Population-density maps of craters in three size ranges (0.6 to 1.2 km, 4 to 10 km, and >20 km in diameter) were compiled for most of Mars from Mariner 9 imagery. These data provide: historical records of the eolian processes (0.6 to 1.2 km craters); stratigraphic, relative, and absolute timescales (4 to 10 km craters); and a history of the early postaccretional evolution of the uplands (> 20 km craters).

Based on the distribution of large craters (>20 km diameters), Mars is divisible into two general classes of terrain, densely cratered and very lightly cratered—a division remarkably like the uplands-maria dichotomy of the moon. It is probable that this bimodal character in the density distribution of large craters arose from an abrupt transition in the impact flux rate from an early intense period associated with the tailing off of accretion to an extended quiescent epoch, not from a void in geological activity during much of Mars' history. Radio-isotope studies of Apollo lunar samples show that this transition occurred on the moon in a short time.

The intermediate-sized craters (4 to 10 km diameter) and the small-sized craters (0.6 to 1.2 km diameter) appear to be genetically related. The smaller ones are apparently secondary impact craters generated by the former. Most of the craters in the larger of these two size classes appear fresh and uneroded, although many are partly buried by dust mantles. Poleward of the 40° parallels the small fresh craters are notably absent owing to these mantles. The density of small craters is highest in an irregular band centered at 20°S. This band coincides closely with (1) the zone of permanent low-albedo markings; (2) the “wind equator” (the latitude of zero net north or south transport at the surface); and (3) a band that includes a majority of the small dendritic channels. Situated in the southermost part of the equatorial unmantled terrain which extends from about 40°N to 40°S, this band is apparently devoid of even a thin mantle. Because this belt is also coincident with the latitutde of maximum solar insolation (periapsis occurs near summer solstice), we suggest that this band arises from the asymmetrical global wind patterns at the surface and that the band probably follows the latitude of maximum heating which migrates north and south from 25°N to 25°S within the unmantled terrain on a 50,000 year timescale.

The population of intermediate-sized craters (4–10 km diameter) appears unaffected by the eolian mantles, at least within the ±45° latitudes. Hence the local density of these craters is probably a valid indicator of the relative age of surfaces generated during the period since the uplands were intensely bombarded and eroded. It now appears that the impact fluxes at Mars and the moon have been roughly the same over the last 4 b.y. because the oldest postaccretional, mare-like surfaces on Mars and the moon display about the same crater density. If so, the nearness of Mars to the asteroid belt has not generated a flux 10 to 25 times greater than the lunar flux. Whereas the lunar maria show a variation of about a factor of three in crater density from the oldest to the youngest major units, analogous surfaces on Mars show a variation between 30 and 50. This implies that periods of active eolian erosion, tectonic evolution, volcanic eruption, and possibly fluvial modification have been scattered throughout Martian history since the formation and degradation of the martian uplands and not confined to small, ancient or recent, epochs. These processes are surely active on the planet today.