (or at a fraction of 2/3). This pattern probably relates to a technological tradition that was followed across two centuries, even though artefact variability within the typological clusters suggests that several vessels diverted from the main norm.

According to the correlation of base diameter to rim diameter for the Kerameikos skyphoi in Figure 10, there appears to be no clear regression pattern as opposed to the one noted with regard to the same proportion for the Agora skyphoi. Although the vessels from Kerameikos are closely congregated in a triangular cluster, their regression line suggests no distinct pattern. It follows the equation y = 0.7209x – 2.1449 and the coefficient of determination (R2=0.5843) shows weak statistical correlation. Despite the presence of a central tendency in this cluster, the possibility of a distinct technological tradition is highly unlikely due to the way scattering takes place. According to the graph, skyphoi with broader base diameters exhibit greater variability in their rim diameters; hence, they spread widely above and below the regression line towards the right side of the graph.

*Figure 10: Scatter-graph of correlation between rim diameter and base diameter with regression line. Athenia skyphoi from the Kerameikos cemetery with complete profiles.*

*The British Museum collections*

The analysis of metrical features and proportions of small drinking vessels from the collections of the British Museum is conducted on 12 skyphoi with complete profiles, recorded in Charts 12 and 13. The assemblage from the British Museum does not contain any kantharoi. All vessels derive from unknown contexts and are characterised as broadly Attic by Coldstream (2010), apart from those of suspected Athenian origin.

Calculated measurements for the British Museum skyphoi were tested in a similar manner as the assemblages of closed ceramic containers and pouring vessels presented in previous articles. During this test, published net height measurements were used to calculate rim diameters through photographs, which were then compared to real rim diameters included in the original publication by Coldstream (2010). According to the test (Chart 14), differences between real and calculated rim diameters range between -0.8 cm and +0.3 cm. Differences between real and calculated proportions of rim diameter to net height range between -9.7% and +4.8%. Similarly to the Kerameikos assemblage, rim diameter differences for the British Museum vessels range below ±1 cm; however, differences in the proportions of rim diameter to net height often exceed 5%, showing that calculated measurements are relatively biased. This is most likely due to the smaller size of these pots: a net height difference of -0.8 cm is insignificant with regard to a 45 cm amphora; however, it is quite significant when referring of a 6 cm skyphos. Such differences up to ±1 cm are likely to produce large variations in the percentages between different metrical features of small drinking cups as opposed to larger vessels. Similarly to the Kerameikos assemblage, the analyses results for the British Museum skyphoi must be treated with caution.

*Figure 11: Scatter-graph of correlation between net height and base diameter. Athenian/Attic skyphoi from the British Museum with complete profiles.*

According to the correlation of net height and base diameter in Figure 11, vessels appear scattered forming a loose central cluster comprised of 10 skyphoi. The only gadrooned sample appears on the cluster’s margins suggesting that is possibly the product of a distinct workshop. Furthermore, two vessels are loners: the first is a typical skyphos, GR1914,0407.1, with almost equal base diameter and net height resembling K10 and K88 from Kynosarges; the second vessel is the only wide skyphos with high lip in the entire assemblage (GR1977,1207.30). Similarly to the previous assemblages, the central cluster from the British Museum does not follow any specific chronological patterns. All vessels date between EGII and LGIa times.

A comparison of mean proportions for the British Museum skyphoi according to the data recorded in Chart 13 suggests high standard deviations and a large degree of artefact variability within the assemblage. The lowest standard deviations relate to the proportions of handle attachment height to net height, and base diameter to rim diameter. The same patterns have been observed in all previous assemblages; however, in this assemblage technological traditions are unlikely:

According to the scatter-graph in Figure 12, the regression line for the proportion of handle attachment height to net height for skyphoi is y = 0.5907x – 0.762 (where y = handle attachment height and x = net height). This means that their handles have been attached at roughly 59% of a vessel’s net height, while the difference of 0.762 cm is small to be considered. The coefficient of determination of this regression line (R2=0.7378) shows weak statistical correlation (73.78%).

Unlike the previous assemblages, the British Museum skyphoi show a weak pattern that is biased for two possible reasons: firstly, calculated measurements could be wrong due to vessel deformations or due to different recording strategies followed in the original publication by Coldstream (2010). Secondly, it could be that a large portion of this material is not Athenian and has been produced by following different conceptualisations compared to distinctively Athenian assemblages. In the second case, it is likely that artefacts from the Agora, Kynosarges and Kerameikos were shaped according to specific conceptions followed only by central Athenian workshops, which were not followed by workshops of the Athenian periphery. In this case, the assemblage from the British Museum, which is of unknown context, is likely to produce a biased regression line due to vessels coming from mixed production sites.

*Figure 12: Scatter-graph of correlation between net height and handle attachment height with regression line. Athenian/Attic skyphoi from the British Museum with complete profiles and recorded handles.*

The same thing is noted in Figure 13 with regard to the correlation between rim diameter and base diameter of the same assemblage. According to the graph, pottery appears highly scattered. The regression line follows the equation y = 0.4103x + 1.7688 and the coefficient of determination (R2=0.3238) shows weak statistical correlation (32.38%).

According to the above statistics, the material from the British Museum is diverse and cannot suggest any clear patterns. There is likelihood that this problem is due to its mixed origin, although calculation bias must not be overruled.

*Figure 13: Scatter graph of correlation between rim and base diameters with regression line. Athenian/Attic drinking vessels from the British Museum with complete profiles.*

### Analysis of fabrics

Fabric analysis is conducted on 12 skyphoi and 9 kantharoi from the Athenian Agora, the results of which are summarised in Chart 15. According to hand specimen examination, all kantharoi and the majority of skyphoi were produced from variant 1 and their Munsell (1975) colours primarily belong to the upper 5YR series. This is the same fabric as the one described for large ceramic containers and pouring vessels, and appears to be dominant across different Athenian fineware groups. Only two skyphoi, P32895 and P22431, have been produced from variant 2. Similarly to the assemblages of large ceramic containers and medium sized pouring vessels, variant 2 is the least popular, comprising 10% of the examined samples (Figure 14). Still, this percentage is significantly higher compared to the percentage of large ceramic containers and pouring vessels produced from the same variant. The only Corinthianising skyphos (P5286) is also made from variant 1. The use of a single fabric for the majority of drinking vessels verifies a strong technological tradition in clay selection, manipulation and tempering processes.

*Figure 14: Fabric division for 12 skyphoi from the Athenian Agora.*

### Analysis of decorative technology

The analysis of decorative technology is conducted on 18 kantharoi and 40 skyphoi (both complete vessels and sherds) coming from the Athenian Agora, the Kynosarges burials and the collections of the British School at Athens. Results are presented in Charts 16 and 17. Such analysis for small drinking pots requires a different approach compared to that for large ceramic containers and medium sized pouring vessels. By contrast to close-shaped finewares, kantharoi and skyphoi are significantly smaller and their shape is open; therefore, decorative elements, coatings and slips appear both on their external and internal surfaces simultaneously.

The vast majority of drinking vessels examined in this chapter were produced with internal coatings and this practice was followed without interruption across the entire Geometric era. Only exception is the SG footed low-handled kantharos P7196, which is completely uncoated and undecorated. Three other skyphoi, K2 and K3 from Kynosarges, and P12112 from the Agora, are not internally coated but internally decorated with various motifs. Still, their internal decoration covers all their surfaces and could have functioned similarly to a thick coating.

Internal coatings might have had practical use in drinking: they probably prevented the liquid content (perhaps wine) to come in contact with the porous clay surfaces of the vessel, resulting to alterations in flavour. Had this been the case though, then the flavour of the liquid content would have been already altered after contact with the clay surfaces of the pouring vessel or larger container that was served from, as such pots were internally uncoated. Another explanation could be that internal coatings served aesthetic purposes or functions related to the broader social notions behind drinking practices. Whatever the practical importance or social agency, coatings were essential elements of the technological chaîne opératoire of small drinking vessels.

The large majority of skyphoi and kantharoi analysed in this project are also coated on the lower half of their external surface, which limits any other form of decoration on the upper half of a vessel’s body. In that sense, all skyphoi and kantharoi (with exception of P7196) carry a form of treatment on both their external and internal surfaces.

According to Chart 16, all decorated kantharoi are painted in black or brown black colours (Colour Group 1). Two vessels are fully coated and they both belong to the early phases of the Geometric era: P19247 (EGI) and P6420 (MG). Finally, it is only 3 out of 18 vessels that carry no coated surfaces (P7476, P7196 and P1765) and all come from the 7th century BC (SG or EPA styles). According to Figure 15, uncoated vessels that have been treated with a thin matte wash comprise 17% of the entire assemblage.

*Figure 15: Comparison of slip or coating quality of Athenian/Attic kantharoi from the Agora, Kynosarges, and the British School’s at Athens collections.*

Despite the small sample, decorative characteristic of kantharoi show some similarities with closed ceramic containers and pouring vessels. Firstly, the dominant colour for decorative elements and coated surfaces is black or brown black. Secondly, although thick coatings were applied on the internal and lower external surfaces of such vessels across different chronological periods, the production of fully coated vessels was restricted during the earlier phases of the Geometric period. The main difference between kantharoi and other vessel classes is that undecorated and uncoated pots are noted after the end of the Geometric era, circa 700 BC; by contrast, the practice of coating in other vessel groups declines right after LGIa.

*Figure 16: Comparison of decorative element and coating colours of Athenian/Attic skyphoi from the Agora, Kynosarges, and the British School at Athens collections.*

The analysis of skyphoi shows that their decoration was less standardised to that of Kantharoi and some effort was carried out to produce bichrome vessels, in colours of Group 1 and 2. According to Chart 6.17 and Figure 28, 82% of the skyphoi are decorated in black or brown black colours. This pattern follows the general decorative norm observed on other vessel classes discussed in previous chapters. A relatively large percentage of skyphoi (10%) is decorated in red or brown red colours (Group 2) and this is not restricted during any specific chronological period. According to Chart 17, the pattern begins as early as MGII. One vessel (P8225) is coated with two different colours (red on its upper and black on its lower half), while two vessels (P8222 and P8223) carry red decorative elements on their upper body and black coatings on their lower body. All three vessels have been found in the same context (well D12:3) and belong to the MGII-LGIa transition; therefore, it is likely to be products of the same workshop if not the same painter.

Macroscopic analysis shows that such bichrome appearance should not be attributed to a random firing accident, rather to the painter’s conscious choice. Although it is not clear if the painter consciously aimed to produce a bichrome effect, it is more likely that there was a deliberate choice to spread thick layers of paint on the lower half of these vessels, and use a less thick and perhaps more diluted paint of similar composition for the upper half.

According to Chart 17, there are only 4 out of 40 skyphoi (10%) that are fully coated, all dating between MGII-LGIa and SG. The main norm for the majority of vessels is to have thick coatings (either lustrous or matte) on their lower halves. Still, this practice follows a distinct chronological pattern: according to Figure 17, during the period before MGII-LGIa 48% of the skyphoi have a lustrous external appearance; however, after LGIb this percentage drops down to zero and the vast majority of vessels carry thick matte coatings. The abandonment of lustrous in favour of matte coatings during the Late Geometric period is unlikely to coincide with a move towards simplicity in pottery production similar to the one noted for pouring vessels. Had this been the case, similarly to the external treatments of amphorae and oinochoai, both coating practices would have been abandoned simultaneously.

*Figure 17: Comparison of coating quality of Athenian/Attic skyphoi from the Agora, Kynosarges, and the British School at Athens collections before and after LGI.*

In conclusion, despite their broader standardisation skyphoi and kantharoi demonstrate a different approach from artisans during the chaîne opératoire of each ware. Skyphoi were more elaborately decorated compared to kantharoi or to any other fineware examined in this thesis, and subject to greater variability and experimentation with regard to colours and external treatments. In that sense, it appears likely that skyphoi painters enjoyed significant freedom compared to the painters involved in the decoration of other wares. By contrast, the decoration of kantharoi was standardised and followed the patterns noted in the decoration of large containers and medium sized pouring vessels, with only one difference: coating practices in the production of this group did not decline in the Late Geometric but during the 7th century BC.

### Summary and discussion of analysis of drinking vessels

The analysis of kantharoi and skyphoi shows that both vessel types follow similar conceptualisations in their shaping characteristics and must be treated as products of the same chaîne opératoire. Kantharoi exhibit greater degree of deformation compared to skyphoi because of having their handles attached adjacent to their rims. Their rim diameter measurements and related proportions needed to be left out due to potential bias.

The analysis of metrical features shows that kantharoi form distinct clusters according to their sizes. In these clusters, small kantharoi with heights below 8 cm and base diameters below 6 cm appear more standardised compared to their larger equivalents. Skyphoi form dense and robust triangular clusters which suggest lesser degree of artefact variability compared to kantharoi and a more standardised production. Some characteristic shapes such as wide skyphoi with stirrup handles and gadrooned skyphoi are plotted at the edges of these clusters, suggesting that their production and their chaîne opératoire might have been distinct. This is more likely for the case of gadrooned skyphoi, the production of which probably copied some conceptualisations from the chaîne opératoire of metallic vessels (see Borell 1978, 93-4 and Markoe 1985, 117-27). Despite this fact, such typological variants never diverted from the technological traditions followed in the production of other skyphoi, particularly with regard to their proportions and fabrics.

The analysis of metrical features reveals that the Kynosarges and British School material might have been produced at different workshops compared to that from the Agora. Furthermore, all drinking vessels from all five assemblages follow no distinct groupings based on their period of production and context of recovery. In that sense, their production must be regarded as homogeneous and standardised across time.

The analysis of proportions shows that some technological traditions might have existed in the partonomy of small drinking vessels, particularly in relation to their handle attachment heights. Regression lines suggest similar proportions of handle attachment height to net height for kantharoi and skyphoi, although the production of the second might have included two traditions that were followed simultaneously. More specifically, the handles of kantharoi were attached at roughly 70-71% of a vessel’s net height, which is translated as a fraction of 7/10. The statistical correlation for the Kerameikos assemblage is weaker compared to other assemblages, while no kantharoi from the collections of the British Museum were included in this article:

The handles of skyphoi from the Agora, Kynosarges and British School collections have been attached similarly at a fraction slightly below 7/10 of a vessel’s net height, and more specifically at 69.5%. However, the handles of skyphoi from Kerameikos show an average handle attachment height at about 67.3%:

Even though this percentage is close to the fraction of 7/10, it is also close to the fraction of 2/3 of a vessel’s net height, which was noted with regard to the same proportion for neck-handled amphorae. In this sense, the handle attachment height of the Kerameikos skyphoi could suggest a second technological tradition, in which handles were attached between 70% and 66% of a vessel’s net height; however, the regression for Kerameikos shows weaker statistical correlation compared to the Agora-Kynosarges-British School assemblage. The only assemblage that does not follow any clear pattern and is likely to suggest no distinct technological tradition is that from the British Museum: not only its average handle attachment height is low (59%), but also its degree of statistical correlation in not satisfactory (73.78%).

The material from the British Museum is problematic. In general, all mean proportions show high standard deviations suggesting a large degree of variability. It is likely that calculated measurements obtained for this material are biased due to different recording strategies followed in the original publication by Coldstream (2010) and the present macroscopic analysis. However, another explanation could be likely: the assemblage from the British Museum comprises of vessels of unknown provenance, some of which are Athenian and others that are broadly Attic. It is likely that the workshops operating in the Athenian periphery followed different conceptualisations in their chaînes opératoires by contrast to the central Athenian workshops. As the material from the British Museum is mixed, its proportional features are likely to divert from the main Athenian norm noted in the Agora and Kerameikos vessels; therefore, regressions are subject to weaker statistical correlation.

The material from the Agora (mainly) suggests a second possible technological tradition with regard to the proportion of base diameter to net height. This is set roughly at 45%; however, the vessels from Kerameikos and the British Museum do not verify the same pattern:

According to the above, the only strong technological tradition in the production of drinking vessels relates to their handle attachment heights. Exceptions are also common suggesting a degree of freedom and flexibility in vessel conceptualisation.

Hand specimen examination shows that the strongest technological tradition in the production of drinking vessels relates to clay selection, levigation and tempering practices. There is a strong preference in the use of fabric variant 1, which is also common in the production of larger wares such as containers and pouring vessels. Variant 2 is less popular and only noted in the production of some skyphoi. The only Corinthianising vessel in the entire assemblage (skyphos P5286) is also produced from variant 1.

In relation to their decorative characteristics, kantharoi appear more standardised compared to skyphoi and follow patterns noted in the decoration of other ware groups. Their decorative colours are black or brown black (Group 1) and their coating characteristics follow two chronological patterns. Firstly, fully coated kantharoi belong to the period before MGII-LGIa. This is also noted with regard to neck-handled amphorae and oinochoai in previous articles. Secondly, all uncoated and undecorated vessels come from the 7th century BC. By contrast, the decline of coating practices for other fineware groups takes place directly after MGII-LGIa, during the period that coincides with the generalised use of figurative style decoration.

The decoration of skyphoi is slightly different compared to kantharoi and shows a larger degree of variability. Even though the dominant decorative colour is black or brown black (Group 1), after MGII there appears a relatively large portion of artefacts decorated in red or red brown colours (Group 2). Even though skyphoi were never produced without coated surfaces, thick lustrous coatings were probably abandoned sometime during the beginning of the Late Geometric period, and were replaced by thick matte coatings. Furthermore, around MGII-LGIa there appear some conscious attempts from painters to produce bichrome vessels, bearing either two different coatings, or different combinations of colours used for coatings and decorative elements simultaneously. In general, this tendency towards polychromic decoration could relate to greater experimentation and freedom in the decoration of skyphoi as opposed to the decoration of other wares. Such variability could also be due to experienced painters involved in the production of skyphoi compared to those involved in the chaînes opératoires of kantharoi or even larger pots. Finally, given the large amount of drinking vessels produced by Geometric workshops, the work of painters and potters involved in such production was probably intense and demanding; however, there was room for experimentation in their technological choices, suggesting a similar degree of artistic freedom in their work.