BackgroundAllosauroidea has a contentious taxonomic and systematic history. Within this group of theropod dinosaurs, considerable debate has surrounded the phylogenetic position of the large-bodied allosauroid Acrocanthosaurus atokensis from the Lower Cretaceous Antlers Formation of North America. Several prior analyses recover Acrocanthosaurus atokensis as sister taxon to the smaller-bodied Allosaurus fragilis known from North America and Europe, and others nest Acrocanthosaurus atokensis within Carcharodontosauridae, a large-bodied group of allosauroids that attained a cosmopolitan distribution during the Early Cretaceous.Methodology/Principal FindingsRe-evaluation of a well-preserved skull of Acrocanthosaurus atokensis (NCSM 14345) provides new information regarding the palatal complex and inner surfaces of the skull and mandible. Previously inaccessible internal views and articular surfaces of nearly every element of the skull are described. Twenty-four new morphological characters are identified as variable in Allosauroidea, combined with 153 previously published characters, and evaluated for eighteen terminal taxa. Systematic analysis of this dataset recovers a single most parsimonious topology placing Acrocanthosaurus atokensis as a member of Allosauroidea, in agreement with several recent analyses that nest the taxon well within Carcharodontosauridae.Conclusions/SignificanceA revised diagnosis of Acrocanthosaurus atokensis finds that the species is distinguished by four primary characters, including: presence of a knob on the lateral surangular shelf; enlarged posterior surangular foramen; supraoccipital protruding as a double-boss posterior to the nuchal crest; and pneumatic recess within the medial surface of the quadrate. Furthermore, the recovered phylogeny more closely agrees with the stratigraphic record than hypotheses that place Acrocanthosaurus atokensis as more closely related to Allosaurus fragilis. Fitch optimization of body size is also more consistent with the placement of Acrocanthosaurus atokensis within a clade of larger carcharodontosaurid taxa than with smaller-bodied taxa near the base of Allosauroidea. This placement of Acrocanthosaurus atokensis supports previous hypotheses of a global carcharodontosaurid radiation during the Early Cretaceous.

We image deep crustal structure using marine seismic refraction data recorded by a linear array of ocean-bottom seismometers in the Gulf of Mexico Basin Opening project (GUMBO Line 3) in order to provide new constraints on the nature of continental and oceanic crust in the northeastern Gulf of Mexico. GUMBO Line 3 extends~524 km from the continental shelf offshore Pensacola, Florida, across the De Soto Canyon and into the central Gulf basin. Travel times from long offset, wide angle reflections and refractions resolve compressional seismic velocities and layer boundaries for sediment, crystalline crust, and upper mantle. We compare our results with coincident multichannel seismic reflection data. Our velocity model recovers shallow seismic velocities (~2.0-4.5 km/s) that we interpret as evaporites and clastic sediments. A Cretaceous carbonate platform is interpreted beneath the De Soto Canyon with seismic velocities >5.0 km/s. Crystalline continental crust thins seaward along GUMBO Line 3 from 23-10 km across the De Soto Canyon. High seismic velocity lower crust (>7.2 km/s) is interpreted as extensive syn-rift magmatism and possibly mafic underplating, common features at volcanic rift margins with high mantle potential temperatures. In the central Gulf basin we interpret thick oceanic crust (>8 km) emplaced at a slow full-spreading rate (~24 mm/yr). We suggest a sustained thermal anomaly during slow seafloor-spreading conditions led to voluminous basalt flows from a spreading ridge that overprinted seafloor magnetic anomalies in the northeastern Gulf of Mexico.

We use air gun data recorded by ocean bottom seismometers to constrain the velocity structure along Gulf of Mexico Basin Opening Line 4, a profile extending from the northwestern Florida peninsula across the Florida Escarpment to the central Gulf of Mexico. Moderately thinned continental crust with a Moho depth of 32-33 km, average sediment thickness of 6 km, and an average crustal thickness of 27 km is interpreted on the northeast end of the profile offshore Florida. Thinned and intruded continental crust is identified over a horizontal distance of 225 km where the crustal layer thins from 25 km to 6-7 km; mean seismic velocities of the crust in this region increase from 6.55 km/s to 6.95 km/s from northeast to southwest and are evidence for increased magmatic input as rifting developed. Oceanic crust with an average thickness of 5.6-5.7 km is observed over a distance of 175 km on the southwest end of the profile, with an extinct spreading ridge with an axial valley morphology imaged on a coincident seismic reflection profile. Anomalously high upper oceanic crust velocities of 6.0-6.7 km/s are interpreted as massive basalt flows and could reflect increased temperatures during emplacement. Integrating well, seismic reflection and our seismic refraction data allow us to estimate a full-spreading rate of 2.2 cm/yr for seafloor spreading along the profile; this indicates that oceanic crust was emplaced at a slow-spreading center.

The opening of the Gulf of Mexico was an important Mesozoic tectonic event that provides new insight in the role of magmatism and lithospheric stresses in the initiation of continental rifting. A new seismic velocity profile based on seismic refraction data in the northwestern Gulf of Mexico offshore Texas, where the basin started opening in the Early Jurassic, shows a rifted margin with strong lateral heterogeneity beneath the shelf and slope. The structure of the thinned crust is consistent with large-scale extensional faulting and moderate amounts of synrift magmatism before continental breakup. These new seismic constraints do not indicate the presence of a volcanic margin along the Texas coast, as has sometimes been proposed based on magnetic data. The Laurentian continental lithosphere of central Texas may have been too thick at the onset of rifting (>100 km) to let magmatic diking control the extension.