Excellent three-dimensional exposures of the Upper Jurassic Salt Wash Sandstone Member of the Morrison Formation in the Henry Mountains area of southern Utah allow measurement of the thickness and width of fluvial sandstone and shale bodies from extensive photomosaics. The Salt Wash Sandstone Member is composed of fluvial channel fill, abandoned channel fill, and overbank/flood-plain strata that were deposited on a broad alluvial plain of low-sinuosity, sandy, braided streams flowing northeast. A hierarchy of sandstone and shale bodies in the Salt Wash Sandstone Member includes, in ascending order, trough cross-bedding, fining-upward units/mudstone intraclast conglomerates, single-story sandstone bodies/basal conglomerate, abandoned channel fill, multistory sandstone bodies, and overbank/flood-plain heterolithic strata. Trough cross-beds have an average width:thickness ratio (W:T) of 8.5:1 in the lower interval of the Salt Wash Sandstone Member and 10.4:1 in the upper interval. Fining-upward units are 0.5-3.0 m thick and 3-11 m wide. Single-story sandstone bodies in the upper interval are wider and thicker than their counterparts in the lower interval, based on average W:T, linear regression analysis, and cumulative relative frequency graphs. Multistory sandstone bodies are composed of two to eight stories, range up to 30 m thick and over 1500 m wide (W:T > 50:1), and are also larger in the upper interval. Heterolithic units between sandstone bodies include abandoned channel fill (W:T = 33:1) and overbank/flood-plain deposits (W:T = 70:1). Understanding W:T ratios from the component parts of an ancient, sandy, braided stream deposit can be applied in several ways to similar strata in other basins; for example, to (1) determine the width of a unit when only the thickness is known, (2) create correlation guidelines and maximum correlation lengths, (3) aid in interpreting the controls on fluvial architecture, and (4) place additional constraints on input variables to stratigraphic and fluid-flow modeling. The usefulness of these types of data demonstrates the need to develop more data sets from other depositional environments.