Settlement limits pose more complex issues than commonly assumed, a point well illustrated by Goodman's (2007) study of Roman urban
Peripheries. In cases she examines, city walls and orthogonal grids suggest urban limits, yet peri-urban facilities are common. Settlement limits are particularly elusive for dispersed occupation at lowland Mesoamerican centers, where the amount of green space is magnified (cf. Magnoni et al., Chapter 5 In this volume), but dispersed occupation occurs more widely. In Mesoamerica, Graham (1999) contrasts lowland "green cities" to "stone cities." In the lowlands, occupational remains may decrease in density with greater distance from a center, continuing into the countryside. Consequently, the peripheries of a settlement have even more open space than the central settlement, and settlement limits become problematic. Much of the inter-residential space within low-density lowland settlements is assumed to have been cultivated in some fashion (Stark and Ossa 2007).
BARBARA L. STARK
The extent to which Mesoamerican lowland settlements can be characterized as low-density urbanism has been challenged (Smith 2005:412). Smith (2005:412) shows that the degree of contrast between the Mesoamerican lowlands and highlands is exaggerated for the Postclassic period. Smith (2005:412) expressed doubt about the reality of greater residential dispersal in lowland Mesoamerica, based on his calculations for population densities in Postclassic Mesoamerican cities because three lowland settlements (Santa Rita, Mayapan, and Naco) fell wIthin the ranges of highland low-density settlements.
The three Postclassic lowland settlements do not afford data consonant with earlier periods, however. Calculations for Classic-period lowland settlements show instances of a considerably lower density than Smith encountered in his late sample, and they also show that densities can vary considerably between the denser core of settlements anD the outlying sectors (Table 11.1) (Culbert and Rice 1990). As noted earlier, calculations of open space versus architectural space could provide a more reliable physical comparison of settlement densities than interjection of demographic calculations, but population densities are the data available.
ANCIENT OPEN SPACE, GARDENS, AND PARKS
Tikal combines residential dispersal wIth delimitation of the settlement by earthworks and ditches (Puleston 1983; Puleston and Callender 1967). The more extensively documented northern earthwork and ditch were first interpreted as a defensive boundary extending between the wetlands located to the east and west. These findings proved problematic with restudy (Webster et al. 2004, 2007). A proposed southern earthwork and ditch was not found; the southeastern earthwork was accompanied by a parallel formation slightly farther in toward the center - a double-ditch alignment; the northern feature proved to be mainly a ditch; a western ditch was detected. Inconsistencies in the depth of the ditches and the height of the earthwork (often absent) cast considerable doubt on the original defensive interpretation, particularly as attacking forces would meet no obstacle to the south. Nevertheless, the constructions are "something" and suggest an emic delimitation, perhaps unfinished. From these Maya examples, we see instances in which no obvious boundary was detected (Dzibilchaltun), a clear surrounding wall did not encompass all the residences associated with the center (Mayapan) (see also Becan [Thomas 1981]), and population densities varied according to a delimiting feature (Tikal).
The Early Classic capital of Cerro de las Mesas in Veracruz offers yet another situation. There, areas without residential traces -"reserve spaces" (presumably green in that environment) - are scattered around the monumental core at a distance of 2 to 3 km (Figure 11.5). Although these spaces require future evaluations, for the moment they are considered a transitional or delimiting feature, somewhat akin to the ditches around Tikal. Residences occur beyond the reserve spaces, but the pattern of spaces seems to represent an effect of planning, as they remained unoccupied by domestic mounds during the Classic period. The reserve spaces in Figure 11.5 Are marked arbitrarily because we do not know the allocation of exterior space around domestic mounds in the vicinity, nor whether subsequent occupations may have encroached on the reserve space. In addition, the survey did not locate occupation north oF the Viejo River, which may also contribute to the peripheral space delimIting Cerro de las Mesas. Comparative data, discussed next, provide a broader context for understanding peripheral-urBan green space.
BARBARA L. STARK
Figure 11.5 Reserve space around the monumental core of Early Classic Cerro de las Mesas. Dashed line shows transition zone of possible settlement limits. Zapotal is predominantly Late Classic, and Sauce is a Middle Postclassic town (drawn By author).
Comparative Examples of Urban Peripheral Gardens
Peripheral palatial or other institutional gardens are relatively common in ancient states and are understandable in part through a historical view of land use, city growth, and land values (Table 11.2). They are only one of a variety of open spaces and associated activities near urban peripheries; for example, Kostof (1992:130-132) notes markets, sports, and other functions (see also Goodman 2007 for the western Roman Empire). A concern with city peripheries has a counterpart in modern contexts. Today, urban planners and other scholars use the notion of a peri-urban transItion area, or urban fringe, to examine multiple effects of urban centers: demand for nearby resources that may lead to agricultural, watershed, or soil changes; disposal of by-products and waste; the expansion of cities with growing populations, especially through in-migration; changes in nearby smaller (rural) settlements; and counter-urbanization as wealthy individuals
Algeria, Fes, 19th-20th centuries
Great houses and gardens predominantly on the periphery (Revault et al. 1992:363).
Assyria, Assur, Nineveh, 704-681 BC
Temple and garden built outside wall of Assur; four extramural royal gardens or "plantations" at Nineveh (Dalley 1993:6; Foster 2004:215).
Austria, Vienna, AD 1683-1720, Beidermeir AD 1812-1848
Nobles had to have a permanent residence in Vienna, and palaces were built outside the walls of the city. During this period, bourgeois building construction remained restricted to the outlying areas of the suburbs (Rotenberg 2008:118).
Aztec empire, Late Postclassic period, Tenochtitlan, Texcoco, AD 1350-1519
Royal gardens and hunting parks, most within 5-12 km of the capital (Evans 2000:209-211).
Byzantine, Thessalonike, Constantinople, AD 1204-1453
Monastery gardens inside and outside the city; suburban Constantinople villas along the sea and outside the city walls; area between two city walls included gardens, tombs, and monasteries (Constantinides 2002:87-89; Talbot 2002:61; Mango 1985:47, 49).
China, Song, and Yuan Dynasties, AD 960-1126, AD 1279-1368
Loyang, gardens on the outskirts (Lifang and Yu 1986:16).
Kaifeng, Northern Song Dynasty, 20 gardens in the city outskirts, with flower, recreational, and residential qualities (Chen and Yu 1986:16; Hammond 2008:45); environs of Kaifeng replete with gardens and orchards, with suburban villas, feasting halls, monasteries, and nunneries (West 2005:297, 298).
China, Suzhou, Jiangnan, Loyang, Ming Dynasty, AD 1450-1650
One of four gardens in gazetteer was within the walls, another on the outskirts, and the other two inside and outside the walls of Kunshan, a smaller subordinate city (Clunas 1996:16); one owner had a garden inside and another smaller one just outside the Loyang city walls; elders built gardens inside walls on vacant land, but some were outside (Hammond 2008:46, 47); survey by Ch'I of Loyang counted 191 in the county, 74 in the city and the rest outside the city (Smith 1992:67).
Dutch, AD 1650-1702
Banlieue became a transitional territory between town and country, near town walls, where kitchen gardens and orchards were rented or purchased by the economically less viable (de Jong 1990:29).
England, Birmingham, also Vienna, Austria, and Paris, France, 20th century
Fringe belts with extensive land use at the edge of an urban area, formed when the city was not growing, with recreational areas, public utilities, allotment gardens, sports clubs, and other institutions (Whitehand and Morton 2006:2048).
England, Medieval
Deer parks on the urban fringe formed, in effect, suburban green belts, in some cases arresting city growth (Creighton 2009:158).
Genoa, Republic of, 16th century
Lavish villas outside city walls, daily commute to city dwellings (Magnani 2008:55).
Greece, Classical
Delos gardens outside city, but elsewhere at Tegea, scattered among houses; Athens gardens
Associated with town periphery, men of property (Osborne 1992:377-379, 381); Athenian educational gymnasia outside of city in garden and grove areas (Carroll 2003:29, 50-52).
India, Mughal Empire, AD 17th and 18th centuries
Suburban gardens and residences for early ruler at Agra, not in citadels (Wescoat 1992:336);
Shahjahanabad had extensive suburbs outside the city wall with gardens, important tombs, markets, and inns (Blake 1991:57-66).
Iran, Isfahan, 17th-18th centuries
Royal palace of Farahabad built in suburb ca. 10 km from the city center by last Safavid Shah, with two earlier rulers also building their own suburban palaces at Isfahan (Brignoli 2007:144); Ettinghausen (1976:7) notes some elaborate gardens in the Islamic tradition were in suburbs or the countryside.
(continued)
Italy, Rome, Renaissance
Rome included area within the walls and adjacent suburban land outside the walls; surrounding land included villas with gardens of people whose activities centered in Rome; elite villas mostly within about 35 km of Rome, but up to 70 km (Coffin 1979 Especially vii).
Northeastern United States, 19th century
Use of rural cemeteries by public as parks, also informal open spaces just outside the developed area of cities (Low et al. 2005:20, 21).
Ottoman, Istanbul, 18th century
New lavish palatial gardens along suburban banks of the Bosphorus and Golden Horn, with return of court of Sultan in 1703; old suburban imperial gardens declined as foci of court life, with various outcomes, disuse, renovations, but some made into gardens for the wider public (Hamadeh 2008:91).
Roman Empire, Rome, 1st century BC-AD 1st century, Pompeii, AD 79
Many wealthy, influential Romans built private villas just outside Rome (Carroll 2003:58); at Pompeii, elite gardens were concentrated toward the southwest edge of the city (Ciarallo 2001:39).
Seljuk Alanya, Turkey, AD 1221-1250
Winter retreat and later de facto capital at Alanya, with seven or more outlying gardens and palaces (also tent pavilions), some royal, some occupied by emirs. Gardens range from 1.9 to 5.9 km from the citadel, mounted travel used. The court, with most emirs in attendance, was seasonally mobile in warfare (Redford et al. 2000:24, 27, 31-32, 40, 42, 54-55, 69).
Sri Lanka, Kandy Kingdom, 18th century
Religious buildings, including temples, built on outskirts of Kandy, also royal garden in suburb 5 miles southwest of capital (Duncan 1990:76).
Yoruba, Ile-Ife, 1388 AD-
Land between the two town walls of Ile-Ife was used for gardens, hunting, wood lots; beyond the outer wall, sacred groves merged into the town farms. Sacred groves occurred within the first wall, the second wall, and outside the second wall, but mainly at the city edge, in part owing to constraints on land (Falade 1984:29, 31, 32, 36).
Use modern transport to reside outside the urban area even if they work there (Simon 2008).
Peripheral eLite gardens attached to elaborate residences as well as other peripheral green spaces are documented in several historic cases, even though urban garden location is seldom the main focus of the landscape and gardening literature (Table 11.2). At Pompeii, for example, elite gardens were concentrated toward the southwest edge of the city (Ciarallo 2001:39). Royal gardens were part of a settlement boundary at Bianliang (Kaifeng) in Song Dynasty China (West 2005); elite gardens also were on the outskirts of Loyang (Lifang and Yu 1986:16). In Renaissance Rome, several elite villas were located outside the city walls, indicating that aristocratic estates may be removed from the denser, more obvious settlement, yet are a functional part of the city, with owners moving between or among residences (Coffin 1979).
Because of land values and land tenure, space for palatial grounds is more readily available at the edges of cities. As Revault
Et al. (1992:363) state for Fez, Morocco, in the nineteenth and twentieth centuries, new palatial residences and grounds frequently were located marginally because of the densely built central city and the difficulty of acquiring appropriate land for a lavish and extensive property. Brignoli (2007:147) remarks on similar factors in respect to the location of a Safavid royal palace. Certainly some demolition and redesign of central space can occur, especially if initiated by powerful rulers or governments, but this process of conversion does not afford the only solution because peripheral locations can be selected.
ANCIENT OPEN SPACE, GARDENS, AND PARKS
Palatial garden estates are only one of the space-hungry facilities that may have gravitated to a peripheral location. Constantinides (2002:87-89) notes the movement of Late Byzantine monasteries from the countryside to cities for security reasons, often to city margins to accommodate their orchards and gardens (see also Talbot 2002:61). Hunting parks may be maintained outside cIties in locations where environmental conditions are suited to game (Constantinides 2002:96; Tabbaa 1992:305; Williamson 1992:74-75). In Aztec times, royal or palatial gardens were found within the capital near the center, but also outside at considerable distances to take advantage of particularly striking or symbolic natural locations or different climate zones (Evans 2000:209-211). Other provocations to establish outlying estates include respite from city life (unconvincing for low-density urbanism) and refuge from city plagues (for Constantinople, Auzepy 1995:360; for Renaissance Rome, CoFfin 1979:9, 84). Thus, extensive gardens or parks can occur scattered into the countryside and present a mosaic of distances despite the fact that the owners who utilize them are city residents.
Some peripheraL gardens have a food-production emphasis rather than social ostentation, a function incorporated in land-use models. Variation in land use with distances from a town motivates von Thunen's (1966 [1875]) isolated state model in which transport costs, market values of products, and cultivation frequency vary in "rings" around a settlement (see summary and discussion of modern contexts at different scales in Bradford and Kent [1977:28-41]). In von Thunen's model, more-distant land costs less to obtain and can be profitable with lower economic yields, but transport costs may negate this extra income, depending on the crop. Intensity of cultivation is adjusted to the distance and transport costs, usually with less-intensive regimes at a greater distance, but modified according to
The weight oF the crop. On the outskirts of Greek city-states (Osborne 1992:381), intensive irrigation gardening of foods reflected von Thunen-Like principles of land use. As Renaissance Dutch merchant families sought country estates, individuals of lesser means rented or bought gardens or orchards on the edges of towns (de Jong 1990: 29).
A more recent expression oF the phenomenon of peripheral green space, land costs, and access involves urban fringe belts. The fringe belt concept dates from 1936 in a study of Berlin by H. Louis and was then emphasized in the work by geographer Conzen (Whitehand 2001:105). Whitehand and Morton (2006:2047) point out that urban fringe belts derive from peripheral installations with associated open space (e. g., sports, health, or educational facilities) that are engulfed by leapfrog urban growth. They note that fringe belts typically have an intermittent or mosaic spatial form, which provides one morphological model for urban peripheries. Once established, despite urban growth, the peripheral uses can be dificult to replace with higher density uses, so the fringe belt remains.
Thus, peripheral open space responds to a complex set of factors. Land costs, claims on land, governmental regulation, and the history of building in the urban settlement, as well as other factors, can conspire to provide incentives for a variety of more spatially extensive land uses at the city margin. Some of these uses may inger out into the countryside, creating an indistinct or mosaic city margin. This possibility has been scarcely considered in archaeological practice. Suficient spatial separation of architectural groups from the core of an urban settlement has normally been grounds for deining a different site. As noted, comparative data concerning peripheral gardens and parks caLl into question archaeological site-deinition procedures.
With an urban green-space perspective, we can be alert to urban boundaries not as hard lines, nor entirely as a gradient, but also as a peripheral mosaic. Faced with the question of where to "draw the line" for a settlement boundary, an arbitrary density limit for artifacts is usually established. At Teotihuacan, a criterion of 300 m without structures or other signiicant materials was applied (Millon 1973:8). At Xochicalco, an interval of 100 m was used (Hirth 2000:54). Hirth (2000:54) separates the concept of settlement deinition from community deinition in discussing Xochicalco. Because he refers to outlying settlements linked to Xochicalco by roads and situated 1-3
Km away as communities, he presumably does not consider them part of Xochicalco (the boundaries of the Xochicalco community are not stated). Such settlements would be possible candidates for accommodating peripheral palace, garden, or estate facilities. Maya settlements such as Caracol and Coba, wIth stone roads to outlying architectural groups, warrant attention in regard to peripheral facilities as well.
ANCIENT OPEN SPACE, GARDENS, AND PARKS
Other archaeological reasoning also can be rethought. In the case of Monte Alban, terraces descending the hill slope had traces of residential walls and domestic artifacts, but eventually, lower terraces did not yield these traces; thus, they were excluded from the city limits (Blanton 1978:8), but such locations would be subject to greater erosional deposition that might mask occupation traces. Commonly, urban studies neglect the possibility of gardens and villas or other installations slightly removed from the built-up portion of the settlement as well as their roles as part of the urban settlement.
One example of outlying installations concerns Tikal, for which "minor ceremonial centers" were identified that fall within the area eventually discovered to be delimited by the ditch and embankment segments (Puleston 1983). In a study of Roman villas during the Renaissance, Coffin (i979:vii) developed a concept of city limits, which includes "all the land surrounding the city which is owned by persons whose political, religious, commercial, or social activities are centered within it." As a result, he accommodates villas used by Roman elites in the vicinity of the more compact portion of that capital; in his view, Rome is not the area within the city walls, but approximately the modern region of Latium or Lazio. The villas range up to 70 km distant from Rome, but most fall within half that distance, approximating ancienT Rome's peri-urban extent in Goodman's (2007:20) analysis. Such distances reflect advantages of animal and wheeled transport different from Mesoamerican technologies, as well as construction of roads leading to Rome. Ethnohistoric documents mention a range oF distances for royal Aztec parks and gardens (Evans 2000:209-211); most were in the Basin of Mexico within 5-12 km of the capitals.
In sum, royal or eLite residences and gardens may be located at some distance from what archaeologists usually take as a settlement boundary, yet the periodic visits or seasonal uses of these diverse properties show a city with a mosaic extent (although still within an accessible area). In a behavioral sense, the city can be seen as the
Array of close places in regular annual use for continuing residential and social interactions connected with city life. From this perspective, the fuzzy limits of low-density tropical settlements are more understandable and constitute a more widespread phenomenon than recognized by any mechanical process of site delimitation in archaeological survey and mapping. Especially elites with peripheral estates may participate in urban affairs on a periodic basis (e. g., Coffin 1979; Tourtellot 1993).
PeripheraL gardens and parks have implications for urban models. Some modern cities support Burgess's (1925) idea of higher income or status groups located toward the periphery, but Sjoberg (1960) described an inverted model for preindustrial cities, with the most powerful elements of society clustered near the urban center (see review in Abbott 1974). Here I call attention to the urban periphery and its potential as a location for social institutions or residences with elevated status that command extensive space. In some respects, peripheral elite gardens are more in keeping with Burgess's original ideas, yet peripheral palaces and grounds coexist with others near the urban core as Sjoberg (1960) contended, supporting a dual model. In some cases, such as the extension of suburban villas from Constantinople (and Istanbul) along the waterfronT (Constantinides 2002:87-89; Mango 1985:47, 49; Talbot 2002:6), the pattern is more in keeping with Hoyt's (1939) radial-sector modeL than a concentric model.
Peripheral gardens and parks suggest that archaeological settlement definitions for major urban centers should exercise caution about peripheries and that urban studies in archaeology will benefit from more attention to open space generally. Investigators should examine what might otherwise be thought of as outlying settlements in the immediate vicinity to see whether some may reflect a mosaic urban extent. A mosaic extent in some cases may resemble modern fringe belts. At stake with palatial grounds is evidence concerning class differentiation and the allocation of labor and resources in Mesoamerican urban societies.
World History
