continued to occupy tripartite houses of the kind familiar from ‘Ubaid sites like
Madhur (Roaf 1989 ), but now were grouped around a common courtyard with build-
ings that probably served non-residential functions. Later third millennium sites
located within the same northern Mesopotamian zone as Habuba Kabira are char-
acterized by radiating “hollow ways” which have been interpreted as the result of
people–and especially animals–moving in and out of the settlements on a daily basis
(Wilkinson 1993 ), but it remains unclear whether there were also hollow ways at the
earlier Habuba Kabira.
A key issue to be determined is whether Protoliterate settlements were similar in
density to the very crowded cities of the later third millennium houses in southern
Mesopotamia (see below). Although, sadly, we have no faunal studies for fourth
millennium sites in Iraq, our understanding of the environmental conditions prevailing
in southern Iraq (Pournelle 2003 , 2007 ; Algaze 2008 : fig. 5 ), fourth millennium textual
sources (Englund 2008 : 155 – 169 ) and art (Kawami 2001 ) all suggest that cattle likely
played a much larger role in the fourth millennium diet than they did in later times
when pig and sheep dominate the faunal records (Desse 1992 ). If fourth millennium
settlements provided shelter for both people and domestic animals, their human
population densities were likely much lower than their later counterparts.
These various strands of data can be brought together to develop the key questions
which need to be answered regarding settlement in fourth millennium and early third
millennium Mesopotamia: Were tripartite houses still the norm at this time, and if so
did they exist as part of larger, more complex compounds? Were domestic animals–
especially cattle–still kept within the households?
The partial plans of three fourth millennium sites (Figure 8. 1 ), together with one
fourth to early third millennium landscape (Figure 8. 2 ), are visible in the Digital Globe
satellite imagery. Tell Umm al-Fargus (Adams 1981 : 272 , site 1096 ) covers some 12 ha.,
spread over three mounds, all of which have surface traces. Toward the northern edge
of the northwestern mound, larger-scale architectural traces are visible which might
indicate the presence of a temple. Beyond this area, although the imagery is not quite
clear enough for us to map a large number of houses, what is clear is that there are no
courtyard houses–since these larger spaces and their regular appearance are always
clearly visible in imagery of similar quality. Some structures seem to reflect tripartite
houses, while others apparently have large open areas surrounded by walls. In one part
of the main mounds there appear to be largish rectangular compounds (ranging from
100 to 300 square meters in size) separated one from the other by narrow alley-
ways. These are not dissimilar to the households at Habuba Kebira identified by Vallet
( 1996 : fig. 4 ).
A second site, similar in organization, is Uruk Survey site 245 (Adams and Nissen
1972 : 229 ). Although at 8 ha. it is somewhat smaller than Tell Umm al-Fargus, it is
similar in having three mounds: two larger ones oriented in a more or less north-
west/south-east direction, and a smaller one to the west. Unlike at Tell Umm-al-Fargus,
the best architectural preservation was on the northwest mound. Adams and Nissen
( 1972 : 229 ) report the presence of many clay cones, some forming a right angle, located
in the northeastern part of this mound–in an area now, sadly, completely looted. This
suggests the presence of an important fourth millennium temple, a suggestion
reinforced by the traces of domestic structures over the rest of this mound (the only
area where such traces are clear). These houses are scattered and, in orientation, radiate
–– Elizabeth C. Stone ––