METHODS

Compressed plant fossils were collected by PW, accompanying a Carnegie Museum of Natural History vertebrate paleontological expedition led by K.C. Beard in November 2000. Approximately 113 specimens were collected from two sites within the Bashi lowstand unit, field numbers PW0001 and PW0002. These are designated as USNM (National Museum of Natural History, where all specimens are deposited in the Division of Paleobotany) localities 43409 and 41312, respectively, and referred hereafter as "site 1" and "site 2" for simplicity. Site 1 (N32.35643°, W88.68343°, GPS ±4 m, NAD27 CONUS datum), within a temporary outcrop excavated for a bridge support pylon related to the construction of a Wal-Mart, produced 58 specimens. The lithology at site 1 is fine-medium, burrowed white sands; leaf fossils with relatively good preservation of venation detail occur in interbedded shale and siltstone units. These specimens were more prone to desiccation and fragmentation than at nearby site 2 (N32.35600°, W88.68430°). This locality produced 55 specimens immediately above the T4 sand, from "overburden" generated by vertebrate collecting in that unit. At site 2, the lithology was finer grained (clay rich) and darker (more organic material); the plant specimens were less fragmented than those from site 1, but fine venation details were generally not preserved (i.e., the opposite attributes to site 1).

At both sites, the fossiliferous matrix was extremely friable and moist. To prevent immediate desiccation cracking after exposure to air, fossil plant specimens were brushed lightly with a 1:1:1 solution of ethanol, glycerin, and water (after Call et al. 1993). Specimens were then immediately wrapped in paper, labeled, wrapped again in aluminum foil, and allowed to dry slowly in storage. Unwrapping of specimens occurred in June 2005. Some damage occurred as expected, but the procedure conserved the majority of specimens.

Plant remains were prepared using air scribes; detailed preparation was done using fine-tipped, hand-held carbide rods. Fine-haired paint brushes and compressed air were used to remove debris from leaf surfaces. Prepared leaves were examined under a Nikon SMZ-1500 stereomicroscope. Up to fifth order venation was preserved, but vein detail was often absent due to coalification or poor preservation. Photographs (excepting most cuticle images, see below) were taken using a Nikon Coolpix 8800 digital camera (Nikon, Melville, New York, USA) and a Nikon DXM-1200F camera mounted on the stereomicroscope.

Cuticle, where present, was first examined on the stereomicroscope with normal light and using an X-Cite 120 epifluorescence illumination unit (EXFO Electro-Optical Engineering Inc., Quebec, Quebec, Canada). Unprepared, uncoated cuticle was observed using an FEI Quanta 2000 ESEM. Cuticle was further examined by directly mounting unprepared cuticle with Cytoseal 60 mounting medium (Richard-Allan Scientific, Kalamazoo, Michigan, USA) for microscopic observation. An Olympus BX61 Epi-Fluorescence Microscope, and an Olympus FV1000 Laser Scanning Confocal Microscope (Olympus America Inc., Melville, New York, USA) were used for digital microscopy and fluorescence imaging of cuticle. The confocal excitation wavelength was 633 nm, and the emitted wavelength captured for the images was 647 nm. Microscopy images were deconvolved, and image stacks were merged using Autoquant 9.3 (Media Cybernetics, Bethesda, Maryland, USA). All images were further processed and composed using Adobe Photoshop (Adobe Systems Incorporated, Seattle, Washington, USA).

Fossil plants were organized into 18 full-leaf morphotypes (RH- prefix), based on shared suites of venation and leaf form characteristics sufficient to distinguish them (Ash et al. 1999), and two distinct cuticle morphotypes, all described below. Exemplary specimens of each morphotype are illustrated and assigned USNM repository numbers. The quantity is noted of any additionally referred, deposited specimens that are not illustrated or individually catalogued. Some morphotypes were assignable to taxonomic groups, and one of these merited full typification as a new taxon. Otherwise, we did not develop new nomenclature or formally designate type specimens because the preservation quality was not sufficient to make comparisons to pre-existing type collections. Margin preservation and number of morphotypes were not adequate for paleoclimate estimation from leaf physiognomy (e.g., Wilf 2000). Descriptions below begin with the distinguishing features that best delineate the morphotype within the assemblage. Classifications of angiosperms follow the Angiosperm Phylogeny Working Group (2003).