Prior to 1934, all Ngatatjara were highly mobile and relatively opportunistic in their settlement pattern. During periods of sustained rains in particular parts of the desert, families congregated to take advantage of the water and to hunt game attracted by improved vegetation growth produced by such rains. Such maximal groups are estimated to have been as large as 150 individuals, but the duration of such aggregations was limited by the amount of game and water available and tended to be only a few weeks. These were major social events, when ceremonies and initiations occurred along with betrothals and curing activities. As drought conditions worsened, extended families departed in search of better hunting, with even smaller family groups setting out for more reliable water sources as drought stress increased. In extreme cases of long-term drought, families would leave their home area altogether and take up temporary residence with related families in areas as far as 500 kilometers away. Particular campsites might not be visited for several years in succession, or they might be visited several times in the same year, depending upon rains and associated plant and animal resources. There was no bounded territory within which such groups confined their foraging, nor were their social groups fixed in size or composition. Minimal social groups consisting of members of related families and totaling about ten to fifteen individuals could be found residing and foraging together around more or less dependable water sources during droughts. Domestic architecture consisted of conical or semicircular bough shelters during the summer, mainly to provide shade, and open-air campsites with linear or semicircular bough windbreaks during winter. Each family campsite had a central hearth that served as the focus for its social activities along with subsidiary hearths for warmth while sleeping. There were also taskspecific sites that included quarries, hunting blinds, woodworking localities, and ceremonial and rock-art sites.