Eastern Hadza describe their country as divided into four regions (Woodburn, 1968b, pp. 103–105). Are the regions so distinct that we should dissect our account of eastern Hadza demography, treating the regions as separate sub-populations? Here I attempt to quantify the description of regions.

Woodburn pointed out that the names of the regions referred to localities, but could also be used to refer to the people living in them. He suggested there was no exclusion of people from a region. Some people spent many years within one region, others moved from region to region. Among the four regions, Siponga, Mangola, Tliika, and Han!abe, the last, Han!abe, appeared to have a less certain status than the others. From his 1930s fieldwork, Kohl-Larsen (1958) seems to have been aware of a similar localization in his references to “the Matete horde” (Mangolabe) and the “Grosse-Wasser horde” (the Sanola area in Tliika). Woodburn's description seemed to apply closely to the situation we found in 1985, and to that shown in Lars Smith's 1977 census and aerial surveys. The Hadza concepts endured unchanged, although the on-the-ground realities had changed a little by the end of my data collection in mid-2000. We might wonder if there were effects of the loss of land that was associated with failed settlement schemes, the occupation and degradation of large parts of Siponga and Mangola, and the eventual arrival of tourism.

Regions are spatial clusters

We recorded Universal Transverse Mercator (UTM) coordinates of 243 Hadza camps all over eastern Hadza country. These represent nearly all the camps that we visited between 1985 and 2000. In our 1985 census, camps were located by compass and 1:50,000 scale maps. From late 1985, camp locations were identified by a hand-held GPS locator (an early model Magellan, then Garmin GPS12). Each camp in our records has also been given a region entry, in accord with our, and our Hadza field assistant's usage of the region names.

I fed the UTM coordinates to a cluster analysis in Minitab15. UTM coordinates measure distance from a very remote point of origin. Camps that are close to each other will have very similar UTM coordinates; camps far apart will have less similar coordinates.