You are on page 1of 155

Masarykova univerzita

Filozofická fakulta

Ústav archeologie a muzeologie


Klasická archeologie

Bc. Tomáš Pavloň

The Ceramic Building Materials and Their Use


in Roman and Late Antique Greece

Magisterská diplomová práce

Vedoucí práce: Mgr. Věra Klontza, Ph.D.

2015
Prohlašuji, že jsem magisterskou diplomovou práci vypracoval
samostatně s využitím uvedených pramenů a literatury.

……………………………………………………………………..
Tomáš Pavloň
Rád bych tímto poděkoval zejména Mgr. Věře Klontze, Ph.D. za cenné rady a
komentáře, které mi poskytla v průběhu psaní magisterské práce. Dále i pracovníkům
Priniatikos Pyrgos Project, kteří mi umožnili účastnit se této archeologické expedice.
V neposlední řadě dekuji své rodině a přátelům za neutuchající podporu a inspiraci.
Tomáš Pavloň
Contents

Contents ................................................................................................................................................... 5
1. Introduction ..................................................................................................................................... 8
2. Method, methodology, aim............................................................................................................ 10
3. CBM phenomenon background ..................................................................................................... 13
3.1. Definition of CBM .................................................................................................... 13
3.2. Chemical background ................................................................................................ 13
4. History of CBM production in Antiquity ...................................................................................... 15
4.1. History of tiles ........................................................................................................... 15
4.2. History of bricks ........................................................................................................ 20
4.3. Administration of production and construction ......................................................... 22
4.3.1. Roman Empire ..................................................................................................... 22
4.3.2. Late Antiquity...................................................................................................... 23
4.4. Production – Kilns ..................................................................................................... 24
5. Typology of Roman and Late Antique CBM ................................................................................ 29
5.1. Bricks......................................................................................................................... 29
5.1.1. Bessalis ................................................................................................................ 30
5.1.2. Pedalis ................................................................................................................. 31
5.1.3. Tetradoron ........................................................................................................... 32
5.1.4. Pentadoron .......................................................................................................... 32
5.1.5. Lydion .................................................................................................................. 32
5.1.6. Sesquipedalis ....................................................................................................... 33
5.1.7. Bipedalis .............................................................................................................. 34
5.1.8. Byzantine brick.................................................................................................... 35
5.1.9. Cuneatus / Solid voussoir .................................................................................... 36
5.1.10. Armchair vousoir ............................................................................................... 36
5.1.11. Triangular brick ................................................................................................. 38
5.1.12. Column brick ..................................................................................................... 38
5.2. Tiles ........................................................................................................................... 40
5.2.1. Roof tiles ............................................................................................................... 40
5.2.1.1. Tegula ............................................................................................................... 41
5.2.1.2. Imbrex .............................................................................................................. 43
5.2.1.3. Ridge tile .......................................................................................................... 45
5.2.1.4. Other roof installations ..................................................................................... 45
5.2.2. Paving tile/brick..................................................................................................... 47
5.2.3. Wall tile/brick for special purpose......................................................................... 48
5.2.3.1. Parietalis .......................................................................................................... 48
5.2.3.2. Tegula mammata .............................................................................................. 48
5.2.3.3. Tubulus ............................................................................................................. 49
5.2.4. Tubes and pipes ..................................................................................................... 50
5.2.4.1. Vaulting tube .................................................................................................... 51
5.2.4.2. Water pipe ........................................................................................................ 51
6. Application – construction techniques .......................................................................................... 53
6.1. CBM in wall construction ......................................................................................... 53
6.1.1. Foundations ......................................................................................................... 54
6.1.2. Brick faced concrete walls................................................................................... 54
6.1.3. Solid brick walls .................................................................................................. 61
6.1.4. Alternate stone and brick bands........................................................................... 62
6.2. Arches and vaults ...................................................................................................... 63
6.3. Floor pavement .......................................................................................................... 67
6.4. Roof construction ...................................................................................................... 68
7. Case studies of CBM in archaeological environment ................................................................... 70
7.1. Kopetra ...................................................................................................................... 70
7.2. Knossos ..................................................................................................................... 75
7.3. Nichoria ..................................................................................................................... 79
7.4. Gortyn ........................................................................................................................ 82
7.5. Oxa ............................................................................................................................ 85
8. Case study Priniatikos Pyrgos ....................................................................................................... 87
8.1. Introduction ............................................................................................................... 87
8.2. History a general information about the site ............................................................. 88
8.3. Investigation of Trench II and related deposits ......................................................... 92
8.4. Methodology ............................................................................................................. 94
8.5. Fabrics ....................................................................................................................... 96
8.6. Shapes...................................................................................................................... 105
8.6.1. Bricks................................................................................................................. 105
8.6.2. Roof Tiles .......................................................................................................... 109
8.6.2.1. Pan tiles .......................................................................................................... 109
8.6.2.2. Cover tiles....................................................................................................... 118
8.7. Signs and other marks ............................................................................................. 120
8.7.1. Signs .................................................................................................................. 121
8.7.2. Finger-marks and grooves ................................................................................. 123
8.7.3. Finger grooves on cover tiles............................................................................. 123
8.7.4. Finger grooves on pan tiles................................................................................ 124
8.8. Hypothetical reconstruction of the roof settings...................................................... 127
9. Conclusion ................................................................................................................................... 132
9.1. CBM in local economy ............................................................................................ 132
9.2. CBM - an archaeological source for interpreting the past ...................................... 136
10. Resumé .................................................................................................................................... 139
11. Bibliography ............................................................................................................................ 144
11.1. Online Sources..................................................................................................... 151
12. List of abbreviations ................................................................................................................ 152
13. Illustrations .............................................................................................................................. 153
13.1. List of Plates ........................................................................................................ 153
13.2. List of Figures...................................................................................................... 153
1. Introduction

The aim of this thesis is to bring wide range of tools, which contemporary and historical

approaches brought into the study of ceramic building material (CBM) in archaeological point

of view, examine its potential to help to understand the architecture, economy, crafts and life

of societies in late antiquity, based on detailed study of ceramic building material from the site

of Priniatikos Pyrgos in East Crete.

The CBM material was, especially in the past archaeological practice, seen as bulky and

therefor left behind completely, or only mentioned briefly in complex studies based on

archaeological survey and excavations, mostly due to its fragmentary state of preservation, huge

quantities without significant differences or illusory “vagueness” compared to other classes of

artefacts. Huge quantities of CBM however lately started to play important role in studying

economic relations of past societies, whose enormous efforts in the meaning of production and

acquiring of such amounts of material can be only understood with contribution of data based

on these structural components.

This study has ambitions to be such contribution, and even though it is not the first study

of this kind, it still occurs as pioneer in this field, and thus is using methods borrowed from

well-established pottery studies (or archaeological ceramology), architectural description and

other approaches to archaeological material, to improve our knowledge on the field of CBM.

In contrast to pottery, which always enjoyed the major interest of researchers (and for

good reasons, of course), CBM is not so rich in varieties, shapes and types, but although the

differences are small, still there are some that cannot be denied. Except from the form, the

second important characteristic of CBM, measurable and comparable is the composition and

8
fabrics, which cannot be investigated via modern methods and experience from other scientific

fields.

The questions raised at the beginning of the study are quite simple, but to find the

responses will be long journey through theoretical to more practical topics: How did the CBM

of Roman and Late Antique Mediterranean look like? What was its role as structural component

in constructions? How can studying of CBM help us to understand the past societies, their

architecture and their economies, technologies?

9
2. Method, methodology, aim

The structure of this study can be concluded as deductive, starting with theoretical

premises obtained from long tradition of architectural-historical approach to the antiquity and

techniques used by ancient builders, through the comparison between the archaeological sites

from the area of interest and finally focusing on CBM from Priniatikos Pyrgos site, which is

carefully studied with contribution of tools and experience previously examined in theoretical

chapters.

The area of interest is Aegean region with focus on archaeological environment Crete.

Sites for comparison were chosen from the island itself and neighbouring areas in Aegean

region, which had similar historical course, and thus the archaeological potential is supposed to

be similar. The study of typologies of CBM and construction techniques are derived from

different parts of Roman Empire and Byzantine tradition. The reason for this borrowing

especially in the case of CBM typology is the disproportion of these typological studies in

eastern and western part of former Roman Empire. For Roman west, the typology is well-

developed, based on research in provinces1 and the region of Italy2. In the east the situation is

complicated by influences from other regions and local architectural tradition – the use of CBM

has older roots than Roman. For construction techniques, the situation is similar – in regions

without highly developed pre-Roman civilization, the process of Romanisation was smoother

also in construction techniques used, with disproportion caused by material available, but in

technologically developed areas, like Greece in our case, the influence of Roman domination

was confronted with strong local tradition, however it definitely left its traces in architectural

1
Notable progress in CBM research was made in Britain, and the methodology from this area (Brodribb 1987;
McWhirr 1979) is adopted by scholars studying other parts of Roman Empire (Dodge 1987; Mills 2013).
2
See for example Helen 1975.

10
concept3, and thus the methods are described on examples found in the whole Roman territory

and then the examples from Aegean region are examined.

Time frame covers the period after Roman conquest of Greece, throughout their

presence here towards the fall of western part of the Empire and transformation of power

structures under the Byzantine rule in Early Byzantine period. The absolute chronological

definition of this time frame is problematic, as the archaeological periods could differ from site

to site even within boundaries of small region like Crete (Gallimore 2011, 33; Klontza-Jaklová

2015, Tab. 1). For archaeological approach in study of Aegean region, I find the most suiting

the chronology proposed by John Bintliff in his summarizing work (Bintliff 2012, 6) 4, with

contribution of chronology based on research at Knossos archaeological site (Sweetman 2004,

317) which puts the break between Late Roman and Late Antique period around the year 500

AD5, and the chronology based on Late Roman to Early Byzantine pottery examination from

Priniatikos Pyrgos site, which clears the break between Late Roman and Early Byzantine phases

(Klontza-Jaklová 2015, Tab. 2)6:

Early Hellenistic 323 BC – ca. 200 BC

Late Hellenistic to Early Roman ca. 200 BC – ca. 200 AD

Middle to Late Roman ca. 200 AD – end of 6th century AD

Late Antique/Early Byzantine end of 5th century AD – beginning of 9th

century AD

Plate 1) Chronological overview for this study.

3
For Roman architectural influence in Greece, see Macready (ed.) 1987.
4
Early Hellenistic 323 BC – ca. 200 BC; Late Hellenistic to Early Roman ca 200 BC – ca. 200 AD; Middle to Late
Roman ca. 200 AD – ca. 650 AD; Early Byzantine ca. 650 AD – 842 AD (Bintliff 2012, pp. 6)
5
Late Antique 5th century – 827 AD (Sweetman 2004, pp. 317).
6
Late Roman 4th century – end of 6th century AD; Early (First) Byzantine phase 1. end of 6th century – mid 7th
century AD; Early (First) Byzantine phase 2. mid 7th century – beginning of 9th century AD (Klontza-Jaklová 2015,
Tab. 2).

11
Chapters about definition and history are focused on CBM as an artefact and structural

component, first emergence and history of its use in investigated area. Various formation

processes standing behind the immense amount of CBM elements are examined in parts about

administration of production and the possible forms of production itself. As the base for further

understanding of CBM, the detailed typology of various forms of particular ceramic elements

is presented summarizing known CBM types used all over the Roman and Byzantine area.

Description, typical attributes, environs of their employment to construction and other than

primary use is studied. Next chapter applies the previously examined information into the more

complex study of particular techniques employing the CBM to form useful structures. This

chapter is crucial to obtain the broader point of view on material examined in case studies,

which are following: At first, the case studies from few chosen sites in Aegean region are briefly

concluded to obtain the better picture of differences between theoretical background and

particular archaeological material, which is often fragmentized and difficult to understand

without wider knowledge about the topic. The chapter about Priniatikos Pyrgos case focuses on

close research of CBM from this site considering the stratigraphy of the site, fabrics and form

of CBM elements, and data from other scientific fields applied in the expedition. The standard

for bibliographical referencing is adopted from Hesperia: The Journal of the American School

of Classical Studies at Athens.

Many thanks belong to the director Barry Molloy, co-directors Joanna Day and

especially to Vera Klontza-Jaklova, who brought me to Priniatikos Pyrgos Project and with the

others helped me to start and finish my research, and Sue Bridgeford who leads the stratigraphic

studies. Many thanks belong also to other participants of the project, who shared their inspiring

thoughts, enthusiasm and positive approach with me.

12
3. CBM phenomenon background

3.1. Definition of CBM

Ceramic building material (CBM) is defined as: “an artificial product made in replicate

units for building construction, each unit capable of being put into position by hand or hands of

one person”. (Harley 1974, 63 in: Mills 2013, 3). In these days the distinction between particular

types is much more clear than in antiquity – we have bricks, roof tiles, floor tiles, wall tiles,

tubes, etc. But for example Roman bricks were much thinner and wider, compared to

contemporary material. For this reason we can use Harley´s formula (Harley 1974, 70 in: Mills

2013, 5), which takes the width plus the length of the piece and divides it by thickness of the

piece. If the result is less than 8, the piece can be named “brick”, if it is more than 8, it is “tile”

(Mills, 2013, 5).

3.2. Chemical background

Chemical changes in the material take place, because of high temperatures towards

which is material exposed. In first stage up to 100 °C, the water (which is bon-chemically

bounded) is vaporized from components in firing due to atmospheric pressure and humidity.

The water transforms into steam and leaves the bricks/tiles. In the next step up to 200 °C, the

organic inclusions, usually straw additives, are burnt and transformed into for example carbon.

In the third step between approximately 400 °C and 700 °C the major change happens, when

the mud develops into terracotta, while the chemically bounded water leaves the material. At

the temperature around 800 °C, the carbon and other elements are burned out during oxidation,

when these elements react with oxygen and leave the material as gasses. At 900 °C the pores

are filled, because the material s vitrified and at first the surface is transformed into glass layer

(Wright, 2005, 114-115). As the output of this processes, the final terracotta material contains
13
changed particles bounded together by vitrification, the connections are stronger than in

mudbricks, terracotta is 15% denser than mudbrick, impermeable and able to sustain more

pressure. At the same time is light compared to stone tiles and tougher than some of them, for

example limestone tiles.

14
4. History of CBM production in Antiquity

4.1. History of tiles

Due to ancient Greek tradition, the first tiles were introduced by king Kinyras from

Cyprus, but true emergence of this type of artefact can be traced to different part of Greece.

Roof tiles can be found in archaeological evidence from Early Helladic period – on Greek

mainland, particularly in Lerna, Tiryns, Asine and Malthi (Wikander 1988, 204). The myth

about Cypriote origin is very unlikely to be based on reality, as the roof tiles emerged on the

island later than in other Greek regions and were used only occasionally, compared to the rest

of the Hellenised areas, where the roof tiles were much more popular. The first Helladic roof

tiles were rather substitutes for schist tiles, as they have the same dimensions as them and design

is very simple, resembling the original stone examples. On Early Helladic sites, the ceramic

tiles are minority in large amounts of stone slates. Roof tiles are small 15-25 cm long and flat,

bearing rectangular or oblong shape. After this period, ceramic roof tiles disappear from

archaeological material until their re-emergence in Mycenaean culture.

Mycenaean roof tiles are remarkably larger – 50 to 64 cm long and 30 to 40 cm wide,

slightly trapezoidal so the higher adjacent tile fit to the lower one. They are flat with raised

borders resembling to some extent the later so-called Corinthian pan tiles. It is not clear if the

gaps between pan tiles were covered with curved cover-tiles, because only fragments of the

latter were found. The sites with these tiles are Berbati, Thebes, Tiryns and Athens. However

the roof tiles were evidently known and used in this period, they were not the main roof covering

material, because they are found in small numbers on these sites.

The third appearance of roof tiles in Greece can be traced to early 7th century BC and

since then the continuity and evolution of this CBM is clearly going through the Classical,

15
Hellenistic, Roman and Byzantine Greece to the Middle Ages. Pan tiles, cover tiles, ridge tiles

and hip tiles were found on sites in Corinth, Isthmia, Perachora and Delphi, emerging in the

same time in many places with remarkable point of development. Almost all described early

tiles from this period belong to the monumental buildings – temples. It can be explained by

increased need to protect these buildings against the danger of fire7, or (and) by economical

connection of roof-tile industry and large building projects to public funds. Yet it does not

mean, that private houses were not roofed with tiles, but archaeological evidence is scarce in

this case (Wikander 1988, 205).

Three main types emerge and are named after the regions, where these particular types

were dominant:

1. Laconian roof combines curved concave pan tile (tegula), covered with semi-circular

imbrex. This type was original found in areas of Peloponnesus and Delphi. Later was very

popular for private buildings in Laconia (Mills 2013, 4). Continuity in use of this roofing

tradition remained strong in Balkan peninsula and its simpler variant can be seen even today

in Balkan countries and Greece, but instead of large pan tile, another reversed imbrex is

used.

7
Thatched roofs were still the most spread roofing system.

16
Fig. 1) Laconian system (from Åkeström 1966, Abb. 64 in Wikander 1988, Fig. 6).

2. Corinthian style is based on flat pan tile with inset flanges on both sides, smaller flange on

the top and “hook” at the lower side from below to attach to another row. Semi-circular

covering imbrex is faceted8, creating three to five planes. Corinthian roofing was originaly

used or monumental buildings and temples in Central Greece, but remained in use till the

end of antiquity. Corinthian style is more fitting for straight roofs without many turns and

is inconvenient for later Roman and byzantine domes.

Fig. 2) Corinthian system (from Åkeström 1966, Abb. 64 in Wikander 1988, Fig. 6).

8
Facets could be on upper and lower surfaces, but also the type faceted only from upper side is
spread.

17
3. Sicilian roofing, which was lately named Hybrid, following the nomenclature by Akeström,

who pointed out, that the name “Sicilian” can be misleading, because the type was invented

as diffusion of previous two styles (Wikander 1988, 213), which did not take place (at least

not only) in Sicily, but was in use in Aeolia, Ionia, Anatolia and Central Italy. The tegulae

are adopted from Corinthian system and are covered with simple semi-circular imbrices of

Laconian tradition. Hybrid style was widely used by Romans, especially by their military

units even in areas, where local population traditionally used different system, as for

example in Bulgaria (Mills 2013, 4).

Fig. 3) Hybrid (Sicilian) system (from Åkeström 1966, Abb. 64 in Wikander 1988, Fig. 6).

Great break-through of tile-roofing then spread over the Magna Graecia quickly – in

about two generations. Ákerström (Ákerström 1966, 275 in Wikander 1988, 206) assumes, that

spread of this technique cannot be carried out simply by trade, but rather by migrating

specialized craftsmen, who took their know-how with them. Wikander (Wikander 1988, 206-

207) is very sceptic about the benefits, which the tiled roofs brought to their users, and finds

the major one in reduction of fire danger. He also names all the disadvantages – weight, which

required solid construction of roof and supporting walls, durability, which is unlikely to be

better, than for other roofing systems, with necessary regular reparations, neither did tiles

18
provide better protection against weather and without another sealing layer didn´t help to keep

the temperature inside the building stable, and as the last one he states the high production price,

which was influenced by the cost of necessary transport. The importance of protection against

fire is stressed in more densely built areas and then it is very likely, that the break-through of

tile roofing can be explained in correlation with emergence of larger cities in this period, where

the need for building materials was growing and couldn´t be satisfied from local sources.

At first simple flat tiles were used, but from 6th century, tegula type emerges together

with imbrex, covering gaps between tegulae. Not only Greece, but also Italy was area of this

technique. Tiles differed in size: the biggest as alrge as 115 x 85 cm, the smallest 36 x 31 cm.

Probably the oldest tegulae fund in situ are those covering the tomb of Hypogean Sacellum in

Paestum, small construction dated to the end of 6th century BC. Weight was the main

disadvantage of ceramic tiles and people did realize that, as I illustrated in the story of death of

king Pyrrhus, who died, when his vertebrae was crushed by roof tile thrown at him by brave

Argian mother(Brodribb 1987, 7).

Importance and spread of tile-roofing is obvious on the fact, that in Tarentum in 89 BC,

no one could had been member of municipal Senate unless he possessed a building tiled at least

with 1500 roof tiles. Cassius Dio informs us about valuation of roof tiles in Rome, when in 43

BC all the senators contributed in the fund for war by four asses for every roof tile they

possessed (Dio Cassius, Hist. XLVI, 31, 3 in Brodribb 1987, 8).

With Roman dominance over the Western Empire and North Africa saw general spread

of Hybrid (Sicilian) roofing system, while in the east the Corinthian style was the ruling

technique of public buildings (Mills 2013, 4).

The evolution of efficient roof tile was completed, when tegulae had developed flanges

and cut-outs at the top and the bottom of each piece. This feature helped to lock tegulae together

19
and the whole roof became much more complex and resistant to slip down from the roof. The

nature of the structure required building from bottom of the roof towards the peak of the roof.

For permanent building the cut-outs emerged also on imbrices in overlapping part of the tile,

under such conditions, with gaps filled with and tiles embedded in mortar, the roof became very

stable (Brodribb 1987, 10).

The use of tiles for roofing declined from the 4th century AD onwards, with decline of

the whole Roman Empire, however it survived in Byzantine fashion, where externally decorated

imbrices appeared and eventually decorated tegulea – with grooves along the vertical axis of

the pan tile. The general use of CBM undergoes hiatus, starting in Late Antiquity, re-emerging

again during 11th and 12th centuries AD, when the most popular system remains that of Laconian

tradition (Mills 2013, 4), which can be proved also by recognition of “Tile Standard” from

Roman period, standing at the Athenian Agora (Philips Stevens 1950, 174-188).

Fig. 4) Restoration of “Tile Standard” from Athenian Agora (from Philips Stevens 1950, pic. 10).

4.2. History of bricks

The invention of fired brick came probably from Mesopotamia, where at first mud bricks

were in wide use, and from the process of sun-drying is close to process of firing, it is a question,

if sporadic experiments with firing of bricks could take place even earlier than firing of pottery

vessels, even though the first examples of fired bricks are documented from period

20
approximately 4000 BC, because surprisingly, archaeologists found evidence for firing lime

stones to manufacture plaster to be used as building material (Wright 2005, 109).

The technique for firing bricks was definitively in Greece at the time, when first tile

occurred, but there is no evidence for them. Probably there was no need to employ CBM for

wall construction and stone, wood and mud were satisfactory enough for this purpose. Also

terracotta plaques were applied to early Greek and Etruscan temples, but this is rather special

purpose and examples of baked bricks are exceptional from Aegean region. From Anatolia,

there were also some rare cases, where burnt bricks wee documented, but after testing they

proved to be simple mud bricks fired during destruction phase of the buildings (Wright 2005,

115). The Classical Greek did not favour bricks, however, we can trace them in Olynthos, or in

construction of vaults in Thracian painted tombs, both examples are dated to 4th century BC

(Dodge 1987, 106). Babylonian Empire, which used burnt brick as the main material of

construction, the tradition continued in Near Eastern region throughout Hellenistic and Parthian

era, and thanks to contacts between Seleucid Empire, the tradition was transferred to Hellenistic

cities in southern Italy, where Romans came to contact with it. In Roman architecture, the burnt

brick gained its importance quite late thanks to the development of concrete technique of

building, where the core of the walls filled with concrete and rubble was faced with stone

components at first. Only later, in the middle 1st century AD Romans realized, that brick facing

of the concrete walls is much more economic. First examples resulted from compromise,

because reused bricks from demolished scrap were employed as facing and proved so effective,

that after that were used as material first-hand. The end of the 1st century AD saw great boom

in brick-works in opus testaceum, opus spicatum and often in opus reticulatum as well. In first

centuries after Christ, bricks developed pre-eminent place in construction of walls, arches and

vaults, once again regained dominant position as in the Babylonian Empire. However this

expansion of bricks ended, with shifting of the Roman capital to Constantinople, where the

21
concrete core method was not in use anymore, but more traditional masonries were applied –

cut stone, mortar with rubble and also the solid brick masonry (Wright 2005, 117). Byzantine

Empire continued in use of bricks for solid walls and for balancing bands in stone masonry. In

Late Antiquity the tendency can be observed to reduce the size of individual bricks, and at the

same time to increase the proportion of mortar used to connect the bands of bricks. As the result

of this pattern, the decorative emerged, for the 4th century AD the ratio between thickness of

brick band and the mortar layer is 1:1 and less, but in the 6th century AD it is up to 2:3 (Dodge

1987, 114-116). Even during Roman Empire, in eastern provinces, the construction from solid

burnt bricks are far more usual than in the west. It is a question then, how deeply is the tradition

of solid burnt brick walls in Constantinople caused by direct impact of Babylonian or better still

Hellenistic influences, in other words, if the tradition came from Rome and was adapted to fit

local taste, or followed natural continuity of Near Eastern tradition (Wright 2005, 116).

4.3. Administration of production and construction

4.3.1. Roman Empire

The resources, labour supply and contracts were managed both by state and private

entrepreneurs with varying approach during examined period

The building and supplying projects were let out to private businessmen by censor or

aedil, using the contract system locatio – conductio, which specified particular obligations for

both sides. Private contractors were called redemptores, as we can read still today at their

funerary inscriptions. Locatio - conductio procedure began at the time of Roman Republic, but

successfully continued through the Roman Empire. Interesting habit is to use more private

contractors on the same building project, which results can be occasionally seen in the form of

small differences in design and execution of the same elements. Instructions for masons were

drawn on papyrus, or cut in stone, so the workers and architect could communicate concept and

22
details. The differences in execution however show, that there was flexibility in following

instructions (Lancaster 2009, 257).

4.3.2. Late Antiquity

The system established by Romans, was still in use in following period, in Byzantine

east, to the Middle Ages. At the start of the process was collecting or quarrying of the clay. This

phase took place in alluvium of rivers or in larger quarries. Then the material was transported

to brick maker’s workshops, if the resources of clay were situated far away from it.

Transportation on short distances was done for example in wicker baskets. Second step was to

put this clay into shallow hole in the ground, mixing with water and puddling of the mixture

with tools, like hoe, or with workers´ feet. To prevent cracking of bricks during firing, additives

were added into the mixture, mostly sand. Organic additives were used mainly in production of

mudbricks, because they would burn during firing, which could also cause cracks in final

product. Then the material in the hole was left to absorb the water properly and to sit. The

duration if this sitting process differed, starting on dozen hours up to few days or even a week.

Material ready to be transformed into building ceramics, was then stored in the dark to prevent

drying and loss of moisture (Krautheimer 1986, 128-130).

The workers in brick-workshops (or ergasterion, with meaning „workshop“ in general)

were otrakaioi (clay workers) and keramopoioi (brick workers), as identified in literary sources.

The division of labour took place in workshops, and each person had his position in the process

chain. One bringing the clay, another putting the clay into the moulds, another one bringing the

moulds to the flat area, which could be covered with layer of sand, where he removed the mould,

and let the brick to dry.

Workshops needed to be situated far away from settlement, because they were large

producers of smoke, while firing bricks and tiles, and they would cause constant danger of fire,

if positioned in dense urban areas. Sometimes they were located close to the seashore, which
23
could be connected to economic transportation of products. With consideration to traditional

workshops still operating today and ethnographic data, it is assumed, that the workshop manned

by three men could produce up to 4,000 bricks per day. Bricks and roof tiles were mostly made

by the same workshop, because the character, process and material is the same or very similar

(Krautheimer 1986, 131-135).

4.4. Production – Kilns

Even though, the bricks played the important role between construction materials since

the times of Mesopotamia, there are very rare examples of purely CBM producing kilns. They

are described from the Roman period, but only few were explored archaeologically, contrary to

the pottery kilns, which are known from excavation from all ceramic periods and places (Wright

2005, 110). This fact is important especially, because we suppose the bricks were produced in

the same, or at least similar way as pottery. As the process of firing is quite expensive, the

production must have been sufficient to cover the expenses. Also the bricks and tiles are needed

in masses and the process of loading, firing and cooling of the final product is long and can be

counted at least in days, but more probably in weeks, the number of pieces fired at once must

be in hundreds or thousands. Compared to the size of the CBM is then hard to imagine, that the

kiln appropriate for pottery would be suitable for the huge amount of bricks.

24
Fig. 5) Roman brick kiln (from Adam 1999, pic. 140).

Two basic types of baking could take place in general:

1) Unbaked bricks could be formed into the pile with many holes between them to allow

the hot air to flow and bake each of them properly. This variant was more economical

in the sense of large quantities of bricks produced in this way in relatively smaller

amount of labour spent. The big disadvantage was uncertain result of this way of firing,

because the bricks at the outer sides of the pile could end up unbaked and many of them

had to be sorted out after firing.

2) Kilns brought more certain result, but required the labour to build them and these kilns

couldn’t contain as much material as free standing piles. Many kilns were discovered

by archaeologists in Greece and other region of Byzantine Empire. Kilns had the shape

of an oval and they were also rectangular and consisted of two levels. The upper level

served as space for building ceramics meant to be fired and under them was space for

fuel. Both chambers were connected by holes or other regular openings allowing the hot

air to rise up.

25
As fuel the charcoal or wood was used. The temperature between 800 and 950 Celsius

degrees was needed. The firing continued for at least 12 hours and another few days up to a

week was necessary for kiln to cool down (Krautheimer 1986, 128).

The excavated kilns from Byzantine period can be subscribed to few different types.

The basic distinction is the oval or rectangular shape. The kiln itself is constructed by fired

bricks and is divided into two separate levels, the lower one for fuel and above it the space for

products of firing. Ideally the flames were not supposed to enter the higher space and only hot

air penetrated he upper compartment. Kilns were often built into the slope for ease with loading

and bringing of fuel, also the surrounding earth conserved the heat better. For more precise

typology the differences in division of two levels is used: The oval shape kiln can be supported

by single central column, or more smaller separate pillars, which hold the floor of the upper

level and thanks to gaps allow the hot air to flow into the firing chamber. The upper floor can

be supported also by semi-pillars attached to the wall of kiln, usually in two rows – one on each

side. With the rectangular kilns, the situation is quite the same – one central column or pillar

can be altered with more smaller pillars spread in the space of lower floor or semi-pillars

stepping out of the body of the walls, or the combination of both – the semi-pillars are

supplemented by pillars standing in the row or more rows in the open space of the kiln. This

last type obviously allowed bigger capacity of the kiln (Krautheimer 1986, 131).

The usual pottery kiln in the shape if beehive would contain only small amount of bricks

if we consider their parameters with gaps allowing the heat to raise. Also the weight of CBM

would be a problem for small kilns, even if they extend in size, because the floor supported

from below would easily collapse into the furnace. This type of kiln would be fitting for smaller

production of maybe a hundred of pieces of CBM (but this practice would not be very

economic), or for workshops which contain more kilns. After George Wright, the other types

of firing were applied for large productions (Wright 2005, 161). In Roman practice, Jean-Pierre

26
Adam expects the use of these pottery-like kilns also for large brick productions, only then, that

they are larger and could consist of more galleries and the furnace space was also partly sunken

into the ground (Adam 1999, 112). The ‘charge chamber’ is loaded with bricks or tiles through

the side entrance, which is then blocked hermetically with other bricks and small holes possibly

with clay layer. In the top is one large hole left in the roof allowing the heat to rise. Depending

on weather conditions, the piles of CBM could be covered with tiles, but always should be the

hole in the top left. This causes that the bricks from upper part are not baked properly and cannot

be used directly, but possibly baked again with sufficient temperature. The large brick kiln

could be also supported by the mounds of bricks from outside to keep the heat inside.

Another possibility can be the sunken kiln which perforations in roof with the CBM

stacked at the top and the whole mound covered with another layer for example made of clay.

In this case the trouble could be again the weight of material – only distinct amount of bricks-

tiles can be put above the furnace without collapsing. This type of kiln would not be so well

preserved in archaeological situations, because was only used once and destroyed after firing.

However, still the fired material would be protected against direct contact with flames (Wright

2005, 112). Jean-Pierre Adam (Adam 1999, 112) suggest this type of firing fitting only for

small productions.

In other approach the flames could directly come to contact with material of firing,

mixed with fuel. Such firing procedure is estimated for some large productions (Adam 1999,

71), when the lime, fuel and bricks are all mixed inside the kiln constructed for one use.

27
Fig. 6) Brick firing in piles in contemporary India (from http://thirdeyerevelation.blogspot.cz/2010/05/brick-

kiln.html).

In contemporary Greece and Turkey (Adam 1999, pic 143) the similar type of firing is

used, when the bricks are stacked into piles, leaving gaps between them for fuel and allowing

heat to flow inside construction. This type would not leave any, or very little, remains for

archaeologists, but would fit for large production of CBM. Bricks were at first put into heaps

in the open field (brickfields), gaps filled with wood and then ignited. Trouble could be not

sufficient firing of pieces, which were placed close to the outer surface of construction. This

error could be fixed by selection after firing and unbaked bricks could be added to next firing

(Wright 2005, 112).

28
5. Typology of Roman and Late Antique CBM

In following chapter, the basic typological types of Roman and Late Antique CBM are

listed and commented. The research is carried out with main contribution of Gerald Brodribb´s

overview (Brodribb 1987) of material from Britain, which is very well structured, and thus

convenient for purpose of this study.

5.1. Bricks

The fundamental function of bricks is to bear a load and sustain the pressure. Also more

effective functions accompany this one, such as protection against weather conditions,

maintenance of temperature, enclosing and division of space, and others, but this original

characteristic makes the difference between brick and tile. This initial function is important

manly in the production process for appropriate choice in the sense of material and method, but

of course, in all environments the need occurred to use bricks and CBM in general in new

setting, which could differ from initial purpose of manufacturer, but when listing the various

types of bricks, we have to focus on this original role of each type to find the differentiating

factor between, sometimes very similar kinds of CBM.

29
Fig. 7) Overview of basic Roman brick typology (from Mills 2013, fig. 2.13).

This distinction between brick types is far easier to track, when examining the Roman

material, because already Romans themselves were very strict with terminology of building

elements. The original names are derived from literary sources, and among them the most

influential one – Vitruvius and his “De Architectura libri X”, written under the reign of emperor

Augustus in the second half of 1st century BC.

5.1.1. Bessalis

Square brick with measurements of complete bessales vary from 23,5 cm to 17 cm, but

the average size is 20 x 20 cm, and thickness from 2,5 cm to 9 cm. The name of this brick refers

to two thirds of unit, where as “the unit” can be understood the length of Roman foot (12 Roman

inches “unicae”) and two thirds are 8 inches – Roman bes or “comprising the number eight”

(Brodribb 1987, 34). The main purpose of bessales was support the floor above the hypocaust,

and so, the bricks were piled in piers or pillars. There is archaeological evidence from Britain

for use of not only square, but also octagonal and round bessales, but they are in minority and

30
usually are combined with common square bricks. Bessales with hole in the middle are known

from Pompeii (Wilson 1979, 33 in Brodribb 1987, 35).

Second important function of bessales apart from forming pilae of bathhouses was their

use in brick-facing masonry. This technique of wall construction was very common especially

in Italian regions. Triangular bricks formed the borders of the wall, which was filled with

building rubble, small stones and mainly mortar. From square bessalis 19,7 cm on side were

easily made two triangular bricks by skilled blow of the hammer. Facing of the walls with

triangular bricks can be traced also in other parts of Roman world, for example in Buthrotum

(Hodges 2004) in modern Albania triangular bricks were used for facing of the wall of

nymphaeum, only that the size of the bricks was not uniform (Dodge 1987, 107). In Greece,

brick-faced wall stood as part of Antonine Baths in Sparta, or in Cretan Gortyna, where

triangular bricks formed the walls of odeion and amphitheatre (Sanders 1982, 57-88). Bricks

could be both stamped or bear other kind of mark or finger-mark.

5.1.2. Pedalis

Square brick with sides 28 up to 34,5 cm, thickness varying between 2,5 to 7 cm, with

average of 4,6 cm. Pes, or Roman foot, was basic unit of measurements and its logical, that

CBM with dimensions of exactly one foot is spread all over the area of Roman influence.

Purpose of pedales was capping of the base and the top of pilae of bessales, in other words to

provide piers holding the hypocaust with more stable support and make the transition between

pillars and floor gradual. Pedales were used in wall facing as well as bessales, for example in

Gortyna, which was mentioned before, the triangular bricks seems to be rather quarter of

pedales, then half of bessales, because complete pieces of pedales found elsewhere in the

island9 bear deep octagonal cross, which was originated most probably to make the division of

9
Many almost complete or easy-to-be-reconstructed pieces were found in Priniatikos Pyrgos site see the
chapter „Case study of Priniatikos Pyrgos“ in this thesis.

31
the brick easier to accomplish, and one quarter of such brick has the same measurements as the

triangular bricks used in some Cretan faced-wall constructions, for example in Gortyna

(Montali 2006, 110-116).

5.1.3. Tetradoron

Tetradoron10 has the same, or almost the same characteristics as pedalis. The square

brick with side 29,5 cm long and varying thickness is mentioned by Vitruvius (II, 3, 3) as one

of three brick types used by Greeks, but for some reason is not compared by him to Roman

system. The name of the brick is derived from the Greek “having the width of four palms”

(Pliny XXXV, 49, 171 in: Brodribb 1987, 36), while the Roman unit palmus measured 7,39

cm, then four palms have approximately 29,5 cm and thus in fact equals the Roman foot.

5.1.4. Pentadoron

Pentadoron11 is another brick of coming from Greek construction tradition, as we are

informed by Vitruvius (II, 3, 3). It has the square shape, with sides around 37 cm long12 (Adam

109-110). As the name tells us, the length was equal to five Greek palms unit. Greeks built their

public buildings with pentadora, while private houses consisted of tetradora. Romans did not

use this type of brick, nor any other type close in dimensions to pentadoron.

5.1.5. Lydion

Size of lydion varies across the Roman Empire: In Britain the average dimensions were

40 x 28 x 4 cm (with extremes 48 x 31, or 33,5 x 23 cm); in Italy, where Vitruvius himself

could see the bricks he is describing as “Lydian bricks”, particularly in Arretium the size is

43 x 28 x 14 cm with notably extreme thickness, which is rare (Brodribb 1987, 40); in Greece

10
In Greek transcription τετρἁδωρον.
11
In Greek transcription πεντἁδωρον.
12
Even though some authors present different size of pentadoron, for example George Wright puts the lenght
of the square to 45 to 46 cm (Wright 2005, pic. 158). Under these circumstances, the Greek foot would be 9 cm
long.

32
and Asia Minor, the square bricks are more common, but oblong bricks had been found in

Olympia, or Eleusis, where the dimensions are 40 – 40 x 30 – 34 cm and thus their classification

as Lydia is questionable. In general, the shorter side of the lydion should be one unit and longer

side one and half unit to serve well for purposes of wall bonding, after Vitruvius the size is one

Roman foot to one and half Roman foot (Vitruvius II, 3, 3). As great Roman architect states,

this type was used by Romans richly in their buildings, and can be labelled as the most

numerous brick type. The name of the brick refers to Lydia, the region in Anatolia, but in

Roman mythology also possible home of Etruscans, which means, that tradition could had been

local, adopted from Etruscans, but also from their use in Greek world, as Vitruvius mentioned

(Vitruvius II, 3, 3). Reason for success of Lydia is the shape, which helps to alternate bonds of

bricks, overlapping part of the body of the bricks in the under-laying course. With this technique

huge public buildings and defences could be built. Also, when the mason found, that he reached

the level, when the working was comfortable for him standing on the ground, he could make

the new level for his feet with course of “Lydian bricks”. Another important function of Lydia

was to bind together the facing of the wall and its core. This was accomplished by using courses

of bricks running through the whole wall, which was otherwise filled with mortar mixed with

stones or rubble and thus slightly unstable, especially when the filling was still soft and thin

outer skin couldn´t hold the bulk of material (Brodribb 1987, 39). These bonds formed by few

brick courses bound the wall in regular intervals on its vertical and later became also rather

decorative element of the buildings.

5.1.6. Sesquipedalis

Square brick approximately 44,4 cm on side (Adam 1999, 293), but occasionally

increasing up to 46 cm or decreasing to 40 cm. Thickness ranging from 4 to 7 cm, with average

of 5,2 cm. Sesquipedalis has the meaning “one and half feet”, which is its true relation to Roman

measurement unit. Vitruvius mentioned sesquipedalis as perfect brick for pavement underlying

33
the pilae (Vitruvius V, 10, 2), and thus forming the floor of hypocaust. Sesquipedales were also

used for wall facing, when from one square brick, eight triangular ones were obtained (Adam

1999, 293). Mass production of such large bricks, which could be easily divided and adapted

for immediate demand made sesquiedales13 one of three most used bricks in large cities as

Rome and Constantinople.

5.1.7. Bipedalis

The largest brick of Roman serial production has square shape with 59 cm on sides, with

average thickness of 6 cm, but rare extremes occurs sometimes, for example pieces from Fréjus

with dimensions 75 x 75 x 7,5 cm weighing about 83 kilogrammes, or another example

measuring 105 cm (Lanciani 1897, 39 in: Brodribb 1987, 42). It can be estimated, that brick

fragments thicker than 7 cm are originally coming from bipedalis14. The name of the brick

refers to dimension of two feet, and contemporary inscriptions often give evidence of wide

range of functions bipedales had. In Corpus Inscriptiorum Latinorum, it bears names

“bipedale”, “bip”, “tegl bpedal”, “bipeda” and The Diocletian Edict uses term “lateres pedum

binum” (Brodribb 41). These huge bricks were much more versatile in purpose, than any other

– they served as the packing of suspensura of bath, while bricks bridged the gaps between

hypocaust piers, which were distributed at exact distance to allow biedales to do so. Vitruvius

also lists them as possible inner cover for timber construction of bath, protecting the wood from

rotting (Vitruvius V, 10, 3) affected by steam coming from bath devices. They fitted for flat

wall bonding, thanks to their proportions, which enabled masons to lead composite course

through the whole wall at once15. In the archthe complete bricks were altered with parts of

13
Togehter with bessales and bipedales (Jeffreys et all 2008, 337).
14
However, there are examples of smaller types of bricks thicker than their standard, but they are rather
exceptional. The thickest brick documented is probably the one from Taormina in Sicily 12 cm thick (Blake
1947, pp. 283 in: Brodribb 1987, pp. 43).
15
The use of complete bipedales for wall bonfing is known from the time of Domitian (Blake 1959, pp. 162 in:
Brodribb 1987, pp. 41).

34
bipedales, and so the whole structure was more stable. When the floor of suspensura had to be

raised, it was done with bricks standing on their smaller space, while the brick was supported

by the wall from side. Exceptionally, the bipedales with hole are found, which were identified

as elements admitting the steam into the bath chamber by some scholars (Brodribb 1987, 42),

but this is unlikely to be correct placement, because the holes were often situated right above

the pilae and the steam would be blocked by them. Other explanation for holes in bricks is, that

they served for mortar to adhere better to the brick, especially when there was need to sustain

side pressure. The holes could also serve to help firing or to help to release gas from the kiln.

5.1.8. Byzantine brick

In general, Byzantine bricks seem to vary even more considerably than the Roman ones,

but there is no strict definition of dimension of particular shapes. Krautheimer assumes

(Krautheimer 1986, 131-132), that this variability occurs due to shrinkage of tiles during drying

and firing process, and thus artisans had to form clay into larger moulds, calculating with

irregular shrinkage. Standard size was around 32 to 36 cm16, with thickness of 3,5 to 5 cm.

However for Constantinople itself, where the building activities were much more intensive than

in other areas, plus, where the materials were scarce and bricks were valuable source for

constructions, we have more precise observations based on architectural chronology of building

projects. Under the rule of Constantine17, the standard measurements was 31 x 31 cm, 5,5 cm

thick; sixth and seventh century saw rise in size – 37 x 37 cm 4,5 cm thick; and at the end of

the 6th century, the size decreased again to 33,5 x 33,5 with thickness 4 cm. After this period of

growth of the Empire, brick production declined and spolia from older monuments served for

new constructions18 (Bardill 2008, 339).

16
Byzantine foot - Greek pous, or πouς measured approximately 31,2 cm, in practice fluctuated between 31
and 32 cm.
17
Reigned from 306 to 337 AD.
18
Exploitation of older buildings is widespread phenomenon in Byzantine architecture in general, so the use of
spolia from the 6th century was only continuation of well-established trend.

35
5.1.9. Cuneatus / Solid voussoir

Cuneatus, or in modern terms “solid voussoir” has wide variety in dimensions –

measurements of the sides has usually length 40, 30 and 15 cm and these are combined together

in few versions: 40 x 40, 40 x 30, 40 x 15, 30 x 30, or 30 x 15 cm. The thickness differs at the

top and at bottom of brick, which is fundamental for brick´s function. Bricks are 6,2 to 8,5 cm

thick at the top to 3,9 cm at the bottom. The tapering does not have the same angle, but slightly

differs from piece to piece19. The name cuneatus is connected to wedge-shape of the CBM.

Vitruvius explains the use of voussoirs for arch constructions (Vitruvius VI, 8, 3-4). The arch

could be made with sufficient result only by usual flat bricks, by adjusting the amount of mortar

between them to form a curve, but it is more convenient to keep the constant thickness of mortar

at the top and bottom, and use tapered bricks. Not only curved arches were constructed with

cuneate, but also flat arches above windows and doors, as we can see in harbour of Ostia

(Brodribb 1987, 45, pic. 18).

5.1.10. Armchair vousoir

For relieving arches in walls, above windows and doors, solid voussoirs were suitable,

but for ceilings of larger spaces, solid bricks were excessively heavy and increasing the danger

of collapse of the construction. For this purpose, lighter construction was introduced – armchair

voussoirs, which combined both, heavy solid bricks altered in rows with lighter combination of

lintels covering the gap between two solid bricks with only one or two flat tiles. Bricks covering

gaps between voussoirs were usually simple flat tiles differing only in length, because for one

arch two types with different dimensions were needed to create two layers – first locked at the

lower part of armchair voussoir and second at the top part of the brick. Shapes could be

different, but lower part is wider, than the upper part of voussoir, body is robust enough to

19
Thinner bricks (for example from Britain) tapers from 4,5 to 3,5 cm (Broadribb, 1987 pp. 45).

36
provide the base to sustain the weight of the other voussoirs piled on it and profile of the brick

has at least two regular steps for placement of gap-covering tiles. Advantage of this construction

was its lightness, because two gap-covering tiles formed hollow space within ceiling

arrangement, but steady support for the roof or next floors. There is no technical term for this

type of brick in Roman sources and thus was referred to simply as cunneatus.

As another which another variant of solid voussoir - tubuli cuneati were hollow, forming

ribs of the arch, with each rib could also provide air circulation for both insulation and heating.

Top and bottom of the tubulus were open to allow the air to flow from voussoir to its neighbour.

Their shape is trapezoidal to form the curve of the arch, and their section is rectangular and

hollow. Advantage arch constructed with these hollow voussoirs is its low weight, but on the

other hand, the building elements are very fragile, and thus only fragments can be collected at

Fig. 8) Examples of “armchair” voussoirs for vault construction found in Britain (from Brodribb
1987, fig. 19).

excavation of collapsed ceilings. Cut-outs on the sides and on outward facing occur sometimes,

and were concerned with setting of the voussoir into the position.

37
Unique function had “springers”, which acted as foot-pieces for pile of voussoirs at the

base of the arch. They are very similar to common voussoirs, but have only one tapering at the

top and are slightly more massive.

Fig. 9) Example of “tubulus cuneatus” (from Bridribb 1987, fig. 35).

5.1.11. Triangular brick

There is no evidence for straightforwardly produced triangular bricks, but huge quantity

of urban buildings were made with brick-facing using this type of CBM. Several other types

served as a good source for triangle. From one bessalis, two bricks could be made with

dimension 30,5 x 21,5 x 21,5 cm, from sesquipedalis up to eight triangles measuring

32 x 22,5 x 22,5 cm could be obtained, bipedalis offered possibility to divide it into 16 bricks

28,5 x 21,5 x 21,5 cm. Since the area of square bipealis is four-times bigger than that of pedalis,

four bricks with same measurements as triangular bricks from bipedalis could be made.

5.1.12. Column brick

Bricks of this special shape were used for columns since the annexation of Sicily in 241

BC. After this event Romans made larger use of bricks in general, but also columns of ceramic

origin were introduced mainly for peristyles and later distribution of idea into Roman world.

Brick columns were cheaper to acquire than stone pieces, plus in particular areas especially in

Roman West, the precious stones with sufficient volume were scarce and had to be transported

on large distance (Wright 2005, 38). In densely built areas, such as Rome itself, the material

resources did not cover public demand and had to be substituted, particularly for ordinary

38
building projects. Except from columns, some pilae in bathhouses were also pilled with circular

or part-circular bricks20 and they could also serve as seats (Brodribb 1987, 55). The entirely

circular bricks are rare with diameter 38 to 40 cm, but common practise was to compose each

drum of the column with more segments, with their joints overlapped by next course.

The semi-circular and quadrant bricks start at 23 cm up to 60 cm in diameter21 and the

thickness from 4 to 8 cm. It seems, that the production method was to prepare the mould for

complete circle and after baking, cut the brick in half as some cuts are very apparent. Section

of the circle is sometimes not exact semi-circle, but slightly smaller, probably thanks to the

defective division. The most elaborated come from Pompeii22, particularly Basilica, where the

columns are not exactly circular, but form the shape of “blossom” of petals, which carries the

stucco fluting on the surface (Adam 1999, pic. 370). The core of the column is created by small

circular cylinder, around which the petals are arranged. However, it can be questioned, if

constructions in Pompeii are representative examples, because many columns are formed no by

bricks, but rather by cut reused tiles, which originated after the earthquake in 62 AD, when the

city was full of collapsed roofs, providing cheap source for CBM which could be easily

reshaped (Adam 1999, 116), and also ancient authors highly recommend constructions of old

roof tiles, which are stronger than that of bricks (Vitruvius II, VIII, IX). The bricks for columns

were not necessarily segment of the circle, because the imperfections in shape could be

improved with huge amounts of plaster and mortar. George Wright shows some examples of

column sections, where the core is made by rubble and mainly the borders of the column are

carefully composed by small brick pieces. On the surface the decorative stucco overlapped the

20
For Cretan examples see for example Grigoropoulos 2008, pp. 312.
21
As „diameter “, understand, the diameter of the full circle composed by more partly-circular sections.
22
More individual columns can be found in Pompeii. Another example of column with core comes from The
House of the Faun (Adam 1999, pp. 143).

39
bricks, and if carefully done, nobody could actually say on the first sight, if the column is made

of stone or some other material.

Fig. 10) Sections of brick columns with strong layer of plaster (from Wright 2005, pic. 208).

Irregular segmented bricks were produced for example for so called “cylindrical

terminal pylons” (Lowther 1937, 34 in Brodribb, 1987, 56), where the circle motif was mainly

the pattern, as the wider parts, or the main base, contained rectangular part adjoining the wall,

no matter, which part was dominant – semi-circular or rectangular.

5.2. Tiles

The difference between tile and brick is in the purpose. The purpose of “tile” is not to

carry heavy burden, or sustain the pressure, but rather to protect the surface, to cover the

construction, or serve as covering layer in general. In opposite to bricks, the main characteristic

of tile is to come into direct contact with changing environments in the sense of weather, light

(but also sight, while examined by observer), short term physical contact (flooring) and more.

Tiles could be grouped according to their purpose on: roof tiles, flooring tiles, wall tiles and

other tiles with special function.

5.2.1. Roof tiles

Terracotta tiles were inherited from Greece, where they were developed in ancient times,

with establishing of three dominant styles in 6th and 7th century BC (Wright 2005, 125-131).

Roof tiles were fired under high temperatures and the fabric needed to be sorted more carefully

than fabric for bricks, because their function was protection against weather conditions,

40
especially the rain, and thus the surface of the roof must have formed waterproof unit without

deleterious voids in the body of the tile.

5.2.1.1. Tegula

Also known as “flat” or “pan” tile, even though the majority is not simply flat, but either

slightly curved in the case of Laconian tile, or bordered with flanges on sides and top ending as

in the case of Corinthian tiles. The shape varied only little, it was both rectangular or bit

trapezoidal, always placed lengthways, overlapping one another, with junction covered with

imbrices. The shape was uniform, the size differed in each region. Recorded Roman tegulae

varies in their dimension: in Ostia from 40,5 to 48 cm x 53 to 72 cm; in Rome from 39 to

49 cm x 46 to 66 cm; in Pompeii from 47,5 to 52,5 cm x 59 to 69 cm (Adam 1999, 436);

Knossos 53 x 73 cm, but also 110 x 38 cm (Sackett 1992, 407–410). In the Western part of

Roman Empire, the Corinthian roofing system was declining in the Late Antiquity in favour of

roofs constructed exclusively with imbrices, applied as pan tiles and cover tiles in Laconian

roofing manner (Adam 1999, 437). In the East on the other hand, the tradition of Corinthian

roofing system remained strong during Late Antiquity, when the Corinthian pan tiles were still

distributed from central production centres23. The meaning of the word is “small covering”

(Brodribb 1987, 5). Apart from roofs, tegulae could have been used for burial cover, these tile-

tombs are well known and spread phenomenon within Roman Empire24, however this function

was not primary. Mainly Corinthian pan tiles served for tile-tombs, but occasional Laconian

took place as well. Another function of re-used pan tiles was to build up floor of drainage

channels (Sackett 1992, 407) and small water streams. Small shards of tiles, which no more

could be employed in their original shape, could still provide useful construction rubble to fill

23
Compare the example Kalavasos-Kopetra in Cyprus (Rautman 2003, pp. 9 - 55) and Nichoria in Peloponnesus
peninsula (Coulson 1983, pp. 384-385).
24
One of the oldest examined tile-burial is in Paestum, so called „Hypogean Sacellum“ (Brodribb 1987, pp. 5),
other examples from Knossos (Sackett 1992, pp. 407 - 409) and more.

41
the core of the walls and small gaps on the surface of the walls, and also could be applied while

repairs of the walls were necessary. Not only broken tiles were reused in the wall constructions

(Adam 1999, 116), but also large pottery shards. This habit can be found in Italy as well as in

Greece, where ceramic shards are complementary material for stone walls formed with irregular

boulders, and thus re-using shards in very convenient way. In some cases, pan tile contains nail

hole, usually the pan tile in the lowest course of the roof, overhanging the eaves. The hole is in

the centre of the tile, around 1 cm large, 5 cm from the top edge, and pierced before firing. The

flanges of Corinthian pan tiles were folded up of the clay still inside the mould, while the surplus

clay was sliced away. This technique can be proved thanks to some broken pieces, where on

the cross section the clay is obviously “curved up to create the flange” (Brodribb 1987, 13), the

average depth of the flange is usually double depth of the flat part of the tile-body, but many

irregularities to this rule occur everywhere. The flanges are stepped at the top end so the

overlapping pan tile could easily be positioned and protected against slipping. The flanges are

known in many variations in shape and thickness, but the top on the inside they are usually

rounded as the result of their formation process. Remarkable variations in profile can be element

of chronological significance, but only within certain areas. For example in French region Bas-

Dauphiné, Chauffin suggests, that in earlier periods of Roman occupation (1st to 3rd century

AD), the flanges were considerably square cut, regular and uniform in shape, than in later

periods, when the diversity increases and wide scale in flange shape occurs (Chauffin 1956, 81-

88 in Brodribb 1987, 14). Philip Mills stresses the distinction between Corinthian and Sicilian

(Hybrid) style tegula – Corinthian types have “inset flange on the upper surface and vestigial

flange at the opposite edge of the lower surface” (Mills 2013, 31), while Sicilian types have

considerable cutaways at the top corners. In secondary use, flanges were removed by chipping

or sawing, and the tiles were placed face downwards to hide the flange-cuts. In Pompeii and

Paestum there are records of such method used for baths (Brodribb 1987, 14). Finger-made

42
grooves appear at the angle between flange and body of tegula, serving probably to assist the

flow of rainwater, or it was simply made at the end of manual formation of flange in the mould.

To lock one tile into another, the cutaways were employed at the bottom and top of the tile. As

the higher overlapped the lower, cutaway at the top were made from inside, or the whole part

of the flange25 was removed, and at the bottom part of the tile was removed from outside. The

top end of the tile is generally wider then bottom, to slot two pieces easily. Cutaways are not

necessary for good function of the roof and it is clear that sometimes they were made more

roughly or even skipped completely. Tiles were fixed on the roof with mortar, bedded into the

clay layer, or nailed on the wooden frame. More of these techniques could be used on one

construction, especially tegulae in the last row needed to be fixed sufficiently, because they

didn´t have natural support from the tile below.

5.2.1.2. Imbrex

Employing of tegulae requires automatically use of tiles covering junctions between

pan-tile rows, making the roof waterproof. Imbrices are smaller than tegulae from the set the

form together and are considerably lighter. They play only supplementary role on the roof,

while the majority of the space is covered with pan tiles. They are narrow compared to pan tiles,

but have the exactly same length and protrude over the tegula they are placed on, to overlap the

imbrex in the next course. Imbrices are curved, sometimes in the shape of semi- and smaller

segment of the circle, in other cases more sharply bended into the shape of converse letter V.

Measurements depend on the size of tegulae, but usually are around 36 to 51 cm long, but can

be a lot longer, according to the length of pan tiles. Width differs on the top and the bottom

edge, while the bottom is slightly wider to lock adequately with underlying cover tile,

measuring from 22 to 13 cm at the wider end, and average 13,5 cm on the narrower end.

Thickness varies around 2 cm, and curiously, the loner tile is not necessarily thicker. Medieval

25
In average 5 cm of the flange was removed in british Roma tile production (Brodribb 1987, pp. 16).

43
and modern practice shows us, that the roof can be composed exclusively of imbrices overlaid

in layers facing up and down, with overlapping edges. Cover tiles could be divided into two

groups: Corinthian and Laconian. Laconian cover tile is simple imbrex in the shape of circle

segment, the surface inside and outside is regularly curved. Corinthian on the other hand is

much more decorative and delicate on the first sight, while the upper surface is faceted (Mills

2013, 32) with three bends dividing upper space into four planes. The surface inside Corinthian

imbrex is basically curved regularly as with Laconian cover tile. In the past, when imbrices

were found and examined by archaeologists, they are refer to under various names: cover tile,

curved tile, joint tile, rounded tile, or gutter tile. The word imbrex refers to the sound of rain

showering on the roof26, the noise of buzzing bees, or clanging together of pots and bricks

(Brodribb 1987, 24). Comparing the value of the cover tile in relation to pan tile, Cato states,

that the price of imbrex should be counted as one quarter of tegula price, but it is more likely

to be true for Corinthian, or better Hybrid (Sicilian) roofing system, because here the imbrices

would be reasonably cheaper than tegulae, which require more expenses, time and effort during

production, than in Laconian system, where pan and cover tiles are of almost the same

construction, differing only in the angle of curvature and volume of clay needed, but they are

far easier to make. Decorative patterns does not appear on imbrices as much as on tegulae, but

still there are some: finger-made lines running the whole length of the tile, finger made lines

over the gable, stamps and signatures of producer, simplified sketches of humans or animals

(Johnson 1979, 375-394) and animal impressions. Interesting remark is made by Brodribb, who

suggests, that deep finger marks on tiles were caused, when the tiles were carried before firing.

Secondary use of imbrices includes bedding for pipes, or direct use for draining channels (Mills

2013, 96), enclosure of hearths and furnaces (Brodribb 1987, 25) and of course using of rubble

for levelling floors and as building material for walls.

26
Imber – in latin „a shower of rain“.

44
5.2.1.3. Ridge tile

The necessary component of saddle roof was the cover for uppermost pan-tiles.

Wikander suppose, that due to lack of this CBM type in Roman deposits, it was probably used

only for finer buildings (Wikander 1988, 211), and commoners used simple imbrices to cover

the gap, but this scarcity, could be caused rather by low numbers of such tiles needed for roof

projects, and difficulties with their recognition in fragmentized archaeological material. Even

though the evidence is scarce, it can be described as another roofing element, which is difficult

to identify, because of variety of shapes. The division between Laconian and Corinthian is also

unclear with ridge tiles, as they were adapted for every single project. Common characteristic

for majority of ridge tiles seems to be openings on their sides to accommodate adjoining cover

tiles, which were underlying this ridge tile. Ridge tiles had to overlap not only tiles in lower

courses but also subsequent ridge tiles on both ends as well, which was possible thanks to

cutaways alternating on top and bottom on the opposite edges of the tile (Wikander 1988, 211),

or with overlapping of slightly tapered tiles27. For Corinthian roofs, the saddle tile appear,

shaped like to ordinary tegulae connected under exact angle at their tops, and their joints are

covered with curved cover tiles of same connected construction (Wikander 1988, 213). For

Laconian roofs, the same semi-circular ridge tiles with openings on long sides occur, sometimes

tapered.

5.2.1.4. Other roof installations

To protect the outmost part of the roof from rain, the exterior

border of last course – pan tile was raised to form a raking sima, but

its decorative form was developed mainly at time of Classical Greeks,


Fig. 11) Gutter tile (from
Wiegand & Schrader [N. 53] than by more pragmatic Romans, which used rudimentary simas on
Abb. 331 in: Wikander
1988, fig. 7).

27
On the other hand, Gerald Brodribb presents the absence of tapering as the main characteristic of ridge tiles
(Brodribb 1987, pp. 28).

45
ceramic roofs only occasionally. Eaves-tiles with drip-edges and lateral simas with water-

spouts (Wikander 1988, 215-216) emerged also soon, in the 5th century BC, and in the time of

Roman Empire and Late Antiquity were rarely used. Gutters on eave-tiles are recommended by

Vitruvius (Vitruvius III, 5, 15) to lead the water from intercolumnia inside the sima, and are

placed at the lower short side of the pan tile in the last course. These were in use during

Hellenistic and Roman period.

Antefixes were developed from protection of openings (plaques) in lowermost cover

tiles and were attached to them28.

Special hip-tiles were required for roofs of irregular shape. Also Apses and tholoi

needed special roofs, constructed with trapezoid tegulae. Scale-shaped tiles occur in Classical

Greece, but later declined. Tegulae colliciarum were solution for roof spaces meeting each other

under the right angle, with stream-channel in their centre, but troubles with combined roof could

be easily avoided by constructing separate roofs at different levels (Wikander 1988, 216).

Part of the roof accessory is chimney pot, especially in

continental areas, where sustaining of temperature inside the building

was

Fig. 12) Chiney pot (from


Brodribb 1987, fig. 15).

crucial. Usually just one

chimney is needed for one


Fig. 133) Roof tiles with ventilation device (after Durm B d R, fig. 364 in:
Wright 2005, fig. 186).

28
Since antefixes are considered to be rather architectural decorative element than CBM and are categorized in
account to their artistic value, I will not focus on them here. For more information about antefixes see for
example (Higgins, R. A. [1954] 1969. Catalogue of the Terracottas in the Department of Greek and Roman
Antiquities, British Museum 1, repr. Oxford.).

46
roof, so it is not surprising, that chimney pots found are incomplete and vary in shape and

design. Important feature is, that the chimney has to be closed from above to avoid raining into

the tube and the holes for smoke to come out have to be on sides. The example here is from

Beauport Park in Britain, with diameter 21 cm and height 47 cm, the windows for smoke are

triangular. The bottom is at the same level all around the pot, so it probably stood on modified

setting on the ridge tiles, the smoke then could be freely released into the room, where gathered

under the saddle roof and went out through the chimney. Since the evidence is scarce, there is

another theory, which explains these objects rather as light or aroma stick covers, but they are

likely to be used for smoke and hot air draining through the roof, because some similar, but

smaller pieces were found attached to the cover tiles (Brodribb 1987, 31 - 33). Another special

devices were included in tile for sky light to shine inside the building. They were usually

incorporated in tegulae, which provided enough flat space for such adjustments (Fig. 13).

5.2.2. Paving tile/brick

CBM paving started quite late in Italy and Mediterranean area, compared to use of

backed bricks and tiles. First experiments with brick paving occurred in Italy in 2nd century AD,

and soon became widespread, especially its form of opus spicatum or herringbone pattern.

Vitruvius refers to CBM used for this technique (Vitruvius VII, 1, 4), as Tiburtina (testacea

spicata), because the town of Tivoli was great producer of small bricks applied for floors. Their

dimension is 15,5 to 7 cm long, 9 to 3 cm wide and 6 to 2 cm thick, with the shape of “playing

card, and other closely resembling a modern fish finger” (Brodribb 1987, 52). According to

Vitruvius, bricks of this type had to be rubbed to fit perfectly to the floor motif, ridges or gaps

were undesirable. The floor itself was also polished and after that, sand had to be applied to

cover small gaps. In Pompeii, the bricks are more variable, taking shape of other geometric

forms. Triangles, squares, rhomboids, hexagons served for brick paving in geometric

47
arrangement, probably imitating more luxurious marble slabs. Occasionally, the ordinary bricks

could be used for flooring, but they were liable to be worn out after short period.

5.2.3. Wall tile/brick for special purpose

To cover the walls, Romans employed special CBM component, because, not only

decorative function was demanded but especially practical – to heat up the walls and thus the

make the temperature in the room comfortable, which was crucial in colder regions, like

northern parts of Empire.

5.2.3.1. Parietalis

To cover the wall with additional layer, Roman used so called parietalis, which can be

found especially in colder regions, where the heating in the room was reasonable. Tiles were

hung inside the rooms and in archaeological material are not very distinctive in any particular

form. Probably the most common feature identifying parietals is lozenge pattern on the surface

made by scoring on the rear side of the tile, and the nail notches for fastening the tile at the

place. Their purpose, except of decreasing the loss of temperature in the room, was to prepare

surface of the wall for plaster (Wright 2005, 124-125).

5.2.3.2. Tegula mammata

For damp places, Vitruvius recommends the use of insulation, to let the walls “breathe”

(Vitruvius VII, 4, 13). It is not clear, if Romans invented square tiles with four bosses in corners,

for insulation, or as part of their bath system, because the same technique is used in both

examples, but obviously, without heating system, the nippled tiles served basically for

ventilation (Adam 1999, 557). Term mammata means “with breasts” and was originally used

by both Vitruvius and Pliny (Vitruvius VII, 4, 2). Broadribb defines two types: A) having round

shallow boss with diameter 4,4 cm in the base and 1,7 cm deep, probably used mainly for

flooring and brick courses, with protrusions for easier bonding with mortar, and B) with conical

48
boss, diameter in the base 5,7 cm and remarkably deeper – around 6 cm, fitting more for wall

insulations. For A type, mammae are rather shallow and roughly made, sometime situated in

the corners or close to the corners of the tile, but very often are distributed irregularly or

chaotically and their number on one tile varies from 1 to 5 mammae. The theory was proposed

by Wheeler (1936, 141), that bosses served to keep the CBM pieces apart in the kiln to let the

air circulate freely, but this is very unlikely to be correct assumption, because bricks and tiles

fired under these circumstances would carry imprints of mammae (Brodribb 1987, 62). Type B

have its bosses distributed more regularly on the body of the tile, and their number is more

uniform, sometime with additional supporting mamma in the centre. Examples from Domus

Transitoria of Neronian period measure 60 x 60 cm and are 4 cm thick (obviously derived from

the length of Roman feet – approximately 30 cm). The method for attachment to the wall was

nailing with T-shaped metal nails, put into the middle of two neighbouring tiles, and thus fasten

them together as well. For this purpose tiles contained small grooves accommodating the nail

Fig. 14) Tegulae mammatae and section of the wall showing their employment (from Wright 2005, pic. 191).

at the place. Long-nippled tiles were suitable for baths and other heating installations indoors,

thanks to the good draught of the air, because if the space for it was thinner, the air would

turbulent without rising, or even revert (Adam 1999, 557).

5.2.3.3. Tubulus

Can be characterized as rectangular pipes (box-tiles), the recorded sections are from 8,5

x 13 cm large to 28 x 31,5 cm (Adam 1999, 557; Brodribb 1987, 74) and length 15,5 to 45 cm.

49
Hollow sides are on the shortest ends of the brick, and in the wall were positioned vertically,

rarely horizontally (Brodribb 1987, 73). The tubuli were piled to form a flue for hot air. They

couldn´t stand alone, but always had to be supported by regular wall, because they are too

fragile to carry the roof or other floor. They were glued to the wall and to each other with

mortar. Some tubuli contained side-openings for regular distribution of air inside the

construction and occasionally openings to room (Brodribb 1987, 70) and the inner space of the

tile could be partitioned in the centre. The compactness was secured by T-nails put into the

brick joints and the whole layer was covered with stucco. Openings for removing hot air and

smoke were left at the top part of the wall, collecting smoke in the chimney, or leading it directly

outside the room. First tubuli emerged in the first half of the first century AD and are used for

example in Stabian Baths at Pompeii (Adam 1999, 557-562). After employment of box-tiles in

bath-heating, it spread also to domestic houses. Earlier heating systems were equipped with

tegulae mammatae, but after their invention, tubuli dominated, because they allowed the air to

circulate without whirls. Tubulus usually contains only one joint on its surface, and thus it is

assumed, that during production, the sheet of clay was wrapped around wooden core, which

was withdrew right after, and the firing took place after sun-drying. Doubts were stated, that

this withdrawal of wooden core would damage the bow tile, but it was proved by experiment,

that the clay is firm enough to stand vertically, and the wood could be carefully removed

(Morgan 1979, 95). Scoring and combing sometimes occur on the surface to attach easier to the

mortar. Complete, or partly complete tubuli were used as drainage (Brodribb 1987, 79).

Special position between tubuli has so called tubulus cuneatus, or hollow voussoir,

described earlier.

5.2.4. Tubes and pipes

Circular hollow elements were not only used for draining water, but also as building

material for vaults.

50
5.2.4.1. Vaulting tube

Building elements in the shape of syringe around 25 cm long, 8 cm in diameter, but the

size is not uniform, as well as relation length/width. Their origins lie probably in Northern

Africa, where they can be frequently found (Mills 2013, 33), but from there, vaulting tubes

were spread into Britain, Lebanon and Sicily. They fit into each other, putting narrow end into

the wider end of another tube. Because the connection is not tight, the narrow end enables the

adjoining tube to tilt to the side and thus the curvature of the vault is achieved. The employment

of tubes for vaults comes probably from the use of amphorae to lighten the roof. This archaic

technique was examined in Stabian bath at Pompeii or Circus Maxentius at Rome (Ward-

Perkins 1981, 424). Pipes could also be easily installed as chimneys, and common was their use

to drain rain-water from the roof to the cisterns. Arrangements for dealing with rain-water are

constructed outside on the walls, as well as integrated into the wall body (Brodribb 1987, 88).

Fig. 15) Sections of vaulting tubes (Mills 2013, fig. 2.14).

5.2.4.2. Water pipe

Famous are lead pipes, which are easy to make, relatively firm and waterproof, but their

disadvantage is high price of material, difficult to obtain in some areas. Material for ceramic

pipes is on the other hand available everywhere and for free. The progress in using pipes was

spreading with Roman culture and hygiene, which required baths, where pipes could be used

for heating systems, water transport and drainage (Wright 2005, 131). Tubes have constriction

51
at one end to fit perfectly with next tube. The diameter is quite large - 13 to 57 cm, with length

3 to 94 cm (Brodribb 1987, 84). To avoid seepage in the joint, these were sealed with mortar

with added oil to keep the joints watertight (Adam 1999, 520). The watertight tubes were crucial

element for system using pressure caused by different level of starting and ending point of

water. This pressure inside the system would not occur, if holes were present. The water was

distributed from central reservoirs to fountains and public facilities, as we can see still today in

Pompeii (Adam 1999, 521). Unusual shapes of tubuli occur – hexagonal in section, bulbous,

trumpet-shape with flared end, tapered in full length, oval in section and syringe shape29.

Fig. 16) Water pipes.1. Hellenistic;2.-5. Roman (from Wright 2005,pic. 196).

29
Ocassionally found in Britain and in Basilica Ursiana at Ravenna, where these tubes formed part of great apse
(Brodribb 1987, pp. 87).

52
6. Application – construction techniques

Following chapter will focus on practical application of CBM – its application in

constructions. Building techniques will be defined and described with various details

concerning construction process, esthetical characteristics, architectural features and

occasionally social and economic aspects. Individual sub-chapter are divided according to

examined construction techniques.

6.1. CBM in wall construction

Bricks were often used in masonry in Roman and Early Byzantine period, but the

techniques of construction vary depending on century, region, taste of the sponsor and the type

of building. Brick was in more densely built areas the cheapest source of material right after

stone30. In this chapter we will go through the evidence to track out basic patterns and rules of

CBM application in constructions.

The advantages of bricks and tiles are apparent: the material is read to hand, components

are of regular shape, easy to form courses I walls, and the surface is smooth and flat enough to

take only small layer of mortar in comparison with difficulties with processing undressed stone.

Another advantage over stone is, that CBM elements can withstand frequent changes in

temperature, under which stone would disintegrate (Webster 1979, 287).

In Greek region and Asia Minor, the bricks were used in slightly different way, than in

Rome itself. Since half of the 1st century, the common habit was to compose the wall from brick

facing and fill it with mortar, from famous Roman pozzolano, which allowed constructions to

30
But stone was not suitable for all types of constructions, for example arch and vaults required qualified
treatment of stone blocks, and thus the architect or constructor had to consider which material would fit the
best.

53
be very light and firm at the same time. As addition to the mortar core, rubble was implemented

to bond material together. This setting has probably its origins in reuse of old roof tiles, which

were re-shaped into triangular form (Wright 2005, 193). In the East, the concrete core technique

is largely surpassed with solid brick walls, alternating courses of stones and bricks, or stone

walls, because the area is rich in suitable resources.

6.1.1. Foundations

Bricks fall apart if placed in moist environment, so they are practical for constructions

above the surface, without direct contact with soil, which would threaten the complexity of the

structure. For this reason, foundations of the walls were usually made by stone rubble or

masonry and first layers of bricks were laid on them.

6.1.2. Brick faced concrete walls

In conventional Roman architecture, the pattern at the bottom of the lowest floor is

applied for the rest of the building to the height of the roof covering the last floor. The primary

elements of the wall are its core, facing (brick facing in this particular study), solids in the sense

of bonding courses and relieving arches, voids, such as doors and windows. The secondary

elements are based on the building function and equipment – heating systems, artistic

decorations of various materials, grooves for pipes, or voids for chimney, and others (Taylor

2003, 94). The facing could be also covered with mortar and painted, or with marble slabs

(Adam 1999, 293).

At the time of Vitruvius, the advantages of brick faced constructions were still not fully

recognized, even though the opus latericium was known at least from the time of Augustus

(Taylor 2003, 97). At that time bricks were already made with additives of pozzolana, or simple

sand.

54
Opus latericium31, also known as opus testaceum was the main architectural technique

composed by bricks. It serves as facing for opus caementicium of which consisted the core of

the wall. The crucial material in Rome and central Italy was pozzolana concrete, light and strong

agent for constructions32. Bricks as facing material allowed masons to build up the concrete

walls without wooden frames, and this caused great boom in brick production in Rome in 1 st

century AD. This wall-skin also provided better conditions for curing of mixture inside the

structure, because concrete required particular level of moisture and stale support on sides,

during the pouring and setting process, otherwise there was the threat of disintegration and

collapse of the structure. With rigid container in the form of wall boundaries, masons could

pour the concrete inside in larger loads, and thus the whole construction took far less time to

accomplish, because the batches could be applied with lesser frequency. As the heritage of

archaic periods of Roman concrete use, the opus mixtum served as the minor technique still in

use side by side with opus testaceum (Wright 2005, 193-194). Opus mixtum combines stone

and brick elements in facing of the core, because fully stone facing33 was first to be used in such

constructions during period of Roman Republic.

Masons avoided cracks by overlapping the bricks over each other and they also came

with principle of using triangular bricks, instead of rectangular, the idea was probably, that if

the concrete core bonds with bricks, the pressure, causing cracks, would be transmitted to the

core as well, making the structure form complex aggregate. The roughness of the break, while

the brick were separated by saw or other cutting tool, adhered perfectly to the mortar (Adam

1999, 293). Tooth-effect on the inner surface created by vertexes of triangular bricks fits

31
Opus latericium is reffered to by Vitruvius, as composed from unbaked bricks (Vitruvius II, III, 2).
32
The difference between mortar and concrete is basically, that concrete is compound – with strong bonds
etween its componnts, firm enough to form the core on itself without additives, like rubble, undressed stones
and other, which mortar contains in various ammounts, and thus is rather mixture. The disadvantage of using
of mixtures is, that under the pressure of heavy loads, they tend to disintegrate the wall horizontally, causing
cracks (Taylor 2003, pp. 98).
33
Mainly opus incertum and opus reticultum contributed on the wall construction (Wright 2005, pp. 191).

55
perfectly for making the bond linked. The principle was both technically convenient and

economic, as the tiles were remarkably more expensive than the same volume of concrete, they

were sawn and put into the place with diagonal – the longest cross section of the brick, covering

the largest possible are on the face of the wall, and thus decreasing the number of bricks

needed34. This setting increased the amount of mortar or concrete required for the structure, but

these were cheaper materials. In Severan period, not only triangular, but also trapezoidal bricks,

made of roughly cut bricks, mainly sesquipedales. George Wright pointed out, that sawing of

each solid brick would be very difficult operation requiring large amount of labour if multiplied

on the vast number of bricks needed for the whole wall and offers explanation, that much easier

would be to divide triangles by knicking and snapping long the crack (Wright 2005, 193). This

process probably took place also at reshaping old reused roof tiles for roof constructions. Bricks

with diagonal leading lines can serve as evidence for use of snapping.

Fig. 17) The division of bricks into triangles (from Adam 1999, fig. 347).

In vertical intervals, there are single brick courses bonding both sides of the brick facing.

They are used from Flavian times till the 3rd century, mainly in Roman Italy. It is very distinctive

34
Bricks laid with diagonal face outside saved up to 40 % of bricks compared with amount needed if laid
orthogonally (Taylor 2003, 98).

56
element, which can be found in wall section. The bonding course is usually composed of large

bipedales bricks, three of them in line, if the wall was six feet wide. Their main purpose is not

quite clear, but they probably served as an interval between concrete pouring phases, securing

underlying concrete stratum by creating the new floor for each level. Because the height of each

stratum closed with bonding course was 1,2 to 1,5 m (4 to 5 Roman feet), it has in round

numbers convenient dimension for workers to lay bricks without difficulties. At the same time

it could mark the one day labour level, sealing the finished batch to protect it for curing, and

this course offered surface for putting logs ad other parts of scaffolding and could be employed

as suspensory space at the time of concrete curing, when other supporting activities took place.

If the courses were placed regularly, they also provided vertical measures, and guides

controlling the regular verticality of the wall. Bonding courses were convenient if the wall had

to sustain the huge loads placed on them, because the wall was reinforced by them, but on the

other hand in the areas with earthquakes, they could cause cracks running along them and thus

slipping of the whole part of the wall, with bonding course acting as sledge. However, it is a

question, how much the freely floating bricks (because only two bricks at the sides, which were

bonded with the facing brick wall were truly locked with the structure, the rest of them was just

floating in the masses of concrete within the core) could prevent the liquid from settling more

than necessary. Because if the concrete really was settling dramatically, it would probably press

the free floating bricks down anyway, only if the cover layer was already cured and steady

enough. Second purpose for bonding layers was to keep in the moisture and maintain relatively

stable temperature, protecting the concrete from frost and sun. Roman pozzolano concrete is

curing much better in moist environment (and was also used for hydraulic structures build in

water), which helps it to harden. It is converting water and lime into hydraulic silicates. This

process benefits from even temperatures and this can be guaranteed by protecting brick layer.

Frequent correlation between the level of bonding courses and foundations of relieving arches,

57
vaults and windows shows, that the bond truly functioned as occasional establishing platform

for supplementary constructions. The relieving arches for example couldn´t be founded without

centering running through the whole thickness of the wall, because it would disintegrate easily.

The contribution of lines s vertical measurement is secondary employment of them, it is hard

to imagine, that they were founded just because of this purpose, especially if in many cases they

are distributed irregularly. Clearly, this diagnostic function emerged from more practical

physical ones (Taylor 2003, 101-106).

Fig. 18) Example of wall with concrete core faced with triangular bricks (from Adam 1999, fig. 348).

During period of concrete curing, masons had to occupy themselves with different parts

of the project – preparing the next batch of bricks and rearrangement of scaffolding. Putlogs of

scaffolding were not placed into the gaps in constructed wall, but were connecting the

scaffoldings on both sides, above the construction level of the wall. Due to this technique, after

finishing each level, the scaffolding had to be adapted to the new settings. The stopping point

for new level of bricks could be highlighted by guidelines. Masons put bricks into the position

course after course, and if necessary, shorten bricks with hammer. Bricklaying continued until

the working space was comfortable for masons, probably to the height of the chest. Concrete

or mortar with rubble were poured inside the brick skin randomly or in horizontal layers, after

that workers removed the air pockets, ramming the filling with timbers. There are traces of

small discrepancies on the brick courses, heights of the courses are slightly shifted. Evidence

58
like this is probably due to different teams or workers, whose areas were meeting at this point

(Taylor 2003, 101).

Construction of windows and relieving arches forced masons working on the opposite

sides of the wall to level their courses at the same height, and thus functioned as regulation and

control for workers. The bottom of the window then begins at the level of bonding course and

the weight pressing on it from above is supported by relieving arch in some distance above the

window. Wooden centerings were used to support the arch before completion. Zone between

window arch and relieving arch was constructed in solid bricks in full thickness of the wall.

The relieving arches could be built on separate centering and the gap, which emerged because

of this setting was filled after completion of the arch, but alternative building of the arch directly

on the brick construction is also possible option. In more complex structures, relieving arches

were built in cylindrical wall, curving not only vertically, but also horizontally in plan, which

were much more unstable, than simple arches in straight walls. Only rarely are these

horizontally curved arches composed by stone, because brick, due to its lightness form steady

support for pressures from above. Special light concrete voussoirs sometimes altered bricks in

the core of relieving arches, which were slightly wider than voids they were relieving. The

distance between relieving arch and the void is not uniform, but in vault construction, there are

commonly used protecting arches to absorb the pressure of the vault. Occasionally, the window

was not equipped with relieving arch, but double course of bipedales bricks in its own arch

instead, which was sufficient protection (Taylor 2003, 107-110).

Opus mixtum generally refers to many types of masonry with basic characteristics to be

composed of stone rubble altered with bricks. The special composition or pattern of these

elements in the wall varies. The technique is known for sure in the period between the Sulla´s

dictatorship and the reign of Augustus, because some buildings in Pompeii from this period are

built in opus mixtum (Adam 1999, 278), but the boom of this technique comes with Flavian

59
period. Decorative function of the facing could play the role in employment, because there is

usual habit to alter regular courses of bricks with regular courses of stone creating elegant

pattern. The technique using the tradition of opus mixtum was dominant method for building

up public monuments and large private projects in Byzantine Empire (Krautheimer 1986, 169)

especially in Constantinople, where other materials were scarce, and from there the technology

was spread as fashionable into other important Byzantine cities. Scaffolding was built into the

walls, which can be roved by tracking the putlog holes. The scaffolding was not necessarily

removed with wall rising, but could also serve for plasterers and painters, at the end the putlog

holes were filled with mortar and the scaffolding finally removed. From literary sources we can

identify three different types of scaffoldings: hung from above, freestanding with its base firmly

on the ground level and in the form of platforms built into the wall (Krautheimer 1986, 185).

The question of chronologically sensitivity is questionable, even though it seems, that

the thickness of the brick course changes in different periods. Arrangement of bricks into opus

spicatum, or herring fish-bone is rare in walls and served mainly for decorative purpose (Adam

1999, pic. 350).

Even though in Greece, the brick as a facing material is not as common as in Rome and

West in general, it still can be traced in this region more often than in Asia Minor where this

technique is very rare. Generally can be Roman architectural influence in Greek region traced

in cities with stronger Roman garrison or proximity of military post (Waelkens 1987, 102). In

Greece, it can be found in Gortyn, where the triangular brick facing is major wall structure of

great building projects there (for example in odeion and amphitheatre), but also in the mainland.

The Antonine Baths at Sparta are built in the same way, and at Buthrotum, the nymphaeum has

mortar rubble core faced with brick-triangles of various shape (Dodge 1987, 107).

60
Fig. 19) Example of opus mixtum, Pompeii around 300 AD (from Adam 1999, fig. 337).

6.1.3. Solid brick walls

In western part of Roman Empire, especially in Italy and Rome itself, the combined

techniques for brick walls dominated, with majority built with concrete core and brick facing,

but in eastern part of Empire, in Greece and Asia Minor, two regions with strong Hellenistic

tradition, slightly different direction of brick construction development took place. The concrete

was known in Greece from Classical period, but lacked the qualities of Roman pozzolano

concrete (Waelkens 1987, 95), which was substituted with lime. The faced walls with lime-

mortar core were used, but this technique did not take the leading role, as in Rome. Between

other techniques, except of stone walls, also the solid brick walls took important part. In the

most important town of Asia Minor, Ephesus, solid brick walls became widespread (Waelkens

1987, 96). Also the ashlar masonry was typical sometimes supplemented with upper part made

of bricks. In opus mixtum setting, the brick courses alternating with stone rubble run through

the whole thickness of the wall.

Examples of walls constructed fully with bricks in Greece are from Athens, Argos,

Isthmia or Olympia. The bricks are whole, sometimes broken, but always faced, and coursed

and joined together with lime mortar. In Asia Minor solid brick walls does not occur so often,

even though one of the most remarkable buildings from Roman period here, the Serapeum at

61
Pergamon is rare example of solid brick construction35 (Dodge 1987, 108). In the second part

of 5th century, the preferred technique used in Constantinople was solid brick masonry,

displayed at the construction of the church of the Theotokos in the Chalkoprateia (Bardhill

2008, 338), but later in the 6th century, the combination with stone returned as the main method.

6.1.4. Alternate stone and brick bands

Bricks bands altered with square stones mortared and with concrete core are sometimes

labelled as Asiatic technique. The main difference between this technique and opus mixtum is,

that the latter refers strictly to the wall faced with bricks and reticulate stones with homogenous

concrete core running through the whole wall, while in this technique, the core is divided by

separating bands into smaller sections. Bricks bands running through the wall are joining the

two facings composed of stones and thus providing coherent structure. The technique was

largely in use in Greece, Asia Minor and Balkan from the beginning of Roman period, but in

Rome and West it was used later, approximately from 2nd century AD. It developed to important

technique of Late Antique Empire, widely used in Constantinople and later Byzantine buildings.

In Athens, Antonine aqueduct to the nymphaeum was built in stones altered with brick courses,

and in Serdica the whole circuit of city walls employed this type of construction. In the period

of Tetrarchy and division of administrative power of the Empire into separate capitals, the

technique became typical for newly founded buildings in state power centres (Dodge 1987,

108-112).

The relation in thickness between brick course and mortar joints is assumed to be of

chronological importance, however it is rather general, than precise tool for dating. For region

of Greece in the 4th century AD, the mortar-to-brick ratio is about 1:1, but in the 5th century

there is considerable increase in mortar layer, which becomes twice the thickness of the brick.

35
Monument also known as „Kizil Avlu“ (Dodge 1987, pp. 108).

62
At some sites there is increase in the thickness of the bricks, such as in Philippi, from 5th to

6th century AD the expansion is from 3 up to 5-6,5 cm. The number of brick- bonds in one

brick-course also changes through the time, but has only supplementary importance for dating

(Dodge 1987, 112-113).

6.2. Arches and vaults

The CBM was often used for great exploitation of arch and vault, which appears in

Rome in the 6th century BC. The strength of this technique lies on the radiating joints between

wedge-shaped voussoirs, which are directing the pressure to the supporting walls on sides

(Lancaster 2009, 257).

The simplest form of arch is corbelling, using stone lintels. But such form is quite

unstable and the pressure is not directed regularly to the piers, thus leaving the significant

weight on the lintel itself resulting in greater risk of collapse of the corbelled arch (Adam 1999,

331). Corbelling was convenient for heavy stone architecture, but not for CBM constructions,

whose advantage is firm, lighter structure composed of small elements. The weight of the true

arch should lie mainly on haunches, drawing the voussoirs together, not on the crown, which is

the most vulnerable part of the arch.

For construction the scaffolding was necessary condition to support the unfinished arch,

because only with the haunches loaded with weight of the wall above the arch is able to stand

alone. The exact curve of the centering was crucial for structure to withstand. The centering

and scaffolding was made of wood, while the tendency of the carpenters was to save as much

material as possible and sometimes the beams were erected from elevated protrusions in the

wall. The protrusions could later serve as part of decorative architectural elements of the wall.

To save wood for scaffolding, the centerings were reused or even transported, and for each row

of voussoirs in the arch, the separate centering construction was employed. Some arches were

63
built without temporary scaffolding, but ceramic cylinders were used instead and then left in

position, while forming the part of the structure. These vaulting tubes fitted into each other and

thus the arch could be erected at the place without any support (Adam 362) and right after

loaded with other layer of solid bricks copying its curve to maintain the shape of the arch. Then

the concrete could be used to bond the structure together.

Fig. 20) Scheme of disposition of weight on the arch (from Adam 1999, fig. 395).

Fig. 21) Terminology of arch elements (from Adam 1999, fig. 401).

64
The early development of vaulting ribs shows Roman interest in controlling the forces

in the vaults. The masons reinforced the concrete vaults with radially set stones or bricks36. The

original stone settings evolved into use of brick for reinforcement during first half of 1st century

side by side with general progress of techniques based on brick-construction37. The biedalis

ribbing decreased at the end of the 2nd century AD with downfall of Imperial power. In periods

after that, the solid ribs were changed with lattice ribbing with bipedales separated by smaller

bricks, so lower amount of large bricks was needed.

Brick vaulting in provinces differed from that in the Roman capital. Radially laid bricks

in vault of Bath in Miletus were supported with haunches filled with stone rubble. Another type:

“The pitched brick vaulting, which required minimal centering, because the bricks were laid on

each other and joined with fast-drying gypsum. This technique was usual in areas scarce in

woods, like Egypt, but in the 2nd century AD, many variations of pitched vault emerge at Argos,

Eleusis or Isthmia, where the span of the vaults is rather small, with exception of Theatre Bath

at Argos, where the span is 11 metres, and the centering was required, so this technique was

used for other unknown reason. At Ephesus, Izmir or Aspendos another style can be examined,

using combination of radial brick vaulting in the haunches and edge-to-edge brick setting in the

crown (Lancaster 2009, 274). In the Roman west, special bricks were used for vaulting – tubuli

cuneate or hollow voussoirs, mentioned earlier, especially for construction of bath houses

(Brodribb 1987, 79), and armchair voussoirs reducing the weight and thus exceeding the

stability of the vault (Brodribb 1987, 46). Tubi fittili (vaulting tubes) enabled masons to avoid

centering, while the tubes were interlocking and secured with fast-drying gypsum, which are

proliferated the most in Roman North Africa. The technique could had been adopted from kiln-

36
Bipedales were commonly employed for this reinforcement for their length, which provided more space for
mortar joint and thus friction securing the brick at the place (Lancaster 2009, pp. 270).
37
The stone i stoo heavy and inconvenient expecially for large vault-spans (the weight of limestone
2,600 kg/m3), while brick is light and fitting for vault-construction (up to 1600 kg/m3) (Lancaster 2009, pp. 272).

65
construction and is anyway strongly connected to pottery production as the tubes could be easily

manufactured at the same workshop as pots (Lancaster 2009, 277). Unique case is vaulting

employing amhporas set into concrete bedding, occurring at first in Hadrianic period, but also

later in the 3rd century. The aim of this setting was probably to save labour and material, because

the common explanation – that amphorae in the vaults should reduce the weight - is false, as

the vessels are placed at the wrong part of the structure without remarkable influence on

reducing the lateral thrust (Lancaster 2009, 278).

The vault of large domes were made of concrete moulded into the wooden frame, again

seated on the proper centering. However the concrete appears as perfect material, its

disadvantage for vaults was paradoxically its monolithic character, which caused immense

tensile stress for lateral walls, which were supporting the vault, and thus increasing the danger

of collapse of the walls outwards. In some domes, this tendency is avoided and no deformation

of the wall in the area, where the haunches are joining it. In these domes, the technique was

employed to divide the concrete mass by courses of bipedales running horizontally through the

thickness of the vault. On extrados of the vault the step-rings were implemented at the ends of

brick courses providing separate relieving surface for the circles above, and thus the lateral

thrusts were not resting against the supporting walls causing their ruptures, but were more

lightly resting on the underlying step-rings (Lancaster 2009, 267).

In Early Byzantine period the vaulting tubes were fully exploited, but no more as

permanent centering, but as regular vaults, without other significant agent that they should

support. Large brick made domes over rotunda or polygon were used from beginning of

monumental building projects in Constantinople, for example in imperial mausoleum or in

hippodrome. The experiments with brick-vaulting led to construction of large dome on

pendentives in Hagia Sophia (Bardill 2008, 341). Also amphorae and usual kitchen pots also

66
served for vault construction in this period, now for relieving the weight in haunches and

support the roof above (Ousterhout 1999, pic 194).

6.3. Floor pavement

Clay as such served for flooring naturally from emergence of architecture. In rural areas

and simple domestic buildings floor was simply “beaten earth” providing sufficient surface.

Bricks and tiles make very functional floor. In rural environs and modest structures, pottery

sherds (Klontza-Jaklova 2014, 139-140) and secondary used CBM rubble could serve as simple

reinforcement of beaten earth floor levelling.

Romans devised very robust brick flooring made by testacea spicata technique, putting

bricks on edge, forming herring bone pattern (Wright 2005,

134). For constructing the herring bone floors we can follow

Vitruvius, who says, that the edges must be carefully levelled

and “The herring-bone pattern,…, must also be carefully

finished, so as to be without gaps or ridges sticking up, but all

flat and rubbed down to rule. When the rubbing down is

completely finished by means of the smoothing and polishing

processes, sift powdered marble on top, and lay on a coating

of lime and sand” (Vitruvius VII, 1, 4). In advantage, brick

floors were in general impervious, and thus could be washed

down with water. Various geometric-shaped bricks formed Fig. 22) Example of opus spicatum
flooring (from Adam 1999, fig. 541).
“mosaic-like brick floor”, imitating precious marble

patterns38. These floor bricks are especially employed in Danubian region (Brodribb 1987, 53).

But far more frequent type is plain-brick pavement, composed of all usual kinds of bricks or

38
This type of floor is conventionally labelled as opus sectile.

67
even pan-tiles. But as the ceramic paving was more brittle than stone, it tended to be wore out

easier (Adam 1999, 475).

6.4. Roof construction

Roofs can be generally divided into two groups according to their shape: pitched roofs

composed of narrow planes (or roofs combining narrow planes), and vaulted roofs covering

rounded space. The latter came to use later in Roman practice and also widely in Byzantine

world.

The pitched roofs were placed on roof timbers. On example from Pompeii, where the

roofs were conserved for examination we can follow the construction process of the timbering:

Horizontal pieces called purlins simply go from one gable of the wall to another. The purlins

hold rafters on which the laths are placed. These can be the straight support for tiles, or smaller

timbers, which re then covered with tiles lengthways. The basic structure was convenient for

usual dwelling houses and buildings with small span, but for larger constructions, it would

require very long beams, which could be gained only with difficulties, or in not really economic

manner. For this purpose the “triangulated truss” was implemented, using self-supporting

principal rafters leaning against each other directing the pressures to the ends of the tie-beam,

which becomes the piece under tension because of the central vertical beam (so called “king

post”), which supports principal rafters at their ridge-end. The triangulated truss thus becomes

very rigid structure (Adam 1999, 419-435). These two techniques served as the base for pitched

roofs throughout the Roman and Byzantine Empire (Bardill 2008, 343), through which the

know-how was transmitted to Middle Ages. In special cases were supplemented with additional

beams due to their dimension or shape.

Alternatively, for vaulted roofs, the direct clay construction was employed. On the

supporting structure the tiles of various materials were used. Very reliable and convenient were

68
metal plates, for their high acquiring price they covered more important buildings, like domed

churches. The ceramic tiling system had to be simplified to this curved settings, and the style

composed of semi-circular tiles set in alternating convex and concave adjustment was applied

(Wright 2005, 135).

On the slanting timber construction of the pitched (or “saddle”) roof, the tiles could

easily become slippery, and so the slope must have been of low angle, around 15 to 25 degrees

(Rook 1979, 295), under which the tiles would rest on the roof by their own weight, supported

sometimes by groove on tile interlocking with the next. Nail-holes appear only exceptionally

and mainly with eaves-tiles. Clay bed under the pan tile was not common practice of fastening

to the surface, but cover tiles could be fixed to the pan tiles by layer of mortar or clay (Wikander

1988, 208).

Fig. 23) The triangulated truss: 1 tie beam; 2 principal rafter; 3 king post; 4 wall-plate; 5 purlin; 6 ridge purlin;

7 cleat; 8 common rafter; 9 boarding (laths) (from Adam 1999, fig. 491).

69
7. Case studies of CBM in archaeological environment

In this chapter, the CBM will be viewed from archaeological point of view, as part of

collection of excavated artefacts. In archaeological settings, CBM usually occur in collapsed

deposits and fragmentary form, especially if CBM was used for construction of elevated

structures, what is after all its primary function. Chosen sites re situated in Aegean region with

particular stress on Crete, because the aim is to understand material from East Crete site, and

thus the closest parallels supposedly lie in proximity of this site. Generally speaking, as the sites

were excavated by different archaeological teams using particular methods, and at different

periods, we can see also different approaches to CBM. In some records, the CBM is only

mentioned at the site, or only the schematic drawing is published, sometimes the records on

CBM is missing, even though the CBM is actually missing at the site only rarely. Fortunately

very often the approach is complex, and the CBM is examined properly, however the features

and characteristics are recorded with different level of involvement39.

For the chosen case studies: At first, the archaeological background, history and related

information are repeated in short and then the most frequent examples of CBM are listed and

described to illustrate their variety and features recorded by authors of the research.

7.1. Kopetra

The site of Kopetra was examined with aim to provide information for period of decline

of the Roman Empire, which is usually illustrated on the examples of large cities, and rural

areas are overlooked. The site lies in the valley of Vasilikos river in western Cyprus,

approximately 4 kilometres from the sea. The survey seasons took place between years 1988

39
For example, even the basic description of inclusions contained in fabrics is missing.

70
and 1992 and gave the researchers the picture of small rural village, which economy was based

mainly on farming in convenient surrounding area, mining of local resources and trade. As

evidence for economic growth can serve building activity, resulting in construction of three

churches and their reconstruction (Rautman 2003, 9-11). Main boom of the village can be dated

to 6th and 7th centuries AD, when the settled area covered up to 4 ha and could serve as home

for around 600 inhabitants (Rautman 2003, 44).

The main building materials were stone, wood and clay. As the bricks were used only

occasionally on the whole island, the CBM includes mainly roofing and paving tiles, which

were dominant roofing element in the island. Three groups of tiles were gathered, belonging to

three different roofing styles. Long, slightly curved tiles of Laconian type, two kinds of

Corinthian tiles with distinctive edges. According to study of fabrics, tiles of Laconian style are

of fine red fabric, 1st Corinthian type is of fine yellow fabric and 2nd Corinthian type is of red

coarse fabric. The coarse 2nd Corinthian type was produced locally, the fabrics of Laconian fine

tiles have origin around Paphos, in round numbers 80 km far, and 1st Corinthian type tiles have

origins in Mesaoria plain some 70 kilometres far, and as the only local variant was the less

numerous one in number of pieces, the author assumes, that the other two types were brought

in large groups, ordered specially for particular building projects (Rautman 2003, 48). The

building technologies are simple and didn´t require the presence of architect of any kind. Within

the structure of roof in churches, the gypsum and mortar were used, creating arches, which were

then holding the roof composed of fired tiles. Mudbricks served as the material for upper parts

of walls of houses and the spolia from older buildings, like shards of tiles were incorporated

into the later walls. The floors are simply made of clay, or as in the churches, from luxurious

opus sectile and marble slabs, which are more difficult to wear out.

Three churches are dated to the 6th and 7th century, while two of them are slightly later,

then the last one covered by red-fabric Laconian tiles. They were covered by series of sloped

71
roofs, with semi-conical shape on the apse. The lower rows of flat pan tiles were laid on the

layer of mud and straw, which was placed on the timber construction. The second rows of semi-

circular tiles covering the gaps between pan tiles were secured at the place by mortar, or were

not secured at all and only laid into positions, holding there each other using their weight.

The Laconian fine tiles and Corinthian fine tiles are spread in all three churches, while

the last – locally made Corinthian coarse tiles are not, which let the author assume, that the

different types were used during original construction and different types for later

reconstructions.

For the Laconian fine tiles (one piece made possible the reconstruction of the shape,

measuring 70 x 34 cm), the weight of the pan tile is calculated on 5 kg, while the covering

imbrex has is estimated to be only one quarter of weight of tegula. The flat pan tiles should

overhang each other in the row 25 cm, the final weight for a square meter is 26,25 kg, while 4,2

tegula pieces used, supplemented by sufficient number of covering tiles (Rautman 2003, 53).

For Corinthian fine tiles the similar calculation is made, while the flat pan tiles are

smaller and heavier, one weighting ca. 6 kg, measuring 44 x 35-40 cm, which would make one

square meter weight 45,5 kg, employing 6,06 tegula tiles. Covering imbrex tile would again

weight only 25percen of tegula.

Corinthian red-fabric locally produces tiles measured 56 x 48 cm (tegula) and weight

14-15 kilograms, which means about 65 kg per square metre and are then the heaviest ones of

the three types used at the site. The advantage of these tiles is obvious – they were easy to get,

while the local sources were exploited, without need to wait for large supplying groups from

distance.

The estimated number of tiles for roofing the three churches (dimensions in rough

numbers: 11 x 19, 15 x 19, 15 x 24 metres, plus additional space, narthexes, etc.) is about 15 000

72
pieces of tiles with weight approximately 60 000 kg. The other CBM material was used for

construction of roofs of houses in the 6th – 7th century village.

As the yellow Corinthian tiles were produced in the region of Salamis, which has been

in the past intensively Hellenised area, Marcus Rautman assumes, that production of this type

of roofing tiles marks continuity in older tradition of monumental buildings, which were roofed

by large heavy tiles. These Corinthian roof tiles were indeed the most popular type used in

Classical Mediterranean, at least for public buildings and other large buildings. These

conservative tradition of older periods can be recorded on the examples of two earlier churches,

while the later one is already roofed by much lighter Laconian tiles (26,25 kg per square metre

compared to 45,5 kg per square metre in the case of yellow Corinthian system). This popularity

of this latter roofing system came probably with late antiquity and continued through the Middle

Ages until these days (Wright 2005, 135), when the modern and reconstructed churches are

equipped with roof, following this Laconian style tradition. It is also tempting to suggest, that

local lighter Laconian tiles were used for repairs, because of their low weight, which would

possibly save the older roof from pressure, which could endanger its perseverance. For repairs

however, could have been used the locally produced red-fabric Corinthian tiles in the times,

when the number of imported tiles was scarce (Rautman 2003, 54-55). The technology of

repairs probably depended on more than one condition and both of the theories could be applied

by builders.

Another important information brought by this survey is that the CBM material (its vast

majority in this example) was transported on large distances even to the rural areas of only local

economic importance, and even when the area of final consumer provided sources for own

production. The CBM was transported supposedly on the small marine vessels and from the

coast on wagons (Rautman 2003, 55), with intention of application on particular building

projects.

73
CBM used at the site:

1. Laconian pan tile. The fabric is red, medium fine, contains limestone inclusions.

Sections of tegulae are slightly concave with raised rim on three sides. Maximum size

is 33 x 70 cm. Tiles are occasionally inscribed.

2. Laconian cover tile. The fabric is red, medium fine, contains limestone inclusions.

Convex on the sides, the sharpest angle in the middle. Maximum size 13 x 19 cm (no

whole piece was found, original estimated size 13 x 70 cm).

3. Corinthian pan tile. Yellow fabric, medium fine, occasionally inclusions. Flat part with

sharply raised side edges, triangular flow directors at the lower art of edges. Maximum

size 43,5 x 35 cm.

4. Corinthian cover tile for yellow fabric pan tile. Red, medium fine fabric with occasional

inclusion. The shape is made in the mould with pitched outer ridge with three edges,

and regularly curved inner angle. The upper border at one end is raised. The maximum

size 42,3 x 10,5 cm.

5. Corinthian pan tile. Coarse, red-brown fabric with many limestone and quartz

inclusions. Flat with sharply raised upper and side edges, triangular flow directors on

the lower endings of the edges. The maximum size 55 x 46,5 cm.

6. Corinthian cover tile, covering for Corinthian red fabric pan tile. Coarse, red-brown

fabric with many limestone and quartz inclusions. Pitched outer ridge, but with only one

edge on the top, the inner surface also pitched with ne edge. The profile in the shape of

“opening V”. The maximum size 28 x 11 cm (no whole piece found, estimated original

size 55 x 11 cm).

7. Water pipe. Pale yellow fabric, large red inclusions. Diameter 14 – 18 cm, maximum

length of the longest preserved piece 37 cm.

74
Fig. 24) CBM from Kalavasos – Kopetra: Laconian tiles 1, 2; Corinthian tiles 3, 4, 5, 6; water pipe

(from Rautman 2005, fig. 5.18).

7.2. Knossos

The site at Knossos in central Crete has been exploited by archaeologists since 19th

century. However the main interest of scientists and before them founders of the tradition of

Cretan archaeology Arthur Evans, were at the beginning more focused on the prehistoric

activities in the area and the upper archaeological strata were not examined so intensively, if

they were examined at all. Even though, our knowledge of later then Minoan settlements is

growing lately and the Roman and late antique artefacts and situations are now being studied

more properly.

The finds of many roof tiles are interpreted in connection to collapse layers, usually the

destruction of the whole roof indicates sudden end of habitation, caused probably by natural

75
conditions, for example earthquakes (Sackett 1992, 406). It is assumed, that the majority of the

Roman houses in Knossos were roofed with tiles, which fact is highly expectable, according to

importance of Colonia Iulia Cnosos, how Roman municipium was called, in the Roman Crete,

while after the capital of province - Gortyn, served as the second most important economic

point on the island. The CB material is also known from secondary use, from enclosures of the

burials (Corinthian tiles), where some whole pieces of tiles have been excavated, or as floor

packing (Laconian tiles) of drainage channel of late Roman phase of so called “North House”.

The tiles in primary use are scarcely found complete, while broken into small shards, during

roof collapse.

On the site the early tiles found are considered to be Hellenistic, due to the pottery

chronology from the same deposit. Corinthian and also Laconian tiles are documented, while

some of them are covered with coloured slip – black to brown (Sackett 1992, 407). There is

apparent similarity between Hellenistic tiles and later Roman types, so the continuity in

production is supposed. The main collection of preserved tiles and their significant parts is dated

to the Roman period of occupation of Crete. Both types, Laconian and Corinthian, were used.

Some of Corinthian pan tiles are stamped, what makes the chronological classification verified.

Corinthian cover tiles are unified in the shape, while the pan tiles differ in rendering of the ridge

on sides and on the top. Ridges documented have broad rectangular section, high and narrow

ridge section, obliquely cut upper part, or with side curving without proper ridge (Sackett 1992,

407). These types variety is coming from very fragmentary material but from well dated

contexts. As a reference to these types can serve whole preserved examples coming from tile-

bordered burials. The variants of Corinthian pan tiles have wide dating range, some types

running back to Hellenistic and Classical Greek tradition, but were obviously still in use in 1st

century AD. Type with flat surface curving up at sides ending with high flat ridge seems to be

76
the major type used. Laconian tiles are fragmentary and the distinction depends mainly on

fabrics, while the Hellenistic example is differentiated by its distinctive red-brown colour slip.

CBM used at the site:

1. Corinthian cover tile. Pink-buff fabric, core is red, surface sometimes more yellow. The

upper surface with angular gable, three breaks in total, inner surface with one angular

gable. Dimensions ca. 62 x 12-14,5 cm. Thickness 2-3,5 cm. Some of these tiles are

dated to Hadrianic and Severan periods, but without significant change in typology of

these tiles.

2. Corinthian pan tile. Red fabric with yellow slip (Augustan) to pink fabric with the slip

of same colour (Severan). Broad flat ridge at sides, small rounded ridge on the front

side. From the reverse side could be groove across the bottom, fitting with the front

ridge of adjacent tile in the next row. Dimensions ca. 70 x 58 cm. Thickness 2-3,5 cm

(body), up to 5,8 cm (ridge). The ridge on the back side is set in by 1 cm. Some of these

tiles are dated, considering the pottery content of context they were found in, to

Hellenistic, Augustan and Hadrianic periods.

3. Corinthian pan tile. Pink gritty fabric with “cream” slip. The body is curving up to the

sides, forming side ridges with flat top. The top edge of the tile contains lower ridge,

the bottom edge from beneath has groove to lock up with tile in next row. Some of these

tiles reused for water conduit leading stream to the cistern. Dimensions approximately

70 x 53 cm. The width of side ridges is very small. Thickness 3 cm (body), 5,6 cm

(ridge). The ridge on the top side is at the very end of the tile surface. Some of tiles from

this group are from contexts dated to Hellenistic, Augustan periods.

4. Corinthian pan tile. Red gritty to yellow fabric with pink to yellow slip. The upper

surface is curving very slightly on sides, forming obtuse angle (if compared to the

previous type). The side ridges are rather wide, the top of the ridge is tapered. The ridge

77
on the top side is not sharply divided from the body of the tile. On the bottom side from

beneath is wide groove for adjacent tile. Dimensions 70 x 63 cm. Thickness 2,8-3 cm

(body), 5-5,6 (ridge). Dating according to contexts to 4th century BC, 1st century BC,

Augustan period.

5. Corinthian pan tile. Pink gritty fabric, with black spots originated during firing. The

upper surface only slightly regularly curving to the side ridges, which are not sharply

accentuated. Dimensions cannot be reconstructed, as the material is too fragmentary.

Thickness 2,2-3 cm (body), 4,8-6 cm (ridge). Considering the contexts dated to

Hellenistic and Hadrianic periods.

6. Laconian pan tile. Pale pink to pink-brown to red fabric with dark grits, yellow to red

slip. Some pieces reused as drainage. Dimensions 110 x 38 (top edge) – 46 (bottom

edge) cm. Thickness 1,6 – 2,5 cm. These tiles were occasionally found complete.

Identified in Hellenistic and Neronian contexts, while Hellenistic tiles are thinner

(1,6 – 2,3 cm), while Neronian are thicker (2-3 cm).

7. Brick/building tile with deeply incised diagonal grooves. Coarse pink fabric, with dark

burnt spots. Very typical CBM type with regular diagonal cross connecting the corners.

Dimensions 27,5 x 27,5 cm. Thickness 3,5 cm. Accoring to content of context dated to

Severan period.

8. Brick/building tile of square shape. Yellow-green fabric, surface smoothed. Dimensions

14 x 18 cm. Thickness 4,5 cm. Dated to the 2nd century AD.

9. Disc tile. Red fabric, burnt spots. Fragment of disc tile used probably under the floor of

hypocaust.

78
Fig. 25) Examples of CBM from Knossos: Laconian pan tiles 7, Corinthian pan tile 2, Corinthian cover tiles 1

(from Sackett 1992, Plate 22-23).

Stamped and incised tiles occurred in excavation. The stamps with letters “CIC”

(Colonia Iulia Cnossos), “CINC” (Colonia Iulia Nobilis Cnossos), or with symbol of labyrinth

are dated to the 1st century AD, when workshop specialized in their production worked in

Knossos area. With stamp was probably approved the quality of the tile, guaranteed by

municipal officials, or relation to public building project, for which these tiles were used. The

inscribed Laconian tiles also occurred, but in this case marks served to specify the position of

particular tile in the construction of the roof or as other builders mark to clarify the function of

the element (Sackett 1992, 410).

7.3. Nichoria

On the Greek mainland site of Nichoria in Messenia region, the excavation were held

by Swedish Institute at Athens40. Exploration of the area for future survey began in 1958 by

McDonald, who focused on transitional period between Mycenaean settlement and early Iron

40
For more information visit the website of the institute http://www.sia.gr/.

79
Age activities (Coulson 1983, 3-4). Another continuous habitation on the site took place from

the Late Roman-Early Byzantine to Middle Byzantine period and were examined in building

area IV by William Donovan (Coulson 1983, 353). The excavation of most of the late antique

and medieval material started in 1969, while the chronology of strata were confirmed by

radiocarbon analyses and typologically thanks to amphora shards found in situ. In 1974 the

sorting out of byzantine roofing tiles was undertaken by Rosser in the areas II and IV.

The settlement on the site, where the location is very convenient for safe harbour, was

at first ruled by Roman government and after its decline continued through the Early Byzantine

period with the époque of prosperity during Middle Byzantine period, when also the rest of

Peloponnesus was flourishing.

The CBM material comes both, from collapse strata inside supposed buildings and

secondly from burials, where the same type of pan tiles, which can be found in collapse debris,

is used for covering the graves (Coulson 1983, 355). Parallels in pottery types allow dating of

the burials (and ergo tiles post quem) to the 5th – early 6th century. The church from 12th century

was identified in Middle Byzantine strata.

Seven basic types of roof tiles were recognized on the site (Coulson 1983, 384-385).

They are all of Laconian-roof construction type. The modern roof tile from nearby village was

brought for comparison, showing the prevailing tradition in roofing style, which is still used

nowadays. In the shape, three types were examined: Laconian pan tile, cover tile for, creating

set with former one, and the ridge tile with more apparent angle in the central part of the slightly

curved shape. This ridged covering tile could be easily interpreted as Corinthian imbrex, but

without presence of Corinthian pan tiles on the site, it is more likely only variant of Laconian

imbrex. Its used at the top of the roof. The provenience of the tiles was not studied, which could

be interesting, considering the fact, that the site served as the harbour, and import of larger

quantities of CBM would be then simplified. But as the all types appear to be of very modest
80
quality, the estimation can be made, that most of them are of local origin, or from neighbouring

regions.

CBM types on the site:

1. Laconian pan tile. Hard yellow reddish fabric, large white grits. Slightly convex, flat

edges. Late Roman-Early Byzantine. Dimensions 80 x 43 cm. Thickness 3,8 cm.

Occasionally three swirly finger marks.

2. Laconian cover tile. Soft yellow reddish fabric, large white grits. Slightly convex,

thicker edges. Late Roman-Early Byzantine. Dimensions 52 x 35 cm. Thickness 2,5 cm.

Occasionally swirl finger mark.

3. Laconian cover tile. Hard yellow reddish fabric, small red grits. Slightly convex,

rounded edges. Late Roman-Early Byzantine. Dimensions 50 x 29 cm. Thickness 2 cm.

4. Ridge tile, the shape between Corinthian and Laconian cover tile. Hard pale brown

fabric, small white grits. Sharply convex with apparent break, flat edges. Probably

Middle Byzantine. Dimensions 45 x 23 cm. Thickness 2,5 cm.

5. Laconian cover tile. Hard pale brown fabric, small white grit, straw impressions.

Convex with flat edges. Probably Middle Byzantine. Dimensions 45 x 23 cm. Thickness

2,5 cm.

6. Laconian cover tile. Thinner version of previous type. Hard yellow reddish fabric, white

grits. Convex, flat edges. Dimensions 45 x 17,5 cm. Thickness 1,5 cm.

7. Laconian pan tile. Hard yellow reddish fabric, white grits. Slightly convex, round edges.

Dimensions 80 x 37 cm. Thickness 2 cm.

(8). Modern cover tile. Hard yellow reddish fabric, small white grits. Convex, flat

edges. Dimensions 40 x 19 cm. Thickness 1,5 cm.

81
Fig. 26) Laconian pan tile 1; Laconian cover tile 5 (from Coulson 1983, fig. 10-77, 10-81).

Laconian pan tiles and cover tiles are here very similar considering the dimensions, the

angle of curvature and thickness. Cover tiles are in general slightly thinner than pan tiles, but

the distinction is small. The crucial element, while determining the type of tile, could be then

marks pictured on convex or concave side of the tile. The side with the mark is supposed to be

the upper side. Another important element in determining the type of tile are residues of mortar,

which was securing the tile at its place on the roof, or other possible methods of fixing CBM at

its place, while the traces of this binding element can be found on the bottom part of the tile.

7.4. Gortyn

Gortyn has major importance in Roman period, as the capital of Crete. Archaeological

exploration of the site started in 1884 with connection to investigation of inscribed law code,

the Italian School of Archaeology started its excavation on the site in 1898 and lasted until

1940. The second wave of research came after 1950 and is still continuing now. The

archaeologists of this second wave focused on Acropolis, Byzantine town, Praetorium and

82
Hellenistic fortifications. The continuity in habitation is from Hellenistic period, through the

Roman Empire till Byzantine period. The CBM was examined from Hellenistic fortifications,

Roman theatre, baths and Praetorium (Montali 2006; Di Vita 2000; Ricciardi, Allegro 1999).

Supposedly, there were workshops in vicinity supporting the numerous building projects within

the city area with CBM.

In Hellenistic fortification (Papadopoulos 1999, 254-269), the roof tiles served to protect

towers against the natural influences. Roof here was composed almost exclusively from

Corinthian type of tiles41. The upper surface of imbrex is ridged with three ridges and the inner

surface is regularly curved. Fabrics of the CBM were not examined.

1. Corinthian cover tile. Yellowish fabric. Curved on the lower surface, three twists on the

upper surface. Dated to Hellenistic period. Length 62 cm, width 16,5 cm, the curve is

7,7 cm high, 2 cm thick. Complete tile reconstructed from several pieces.

2. Laconian cover tile. Simple sharply curved cover tile. Dimensions 10 x 13 cm, 2 cm

thick. Not complete. Dated to Hellenistic period.

3. Corinthian pan tile. Length max 66 cm, width 53 cm, body 2 cm thick, side ridges 5,4 cm

thick, side ridges lowered on the upper end for 8 cm. Dated to Hellenistic period. Tile

completed from several pieces.

41
Originally „tegoli piani“ and „kalypteres“ (Papadopoulos 1999).

83
Fig. 27) Examples of CBM from Gortyn: Corinthian pan tile 3; section of irregular Corinthian pan tile

(from Papadopoulos 1999, fig. 227, 242).

For construction of Roman theatre in Gortyn, the main material for supporting structures

was baked brick (Montali 2006, 103 - 116), used in typical manner of Roman engineering. CBM

was studied, but more in relation to the composition of structures, combining mortar and other

materials using various techniques. The foundation of the building is dated to 2nd century and

the reign of Antoninus Pius and Marcus Aurelius (Montali 2006, 110).

4. Brick of pedalis type. Brick bears typical diagonal grooves, dividing the brick into four

triangles. Side is 29,6 cm long, thickness 4-4,2 cm. Dated to 2nd century A.D.

5. Brick of bipedalis type. Brick is 59,2 cm long on sides, with thickness 6-6,3 cm. After

the building dated to 2nd century A.D.

6. Brick of bessalis type. Side of the brick is 19-21 cm long, with thickness 3,3-3,5 cm.

Dated to 2nd century A.D.

Fig. 28) Example of pedalis brick 4 with mark from Gortyn (from Montali 2006, TAV. VIII).

84
The set of water pipes was collected from excavations of Praetorio (Rizzo 2000, 562).

7. Water pipe - tubuli. Maximal length of complete piece 60 cm. Part about 10 cm long on

one end is remarkably thinner to allow pipes to interlock.

8. Water pipe - tubuli. Maximal length of complete piece only 28 cm. Almost half of the

pipe is thinner to interlock with adjacent one.

Fig. 29) Examples of tubuli from Gortyn (from Rizzo 2000, fig. 208).

7.5. Oxa

The Oxa peak lies about 1 kilometre south-west from Elounda, in Eastern Crete. From

the historical point of view, the peak played the important role as bordering area for Classical

Greek poleis Olous, Lato and Dreros and as such, appeared as the mean for disputes between

these municipalities. The site was not put under the scientific investigation so far, and only brief

survey occurred there, which was recently published (Klontza-Jaklova et all., in press). The site

served as fort or acropolis (Geisler 2015) due to its strategic placement and convenient

geographical circumstances. The presence of Hellenistic, Roman and Byzantine activities on

the site is clear from pottery finds distributed over the area (Klontza-Jaklova et all., in press),

but yet, requires further focus. The CBM material is also detectable on the site, but only

preliminary conclusions can be made, comparing the material on the site with examined

locations in site´s proximity.

85
1. Laconian pan tile. Brown yellowish fabric, medium coarseness, limestone and quartz

inclusions. Section is slightly regularly convex. The edges are rounded, but slightly

protruding on the lower part of the edge. The maximum width of the sherd is 26 cm,

length is far from being complete – 17 cm. Thickness 2,2 cm. lower end of upper surface

is decorated with finger-made wavy line running along the edge, about 1 cm wide.

2. Rectangular brick of pedalis type. Red fabric, coarse, limestone and quartz inclusions.

The side is 31 cm long, thickness 3,6 cm. There are diagonal grooves connecting corners

of the piece. Both upper and lower surface carries remarkable traces of mortar.

Fig. 30) Fragment of Laconian pan tile 1 with decoration from Oxa (photo by author).

Fig. 31) Fragment of pedalis brick 2 (photo by author).

86
8. Case study Priniatikos Pyrgos

This chapter deals with CBM from Priniatikos Pyrgos site, which is published here for

the first time. The material is examined with informational background delivered by previous

chapters. After brief introduction, the methodology of research is explained and then the results

in the form of fabrics classification, recognition of CBM shapes and other features, like signs

and marks. At the end the recognized facts are used to hypothetical reconstructions of the CBM

structures. Conclusions on the material are stated in the last chapter of the thesis.

8.1. Introduction

During three post-excavation seasons (2013-2015), I was allowed to study CBM from

Priniatikos Pyrgos, mainly focused on Trench II, where huge amounts of tiles and bricks were

gathered. Most of the material is in literally fragmentary form and only few were found in the

one piece. Fortunately, the large numbers of CBM shards provided enough material to recognise

repetitions in shapes and fabrics an allowed me to distinguish main types used on the site. The

aim of my study was not to create new typology, but using the knowledge of ancient CBM,

bring to light new material, create methodology to explore it and eventually show the possible

way for future studies on CBM, which is, as archaeological tool, very often misremembered,

but which can be very useful in answering the questions about architecture of the sites explored

and solving the problematics of their destruction.

87
Fig. 32) The view on Mirabello bay in east Crete with Priniatikos Pyrgos site at the small peninsula protruding
from the sandy shore. (photo by author).

8.2. History a general information about the site

As an example for studying the CBM of antiquity can serve the material gathered during

few excavation seasons on the Cretan site Priniatikos Pyrgos in Mirabello Bay. The alone

standing small peninsula, bordered with cliffs provided the safe habitation place for its

communities from Late Neolithic period, through the Early, Middle and Late Minoan societies.

The site was occupied during formation of Iron Age in Crete, saw the emergence of classical

Greek polis, with its centre on the, few hundred metres along the coast standing, Nisi

Pandeleimon. Hellenistic period is recorded on the peninsula by lively building activity. The

settlement continued through the period of Roman occupation of Crete as part of province

Cyrenaica, when the harbour was in use in the area between Priniatikos Pyrgos and Nisi

Pandeleimon, and during Byzantine times served for dwelling and burying of dead members of

community next to sacral district with possible church with apse. Priniatikos Pyrgos is

an unique site for understanding the Early Byzantine period in Crete, and thus all new

information about this period are crucial. The times of Christian society at the site ended

88
probably with invasion of Arabic pirates, which were threatening coast of Mediterranean

islands since 8th century. As evidence, that the site was abandoned in haste can serve the lead

chalice, buried intentionally in the corner of possible church (Klontza-Jaklova 2014, 141). After

Arabs, Byzantines included their island back into the Empire by force and later sold it to

Venetians, who also left the traces of their presence in Priniatikos Pyrgos. After Venetians came

Ottomans in 1669 and remained masters of the island till 19th century.

The first excavation at the near site Vrokastro were led by Edith Hall and Richard Seager

in 1911-1912. The main focus of this expedition was to fill the gap in the knowledge of Iron

Age habitation, because the excavators at that time were interested mostly in Bronze Age,

Classical Greece an Roman period. During Vrokastro excavation, a small testing survey took

place in Priniatikos Pyrgos to find out, if the site is worth excavating, which was the common

practice at that time. Finds showed habitation from various periods and also connections of the

site with Mediterranean basin, unfortunately the Great War broke out in the autumn right after

the second season, and it made next excavations impossible (Betancourt 2014, 10-14). After

this short expedition, there was almost three quarters of century gap, until Barbara Hayden (The

University of Pennsylvania Museum) continued with another archaeological survey in the area

in eighties, focused on Vrokastro hill with late Bronze Age refugium and its environs. In 2002

geophysical survey took place, aiming to explore the coastal area and during that time

discovered two Bronze Age kilns on the site. The rescue excavation survey in 2005-2006 under

the director Barbara Hayden and Metaxia Tsipopoulou (Hayden 2012) were supported by

INSTAP (Institute for Aegean Prehistory), from 2007 Barry Molloy leads the survey with

contribution of Institute of Hellenic Studies at Athens.

From Neolithic and Early Bronze Age phases of habitation, vast numbers of processed

obsidian were found, which documents contacts with Aegean area (Milič 2014, 118-124).

Bronze Age settlement has continuity starting in Final Neolithic, through the Early Minoan and

89
Middle Minoan phase, where the destruction deposits are documented, but the buildings were

rebuilt immediately and occupation of the site continued. From this phase also the structures

which can be assigned to local elite are known, containing vessel for feasting rituals (Hayden

2012, 560-565). Traces of existence of craftsmen includes pottery kilns and metalworking

(Filippaki 2014, 105-112). The Late Minoan period shows slow shifting of settlement from the

coastal zone to the mainland, and in correlation, rise and decline of the harbour. In Early Iron

Age, the general pattern of Cretan settlement is moving into the extreme locations, so-called

refugies, situated at the top of the hills and accessible only with difficulties. Refugium in the

area of Priniatikos Pyrgos is the peak of Vrokastro more than 300 metres high, above the sea

level, facing the sea with steep slopes.

During 7th and 6th centuries BC, the Greek polis emerged near Vrokastro and east from

Priniatikos Pyrgos, with its centre on Nisi Pandeleimon and todays Istron town. Settlers

cultivated fertile soil in the valley of Istron river and the slopes were used for herding (Hayden

2014, 15-21). Classical deposits contain ash layers, probably resulting from ritual feasting on

the site (Erickson 2010, 305-309). However the center of polis of Istron lied few hundred metres

to the east at Nisi Pandeleimon. Field survey showed that suburbs of the city were spread also

to the area between Priniatikos Pyrgos and Nisi Pandeleimon and to the small peninsula itself.

The harbour was always important feature of the coastal settlement and through it contacts with

the rest of the Greek world were secured, and provide gateway for import of goods as well as

ideas documented by both – Athenian black, red wares and their imitations (Erickson 2010,

339-344). Settlement in the area continued through the late Classical to Hellenistic period, many

buildings were reoccupied and some of them re-modelled. The existence of the polis of Istron

ended at 180 BC, when the large city of Hierapytna conquered the northern coast, but obviously

the inhabitants lived here even after this event (Hayden 2012, 555-556). There is scarce activity

documented in archaeological material before 150 AD, because the general pattern recorded for

90
early Roman period was to settle in remote upland areas, here in Istron river valley, Prina and

Meseleroi. Breaking point comes with middle and later Roman periods42, when the coastal areas

are being populated (Hayden 2003, 209-210). The character of habitation is rural, but the

harbour is shifted from western side of Nisi Pandeleimon, where the level of sand was probably

too high for ships to land, and the new installation moved in western direction to Priniatikos

Pyrgos. In Early Byzantine époque, the importance of the site as the centre of rural area is

pinpointed by existence of at least three churches – ecclesiastical complex with cemetery on

the hilltop of Prinitikos Pyrgos, Agia Barbara to the southwest and another one, which probably

replaced older Roman temple at Marmara norther from the Pyrgos village (Hayden 2014, 20).

Coastal settlement was protected by watch stations situated on the hilltops of Vathi, end of

Ioannimiti and on the peak of Vrionisi (Hayden 2012, 563). Activities of Arabic pirates

threatened coast of Crete and caused end of the Early Byzantine occupation in the beginning of

8th century, what can be illustrated with deposit of pewter chalice, supposedly executed in hurry,

at the corner inside the ecclesiastical complex (Klontza-Jaklova 2014, 141). Arab rule lasted

from 824 to 961 AD (Makrakis 2011, 55) Byzantine centre in Priniatikos Pyrgos came into

importance again during 11th-13th century, after reconquering of the island from the hands of

Arabs. Venetians bought Crete in 1204 AD for their commercial interests in the East, and after

them the Ottomans conquered the island in 1645 AD.

42
Continuing into Early Byzantine period, because there is no strict division between Late Roman and Early
Byzantine periods in terminology of Vrokastro Area Survey, and these two periods are usually mentioned
together.

91
Fig. 33) Priniatikos Pyrgos, site plan with highlighted Trench II, the main source of studied CBM.
(Courtesy of Priniatikos Pyrgos Project).

8.3. Investigation of Trench II and related deposits

Trench II provided enough CBM material for further focus, and so this part of the site

was chosen as the base for CBM study. In this trench mainly Byzantine material appears, but

in lower strata, also Bronze Age deposits were found. The Byzantine layers come also from

Trenches I, IV and A 4000, A 6000. The main Byzantine features of Trench II are Buildings

1 and 2, five graves, which served as ossuary. The disposition of walls in trenches A 6000 and

A 4000 let us assume that they could be related to constructions in Trench II, however for better

understanding of this relation further excavations are necessary (Klontza-Jaklova 2014(b),

800). Byzantine Building 1 is, according to finds, multi-phase installation with distinctive

element – the apse, which could point to ecclesiastic nature of the building. Around the apse

are three other rooms related to it and rich in byzantine sherds and more artefacts. Byzantine

Building 2 is situated to the north from ByzB1, is partly collapsed and rich in cooking ware and

92
amphora sherds (Klontza-Jaklova 2014, 136). Part of Byzatine building was uncovered also in

Trench IV, and the floor from the same period in Trench I (Klontza-Jaklová 2014, 136).

Fig. 34) Priniatikos Pyrgos, map of architecture in the Trench II. (Courtesy of Priniatikos Pyrgos Project).

93
Byzantine Building 1 in Trench II is according to the pottery typology and numismatic

dating built in Early Byzantine phase (7th – 8th century AD). The building is not fully excavated,

but due to its shape and due to the finds, such as the goblet

made of lead, which has analogies in other goblets typically

used during celebrations of orthodox mass, it is assumed,

that could serve as an ecclesiastic building, at least for part

of its existence. When excavating, many building ceramic

shards were gained. The vast numbers of reused ceramics

were obviously used for creating the floor packing of newly Fig. 35) Context 706. Amphora sherd
secondary used in wall construction,
built construction phases, and the CBM, due to type variety still covered with mortar (photo by
author).

was gathered in the area, employing the ruins of older

buildings from period of Roman occupation. Spolia were also used as building material for the

walls (Fig. 35), not only CBM spolia, but also large shards of amphorae and pithoi, many of

them found in situ, still bearing the marks of mortar bounding them together. Important feature

in ByzB1 is floor packing layer C 26 in Room 143, which contains large numbers of CBM

sherds, many of them in good condition and also variety in shapes and CBM types. Importance

of this context is highlighted by fact, that it provides us with terminus post quem, because the

coin of Leon III44 was found in it.

8.4. Methodology

The CBM was the important material used on the site in many periods as supplementary

material for stone, which is very well available in the whole area. Mudbricks are recorded

already for Minoan settlement. The roof tiles and bricks are supposed for Greek, Roman,

43
For further reading on building and habitation phases of Early Byzantine period in Trench II, see Klontza-
Jaklova 2014, 136-140.
44
The coin was released between 717-747 Klontza-Jaklova 2014, 140.

94
Byzantine and later periods, but the recognition of them is complicated by fact, that CBM is not

chronologically sensitive material and the archaeological layers on the site are mixed.

As methodology for my work I borrowed the basics from very well established approach

for pottery studies. The material, which is now collected in depository in Agios Nikolaos, Crete,

was carefully studied and documented, using methods of drawing, photography, literary

description, comparison. At first, the different fabrics were sorted, according to their inclusions,

coarseness, relative quality of firing, pores, colour and colour according to Munsell colour

scale. During sorting I have met many unexpected difficulties, which I will comment later. At

the end, the different sorts of fabrics were gathered to more general groups, to provide the more

simplified and synoptic list of fabrics used for CBM on the site. So far no microscopic

petrographic analysis was made for this CBM material from Priniatikos Pyrgos, however

hopefully in the close future it will be done to clarify the results of this case study. The

inclusions were then recognized according to geology of the area and comparisons with pottery

fabrics, which have been studied carefully in the past on Priniatikos Pyrgos (Nodarou, Moody

2014). Second aim of the CBM survey was to explore the form of CBM material on the site.

The basic categories of interest were: study based on variety of shapes of bricks and tiles, study

of marks, signs, stamps and other intentional traits on CBM, recognition of imprints caused

during production processes. As the method was used observation and literary description,

photography, drawing and comparison. For the typology of CBM were used studies on the topic

(Mills 2013) and studies on traditional Roman materials (Brodribb 1987). A the end, the basic

shapes of CBM, gathered at Priniatikos Pyrgos site, were defined, and intentional and

production marks were categorized.

95
8.5. Fabrics

For exploration of fabrics was used macroscopic observation with magnitude lance

(magnification 15x), measure, the precise eye of the observer and the Munsell colour scale. If

examined macroscopically, the best way of observation of shards was studying of their breaks,

unfortunately the best breaks for observing were the fresh ones, caused by excavators, or during

other later incautious manipulation with CBM artefacts. On the other side, the freshly broken

material provided good chance to see the real composition of fabrics, not so largely affected by

post-deposition processes, as on the original surface of artefacts. Part of material was washed

manually, but this kind of manipulation necessarily causes changes on the surface of the

material, especially in softer materials, brush makes misleading lines and cracks, the pores are

fouled, some inclusions from the surface fall out, plus too fiery washing changes the shape of

CBM artefacts. The breaks caused by incautious manipulation with artefacts are without

question inevitable little tragedies of every archaeological survey, but if we are studying large

quantities of material, we can conveniently use these tragedies to gather some extra information

about examined material without necessity to employ the new invasive irreversible

interventions into artefacts. For further petrographic analyses is, of course, possible to choose

few examples of fabric types, but “first” division and definition of these types can be done with

contribution of unwanted breaks. If possible, the CBM material from Priniatikos Pyrgos was

examined in the areas of fresh breaks of the shards, where the original fabric could be examined

the best. The majority of CBM fabrics was studied on the washed surface of shards. The

inclusions which are naturally present in the area of Priniatikos Pyrgos and can be then qualified

as local are quartz, limestone, granodiorite, biotite (golden mica), siltstone (Nodarou, Hayden

2014). These minerals could have been present in the clay or quarried separately, when

exploited in the close distance of the site.

96
Aplastics Sorting Incl. Incl.max Incl.spher. Incl.shape Inclusions Voids shape Voids Munsell colour Colour
PP_CBM_01 25% poor 2 mm 3 mm high sphericity angular gold mica, quartz, limestone rounded 1% 2,5YR 6/8 light red
PP_CBM_02 5% good 1 mm 2 mm high sphericity angular gold mica, granodiorite, phyllite rounded 2% 10YR 7/6 yellow
PP_CBM_03 10% good 1 mm 1 mm high sphericity sub-rounded granodiorite, limestone residues rounded 5% 5YR 6/3 light reddish brown
PP_CBM_04 1% very good 0,5 mm 0,5 mm high sphericity rounded gold mica, granodiorite sub-rounded 5% 7,5YR 5/4 brown
PP_CBM_05 10% good 1 mm 1 mm high sphericity sub-angular granodiorite, limestone rounded 5% 2,5YR 7/4 pink
PP_CBM_06 10% good 1 mm 4 mm high sphericity angular black vulcanic stone (basalt?), limestone, gold mica rounded 5% 5YR 6/6 reddish yellow
PP_CBM_07 30% very poor 1 mm 5 mm low sphericity very angular phyllite, limestone sub-rounded 5% 2,5YR 5/8 red
PP_CBM_08 20% poor 1 mm 10 mm high sphericity sub-rounded quartz, limestone, schist rounded 2% 7,5YR 7/4 brown
PP_CBM_09 10% fair 1 mm 2 mm high sphericity sub-rounded quartz, gold mica, granodiorite, limestone rounded 2% 2,5YR 6/8 light red
PP_CBM_10 10% good 0,5 mm 1 mm high sphericity rounded limestone residues rounded 1% c. 5YR 4/1, s. 10R 5/8 grey, red
PP_CBM_11 5% good 0,5 mm 2 mm high sphericity sub-angular granodiorite long, sub-angular 10% 2,5YR 7/4 pink
PP_CBM_12 20% fair 1 mm 2 mm high sphericity angular phyllite, limestone rounded 1% 10R 6/8 light red
PP_CBM_13 20% fair 1 mm 3 mm high sphericity sub-angular silver mic, limesone, quartz rounded 1% c. 10YR 6/1, s. 2,5YR 6/6 grey, light red
PP_CBM_14 10% poor 1 mm 10 mm high sphericity angular limestone, brown volcanic stone rounded 5% 10YR 7/4 very pale brown
PP_CBM_15 20% very poor 2 mm 4 mm high sphericity sub-rounded rock crystal, granodiorite, quartz rounded 5% 5YR 7/8 reddish yellow
PP_CBM_16 10% good 0,5 mm 1 mm high sphericity sub-angular silver mica, limestone rounded 1% 5YR 8/1 white

Plate 2) The list of CBM-fabric groups from Priniatikos Pyrgos and their characteristics (abbreviations: Incl. – inclusions, sphere. – sphericity, c. – core, s. – surface).

97
PP_CBM_01: Together with PP_CBM_02, PP_CBM_09 and PP_CBM_11 the most

common building material from excavation. These groups are around 90% of all the material

explored. The group 01 makes some 15% of it. Red to red-pink fabric, with rounded inclusions,

probably simple sand. According to the variation of inclusions, this fabric is of local origin,

with clay quarried in the alluvium in proximity of the site of Priniatikos Pyrgos. The shards of

this fabric are usually slightly curved, belonging to Laconian pan tiles or sharply curved cover

roof tiles. No complete piece of CBM element of this fabric was found, but many shards of

medium dimension. This fabric was used for CBM production for long period starting probably

already in Hellenistic period.

PP_CBM_02: Fine sorted fabric of very light pink to yellow colour. Shards of this fabric

form some 15% of CBM finds from examined collection. Inclusions are scarce and if they are

present, they are very small and sparse. Material was not very highly baked. The shapes made

of this fabric are Corinthian flat pan tiles, from which corners with rim on two sides are

common. Also some large pieces of Laconian slightly curved pan tile were found, and sharply

curved cover tiles. This fabric was used for CBM production for long period starting probably

already in Hellenistic period.

PP_CBM_03: Very rare type of fabric, usually found with pottery shards

chronologically dated as Hellenistic. Good sorting of the fabric, highly baked due to limestone

residues, which are apparent on the shard break. The colour is light yellow, with pinkish dots,

the surface was possibly covered with yellow, light slip. Inclusions are rarely noticeable on the

surface. One shard of Corinthian pan tile with finger-groove was examined, the width is

unusually thin.

PP_CBM_04: Very fragmentary pieces of mudbricks. Inclusions are rare, probably were

washed away. Sometimes the imprints of organic additives are visible - straws of grass. The

98
fabric is fine, soft with pink to brown colour. Shards are small, fragmentary, rounded by impact

of natural elements. Mudbrick pieces were found already in Bronze Age deposits in large

quantities, they are scarcer in Roman and Late Antique deposits. The shape of original CBM

elements cannot be reconstructed due to the condition of material, but it is supposed that bricks

were made of this fabric, as the non-baked material would not fit for purposes of roof

construction, neither for paving tiles.

Fig. 36) CBM fabric samples: PP_CBM_01, PP_CBM_02, PP_CBM_03, PP_CBM_04 (photo by author).

PP_CBM_05: Rare type of fabric, very close to the type PP_CBM_03, could be only

variant of the same, the difference is in the lack of limestone residues. The fabric of light yellow

colour, carefully sorted. The usual shape for this fabric is curved tile – Laconian pan tile, but

more probably, as the examined examples are very thin, the cover tile.

PP_CBM_06: Very distinctive fabric, heavy and highly baked. The colour is dark pink

to very light reddish. Sometimes with „sandwich baking“ pattern, the core in the middle is more

black and the core of the shard closer to the surface becomes pink to light reddish. This fabric

is connected with the shape of brick with two regular grooves, each one joining the corners of

99
the brick. Dimensions of the brick are very precise, varying between 31 to 31,5 cm on the side.

The width is 3,4 cm, but more commonly 3,6 cm, slightly thinner in the middle. The

reconstructed weight of the whole brick is around 5000 g. As the grooves are very regular, they

were most probably pressed by some tool. The inclusions are small and sparse. The fabric is

strongly related to pedalis production and as such was used mainly during Roman periods.

PP_CBM_07: Coarse material of red colour, rich in inclusions, which consist almost

exclusively of flat, long scales of phyllite and rarely with addition of small limestone dots. This

fabric was obviously fragile and the shards are fragmentary, but spread in large numbers in all

examined trenches. The large shards allow us to suggestion, that this material was used for

production of Laconian pan tiles, because the surviving pieces are slightly curved and relatively

thick. The question is provenance of this fabric, as the phyllite is not natural resource in the

area and the closest known sources are in the region of Hierapetra, some 20 kilometres from

Priniatikos Pyrgos. According to pottery from the same contexts, this fabric was used in Early

Byzantine and later periods.

PP_CBM_08: Coarse fabric, with large inclusions. Clay for this type of fabric was

poorly sorted, but still, there are not many voids. Material is heavy and baked in high

temperatures, as it is very tough and solid. The thickness of preserved pieces help to suggest,

that this fabric was used mainly to form bricks. The surface of bricks is very lightly processed,

with lot of cracks caused by firing. In few cases the upper surface is marked with double grooves

made by fingers of the artisan diagonally, forming the double cross pattern. The colour of fabric

is pink to weak brown.

100
Fig. 37) CBM fabric samples: PP_CBM_05, PP_CBM_06, PP_CBM_07, PP_CBM_08 (photo by author).

PP_CBM_09: The largest group of CBM artefacts on Priniatikos Pyrgos is the one with

this fabric (approximately 30% of CBM). Light red colour on the break, with small inclusions

of quartz, granodiorite, limestone and gold mica, very rarely was present phyllite in microscopic

dots. According to the inclusions, we can assume, that the material could have been produced

locally45. The main CBM shape for this fabric is curved tile, both Laconian pan tile and cover

tile, but also Corinthian cover tiles and pan tiles were identified. This fabric was used for CBM

production for long period starting probably already in Hellenistic period.

PP_CBM_10: Highly baked fabric, with wide range of colour on the break. The

inclusions are not dense, but rather really small and numerous, they contain white bubbles,

identified as limestone residues after firing in high temperatures. The material is very tough

then and preserved in larger pieces. The large shards are compact and it appears, that this fabric

was connected strictly with flat Corinthian pan tiles. The colours of fabric are dark grey to black

45
For comparison with local pottery fabrics see Nodarou E. – J. Moody 2014.

101
in the core, getting lighter and redder to closer to the surface, up to red-orange colour.

Assumption is that this fabric was employed during Roman periods.

PP_CBM_11: This is the second most common type of fabrics (approximately 25% of

CBM), after PP_CBM_09, on the site. The coarseness and voids are very similar to the major

type, the main difference are inclusions and colour on the break. As inclusion was identified

granodiorite. Voids are thin and long. The fabric was examined almost strictly on curved roof

tiles, both Laconian pan tiles and cover imbrex tiles. The colour of the fabric is pale yellow.

This fabric was used for CBM production for long period starting probably already in

Hellenistic period.

PP_CBM_12: Some roof tiles from modern times have been excavated from upper

layers of debris. Their fabric is fairly sorted, with large amount of small inclusions of limestone

and phyllite. The inclusions were most probably quarried by purpose as additive, because they

are more angular, than inclusions in other types of fabric. The voids are very rare, as the material

is very well processed. The colour is light red and is uniform for the whole piece, same on the

break as on the surface.

102
Fig. 38) CBM fabric samples: PP_CBM_09, PP_CBM_10, PP_CBM_11, PP_CBM_12 (photo by author).

PP_CBM_13: Only few shards of this type were examined. The fabric contains huge

amount of inclusions, with very apparent silver mica, limestone and quartz. The colour of the

shard is dark grey to black in the core and red-brown close to the surface. Only small shards,

very difficult to read were found and identified as CBM, but they can be misinterpreted shards

of pithoi, which would be possible also if we consider the large content of silver mica.

PP_CBM_14: The rare fabric of CBM material on the Priniatikos Pyrgos site. In colour

and pattern close to groups PP_CBM_02 and PP_CBM_11, but compared to them, much

coarser, with poor sorting. Voids are rounded, creating up to 5% of fabric. There are imprints

of organic inclusions, probably grass straws. Inclusions are various in dimensions, consisting

of limestone and unidentified brown inclusions of volcanic origin. Shards belonging to this

group are of slightly curved roof tiles – Laconian pan tiles or curved cover tiles. The colour is

very pale brown to yellow.

PP_CBM_15: Unique type of fabric, because it contains inclusions of pure, transparent

rock crystal. A large percent of aplastic elements, around 20%. Sorting is very poor. Except of

103
rock crystal, fabric contains granodiorite and quartz. This fabric was probably used for making

of Laconian pan tiles, as the preserved pieces are around 3 cm thick and slightly curved on

upper and lower surface. Voids are scarce and rounded, the colour is reddish yellow to yellow.

The question is, if the rock crystal was added by purpose to the fabric. This is unlikely to be

true, if we assume, that it was so to make final product flashier, reflecting the sun rays on small

rock crystal cobbles, because, the pan tiles were partly hidden under the space of cover tiles. I

suppose, that the crystal rock was spread in the source of sand, which was used as additive to

the clay, and so its presence in the fabric is probably incidental.

PP_CBM_16: The last group of CBM is not actually made of clay, but often was sorted

as building material and mixed in evidence with tiles and bricks. It is mortar and its huge pieces.

For formal accuracy was also examined in the same way as previous fabrics, but of course has

different purpose in construction process. Sorting of this fabric is good, with mica and limestone

inclusions up to 10%, with very few voids. The colour is white. The mortar often forms regular

pieces with imprints of adjacent CBM elements or stone masonry.

Fig. 39) CBM fabric samples: PP_CBM_13, PP_CBM_14, PP_CBM_15, PP_CBM_16 (photo by author).

104
8.6. Shapes

Distinction between forms more difficult with fragmentized material, but even the small

pieces can be distinguished from each other and put into two groups: tiles, which are thinner,

in general up to 3 cm, but the majority around 2,2 cm, and bricks, which are thicker – over 3

cm, the majority around 3,6 cm thick, but the thickness is variable and not so uniform as in the

case of roof tiles. The question remains, how to identify the paving tiles, which are typical with

smoothened upper surface of the pavement element. It seems, that on the site Priniatikos Pyrgos

were bricks as paving elements in some parts, because these are in few cases obviously worn

out on one side.

8.6.1. Bricks

The brick forms are described in ancient sources (e.g. Vitruvius) and they remained

constant through the antiquity. They are usually referred to in their Latin names, which are

derivates of their dimensions compared to the dimension of ancient measurement system based

on Roman foot, which varies from 263 – 310 mm (Mills 2013, 32). Two distinctive types were

identified, according to the complete or nearly complete pieces and few more types are assumed

according to examined fragments.

PP_BR_01: The majority of CBM identified as bricks belongs to this shape.

Dimensions: 31-32 x 31-32 cm, thickness 3,4-3,6 cm and are very uniform, made in one fashion.

The reconstructed weight of the whole piece was approximately 5 kg. Remarkable diagonal

grooves, probably impressed by regular tool from above, while the clay was still in the mould.

Slightly thicker closer to the edge, but very regular in thickness, the difference is only 1-2 mm.

Two variants appeared in the examined material – first has the edge 3,6 cm thick. This one is

more usual, than the second variant, which is only 3,4 cm thick. The fragments are usually one

quarter segment of the whole brick, the cracks emerged always following the diagonal grooves.

105
The cracks are clear and sharp, fragments of one piece can be easily put together. There can be

more purposes for which the deep grooves were made into the brick – firstly, the mortar would

more easily penetrate the brick and join the adjacent brick, or secondly, if the partition of the

brick was needed for irregular parts of the wall or floor, the break was easily done following

the groove, where the brick was the most fragile. The former could be backed up by residues

of mortar, which were examined on few pieces. The clay for this type of CBM is well-sorted,

all the pieces appear to be made of one fabric – PP_CBM_06. It is quite possible, that this type

was produced in one highly specialized workshop, following the same procedure. All the pieces

can be classified as pedales, while the side length it’s the characteristic of this Roman shape

(Mills 2013, fig. 2.13). The type evidently follows the Roman tradition of production, in

Knossos, very similar type was identified as Severan in origin, and it is highly probable, that

also these elements from Priniatikos Pyrgos could be assigned to Roman period. As the type

was evidently manufactured in specialized production, we can assume, that present pieces were

imported to the site from area with higher need for production, which could be for example

Hierapytna, the most important city of the region in Roman times.

It seems, that these bricks were also used for purpose of floor paving, as some pieces

are remarkably smoothened by contact with shoes and feet. Affected is usually the upper side

of the brick – the one with incised diagonal cross.

Fig. 40) Trench II, context 504, 52: Drawings and section of pedalis sherds (made by author).

106
Fig. 41) Trench II, context 764: The upper view of part of pedalis brick (photo by author).

Fig. 42) Trench II, context 764: The bottom viwe of part of pedalis brick (photo by author).

PP_BR_02: The complete piece of this type was found on the site. The fabric is very

coarse, surface is marked by many small cracks caused both during firing and by natural causes.

Thickness varies from the edge 4,5 cm to the middle 3,4 cm, while the edges are slightly

elevated, while the clay was pressed to the mould. Dimensions of the sides are 28,5 x 28,5 cm.

The weight of the whole piece is 4,55 kg. Very distinctive sign is diagonal double-cross,

obviously made by pulling two fingers from corner to corner. This pulling was probably made

107
right after the pressing o clay into the mould was done, as the finger grooves overlapped the

small lines, made unintentionally by some kind of brush, while smoothening the upper surface

of the produced piece in the mould. The fabrics of this type are PP_CBM_08 and PP_CBM_09,

which are of local origin in Priniatikos Pyrgos, also the simple design of the bricks shows not-

specialized occasional production, these facts lead to conclusion that this type of CBM could

have been produced in proximity of the site. The finger grooves seem to serve for better join of

adjacent row of bricks, at the same time marking the upper surface of the brick. The shape can

be classified as pedalis tradition, but in rough design and due to pottery from the same contexts

belongs to Early Byzantine period.

Fig. 43) Trench II, context 23: The edge of brick with diagonal double finger-cross (made by author).

Fig. 44) Trench II, context 23: The edge-sherds of brick with diagonal double finger-cross (photo by author).

108
PP_BR_3: Very similar to previous piece is the brick with double finger-cross. Only

one specimen was found, which differs in dimensions from PP_BR_02 – thickness is uniform

3 cm, sides 20,5 x 20,5 cm. Rough design is the same as of previous group, the fabric is of

PP_CBM_09 type. Weight approximately 2 kg. The shape is closer to bessalis type (198 mm

on side based on Mills 2013, fig. 2.13). According to fabric and design, could be of local origin.

Fig. 45) Trench II, context 909: Sherd of smaller brick with diagonal double finger-cross (photo by author).

Many smaller fragments of other local types were discovered, with thickness from

3,6 cm till 5 cm, but shape cannot be reconstructed, as the shards are too fragmentary. They are

composed of fabrics PP_CBM_02, PP_CBM_08, PP_CBM_09 and PP_CBM_11. These bricks

seem to be of local origin as well.

8.6.2. Roof Tiles

8.6.2.1. Pan tiles

PP_PT_1: Laconian pan tile. Slightly curved on the whole surface, both sides are curved

in the same angle. The curvature is not so remarkable closer to the side edges, where the lower

surface of the tile tend to be flat or even slightly concave (C26 002-149). The side edges are

rounded in connection to upper surface. The side edges are a bit thicker than the rest of the tile.

109
The minimal thickness is 2,1 cm and the maximal (at the edge) is 3,5 cm. No complete pieces

is preserved. The largest shard is 35 cm long and 34 cm wide, the original dimensions are

assumed as 45 x 70 cm approximately. Tile is of PP_CBM_02 fabric, considered to be local. It

comes from context identified as rubble used as paving layer of Building II and so it was used

in 8th century or earlier. This kind of large pan tiles must have been very well embedded on the

roof, not only on simple wood construction, but probably laid into more flexible material (clay

or mortar), otherwise the elements of the roof would become wobbly. The shape is Hellenistic

to Late Roman due to contents of the contexts with sherds of this type of tile.

Fig. 46) Trench II, context 26: Section of Laconian pan tile (made by author).

PP_PT_02: Laconian pan tile. Common type of Laconian pan tile on Priniatikos Pyrgos.

Tiles are thinner than previous type and the curvature is regular, the edges does not differ in

thicknes from the rest of the body of the tile. The side edges are sharply cut, what is more

apparent on the lower surface of the tile. The thickness varies from piece to piece between 1,9

to 2,6 cm. This simple type of pan tile is covered by fabrics PP_CBM_01, PP_CBM_02,

PP_CBM_09, PP_CBM_11, which are local and in PP_CBM_07, whose origin is not quite

clear. The dimensions of these tiles are not known, as no complete piece was found and the

large shards are also scarce. The largest shard is 13 cm wide and 20 cm long, however estimated

width is at least 30 cm, because if thinner with the low curvature, the water could have left the

stream bed and penetrate the roof. It is likely, that some of these pan tiles were fitting with

cover tiles PP_CT_2, and so the length is estimated at least on 48 cm. Pieces of this pan tile

type are very easy to be confused with Laconian cover tiles, especially, when only small shards

are present, but there are small differences, which can help to distinguish: The edge of pan tile

110
is sharper, while of cover tiles are rounded, and second trace is the angle of curvature, which is

higher for cover tiles. The tradition of production of this tile-type is very long and cannot be

related to exact period.

Fig. 47) Trench II, sub-context 002-056: Section of Laconian pan tile (made by author).

PP_PT_03: Corinthian pan tile. This type can be subscribed to Corinthian group of

roofing systems, however is exceptional in design, as the rim exceeds the edge and continues

under the angle of 40° out of the mass of tile. This edge is also remarkably thinner than the rest

of the tile, as the body has regular thickness of 2 cm, up to the point, where the rim starts, the

thickness decreases to 1,2 cm, continues like this another 6,5 cm and at the end spreads again

to rounded edge 1,8 cm thick. Only one large shard of this type of tile was found (A5009.6).

Estimated width is 35 cm and length around 50 cm. The weight of the complete tile is 6 kg.

Parallel to this type can be found in Hellenistic contexts in Gortyn (Papadopoulos 1999, Fig

242, tile 400).

Fig. 48) Trench A5000, context A5009.6: Profile view of unusual Laconian pan tile (photo by author).

111
PP_PT_04: Corinthian pan tile. Very well fired type, which is usually connected to

fabric PP_CBM_10. The design is uniform and shards of this shape are very remarkable in

CBM on the site. The tile has three raised edges - two on sides and on the upper surface of

upper end of tile. The thickness is 2 cm in the body, the side rims are 4,4 cm and the upper end

rim is 3,5 cm thick. The space of raised side edge is flat, with sharp borders and its transition

to the body sector is moderate. The outer side of the raised edge is slightly concave inwards,

but its bottom is again sharply connected to the lower surface of the tile. The raised edge on the

upper end of the tile is 1 cm intended from the edge of the tile, on the outer side is steep and on

the inner side is sloped. The shards can be found mainly in Roman and Early Byzantine

contexts. Close analogies to this type can be traced to Roman tiles used in Knossos (Sackett

1992, Plate 23, tile C and D). And because also the quality of the tile is high, it can be subscribed

to well organised production in some larger workshop of Roman period.

Fig. 49) Trench II, context 26: Sections of Corinthian pan tile (made by author).

Fig. 50) Trench A2000, context A2005.1: Sherd of Corinthian pan tile (made by author).

112
Fig. 51) Trench II, context 26: Large sherd of Corinthian pan tile (photo by author).

PP_PT_05: Corinthian pan tile. This type is remarkable for its massive edge rim, which

is rounded at least partly. Composed mainly by fabric PP_CBM_02 of supposedly local origin.

No complete piece was found, mainly small and medium size shards. Body of the tile is 2 cm

thick and the rim at the edge up to 4,8 cm thick. The rim is 4,5 cm wide and can be partly

rounded on the outer side, but its inner border is sharp, with transition to the body also under

the sharp, close to right angle. There are two other variants of this type: The second has

approximately same disposition, but transition to the body is not sharp, but the angle is obtuse

and surface of the rim and the body are connected gradually. The third version has the most

rounded rim, connected to the body under very gentle gradient. The thickness of the rim is again

4,8 cm, but is difficult to say, where does the rim ends and the body starts. The original shape

is hard to distinguish, because, of course, it could be affected by the influence of natural powers

and erosion.

113
Fig. 52) Trench II, sub-context 002-028; context 505: ridge-sections of Corinthian pan tiles (made by author).

Fig. 53) Trench II, sub-context 002-056; context 706: The corner-sherds of Corinthian pan tile (made by

author).

Fig. 54) Trench II, context 002-056: The corner sherd of Corinthian pan tile (photo by author).

PP_PT_06: Corinthian pan tile. Tile with regular high rim, which ends with flat space

on the top. The right angle is included on outer surface of the edge and lower surface of the tile.

The rim is joined to the body with regular curvature. The body is 2-2,4 cm thick and the rim is

6 cm thick. It is the question, if some of these tiles could serve as gutter tiles (Wikander 1988,

114
214) in the last row on the roof, to drain water to the corner or other suitable part of the roof,

where could be also collected into the cistern or other vessel. For some pieces it is very likely,

because when the rim was so lifted up, it would require also very high cover tile to complete

the roof. Since no complete tile was found, I also offer explanation, that possibly only one side

edge of the tile was this high and this tile could have been used on the side border of the roof

and then no cover tile was needed to protect this edge.

Fig. 55) Trench II, context 69, 871: Ridge-sections of Corinthian pan tiles (made by author).

PP_PT_07: Corinthian pan tile. Rare sherd of the edge was found. The ridge is made

separately and attached to the body of the tile before firing. This technique is unusual46, because

cracks are likely to appear on the sticking of former separate pieces. The edge is rounded due

to production method. Thickness of the body is 2,9 cm, on the edge it is 5 cm. The fabric of this

type is PP_CBM_03.

Fig. 56) Trench II, context 509: Ridge-section of Corinthian pan tile (made by author).

46
For comparison with experimental reconstruction of tile production see e. g. Rook 1979, pp. 298-301.

115
PP_PT_08: Corinthian pan tile. Only ridge-sherds of this common type were found on

the Priniatikos Pyrgos site. The edge is carefully made with very straight borders. Supposedly

this regularity was achieved by proper use of mould. The thickness of the body is 1,9 cm and

of the edge is 4,6 cm. Occasionally variant with same dimensions appear, but with inner side

of the edge slightly curved inwards. The fabric is connected to PP_CBM_09 type, which is

most probably of local origin and thus, the tiles were manufactured in the proximity of the site.

Fig. 57) Trench II, context 39: Ridge-section of Corinthian pan tile (made by author).

PP_PT_09: Corinthian pan tile. Again only sherds of this type were found without

possibility to reconstruct the full dimensions of the tile. The transition between the body and

the edge is gentle, only slowly rising from it. The outer side of the edge is curved outwards with

the outmost point in the lower part of the edge. Inner side of the edge is cured inwards, but not

crossing the vertical axis of the edge, but rather descending to the body of the tile. Thickness is

2 – 2,2 cm in the body and 4,4 cm in the edge. Remarkable lengthways traces of surface

polishing were detected on few pieces, these are probably remnants of removing the excessive

clay. The fabric is PP_CBM_02.

Fig. 58) Trench II, sub-context 002-061, context 9: Ridge-sections of Corinthian pan tiles (made by author).

116
Fig. 59) Trench II, sub-context 002-061: Ridge-sherd of Corinthian pan tile (made by author).

PP_PT_10: Corinthian pan tile. Exceptional large sherd of coloured tile. The edge is

rectangular in section with inner side of the edge smoothly descending to the body. The outer

side and the top of the edge are flat, the edge itself is 2,8 cm wide. Thickness of the body is 2,2

cm, the edge is 4 cm thick. The upper side of the body is painted in brownish red colour, which

is now partly faded away.

Fig. 60) Trench G6000, context G6006.2: Ridge-section of Corinthian pan tile (made by author).

117
Fig. 61) Trench G6000, context G6006.2: Ridge-sherd of Corinthian pan tile (photo by author).

8.6.2.2. Cover tiles

PP_CT_1: Corinthian cover tile. One almost complete piece was found and few more

shards of another pieces. Minimal length is 67 cm (estimated 75 cm for complete tile), width is

13 cm. The upper surface has three breaks, the lower surface is regularly curved. The maximum

thickness in the topmost point of the curve is 3 cm and the minimal is 1,7 cm. The height of the

tile with arc is 6,5 cm. Weight of the preserved piece is 2,9 kg. If the estimation of complete

dimension is correct, the shards covers almost 90% of the tile and full weight is approximately

3,5 kg. Material is very well fired, which also allowed the piece to remain together until

excavations. The upper surface bears the sign made before firing. The sign resembles M/W/ε/Σ,

depending on the point of view and served probably to mark the position of the tile on the roof.

Fabric is of local origin (PP_CBM_09). This type can be subscribed to Classical to Roman

period due to pottery analyse from the same context.

118
Fig. 62) Trench A5000, context A5013.4: Section of Corinthian cover tile (made by author).

Fig. 63) Trench A5000, context A5013.4: Almost complete Corinthian cover tile (photo by author).

PP_CT_2: Laconian cover tile. Remarkable part of the tile, the width is complete 16,5

cm, minimal length 48 cm, weight 2,7 kg. Both surfaces are curved, however not very regularly.

The thickness at the top of the curve is 2,2 cm, and maximal close to the edge 2,4 cm. The

height of the tile with arc is 5,5 cm. At the complete end of upper surface are two finger grooves

parallel to the short edge. The most probably served for better fixing of the adjacent cover tile,

and so this edge is the upper edge of the tile, which was originally covered by the next row of

tiles. The fabric is local (PP_CBM_09). This type of tile has definitely very long tradition of

production, but certainly was used during Late Roman and Early Byzantine periods.

Fig. 64) Trench II, context 544: Section of Laconian cover tile (made by author).

119
Fig. 65) Trench II, context 544: Incomplete Laconian cover tile (photo by author).

PP_CT_3: (Catalogue number #09-6552) Laconian cover tile. Exceptional part of

painted tile. Piece is very thin, only 1 cm, one edge is preserved, the curving is more apparent

closer to the broke edge. The fabric is local (PP_CBM_09), but sorting is finer. The painted slip

is brownish red, colour is deeper on the upper surface at the top of the curve, lower surface and

upper surface close to the original edge is reddish. This particular tile can be related with

Hellenistic period due to pottery analyses from the same context.

Fig. 66) Trench II, context 558: Edge-section of Laconian cover tile (made by author).

8.7. Signs and other marks

Stamps and signs on CBM served for different reasons: They could mark the origin of

the tile, in the Roman world usually the military unit or private workshop, which produced the

building element. Second meaning is for the process of completion of the construction, where

the sign helps to put the element into the correct place or position. This happened often in the

120
completion of roofs, where the tiles were marked according to their planned order on the

construction. The order could have been marked with number, but more often with letter. Other

signs and patterns without actual meaning served as adjustment of the surface of CBM for better

placing into the position, or connection with next pieces, or, as with the example of type

PP_BR_01, for easier and more precise division of the bricks. Some of these patterns could also

be meant as decorative element on the brick47 or tile. Fourth group of marks can be classified

as not intentional and includes regular patterns resulted from process of creation of the CBM

element, like forming in the mould, and patterns originated by chance48.

From all above mentioned types of marks, last three are present, so far no stamped CBM

element was found. Lot of grooves on tiles are made simply using finger or fingers and due to

the width of the grooves we can assume, that they were made by man´s hand, rather than

female´s or child’s. Some marks caused by pressing the clay into the mould prove, that these

pieces were made in the wooden moulds constructed from more small boards, which were

sometimes differing in the elevation or with slight gaps between each other.

8.7.1. Signs

As mentioned before, these signs could serve for making the position of CBM element

clearer for constructor. Sign in the shape of letter was found on one piece of Corinthian cover

tile, on the upper surface. The letter M, Σ or ε is depicted, depending on the point of view. It is

6,5 cm long and 3,3 cm high, and was carved into the surface of the tile with sharp, 2-3 cm wide

tool. The carving took place before firing. Belongs to the context A 5013.4, presumably

emerging during Early Byzantine period. If we accept this presumption, the letter is in Greek

47
It is a question, if the double cross pattern, used on the upper surface of Byzantine bricks, was commonly
employed for decorative purpose.
48
For example animal footprints emerging on tiles, during drying up on the sun in open areas (Sackett 1992,
Plate 223, picture 45).

121
alphabet, as this was in use in Crete at that time (Harrison 2003). The fabric is PP_CBM_09, of

local origin.

Fig. 67) Trench II, context 544: Sign on the side of Corinthian cover tile (drawing and photo by author).

Another example of intentionally made sign with meaning is triangular carving in the

flat piece of CBM, probably of flat Corinthian tile, but due to the fragmentary state of shard, it

is difficult to reconstruct the original shape of the element. The carving takes the shape of

triangle with sides 4,8 x 4,5 x 3,1 cm long. The function and origin of the sign is more

problematic here, as the edges of the carving are not raised by excess clay pushed away by

carving tool, the sign could have been made before or after firing and even after the original

use of CBM element – in secondary use. In context C 17, where the shard was found, there are

more ceramic spolia, used as filling in later wall49.

49
Another spolia used in this context is e. g. the rim of archaic – classical pithoi, found in situ with large amount
of mortar attached to it.

122
Fig. 68) Trench II, sub-context 002-055: CBM with triangular mark (photo by author).

8.7.2. Finger-marks and grooves

Some signs and patterns mainly on tiles are made without tools, simply by fingers.

Special group form bricks of pedales type, which are supposedly always signed with diagonal

cross, composed by two finger grooves, and this cross is distinctive for this type, firmly related

to the shape. Double-cross bricks are however described earlier in this work and here we will

focus on other more irregular ornaments and patterns. Except of double-cross brick, the finger

grooves are displayed on roof tiles, forming usually straight lines of one, two or three grooves

or occasionally other symbols.

8.7.3. Finger grooves on cover tiles

Functional purpose have grooves close to the edge of the upper surface of the Laconian

cover tile. Two parallel grooves, both approximately 1,8 cm wide served for better connection

with adjacent tile, which overlapped this end of the tile and was joined to it with mortar. The

width of the grooves matches with average width of man´s fingers, supposedly index and

middle finger. Only two examples of this kind of technical improvement were found in the

CBM of Priniatikos Pyrgos and we can assume then, that it was not the common practice. The

example belongs to the context C 544.

123
Another piece with use of this technique comes from C 706, where only corner of the

Laconian cover tile is preserved. This tile was provided with three finger grooves, again parallel

with each other and parallel to the upper end of the tile

Fig. 69) Trench II, context 544, 706: Laconian cover tiles with grooves (photo and drawing by author).

8.7.4. Finger grooves on pan tiles

On pan tiles use of these patterns has rather marking than functional purpose. There are

occasionally two short grooves in the corner on the upper surface of Laconian pan tile. Reason

for this sign remains unclear, because they are parallel to the sides of the tile, not to the

upper/lower end. The meaning of the signs could, again, have connection to the position of the

CBM element on the roof. The width of the grooves is around 1,8 cm, and their length around

7,5 cm. Due to regular dimensions of this pattern it is obvious, that the mark had some meaning

and didn´t appear on the pan tiles by chance. Important information is also, that this type of

double groove can be found only on upper surface of Laconian pan tiles. Grooves are carved 3-

4 mm deep.

124
Fig. 70) Trench II, context 13, 723: Laconian pan tiles with grooves (photo by author).

One examined piece of pan tile contains single finger-groove, which is longer than

10 cm and the same width as usual. The purpose was probably the same as in the case of

Laconian pan tiles with double groove in the corner.

Fig. 71) Trench II, sub-context 002-049: Laconian pan tile with groove (made by author).

Exception in the finger-groove patterns is sign in the shape of irregular cross in the

corner of upper surface of Laconian pan tile. The width of the groove is 1,3 – 1,5 cm. The sign

is very lightly carved into the body of the tile, about 1-2 mm. This exceptional tile come from

C 26 context and is typical representative of PP_PT_02 type.

125
Fig. 72) Trench II, context 26: Laconian pan tile with cross mark (photo and drawing by author).

Diagonal finger grooves are on the upper surface of both – Corinthian and Laconian pan

tiles. The width of the groove varies between 1,5-1,9 cm. One or two grooves appear on the

material.

Fig. 73) Trench II, context 851, 16: Tiles with diagonal grooves (photo and drawing by author).

Few other pieces of CBM with remarkable signs and marks show no regularity and can

be then labelled as others. Shard belonging most probably to category of Laconian pan tiles

contains deep hole, obviously made with finger. The hole is approximately 1,8 cm wide, which

correlates with width of finger grooves. Still this example is rare on the site and can be assigned

to accidental origin or at least not related to production and construction techniques. Another

126
unique case of uncommon mark is curved groove on the body-shard from tile (most probably

of Corinthian pan tile) with curved groove 12 cm long, but continuing behind the borders of the

shard. The width of the groove is 1,3 cm and as the curvature of the line is regularly made, it

could have originally form more complicated motif, yet because the only fragment of this tile

is preserved, other assumptions would be pure guessing.

Fig. 74) Trench II, context 590, 89: CBM with irregular marks (photo by author).

8.8. Hypothetical reconstruction of the roof settings

Even though it is not quite problematic task it is possible to make estimations concerning

the dimensions of the roof settings. Another issue we meet, when trying so, is the set of pan and

cover tile. From the examined material it is clear, that as the main cover tile on the site, curved

imbrex (or Laconian cover tile) was used, and the Corinthian cover tiles are in minority in the

collection of finds. It seems, that Corinthian cover tiles preserved are rather from later contexts,

supposedly Early Byzantine or even later, because Romans preferred pragmatic “Hybrid”

roofing system combining flat pan tiles with raised edge (tegulae) and simple curved cover tile

(imbrex). The amount of Laconian pan tile-sherds is difficult to measure, because fragments are

usually small, and thus undistinguishable from fragments of imbrices, but if we divide the

amount of tiling material from the site to two groups: flat tiles and curved tiles, then the curved

127
tiles would form about two thirds of the unit and flat tiles about one third. Roughly counting

from such statistic, the amount of Laconian pan tiles must had formed substantial part of the

collection, because the number of imbrices must be equal to the number of tegulae50. The

estimations for dimension of the tiles and the weight of the roof can be made separately for

each type of tile examined, with contribution of analogical studies on CBM, and using general

knowledge about antique (understand Roman) construction techniques. The estimations are

based on the data gathered from pan tiles, because these were supposedly closely placed to each

other on the roof construction, and cover tiles served only to protect the connection between

pan tiles51. For counting the proportion between the weight of pan tile and cover tile, I use the

index proposed by Marcus Rautman (2003, 53-55), suggesting, that the weight of the cover tile

can by in round numbers counted as 25% of the pan tile weight52.

1. Corinthian pan tile, context 5013.4, fabrics PP_CBM_09, shape PP_PT_04. Dimensions

of the sherd max. 56 x 33 cm, weight 5,1 kg. Reconstruction of the shape according to

similar finds from Knossos (Sackett 407-409). Reconstructed dimensions 53 x 75 cm,

weight approximately 15 kg. Weight for 1 sq. meter 37,75 kg plus 25% for cover tile

weight makes final estimation approximatelly 47 kg/1 sq.meter, using 2,5 pan tile +

cover tile sets.

50
However these conclusions from the statistics provide only supplementary, not precise information, because
the statistics do not consider possibility of differences in preservation of pan tiles and cover tiles. The former or
the latter could be reused under different circumstances, and so the final collection, which was excavated,
most probably does not correspond with the real proportion of material used for roofing.
51
In other words: The pan tiles were fully covering the surface of the roof, and thus the surface they cover is
equal to the surface of the roof, and cover tiles were only supplementary covering of their connections
(Ratuman 2003, pp. 53-55).
52
This calculation Marcus Rautman used for enumerating the hypothetical roof weight of the three Early
Byzantine churches at the Kalavassos-Kopetra site.

128
Fig. 75) Trench 5000, context 5013.4: Large part of Corinthian pan tile (photo by author).

2. Corinthian pan tile, context 5009.6, fabric PP_CBM_11, shape PP_PT_03. Dimensions

of the sherd max. 27 x 16 cm, weight 1,5 kg. Reconstruction of the shape according to

the similar find from Gortyn (Gortina IV hell 269), but due to the size of the lower edge,

which is well preserved, the tile is definitely narrower then the example from Grotyn53.

Reconstructed dimensions 34 x 50 cm, weight approximately 6,4 kg. Weight for 1 sq.

meter 37,5 kg plus 25% for cover tile makes 47 kg/square meter, using almost 6 pan tile

+ cover tile sets.

Fig. 76) Trench A5000, context A5009.6: Profile view of unusual Laconian pan tile (photo by author).

3. Laconian pan tile, context C16, sub-context 002-105, fabric PP_CBM_09, shape

PP_PT_02. Dimensions of the sherd max. 28 x 24 cm, weight 1,5 kg. Reconstruction of

53
Tile from Gortyn is 53 x 65,5 cm large (Papadopoulos 1999, pp. 269).

129
the shape is quite problematic, because Roman Laconian tiles found in Knossos have

dimension up to 42 x 110 cm (Sackett 407-409), but on the other hand, examples from

Kalavassos-Kopetra show smaller dimension 33 x 70 cm (Rautman 2003, 177). Close

to the average between these two extremes are tiles from Nichoria (Coulson 1983, 384-

387) measuring 43 x 80 cm, and these are taken as example for reconstructing the full

shape. Reconstructed dimensions 40 x 80 cm, weight approximately 7,2 kg. Weight for

1 sq. meter 22,5 kg plus 25% for cover tile makes 28,2 kg/square meter, using about 3,2

pan tile + cover tile sets. For Laconian tiling system, assumption of cover tile being only

25% of the pan tile is quite low. Taking in account examples from Nichoria, the volume

of material in cover tile is around 50% of the volume of pan tile. If we change the

calculation regarding this fact and assuming the weight of cover tile to be 50% of pan

tile, the number for square meter of the roof would change to 33,75 kg, which I find

more likely to be correct.

Fig. 77) Trench II, context 16: Large part of Laconian pan tile (photo by author).

From the rough numbers it is clear, that the weight of the whole roof must had been

remarkable, and in calculation is not included the weight of the mortar binding the tiles or

providing the layer for embedding (Brodribb 1987, 11-12). Moreover the roof was

130
overlapping the ground plan of the building and if there were some decorative elements

placed on the roof, the final weight was exceeding. It seems, that the weight was not so

much affected by the size of the tiles, and thus the architect must had always count with

immense pressure to the construction. So far no nail holes were detected on the tiles from

Priniatikos Pyrgos, and thus the tiles could be supposed to hold at their place only by their

own weight, eventually secured with mortar, which can still be found on the tiles. The

sufficient slope of the roof for tiles to withstand is 20° degrees, 30° for churches covered

with vault (Rook 1979, 295).

131
9. Conclusion

As the aim of the thesis can be divided into two lines, the final part also focuses on them

separately. At first the examination of CBM from Priniatikos Pyrgos site, held in previous

chapter is summarized and appropriate conclusions are drawn considering the site as the central

binding point of these. In the second more general section, the broader assumptions are

pronounced with focus on CBM as uniting factor for interpreting the past. Because the CBM in

archaeological environment is still not fully investigated source of data about the past and the

sources of archaeology in general terms provide us only with limited information about the

economy and society, upcoming conclusions fully acknowledge these limitations, but at the

same time are eager to touch the field of theoretical hypothesis based on interpretations of

studied material.

9.1. CBM in local economy

The site of Priniatikos Pyrgos offers various possibilities to study CBM material,

because it was populated almost continuously through the prehistory and history (Molloy and

Duckworth (eds.) 2014, Hayden and Tsipopoulou 2012) and the character of CBM is such that

it survives, even though in terrible condition, for long time, at least from periods of baked CBM.

But even when the shape cannot be read anymore, the fabrics can still provide useful

information.

The fabrics of studied material were only examined macroscopically, with hep of

magnifying glass, but preliminary report can be already made, as the inclusions are identifiable.

For area of the site, the numerous pottery studies helped to distinguish local and other

inclusions, based on geological arrangement of the region. As local inclusions were identified:

quartz, limestone, granodiorite, biotite (golden mica), siltstone (Nodarou, Moody 2014). If we

132
consider this assumption, the majority of CBM fabrics could be of local origin, but of course,

this information is not definite, because the same clay could be collected elsewhere on the spot

with similar geological disposition. Statistically about 90% of studied CBM material has

inclusions of local origin54. Remarkable fact is that fabric group PP_CBM_07 with eminent

content of phyllite inclusions, is certainly not from the local area in origin55. The question

occurs then, whether the finished building elements were imported, or rather the raw rocks.

Following the example from Kallavasos-Kopetra site (Rautman 2003, 355) I suggest that the

complete CBM elements were transported, because the imports were needed in first place, it

was most probably because local economy was not able to secure its own sufficient production

of suitable material in the proximity of the site, and so it was more convenient to acquire the

complete CBM from larger more specialized workshop, than to establish new manufacture

requiring wide economic relations in obtaining necessary resources and man-power.

Macroscopic examination of fabrics give us schematic information about production of material

and quarrying of clay, but for further understanding of economic processes, the petrographic

thin section samples should lead to place the PP CBM into the wider context within the

boundaries of the island.

Variety of CBM shapes found in excavated trenches can help us to indicate what

construction techniques were used at the site, however for full image larger excavated space is

required to make the architectural outlines clear. So far the appearance of the site cannot be

reconstructed as well as disposition of CBM, however it provides some hints of it. Also the

chronological order of CBM elements is problematic. So far, we may assume the presence of

Corinthian and Laconian roofs at the site, or in its proximity and few types of bricks. The

Corinthian tiles suggest flat roof, probably sitting on timber structure. The Laconian tiles could

54
CBM groups PP_CBM_01, PP_CBM_02, PP_CBM_09, PP_CBM_11.
55
In area of Mochlos and area of Kentri – Anatoli are sources of phyllite exploited already in ancient times.

133
indicate both the presence of flat roof, or roof with rounded surface, possibly apse, some of the

tiles with distinctive remnants of mortar on the bottom part which can mean mortar or clay

structure beneath. For Corinthian roof defining feature is Corinthian cover tile, and for Laconian

on the contrary the Laconian pan tile, so the presence of both is proved for sure. The existence

of Hybrid roof (“Sicilian system”), the major system used in Roman period, cannot be proved

or denied, because the system contains Corinthian pan tiles and Laconian cover tiles. But as the

Roman architectural influence is strong on the site in the form of Roman-style brick tiles,

probably also the Hybrid roofing system was employed. Tiles appear as very useful material as

we can see on the examples of secondary56 use - their life cycle continued, and often were re-

used more than once, even after the loss of original shape, they could be still used for floor

levelling or to fill small gaps in walls, also for reconstructions. Tiles are variable in function

and didn´t have their firm position within the structure, but their status differed even inside the

group of one shape – the brick could had been used for wall construction, as well as for flooring

the space. Tiles could be useful part of the roof as well as of the wall, or covering of the grave57.

Bricks discovered in the deposit layers are of three types, first with very uniform look and

octagonal grooves, which can be subscribed to tradition of Roman production as they appear in

vast numbers in architecture of 2nd century AD Gortyn (Montali 2006, 103-116) and can be

described as pedalis type with grooves enabling easier division of tiles into triangular shape for

walls construction with concrete core. Some of the pedalis bricks from PP are truly divided, but

their primary use for walls construction is questionable, as no concrete-core wall58 was

excavated. Some of the bricks are remarkably worn out on one side, and are not canonically

divided into triangles, but rather irregularly, probably unintentionally due to post-deposition

56
But for example for using the tile-sherds in wall construction, the employment could be tertiary or even
further in the life-cycle of the tile.
57
But in this case, it is tempting to suppose the symbolic function, some “eternal house” of the deceased.
58
Mortar was definitely used to bind stones together in the wall structure, but no trace of true opus
caementicium was discovered.

134
processes, and so it seems, that at least some of them were used for floor construction and were

worn out by exposition to traffic. Some of them also carry traces of mortar on both sides and

thus were employed as part of wall structures. Other two brick types are of very rough

appearance with many small cracks. Their dimensions show, at least for PP_BR_02 tradition

of one feet long bricks, but are supposedly from not specialized workshops, and eventually

could be made by local workers in smaller or even household production. The purpose and

explanation of double finger-groove running in diagonals of the brick is task for further work,

because it obviously didn´t serve for functional purpose as did the deep groove in pedalis brick

of Roman production. So far no CBM with stamps was found on the site, but it is highly

probable that it will be in the future, which would prove the wider economic relations of the

region, because supposedly stamps were added only in larger productions run by entrepreneurs

with purpose of income, or in workshops owned by government. Other various marks on tiles

can be identified as tally marks and other notes made by workers to clear the placement of the

element on the roof, or could serve as grooves for mortar application, which adhered better to

these. Even when the CBM lost its shape and couldn´t be used for its primary purpose anymore,

it still could be employed as debris for floor levelling, which happened in Byzantine Building

1 on PP site, particularly the context C26 in Trench II is this floor mass (Klontza-Jaklová 2014,

138-139), where the material was rammed down. Hypothetical roof-reconstructions should help

us to imagine the appearance of the site, but also determine demands on supporting construction

carrying the immense weight of the roof. Some fabrics and shapes can be certainly related with

particular periods, but presumably there was no strict evolution of CBM at the site.

Further work with CBM on the Priniatikos Pyrgos site should contain the material from

so far unexplored areas, and with contribution of new data about architecture of the site it can

provide firmer reconstructions considering the amounts of CBM required, and thus produced

here or transported from elsewhere. The petrographic thin section research, which could be

135
compared to other sites, should take place, and then the outlines of economic relations in the

region sketched in this study could be concluded into more complex image.

9.2. CBM - an archaeological source for interpreting the past

With help of case studies and overview of theoretical architectural handbooks, in this

thesis I offered complete example, how ceramic building material can serve us as tool for

revealing past human activities in archaeological environment. Past publications on CBM topic

focused mainly on understanding the production59, or some distinctive features, like stamps60.

Complex study of CBM from Beirut and Carthage brought Philip Mills in his monography

(Mills 2013), where he also sketches the importance of maritime trade and transport of CBM

even on large distance. His work served in many respects as leading line for this study. For

research in typology of CBM especially for Roman period I would emphasize the summarizing

research of british material carried out by Gerald Brodribb (1987). The two main handbooks,

first about antique architecture (Wright 2005) and second focused on Roman architecture

(Adam 2005) provided basis for chapters about building techniques between other studies.

Interesting remarks were drawn from Wikander´s article about antique roof tiles

(Wikander 1988). Byzantine architectural environment was covered by works of Robert

Ousterhout (2008) and Richard Krautheimer (1992). Many other works, which influenced this

study are stated in bibliographical part.

Philip Mills in his work proved (Mills 2013, 1-3; Mills 2013, 116) , that the “consumer

city” model offered by Finley (1973), which sees the city as the centre of consumption with

basic commodities produced locally and only luxurious good transported on large distances, is

not necessarily true for CBM, which can be seen as ordinary goods, but still could be transported

59
See for example McWhirr 1979.
60
See for example Kurzmann 2006 for Roman Empire, or Bardill 2004 for Byzantine environment.

136
on inter-regional dimension. Even though Finley assumes, that the CBM production was seen

in the eyes of Romans rather as part of agricultural activities (Finley 1973, 58), then as

profitable entrepreneurship, or even better as honourable way of acquiring in comparison to

other non-agricultural businesses, the situation was probably much more complicated especially

in the rural areas (such as Priniatikos Pyrgos in Late Antiquity is), where the society had

basically two options: to acquire high-quality material from specialized workshop and pay for

it appropriate price, or manufacture their own CBM with lower quality, but with corresponding

save up. This is rather black and white point of view, because works out mainly for small

economical subjects, which had only occasional need for CBM, and the factual diversity was

probably larger, as Gerald Brodribb assumes (Brodribb 1987, 139), when he offers variety of

production models from household production, through the local brickyards established to

supply local needs; seasonal workshops related to periods with minimal requirements for

farming; workshops centred in area of convenient clay-source; workshops related to private

estates; and finally established by municipality, government or military unit.

Except for creating economic models, the CBM offers tool for reconstructing general

architecture on the archaeological site, because certain CBM types are strongly connected to

certain techniques, as we can see on examples from Roman architectural tradition. The presence

of triangular bricks would suggest former wall structure with concrete core and brick facing.

Voussoirs would indicate that the vault was part of the structure. Circular or rounded bricks can

identify the bath building in vicinity, or presence of suspensura in general. Tiles will help to

reconstruct the image of the roof and eventually requirements for supporting structure.

Considering of course the possibility of other employment of CBM than its original purpose,

secondary and substitutive function.

This study was carried out as complex examination of ceramic building material from

site Priniatikos Pyrgos in East Crete, with chapters introducing into general problematics of

137
CBM as comprehensive phenomenon, which can be met in archaeological projects, and as

structural element of architecture of past societies, their settlements and economies. The CBM

as a source of information offers potential, which is still not adequately acknowledged and can

bring new perspectives in global recognition of building materials. Hopefully this thesis will

serve as a source of particular data and inspiration in the means of methodology and approach

for future studies carried out on the field of CBM.

138
10. Resumé

Základní data pro tuto studii poskytl soubor stavební keramiky získaný archeologickými

výzkumy v rámci „Priniatikos Pyrgos Project“, zkoumající lokalitu Priniatikos Pyrgos v zátoce

Mirabello na východní Krétě. Tímto bych chtěl také vyjádřit své poděkování účastníkům

projektu, a především jeho vedoucím pracovníkům: Dr. Barry Molloy, Dr. Joanna Day, Dr. Sue

Bridgeford, a hlavní díky zaslouží Dr. Věra Klontza-Jaklová, která mne do projektu zapojila.

Keramický stavební materiál byl v minulosti opomíjeným zdrojem archeologickým dat

o antické kultuře, ekonomice a společnosti, který se s rozvojem technických metod při

zkoumání archeologického materiálu získává své místo ve vědeckých studiích.

Cílem této práce je poskytnout širší souvislosti v možnostech studia keramického

stavebního materiálu z prostředí antické civilizace, formou typologických rozborů materiálu

a následné komparace s informacemi, které zprostředkovali antičtí autoři, jako například

Vitruvius a jeho deset svazků „De architektura“. Další úhel pohledu přináší tradiční rozbory

antické architektury se zaměřením na stavební prvky, jako na součást celkové výpovědní

hodnoty stavitelství. Třetím úhlem pohledu je pak archeologický přístup ke keramickému

stavebnímu materiálu, který na něj pohlíží jako na součást nálezových celků, které poskytují

doplňkové informace nejen o pozici jednotlivých prvků v původních strukturách, ale také

o jeho následném využití a archeologizaci v kontextu jiných druhů artefaktů umožňujících jeho

funkční, časové i prostorové umístění. Cílem práce je zjistit, jaký je informační potenciál

keramického stavebního materiálu, respektive na jaký druh otázek týkajících se antických

společností nám může tento materiál pomoci nalézt odpovědi, a to jak v obecném kontextu, tak

v konkrétním případě lokality Priniatikos Pyrgos.

139
Časově je tato práce vymezena pozdně helénistickým obdobím a obdobím pozdní

antiky. Prostorově se studie zaměřuje na ostrov Krétu, a oblast Egejského moře, se

srovnávacími paralelami z celého území Římské říše, která v rámci romanizace šířila jednotící

tendence v architektuře a tedy i produkci stavebního materiálu.

Keramický stavební materiál je definován, jako „umělý produkt (keramický) lidské

činnosti, opakující se formy, jehož jednotky jsou určeny ke stavební konstrukci, a jehož

jednotky je možné umístit do pozice pomocí rukou jedné osoby“61, a který prošel procesem

výpalu. Historicky se první tašky a cihly dostaly do Středomoří z oblasti Mezopotámie, kde

figurovaly, jako hlavní stavební materiál. V Řecku byly potom tašky užívány od doby časně

helladské, jak dokazují nálezy z Lerny (Wikander 1988, 204). V Archaickém období pak došlo

k ustálení střešních systémů, které byly nadále užívány skrze zkoumané období. Cihly začaly

být využívány později, a v plné míře se uplatnily až v římské stavební tradici, kde došlo

k boomu cihlových staveb v 1. st. n. l.

Římské stavitelství se zakládalo na pevné typologii keramického stavebního materiálu

založené na římské délkové míře - stopě, přičemž drobné rozdíly v rozměrech stavebních prvků

jsou často dané geografickou rozlohou Impéria a lokálními stavebními tradicemi. Mezi

nejpoužívanější typy patří cihly bessalis, pedalis a bipedalis. Pro střešní konstrukce byly

užívány tašky odvozené od korintského a lakónského typu v řecké architektuře, v římském

prostředí známé jako tegula a imbrex. Pro stavby zvláštního účelu byly užívány keramické

prvky speciální konstrukce, jako klínové cihly pro klenby kulaté sloupové cihly, tašky

s ventilací nebo otvory propouštějícími světlo, obkládací dlaždice a cihly určené pro konstrukci

systému vytápění místnosti, vodovodní potrubí nebo keramické trubky pro odlehčenou

konstrukci klenby.

61
V originále “an artificial product made in replicate units for building construction, each unit capable
of being put into position by hand or hands of one person,” (Harley 1974, 63 in: Mills 2013, 3).

140
Techniky konstrukce a typy struktur, pro které byl využíván keramický stavební

materiál ukazují jeho široké využití v různých prostředích. Velký úspěch zaznamenala římská

technika kombinující opus caementicium a ztracené bednění tvořené často právě cihlami (opus

testaceum) nebo kombinací cihel a kamene (opus mixtum). V této technice byly užívány

trojúhelníkové cihly, vyrobené lámáním čtvercových cihel ve vyznačených drážkách. Užitím

trojúhelníkových cihel bylo šetřeno jejich množství a cement je lépe zafixoval. Zdi tvořené

výhradně cihlami byly stavěny jen ve výjimečných případech. Kombinace kamenných

a cihlových vrstev se stala velice oblíbenou při stavbě Konstantinpole a později hrála

významnou roli v Byzantské architektuře, kde tato technika fungovala zároveň jako dekorační

prvek staveb. Počet cihlových řad v jedné vrstvě a vztah tloušťky cihly a cementové vrstvy je

v jistých případech chronologicky citlivý faktor, který může datovat stavbu budovy (Dodge

1987, 112-113). Cihly se staly výhodným materiálem pro konstrukci oblouků a kleneb, kde

bylo užíváno jak cihel standardní velikosti, tak speciálních dutých a odlehčených cihel které

s pomocí lešení utvořily velmi pevné struktury. V oblasti severní Afriky byly v římském období

oblíbeny keramické trubky, které umožnily stavbu klenby bez podpůrného lešení, jelikož držely

relativně stabilní pozici zasazeny jedna do druhé. V mnoha případech jsou cihly užity také jako

podlahová krytina v luxusnějších vilách komponované do vzoru „obilného klasu“ (opus

spicatum). Velké uplatnění našla stavební keramika při konstrukci střech, kdy byla pokládána

na dřevěné krovy nebo, jako v případě klenutých střech, do vrstvy cementu nebo jílu přímo nad

stropní konstrukcí.

Pro lepší orientaci v problematice keramického stavebního materiálu v archeologickém

prostředí bylo vybráno několik případových studií z výzkumných projektů v Egejské oblasti,

konkrétně: Gortyna, Knossos, a Oxa na ostrově Kréta; Kalavasos-Kopetra na Kypru; a Nichoria

na Peloponésu. Lokality byly vybrány vzhledem k charakteru lokality Priniatikos Pyrgos, tedy

– Nichoria a Kalavasos-Kopetra pro srovnání s nálezy v podobném rurálním prostředí v období

141
přelomu římského a raně byzantského období; Knossos a Gortyna pro srovnání materiálu se

stavební keramikou z předpokládaných velkých produkčních center římského a byzantského

období na Krétě; Oxa pro srovnání materiálu s blízkou byzantskou lokalitou.

Samotná studie stavebního keramického materiálu z lokality Priniatikos Pyrgos se opírá

o pozorování složení keramického těsta a tvarové variability materiálu. Bylo rozlišeno deset

skupin složení hmoty, z nichž 4 skupiny tvořící asi 90% stavební keramiky na lokalitě má

předpokládaný lokální původ, určený na základě příměsí hornin v keramickém těstu. Tašky

s příměsí velkých šupin fylitu, který nemá v regionu žádný zjištěný zdroj, byly s největší

pravděpodobností importovány jako hotové zboží z většího produkčního centra v raně

byzantském období. Z hlediska tvaru byly zjištěny tři typy cihel: robustní cihly typu pedales

římské výrobní tradice, produkovány ve větším výrobním centru a na lokalitu dopravovány;

a dva ojedinělejší druhy cihel hrubšího provedení, pravděpodobně vyráběny v blízkosti

lokality. Identifikováno bylo deset typů „spodních“ tašek62 - dvě tzv. lakónské a osm tzv.

korintských; a dále tři typy „krycích“ tašek63 - jedna korintská a dvě lakónské. Některé z typů

vykazují přímou korelaci se zjištěnými typy keramického těsta, jako například typ korintské

„spodní“ tašky, který se vyznačuje vysokým stupněm výpalu a inkluzemi vápencových zrnek.

Bylo zaznamenáno několik typů značek otisků na povrchu studovaných keramických

prvků. Jejich význam lze chápat jako značky určující pozici prvku v konstrukci (značka ve tvaru

„M“), rýhy usnadňující přilnutí cementu ve spoji s dalšími částmi konstrukce (prstové rýhy na

okrajích lakónských spodních tašek), doklady výrobních postupů (otlaky dřevěných forem

a uhlazování povrchu nástrojem) nebo čistě symbolicky (dvojitý kříž na povrchu hrubých cihel;

trojúhelníková značka vyrytá do povrchu tašky). Rekonstrukce celkových rozměrů tašek

62
Anglicky „pan tile“.
63
Anglicky „cover tile“.

142
napovídá, že při užití různých typů střešních tašek se celková váha střechy mohla mírně lišit,

ale v každém případě vyžadovala velmi pevnou podpěrnou konstrukci.

V konkrétním případě archeologické lokality Priniatikos Pyrgos přináší studium

keramického stavebního materiálu poznatky o konstrukčních technikách a tedy celkové podobě

budov na lokalitě v různých periodách antiky. Ačkoliv chronologická citlivost stavební

keramiky je nízká, lze sledovat určité změny a odlišit lokální produkci od importovaného

materiálu, což vede k prohloubení znalostí o ekonomických vazbách v regionu. Další bádání

by mělo vést ke studiu stavebního keramického materiálu z dosud nezkoumaných částí lokality,

jeho srovnání s popsanými stavebními prvky a zasazení do celkového kontextu regionu.

Z obecného hlediska přináší keramický stavební materiál nový pohled na hospodářské

vztahy v antickém světě, případně míru regionální samostatnosti sídlišť venkovského

charakteru. V římském světě byla výroba stavební keramiky řazena spíše k zemědělství než

k výnosným podnikatelským činnostem (Finley 1973, 58), a tedy byla nahlížena jako morálně

nepoškozující svého provozovatele. Jak je ovšem patrné z archeologického materiálu, obchod

se stavební keramikou probíhal i na větší zeměpisné vzdálenosti a otázka výnosnosti je tedy

nasnadě. Ve vztahu archeologického materiálu k lokalitě pak může stavební keramika pomoci

utvořit obraz užitých konstrukčních technik a celkové architektonické podoby sídel, a vztahu

obyvatel k prostředí na příkladu sekundárního využívání stavební keramiky.

Na závěr, lze dodat, že keramický stavební materiál má velký informační potenciál

vzhledem k interpretaci minulých lidských společností, a pro globální obraz jeho využití je

třeba jej pečlivě studovat v rámci archeologických projektů.

143
11. Bibliography

• Adam, J.-P. 1999. Roman Building. Materials and Techniques. London; New York.

• Bahník, V. et al. 1974. Slovník antické kultury, Praha.

• Bardill, J. 2004. Brickstamps of Constantinople,"Oxford.

• Bardill, J. 2008: „Building Materials and Techniques“, in: The Oxford Handbook of Byzantine

Studies, eds. E. Jeffreys, J. Haldon, and R. Cormack, Oxford; New York, pp. 335-352.

• Betancourt, P.P. 2014. „Priniatikos Pyrgos in 1912: The Last Foreign Archaeological

Excavation in Independent Crete,“ in: A Cretan Landscape through Time: Priniatikos Pyrgos

and Environs, (British Archaeological Reports International Series 2634), eds. B. Molloy, C.

Duckworth, Oxford, pp. 8-14.

• Bintliff, J. 2012. The complete Archaeology of Greece: From Hunter-Gatherers to 20th

century AD, Oxford.

• Brodribb, G. 1987. Roman Brick and Tile, Gloucester.

• Coulson, W.D.E.,W.A. McDonald, J. Rossner, eds. 1983. Excavations at Nichoria in

Southwest Greece III: Dark Age and Byzantine Occupation, Minneapolis.

• Darwill, T., A. McWhirr, 1984. „Brick and Tile Production in Roman Britain: Models of

Economic Organisation,“ World Archaeology 15, pp. 239-261.

• Dodge, H. 1987. „Brick Construction in Roman Greece and Asia Minor,“ in: Roman

Architecture in the Greek World, eds. S. McReady, F.H. Thompson, London. pp. 106-116.

144
• Erickson, B.L. 2010. „Priniatikos Pyrgos and the Classical Period in Eastern Crete: Feasting

and Island Identities,“ Hesperia: The Journal of the American School of Classical Studies at

Athens 79, pp. 305-349.

• Filippaki, E., Y. Bassiakos, B.J. Hayden, 2014: „Metalworking at Priniatikos Pyrgos,

Mirabello Gulf, Crete,“ in: A Cretan Landscape through Time: Priniatikos Pyrgos and

Environs, (British Archaeological Reports International Series 2634), eds. B. Molloy, C.

Duckworth, Oxford, pp. 105-112.

• Finley, M. 1973. The Ancient Economy, London.

• Gallimore, S.C. 2011. „An Island Economy: Ierapetra and Crete in the Roman Empire,“ (Diss.

State University of New York, Buffalo).

• Geisler, A. 2015: „Fortification system of the Byzantine site of Oxa (Crete),“ (master thesis

Masaryk University, Brno).

• Gerstel, S.E.J. 2009: The Nikomedia Workshop: New Evidence on Byzantine Tiles, The

Journal of the Walters Art Museum, Vol. 66/67 (2008-2009), pp. 5-53.

• Gortina IV = M. Ricciardi, N. Allegro, eds., Le fortificazioni di etá ellenistica (Gortina IV,

Monografie della Scuola Archeologica Italiana di Atene X), Padua 1999.

• Gortina V = A. Di Vita, ed., Lo scavo del pretorio (1989 –1995) (Gortina V, Monografie della

Scuola Archeologica Italiana di Atene XII) Padua 2000.

• Grigoropoulos, D. et al. 2008. „Baths in Roman and Late Antique Chersonissos: Preliminary

Remarks on their Topography, Architecture and Building History,“ Creta Antica 9, pp. 303-

318.

145
• Hallager, E. and B.P. Hallager, eds. 1997. The Greek Swedish Excavations at Agia Aikaterini

Square Kastelli, Khania 1970-1987 1: From Geometric to the Modern Greek Period (Acta

Instituti Atheniensis Regni Sueciae Series 47), Stockholm.

• Harrison, G.W.H. and J. Francis, 2003. „Gortyn: First City of Roman Crete,“ American

Journal of Archaeology 107, pp. 487-492.

• Hayden,B. J. 2003. Reports on the Vrokastro Area, Eastern Crete 3: The Vrokastro Regional

Survey Project: Sites and Pottery (University Museum Monograph 123), Philadelphia.

• Hayden, B.J. and M. Tsipopoulou, 2012. „The Priniatikos Pyrgos project. Preliminary report

on the rescue excavation of 2005–2006,“ Hesperia: The Journal of the American School of

Classical Studies at Athens 81, pp. 506-84.

• Hayden, B. 2014. „Priniatikos Pyrgos and its Territory: Results of Survey and Excavation,“ in:

A Cretan Landscape through Time: Priniatikos Pyrgos and Environs, (British Archaeological

Reports International Series 2634), eds. B. Molloy, C. Duckworth, Oxford, pp. 15-21.

• Hayes, J. W. 1972. Late Roman Pottery, London.

• Helen, T. 1975. Organisation of Roman Brick Production in the First and Second Centuries

AD: an Interpretation of Roman Brick Stamps (Annales Academiae Scientiarum Fennicae:

Dissertationes Humanarum Litterarum 5), Helsinki.

• Higgins, M. and R. Higgins. 1996. A Geological Companion to Greece and Aegean, London.

• Higgins, R. A. [1954] 1969. Catalogue of the Terracottas in the Department of Greek and

Roman Antiquities, British Museum 1, repr. Oxford.

• Hodges, R. 2004. Byzantine Butrint : excavations and surveys 1994-1999, Oxford.

• Jeffreys, E., J. Haldon, R. Cormack eds. 2008. The Oxford Handbook of Byzantine Studies,

Oxford; New York.

146
• Jeremic, G. 2009. Saldvm, Roman and early Byzantine fortification. Belgrade.

• Johnston, D. and D. Williams, 1979. „Relief Patterned Tiles – A Reappraisal,“ in: Roman

Brick and Tile. Studies in Manufacture, Distribution and Use in the Western Empire, ed.

A. McWhirr, London, pp. 375-394.

• Klontza, V. 2014. „The Byzantine Sequences at Priniatikos Pyrgos: Preliminary Observations

on Ceramic Chronology and Architectural Phasing,“ in: A Cretan Landscape through Time:

Priniatikos Pyrgos and Environs, (British Archaeological Reports International Series 2634),

eds. B. Molloy, C. Duckworth, Oxford, pp. 135-142.

• Klontza-Jaklova, V. 2014(b). „Transport and storage pottery from Priniatikos Pyrgos - Crete:

a preliminary study,“ in: LRCW 4. Late Roman Coarse Wares, Cooking Wares and Amphorae

in the Mediterranean Archaeology and archaeometry The Mediterranean: a market without

frontiers (British Archaeological Reports International Series 2616), eds. N. Poulou-

Papademetriou, E. Nodarou, V. Kilikoglou, Oxford, pp. 799-810.

• Klontza-Jaklová, V. 2015. „History hidden in broken pots or broken pots hidden in history?

The Late Roman – Early Byzantine stratigraphy at Priniatikos Pyrgos (Crete),“ Studia

archaeologica Brunensia 20, pp. 139-164.

• Klontza-Jaklová, V. et all., in press. „Oxa – Kalos Lakkos region (Crete): results of

preliminary survey (2013 – 2014),“ Památky archeologické 106.

• Krautheimer, R. 1986. Early Christian and Byzantine Architecture. New Haven.

• Lancaster, L. 2009. „Roman Engineering and Construction,“ in: The Oxford Handbook of

Engineering and Technology in the Classical World, ed. J.P. Oleson, Oxford, pp. 257-284.

• Makrakis, M. 2011. Elounda, Agios Nikolaos, Spinalonga: their history, Athens.

• Mango, C. 1950. „Byzantine Brick Stamps,“ American Journal of Archaeology 54, pp. 19-27.

147
• Mango, C. ed. 2002. The Oxford history of Byzantium, Oxford.

• Mattingly, D. and J. Salmon eds. 2001. Economies Beyond Agriculture in the Ancient World,

London; New York.

• McReady, S. and F.H. Thompson, eds. 1987. Roman Architecture in the Greek World,

London.

• McWhirr, A. (ed.) 1979. Roman Brick and Tile. Studies in Manufacture, Distribution and Use

in the Western Empire (British Archaeological Reports International Series 68), Oxford.

• Mills, P. J. E. 2013. The Ancient Mediterranean Trade in Ceramic Building Materials: A Case

Study in Carthage and Beirut (Roman and Late Antique Mediterranean Pottery 2), Oxford.

• Molloy, B. and C. Duckworth eds. 2014. A Cretan Landscape through Time: Priniatikos

Pyrgos and Environs (British Archaeological Reports International Series 2634), Oxford.

• Montali, G. 2004. „Il grande teatro di Gortina,“ in: Creta Romana e Protobizantina. Atti del

Convegno Internazionale, Iraklion, 23–30 settembre 2000 (Scuola Archeologica Italiana di

Atene III/1), Torino – Padova, pp. 709-724.

• Montali, G. 2006. Il teatro romano di Gortina (Studi di Archeologia Cretese IV), Padua.

• Morgan, G. 1979. „Experiments in Making and Firing Box-Flue Tiles,“ in: Roman Brick and

Tile. Studies in Manufacture, Distribution and Use in the Western Empire, ed. A. McWhirr,

London, pp. 375-394.

• Munsell Color Company. 2009: Munsell Soil Color Charts: With Genuine Munsell Color

Chips.

• Nodarou, E. and J. Moody 2014. „Mirabello Fabric(s) Forever: An Analytical Study of the

Granodiorote Pottery of the Vrokastro Area from the Final Neolithic Period to Modern

Times,“ in: A Cretan Landscape through Time: Priniatikos Pyrgos and Environs, (British

148
Archaeological Reports International Series 2634), eds. B. Molloy, C. Duckworth, Oxford, pp.

91-98.

• Oleson, J.P., ed. 2009. The Oxford Handbook of Engineering and Technology in the Classical

World, Oxford.

• Ousterhout, R. 1999. Master Builders of Byzantium, Princeton; New Jersey.

• Papadopoulos, J. 1999. „Laterizi,“ in: Gortina IV = M. Ricciardi, N. Allegro, eds., Le

fortificazioni di etá ellenistica (Gortina IV, Monografie della Scuola Archeologica Italiana di

Atene X), Padua 1999, pp. 254-269.

• Peacock, D. P. S. 1979. „An Etnoarchaeological Approach to the Study of Roman Bricks and

Tiles,“ in: Roman Brick and Tile. Studies in Manufacture, Distribution and Use in the Western

Empire, ed. A. McWhirr, London, pp. 5-10.

• Peña, J.T. 2007. Roman Pottery in the Archaeological Record, Cambridge; New.

• Philips Stevens G. 1950. „A Tile Standard in the Agora of Ancient Athens,“ Hesperia: The

Journal of the American School of Classical Studies at Athens 19, pp. 174-188.

• Poulou-Papademetriou, E. Nodarou, V. Kilikoglou, eds. 2014. LRCW 4. LRCW 4. Late

Roman Coarse Wares, Cooking Wares and Amphorae in the Mediterranean Archaeology and

archaeometry The Mediterranean: a market without frontiers (British Archaeological Reports

International Series 2616), Oxford.

• Poulter, A.D. 1995. Nicopolis ad Istrum, a Roman, Late Roman, and early Byzantine city:

excavations 1985-1992, London.

• Rautman, M. L. et all. 1999: Amphoras and Roof-Tiles from Late Roman Cyprus: A

Compositional Study of Calcareous Ceramics from Kalavasos-Kopetra, Journal of Roman

Archaeology, vol. 12, pp. 377-391.

149
• Rautman, M. L. 2003. A Cypriot Village of Late Antiquity. Kalavasos-Kopetra in the Vasilikos

Valley (Journal of Roman Archaeology Supplementary Series 52), Portsmouth, Rhode Island.

• Rizzo, M.A. 2000. „Laterizi,“ in: Gortina V = A. Di Vita, ed., Lo scavo del pretorio (1989 –

1995) (Gortina V, Monografie della Scuola Archeologica Italiana di Atene XII) Padua, pp.

562-566.

• Sackett, L.H. 1992. Knossos: from Greek City to Roman Colony. Excavations at the

Unexplored Mansion II, 2 vols. Athens; London.

• Sanders, I.F. 1982. Roman Crete: an Archaeological Survey and Gazetteer of Late Hellenistic,

Roman ans Early Byzantine Crete, Warminster, Wilts.

• Saradi, H. 1997. „The Use of Ancient Spolia in Byzantine Monuments: The Archaeological

and Literary Evidence,“ International Journal of the Classical Tradition 3, pp. 395-423.

• Sweetman, R. J. 2004. „Late Antique Knossos. Understanding the city: Evidence of mosaics

and religious architecture,“ Annual of British School at Athens 99, pp. 315–354.

• Taylor, R. 2003. Roman Builders. A Study in Architectural Process, Cambridge.

• Tsougarakis, D. 1988. Byzantine Crete. From the 5th Century to the Venetian Conquest

(Historical Monographs 4), Athens.

• Rook, T. 1979. „Tiled Roofs,“ in: Roman Brick and Tile. Studies in Manufacture, Distribution

and Use in the Western Empire, ed. A. McWhirr, London, pp. 295-302.

• Rostoker, W. and E. Gebhard, 1981. „The Reproduction of Rooftiles for the Archaic Temple

of Poseidon at Isthmia, Greece,“ Journal of Field Archaeology 8, pp. 211-227.

• Vroom, J. 2005. Byzantine to Modern Pottery in the Aegean. An Introduction and Field

Guide, Utrecht.

150
• Waelkens, M. 1987. „The Adoption of Roman Building Techniques in the Architecture of

Asia Minor. in: Roman Architecture in the Greek World, eds. S. McReady, F.H. Thompson,

London, pp. 94-105.

• Ward-Perkins, J.B. 1981. Roman Imperial Architecture, Harmondsworth; New York.

• Ward-Perkins, J.B. 1994. Studies in Roman and Early Christian Architecture, London.

• Webster, G. 1979. „Tiles as a Structural Component in Buildings,“ in: Roman Brick and Tile.

Studies in Manufacture, Distribution and Use in the Western Empire, ed. A. McWhirr,

London, pp. 285-294.

• Wikander, O. 1988. „Ancient Roof Tile – Use and Function,“ Opuscula Atheniensia 17, pp.

203-216.

• Williams, Ch.K. 1994. „Roof Tiles from Two Circular Buildings at Corinth,“ Hesperia

Supplements 27 (Proceedings of the International Conference on Greek Architectural

Terracottas of the Classical and Hellenistic Periods, December 12-15, 1991), pp. 53-60.

• Wright, G.R.H. 2005. Ancient Building Technology, Volume 2: Materials, Leiden; Boston.

• Young, Ch. 1979. „The Processing of Roman Tile,“ in: Roman Brick and Tile. Studies in

Manufacture, Distribution and Use in the Western Empire, ed. A. McWhirr, London, pp. 401-

404.

11.1. Online Sources

• Hicky Morgan, M. 1914: Vitruvius. The Ten Books on Architecture. Cambridge. Available

on: https://en.wikisource.org/wiki/Ten_Books_on_Architecture. Date of searching:

23. 9.2015.

• Perseus Digital Library. Available on: http://www.perseus.tufts.edu. Date of searching: 20. 4.

2014.

151
• Third Eye Revelation: Brick Kiln. Available on:

http://thirdeyerevelation.blogspot.cz/2010/05/brick-kiln.html. Date of searching: 15. 11. 2015.

• Swedish Institute at Athens. Available on: http://www.sia.gr. Date of searchinng: 2. 9. 2015.

12. List of abbreviations

g gram

kg kilogram

cm centimetres

m metres

no. number

fig. figure

cc. circa

152
13. Illustrations

13.1. List of Plates

Plate 1) Chronological overview for this study. .................................................................................... 11


Plate 2) The list of fabric groups and their characteristics (abbreviations: Incl. – inclusions, sphere. –
sphericity, c. – core, s. – surface). ......................................................................................................... 97

13.2. List of Figures

Fig. 1) Laconian system (from Åkeström 1966, Abb. 64 in Wikander 1988, Fig. 6). .......................... 17
Fig. 2) Corinthian system (from Åkeström 1966, Abb. 64 in Wikander 1988, Fig. 6). ........................ 17
Fig. 3) Hybrid (Sicilian) system (from Åkeström 1966, Abb. 64 in Wikander 1988, Fig. 6). .............. 18
Fig. 4) Restoration of “Tile Standard” from Athenian Agora (from Philips Stevens 1950, pic. 10). ... 20
Fig. 5) Roman brick kiln (from Adam 1999, pic. 140). ........................................................................ 25
Fig. 6) Brick firing in piles in contemporary India (from
http://thirdeyerevelation.blogspot.cz/2010/05/brick-kiln.html). ........................................................... 28
Fig. 7) Overview of basic Roman brick typology (from Mills 2013, fig. 2.13). .................................. 30
Fig. 8) Examples of “armchair” voussoirs for vault construction found in Britain (from Brodribb
1987, fig. 19). ........................................................................................................................................ 37
Fig. 9) Example of “tubulus cuneatus” (from Bridribb 1987, fig. 35). ................................................. 38
Fig. 10) Sections of brick columns with strong layer of plaster (from Wright 2005, pic. 208). ........... 40
Fig. 11) Gutter tile (from Wiegand & Schrader [N. 53] Abb. 331 in: Wikander 1988, fig. 7). ............ 45
Fig. 12) Chiney pot (from Brodribb 1987, fig. 15). .............................................................................. 46
Fig. 13) Roof tiles with ventilation device (after Durm B d R, fig. 364 in: Wright 2005, fig. 186). .... 46
Fig. 14) Tegulae mammatae and section of the wall showing their employment (from Wright 2005,
pic. 191)................................................................................................................................................. 49
Fig. 15) Sections of vaulting tubes (Mills 2013, fig. 2.14). .................................................................. 51
Fig. 16) Water pipes.1. Hellenistic;2.-5. Roman(from Wright 2005,pic. 196). .................................... 52
Fig. 17) The division of bricks into triangles (from Adam 1999, fig. 347). .......................................... 56
Fig. 18) Example of wall with concrete core faced with triangular bricks (from Adam 1999, fig. 348).
............................................................................................................................................................... 58
Fig. 19) Example of opus mixtum, Pompeii around 300 AD (from Adam 1999, fig. 337). ................. 61
Fig. 20) Scheme of disposition of weight on the arch (from Adam 1999, fig. 395). ............................ 64
Fig. 21) Terminology of arch elements (from Adam 1999, fig. 401).................................................... 64
Fig. 22) Example of opus spicatum flooring (from Adam 1999, fig. 541). .......................................... 67
Fig. 23) The triangulated truss: 1 tie beam; 2 principal rafter; 3 king post; 4 wall-plate; 5 purlin; 6
ridge purlin; 7 cleat; 8 common rafter; 9 boarding (laths) (from Adam 1999, fig. 491). ...................... 69
Fig. 24) CBM from Kalavasos – Kopetra: Laconian tiles 1, 2; Corinthian tiles 3, 4, 5, 6; water pipe
(from Rautman 2005, fig. 5.18). ............................................................................................................ 75
Fig. 25) Examples of CBM from Knossos: Laconian pan tiles 7, Corinthian pan tile 2, Corinthian
cover tiles 1 (from Sackett 1992, Plate 22-23). ..................................................................................... 79
Fig. 26) Laconian pan tile 1; Laconian cover tile 5 (from Coulson 1983, fig. 10-77, 10-81). .............. 82
Fig. 27) Examples of CBM from Gortyn: Corinthian pan tile 3; section of irregular Corinthian pan tile
(from Papadopoulos 1999, fig. 227, 242). ............................................................................................. 84
Fig. 28) Example of pedalis brick 4 with mark from Gortyn (from Montali 2006, TAV. VIII). .......... 84
Fig. 29) Examples of tubuli from Gortyn (from Rizzo 2000, fig. 208). ................................................ 85
Fig. 30) Fragment of Laconian pan tile 1 with decoration from Oxa (photo by author). ..................... 86
Fig. 31) Fragment of pedalis brick 2 (photo by author). ....................................................................... 86

153
Fig. 32) The view on Mirabello bay in east Crete with Priniatikos Pyrgos site at the small peninsula
protruding from the sandy shore. (photo by author).............................................................................. 88
Fig. 33) Priniatikos Pyrgos, site plan with highlighted Trench II, the main source of studied CBM.
(Courtesy of Priniatikos Pyrgos Project). .............................................................................................. 92
Fig. 34) Priniatikos Pyrgos, map of architecture in the Trench II. (Courtesy of Priniatikos Pyrgos
Project). ................................................................................................................................................. 93
Fig. 35) Context 706. Amphora sherd secondary used in wall construction, still covered with mortar
(photo by author). .................................................................................................................................. 94
Fig. 36) CBM fabric samples: PP_CBM_01, PP_CBM_02, PP_CBM_03, PP_CBM_04 (photo by
author). .................................................................................................................................................. 99
Fig. 37) CBM fabric samples: PP_CBM_05, PP_CBM_06, PP_CBM_07, PP_CBM_08 (photo by
author). ................................................................................................................................................ 101
Fig. 38) CBM fabric samples: PP_CBM_09, PP_CBM_10, PP_CBM_11, PP_CBM_12 (photo by
author). ................................................................................................................................................ 103
Fig. 39) CBM fabric samples: PP_CBM_13, PP_CBM_14, PP_CBM_15, PP_CBM_16 (photo by
author). ................................................................................................................................................ 104
Fig. 40) Trench II, context 504, 52: Drawings and section of pedalis sherds (made by author). ........ 106
Fig. 41) Trench II, context 764: The upper part of pedalis brick (photo by author). .......................... 107
Fig. 42) Trench II, context 764: The bottom part of pedalis brick (photo by author). ........................ 107
Fig. 43) Trench II, context 23: The edge of brick with diagonal double finger-cross (made by author).
............................................................................................................................................................. 108
Fig. 44) Trench II, context 23: The edge-sherds of brick with diagonal double finger-cross (photo by
author). ................................................................................................................................................ 108
Fig. 45) Trench II, context 909: Sherd of smaller brick with diagonal double finger-cross (photo by
author). ................................................................................................................................................ 109
Fig. 46) Trench II, context 26: Section of Laconian pan tile (made by author). ................................. 110
Fig. 47) Trench II, sub-context 002-056: Section of Laconian pan tile (made by author). ................. 111
Fig. 48) Trench A5000, context A5009.6: Profile view of unusual Laconian pan tile (photo by author).
............................................................................................................................................................. 111
Fig. 49) Trench II, context 26: Sections of Corinthian pan tile (made by author). ............................. 112
Fig. 50) Trench A2000, context A2005.1: Sherd of Corinthian pan tile (made by author). ............... 112
Fig. 51) Trench II, context 26: Large sherd of Corinthian pan tile (photo by author). ....................... 113
Fig. 52) Trench II, sub-context 002-028; context 505: ridge-sections of Corinthian pan tiles (made by
author). ................................................................................................................................................ 114
Fig. 53) Trench II, sub-context 002-056; context 706: The corner-sherds of Corinthian pan tile (made
by author). ........................................................................................................................................... 114
Fig. 54) Trench II, context 002-056: The corner sherd of Corinthian pan tile (photo by author). ...... 114
Fig. 55) Trench II, context 69, 871: Ridge-sections of Corinthian pan tiles (made by author). .......... 115
Fig. 56) Trench II, context 509: Ridge-section of Corinthian pan tile (made by author). ................... 115
Fig. 57) Trench II, context 39: Ridge-section of Corinthian pan tile (made by author). ..................... 116
Fig. 58) Trench II, sub-context 002-061, context 9: Ridge-sections of Corinthian pan tiles (made by
author). ................................................................................................................................................ 116
Fig. 59) Trench II, sub-context 002-061: Ridge-sherd of Corinthian pan tile (made by author). ....... 117
Fig. 60) Trench G6000, context G6006.2: Ridge-section of Corinthian pan tile (made by author). ... 117
Fig. 61) Trench G6000, context G6006.2: Ridge-sherd of Corinthian pan tile (photo by author). ..... 118
Fig. 62) Trench A5000, context A5013.4: Section of Corinthian cover tile (made by author). .......... 119
Fig. 63) Trench A5000, context A5013.4: Almost complete Corinthian cover tile (photo by author).
............................................................................................................................................................. 119
Fig. 64) Trench II, context 544: Section of Laconian cover tile (made by author). ............................ 119
Fig. 65) Trench II, context 544: Incomplete Laconian cover tile (photo by author). .......................... 120
Fig. 66) Trench II, context 558: Edge-section of Laconian cover tile (made by author). ................... 120
Fig. 67) Trench II, context 544: Sign on the side of Corinthian cover tile (drawing and photo by
author). ................................................................................................................................................ 122
Fig. 68) Trench II, sub-context 002-055: CBM with triangular mark (photo by author). ................... 123

154
Fig. 69) Trench II, context 544, 706: Laconian cover tiles with grooves (photo and drawing by author).
............................................................................................................................................................. 124
Fig. 70) Trench II, context 13, 723: Laconian pan tiles with grooves (photo by author). ................... 125
Fig. 71) Trench II, sub-context 002-049: Laconian pan tile with groove (made by author). .............. 125
Fig. 72) Trench II, context 26: Laconian pan tile with cross mark (photo and drawing by author). ... 126
Fig. 73) Trench II, context 851, 16: Tiles with diagonal grooves (photo and drawing by author). .... 126
Fig. 74) Trench II, context 590, 89: CBM with irregular marks (photo by author). ........................... 127
Fig. 75) Trench 5000, context 5013.4: Large part of Corinthian pan tile (photo by author). .............. 129
Fig. 76) Trench A5000, context A5009.6: Profile view of unusual Laconian pan tile (photo by author).
............................................................................................................................................................. 129
Fig. 77) Trench II, context 16: Large part of Laconian pan tile (photo by author). ............................ 130

155

You might also like