Ten almost complete specimens have been studied in detail; nine coming from C. D. Walcott's original collection, one from the recent re-investigation. The cuticle is preserved as a thin, dark layer; the external surface was apparently smooth, except for striations on the frontal process and adjacent to the mouth. Dorsally on the short cephalon were five eyes, thought to have been compound, the inner and outer pairs pedunculate, the median not stalked. From the anteroventral slope of the cephalon arose a long, flexible frontal process, divisible into a longer, cylindrical proximal portion, and a shorter, broad distal portion. The latter was divided longitudinally, each half bearing a group of long spines, directed inward and forward. The process probably contained a median, fluid-filled canal. The mouth was situated on the vertical, posteroventral wall of the cephalon, the alimentary canal U-shaped. The cylindrical axial region of the trunk tapered slightly backward, the alimentary canal situated ventrally and extending to the tip. The trunk was divided into a main portion of 15 segments, subequal in length, and a short posterior portion lacking segmentation. The junctions between segments gave a limited flexibility to the body. Each segment of the main portion of the trunk bore a pair of thin lateral lobes, directed downward and outward, overlapping, of maximum width medially, the lobes progressively more strongly prolonged backward. Dorsal to lobes 2-15, a paddle-shaped gill was attached near the base of the lobe. The ventral surface of the gill was flat, the outer, dorsal surface bearing imbricated, thin lamellae. The gills lay between adjacent, overlapping lateral lobes. Internally, in the main portion of the trunk what may have been diverticula of the gut are preserved, extending into the proximal portions of the lateral lobes. The posterior portion of the trunk bore three pairs of thin, lobate blades, directed upward and outward, overlapping in the opposite sense to the lateral lobes, the entire structure forming a tail fan. The dorsal margin of the tip of the axial region of the fan appears to have borne a pair of spines. The body is preserved with thin layers of rock between such parts as left and right eyes of a pair, adjacent lateral lobes, between gills and lobes, and between gill lamellae. The parts of the bodies are shown to have been entombed at varied angles to the horizontal bedding planes, and are greatly compressed. It is therefore considered that individuals were trapped in a cloud of sediment in suspension, moving along the sea bottom, and buried as it settled out. If so, the animal was benthonic in habit. Opabinia regalis may have ploughed shallowly in the bottom mud, propelled by movement of the lateral lobes. The eyes are presumed to have been capable of detecting movements in the surrounding waters, and the frontal process to have been used to explore the mud for food and bring it to the backward-facing mouth. The posterior region of the trunk may have aided in producing water currents over the dorsal surface of the body, or have aided in steering if the animal was capable of swimming. No structures that appear to have been antennae, and no other jointed appendages, have been observed, and the gills are not trilobite-like. O. regalis is not considered to have been a trilobitomorph arthiopod, nor is it regarded as an annelid. It may be descended from segmented animals from which arthropod phyla and/or annelids were derived.