Low islands are by themselves a natural paradox: a tiny surface of a mere few square kilometres floating up to 7 meters upon the ocean, a basic soil made of coral debris scattered upon a calcified beach rock, no visible fresh water resources and an exceptional isolation. They present nothing, on first appearance, which would support or incite any human settlement. This is probably the reason why the first European visitors, staggered by the aspect of the islanders, thought they were just lost survivors of wreckages or outcasts rejected from high islands. In fact, archaeological surveys in the Tuamotu atolls have revealed a rather complex distribution of holes and trenches dug into the coral sand and pebble substrate and spread over hectares. Oral tradition and collected memories, at least in the Tuamotus, have shown that these structures were cultivation pits in which different varieties of taro were grown. Then the reconstruction of cultivation techniques and social structures associated with these elementary food resources has been confirmed by observations of still-living systems in the Cook Islands and the atolls of Micronesia, especially Kiribati. They proved that a very considerable knowledge of the natural resources of low islands had been observed and used; for example, the practice of mulching with material from specific trees or creeping plants thus providing phosphates or other necessary nutrients. And most important, they had observed that a thin fresh water lens was present almost everywhere on any motu (islet), which provided moisture for cultivation pits and drinking water for the islanders. Therefore the origin of the very first settlers of atolls has still to be sought elsewhere than adjacent high islands and is more probably towards the most western islands of the Pacific.