Virgin soil was not reached, but the levels exposed at plain level correspond to the last phase of occupation at Adam Buthi. In addition to purplish slipped unpainted pottery, Togau B and Kechi Beg pottery was found (36). Two trenches were opened in 1999 (37, 38, 39). During the early 3rd millennium BCE. the site reached its maximum extension. Well-built stone and mud brick architecture was exposed in the sections and on the surface. In the east, several stone-lined hearths and dump pits containing animal bones and a large number of discarded and broken pots were excavated (40, 41, 42)
Apart from the typical buff "Nal"-pottery with black paint, fine orange and coarser household wares (43, 44), polychrome vessels, partly still complete, were unearthed (45, 46, 47, 48).
A single Faiz-Mohammad Grey ware sherd (49) and a chlorite fragment with an imbricate design are important finds since they provide cultural links to the north and the west. The pottery changes through the levels. Polychrome sherds are outlasted by monochrome Nal wares and in the upper layers of trench II carinated bowls with hammer-head rims and reddish-brownish slips foreshadow the later Kulli pottery (50). A typical motif is the single-bracket design which becomes a hallmark of the late 3rd millennium BC occupation (51, 52).
Balakot, which is located in the southeastern Bela plain, was excavated between 1973 and 1976 by G.F.Dales, of the University of California, Berkeley. It is the only properly excavated site in the region. Despite its small size (ca. 4.5 hectares), the site is thus of crucial importance due to its long Early Harappan cultural sequence which is now dated to between 3100/3000 and 2600 BCE. It is the southernmost findspot of Quetta- and Nal-pottery, but has also many affinities to Amri in Lower Sindh.
Although the transition to the Harappan period (II) is stratigraphically not very clear, there appears to be a gap. Despite some pottery forms which continue into the later third millennium BCE, the classical Harappan pottery appears suddenly and fully fledged at the site. Kulli elements are also present, but not as pronounced as at Nindowari or the many Kulli sites found in the Kanrach, Hab- and Saruna Valleys.
Murda Sang is the largest prehistoric site in the Kanrach Valley (53). It was discovered in 1997 and trial trenched in 1998. The nucleus of the settlement consists of houses grouped along lanes and streets. This central portion is about 6ha large, but scattered occupation and a kiln area cover altogether ca. 35ha. The eastern edge is eroded by the Kanrach River (55). Two dams were found to the north of the site and we assume that fields were located there. The site and the whole valley are overlooked by a fortification built on top of a terrace hill at the southern edge of the site.
The soundings revealed two main periods of occupation, the lower with three very compact building phases, the upper one with two. The ground was terraced with gravel and pottery before construction.
The ground was terraced with gravel and pottery before construction. Houses have a stone foundation, but mudbricks were also used (54, 56), the roof was covered with mud-smeared reed. The pottery from the earlier occupation is very similar to that from the earlier levels of Balakot I (57, 58, 59, 60).
An AMS date run on charcoal suggests a dating into the very early third millennium BCE. After 2700/2600 the site was abandoned. The uppermost, badly preserved occupation dates to the later Kulli period. Very small parts of the site were re-used during the late Islamic or British period. A very large platform-house site of the Historic Period was built over scattered houses and possibly fields north of the main settlement.
A sounding revealed a sequence of finely banded sand and mud layers (61). This evidence and the accumulation of humus above the old gravel surface indicate frequent flooding (62). Most probably, the river and wadis which have deeply cut their bed into the rock, flew at a very different level 5000 years ago. A substantial change in the topography and drainage pattern since the 3rd millennium BCE thus appears likely.