base CORSO PORTA RENO, FERRARA (NORTHERN ITALY): A STUDY IN THE FORMATION PROCESSES OF URBAN DEPOSITS Mauro Cremaschi1 & Cristiano Nicosia1 1 Dipartimento di Scienze della Terra “A. Desio”, Università degli Studi di Milano, Via Mangiagalli 34, 20133, Milano, Italy. (mauro.cremaschi@unimi.it; cristiano.nicosia@unimi.it) Corresponding author: M. Cremaschi ABSTRACT: Cremaschi M. & Nicosia C., Corso Porta Reno, Ferrara (Northern Italy): a study in the formation processes of urban depo- sits. (IT ISSN 0394-3356, 2010) The city of Ferrara (Emilia Romagna, northern Italy) is located on the left bank of the branch of the medieval channel of the Po River, the “Po of Ferrara”. This paper presents the results of the geoarchaeological study of the stratigraphic sequence exposed during excavations in Corso Porta Reno, located at an intermediate position between the palaeo-Po alluvial ridge and the adjacent floodplain. The area, also referred to as the “San Romano complex”, was excavated between 1981 and 1984, revealing an exceptionally well-pre- served sequence, comprising occupation deposits and structural remains dated between the 8th and the late 13th century AD. Geoarchaeological observations in the field and the micromorphological study of soil thin sections allowed the reconstruction of the natural and anthropic processes involved in the formation of the site stratigraphic sequence. This is characterised by an alternating succession of anthropogenic units (trampled domestic deposits, layers deriving from construction, destruction and/or refurbishing of buildings, materials from decantation within anthropic structures) and of natural alluvial deposits (light-coloured sandy layers often showing cross laminations, linked to a crevasse-splay context active during the human dwelling). The waterlogged nature of the site, the rapid rate of sedimentary accretion, and the sealing of the sequence due to the construction of a brick house in the 13th century avoided the well-stratified deposit to be heavily homogenized by bioturbation or reworked by later human activities. The model envisa- ged at Corso Porta Reno could partly explain the strong vertical accretion of urban deposits during the Middle age and provide a pos- sible mode of formation for thick “Dark Earth” sequences in medieval cities in active alluvial settings. RIASSUNTO: Cremaschi M. & Nicosia C., Corso Porta Reno, Ferrara (Italia settentrionale): studio dei processi di formazione di depositi urbani. (IT ISSN 0394-3356, 2010) L’articolo presenta i risultati dello studio geoarcheologico della sequenza stratigrafica esposta durante gli scavi archeologici in Corso Porta Reno a Ferrara, in un’area posta tra il dosso del paleo-Po e la piana alluvionale adiacente. Il sito, conosciuto anche come “com- plesso di San Romano”, fu scavato tra il 1981 e il 1984 e restituì depositi d’abitato ed elementi strutturali datati tra l’VIII ed il XIII secolo d. C., in eccezionali condizioni di conservazione. Lo studio geoarcheologico sul campo, integrato da quello di sezioni sottili, ha per- messo di ricostruire i processi naturali e le attività antropiche responsabili per la formazione della sequenza stratigrafica del sito. Essa è caratterizzata dall’alternanza di unità antropogeniche (depositi di origine domestica soggetti a calpestio, unità derivanti dalla costruzio- ne e/o dalla distruzione di edifici, sedimenti decantati all’interno di strutture negative) e di depositi alluvionali naturali (strati sabbiosi di colore chiaro, spesso con laminazioni incrociate, legati ad un ambiente di lobo di rotta attivo durante la frequentazione antropica del sito). Le condizioni di saturazione idrica, il rapido accrescimento sedimentario e la costruzione di un edificio in mattoni nel XIII secolo hanno evitato che la stratificazione originaria del sito venisse omogeneizzata dalla bioturbazione o sconvolta da attività antropiche nei secoli successivi. Il modello delineato in Corso Porta Reno potrebbe in parte spiegare la forte crescita verticale delle stratificazioni urbane durane il Medioevo e fornire un modello per la formazione di spesse “Terre Nere” in città poste in ambienti alluvionali attivi. Keywords: Ferrara; Po river; urban archaeology; “Dark Earths”. Parole chiave: Ferrara; fiume Po; archeologia urbana; “Terre Nere”. Il Quaternario Italian Journal of Quaternary Sciences 23(2Bis), 2010 - Volume Speciale - 373-386 1. INTRODUCTION This paper reports the geoarchaeological study of a stratigraphic sequence exposed in Corso Porta Reno in Ferrara (Emilia-Romagna region, northern Italy) – (Fig. 1). At this site, also referred to as the “San Romano complex”, archaeological excavations were carried out between 1981 and 1984 (GADD & WARD-PERKINS, 1991; VISSER TRAVAGLI, 1992; VISSER TRAVAGLI & WARD-PERKINS, 1981). Unlike many major Italian towns, Ferrara is not a city of Roman origin, but was settled as Byzantine castrum in the 7th century AD (VISSER TRAVAGLI & WARD- PERKINS, 1981) on the left bank of the branch of the Po River active during the early medieval period (the “Po of Ferrara”; BONDESAN et al., 1995). Due to its position, the city acted as an important trade centre connecting the Po plain and northern Italy to the Adriatic Sea and con- sequently flourished since the Middle Ages becoming a main centre in the Renaissance (VISSER TRAVAGLI & WARD-PERKINS, 1981). Ferrara is located at the apex of the delta plain of the Po River (CASTIGLIONI & PELLEGRINI, 2001), where the distributary channels are highly unstable and likely to switch frequently their course. A major avulsion, known as the “rotta di Ficarolo” (mid-11th century – end of the 12th century AD), caused the main channel of the Po River to migrate to the north, reaching the present posi- tion, c. 5 km north of the modern town of Ferrara (VEG- GIANI, 1985; BONDESAN et al., 1995). After the “rotta di Ficarolo”, the Po of Ferrara and its two eastward bran- ches (Po of Volano and Po of Primaro) stayed active and able to generate floods, also due to the input from the Apennine streams, until the beginning of the 16th 374 M. Cremaschi & C. Nicosia Fig. 1 - Geomorphological setting of the city of Ferrara: (1) Alluvial ridge; (2) Uncertain trace of alluvial ridge; (3) Palaeo-river channel, level of the surrounding plain; (4) Crevasse splay; (5) Ancient crevasse channel; (6) Canals; (7) Present-day river levee (after BONDESAN et al., 1995). Inquadramento geomorfologico della città di Ferrara: (1) Dosso fluviale; (2) Tracce incerte di dosso fluviale; (3) Paleo-alveo al livello della pianura circostante; (4) Lobo di rotta; (5) Antico canale di rotta; (6) Canali; (7) Argine fluviale attuale (da BONDESAN et al., 1995). century AD. At this time, the Po of Ferrara was cut off from the active hydrographic network of the area by human intervention (BONDESAN et al., 1995). The position of the Po River channel crossing Ferrara during the medieval period is still evident in the modern topography of the town (STEFANI, this volume). The river flowed, similarly to most of the water courses in this part of the Po plain, on top of a sandy alluvial ridge delimited by natural levees (BRIDGE, 2003), about 4–5 m above the surrounding floodplain (Fig. 1). The latter consists of highly organic clays and peats, as observed by CREMASCHI (1992) in the archaeological excavations at Piazzetta Castello, an area inhabited only after the 13th century AD north of the Po (Fig. 1). Crevasse splays originating from the left side of the river channel were common and can be detected from the elevation models published by BONDESAN et al. (1995). The site of the archaeological excavation at Corso Porta Reno is located at the lower fringe of the Po river ridge, in an intermediate position between the latter and the floodplain, possibly at the outer margin of a crevasse-splay fan. The peculiar position of the site allowed an excellent preservation of the archaeological deposits. Wood remains, pertaining to a series of supe- rimposed structural phases were exposed during exca- vations. These phases were dated by means of the associated material culture between the 8th and the late 13th century. Based on GADD & WARD-PERKINS (1991) and VISSER TRAVAGLI & WARD-PERKINS (1981), the earliest structures exposed at Corso Porta Reno have been interpreted as open-sided temporary sheds, followed by huts with a raised living space above an open-sided ground floor used for storage. A later phase was cha- racterised by timber-framed houses, possibly with an upper storey for domestic activities, with evidence of repeated destruction and reconstruction. Lastly, in the late 13th century, the last timber-framed house was replaced by a brick building. All of the above-mentio- ned structural phases are characterised, according to the excavators, by an alternating sequence of: “floors [...] of blue-grey silt and [...] accumulated tread layers interleaved with patches of clean silt reflooring” (GADD & WARD-PERKINS, 1991, p. 108); or of “a series of very pure silt deposits separated by charcoal-rich tread layers” (VISSER TRAVAGLI & WARD-PERKINS, 1979/80, p. 59); or again it was stated that “these [...] houses underwent refurbishment, indicated by [...] the gradual build-up of tread and reflooring layers” (GADD & WARD-PERKINS, 1991, p. 112). Making use of soil micromorphology and field observations, the present study is focused on deciphering the sedimentary processes that establi- shed the stratigraphic sequence, characteristic of a peri-fluvial urban environment. In particular, this investi- gation is aimed at identifying the formation processes of both the light-coloured layers, previously interpreted as man-made floors, as well as of the dark anthropoge- nic layers. 2. MATERIALS AND METHODS Samples were collected from the stratigraphic profile (Sect. 2) exposed during the 1982 excavation and subsequently integrated by further observations in 1983 (Figs. 2, 3). Sampling has been aimed especially at interpreting the sequence from the stratigraphic and 375Corso Porta Reno, Ferrara (Northern Italy): a study ... sedimentary standpoint, avoiding anthropic structures. Thin sections were manufactured from air-dried, undisturbed and oriented blocks according to standard methods (MURPHY, 1986). For thin section description the terminology of STOOPS (2003) was employed, together with the concept of “Soil Microfabric Type1” (hereafter SMT) (MACPHAIL & CRUISE, 2001; GOLDBERG & MACPHAIL, 2006) 3. RESULTS 3.1 The stratigraphic section The studied stratigraphic section is delimited at both sides by brick walls (Figs. 2, 3, 4). The strati- graphic units observed in the field are hereafter briefly described: 3.1.1 Lower layers They consist of two sub-units: (a) planar orga- nic loamy sand layers, massive or discontinuously lami- nated with mottles (5Y 4/1dark gray) (Samples 9, 3 – Fig. 3); (b) laminated organic-rich loam layers (5Y 2.5/1, black – 10YR 3/1, very dark gray) rich in charcoal, wood fragments, vegetal residues; they also include large brick fragments and discontinuous lenses of subroun- ded lithorelictcs with sandy texture. Intercalated to these layers there are lenses of sediments rubefied and hardened due to exposure to fire (Samples 1, 4 – Fig. 3). 3.1.2 Ditch, cesspit and associated deposits In the stratigraphic profile of Fig. 3, the lower layers are cut by a cesspit related to the timber-framed house. The basal deposit of the ditch consists of peat (Sample 10), rich in wood fragments, sealed by a sandy layer (Sample 8). The cesspit is filled by charcoal and organic rich layers in the lower portion, then by alterna- ting sand and organic-rich sub-units, described below. These alternations are associated with the timber-fra- med house deposits. 3.1.3 Timber-framed house related deposits Moving upward in the profile in Fig. 3, we observe a series of planar lenticular layers, concave at the base, as they seal the depression inherited from the underlying ditch. These layers include the structures, the horizontal beams, and the posts pertaining to the timber-framed houses. Two sub-units of contrasting lithology were observed: (a) Medium to fine grained loamy sands, (2.5 YR 5/2 greyish brown); planar lamina- tions, cross laminations and small ripples are often put in evidence by diagenetically oxidised laminae (sample 5). These sub-units do not display evidence of pedolo- gical alteration and are devoid of any archaeological material. (b) Loam to loamy silts (from 2.5 Y 3/2, very dark greyish brown, to 5 Y 4/3, olive), massive or displaying discontinuous planar lamination (samples 2 and 6). They include charcoal, wood fragments and archaeological materials (mammal bones and brick 1 Each SMT corresponds to a set of observed micromorpholo- gical characteristics which can occur within different thin sec- tions and are labelled with progressive numbers. Variations within the same SMT are indicated by a small letter (for exam- ple SMT 3, 3a). 376 Fig. 2 - Left: location of the Corso Porta Reno excavation in the city of Ferrara. Right: excavation plan with position of the studied pro- file; (a) modern (19th century) structures (walls, cistern, well), (b) late-13th century brick-house walls; (c) limits of excavated area. A sinistra: posizione del sito di Corso Porta Reno nella città di Ferrara. A destra: pianta di scavo con posizionamento del profilo esaminato; (a) strutture moderne (XIX secolo): muri, cisterna, pozzo; (b) muri in mattone della case della fine del XIII secolo; (c) limiti dell’area scavata. Fig. 3 - The studied stratigraphic profile: (A) Top soil; (B) Alternating lenses of sand (dotted) and loamy organic layers (light grey); (C) Peat deposits and wood; (D) Wood posts, lenses reddened by burning and brick fragments. Numbers on the profile indicate location of thin sections: 1, 9, 3, 4: lower layers; 10, 8: ditch, cesspit and associated deposits; 2, 5, 6: timber-framed house related deposits; 7: top soil. Il profilo stratigrafico esaminato: (A) Top soil; (B) lenti di sabbia (puntinato) alternate a strati organici a tessitura franca (grigio chiaro); (C) Depositi torbosi e legni; (D) Pali lignei, lenti rubefatte per scottatura e frammenti di mattone. I numeri sul profilo indicano la posizione delle sezioni sottili: 1, 9, 3, 4: strati inferiori; 10, 8: fossato, latrina e relativi depositi; 2, 5, 6: depositi relativi alle strutture abitative in legno; 7: top soil. M. Cremaschi & C. Nicosia 377 fragments). Sediments rubefied and hardened by expo- sure to fire are common, particularly on top of these lenses. In the top part, the stratification dips slightly towards the east and the loamy sandy layers (sub-unit a) increase in number and markedly in thickness. 3.1.4 Topsoil The topsoil is separated from the underlying unit by an erosional discontinuity. The topsoil consists of silty-clay with common brick and tile fragments; it has mottles (2.5 Y 4/2, dark greyish brown; 7.5 YR 4/4 brown/dark brown), and a weakly separated angular blocky structure with common voids (Sample 7 - Fig. 3). 3.2 Thin section analysis Tables 1, 2, 3 and 4 provide a brief definition of the observed fabrics and of microstructures, followed by the description of the organic and anthropogenic components, of the groundmass and of the pedofeatures for each SMT occurring in a thin section2. An outline of the main characte- ristics of each Soil Microfabric Type (SMT) identified during thin section analysis is provided below. SMT 1 : sands and silts, often showing sedimentary structures (graded fabric, lami- nations), almost completely devoid of archaeological mate- rials. It is interpreted as the cha- racterizing microfabric of Po flood deposits, not perturbed by anthropogenic intervention. It has been observed in thin sec- tions 1, 3 (lower units) and 5 (timber-framed house-related deposits). Thin section 9 (lower units) is characterised by SMT 1a, consisting of alluvial sandy- silts reworked by biological acti- vity. SMT 2: composed almost exclusively of anthropogenic materials (charcoal, mammal and fish bones, excrements, ash, vitrified phytoliths, pottery) with coarser inclusions showing strong horizontal orientation. This microfabric derived from the trampling of deposits produ- ced by domestic activit ies (hearths, food preparation, dum- ping of waste etc). The presen- ce of intercalated thin fine-grai- ned lenses and discontinuous laminations testifies the periodic reworking by waters of the exposed dwelling surfaces, pro- bably during floods. SMT 2 has been observed in thin sections 1 (lower units) and 6 (timber-framed house-related deposits). SMT 3: characterised by the admixture of reworked aggregates deriving from both alluvial layers and from occupation deposits, with components distri- buted chaotically. Interpreted as the result of the reworking of underlying stratigraphic units during pha- ses of building construction, destruction and/or refurbi- shing. Excrements and secondary phosphates deriving from cesspit reworking are common. SMT 3 has been Fig. 4 - The studied stratigraphic profile during excavation, with the alternate sequence of light- and dark- coloured layers. To the left the brick wall of the late 13th century AD house can be seen along with wooden structural elements. Il profilo esaminato durante le fasi di scavo archeologico. Si noti l’alternanza di strati a colora- zione chiara e scura. Sulla sinistra sono visibili il muro in mattoni della casa della fine del XIII secolo e degli elementi strutturali lignei. 2 Abundance of fabric units in Tables 1, 2, 3 and 4 is expres- sed according to STOOPS (2003) as follows: very few (<5%), few (5-15%), common (15-30%), frequent (30-50%), dominant (50-70%). Corso Porta Reno, Ferrara (Northern Italy): a study ... 378 observed in samples 4 (lower units), 2 and 6 (timber-fra- med house-related deposits), 7 (topsoil). SMT 3a has been assigned in thin section 6 to a sub-unit showing traces of trampling and compaction. SMT 3b derived from the weathering and bioturbation of SMT 3 and was observed in thin section 7 (topsoil). SMT 4: Silty-clay textured material with common plant remains showing strongly-expressed parallel hori- zontal orientation. Intercalated lenses that can be silty- clayey or made up by phosphates or ash. Frequent archaeological components (charcoal, bones and fish- bones, excrements). Interpreted as the outcome of sta- gnation within the ditch accompanied by the occasional dumping of waste. It has been observed in thin sections 8 and 10 (ditch, cesspit and associated deposits). 3.2.1 Lower layers (thin sections 1, 9, 3, 4 – Table 1) Thin Section 1 is an example of the sequence interpreted by GADD & WARD-PERKINS (1991) and VISSER TRAVAGLI & WARD-PERKINS (1979/80) as the alternation of dark “thread” layers and yellowish “reflooring”. Under the microscope (Fig. 5A), the sample is made up of 2- 15mm thick levels of well sorted laminated sandy silts and silty clays, sometimes showing graded fabric (SMT 1), intercalated to 2-5mm-thick trampled occupation layers (SMT 2). The presence of sedimentary structures (Fig. 5C) allows us to interpret the material labelled as SMT 1 as a series of thin alluvial levels, natural in origin, rather than as man-made reflooring layers. Only one of these layers appears devoid of sedi- mentary structures (thin section 1), possibly due to Table 1 - Lower layers: description of thin sections 1, 3, 4, 9 and associated Soil Microfabric Types (SMT). Unità inferiori: descrizione delle sezioni sottili 1, 3, 4, 9 e relativi Soil Microfabric Types (SMT). M. Cremaschi & C. Nicosia 379 Table 2 - Ditch, cesspit and associated deposits: description of thin sections 8 and 10 and associated soil microfabric types (SMT). Fossato, latrina e relativi depositi: descrizione delle sezioni sottili 8 e 10 e relativi Soil Microfabric Types (SMT). Fig. 5 - Lower layers, thin section 1. The scanned thin section (A) shows the alternating sequence of laminated alluvial sands and silty clays characteri- sed SMT 1 (C, 20x) and of occupation deposits cha- racterised by SMT 2 (B, 20x). In the latter, the strong parallel horizontal orientation of coarser constituents and the mas- sive microstructure are linked to trampling and compaction. Unità inferiori, sezione sottile 1. La sezione sottile scannerizzata (A) evidenza l’alternanza di sabbie/ argille limose alluvionali caratterizzate da SMT 1 (C, 20x) e di depositi di frequentazione caratteriz- zati da SMT 2 (B, 20x). In questi ultimi l’iso-orienta- mento delle componenti grossolane e la micro- struttura massiva sono legate al calpestio e alla compressione. Corso Porta Reno, Ferrara (Northern Italy): a study ... 380 Table 3 - Timber-framed house related deposits: description of thin sections 2, 5, 6 and associated soil microfabric types (SMT). Depositi associati alle strutture abitative in legno: descrizione delle sezioni sottili 2, 5, 6 e relativi Soil Microfabric Types (SMT). M. Cremaschi & C. Nicosia lowish (plane polarized light - PPL) aggregates, optically isotropic under crossed nichols, containing phytoliths and occasional small fragments of bone, are interpreted as carnivore/omnivore excrements (MACPHAIL 2000), possibly of human origin (Fig. 6A). Orange-yellowish (PPL) optically isotropic orthic and disorthic nodules, often associated with vivianite crystal intergrowths, are also present in thin sections 4 and 2, and can be inter- preted as Fe-phosphatic nodules (Fig. 6B). These nodu- les, the crystal intergrowths and the excrements derive most likely from the destruction or rebuilding of ces- spits and latrines. Several phases of cesspit construc- tion and refurbishing have indeed been recorded during archaeological excavations (GADD & WARD-PERKINS, 1991). 3.2.2 Ditch, cesspit and associated deposits (thin sec- tions 8, 10 – Table 2) Thin sections 8 and 10 are mainly composed of elongated plant fragments, among which leaves are fre- quently recognisable (Figs. 7A, 7B - SMT 4 microfabric). These vegetal residues show a horizontal parallel orien- tation, locally resulting in a platy microstructure. Discontinuities within plant residue levels are marked by ash-rich or phosphatic 1-2mm thick discontinuous layers (especially in thin section 8), or by continuous, laminated or graded thin silty-clay lenses (thin section 10 – Fig. 7C). The lenses most likely correspond to fine- grained mineral layers, intercalated to the organic mate- rials through deposition in a low-energy environment. We interpret this portion as the deposition in standing waters inside the ditch cut from the base of the timber- framed house-related deposits (Fig. 3). Along the sides of the ditch dumping of ashes, cess material and con- struction rubble (observed in thin section 8) took place. 3.2.3 Timber-framed house-related deposits (thin sec- tions 2, 5, 6 – Table 3) Micromorphology shows that the deposits asso- ciated to the timber-framed house consist of alternation of materials deriving from trampled occupation layers and from building destruction and refurbishing. These 381 Table 4 - Top soil: description of thin section 7 and associated soil microfabric type (SMT). Top soil: descrizione della sezione sottile 7 e relativo Soil Microfabric Type (SMT). trampling disturbance (Fig. 5A). The occupation layers with SMT 2 microfabric are rich in plant residues and in anthropogenic components. These consist of large quantities of charcoal of all sizes, burned bone, ash aggregates and vitrified phytoliths (deriving from the melting of biogenic opal - COURTY et al. 1989), sympto- matic of materials from the functioning and maintenan- ce of domestic hearths. Fishbones are widespread in all thin sections from occupation deposits, suggesting the exploitation of the nearby river Po for subsistence, even though detailed studies on the faunal remains are not yet available. The massive microstructure and the hori- zontal orientation of coarser constituents (charcoal, elongated plant material, bones) indicate that these levels were subject to trampling and compaction (Fig. 5B). Evidence of flood reworking on the exposed dwel- ling surfaces were also observed, such as rare fine- grained lenses and the sorting of charcoal fragments toward the top of each layer. Thin sections 3 and 9 are from sandy-silt layers almost devoid of anthropogenic inclusions. Thin section 3 is interpreted as a natural alluvial deposit (SMT 1), confirming the occurrence of floods during this phase. The origin of the sandy silts in thin section 9 is more dif- f icult to establish, since the sample is strongly reworked by biological activity, as shown by the chan- nel microstructure and the widespread passage featu- res (SMT 1a). Anthropogenic components such as charcoal, burned bone and excrement fragments are however very rare in this thin section. Thin section 4 is characterised by SMT 3 microfa- bric, consisting of an admixture of different materials, with a chaotic organisation of the components and no trace of trampling or compaction (Fig. 6). This microfa- bric type encompasses reworked aggregates of the occupation layers described above (SMT 2 microfabric), together with frequent subrounded aggregates of allu- vial sediments (SMT 1 – Fig. 6A). This material can thus be interpreted as the outcome of the destruction by digging and the reworking of underlying deposits. This process resulted in a chaotic mixture of aggregates and fragments from different stratigraphic units. Orange-yel- Corso Porta Reno, Ferrara (Northern Italy): a study ... 382 Fig. 6 - Lower layers, thin section 4. SMT 3 derives from the reworking of underlying deposits with inclusions of cesspit materials. In the scanned thin section (A) arrows point at the reworked aggregates from underlying alluvial layers (SMT 1 material). The occurrence of carnivore/omnivore excrement fragments (B, 40x PPL) and of Fe-phosphatic nodules (C, 20x PPL) testifies to the input of materials from cesspits. Unità inferiori, sezione sottile 4. SMT 3 deriva dal rimaneggiamento dei depositi sottostanti, con immissione di materiale derivante da latrine. Nella sezione sottile scannerizzata (A) le frecce indicano aggregati rimaneggiati provenienti dalle unità alluvionali sottostanti (materiale con SMT 1). Frammenti di escrementi di carnivori/onnivori (B, 40x PPL) e noduli ferro-fosfatici (C, 20x PPL) testimoniano la presenza di materiali derivanti da latrine. Fig. 7 - Ditch, cesspit and associated deposits, thin section 10. In the scanned thin section (A) it is possible to observe the layering that characterises the filling of the ditch. (B): layered plant remains and charcoal, characteristic of SMT 4 (20x PPL). (C): detail of one of the fine sediment lenses intercalated to the organic layers in SMT 4 (40x PPL). Fossato, latrina e depositi associati, sezione sottile 10. Nella sezione sottile scannerizzata (A) è possibile osservate la stratificazione che caratterizza il riempimento del fossato. (B): residui vegetali e carboni stratificati, tipici di SMT 4 (20x, PPL). (C): dettaglio di una delle lenti di sedimento fine intercalate agli strati a componenti organiche in SMT 4 (40x PPL). M. Cremaschi & C. Nicosia deposits bear traces of reworking by flood water flows. Thin section 2 provides another example of a layer deri- ving from digging and reworking of underlying deposits during building destruction or refurbishing. The layer shows the chaotic disposition of coarse organic and anthropogenic constituents (SMT 3) and the frequent occurrence of reworked sandy-silty aggregates of allu- vium (SMT 1 material). Destruction and refurbishing of buildings followed by occupation was observed also in thin section 6. Here, the admixture of reworked mate- rials from underlying units shows traces of trampling and compaction, such as the parallel horizontal orienta- tion of coarser constituents and a massive microstruc- ture (SMT 3a). In the same thin section, there are traces of further deposition of SMT 3 material and then again of the formation of a trampled occupation layer. Thin section 5 pertains to silty-sands interpreted as a natural alluvial deposit, as shown by the presence of internal laminations and graded fabric. The field observation of planar laminations, small ripple forms and cross-laminations fits with the microscopic eviden- ces. 3.2.4 Topsoil (thin section 7 – Table 4) This layer predates the construction of the brick house at the end of the 13th century and contains abun- dant coarse construction debris, such as brick and tiles. Thin section 7 is a chaotic admixture of materials, deriving from the destruction and refurbishing of buil- dings, and has a weakly developed channel microstruc- ture. The latter suggests that biological reworking by earthworms took place during a phase of topsoil forma- tion, during which the surface was dry and stable (SMT 3b). 4. DISCUSSION The site of Corso Porta Reno in Ferrara offered an important opportunity to investigate the modes of for- mation and accretion of an urban stratigraphic sequen- ce in a medieval town. In cities throughout Europe, archaeological excavations have often exposed depo- sits referred to as “Dark Earths” (see CARVER, 1987; BROGIOLO et al., 1988; YULE, 1990; MACPHAIL et al, 2003; GALINIÉ, 2004). Dark Earth is a term used to designate thick, seemingly homogeneous and dark-coloured deposits rich in archaeological material, generally ascri- bed to the post-Roman period. Since the early 1980s (MACPHAIL 1981, 1983), geoarchaeological and especial- ly soil micromorphological analyses have demonstrated that in many cases Dark Earth represents an once- complex deposit that has been heavily homogenised by pedological processes, especially bioturbation and weathering (MACPHAIL, 1994; MACPHAIL & GOLDBERG, 1995). Geoarchaeological analyses revealed that many diverse, superimposed activities and processes are responsible for the formation of urban Dark Earths. These correspond, for example, to the repeated dum- ping of waste (CAMMAS, 2004), cultivation (CREMASCHI, 1992), human dwelling, stabling of animals and decay of insubstantial buildings (MACPHAIL & GOLDBERG, 1995), among others. Due to human reworking, to the action of rodents, earthworms, enchytraeds, plant roots, and weathering of organic and earth-based construction materials, any macroscopic original sedimentary struc- ture and stratification has been obliterated. At Corso Porta Reno the persisting waterlogged conditions and the rapidly aggrading sedimentary environment preven- ted the deposit from being disturbed and homogenized by soil fauna or by roots. In a later phase, the presence of the topsoil (sample 7) witnesses the interruption of sedimentary processes, replaced by weathering. The whole sequence was then sealed in the 13th century AD by the construction of a brick house that remained in place until its destruction in the last war (GADD & WARD- PERKINS, 1991). All these factors enabled the original stratification of the deposit to be preserved. Macro- and micro-morphological investigation indicates that the anthropic accretion at the site is mostly the outcome of: (a) trampling of domestic depo- sits (SMT 2), generated by the activity of fireplaces as also shown by the occurrence of surfaces reddened by burning. Flood reworking occasionally took place (see below); (b) construction, destruction or refurbishing of buildings (SMT 3) marked by digging of underlying deposits (post-holes, pits, cesspits); (c) sediment set- tling inside negative structures (ditch, cesspits) produ- cing organic levels with intercalated fine-grained sedi- ment lenses (SMT 4). Within such structures occasional dumping of construction rubble, ash and cesspit mate- rial could take place, as well as the filling by fluvial sedi- ments. The natural alluvial deposits are intercalated to anthropogenic ones mainly in the phase of the timber- framed house, their thickness increasing markedly towards the top of the sequence. Alluvial deposits can be classified into two types: (a) layers of well sorted medium-fine sand or silty-sand with cross-laminated structures, deposited in crevasse splays, due to floods breaching from the Po River chan- nel (SMT 1); (b) organic-rich silt and fine sand, including charcoal, wood fragments deriving from decantation and redistribution of materials in a waterlogged environ- ment (fine-grained lenses in SMT 2). These deposits are often intercalated to anthropogenic surfaces related to trampling, recording water content fluctuation, possibly on a seasonal base. In the study area, floods are mainly expected during the autumn and early spring. These stratigraphic information fit well with the observations made during the archaeological excava- tion, which indicate that before the brick house of the late 13th century, buildings were mostly two-storey tim- ber houses, with an open ground floor. Fireplaces at the ground floor were frequent and often rebuilt and moved. The houses underwent several phases of rebuilding, refurbishing and even destruction by fire (GADD & WARD-PERKINS, 1991; VISSER TRAVAGLI & WARD- PERKINS, 1979/80). Not all of the “clean” silty and sandy- silty layers appear to be man-made reflooring layers (see § 1). Based on field and microscopic evidence, these layers could rather be interpreted as natural allu- vial deposits intercalated to occupation or “thread” layers. The presence of flood beds intercalated to occupation deposits has important bearings for the understanding of the urban stratigraphy. It shows that in cities located close to active rivers (e.g. on top of active alluvial ridges or in floodplains) part of the accre- tion can be linked to the episodic input of alluvium or to deposition and redistribution in standing waters. 383Corso Porta Reno, Ferrara (Northern Italy): a study ... Alluvium and occupation deposits will normally be subject to weathering, homogenisation by syn- and post-depositional bioturbation and human reworking (e.g. cultivation, digging, construction of buildings), unless these processes are hampered. While in need of further research, this model could partly account for the strong vertical accretion and for the formation of homo- geneous “Dark Earths” in medieval cities within active fluvial environments. 5. CONCLUSIONS The geoarchaeological investigation of the strati- graphic sequence of the Corso Porta Reno site, in Ferrara, has proven useful for both archaeological- historical and geomorphologic purposes. It elucidated the natural and anthropic processes leading to the for- mation of thick urban stratigraphic sequences. In doing so, it revealed that accretion derives from the superim- position of trampled domestic occupation deposits and activity surfaces, of chaotic layers deriving from buil- ding construction activities, of organic-rich facies for- med within negative structures with standing waters, and of flood deposits from the Po of Ferrara. Given the proximity to the alluvial ridge, the alluvial deposits, with their generally sandy texture and cross-laminated sedi- mentary structures, can be explained in a crevasse splay context, which has been observed in all the units of the stratigraphic sequence, except for unit “topsoil” (Unit A in Fig. 3, see § 3.1.4). The waterlogged nature and the strongly aggra- ding environment of the site avoided the effects of bio- turbation, allowing this exceptional sequence to be very well preserved and readable. It therefore seems clear that at the site the vertical growth of the urban deposit resulted from the interplay between alluvial sedimenta- tion close to the Po ridge and repeated episodes of human occupation, possibly with a seasonal character. The model of formation envisaged in Corso Porta Reno should be investigated for urban sites with Dark Earth- like deposits located in an active alluvial setting. The results of the study of the Corso Porta Reno stratigraphic profile, integrated with data from archaeo- logical excavations, showed that, in the early medieval period structure types, construction techniques and modes of occupation were adapted to a peri-fluvial environment. The change in the style of the sedimenta- tion towards the top of the sequence, highlighted by the bioturbated and well-drained character of the topsoil (Fig. 3), seems to be correlated with an impor- tant change in the sedimentary environment of the urban area. This corresponds to the shift from a setting subject to seasonal floods to one in which sedimentary processes appear to be somewhat under control. This latter process might have an anthropic origin, like the increased control of levees, but might also be related to the major change in the hydrography of the Po flood- plain associated with the “rotta di Ficarolo” avulsion (mid-11th century – end of the 12th century AD - VEGGIANI, 1985; BONDESAN et al., 1995). This avulsion led in fact to the progressive abandonment of the channel of the Po of Ferrara, which flowed through the city until that moment. 6. ACKNOWLEDGEMENTS For the involvement in the Corso Porta Reno excavation Mauro Cremaschi wishes to express his gratitude to Dr. A. M. Visser Travagli. 7. REFERENCES BONDESAN M., FERRI R., STEFANI M. 1995 - Rapporti fra lo sviluppo urbano di Ferrara e l’evoluzione idrografi- ca, sedimentaria e geomorfologica del territorio. In: VISSER TRAVAGLI A. M. (Ed.), Ferrara nel Medioevo. Topografia storica e archeologia urba- na, Grafis Edizioni, Casalecchio di Reno, 27-42. BRIDGE J. S. 2003 - Rivers and floodplains: forms, pro- cesses and sedimentary record. Wiley-Blackwell, Malden, MA, U.S.A. BROGIOLO G.P., CREMASCHI M., GELICHI S. 1988 - Processi di stratificazione in centri urbani (dalla stratificazio- ne “naturale” alla stratificazione “archeologica”). Archeologia stratigrafica dell’Italia Settentrionale, 1. Edizioni New Press, Como, 219-226. CARVER M.O.H. 1987 - The nature of urban deposits. In: SCHOFIELD J., LYNCH R. (Eds.) - Urban Archaeology in Britain, CBA Research Report 61, London, 9- 26. CASTIGLIONI G. B., PELLEGRINI M. (Eds.) 2001 - Note illu- strative della carta geomorfologica della Pianura Padana. Suppl. Geografia Fisica e Dinamica Quaternaria, IV, 1-204. COURTY, M. A., GOLDBERG P., MACPHAIL R.I. 1989 - Soil micromorphology in archaeology. Cambridge Manuals in Archaeology, Cambridge University Press, Cambridge. CREMASCHI M. 1992 - Caratteristiche geoarcheologiche della successione stratigrafica posta in luce nello scavo archeologico di Piazzetta Castello. In: GELI- CHI S. (Ed.), Ferrara prima e dopo il Castello. Testimonianze archeologiche per la storia della città, Spazio Libri Editori, Ferrara, 58–65. GADD D., WARD PERKINS B. 1991 - The development of urban domestic housing in North Italy. The evi- dence of excavations on the San Romano site, Ferrara . The Accordia research papers, 2, 105–127. GALINIE H. 2004 - L’expression “terres noires”, un con- cept d’attente. In: VERSLYPE L., BRULET R. (Eds.), Terres noires - dark earth. Actes de la table-ronde internationale tenue à Louvain-la-Neuve les 09 et 10 Novembre 2001, Universitè catholique de Louvain, Louvain-la-Neuve. GOLDBERG P., MACPHAIL R. I. 2006 - Practical and theore- tical geoarchaeology. Blackwell Publishing, New York. MACPHAIL R.I. 1981 - Soil and botanical studies of the “Dark Earth”. In: JONES M. & DIMBLEBY G.W. (Eds.), - The environment of man: the Iron Age to the Anglo-Saxon period. British series, 87. British Archaeological Reports, Oxford, 309–331. MACPHAIL R.I. 1983 - The micromorphology of dark earth from Gloucester, London and Norwich: an analy- sis of urban anthropogenic deposits from Late Roman to early Medieval periods in England. In: BULLOCK P., MURPHY C.P. (Eds.), Soil Micromor- 384 M. Cremaschi & C. Nicosia phology. AB Academic Publishers, Berkhamsted. MACPHAIL R.I. 1994 - The reworking of urban strati- graphy by human and natural processes. In: HALL A.R. & KENWARD H.K. (Eds.), Urban-rural con- nexions: perspectives from environmental archaeology. Oxbow Monographs, 47, Oxford, 13–43. MACPHAIL R. I. 2000 - Soils and microstratigraphy: a soil micromorphological and microchemical approach. In: LAWSON A.J. (Ed.), Potterne 1982-5: animal husbandry in later prehistoric Wiltshire, Wessex Archaeology Report, 17, Salisbury, 47–71. MACPHAIL R. I., GOLDBERG P. 1995 - Recent advances in micromorphological interpretations of soils and sediments from archaeological sites. In: BARHAM A. J., MACPHAIL R. I. (Eds.), Archaeological sediments and soils: analysis, interpretation and manage- ment, Institute of Archaeology, London, 1–24. MACPHAIL R. I., CRUISE J. 2001 - The soil micromorpholo- gist as a team player – a multyanalytical approach to the study of European microstratigraphy. In: GOLDBERG P., HOLLIDAY V. T., REID FERRING C. (Eds.), Earth Sciences in Archaeology, Kluwer Academic/Plenum Publishers, 241–267. MACPHAIL R.I. GALINIÉ H., VERHAEGHE F. 2003 - A future for dark earth?. Antiquity, 77 (296), 349–358. MURPHY C.P. 1986 - Thin Section Preparation of Soils and Sediments. AB Academic, Berkhamsted. VEGGIANI A. 1985 - Il delta del Po e l'evoluzione della rete idrografica padana in epoca storica. In: Atti della tavola rotonda: Il Delta del Po, Sezione Geologica. Accademia delle Scienze dell'Istituto di Bologna, 39–68. VISSER TRAVAGLI A. M., WARD PERKINS B. 1979/80 - Archeologia urbana a Ferrara: scavi nel comparto di S. Romano, 1981. Bollettino Annuale dei Musei Ferraresi, 9/10, 55–58. VISSER TRAVAGLI A. M. 1992 - Lo scavo nel comparto di S.Romano: analisi di un’ esperienza archeologica. In: GELICHI S. (Ed.), Ferrara prima e dopo il Castello. Testimonianze archeologiche per la sto- ria della città. Spazio Libri Editori, Ferrara, 248–259. STOOPS G. 2003 - Guidelines for analysis and description of soil and regolith thin sections. Soil Science Society of America, Madison, WI. YULE B. 1990 - The “dark earth” and late Roman London. Antiquity, 64, 620–628. 385 Ms. ricevuto il 2 agosto 2010 Testo definitivo ricevuto il 16 ottobre 2010 Ms. received: August 2, 2010 Final text received: October 16, 2010 Corso Porta Reno, Ferrara (Northern Italy): a study ...