Water and production reflections on the water supply to urban workshops in roman times

Research into water management in Roman towns has traditionally focused on its supply by means of aqueducts and its use in the thermal baths and domus of the elites. However, the urban functions of water included uses that went far beyond recreational, social activities and also included its supply to production facilities. Numerous artisanal activities required large amounts of water, including pottery, textile activities, wickerwork and manufacturing bone or metal instruments. All these activities are commonly identified in Roman towns. In each case, the workshops could have obtained water not only from wells and cisterns, but also through a connection to the urban supply network, thanks to an aqua publica concession. In this study, we compile both archaeological and textual evidence, including literary and legal texts, on the private use of aqua publica for productive activities in urban contexts.


Introduction
Studies of water management in the Roman period have traditionally focused on analysing urban supply systems, mainly of the monumental type (aqueducts). Non-urban destinations of water and other methods of obtaining water (using wells or cisterns) have been largely overlooked, as has its distribution and use in the towns themselves. This last aspect, that of water use, has only been analysed in detail in the case of the thermal baths and, more recently, in domestic areas connected to the urban network. In this way, despite a general appreciation of the importance of water for carrying out certain activities, particularly those of an artisanal nature, there have been very few reflections on the true role water played in them. 1 Though there is no doubt about the fundamental role played by water in processes like the preparation of clay for pottery production, the manufacture of the mortar essential for Roman construction or the preparation of textile dye baths, the role of water in these processes has not been studied in any depth. As such, hardly any reflections on the specific processes in which water was essential and what its function was in them are available. Neither have questions concerning the quantities of water required and where it came from been adequately addressed. Likewise, rarely have aspects such as water management within the productive facilities been taken into account -for example, whether it was possible to reuse the water in different phases of the process, and if not, whether this was because it was lost, for example, through evaporation, or because it was too dirty and had to be discarded. In this last respect, there has been little by way of systematic analysis of water drainage systems in this context.
We also receive little help from the Roman literary sources, as they generally pay scant attention to the detail of water use in artisanal processes. Suffice it to mention the total silence of the texts in relation to the activities that took place in the basins of the fullonicae, the main archaeological element documented in that type of facility.
Regardless of the purpose of the water or the amount needed, another topic seldom addressed is the provenance of this water. Workshops could have obtained it not only from wells and cisterns, but also, as shown by the archaeological record, through a connection to the urban supply network. 2 Following the aims of the workshop Water and Waterways Management in the Roman Empire, held at the University of Edinburgh, where a first draft of this paper was presented and discussed, the private use of aqua publica in productive activities in Roman urban contexts is examined here by combining both archaeological and textual evidence. Even if Roman authors generally overlooked such a mundane topic as industrial uses of water, some evidence can still be traced through literary and legal texts.

The town that produces
In contrast to the traditional position taken, for example by Finley (1973), that the Roman town is a consumer and almost parasitic in quality, many scholars now settle upon a model of a "producing town" with an active artisanal presence; more in line, for example, with the position of, among others, Toynbee (1970). In the words of Morel (1989, 253), "l'artigiano romano è fundamentalmente un uomo di città".
Sometimes the geographical location of the urban centres leads to an overrepresentation of certain activities. This is true of the salted fish and fish sauce production facilities in some of the coastal towns in the west of the Empire, such as those at the so-called "industrial quarter" of Baelo Claudia and the peri-urban factory of El Majuelo in Almuñécar (both in the south of the Iberian peninsula). As a general rule, however, a wide catalogue of productive workshops and spaces would have been present in the majority of towns of all types, the productive capacity of which would have been aimed at local consumers (Volpe 2018). 3 A good example of this may be found at Augusta Emerita, the provincial capital of Lusitania, which has been the subject of a project to study its urban artisanry and identify workshops involved in textile weaving, pottery, metallurgy, bread baking and the manufacture of glass and bone items (Bustamante 2013; Bustamante and Detry 2019).
This intensive production work would have taken up a considerable area in such towns, as is shown by the approximately 600 tabernae identified in Pompeii ( Fig. 1) (Monteix 2010a, 148). Although not uncontroversial, in recent decades various hypotheses have been put forward regarding the functional attribution of the Pompeian artisanal zones (Eschebach and Müller-Trollius 1993;Monteix 2010b;Ellis 2018). They have contributed to creating a clearer picture of the variety of productive activities that took place in towns: the production of olive oil, perfumes, wine or salted fish; workshops devoted to leather tanning, basket weaving and metalworking; bakers; different textile-related activities, including wool washing and dyeing and the treatment and washing of fabrics, etc. There would also have been different types of market gardens cultivating flowers, fruit trees, grapevines and vegetables (Monteix 2010a(Monteix , 2017, to which we can add other activities identified in the immediate peri-urban area, such as pottery making.

Urban artisanry as a consumer of water
Although water was essential in many of the aforementioned activities, the role it played in each of them would have varied enormously. In the manufacture of pottery, for example, water was necessary in preparing the clay. The raw material extracted from the quarry was mixed with large amounts of water until it formed a liquid paste. The components with a larger particle size, which would have hindered the modelling of the pieces and subsequently caused them to crack when fired, were deposited at the bottom of decantation basins. Once free of those elements, the liquid clay was left to stand until enough water had evaporated to turn it into a mouldable dough. 4 We know the ancients were aware of the large amounts of water they used from pottery workshop rental contracts in Roman Egypt, under which -in addition to kilns, potter's wheels, storage space and different types of clay -the owners were obliged to guarantee water in the cisterns (Mayerson 2000).
Textile manufacture also needed large quantities of water. 5 If we focus on wool processing, the furnaces found in the so-called officinae lanifricariae 6 suggest that the procedure began with washing the fibres in tepid or warm water. The water removed any dirt stuck to the lanolin, the natural oil found in the skin of sheep. Water was also essential for dyeing the fibres, a process that could include three different phases: whitening, mordanting (needed when using certain dyes) and finally the dyeing itself. These were all actions involving immersion in baths in which different components were diluted. Pliny (Nat. Hist. 19,49) recommended the use of soapwort root macerated in water for whitening, although other sources, such as the Holm. 112 papyrus, describe washing with ashes accompanied by rinsing in water. For the mordanting, the fibres were submerged in a solution of water and different elements of mineral or biological origin, among them urine, nutgall and aluminium or copper salts. Finally, the dyeing was carried out by submerging the fibres in dye baths obtained by macerating the colourant substances in cold or lukewarm water, decocting (in boiling water) or fermenting them, using reductions in alkaline agents (for example, urine or lime) (Martínez García 2011, 190-4;Martínez García 2016, 2-6).
Perhaps the textile activity most commonly associated with the use of large amounts of water is that of the officinae fullonis: that is, the establishments whose business was the washing of fabrics (Fig. 2). 7 In them, following a first phase of lathering with urine and fuller's earth, to which a small amount of water was added (Flohr 2013b, 104), the fabrics were rinsed in the omnipresent basins that are always associated with this type of facility.
Leather tanning was also in need of large quantities of water. First the skins were washed, an immersion that in the case of previously dried or salted skins could last for several days, during which the water had to be renewed periodically. Once clean, the skins were submerged in hot water, to speed up the natural decomposition of the hair, or in basins in which different substances were fermented (rye bran, urine, excrement). Following the mechanical removal of the remains of adhered hair, flesh, and fat, the skins were rewashed before tanning. This consisted of a new immersion, in this case adding alum or holm oak bark powder, and a final rinsing with clean water (Brun and Leguilloux 2014).
Another productive activity in which water was fundamental was food production. The preparation of salsamenta (preserved fish), for example, began with the washing of the fish (Trakadas 2015, 10-11), while in the preparation of garum or liquamen, type fish sauces, it was also necessary to add a certain amount of water to the mixture of spices, salt, and fish ( Fig. 3). 8 Water was also an essential resource for baking bread that, by definition, is made by mixing flour and water, with the end result depending to a large extent on the quality and quantity of the latter ingredient, as we can see from Pliny's explanation of panis aquaticus or parthicus (Pliny, Nat. Hist. XVIII, 27).
The plan published by Monteix (2017, 218-9, Fig. 7.2) of the production zones identified in Pompeii, as well as different workshop types, also shows that there were large areas devoted to cultivation inside the town walls. Flower gardens, vegetable patches, orchards of fruit trees, and vines all needed irrigation infrastructure. One example is that of the cistern excavated in the vegetable garden of the House of Pansa (VI 6, 1) (Jashemski 2017), from which the different furrows would have been irrigated by flooding the earthen channels that bordered them.
We end this brief tour with another activity that is not linked to the workshops, but that in recent years has become one of the fields studied in the archaeology of production: the manufacture of the different types of mortar used by the Romans, which also required large 8 For a general reflection on the role of water in halieutic production, consult Sánchez López (2018). [Down] Water channel through the facilities. Photos by the author amounts of water (Vitruvius II, 5) 9 . In this case it was necessary to slake the lime, which involved mixing the quicklime (obtained by cooking limestone in a fornax calcaria) with water. This process could either have been carried out at the building site itself or somewhere else, as appears to be shown by the presence of an amphora full of lime at the Casa della Soffitta in Pompeii (V 3, 4), which was under construction when Mount Vesuvius erupted (Adam 2002 , Fig. 160). The lime was then mixed with different types of caementa to make the various kinds of mortar (stucco, opus caementicium, opus signinum), which also could have required the addition of certain amounts of water. To this we have to add the water used to spray the mortar during hardening to prevent the appearance of cracks.

The origin of the water used in urban workshops
The water needed for these and other productive activities identified in Roman towns likely came from diverse sources. Probably workshops would have installed their own systems for obtaining or storing water. It could have come from wells that took advantage of the water available in the subsoil or cisterns used to store rainwater runoff from building roofs. This can be seen in the cisterns documented at salted fish and fish sauce production sites, such as Industrial Complex IV at Baelo Claudia (Spain) (Bernal et al. 2007) or the cetariae of Nabeul (Tunisia) (Slim et al. 2007). As for wells, these have been discovered at factories like those of the Teatro Andalucía in Cádiz (Spain) (Bernal 2005, 141), here in addition to the cisterns.
Among the Pompeian examples of facilities with this type of infrastructure, the tannery at I 5, 2 possessed a well that over time was fitted with a waterwheel (Bruun et al. 2010, 272-274). The water it raised was channelled via different conduits to the basins and dolia distributed around the workshop (Brun and Leguilloux 2014, 152). On the other hand, the potteries in Via dei Sepolcri 29 and Via Nocera (I 20, 1) obtained their water from cisterns (Cavassa et al. 2014;Peña and McCallum 2009).
In other cases, however, it has been hypothesised that the water was brought in from outside the workshop. This was the case of the officinae infectoriae (wool-dyeing facilities) in Pompeii studied by Borgard and Puybaret (2004, 55-56). They suggested that the workers at the establishment at I 8, 2.19 could only have obtained water from a cistern located in the neighbouring house, transporting it in buckets or pottery containers. Meanwhile, the workshops at V 1, 4 and V 1, 5 probably obtained their water from a public fountain in the street.
One final option was to establish a connection to the urban aqua publica 10 distribution network. Although this system is difficult to identify at many archaeological sites, because the lead fistulae have rarely survived, it was the most common solution used in another of the textile activities identified in Pompeii; that is, fabric washing in the officinae fullonis. In fact, all but one of the fullonicae studied to date had direct connections to the urban water supply system (Bogard and Puybaret 2004, 56;Flohr 2013b, 137). This option has also been identified in Pompeian workshops involved in other activities, such as the pistrinae at premises VII 12, 13 (Dessales 2013, 237) and IX 3, 19-20 (Monteix 2010c).

The written sources and the productive use of aqua publica
Despite the archaeological evidence for the supply of water from the urban distribution network to different types of artisanal establishment, Vitruvius (De Arch. 8.6.2) only mentions three destinations for the water channelled into the town by aqueducts: public fountains (lacus et salientes), baths (balneas), and houses (domus privatas). In contrast, the productive uses of water from aqueducts are briefly mentioned in the discussions concerning the seruitus aquae ductus in the Digest 11 , though these mainly refer to uses associated with rural contexts, such as irrigation or stockbreeding (D. 43.20.3, Pomponius, 2nd c. AD). Even so, Ulpian (early 3rd century AD) specified that the considerations made were also applicable to water distribution in urban areas (D.43.20.1.11), where water use has mainly been associated with leisure in recent historiography. According to Saliou (1994, 135-136) -based maybe on Ulpian's affirmation that the amoenitas were among the utilitates of the water transported by a conduit (D. 43.20.3) -the big difference between the rural and urban use of water is the type of need satisfied: productive in the first case and domestic (citing as an example a nymphaeum) in the second.
As it stands, the majority of current researchers link the urban distribution of water -for uses other than fountains and baths -largely to domestic contexts (e.g. Saliou 1994, Ronin 2015, which draws attention away from the productive uses of water in towns. This position is not only in line with the Vitruvian sentence, but also with epigraphic texts such as the inscription of Thysdrus (El Djem, Tunisia) (CIL, VIII, 51 = D. 5777) dated to the second half of the 3rd century AD, which specifies that under certain circumstances water could also be distributed to some houses (

[Aqua] (…) domibus etiam certa condicione concessa (…)).
In the majority of cases, however, Frontinus (1st c. AD) uses more generic terminology to refer to water concessions in Rome. Here, the use of the word privatus indicates that these were not necessarily linked to the supply of houses. 12 Take, for example, the priuatorum usibus of the water of the Aqua Alsietina (De Aq. XI, 1), the privati that ex riuis publicis aquam ducere (De Aq. CVI, 1) or the privati to which, according to his data, 38% of the water distributed by the aqueducts corresponded (De Aq. LXXVIII, 3). 13 The same is true of the Lex Ursonensis (the law of the colony of Urso, present-day Osuna (Spain), dated to the time of Caesar) with the aquam in priuatum (…) ducere (Lex Col. Gen., C) that does not appear to be linked in any case to a use that can be considered exclusively domestic. This is 11 In light of the reflections on this point, I will combine information contained in texts referring to different towns and periods, since the comparison of literary and juridical texts (either contained within the Digest or transmitted by some epigraphic medium) can help shine a certain amount of light on the uses of water in general in the urban contexts of the Roman period. In fact, except for certain minor differences, such as those linked to the identification of the actors involved (the Emperor or the Senate in Rome and the duumviri or ordo decurionum in the provinces), the water management strategies, as well as the applicable regulations, appear to be consistent throughout the whole territory (Gerez Kraemer 2020). 12 A term he uses, for example, when -referring to Marcus Agrippa's management of the aqueducts of Rome -he divides the destinations of the water transported by aqueducts to the city into three: public infrastructures, fountains and private individuals (De Aq. XCVIII, 2); rather than to fountains, baths and houses as done by Vitruvius (De Arch. 8.6.2). 13 According to Bruun's calculations (2000, 588), based on the figures given by Frontinus, 44% of the water brought into the city of Rome by aqueducts would have been for public use, 17% for the needs of the emperor (including sometimes public baths and military barracks) and 38% for private use. We would consider this last percentage too high if it really only referred to the supply to the domus, bearing in mind that such concessions only included in domos principum ciuitatis (De Aq. XCIV, 5). even true of the priuatos usos of water in the Novellae of Valentinian III in the 5th century in Constantine (Numidia, North Africa) (Nov. Val. VIII,9). 14 In a similar vein, Frontinus mentions -although within his commentaries to the Aquis discourse by Marcus Caelius Rufus (1st c. BC) -the connection to the urban network of agros, tabernas, cenacula etiam (De Aq. LXXVI, 2), subsequently complaining of the illegal connection of a multitude of negotiationes (De Aq. CXV, 3). These are areas and establishments that could easily be included among those users and uses defined as private, some of which we can link directly to the production areas we are interested in here.
There is little doubt that the reference to inriguos agros (flooded fields) (De Aq. LXXVI, 2) recalls the technique identified by Jashemski (2017) for irrigation using furrows separated by irrigation channels as could be found in urban gardens producing, for example, flowers, fruit, grapes, and vegetables for sale) (Fig. 4). This agricultural use of urban aqueducts has been proposed by several scholars (Wilson 1999;Bannon 2009;Kamash 2010;Sánchez 2015) who consider that the cost of having to pay for access to this water would have been offset by increased production (Wilson 1999;Bannon 2009). These hypotheses concerning irrigation and aqueducts, however, have always been made in connection with erogationes extra urbem and not the cultivated areas identified in the urban framework of towns such as Pompeii.
For its part, the concept of taberna is currently associated with a well-defined archaeological structure of a quadrangular ground-floor room with a wide entrance that opened directly onto the street front and was mainly devoted to marketing (Holleran 2012; Ellis 2018). However, this definition may not coincide exactly with the original concept we find in the sources. The ancient texts describe a wide variety of functions for these premises, including -in addition to their function as shops -administration, dwellings, production, and services (Holleran 2012, 113-158). The identification of tabernae as workshops has been confirmed archaeologically by the presence of specific structures including kilns, ovens, furnaces or basins. The sources also refer to tabernae ferrariae 15 (D. 31.1.88.3), pellesuinae 16 (Var. L. 8.55) and sutrinae 17 (Tac. Ann. 15.34), among others. In fact, one of the main production areas in a Roman town was the taberna, which often functioned as both a workshop and a shop (Holleran 2012, 121-124). Other types of establishment categorised by Holleran (2012) as offering different services can be associated with the term taberna. They include the workplaces of barbers or doctors, as well as bars, fullonicae and bakeries (although these last two are often associated with premises that are much larger and more complex than those archaeologically linked to the term taberna).
It is to one of these latter activities, the washing of fabrics once they had been spun and woven, and the laundering of used clothes (Flohr 2013) in the officinae fullonis, that Frontinus connects the use of aqua publica in the town. According to what we read in the text on the management of the aqueducts of Rome, during the Republican period the only activities authorised to draw excess water from the public fountains were balnea and fullonicae (that, moreover, are described as privatus or private supply) 18 (De Aq. XCIV, 3-4): a use that 14 For a reflection on the management of the water concessions in Valentinian's Novellae, see Ronin 2018. 15 Blacksmiths. 16 Leather workshop. 17 Shoe factory. 18 The testimonies to the direct relationship between fullonicae and the water from a fountain (fons) are not limited to Frontinus' text, but are also found in the Digest (Ulpian, Dig. 39.3.3. pr.) and, according to some scholars, in the so-called Lis Fullones (CIL VI 266) dated to the mid-3rd century AD, specifically to 244 (Tran 2007). explicitly differentiates them from the supply in domos principum ciuitatis (De Aq. XCIV, 5). The term used by Frontinus to refer to the water that overflowed from the fountains, aqua caduca 19 , is also used in other sources -including the aforementioned Lex Colonia Genetiua Iulia Ursonensis -when speaking of the private derivation of aqua publica by the colonists (Si quis colon(us) aquam in priuatum caducam ducere uolet (…). Si decuriones m(aior) p(ars) qui tum atfuerint, aquam caducam in priuatum duci censuerint) (Lex Col. Gen.,C,(9)(10)(11)(12)(13)(14)(15)(16).
Three rather unusual fountains in Pompeii have also been linked to the use of excess water. They are different because they do not have a receptacle for the water to pour into, but rather a stone slab fitted vertically into the pavement with an orifice to support a lead pipe (Dessales 2008(Dessales , 2011 To be more specific, the fountain at V, 1, 3 has been linked by H. Dessales to another nearby fountain, from which it received the overflow and which would have supplied a neighbouring dye workshop (located at V 1, 4) (Dessales 2008, 62).
Except for the references to the Republican period, however, Frontinus' treatise does not link the concession of water to private individuals with aqua caduca 21 , but rather with the derivation of water from the urban distribution network through pipes that would not have been connected directly to the main pipes of the network, but to the castella 22 (De Aq. CVI, 1) 23 . Moreover, that concession must have been authorised by the competent authority, which in the case of Rome, would have been the emperor (De Aq. CIII,2;Ulpian D. 43. 20. 1. 42). In the provinces, the local authorities appear to have been in charge of water, or at least we can glean as much from the Edictum Augusti de aquaeductu Venafrano (CIL X, 4842) that names the duouiri as being responsible for authorising concessions for water. In the case of the Lex Ursonensis, the only specification in this respect refers to aqua caduca, specifying in its Article C that the application must be made to a duouir, who would present it to a meeting of the decurions responsible for taking the decision.
The question of payment for these concessions is particularly controversial. Both Vitruvius (De Arch. 8.6.2) and Frontinus (De Aq. XCIV, 4; CXVIII, 1) refer to the payment of a vectigal; however, it has long been debated whether that tax was for the maintenance of the conduits or payment for the water (Bruun 2000;Maganzani 2004;Ronin 2015, 70). Some researchers have suggested that, at least in the city of Rome in the Imperial period, water would have been free of charge for those citizens who used it for amoenitas, while artisans and all those who "made business" with it would have been the only ones obliged to pay the vectigal (Malissard 1996, 286;Bruun 2000, 589). The situation could have been different in less wealthy provincial municipalities, where they would have been obliged to charge for the water (Malissard 1996, 286), as appears to have been the case in Venafrum (CIL X, 4842, 4). 19 We understand from Frontinus' text that in Republican Rome aqua caduca was the water that overflowed from the fountains (De Aq. XCIV, 3). In the Imperial period the meaning of the term appears to have been expanded to include water overflowing from cisterns (castella) and leaks from pipes (ex manationibus fistularum) (De Aq. CX, 1-2) (Dessales 2008, 55). 20 Located in VII 15, 12; VIII 7, 25; V 1, 3 (Dessales 2008, 60-62). 21 During the Republic, concessions for the use of aqua publica were only envisaged in relation to aqua caduca or surplus water. This changed in the Imperial period when the diversion of water from the aqueducts was allowed, even before they reached the town. 22 The castellum should be understood as a reservoir supplied by an aqueduct and the cisterna as a deposit for storing rainwater. 23 A link between the concession of water and deposits maintained by Ulpian in the Digest (D. 43.20.1).

Some final reflections
Despite the aforementioned archaeological evidence and possible references in the classical texts to the supply of urban artisanal zones from the water distribution network, the historiography has almost exclusively linked the private use of water to domestic supply (essentially for balnea and water features in courtyards and gardens) -even in those contexts with evidence of uses that would be difficult to define as domestic. A good example of this is the connection of some tabernae in the Pompeii area to the intake of the domus of which they were a part. We can see this, for example, in the fullonica of the Casa della Regina d'Inghilterra (Pompeii VII 14,5.17), where the fistula that supplied the house also had a branch that led to the workshop (Dessales 2013, 236), and in one of the tabernae in the Casa di Nettuno e Anfitrite (Herculaneum V 6-7) (Fig. 5). Here, the commercial establishments of Taberna 6 were supplied from a fistula that entered through the intake of the dwelling and bifurcated towards the caupona (Dessales 2013, 236), where it appears to have supplied a water heating facility.
If it is true that in Rome water for domestic use was free, while artisanal use was subject to the payment of a vectigal (Malissard 1996, 286;Bruun 2000, 589), the above-mentioned examples could have constituted a fraud that we can relate to Frontinus' complaints about the discovery of agros, tabernas and cenacula etiam (De Aq. LXXVI, 2) connected to the urban network.
In any case, it is clear that, as today, water was an essential resource for carrying out numerous artisanal and productive activities. It was indispensable for refining clay and turning it into a mouldable dough; for washing and dyeing fibres and fabrics; for treating animal skins and extracting the lanolin used as the basis for manufacturing perfumes and unguents; for tempering iron; for manufacturing mortars; for re-hydrating olive residue and obtaining olive oil for lamps. Therefore, as with the rest of the raw materials, the artisans would have had to plan for its supply, which was further complicated if provision had to be made for transport and storage. Clay, sand, wool, wood or animal skins could be transported relatively easily on the backs of pack animals, in carts or by boat, secured by ropes and perhaps tarpaulins, and then simply piled up ready to be used in the workshops. . Lead pipes connected to the urban water distribution system were used to supply water to the domus. Once in the house, a branch supplied the taberna located in the front of the house. From Dessales 2011, 151, Fig. 6 Water, however, required more sophisticated infrastructure. Its supply, often in very large amounts, as reflected by the size of the basins used for rinsing the fabrics in fullonicae, would have depended on digging deep wells, building cisterns, installing pipes or the laborious task of transporting it from nearby fountains in buckets or pottery vessels. In these last two cases, its management within the workshop would also have included the construction of deposits in which to store the water transported in that way. We should not forget that this use of aqua publica was subject to the payment of the vectigal, which would have increased the operating costs of the businesses. Perhaps this explains the establishment in some Pompeian fullonicae of operational models that made it possible to reuse the water. Thus, in keeping with M. Flohr's proposal (2013, 144), in the fullonicae with several basins, those with the cleanest water were nearest to the workshop's only water intake point, while those in which the water was dirtiest were at the far end of the chain, near to the drainage point. The rinsing process would therefore have been carried out in the last of these basins, which were also fitted with work surfaces to facilitate the task of removing the detergents, fat, and other types of dirt. The fabrics would have passed progressively towards the basins with the cleanest water, where the procedure would have consisted only of a light final rinsing.
This planning of the water supply must have affected the siting of some of the production areas. Thus, for example, for Pompeii, N. Monteix (2010a, 157) has proposed that the concentration of production areas with the highest water consumption along the axes of the Via Vesuvius and the Via Stabiana (the town's best supplied sector, owing to the high proportion of castella) can be explained with reference to their water needs. Water needs for productive activities may have been important elements in Roman urban development.