Shell-bed development can be a product of complex sedimentological and biological factors. The Upper Ordovician sediments near Cincinnati, Ohio constitute a succession of thinly interbedded shelly carbonates and mudrocks. Despite years of study, the development of Cincinnatian shell beds and metre-scale cycles has, until recently, been attributed solely to storm reworking. This “storm-winnowing model” treats shells as passive sedimentary clasts, ignoring other factors of shell-bed development. A recently proposed alternative is Brett and Algeo’s idea that these shell beds grew during long periods of normally low sedimentation, while most mud accumulated during brief periods of high sedimentation. Under this “episodic starvation model,” any storms would winnow pre-existing muds and shell beds alike. We tested both models in the Edenian–Maysvillian (early to mid Katian) strata of the Cincinnati region by compiling observations on their petrologic, taphonomic, and paleoecologic characteristics. The storm-winnowing model does not explain several observed features that the episodic starvation model does, including (i) storm-related sedimentary structures in mudrocks and limestones; (ii) lack of a sufficiently fossiliferous precursor deposit to winnow; (iii) deep-water faunas in grainstones; (iv) mixed taphonomic conditions of shell-bed fossils; (v) ubiquitous discontinuity surfaces; (vi) carbonate concretion horizons; (vii) unwinnowed shell beds; and (viii) micrite in packstones. Episodic starvation is a superior explanation because it explains all of these features and allows for the complex interplay of other environmental and biological factors that contribute to shell-bed growth. It may also be applicable to other deposits, previously interpreted as tempestites.