image/svg+xml
143
XII/2/2021
INTERDISCIPLINARIA ARCHAEOLOGICA
NATURAL SCIENCES IN ARCHAEOLOGY
homepage: http://www.iansa.eu
Orientation Patterns Characteristic for the Structure of the Ceramic Body
of Wheel-thrown Pottery
Richard Thér
1*
, Petr Toms
2
1
Department of Archaeology, Philosophical Faculty, University of Hradec Králové, Rokitanského 62, 500 03 Hradec Králové, Czech Republic
2
Private researcher, Machovská Lhota 71, 549 63 Machov, Czech Republic
1. Introduction
During the past decade, we have been developing
a methodology based on quantifcation of the orientation and
alignment of the components of a ceramic body as one of
the principal features refecting pottery-forming techniques
that are theoretically observable on every sherd (Thér,
2016; Thér
et al.
, 2019; Thér and Toms, 2016). Many of the
phenomena that occur on the surface of pottery fragments
and can be related to pottery-forming practices are randomly
preserved, and their interpretation is further complicated by
the common practice of combining several techniques during
the forming and fnishing of vessels. One diagnostic attribute
can, at least theoretically, be observed on every ceramic sherd
– the orientation of the structure of the ceramic body. The
relationship between forming techniques and the orientation
of the components of the ceramic material has long been
recognised (Balfet, 1953; Bordet and Courtois, 1967; Felts,
1942; Giford, 1928; Linné, 1925, p.33; Shepard, 1956,
pp.183–184). The application of physical force to the plastic
clay during forming is the main factor afecting the alignment
of the components. The resulting orientation and alignment
are characteristic of each forming method, although some
orientation patterns might result from more than one
fabrication process (for an overview of the assumptions for
particular techniques see Berg, 2008, Figure 1; Carr, 1990;
Courty and Roux, 1995, Table 1; Livingstone Smith, 2007,
pp.88–146; Middleton, 2005, Figure 4.8; Pierret, 1995,
pp.46–50; Roux, 2019, Figure 3.20; Rye, 1981, pp.58–89;
Thér, 2020, Figure 9; Whitbread, 1996).
Measurement of the orientation refnes the analysis of
preferred orientation by defning the exact intervals of
orientation variability for the individual forming techniques
and their combinations. For the measurements, we selected
two basic sections: sections perpendicular to the wall
surface in the plane parallel to the vessel height (hereinafter
referred to as a
radial section
) and sections tangential to the
vessel wall cut through a core zone of the wall (hereinafter
referred to as a
tangential section
). Originally, we captured
three transects approx. 6 mm wide in each thin section at
a magnifcation of 40 times in plane-polarised light using
a standard petrographic microscope. The resultant images
have a resolution of 1.09 μm. Then inclusions and voids were
extracted using object extraction and separation methods in
Volume XII ● Issue 2/2021 ● Pages 143–154
*Corresponding author. E-mail: richard.ther@uhk.cz
ARTICLE INFO
Article history:
Received: 19
th
February 2021
Accepted: 6
th
October 2021
DOI: http://dx.doi.org/10.24916/iansa.2021.2.3
Key words:
orientation analysis
wheel throwing
pottery forming
image analysis
thin section petrography
ABSTRACT
The described analysis follows recent fndings related to the orientation of particles and voids in
a ceramic body that is characteristic for wheel-made pottery. The analysis is focused on the potential
variability within wheel-throwing method and is based on an experimental collection that combines the
factors of the experience and motor habits of individual potters and the vessel shape. The orientation of
the components of a ceramic body is calculated for two sections: radial and tangential. The sections are
analysed using optical microscopy. The calculated orientation and alignment refect the throwing style
of potters using the same forming method.
image/svg+xml
IANSA 2021 ● XII/2 ● 143–154
Richard Thér, Petr Toms: Orientation Patterns Characteristic for the Structure of the Ceramic Body of Wheel-thrown Pottery
144
Table 1.
Orientation analysis results for experimental samples taken in tangential and radial sections. MD – Mean direction, CSD – Circular standard
deviation.
Sample
Min.
thickness
Max.
thickness
Dif. in
thickness
ShapeAuthorWheel
Radial sectionsTangential sections
MDCSDMDCSD
1379749101113BowlHenryMotorised735
28
36
2
4048
5001953BowlHenryMotorised5331930
339455125
1180
BowlHenryMotorised33627
38
4
48065807
1001BowlHenryMotorised3352731
5
4833
5617
784
BowlHenryMotorised3394123
65022
5587
565BowlHenryMotorised4353437
7
3809
4265456BowlHenryMotorised4392231
8
34924377
885
BowlHenryMotorised136
38
37
936724415743BowlHenryMotorised4332534
10
4080
4550470BowlHenryMotorised5335131
114026
4318
292BowlHenryMotorised733
38
35
1240434393350BowlHenryMotorised1354231
13
3830
4021191BowlHenryMotorised13239
38
1436723979307BowlHenryMotorised6314235
15
3784
4373
589
BowlHenryMotorised7302330
1636164229613Conical v.HenryMotorised15
28
2133
1733113742431Conical v.HenryMotorised19263734
18
35104471961Conical v.HenryMotorised14273135
19437256511279Conical v.HenryMotorised143439
38
20395651561200Conical v.HenryMotorised9334533
2140405329
1289
Conical v.HenryMotorised13323535
2243954700305Conical v.HenryMotorised3042
18
17
2339344593659Conical v.HenryMotorised24391921
244197
4538
341Conical v.HenryMotorised25401920
2544464970524Conical v.HenryMotorised13293231
2645495163614Conical v.HenryMotorised15
28
4341
2744765379903Conical v.HenryMotorised143143
38
28
321351751962Conical v.HenryMotorised17352324
2945545529975Conical v.HenryMotorised22322622
30461759771360Conical v.HenryMotorised16332629
31
2680
3270590BowlPeterMotorised5292933
323150317727BowlPeterMotorised53216
38
3331513617466BowlPeterMotorised
8
321934
3439714362391BowlPeterMotorised9293736
3532944134
840
BowlPeterMotorised
8
324540
363417
3859
442BowlPeterMotorised10304740
3737434146403BowlPeterMotorised0
38
24
28
383658
4372714BowlPeterMotorised232
38
43
393666
4086
420BowlPeterMotorised
8
35
28
30
40
3738
4265527BowlPeterMotorised
8
33
18
29
41
3598
4154556BowlPeterMotorised11342433
42
38714282
411BowlPeterMotorised12312432
43
3087
42341147BowlPeterMotorised7
38
2032
443691
4187
496BowlPeterMotorised9
38
3225
453727
4185458
BowlPeterMotorised11342726
46
4849
5931
1082
Conical v.PeterMotorised13372120
47
4885
5,10E+03215Conical v.PeterMotorised12
28
2624
48
42275013
786
Conical v.PeterMotorised6303523
4940945015921Conical v.PeterMotorised12302130
5045505016466Conical v.PeterMotorised36452130
5146435404761Conical v.PeterMotorised
8
343227
image/svg+xml
IANSA 2021 ● XII/2 ● 143–154
Richard Thér, Petr Toms: Orientation Patterns Characteristic for the Structure of the Ceramic Body of Wheel-thrown Pottery
145
Sample
Min.
thickness
Max.
thickness
Dif. in
thickness
ShapeAuthorWheel
Radial sectionsTangential sections
MDCSDMDCSD
5246205395775Conical v.PeterMotorised26342021
53
40384897859
Conical v.PeterMotorised223013
18
5444905162672Conical v.PeterMotorised
18
441326
5550616,15E+031091Conical v.PeterMotorised13
38
2229
5646195163544Conical v.PeterMotorised5272121
5746015274673Conical v.PeterMotorised11252124
583890
4120230Conical v.PeterMotorised
18
402627
5947165365649Conical v.PeterMotorised27512722
60
4783
5099316Conical v.PeterMotorised7313120
61
88249686862
Conical v.PeterFlywheel21341735
62
8832
9545713Conical v.PeterFlywheel193615
28
63
8981
9604623Conical v.PeterFlywheel2240
1828
64
8493
101591666Conical v.PeterFlywheel1336427
65
78708400
530Conical v.PeterFlywheel16351231
66
860711098
2491Conical v.PeterFlywheel17331731
67396560192054BowlThomasMotorised1030641
68
39165497
1581
BowlThomasMotorised12341639
69355455111957BowlThomasMotorised334532
70421664512235BowlThomasMotorised9351734
71434964752126BowlThomasMotorised
18
331132
723926
58161890
BowlThomasMotorised163517
28
73370271273425BowlThomasMotorised144013
28
74394766062659BowlThomasMotorised15371132
75377163352564BowlThomasMotorised23361132
76
4282
52911009BowlThomasMotorised9401933
774256
5658
1402BowlThomasMotorised13441331
78
4364
6078
1714BowlThomasMotorised11412029
79327357422469BowlThomasMotorised
8
32736
80
29545773
2819
BowlThomasMotorised
8
351135
81
354264772935BowlThomasMotorised1139935
82
6513
81941681
Conical v.ThomasMotorised13403539
83
62197205
986
Conical v.ThomasMotorised2241
18
36
84
6471
8063
1592Conical v.ThomasMotorised11361029
85
67737352579Conical v.ThomasMotorised15401737
86
67157233
518
Conical v.ThomasMotorised20
372436
87
71207343223Conical v.ThomasMotorised203716
38
88
7179
8131
952Conical v.ThomasMotorised19401130
89
7535
8166
631Conical v.ThomasMotorised16412937
907305
8214
909Conical v.ThomasMotorised19422537
91
6786
7729943Conical v.ThomasMotorised1139
8
39
9267757370595Conical v.ThomasMotorised
8
361640
9369697340371Conical v.ThomasMotorised
8
36
18
34
94792390601137Conical v.ThomasMotorised
8
3514
28
95747791141637Conical v.ThomasMotorised21332230
96
84768922
446Conical v.ThomasMotorised19331739
Table 1.
Orientation analysis results for experimental samples taken in tangential and radial sections. MD – Mean direction, CSD – Circular standard
deviation. (
Continuation
)
JMicroVision software (Roduit, 2014). Two basic measures
were chosen to express the object orientation: (a) mean
direction (MD) – average orientation of objects, and (b)
circular standard deviation (CSD) – the dispersion of the
values from the average (Fisher, 1993, pp.75–78; Mardia and
Jupp, 2000, pp.15–19).
In the frst experimental collection, we found several
signifcant markers distinguishing wheel fnishing,
image/svg+xml
IANSA 2021 ● XII/2 ● 143–154
Richard Thér, Petr Toms: Orientation Patterns Characteristic for the Structure of the Ceramic Body of Wheel-thrown Pottery
146
wheel shaping, and wheel throwing as basic levels of the
contribution of rotational movement in pottery forming
1
,
especially in the mean directions in core areas of radial
sections, in CSD in core areas of radial sections or the mean
direction in tangential sections (Thér, 2016).
In the second experimental dataset, we focused directly
on the distinctions among diferent uses of the potter’s
1
There are two basic ways to classify variants of the application of rotational
movement in the pottery-forming sequence. The frst approach classifes
individual combinations of the techniques applied at diferent stages of
the forming. The forming methods are then referred to as, for example,
wheel coiling or wheel moulding (Berg, 2009; Roux, 2019; 2017; Rückl
and Jacobs, 2016; Thér and Toms, 2016). An
alternative approach is to
separately defne the variants of the use of rotational movement and defne
them independently of the other techniques (Berg, 2008; 2007; Choleva,
2012; Courty and Roux, 1995; Henrickson, 1991; Roux, 2003; Roux and
Courty, 1998; Thér, 2016; Thér
et al.
, 2017; Thér and Toms, 2016). The
diferences in the contribution of rotational movement to the whole forming
sequence are the main criterion in this classifcation:
(a)
Wheel fnishing
. The vessel is formed by some hand-building technique
and subsequently the rotational movement is used for surface modifcations
and minor shape corrections,
i.e.
only in the fnishing stage.
(b)
Wheel shaping
. A roughout of the vessel is formed by some hand-
building technique and subsequently rotational kinetic energy (RKE) is used
to shape and thin the vessel walls. This technique can be used in assembling
and fnishing the vessel.
(c)
Wheel throwing
. The entire forming sequence is performed using RKE.
The main interest of the orientation analysis is to defne the relation between
the contribution of rotational movement in forming and orientation patterns:
thus, we use the second approach to classifcation.
wheel. In this dataset, we evaluated the efect of the degree
of transformation of the clay mass, the shape of the vessel,
the velocity of rotation or the individual experience and
skills of the potter. The principal fnding of the analysis of
the second experimental collection was that the specifc
characteristics of the orientation of wheel-thrown samples
are developed especially in the lower parts of the vessels.
The signifcant diference between the results obtained from
lower and upper parts of the experimental vessels can be seen
especially in the tangential sections. The diference is due to
the fact that the lower part of the vessel undergoes a strong
transformation when the potter creates a basic form prepared
for lifting. While she/he lifts the clay mass upward, the rest
of the clay is lifted above the fngers but is not afected by
their movement (Thér and Toms, 2016, pp.38–39).
The analysis of the second experimental series also
confrmed the observation made in the frst experimental
series, namely that the upper ends of the objects in the
marginal zones of wheel-thrown
pottery incline inwards
towards the core of the wall (Figure 1). We called this
phenomenon “imbricate pattern” and suggested that
this pattern is caused by shear stress induced by upward
movements of the fngers during wheel throwing. The clay
mass in the margins moves more quickly during lifting than
the mass in the core of the wall. Therefore, marginal zones
can be seen as shear zones with a predominance of shear
stress. The comparison of internal and external areas shows
that the inclination of the inclusions and voids inwards is
more strongly developed in the external area. We explained
this phenomenon by the disproportion of the forces required
on the interior and exterior of the vessel, which causes larger
shear deformation on the exterior area of the vessel wall and
subsequently a more pronounced imbricate pattern in this
area (Thér and Toms, 2016, p.38).
In the third experimental series described in this study,
we focused solely on the orientation patterns resulting from
wheel throwing and especially on those variables whose
signifcant efect became the subject of hypotheses after
evaluating the previous series.
a) Above all, the shape of the vessel is important. The
analysis suggested that the shape signifcantly infuences
the orientation parameters. Samples taken from the oblate
ellipsoid fashioned in the second experimental series showed
below-average CSD values in radial sections from the lower
parts of the vessels but, more importantly, a signifcant
increase in CSD and lesser deviation from the horizontal
axis in tangential sections (Thér and Toms, 2016, Figures 5
and 7). The distortion from typical wheel-throwing values
for conical shapes could be hypothetically proportional to
the degree of transformation that is required to fnish the
shape of the vessel extra to the lifting of the clay.
b) The second experimental series also showed that the
orientation patterns refect the equilibrium established between
the potter´s actions and tools she/he uses during forming.
If the potters use an unfamiliar clay or rotational device or
throw an unusual shape, they disturb the equilibrium gained
by experience and thus also the alignment typical for the
Figure 1.
Imbricate pattern – orientation pattern typical for wheel throwing
observed in radial sections. The upper ends of the objects in the marginal
zones of wheel-thrown pottery incline inwards towards the core of the wall.
image/svg+xml
IANSA 2021 ● XII/2 ● 143–154
Richard Thér, Petr Toms: Orientation Patterns Characteristic for the Structure of the Ceramic Body of Wheel-thrown Pottery
147
technique. This especially applies to the beginner for whom
all the components of the technique are new (Thér and Toms,
2016, Figure 7). In this current, third experimental series we
compared three professional potters who routinely produce
pottery, to see whether the results are comparable when the
potters have (a) a similar, high level of skill, (b) create shapes
that do not difer signifcantly from what they are used to
forming on a wheel, and (c) use familiar tools,
i.e.
potters are
in equilibrium with their working environment.
2. Materials and method
The third experimental collection is focused on the variability
of orientation patterns within the wheel-throwing method.
So far, one principal experienced potter with 23 years of
experience in wheel throwing, Peter Toms, was employed in
our experiments. Along with Petr Toms (hereinafter referred
to as Peter) we included two other professional potters: Jiří
Lang (hereinafter referred to as Henry) and Tomáš Macek
(hereinafter referred to as Thomas).
Two diferent vessel shapes were replicated: a simple
conical vessel 180 mm in height and 200 mm in diameter at
the top and an S-shaped bowl 85 mm in height and 200 mm in
diameter at the top (depicted in Figure 2). The S-shaped bowl
was chosen because, in our application of the methodology,
we are dealing mainly with Late Iron Age pottery in Central
Europe, and this is the most common shape of wheel-made
pottery in this context.
Each potter formed 15 slightly conical pots and 15 S-shaped
bowls. The target wall thickness for all the containers was
5 mm. No other parameters of the forming method were
specifed in order not to force the potters to employ motions
that are not “natural” for them. All the potters used their
wheels (motor-driven) and the same fne-grained commercial
clay – Witgert 10. The experimental collection was created
during one session in one pottery workshop after the potters
became acquainted with the selected pottery shapes. The
speed of the wheels was measured by a laser tachometer.
The dataset was complemented by six conical vessels
thrown by Peter on a replica of a fywheel made of a wooden-
spoked wheel. The device is located in the Archaeological
park of prehistory in Všestary (Czech Republic). Peter
does not work on this wheel on a regular basis and there
was a minor technical problem related to ftting the wheel
socket in the axis which caused vibrations of the wheel when
a certain speed was reached.
Two oriented thin sections were cut from the lower
body of each experimental vessel: tangential and radial
(Figure 2). The entire area of each thin section was recorded
at a magnifcation of 200× using a Keyence VHX6000
digital microscope. The resultant images have a resolution of
1.11 μm. The analysis followed the published methodology
(Thér, 2016; Thér and Toms, 2016), except for the software
treatment. The components of the ceramic materials were
extracted using automatic area measurement tools available
in the Keyence VHX6000 measurement software. The range
of threshold values chosen to separate inclusion and void
representations was based primarily on colour saturation,
which shows the best results for the thin sections with
uneven thickness (resulting in uneven brightness of the
captured image).
The extracted objects in the radial sections were analysed
only in the external zones of the section (one-third of the
thickness adjacent to the outer edge). The focus on the
external area follows the results of the analysis of the