In these natural test tubes, the complex molecules needed for life could have evolved in safety, taking its building blocks from the water washing over the rock and from the minerals within.

New research argues that the pores provide the perfect sheltered environment for the chain of chemical reactions necessary to evolve the first bacteria. The proposal, from scientists at the Universities of Edinburgh and Chicago, also explains how cell walls first developed: as lids on the pores used to batten down the hatches during droughts.

The work was welcomed by Dr Graham Cairns-Smith, an expert on origin of life at the University of Glasgow. "It's a fascinating piece of mineralogy and I'm all for it," he told BBC News Online.

"The idea that the critical organic reactions would take place in the open primordial oceans is now very hard to believe."

These tiny pores may have been the vessels in which life began

The new idea follows a painstaking analysis of the microscopic structure of alkali feldspar, a common mineral in rocks. It solves many of the problems with current ideas of how life first began.

Life blossoming in primordial pools, rich in organic molecules, is now thought to be unlikely. The molecules would easily be dispersed by waves or currents, stopping the chain of reactions. Furthermore, the intense ultraviolet light shining on the early Earth would have quickly destroyed any embryonic life.

But a cosy, safe haven would prevent these difficulties. It has been suggested that clay minerals could have acted as templates for life.

But as the team leader Professor Ian Parsons at Edinburgh University told BBC News Online, "That's a good start but most minerals, including clays, don't work because they repel rather than attract organic molecules. However, there is a mineral called zeolite, one rare type of which does attract organic molecules, and this could occur on the surface of feldspars."

The first cells

Best of all, the new work presents a plausible way to develop cell walls, a critical step which allows living organisms to travel safely in search of food or away from predators.

Professor Parsons explains: "The reactions for life will not work in the right order without a container, but you cannot get a container, or cell wall, until you have the right reactions - it's Catch 22."

A cast of the pore network shows how the first organisms could have spread

He, along with colleagues Dr Martin Lee and Professor Joseph Smith, believes that if the surface of a rock with the first glimmers of life hidden inside dried out occasionally, this would encourage the proto-organism to develop some form of protection.

A fatty lid would prevent it from perishing through evaporation. Through time, this could expand outwards, like an inflating balloon, so the organism could capture passing nutrients. Eventually the organism could leave the pore entirely, surrounded by the protective layer. This budding process would then have created the first cell.

Most likely place for life

Dr Cairns-Smith sees merit in this. "Feldspars are found in terrestrial rocks, where ocean, air and earth all come together. This is the best place for life to begin as a more complicated environment will create more selection pressures for interesting things to develop."

The pores form when the mineral is exposed to the dilute acids contained in rain. A regular honeycomb of pores, with a huge surface area, begins to develop.

Each pore is less than a thousandth of a millimetre wide and reaches less than one hundredth of a millimetre into the rock.

The researchers' study, published in the Proceedings of the National Academy of Sciences on Tuesday, shows that there are a million potential micro-bioreactors on every square millimetre of rain-weathered feldspar.

The next step, according to Professor Parsons, is to understand the crystal structure of the pore walls. "If atomic structure resembles known catalysts for these organic molecule reactions, it would really all hang together."