The surface of Venus represents a unique record of the impact
cratering process. Venus' massive atmosphere significantly affects
the cratering process and effectively filters out small
meteorites, leaving the surface deficient in craters with
diameters less than 
km [ Schaber et al.,
1992; Phillips et al., 1992]. Because of
extremely low erosion rates, impact craters are almost completely
undisturbed except for tectonic or volcanic processes. But the two
most striking aspects of the cratering record are that the spatial
distribution of craters is indistinguishable from a random
distribution and that few craters display signs of significant
modification by either tectonic or volcanic processes (Table 1).
These two observations, in particular, place stringent constraints
on the resurfacing history of Venus [ Strom et al.,
1994], although interpretation of these constraints remains
controversial.
Models for the surface age of Venus can be considered as lying in between two extreme, or end-member models. In one extreme, the crater data can be interpreted to indicate that all significant geologic activity ceased on Venus as of 300--500 Ma (million years ago), and that prior to that time, activity was rapid and no craters from this previous epoch survive on the surface today. The other end-member model suggests that geologic activity is ongoing and that no steep decline in rates of tectonism and volcanism is required.
Noting the spatial randomness and the unmodified appearance of
impact craters, Schaber et al. [1992; also
Strom et al., 1994] suggest that they represent a
production population, one that has undergone little or no
modification due to volcanism or tectonism. Prior to 
Ma,
they suggest that the rate of volcanic and tectonic activity was
high in comparison to the cratering rate. This activity
constitutes global resurfacing, and may be a recurring event in
Venus history [ Arkani-Hamed et al., 1993;
Parmentier and Hess, 1992]. Subsequent to
resurfacing was a short period (approximately 10 m.y.) of steep
decline in the rate of geologic activity. Geologic activity since
the end of this ``great decline'' has been comparatively minor.
Phillips et al. [1992] argue against an
end-member global resurfacing model for reasons that involve
spatial correlation of modified craters with geological features
or with one another. These workers formulated an equilibrium
resurfacing model which considered a characteristic area and
recurrence interval for resurfacing events. The model only
satisfies the requirement of spatial randomness for events
covering less than 140,000 km
(0.03% of Venus' surface) that
recur at intervals of less than 150,000 years or for events
covering more than 10% of the planet (
km
)
with recurrence intervals greater than 50 m.y. This simple model
encountered difficulties explaining the numbers of embayed and
tectonized craters, and Phillips et al. [1992]
note that no single value for the size of resurfacing events was
completely consistent with observations. Further tests of such
models [ Bullock et al., 1993] appear to indicate
that large resurfacing events will not satisfy the observed crater
distribution, and that small, more frequent events produce more
embayed craters than are observed. Although observations probably
allow for particular equilibrium scenarios, the impact crater
distribution appears to be most consistent with models that call
for a near-complete resurfacing of the planet prior to 300--500
Ma. Subsequent to this period of extreme activity, process rates
declined and impact craters began to accumulate, with only minor
modification and resurfacing since.