Received 15 April 2005; revised 16 September 2005; accepted 21 September 2005; published 2 December 2005. (1) To explain the much higher denudation rates and valley network development on early Mars (>� 3.6 Gyr ago), most investigators have invoked either steady state warm/wet (Earthlike) or cold/dry (modern Mars) end-member paleoclimates. Here we discuss evidence that highland gradation was prolonged, but generally slow and possibly ephemeral during the Noachian Period, and that the immature valley networks entrenched during a brief terminal epoch of more erosive fluvial activity in the late Noachian to early Hesperian. Observational support for this interpretation includes (1) late-stage breaching of some enclosed basins that had previously been extensively modified, but only by internal erosion and deposition; (2) deposition of pristine deltas and fans during a late stage of contributing valley entrenchment; (3) a brief, erosive response to base level decline (which was imparted as fretted terrain developed by a suite of processes unrelated to surface runoff) in fluvial valleys that crosscut the highland-lowland boundary scarp; and (4) width/contributing area relationships of interior channels within valley networks, which record significant late-stage runoff production with no evidence of recovery to lower-flow conditions. This erosion appears to have ended abruptly, as depositional landforms generally were not entrenched with declining base level in crater lakes. A possible planetwide synchronicity and common cause to the late-stage fluvial activity are possible but remain uncertain. This increased activity of valley networks is offered as a possible explanation for diverse features of highland drainage basins, which were previously cited to support competing warm, wet and cold, dry paleoclimate scenarios.