Certain districts are still uninhabited on account of the lack of water. Some of the sources of water are permanent. The water issues at the base of the volcanic rocks, when the underlying rock is impermeable, and above the various levels of the marl in the Patagonian swamps; for instance, in the _canadones_ round the Gulf of San Jorge. Besides this, the rain and melting snow leave on the surface of the tableland a great number of pools, which evaporate in the dry season. These are temporary supplies, the _manantiales_, to which the breeders are reduced over large areas of the tableland. Most of the stagnant sheets of water which are permanent are saline. The proportion of salt in them is very variable, and changes in each case according to the cycle of dry and wet years. The water of the Carilaufquen was fresh in 1900, and in 1914 it had become brackish, though it could still be used for the flocks. Finding permanent sources of water is the first concern of the breeder. In some districts he has succeeded in tapping sheets of fresh water by means of wells. There are none of these wells in the crystalline zones, the closed hollows, where the sheets of water are often large, but they are always saline. Neither are there any in the red sandstone district, the dryest of all. In the western region the wells are sunk in the arid valleys, along the track of the underground stream. Thus the Picun Leufu, the visible course of which is lost seventeen miles above its confluence with the Limay, may be traced by a continuous line of wells. It is especially in the coastal districts that the wells have transformed the conditions of breeding. Water was first discovered at the foot of the dunes, along the coast itself (district of Viedma, San Jose, etc.). Since then deep borings have been made over the whole of the Tertiary platform on both sides of the lower part of the Rio Negro, north of San Antonio. There every ranch has its sheet-iron tank, sheltered by a clump of tamarinds, with a windmill to fill it.