Even without subdivision of predominantly green versus gray open spaces, calculations of open space provide a basis to compare settlements. The degree of nucleation of settlement has implications for a variety of urban issues, such as quality of life, sanitation and health, and urban food production. For the most part, archaeologists have compared population densities of ancient cities to analyze the degree of nucleation, rather than the proportions of open space. Whereas population igures for modern cities are readily available, such information is hard-won for ancient cities. Population calculations make a number of assumptions about household size or numbers of people in relation to architectural remains, such as residences or sleeping space (e. g., Rice and Culbert 1990), or to artifact densiTies (e. g., Sanders et al. 1979:34-40). Population density is valuable for comparing urban settlements, but the proportion of open space versus architectural space involves fewer assumptions. Of course, not all ancient settlements allow assessment of open spaces because surface architecture is not reliably visible.
Where we have information, the proportion of urban open space versus roofed architecture or structures covers a wide spectrum. For modern cities, Huang et al. (2007) used satellite imagery to determine
The percent of open space in 47 cities in developing countries, which had a standard deviation range of 13.56-39.6 percent open space. For 30 cities in developed countries, one standard deviation encompassed 9.23-24.88 percent open space. Combined, the 77 modern cities span approximately 9-40 percent open space across one standard deviation. Occasionally, ancient cities have been evaluated for open space. JashemsKi (2008:15) characterized Roman Pompeii as just over one-third open space approximately equally divided between green and gray space. Jim anD Liu (2001:359, citing Zeng 1991) state that arounD AD 300, gardens covered half of Guangzhou, China, somewhat higher than the modern range. Not well represented in the Huang et al. (2007:185) modern sample were tropical and mountain areas where cloud cover was an impediment to imagery; thus, low-density tropical urbanism (Evans et al. 2007) is not adequately represented. At the tropical Mesoamerican capital of Cerro de las Mesas in Veracruz, where dispersed occupation is characteristic, conservatively 90 percent or more of the city is open space (mapped in my fieldwork). Clearly, marked variation characterizes open spaces among cities, wIth low-density tropical cities at one extreme, possibly as part of a continuum. Both Creekmore (Chapter 2 in this volume) and Fisher (Chapter 6 in this volume) remark on the paucity of open space for cities in third-millennium Upper Mesopotamia and Late Bronze Age Cyprus respectively, perhaps exhibiting extremes at the other end of the spectrum.
ANCIENT OPEN SPACE, GARDENS, AND PARKS
In the next section, I address one aspect of Mesoamerican urban open space - the archaeological potential for palatial gardens. Such gardens were recorded ethnohistorically for the Aztecs. Although other open spaces are important as well, the investments in royal and elite gardens deserve attention because of their potential labor investment and cultural and social roles. Instead, attention has focused to a considerable extent on the important subject of palaces, usually without consideration of associated open space (e. g., Christie 2003; Christie and Sarro 2006; Evans and Pillsbury 2004; Inomata and Houston 2000, 2001).
World History
