Savanna sites are idealized as exhibiting a demographic “bottleneck” physiognomy comprising a lower stratum of abundant resprouting persistent “juveniles” (albeit of indeterminate age), a mid-stratum comprising relatively few released “saplings,” and a canopy-layer cohort of “adults.” The magnitude and frequency of disturbance is considered to influence the critical transition from juvenile into adult phases. Under fire-prone Australian savanna conditions, an extensive suite of both observational and manipulative studies have explored the responses of tree recruitment to fire disturbances. These studies oftentimes have produced seemingly highly disparate responses, particularly with respect to the differential responses of relatively fast-growing eucalypts versus non-eucalypts under different fire regime, and overstory competition, conditions. This study contrasts the responses of tree recruitment height classes to (1) the effects of total canopy removal from severe Cyclone Monica in 2006 over a subsequent 10-yr period, (2) with observations from long-term monitoring sites under relatively stable overstory conditions at Litchfield National Park over a six-year period, (3) under ambient, frequent fire occurrence (mean > 0.5 fires/yr) at both locales including relatively severe late dry season fires. Recruitment at both study sites was represented mostly by resprouting, clonally reproducing juvenile trees <2 m tall, around half of which died over respective assessment periods. At post-cyclone assessment plots, there was substantial release of eucalypts, including within the first five years, into the >5 m height class, with negligible corresponding release of non-eucalypts. At Litchfield plots, there was negligible release of both eucalypts and non-eucalypts. In discussion, we contrast these results with findings from relevant regional studies. We contend that collective disparate observations feasibly can be reconciled as reflecting significant interactions between fire regime characteristics and variable site overstory competition effects, such that the rate of recruitment of fast-growing savanna eucalypt individuals into the midstory is relatively independent of the fire regime, but is significantly regulated by resource competition interactions especially with the overstory, whereas recruitment of non-eucalypts is relatively independent of overstory competitive effects, but is suppressed under fire regimes dominated by frequent, especially severe fires.