In 1912 Hudson described a Middle Ordovician starfish as Protopalaeaster narrawayi and made it the type of his new genus. In 1914 Spencer suggested that it was conspecific with Palaeaster caractaci Gregory. Also in that year, Schuchert transferred Hudson's species to Hudsonaster, for which Stuertz, author of the genus, had selected Palasterina rugosa Billings as type species in 1900. In 1916, Spencer followed Schuchert in placing narrawayi in Hudsonaster and used Palaeaster caractaci as the type species for his new genus Caractacaster, thereby refuting his previous suggestion of synonymy. The ambulacral plates of Protopalaester narrawayi are closely set, so that the "podial openings" (notches for the admission of the ampullae) are mere slits, if indeed they exist at all. In contrast, those in Hudsonaster matutinus (Hall) have very large openings. According to Schuchert, the ambulacrals of Hudsonaster incomptus (Meek) are "so far as can be determined, like those in P. matutinus." The ambulacral plates of the type species, Hudsonaster rugosus, are still unknown. Because of the strong aboral resemblance of H. rugosus to H. incomptus, however, it seems likely that these 2 Richmond contemporaries are congeneric. Protopalaeaster has not been shown to be the same as Hudsonaster, and should be restored, as proposed by Sardeson in 1928. In the author's present understanding, it is a valid genus - but not because of the characters attributed to it by the original author, who confused oral and aboral sides in the halved, immature holotype. Details of the aboral side of P. narrawayi are revealed in previously unstudied specimens. Its flat-topped disk is surmounted by a coronet of 10 stellate plates: 5 first radials and 5 first supramarginals. These plates are strongly convex, extending above the general level of the summit. A distinct centro-dorsal plate is separated from each first radial by 2 centroradial plates. Proximal to each first supramarginal and between the centroradial plates of adjacent radii is a centrointerradial. The second supramarginals are also stellate and joined at their tips, except in the interradius containing the madrepore plate. The madrepore plate fits into a recess between modified second supramarginals and slopes down from its contact with the first supramarginal. It is traversed by numerous small pores spaced along narrow, anastomosing, more or less radial grooves. In the interradius to the right of the madrepore, a low pyramid of triangular plates between the centro-dorsal and the centrointerradial plates is interpreted as the anus. The ambulacral plates form a high vault within each arm, but they are not appressed against the radials or supramarginals.