Several theories have been proposed for their formation. The more common theory is that they form by the gradual removal of the sand, dust and other fine grained material by the wind and intermittent rain leaving only the larger fragments behind. This does not continue indefinitely, however, because once the pavement has been formed it can act as a barrier to further erosion. Secondly, it has been proposed that desert pavement forms from the shrink/swell properties of the clay underneath the pavement; when precipitation is absorbed by clay it causes it to expand and later when it dries it cracks along planes of weakness. This geomorphic action is believed to have the ability to transport small pebbles to the surface over time; it stays this way due to the lack of abundant precipitation that would otherwise destroy the pavement development through transport of the clasts or excessive vegetative growth.
The newest theory of pavement formation comes from careful studies of places like Cima Dome, in the Mojave Desert of California, by Stephen Wells and his coworkers. Cima Dome is a place where lava flows of recent age, geologically speaking, are partly covered by younger soil layers that have desert pavement on top of them, made of rubble from the same lava. Obviously the soil has been built up, not blown away, and yet it still has stones on top. In fact, there are no stones in the soil, not even gravel.
There are ways to tell how many years a stone has been exposed on the ground. Wells used a method based on cosmogenic helium-3, which forms by cosmic ray bombardment at the ground surface. Helium-3 is retained inside grains of olivine and pyroxene in the lava flows, building up with exposure time. The helium-3 dates show that the lava stones in the desert pavement at Cima Dome have all been at the surface the same amount of time as the solid lava flows right next to them. It's inescapable that in some places, as he put it in a July 1995 article in Geology, "stone pavements are born at the surface." While the stones remain on the surface due to heave, deposition of windblown dust must build up the soil beneath that pavement.
For the geologist, this discovery means that some desert pavements preserve a long history of dust deposition beneath them. The dust is a record of ancient climate, just as it is on the deep sea floor and in the world's ice caps.
Desert pavement surfaces are often coated with desert varnish which is a dark brown, sometimes shiny coating that contains clay minerals. In the USA a famous example can be found on Newspaper Rock in Canyonlands, Utah.
According to Ronald I. Dorn and Theodore M. Oberlander (Science Volume 213, 1981), desert varnish is a thin coating (patina) of manganese, iron and clays on the surface of sun-baked boulders. Desert varnish is formed by colonies of microscopic bacteria living on the rock surface for thousands of years. Desert varnish is also prevalent in the Mojave desert and Great Basin geomorphic province.