CRUSTUMERIUM: A GEO-ARCHAEOLOGICAL APPROACH TOWARDS THE DEVELOPMENT AND ABANDONMENT OF THE SETTLEMENT AREA FROM IRON AGE TO ROMAN TIMES AND AFTER

(1)

1 Nikolaas Noorda (s1816853)

DEPARTMENT OF ARCHAEOLOGY, University of Groningen

CRUSTUMERIUM:

A GEO-ARCHAEOLOGICAL APPROACH TOWARDS THE

DEVELOPMENT AND ABANDONMENT OF THE SETTLEMENT

(2)

(3)

1. Introduction

In this thesis I will discuss the results of a

geoarchaeological analysis based on the samples, results and interpretations of research that was conducted as part of the International Crustumerium Project. This project is a collaboration of multiple institutions including the Soprintendenza Speciale per il Colosseo, il Museo Nazionale Romano e l’Area

Archeologica di Roma (SS-Col) and the Groningen Institute of Archaeology (GIA). The project has a strong focus on the funerary record of the site. I joined the GIA excavations at Crustumerium in 2012 and from 2013 onwards I became involved in the

geoarchaeological aspects of the project and the settlement area of Crustumerium in its entirety. Since then I have participated in many of the campaigns that have led to the geoarchaeological dataset used in this thesis; The dataset is comprised of data gathered during large-scale geophysical surveys, systematic corings, and soil samples from stratigraphical contexts, which were excavated during campaigns in 2013, 2014 and 2015. These surveys, corings and excavations were part of a clear fieldwork strategy to establish the erosion and sedimentation patterns on and around the site on a landscape scale, and to understand the

development of local settlement strategies. During these campaigns I worked closely with professor Peter Attema from the GIA and emeritus professor Jan Sevink from the University of Amsterdam, and when it was decided to do this interdisciplinary analyses as part of my thesis it was with the help of professor Sevink that I came to work at the Institute for Biodiversity and Ecosystem Dynamics (IBED), where I was able to conduct these analyses under the guidance of lab technician Leo Hoitinga.

1.1. The site and the project

The site of Crustumerium is situated on one of the tuff plateaus 16 km north of modern day Rome alongside the Northern Tiber valley. (map 1) It emerged as early as the 10th_{or 9}th_{century B.C. - the}

earliest archaeological evidence of habitation of the area dates to the Latial period IIB21 - and developed

into a flourishing settlement between the second half of the 8th_{and the 6}th_{century B.C. During this}

period the settlement extended considerably, reaching up to ± 63 ha, and became one of the largest centres in the region. It gained control over a part of the Via Salaria and the route that led from Etruria

1_{Willemsen, 2014.}

(6)

2 to Campania.2_{The control over these routes enabled close contact with other settlements in the}

Etruscan, Faliscan and Sabine regions, as well as with the southern part of Latium Vetus.3_Crustumerium

is referred to in historical sources4_{, specifically in relation with its conquest by the city of Rome in 499}

B.C. After Rome took control of the area, it was completely abandoned by its original inhabitants and turned into a military outpost that shielded the city from attacks from Rome’s enemies. After the defeat of Veii in 396 B.C. Crustumerium lost its strategic position. Finally, at the beginning of the first century A.D. Crustumerium was added to the municipium of Fidenae.5

Archaeological research at Crustumerium started between 1974-1976 with an extensive artefact survey by Quilici & Quilici-Gigli6_{that covered the entire territory of the ancient city, including the burial}

grounds. The latter surround the settlement area on the north, northeast, southeast and the westside, and some of the burial grounds have been in use from the late 9th_{century B.C. onwards.}7_{The large}

quantity of tombs8_{offers a good insight into the changing social make-up of the community during the}

rise of the settlement. The tombs show clear developments in their architecture and spatial

configuration that match changes in funerary gifts.9_{The different sepulchral areas are named after the}

toponyms of their location; respectively Campo Grande, Sasso Bianco, Monte Del Bufalo, Cisterna Grande and Marcigliano.10_{(map 2) The first excavation of the burial grounds started in 1987 when the}

SS-Col excavated several tombs at Sasso Bianco and Monte Del Bufalo. The latter became the main research area for the team of the GIA, when they joined the project in 2006.11

The Monte Del Bufalo burial area is situated south of the settlement area and it covers a small hill that slopes downwards towards the Fossa della Formicolo in the south.12_{During the survey of the Quilici’s in}

the 80s of the last century (funerary) material dating to the Orientalizing and Archaic period confirmed the use of the area as a burial ground. The Monte Del Bufalo burial ground can be subdivided into two subareas; respectively the Southern Area and the Fossato Area.13_{The Fossato Area, which is the area}

where the GIA gathered most of its funerary data. It is the part of the site from where the data used in this research was collected. The area is delimited by the defensive fossato14_{, which forms the boundary}

between the settlement and the burial ground, in the North and a deeplying road that lays just north of

2_{Attema & Nijboer, 2010; Willemsen, 2014.} 3_{Willemsen, 2014.}

4_{Livy II, 19; Attema & Nijboer, 2010; Willemsen, 2014.} 5_{Attema & Nijboer, 2010; Willemsen, 2014.}

6_{Quilici & Quilici-Gigli, 1980.} 7_{Willemsen 2014, p.1}

8_{The settlement is one of the few known sites in Latium Vetus that, up to now, has many tombs dating to the 6}th

century B.C. (Attema & Nijboer, 2010)

9_{Attema & Nijboer, 2010.} 10_{Willemsen 2014, p.1}

11_{Attema & Nijboer, 2010; Willemsen, 2014.} 12_{Willemsen 2014, p.16.}

13_{Willemsen 2014, pp. 7/9-10.}

14_{The fossato is part of a defensive system called ‘agger and fossa’ and, as present in for example in Ardea,}

(7)

3 the modern dirt road.15_{The tombs in the Fossato Area date generally from the middle of the 7}th_century

B.C. in Latium16_.

In 2010 the GIA initiated a new project, funded by NWO17_{, “The people and the state. Material culture,}

social structure and political centralization in Central Italy (800-450 BC).” with Crustumerium as case-study. The aims of this new project were to establish “in what ways did early state formation in central Italy, with as its main protagonist the city of Rome, affect the social fabric of community life”, and determine “how did central Italian habitants of minor urban centres respond to the changing socio-political situation due to socio-political centralization”.18_{The project, which was divided into two subprojects,}

provided the financial opportunity to conduct large scale research, such as the geophysical survey, corings, and stratigraphical research as part of subproject 1 that was dedicated to investigations of the settlement and territory of Crustumerium. 19_{This thesis contributes also to this subproject.}

15_{Noorda & Attema, 2013.}

16_{Willemsen, 2014, p.10.}

17_{NWO is the Netherlands Organisation for a Scientific Research; a national research organisation with an active}

contribution to various parts of the national science and innovation policies. Its main task is to fund scientific research at public research institutions in the Netherlands, especially universities. (www.nwo.nl).

18_{Attema & Nijboer, 2010.}

19_{The focus of the other subproject was on the funerary aspect and covered the cemeteries of Crustumerium in}

order to conduct a social and artefact based analysis. (Attema & Nijboer, 2010)

(8)

4

1.2. Aims and theory behind this research

As will be made clear in chapter three, in recent years the focus of new research at Crustumerium shifted more and more from research of tombs towards the settlement area, and with this shift came an

increasing need for a more interdisciplinary approach. Especially knowledge of geology and soils became much more important in order to get a better understanding of what happened to the settlement since its abandonment, especially when it became apparent that large parts of the settlement had endured a high amount of erosion. First of all, it became important to establish the geomorphic conditions of the site, as these not only form and characterize a landscape, but also determine the stability and

preservation of the landscape.20_{To try to reconstruct the ancient environment and the processes active}

during the habitation of the settlement became a primary aim in the GIA’s research. Also the changes in environment and processes active after the abandonment of the settlement were deemed of great importance, as these changes can have the ability to change the archaeological record as well.21

Furthermore, modern processes and their spatial distribution can also provide equivalents for understanding past associations.22_{The nature, the completeness and the interpretation of the}

archaeological record parallels and depends on the nature and the completeness of the geological and geomorphological sequence and the understanding of the natural context of the archaeological context.23_{This asks for a broad approach, something which became clear during the excavation of the}

trial trenches dug over the past years, and this realization instigated the analyses conducted for this thesis. These processes are often divided into two types of processes: Non-cultural processes, or in other words natural processes, and cultural processes.

1.2.1. Non-cultural (i.e. natural) processes

The non-cultural processes encompass all the natural events that govern both the burial and survival of the archaeological context24_{; for instance, processes that lead to sedimentation over and burial of}

material, such as colluviation or fluviation, but also processes that lead to erosion and have the

destruction of the archaeological context as a result. These types of processes can be caused for instance by local changes in water management, type of soil, or seismic activities. On a large scale, environment and climate are important variables that should not be overlooked, as they not only have an effect on the formation and movement of soils and sediments, but also influence human behaviour. In both ways they have an effect on the archaeological context, for instance via movement of soils, but also via degradation of soil, which could subsequently cause the degradation of living conditions and ultimately abandonment.25_{Changes in the use of a place or site through time may indeed be a result of larger}

changes in the landscape, caused, for instance, by changes in the environment.26_{The changes in the}

environment have effect on different scales and factors range from site-specific locations, a scale where archaeologists often operate on, to much broader, zonal factors that affect entire regions. Depending in

20_{Schiffer 1983, pp. 691-692; Waters & Kuehn 1996, pp.485/495.}

21_{Wood & Johnson 1978, p. 317; Waters & Kuehn 1996, pp. 483-484; Barton et al. 2002, p.176.} 22_{Gladfelter 1977, p.519.}

23_{Gladfelter 1977, p. 519; Schiffer 1983, p. 691; Waters & Kuehn 1996, p. 484; Barton et al. 2002, p.176.} 24_{Schiffer, 1983.}

(9)

5 part on the period, but also on the degree of environmental change, data collected on these different scales can differ from the modern environment.27_{Pressing once more the importance of a more holistic}

approach; one cannot assume that the modern environment and the processes active in this environment are similar to or representative of the environment and processes in the past and vice versa. The varying distributions and effects of surface processes through time and space account for the dynamic nature of landforms. This has the result that the surface of the earth can represent a composite of past and ongoing processes (a polygenetic form), or it reflects a simpler equilibrium between

contemporary processes and its form.28_{So in order to reconstruct the ancient environment and the}

processes active during the habitation of a settlement it is important to retrace the development and changes made on a site-scale, such as landscape changes, as well as on a larger, regional scale, such as climate changes. Even though it is not always possible in smaller research projects to investigate both scales, it should not be ignored that without both scales a representative reconstruction is not possible.

27_{Gladfelter 1977, p.518.}

28_{Gladfelter 1977, p. 519. Furthermore, attempts to explain regional change often disregard intraregional}

variability, while landscapes are rarely environmentally uniform. (Lillios 1993, p. 10/117; Nelson & Hegmon 1993, p. 214)

29_{There are some conflicting hypotheses about whether the agricultural activities in Crustumerium took place}

inside the settlement area or not. It has long been believed that the entire settlement area was built upon and that therefore agriculture would have taken place outside the settlement boundaries; this was based mainly on survey data from the western side of the settlement and some pottery finds from Quilici O, whereby the results from the western side were extrapolated to the eastern side (which is the focus area of this thesis). However, according to Jorn Seubers MA and myself this hypothesis is unlikely and is not corroborated by recent results produced by the surveys, excavations and the results from this thesis. Our hypothesis is that habitation in Crustumerium was most likely focussed on the western side of the settlement (as supported by the survey data) and that the eastern side could very well be used as an area for agricultural activities. (Jorn Seubers will discuss this further in his

dissertation). This is supported by the marginal number of archaeological remains found in the soils on the eastern slopes during the excavations; the number was much lower than would be expected from an urban area. (Noorda & Attema, 2013) Furthermore, the amount of colluvial soil deposited in the lower parts of the eastern side is much higher than would be expected from erosion of an urban area. This implies that the eastern part remained used as an agricultural area (probably horticulture). This would explain why the number of archaeological remains in the colluvial sediments in trench Saggio 2 and 4 is so low (see chapter 3.1.).

Table 1 The main anthropogenic developments that caused changes to the landscape in Crustumerium in order of time.

Period Development

Iron Age Reclamation of the plateau in favour of habitation

and small-scale agriculture.

Orientalizing period Intensification of the habitation on the plateau and intensification of the agriculture on and around the plateau.29

Archaic period

5th_{century B.C. onwards} _{Abandonment of the settlement, and gradual}

development of the Roman countryside and introduction of Roman agricultural methods (plough cycles).

20th_{century B.C. onwards} _{Mechanisation of agriculture, especially during}

(10)

6

1.2.2. Cultural (i.e. anthropogenic) processes

Another important factor that will be addressed in this thesis is the impact of human behaviour on all processes and all scales, in size as well as through time. The environment, which plays a large role in the formation of soils and sediments and the behaviour of humans living in it, is itself also influenced by human action. Environmental changes, together with taphonomic processes, are often caused or at least enhanced by the actions of humans. For instance, erosion as a result of past agricultural endeavours, such as ploughing or the cutting of trees, can significantly influence an area.30_{Besides a direct visual}

change to the landscape, it can also change how the landscape reacts to the weather, and which post-depositional processes will occur. Semi-arid areas show an increase in erosion during the first

years/decades after the abandonment; the recovery of semi-natural vegetation in the area is limited due to water and seed dispersal and results in a low vegetation cover which makes it vulnerable to erosion.31

The increase in erosion due to intensive (deep) ploughing can also lead to a barren landscape, which then will require an even more intensive human input to be usable for agriculture, causing a continuous spiral of soil depletion, landscape change and eventually environmental change. This leads to the second question whether we can determine to what extent human interaction has (passively) influenced the processes and sedimentation of the soils and how this has changed the landscape. Table 1 shows the main anthropogenic developments during and after the habitation of Crustumerium that must have had an effect on the landscape. In chapter two the anticipated effects based on these developments are discussed. From a more archaeological standpoint it is important to determine whether any of the sediments and fillings are the direct result of human endeavour. Of course, all sedimentation on a site, since the start of habitation, will be influenced by human activities, especially with the introduction of agriculture. However, massive fillings that are the direct result of human activity narrate a different story and can be used in the research of another process influencing the settlement and the archaeological context, namely abandonment.

1.2.3. Abandonment theory

Abandonment, in the colloquial meaning of the word, was often explained as an ‘act of giving up’, expressed by leaving or ceasing to operate a specific thing or area. The focus of the research into abandonment was therefore generally on either gradual regional abandonment (slow and large) or on spectacular events of rapid abandonment at specific sites, like Pompeii (fast and small).32_{In recent years}

the focus and interpretation of abandonment changed towards the view of abandonments as a part of normal processes active in settlements. Although abandonment can be a result of a large incident, like a volcano eruption, physical abandonment often begins on the smallest levels and occurs first as local processes, which can eventually evolve into regional depopulation.33_{With that in mind, current research}

focusses more on the agents of change and continuity and the process of abandonment as a choice, based on the role of individuals acting within a range of natural and social contexts.34_{This new focus}

went hand in hand with the identification of so-called site formation processes, such as the erosion

30_{Sluiter & De Jong 2006, p.561.} 31_{Lesschen et al., 2008.}

32_{Cameron 1993, p. 3; Nelson & Schachner 2002, p. 168; Colwell-Chanthaphonh & Ferguson 2006, pp. 37/40.} 33_{Schlanger & Wilshusen 1993, p. 85.}

(11)

7 processes discussed above, and the role these processes had on abandonment and vice versa.35

Formation processes are described by Schiffer36_{as the factors, both cultural or non-cultural, that create}

and control the existence of the archaeological context that is studied by archaeologists. However, these processes vary between areas, materials and environments, and have different effects in different regions and on different materials.37_{What the situation was in Crustumerium is still unclear, with as only}

certainty that the settlement was abandoned in the 5th_{century B.C. and taken over by the Romans.}

Crustumerium is thought to have been inhabited continuously since its origin until the abandonment. However, in my opinion, recent results that will be discussed in this thesis suggest abandonment of (at least) certain parts of the settlement during its existence. Especially the presence of colluvial sediments in the deeplying road system, which was excavated in 2013, beg the question which processes were active at the site and how people dealt with them.

1.3. Main research questions and methodology

This thesis aims at a broad interdisciplinary study into establishing which cultural and non-cultural processes were active during and after the habitation of Crustumerium. By using results from archaeological, geophysical and soil research it aims to answer several research questions:

- What were the main post-depositional processes active in the area of Crustumerium and what was the effect of these processes on the landscape of Crustumerium.

- Can we determine to what extent human interaction was responsible for the post-depositional processes and sedimentation of the soils and how these processes have changed the landscape.

- Can the analyses of the soils and reconstruction of the post-depositional processes help establishing a more detailed development and abandonment history of Crustumerium; can we determine for instance if Crustumerium was truly inhabited continuously and in what condition and under which circumstances it was abandoned.

To answer these questions, I have examined, compared and analysed data that was collected between 2011 and 2015 with the following methodology:

- Fieldwork: At the basis of this thesis lays all the data that was gathered through archaeological research by the GIA38_{and SS-Col, geophysical research by Eastern Atlas}39_{and GeoRes}40_{, and soil research by Jan}

Sevink41_{and Michael den Haan}42_{during fieldwork campaigns since 2011.}

- Particle Size Distribution Analysis: Besides examining the data collected during the various fieldwork

35_{Schiffer 1983, p. 677; Cameron 1993, p. 3; Joyce & Johannessen 1993, p. 138.} 36_{Schiffer 1972, p. 157; Schiffer 1983, pp.675/677-678.}

37_{Baker 1975, p.10; Schiffer 1972, p. 157; Schiffer 1983, pp.675/677-678; Horne 1993, p. 47; Shott 1998, pp.}

257/322.

38_{Noorda & Attema, 2013; Noorda & Attema, 2016} 39_{Ullrich & Pilz, 2012; Ullrich, 2013; Ullrich, 2015} 40_{Vercelli, 2015.}

(12)

8 campaigns, I also conducted analyses on the samples taken from soils encountered during the

excavations. One of these is particle size distribution analysis, which is used to determine the general size distribution of the individual sediments, with a focus on the texture of the sediments, in order to compare and possibly match sediments from different locations that have the same characteristics. - General mineralogical examination: In addition to the particle size analysis I conducted, I performed a general examination of the minerals in the samples, to assess which minerals were present and in which condition they were in order to get an idea of the nature of the post-depositional processes that created the sediment, and possibly the origin of the sediments.

Finally, this thesis is also a test case to determine whether analyses, such as particle size distribution analyses, yield positive enough results in relation to the costs (time and money) and whether it should become a more standard approach for archaeological investigations in the future.

1.4. Outline of this research

This thesis is divided into five chapters, including this chapter with the introduction to the site and the research aims, and will be followed by the conclusion and a short discussion on the methods and results. In chapter two an introduction will be given into the geology and the soils present at Crustumerium. This chapter together with chapter three form the basis on which this research was constructed and

conducted. Much of the information used in this chapter has to be credited to prof. Sevink, who, since 2012, has helped the project to conduct soil research to form a more detailed knowledge of the geology and soils present at Crustumerium. Chapter three is deals with the results of several excavations that have been conducted at Crustumerium since 2013: respectively the excavation of trenches Saggio 2 and Saggio 4 in the summer of 2013 (number I and II on map 6 and appendix 9.1.1.) that uncovered a complex system of successive roads in a deeplying trench. This is followed by the D1-trenches in the

(13)

9

2. The soils of Crustumerium

The settlement of Crustumerium is located on one of the hill plateaus along the wide Tiber river plain, north of Rome, at a height of 90-120m asl. These plateaus consist of a thick series of pyroclastic volcanic rock that originates largely from the Sabatini volcanoes, lying over pre-volcanic fluvial deposits. The latter consist largely of conglomerate beds that, in their turn, rest on a thick system of coarse to fine textured marine and perimarine deposits. These deposits are part of the Monte Mario and Monte Vaticano formations, and date from the Pliocene to Early Pleistocene.43_{They include the Lower}

Pleistocene Monte Mario Formation (MTM) and the Middle Pleistocene Fosso Della Crescenza Formation (FCZ).44_{The latter consists of fluvial conglomerates, which are composed of rounded, well-sorted, white}

limestone gravel, and sand and clay. They are still visible on the lower slopes north of the settlement area. The series of rock, which overlies these pre-volcanic fluvial deposits, ranges from rather loose ash to dense, hard tuffs.45_{They consist of well-consolidated tuffs, mostly ignimbrites formed by glowing}

avalanches or pyroclastic flows, to loose ash fall deposits, which are predominantly present in

Crustumerium. This is most likely because of the relatively large distance between the volcanoes and the area of Crustumerium, which means that only the lighter, finer ash can travel this distance, whilst heavier, denser ash will deposit closer to the volcanoes. Only the major volcanic units are reported and indicated for the region of Crustumerium on the geological map (map 3 and appendix 9.1.5.). However, it is likely that there have been many minor

eruptions that were not or only marginally recorded as pyroclastic strata.46_{One of the}

largest of the major series of pyroclastic volcanic rock present in Crustumerium are the Tufi Stratificati Varicolori di Sacrofano (SSKF). This series is made up of an alteration of fine ash and lapilli-sized fallout beds, of which some are altered into paleosols. The series is subdivided by Karner et al.47_{into three main units.}48_{The oldest}

of the three is the Tufo Giallo di Prima Porta (TGPP), which is a complex unit that includes a lithified upper unit. It is a yellow pyroclastic flow deposit with yellow pumice and rare grey scoria, and is most common in the Tiber valleys.49_This

unit is followed by the Grottarossa Pyroclastic Sequence (GPRS) and by the Tufo Terroso con Pomici Bianche (TTPB) unit.

43_{Karner et al., 2001; Funiciello et al., 2008; Sevink et al., 2016.} 44_{Giordano, 2008; Funiciello et al., 2008.}

45_{Karner et al., 2001; Funiciello et al., 2008.} 46_{Sevink et al., 2016.}

47_{Karner et al., 2001.} 48_{Sevink et al., 2016.}

49_{Karnet et al., 2001; Sevink et al., 2016.}

(14)

10 The latter is, according to Sevink et al.50_{, the best stratigraphic marker bed for either the Monte Sabatini}

or the Alban hills, as it is composed almost totally of a metres thick brown ash that holds two white to yellow pumice horizons in the centre. Officially no pyroclastic rocks from the Alban Hills have been reported for the area of Crustumerium, yet it cannot be excluded that ashes from the Alban Hills reached the area and have never been noticed.51_{In time significant pyroclastic depositions stopped taking place}

and the Tiber river began to incise the bedrock and cut deep valleys through it, which resulted in a landscape with the distinctive plateaus like the one on which Crustumerium lays. Although the valleys were very steep the slopes formed by the incising Tiber river were very stable. This is due to the fact that pyroclastic deposits are generally very permeable and when interbedded with dense and hard tuffs, surface erosion due to water had little effect. Additionally, when the slopes were covered with any type of vegetation erosion was most likely limited to the longer periods with dry and cold conditions in which plant growth was limited. For instance, during glacial periods. However, in the Late Holocene use of the land by humans e.g. for agriculture increased, which changed this stable environment and induced massive anthropogenic erosion since.52

Soil formation on the plateaus started already during the deposition of the pyroclastic rocks, alternating with the ash deposits, which resulted in the SKF and LTT units. According to Sevink et al.53_{these units}

demonstrate the general trend of soil formation in these highly weatherable and porous rocks in the

Mediterranean: Fast development of soils due to the rapid weathering of the instable volcanic glass and other weatherable components that are abundant in tuff. The weathering of these items generates the formation of clay, which in this way forms the characteristic reddish-brown argic B-horizons of the Mediterranean.54_{(map 4}

and appendix 9.1.6.) The reddish brown coloured clays are furthermore evidence for the accumulation of insoluble components, such as Iron and manganese, which are visible furthermore as black nodules and coatings in the soils. After the deposition of ashes declined, in the Mid-Pleistocene, the rapid weathering of the volcanic glass increased and proper soils started forming. They classify as deep chromic luvisols55_{with a}

50_{Sevink et al., 2016.}

51_{Sevink et al., 2016.} 52_{Sevink et al., 2016.} 53_{Sevink et al., 2016.}

54_{Ollier, 1984; Wada, 1987; Ugolini & Dahlgren, 2002; Sevink et al., 2016.}

55_{“Luvisols are soils with a pedogenetic clay differentiation (especially clay migration) between a topsoil with a}

lower and a subsoil with a higher clay content, high-activity clays and a high base saturation at some depth. Their parent material consists of a wide variety of unconsolidated materials such as glacial till, and aeolian, alluvial and colluvial deposits. Luvisols are most common in flat or gently sloping land in cool temperate regions and in warm regions (e.g. Mediterranean) with distinct dry and wet seasons.” (WRB, 2015)

(15)

11 prominent argic B-horizon to nitisols56_{with a very deep clayey argic B-horizon.}57_{Under the influence of}

severe erosion these soils tend to develop a dark coloured A-horizon that can be classified as luvic phaeozems58_{or even vitric mollic andosols}59_{. This is caused by the stabilization of organic material by the}

particles that release during the weathering of the underlying pyroclastic rocks.60_{The original soils were}

largely in the LTT-unit, and were originally very deep and with well-developed reddish brown argic B-horizons with a fine sub angular blocky structure (Nitisols).61_{However, the results from coring campaigns}

in 2012, 201362_{and 2015 illustrate that the thickness of the A-horizon and B-horizon, the solum, is}

nowadays very limited, and that the soils are severely eroded, leaving only a thin dark topsoil on top of the hard tuffs. This is observed in several archaeological studies63_{, and is especially evident after severe}

rainstorms. During these extensive erosion occurs, associated with the development of rills and gullies, causing the thin top layer to be quickly transported, revealing fresh, underlying tuff. The harder, more lithoid tuffs are relatively impermeable to water, which prevents fast infiltration of large quantities of water. This causes massive topsoil erosion. Intensive ploughing and sowing, together with the heavy Mediterranean rainfall that occurs at the end of the summer, creates a similar situation. Without a well-developed root system, the topsoil is very loose and fine soil particles are easily washed downslope. Only on adjacent, less eroded plateaus to the east and south of the settlement deep nitisols were

encountered with a solum up to 1.5 meters or more. Here the topsoil still retains characteristics of the original topsoil and holds a fair residual siliceous coarser fraction of sand and rounded fluvial gravel.64

Subsequently, in the valleys and on the lower foot slopes thick colluvial deposits were encountered that are created by the erosion. Especially the colluvial fill of the ancient roads65_{and the fossato}66_{(a meters}

deep filled-in moat) at Crustumerium, of which the latter has up to five meters of fillings, are a good example of the magnitude of this soil erosion. The filling of the fossato is characterised by compact earth, dark in colour, and in the vicinity of the southern limit, tuff blocks of medium size were present on the surface. It is unclear whether the filling was deposited in one rapid episode, or slowly through time,

56_{“Nitisols are deep, well-drained, red tropical soils with a clayey nitic horizon of high aggregate stability that has}

typical angular blocky structure breaking into polyhedral or flat-edged or nut-shaped elements with, in moist state, shiny aggregate faces. The parent material of nitisols consists of finely textured weathering products of inter-mediate to basic parent rock, in some regions rejuvenated by recent admixtures of volcanic ash. Nitisols are predominantly found on level to hilly land under tropical rainforest or savannah vegetation.” (WRB, 2015)

57_{WRB, 2007.}

58_{“Dark soils rich in organic matter.” (WRB, 2015)}

59_{“Andosols accommodate soils that develop in glass-rich volcanic ejecta under almost any climate (except under}

hyper-arid climate conditions). However, Andosols may also develop in other silicate-rich materials under acid weathering in humid and peri-humid climates. They are typically dark soils of volcanic landscapes formed by glass-rich volcanic ejecta (mainly ash, but also tuff, pumice, cinders and others) under almost any climate or other silicate-rich material under acid weathering in humid and peri-humid climates. Andosols have a high potential for agricultural production, but many of them are not used to their potential. Andosols are generally fertile soils, particularly Andosols in intermediate or basic volcanic ash and not exposed to excessive leaching.” (WRB, 2015)

60_{Sevink et al., 2016.} 61_{Sevink et al., 2016.} 62_{Den Haan, 2013.}

63_{Quilici & Quilici-Gigli, 1980, p.96; Amoroso, 2002, p.293; De Gennaro et al. 2004, p.148; Attema et al. 2013, p.55;}

Barbaro et al., 2013, pp.21-27; Willemsen 2014, p.28; Seubers & Trienen 2015.

64_{Sevink et al., 2016.}

(16)

12 but Barbaro et al.67_{do distinguish three separate fillings. The lowest two strata were darker in colour,}

inclined and convergent towards the centre of the moat; both contain no visible archaeological remains. The youngest fill is characterized by a light soil and the aforementioned tuff blocks. In this stratum fragments of seventh century B.C. pottery were found.

It is clear that the modern plateau has been largely stripped off its former deep soils as a result of long continued anthropogenic activities. Particularly, massive erosion as a result of the modern and intensive mechanical (deep) ploughing in the 20th_{century that caused this stripping and concurrent disappearance}

of superficial archaeological phenomena. As a consequence, the current soil over large areas consists of a plough layer (Ap), largely composed of a mixture of more or less crushed tuff fragments and material

originating from the original argic B horizon, over rather fresh tuff.

This creates the image that the landscape of Crustumerium eroded easily and was vulnerable. However, originally the tuff slopes and the soils were very stable and if the slopes were covered with any type of vegetation erosion was most likely limited to the longer periods with dry and cold conditions in which plant growth was limited. Furthermore, soils types such as the nitisols and luvisols are known for their aggregate stability.68_{It is only in the Late Holocene when use of the land by humans e.g. for agriculture}

increased that this stable environment changed and massive anthropogenic erosion initiated. Looking at the development of agriculture given in table 1 and comparing it with the original soils, it are especially the last phase of the settlement, the abandonment of the area in the 5th_{century B.C. and the}

mechanisation of agriculture in the 20th_{century that have caused most of the erosion at Crustumerium;}

especially luvisols are susceptible to structure deterioration where tilled when wet or with heavy machinery.69_{The initial small-scale agriculture during the Iron Age, and even the intensification of}

agriculture during the Orientalising and Archaic period would not have resulted in massive erosion of these stable soils. Even though the topsoil would have been ploughed, especially following the

intensification, there would have been enough soil underneath to absorb (excessive) rainwater, and as long as the soils were planted they would have stayed very stable; The deep and porous solum and the stable soil structure of Nitisols permit deep rooting and make these soils quite resistant to erosion.70

Besides the downslope displacement of soil during ploughing, erosion of the soil would have remained low in this period. However, when agricultural land is abandoned and the vegetation cover becomes limited or non-existent erosion will increase. Finally, after most of the soil has eroded away, combined with (deep) ploughing, only a thin stratum of soil would have been left between the topsoil and

underlying bedrock, which will no longer absorb excessive rainwater. The underlying bedrock would then have acted as an aquifer, resulting in rill and sheet erosion. These differences in the degree of erosion would most likely also result in different characteristics of sediment; from the residual, mainly coarser grained, weathered particles released by the slow erosion of the stable topsoil during the early stages of the settlement. This would have been followed by more clayey material when ploughing intensified and material from the underlying soils was ploughed up. Finally, the mechanised, deep ploughing would have caused massive erosion of all the soil on top of the bedrock, most likely including freshly ploughed bedrock, resulting in a mixed clayey sediment with large quantity of fresh minerals.

67_{Barbaro et al., 2013.}

68_{WRB, 2015; personal communication with Jan Sevink.} 69_{WRB, 2015.}

(17)

13

3. Focus areas

3.1. Previous research

As stated in the previous chapter, from the summer of 2012, research started to focus also on the soil of the settlement of Crustumerium, starting with a short coring campaign conducted by Den Haan with the help of several students. 71_{The aim was to gather a general knowledge of the soils on and around the}

settlement area with the use of randomly placed corings. This was followed shortly after with a larger, additional systematic coring campaign in the spring of 2013. During this campaign the focus was mainly on establishing the condition of the soils and the level of erosion present on the site, whilst adding to the knowledge gathered in 2012. This was achieved by the coring of long transects, which extended over the entire length and width of the settlement, and conducting additional corings on specific locations under the supervision of Jan Sevink. (appendix 9.1.4.) During these coring campaigns it became evident that a large part of the settlement area, especially around the tops of the hills, had suffered severely from erosion, and that often only a shallow solum was present on top of the tuff bedrock. However, in the lower sections of the slopes and in the valleys there was often a thick layer of colluvial soil present, with hardly any signs of soil development.

71_{Den Haan, 2013.}

Map 5 Overview of all areas where magnetic prospection was done by Eastern Atlas. (Ullrich, 2015). The excavations discussed in this thesis are located in the southern part of area A, which was surveyed by EA in 2011. (Ullrich & Pilz, 2012) The green lines show the resistivity profiles over Quilici O that conducted by EA in 2015, while the red lines show the profiles conducted by GeoRes (Vercelli, 2015)

(18)

14 This resulted into the conclusion that a large part of the archaeological remains most likely had eroded down the slope together with the colluvium. Subsequently, the approach towards choosing locations for excavation had to be based on prior knowledge of erosion and sedimentation, as most of the areas would likely not contain any remains of archaeology just below the surface, while other areas were likely to contain archaeological strata. Therefore, the focus shifted towards geophysics and the results of various geophysical campaigns conducted by the German company Eastern Atlas GmbH Co. KG. The results show anomalies that could be archaeological features or contain archaeological material, and possible even old soil stratigraphy that was not affected by the same level of erosion that had changed the rest of the settlement area. In particular, the geomagnetic surveys72_{which cover the entire}

settlement of Crustumerium (map 6), became the basis of the excavations since 2013, including the excavations the data whereof is used in this research. The excavations that form the basis of this analysis started in 2013 and were carried out in the southeast part of the settlement area, east of the old Road Trench, and in the Monte Del Bufalo necropolis. (map 6 and appendix 9.1.1.) The latter has been the focus area of the GIA’s research of burial practices since 2006 and every summer several tombs are excavated in this area as part of the GIA’s summer field school. Since 2013 excavations have been taken place at three different locations: In the summer of 2013 both trench Saggio 2 and Saggio 4 (numbers I and II on map 6, map 7 and appendix 9.1.1.) were excavated in order to verify, examine and classify two large, oblong geophysical anomalies that run parallel to the modern dirt road leading up to the

excavation house. This was followed in the autumn of the same year with the excavation of the D1

-trenches (number IV on map 6, appendix 9.1.1 and 9.1.2.), which were dug to verify, examine and classify an elongated, curved geophysical anomaly on the slope north of the Monte Del Bufalo burial ground. In both 2014 and 2015 the excavation was focussed on Quilici O (the artificial hill), with trench R7 (number V on map 6, appendix 9.1.1 and 9.1.3.) being excavated in 2014 and trench QO_15 (number VI on map 6, appendix 9.1.1 and 9.1.3.) in the following year. Trench R7 was excavated in order to verify, examine and classify the results from the geomagnetic survey of Eastern Atlas from 2011 and the

additional resistivity survey conducted by Eastern Atlas in 2013; trench QO_15 was a continuation of this trench after the excavation of 2014.

3.2. Test trenches 2013

3.2.1. Trench Saggio 2 and trench Saggio 4

The first two in the series of excavations took place in the summer of 2013 and were located just north of the modern dirt road that leads to the excavation house and south of the so-called fossato-area (numbers I and II on map 6, map 7 and appendix 9.1.1.). The area gives place to a cluster of tombs in the Monte Del Bufalo necropolis that lays just south of the fossato, the remaining part of the ancient defensive system of an agger (earthwork) and a fossato (moat). The fossato, which is filled up with colluvial material, marks the southern border between the settlement and the sepulchral areas. On the geomagnetic map of Eastern Atlas (map 6 and appendix 9.1.1.) the fossato is visible as one of the large

72_{Results are fully discussed in both the internal reports from Eastern Atlas as well as the internal excavation}

(19)

15 linear anomalies in the Monte Del Bufalo area, in this case the one that runs along the entire southern border of the settlement.

The existing knowledge of the location of the fossato (appendix 9.1.1.) and the presence of an anomaly at this location on the geomagnetic map confirmed the value of the geomagnetic surveys and instigated further research of two other linear features north of the modern dirt road. Respectively 120m and 160m long, both anomalies appear to run parallel to the dirt road over a considerable distance before they disperse slightly. The top one is visible as a mainly white signal, whilst the lower anomaly is visible as a dark grey/black signal.73_{In order to verify the existence of a feature or features at the location of the}

geomagnetic anomalies and to examine what causes the difference between the white and dark grey/black signal it was decided to excavate.

To research both of these anomalies, two trenches74_{were placed perpendicular to the signals (map 7);}

Trench Saggio 2 was planned at the easternmost end of the excavation plane of 2013, whilst trench saggio 4 was planned further to the west. Both cut through the strongest parts of the signals. During the excavation it quickly became evident that the stratigraphy (appendix 9.2.1. and appendix 9.2.2.) was much more complex than could be observed in the corings, and that more than just the plough layer became visible. Furthermore, it showed that the modern landscape of Crustumerium, which consists of rolling hills and flat areas, does not represent the ancient landscape, which was more accentuated. Indeed, the landscape must have been much more pronounced with deep cuts and plateaus of tuff; for

73_{Ullrich & Pilz, 2012.} 74_{Noorda & Attema, 2013.}

(20)

16 instance, the excavation of trench Saggio 2

and 4 revealed a deeplying trench with three road levels, quarries and drainage channels, while nowadays the area is relatively flat.75

The top layer in both trenches (US 140 in saggio 2 and US 244 in saggio 4) is a thick plough layer that still shows marks from deep ploughing done in the 20th century.76

The compactness of this soil is high, especially in the dry summer months, and even for a large part of the autumn and winter the water has difficulty penetrating deeply into the soil. This top layer, which is similar over the entire settlement, has colluvial characteristics combined with

organic material from recent cultivation of Erba Medica (Alfalfa/Lucerne). The archaeological remains range from Archaic to Modern period, although a large part dates to Roman to modern times. Right below this top layer is the youngest stratum of colluvium (US 138 in saggio 2 (figure 1; appendix 9.2.1.) and US 200 in saggio 4 (appendix 9.2.2.)) in the stratigraphy. It covers the first and most recent of the three road levels. The colluvial layer consists of a dark compact soil that is similar to the plough layer but has a slightly lighter colour, which is likely due to the lack of organic material or a lower water saturation. The most recent road was only excavated partly in both of the trenches, but appears to be composed of multiple sections and is likely to have undergone modifications during its use. It is built from large white tuff stone blocks, with some gravel fillings in the extensions, and has an absolute width of 3,40 meters. (picture 1) The roadbed was built on several foundation layers that were dug into the underlying stratum, and consists of at least two layers of compact black soil that only differ as to the blackness of the colour. In saggio 2 it was also

visible that a small channel was dug directly next to the road on the southern side. However, it is unclear whether water ran through the channel or only over the thin stratum of tuff chunks that covered the channel. The filling of this feature consisted of moist clayey soil. (US 153 in saggio 2) No diagnostic archaeological material was found in the roadbed or in the thin gravel layer that covers the entire roadbed (US115 in saggio 2). 75_{Noorda & Attema, 2013.}

76_{Plough marks up to a meter deep are visible in section 1 of Saggio 2. (figure 1 and appendix 9.2.1.) (Noorda &}

Attema, 2013)

Figure 1 The Roman level in trench Saggio 2. (appendix 9.2.1.) (Noorda & Attema, 2013)

(21)

17 Nonetheless, based on a combination of

arguments it is most likely a Roman road. For instance, the road had a direction towards a known Roman villa east of Crustumerium, and the distance in between the wheel ruts visible in the middle of the road falls within the standard Roman dimensions known from Adam77_{. But first and foremost, it is}

covered by the aforementioned youngest colluvial stratum that contained mainly Roman pottery. The foundation of the Roman road was dug into a second layer of even more compact colluvium (US 101 in saggio 2 and US 242 in saggio 4) that only contained few pottery sherds dating

to the Archaic period. Subsequently, this stratum, itself dating to the (post)Archaic period and running from the northern edge of the trenches to their southern edge, dates all underlying strata and features as (pre)Archaic.

After opening a part of the sepulchral area north of the trenches it became clear that the bedrock was very near to the surface in this part, and that it was easy to follow the bedrock from this point onwards to the northside of the trenches. In trench saggio 2 the bedrock slopes downwards within the first two meters of the trench, and a transition is visible from the yellowish LTT-tuff, which is near the surface, into a blue/grey tuff (US104) that is much finer grained. In trench saggio 4 the situation is more complicated. Here signs of quarrying were observed by Belelli Marchesini from SS-Col. On figure 2 it is visible that several vertical, straight cuts are made in the bedrock, as well as steps in the slope. These cuts resemble marks that were found during previous excavations at Crustumerium and also interpreted as evidence for quarrying activities.78_{In the northern corner two parallel cuts are visible with an}

interspace of circa one meter, which most likely are from a tomb. It is well known that the quarries were reused as burial ground after quarrying had stopped, and although (direct) dating is not possible, the filling of the quarried area contains a discarded cippo (grave marker), and is therefore believed to be from the Orientalising/Archaic period. During this period many gravemarkers were removed from the sepulchral areas and/or destroyed. In trench saggio 2 the sloping side of the blue/grey tuff changes for an even darker blue, and finer grained tuff (US108) that forms a plateau at the bottom of the slope. In between the bottom edge of the slope and the dark blue tuff plateau, a small (drainage) channel was dug, which shows some possibly original cut marks. In saggio 4 the situation is similar, yet the channel is dug into a different type of tuff; a very fine grained, grey/brownish tuff with white speckles (US187), which forms a tuff plateau also on the southside of the channel. It is interesting to see that the general stratigraphy is similar in both trenches, yet revealing a different succession of tuff. This confirms the

77_{The distance between the wheel ruts is ca. 1.25m, which coincides with figures given by Adam (1984), who}

measured ruts in Roman roads and concluded their average distance/width is 1.3m.

78_{Based on personal communication with Belelli Marchesini during the excavation.}

(22)

18 statement of Sevink et al.79_{that the tuff stone diversity in and around the settlement of Crustumerium is}

much broader than the geological map of Rome shows. The filling of the drainage channel appears also to differ between both trenches, as saggio 2 shows only two distinctive fillings, which both consist of a very homogeneous mix of sand and clay, whilst in saggio 4 five layers are visible inside the channel. The latter appear to be a succession of thin gravel layers with a varying amount of gravel between the layers. Nevertheless, in both trenches the fillings seem to be naturally deposited and support the hypothesis of a drainage function. Further differences between both trenches are visible between the tuff plateau with the drainage channel and the aforementioned Roman road that lays several meters higher and more to the south, resulting in a very complex stratigraphical succession. Different from the Roman road, which appears to be more or less similar in both saggio 2 and saggio 4, it is visible in map 7 that the location of the second, Archaic, road level has changed between the two trenches. Moreover, the height between both road levels differs as well, with in saggio 2 the second road being on the same level as the tuff plateau mentioned earlier, whilst in saggio 4 the road is located almost directly below the Roman road. (visible between appendix 9.2.1. and 9.2.2.) The Archaic roadbed is made of a stamped gravel surface, called battuto, of which the amount of gravel again differs between the trenches. The battuto is bounded by a row of larger tuff blocks. The road is built on top of a stratum of blue/grey gritty soil (US 157 in saggio 2 and US 207 in saggio 4) that runs from the top of the sloping bedrock to the lower side of the Archaic road, and covers the drainage channel and tuff plateau. The stratum appears to be placed as a reinforcement layer on top of yet another stratum of colluvium (US 209 in saggio 4). The latter is the oldest of the three colluvial layers and is only really visible in saggio 4. (figure 3) Here it separates the plateau of tuff bedrock and the drainage channel with the level of the archaic road. In saggio 2 it seems only to be the Archaic foundation layer (US 157) that separates the two. Below the third layer of colluvium lays not only the tuff plateau and the drainage channel but also another road. The deepest, third road is also made of a (a

series of) battuto, however without the boundary of tuff blocks. In saggio 2 this roadlevel lays almost directly below the Archaic road, but in saggio 4 the Archaic road and the underlying road are separated by a thick section of the third colluvial stratum. The dating of the third and deepest road is very difficult, as no (dateable) archaeological remains were recovered at this level. However, based on the stratigraphy it is possible to link the third road to the plateau of tuff80_{and the drainage channel}

situated at the bottom of the tuff 79_{Sevink et al., 2016.}

80_{In trench Saggio 2 this is a plateau made from a dark blue, fine grained tuff (US108), while in Saggio 4 it is made a}

very fine grained, grey/brownish tuff with white speckles (US187).

(23)

19 slope on the north side of both trenches.

Because the foundation layer of the Archaic road covers the quarry it is most likely that the quarry is also linked to the third and deepest roadlevel. This resulted in the hypothesis that these features and

intermediate strata together form the earliest phase of the structure and most likely date to the Iron Age.

This results in a complex stratigraphical system of successive roads cut into the tuff bedrock that are separated by strata of colluvial sediment; starting with a (presumed) Iron Age level at the bottom of a large tuff slope, similar to the deeplying road in the middle of the settlement (the so-called trincea viaria). This phase consists of a (series of) battuto road surface next to a flat section of tuff in which a drainage channel was cut. In trench Saggio 4 also traces of quarrying have been found that most likely date to the earliest phase of the sequence of roadbeds. The successive, Archaic road lays either directly on top of the Iron Age road, as is the case in trench Saggio 2, or separated by a stratum of colluvial sediment, as is the case in trench Saggio 4. The Archaic road consists also of a battuto bounded by large white tuff blocks, which is constructed on top of a thick foundation layer, which forms a clear transition between the Iron Age phase and the successive Archaic phase. Finally, the Archaic road is covered by another stratum of colluvial sediment before a Roman road was constructed on top of the colluvium. Finally, also this road level is covered by colluvium after which the deeplying road system completely disappeared.

3.2.2. D

1

-trenches

After the success of this new approach towards excavations on the basis of geophysical observations, a follow-up campaign was planned in the late autumn of the same year.81_{Again the focus was to verify,}

examine and classify geophysical anomalies visible on the geomagnetic map. This time the anomalies were located higher up the slope just east of the old road trench. Here two linear anomalies are visible on the map (number IV on map 6 and appendix 9.1.1. and 9.1.2.), respectively 110m and 250m long, that follow the relief of the slope. The geophysicists interpret these anomalies as possible ditch-like

structures82_{, a type of anomaly that}

seems to be present at more locations in Crustumerium, especially near the tops of slopes/hills. During this campaign four trenches, labelled the D1-trenches, were

dug to research the most eastern one of the two anomalies. (appendix 9.1.2.) The phases of excavation in which the 81_{Noorda & Attema, 2013.}

82_{Ullrich & Pilz, 2012.}

(24)

20 trenches were left at the end of the campaign differ, based on the usefulness of the results compared to the other trenches, or, as was the case with trench D1_5 (picture 2), because it was decided that a part of

the trench would stay in situ for a larger excavation at a later date. All the trenches gave the same general image, and confirmed the hypothesis of a ditch, most likely a drainage channel with several different fillings that became covered by three other strata.

In trench D1_1, which shows the most extensive stratigraphy of the four trenches, there are three strata

visible that cover the ditch. The top two strata are similar to the situation in saggio 2 and saggio 4, with a plough layer on top (US 01 in D1_1) followed by a mixed layer of colluvium (US 09 in D1_1). Both contain

remains of Roman pottery, charcoal, and some stones. Below these two layers there is stratum of compact soil that contains many pottery sherds, charcoal and tuff chunks. Different to the top layers the pottery in this stratum is dated to the 6th_{century B.C., suggesting that the ditch must have been filled in}

or after the 6th_{century B.C., making the layer very similar to the colluvial layer below the Roman road in}

saggio 2 and saggio 4. Directly below this colluvial deposit is a ‘pavement’ of tuff blocks that closes off the channel. The blocks have a width of almost two meters and are slightly cemented together, which could indicate that the stones were exposed to weathering. In trench D1-5 the pavement appeared to be

the best preserved and therefore in this trench only a small depression next to the pavement and the larger V-shaped drainage ditch D1 was excavated and the remaining pavement was not further

excavated. It was decided that at a later stage it was possible to uncover a larger part around the trench in order to verify the hypothesis of a pavement. Underneath the packing of tuff blocks the bedrock slopes downwards on both sides, forming a V-shape. The top filling of the ditch is made of very compacted, clean, homogeneous soil (US 03 in D1_1), without any archaeological remains except for

several tuff blocks. Further down follows a thin layer of very compacted clay (US 06 in D1_1) at the

bottom of the ditch. The oldest layer of the ditch is a thin layer of limescale (US 07 in D1_1) that covers the

slopes of the ditch from a depth of 130cm to the bottom. This could be caused by the evaporation of standing water; however, it is also possible that the calcium is released by weathering of the tuff; If drainage, due to denser underlying tuffs, is impeded it can lead to the accumulation of secondary lime.83

Nevertheless, it is very likely that the channel held water, or transported water during the (early) use of the feature.

All the trenches dug through the D1-feature show that we are dealing with a V-shaped channel that likely

had a function as drainage, supported by the signs of water transportation or stagnation. During its use or after it became in disuse the channel filled up and was overlain with large tuff blocks. Whether the channel was intentionally filled or naturally filled by material that was flushed down the slope has still to be answered. However, the pottery found in between the large tuff blocks on top indicates that in the late 6th_{century or early 5}th_{century B.C. the ‘pavement’ was laid and the channel was no longer visible.}

(25)

21

3.3. Test trenches 2014 and 2015

3.3.1. Trench R7 and trench QO_15

In 2014 the focus diverted further to the east, towards the artificial hill of Quilici O. (map 6 and appendix 9.1.1. and appendix 9.1.3.) Similar to the previous locations, on and around Quilici O several geophysical anomalies are visible on both the geomagnetic map84_{and on the resistivity profile}85_{produced by Eastern}

Atlas. After examining the results of this non-invasive research it was decided to excavate a test trench through Quilici O in order to verify, examine and classify the geophysical results, and to research

whether the initial hypothesis of a tumulus was indeed correct.86_{At first glance Quilici O appears as one}

of the rolling hills natural to Crustumerium, although located at an odd angle. However, during the excavations of 2014 and 21587_{it became clear that, similar to the anomalies examined in trenches saggio}

2 and saggio 4, it had a complicated stratigraphy and the need for more knowledge of the soils became even more pressing. After processing the data recovered in both campaigns three main phases became distinguishable in the build-up of the hill, but at the time of writing this thesis it was not yet possible to create a more detailed chronology due to the complicated stratigraphy inside the hill.88_{On the}

geomagnetic map from Eastern Atlas three linear anomalies are visible on the top of Quilici O. One has a southwest-northeast

orientation, and the other two have a northwest-southeast orientation. (map 6 and appendix 9.1.2.) Together the three lines seem to form a large rectangular feature of circa 40m by 60m. Inside this rectangle is a smaller rectangular anomaly of 27m by 27m, the eastern edge of which is unclear and most likely eroded.89_{Two trenches}

were planned based on these anomalies, and, although the initial plan was to cut through the entire length of the hill, they remained limited to the

southwest part.90_{First, trench}

85_{Ullrich, 2013.}

86_{Quilici & Quilici-Gigli, 1980.} 87_{Noorda & Attema, 2016.}

88_{There are archaeological remains collected during the excavations, such as pottery; during the summer campaign}

of 2016 all material has been studied, drawn and described. However, at the time of writing this thesis the results of this study were not yet available.

90_{This was due to the quick discovery of the stone pavement on top of Quilici O (labelled US 03) after which it was}

decided not to go through the stones, but to follow the pavement towards the south. After 23 meters the pavement stopped and it was possible to go deeper, resulting in a trench in the southwestern part of Quilici O.

(26)

22 R7 (number V on map 6, appendix 9.1.1 and 9.1.2.)was excavated in 2014, following the resistivity profile from Eastern Atlas.91_{After a successful campaign with encouraging results it was decided to continue in}

2015 with an enlargement of the 2014 trench that resulted in a trench of quadruple the size, named QO_15 (number VI on map 6, appendix 9.1.1 and 9.1.2.). This enlargement and continuation in the same area resulted in the discovery of new features, as well as changes and adaptions to hypothesis made in 2014. Besides the three phases that are directly linked with the artificial hill, several other phases are visible that pre- and postdate the formation of Quilici O. Below I review the main phases:

The oldest phase is a combination of a gully cutting through the bedrock and its fillings in the lower southeast part of the hill. At first it was thought to be linked to the fossato defensive system. However, after examination of the fillings by Sevink during the campaign of 2015 this interpretation changed. He states that these fillings (US 18 to US 22) are still in situ and may date to the Eemian. They appear to be alluvial deposits with marks of water erosion and show presence of limescale in the lower fillings. This would mean that the gully was already filled and inactive before the settlement was founded.

Nonetheless, it gives an example of the changes that have occurred in the landscape even before human intervention.

The first phase of Quilici O is visible just north of this (paleo)gully, on the southern side of the hill, and forms the centrepiece of both trenches. It consists of a structure of eight by eight by one meters and is constructed directly on top of the bedrock. It seems to be built up of three (different) soil layers (US 08, US 10, and US 11), which together form a mound. The soils of these three layers are very clean,

homogeneous, and sterile. They seem only to differ as to the colour of their soil. Although only a fraction of the mound has been excavated so far, it seems that solely the top layer contains pottery, which in this case consists of several burnished dark impasto sherds dating to the Iron Age. The lower

91_{Ullrich, 2013.}

(27)

23 two layers are almost without any archaeological remains. The sides of the mound are covered by a thin layer of gravel (US 13), most likely as a protection against the elements.

Initially the mound was thought to be the remains of an agger (earthwork), mainly because of the nearby (paleo)gully and the defensive fossato. However, after confirmation by Sevink of the age of the

(paleo)gully together with the fact that instead of only two sides, which is normal for an agger, all sides of the mound sloped downwards, this hypothesis changed. After a specific cut became visible in the bedrock underneath the centre of the mound, which looked similar to known marks that indicate tombs, it became clear that the mound was actually a tumulus dating to the Iron Age. This interpretation was further reinforced by the discovery of twelve large, white tuff blocks in the top stratum of the mound. (picture 4 and figure 4) Together the blocks form an apsidal structure in the middle of the tumulus. Their function is still unclear. Thanks to the cover of Quilici O, the Iron Age funerary architecture has been preserved. Elsewhere in the burial ground it will most likely have been destroyed by ploughing. However, one of the hypotheses is that the sepulchral area would have had many of these tumuli from the Iron Age, but that through time they all disappeared. The idea therefore is that the tumulus found during the excavation of Quilici O is only one of many, and that under Quilici O still several tumuli must be present, preserved by the cover of Quilici O. The next phase is the formation of a platform in the

Orientalizing/Archaic period, which covered the Iron Age tumulus/tumuli. During this phase the area is transformed into one significant feature in the area of Monte Del Bufalo. The platform is constructed of a succession of thin(ner) gravel layers that surrounds and partly covers the Iron Age tumulus and the area northwest/northeast of the tumulus. Eventually this created a new, level surface one meter higher than the original bedrock. This platform closed off all the Iron Age tombs, but was still used for burials, as attested by outlines of tombs cutting through the stratigraphy. During this procedure it is likely that the top of the Iron Age tumulus has been altered, attested by the lack of a real top, and the single row of tuff blocks. The final phase of Quilici O is the transformation of the Orientalizing/Archaic platform to the tumulus or artificial hill that is visible in the present. During this phase an enormous, thick stratum of reddish soil (US 02) filled with tuff stones, mudbricks, and charcoal, was deposited. This raised the surface at least 1,60m-1,80m higher than the previous platform, which was already one meter above the bedrock. Pottery found in this stratum suggests that is was raised during the 6th_{or early 5}th_{century B.C.,}