Hyperelongate neural spines forming a prominent dorsal “sail” are known in eight genera distributed between two families of pelycosaurian-grade synapsids. Although the function(s) of the sail remain disputed, most researchers assume that resilient soft tissue stretched between the elongate neural spines, extending to the distal tips. Hypotheses to explain the purpose of the sail have included thermoregulation (Romer & Price, 1940; Bramwell & Fellgett, 1973; Haack, 1986; Tracy et al., 1986; Bennett, 1996; Florides et al., 1999) and sexual selection (Tomkins et al., 2010). In this paper, we analyze the natural pathologies found in the neural spines of a very large pelycosaur, Dimetrodon giganhomogenes, as a natural experiment: What would ensue in the event of sail breakage and what does that tell us about sail structure, development, maintenance, and the orientation of the sail?

A series of seven associated neural spines from fmnh UC 1134 demonstrate subtle though distinctly abnormal rugosities, a sign most often indicative of a well-healed hard callus of bone fracture. Microstructural examination revealed surprising facts: not only did the abnormal bone areas prove NOT to be fracture hard callus, but the abnormal tissue reflected underlying material failure resulting from slippage between adjacent lamellae of bone. Moreover, the characteristic cranial and caudal orientation of the deep longitudinal grooves contributing to the classic dimetrodont figure-8 spine cross section was rapidly reestablished in vivo by a combination of osteoclastic resorption and additional lamellar deposition of bone to regain the “correct” pre-injury orientation, underscoring the architectural importance of the dumbbell shape in resisting lateral bending. This bone disruption and repair occurred at least five seasons before death, which explains the well-healed external appearance of the lesions. The absence of vascular communicating canals casts doubt on the widely held hypothesis that these grooves contained blood vessels that supplied a thermoregulatory sail. Furthermore, the distal morphology of spines in more complete specimens, including the type fmnh UC 112 and omnh 01727, suggests that the dorsal margin of the sail was located well proximal to the tips of the elongate neural spines. The cross-sectional architecture of the spines suggests a further hypothesis: that the proximal portion of the sail may have also functioned as an energy storage device, facilitating fast locomotion in this top predator.