Most approaches to interpreting alluvial records in drylands fall into one of two categories: (1) The "arroyo problem," wherein workers study cycles of streambed aggradation and degradation in broad, unconstricted alluvial valleys; and (2) paleoflood hydrology, where alluvial sequences in constricted bedrock canyons are interpreted as paleoflood deposits from streams with stable channel grade and geometry. Both approaches can be valid in their end-member settings, but there is confusion about how the two record types relate in a single drainage. We address this disconnect in Buckskin Wash, an ephemeral stream that consists of a broad alluvial reach draining into a tightly constricted slot canyon. By employing detailed sedimentology, stratigraphy, and geochronology in both the alluvial and constricted reaches of the watershed, we test the hypothesis that the slot canyon deposits are anticorrelated to valley-fill deposits upstream, implying that arroyo cutting is driven by episodic flooding. Alluvial reach deposits are characterized by stratal packages representing incremental, long-term aggradation bound by erosion surfaces representing channel entrenchment. At least four packages younger than ~3 ka are present, the youngest spanning ~0.7 - 0.15 ka. Each is composed of interfingering imbricated gravels, laminated sands, and massive silty clays. Constricted reach deposits consist of five discrete packages, each composed of tabular beds of laminated silty sand that were deposited relatively rapidly. The oldest package dates to ~1.9 - 1.1 ka, whereas the rest of the deposits are younger than ~0.15 ka. Traditional paleoflood techniques would suggest that the constricted reach deposits record a ~1000-year absence of paleofloods followed by ~100 years of frequent, high-magnitude flooding that indeed correlate to arroyo cutting upstream. We argue instead that the constricted reach deposits record an episode of higher preservation potential. Transport of sediment from the alluvial reaches during historic arroyo cutting likely led to a pulse of sediment storage and changed stage-discharge relations in the slot canyon downstream, allowing even moderate floods to overtop existing deposits and be preserved. This new interpretation suggests that, because preservation may be a function of episodic sediment loading from upstream, constricted-reach deposits may not accurately record the paleoflood history of a stream.