OF THE UNIVERSITY OF LIBRARY SMITHSONIAN INSTITUTION UNITED STATES NATIONAL MUSEUM Bulletin 88 REVISION OF PALEOZOIC STELLEROIDEA WITH SPECIAL REFERENCE TO NORTH AMERICAN ASTEROIDEA BY CHARLES SCHUCHERT Professor of Paleontology, Yale University Sew Haven WASHINGTON GOVERNMENT PRINTING OFFICE 1915 SMITHSONIAN INSTITUTION UNITED STATES NATIONAL MUSEUM Bulletin 88 REVISION OF PALEOZOIC STELLEROIDEA WITH SPECIAL REFERENCE TO NORTH AMERICAN ASTEROIDEA BY ' '*" CHARLES SCHUCHERT Professor of Paleontology , Yale University New Haven WASHINGTON GOVERNMENT PRINTING OFFICE 1915 PALEONTOLOGY BULLETIN OF THE UNITED STATES NATIONAL MUSEUM. ISSUED MARCH 20, 1915. ADVERTISEMENT. The scientific publications of the United States National Museum consist of two series, the Proceedings and the Bulletins. The Proceedings, the first volume of which was issued in 1878, are intended primarily as a medium for the publication of original, and usually brief, papers based on the collections of the National Museum, presenting newly-acquired facts in zoology, geology, and anthro- pology, including descriptions of new forms of animals, and revisions of limited groups. One or two volumes are issued annually and dis- tributed to libraries and scientific organizations. A limited number of copies of each paper, in pamphlet form, is distributed to specialists and others interested in "the different subjects as soon as printed. The date of publication is printed on each paper, and these dates are also recorded in the tables of contents of the volumes. The Bulletins, the first of which was issued in 1875, consist of a series of separate publications comprising chiefly monographs of large zoological groups and other general systematic treatises (occasionally in several volumes), faunal works, reports of expeditions, and cata- logues of type-specimens, special collections, etc. The majority of the volumes are octavos, but a quarto size has been adopted in a few instances in which large plates were regarded as indispensable. Since 1902 a series of octavo volumes containing papers relating to the botanical collections of the "Museum, and known as the Contribu- tions from the National Herbarium, has been published as bulletins. The present work forms No. 88 of the Bulletin series. RICHARD RATHBUN, Assistant Secretary, Smithsonian Institution, In charge of the United States National Museum. WASHINGTON, D. C., January 27, 1915. 3 M650670 TABLE OF CONTENTS. Page. Preface 9 Terminology as applied to the skeleton of Paleozoic Asteroidea 13 Finding list of Paleozoic Stelleroidea 18 Remarks on the development of Paleozoic Stelleroidea 27 Geological 27 Distribution 27 Preservation and occurrence 28 Asterid evolution 30 General observations 30 The most primitive fossil starfish 31 Origin of the wriggling type of starfish 32 The asterid radicle 34 Ontogeny 36 Regeneration 37 Number and character of rays 38 Development of the parts 39 Anus 39 Madreporite 39 Spinosity 41 Ambulacralia 41 Adambulacralia 43 Inframarginalia 43 Supramarginalia 45 Radialia 46 Accessory plates 47 Ocular plates -^, 48 Centro-dorsal 48 Primary disk-plates 49 Axillary ossicles 49 Interbrachial areas 50 Classification 51 Catalogue of Paleozoic Stelleroidea 51 Description of genera and species 51 Class Stelleroidea 51 Subclass Asteroidea 51 Phanerozonia 51 Family Hudsonasteridae 53 Hudsonaster 53 Siluraster 65 Family Palseasteridse 66 Palseaster , 67 Australaster 72 5 6 TABLE OF CONTENTS. Description of genera and species Continued. Class Stelleroidea Continued. Subclass Asteroidea Continued. Phanerozonia Continued. p age . Family Promopalaeasteridae 73 Subfamily Mesopalaeasterinae 74 Mesopalseaster 74 Spaniaster 95 Miomaster 96 Devonaster 97 Subfamily Promopalasasterinse 102 Promopalasaster 102 Subfamily Anorthasterinae 125 Anorthaster 125 Family Xenasteridae 128 Xenaster 129 Agalmaster 131 Rhenaster 133 Eifelaster 133 Trimeraster 134 Family Neopalaeasteridse 134 Neopalseaster 134 Family Palasterinidae 138 Petraster 138 Lindstromaster 148 Palasterina 150 Uranaster 154 Palseostella 155 Pseudopalasterina 156 Family Lepidasteridae 157 Lepidaster 158 Helianthaster 159 Lepidasterella 160 Family unknown 161 Astropecten 161 Ataxaster 161 CryjDtozonia .- 162 r~Family Stenasteridae 163 / Stenaster 163 Tetraster 167 Family Monasteridaa 170 Monaster 170 Family Urasterellidee 172 Urasterella 173 Family Calliasterellidaa 190 Calliasterella 190 Family Compsasteridse 191 JaeJcelaster 192 Compsaster 192 Family Schuchertiidae 194 Schuchertia 195 Family Palasteriscidae 199 Palasteriscus 200 Echinasterella... 200 TABLE OF CONTENTS. 7 Description of genera and species Continued. Class Stelleroidea Continued. Subclass Asteroidae Continued. Cry ptozonia Continued . Family Palasteriscidae Continued. Page. Loriolaster 201 Cheiropteraster 202 Family Schoanasteridae . 202 Schcenaster 202 Family Palseosolasteridae 207 Palxosolaster 209 Echinasterias 211 Echinodiscaster 211 Echinostella 212 Medusaster 212 Subclass Auluroidea 213 Order Lysophiurae 215 Family Protasteridae 215 ^fseniaster 216 [_ Eophiura. . 222 Bohemura 223 L-Palseum 223 Protester 224 Akpidaster 228 Gregoriura 233 Bunderibachia 234 Palxophiomyxa 234 Family Palaeophiuridae 235 Palseophiura 235 Sturtzura 236 f-JSygasterella 237 ; Ptilonaster 239 Family Encrinasteridse 241 Encrinaster 241 Order Streptophiurae . 246 Family Ophiurinidae 246 Ophiurina 246 Tremataster 247 Family Lapworthuridae 248 Squamaster 249 ^Lapworthura 250 r Sturtzaster 252 _ Bdellacoma 254 THhopalocoma 254 Hallaster 254 Sympterura 256 Furcaster 261 Palastropecten 261 Family Eoluidiidae 262 Eoluidia 262 Eospondylus 263 Miospondylus 263 Family Aganasteridae 263 Aganaster 263 Family Cholasteridae 265 Cholaster. . . 265 8 TABLE OF CONTENTS. Description of genera and species Continued. Class Stelleroidea Continued. p age . Subclass Ophiuroidea 267 Family Onychasteridee 268 Onychaster 268 Forms whose relationships are unknown 274 Cribellites carbonarius 274 Order Ophiocistia , 275 Family Eucladiidse 276 Eucladia 276 Euthemon 279 Bibliography 281 Explanation of plates 289 Index.. 303 REVISION OF PALEOZOIC STELLEROIDEA WITH SPECIAL REFERENCE TO NORTH AMERICAN ASTEROIDEA By CHARLES SCHUCHEET Professor of Paleontology, Yale University, New Haven. PREFACE This memoir had its inception in 1896, and came about through the force of circumstances. The writer, at that tune assistant curator in the United States National Museum, was made aware of the desire of Mr. I. H. Harris, of Waynesville, Ohio, to present to that institu- tion his extensive collection of Ordovicic fossils. He had long known Mr. Harris, having met him as a young man, about the year 1876, at which time the Harris collection was already widely known, not only for its quantity, but more especially for its many choice speci- mens of the rarer species of starfishes, crinids, and trilobites. Twenty years later came Mr. Harris's desire to present to the National Museum the collection upon which he had been at work for fifty years. This gift brought to the National Museum not less than 40 asterids, 35 (31 on one slab) ophiurids from the Cincinnatic strata (the great majority from the Waynesville formation), besides an asterid and 12 specimens of Ony chaster flexilis from the Keokuk formation at Crawfordsville, Indiana. Some years later the National Museum acquired the very valuable E. O. Ulrich collection, in which there were also many good starfishes. Accordingly, it is safe to state that in no other museum are there so many Ordovicic asterids, in so great a variety and in such excellent preservation. The material of the Harris collection seemed to show that its com- plete study would not only reveal much new knowledge, but give an insight as well into the probable evolution of the Paleozoic starfishes. When these facts were made known to the then director of the National Museum, Dr. G. Brown Goode, he asked the writer to invite Mr. Harris to join him in a work on the Ordovicic starfishes. This was done and Mr. Harris was greatly pleased to undertake the joint authorship. He did all he could to further the work, but it went very slowly, because the present writer was so much occupied with curatorial work as well as the arranging of museum exhibits for the expositions in which the United States Government participated and which came 9 10 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. nearly every year. Unfortunately, Mr. Harris died on October 17, 1897, when the work was still in the first stages. At odd times during subsequent years the writer kept up his studies and in the spring of 1904 the work on the Asteroidea was practically completed. Then came his appointment to the chair of paleontology at Yale University. Here new duties and new lines of thought kept him away from the study of the Paleozoic ophiurids which had been planned, but the hope was strong that somehow the time would be found to complete the work. However, this hope has not materialised, others have gone to work on the starfishes, and one genus and general- ization after another has slipped away. Therefore it was thought best to revise what had been done, bringing the asterids up to date. In regard to the ophiurids, it was found impossible to restudy the specimens, but as the writer had kept abreast of the literature, this was put into order and is here presented, in the hope that subsequent workers will find it useful in enabling them to find the references on any species and the taxonomic status of any form. A study of the specimens in the Harris collection made it plain that no satisfactory taxonomic and systematic results would follow unless the undertaking also embraced many other Paleozoic forms, of Europe as well as of America. Species and genera had been pro- posed in a very loose manner; in fact, most of the Paleozoic litera- ture on asterids and ophiurids up to the tune this work began had no other value than to force the reviser under the rules of nomencla- ture to recognize the names proposed. The International Rules of Zoological Nomenclature have here been strictly followed. Under these circumstances, all of the type-specimens had to be seen, and most of them have been accessible. In this way the work finally came to be extended to all Paleozoic Stelleroidea. To make it plain how little the American Paleozoic starfishes were known in 1889, it will suffice to give one example: Miller in his North American Geology and Palaeontology lists 23 species under Palseaster'j in the present work there is but a single form of this genus, the genotype, P. niagarensis, the remainder having been distributed among seven genera, all new and proposed in the past few years or at this tune. As a result, every paleontologist will be obliged to disregard what he has learned re- garding American Paleozoic starfishes and start anew. To facilitate this, a list has been prepared and is here presented of all the names so far given to Paleozoic asterids and ophiurids, with cross references to the names adopted in this work. Now that this study is finished, at least for the present, it is plain to the author that his species and genera are also in some cases too comprehensive. The future student will restudy the specimens along with the new material and go more deeply into the detailed structure of the parts. He who attempts this, along with painstaking recon- EEVISION OF PALEOZOIC STELLEROIDEA. 11 struction of the skeleton as it was in life, will learn where to separate and what to reject. During the past 10 years, Schondorf has revised most of the German asterids and has studied deeply into the structure of the ophiurids. His work is very detailed and the best extant on the Paleozoic forms. Every student of Stelleroidea intending to do thorough work will have to examine his results. In classification he has laid the founda- tion, and has shown that some asterids are not Asteroidea at all, but that they and the bulk of the so-called ophiurids of the Paleozoic must, because of their peculiar structure, ambulacral and otherwise, be referred to another class, the Auluroidea. In regard to the evolution of the Asteroidea, the writer hoped to find the time to present his views concerning it in detail in this memoir, but that also is not possible. However, the main lines and often the generic directions of organic change are stated and will be found either in the introductory pages or scattered through- out the work, generally in the remarks under the generic discussions. These results in connection with the work of Schondorf, it is thought, ought soon to place the Paleozoic Stelleroidea in such order that a final classification, based on ontogeny, chronogenesis, and phyl- ogeny, and embracing not only the Paleozoic forms but those of Mesozoic, Cenozoic, and Recent time as well, can be made. Many museums and individual paleontologists have loaned the writer the specimens in their collections or in their keeping, and he wishes here to thank them, one and all, for these loans. The greatest bulk of the material studied is of course in the United States National Museum in the Harris and Ulrich collections; a number of Trenton specimens and some from the English Siluric have also been given to the National Museum by Mr. Walter R. Billings, of Ottawa. Since the writer has been in New Haven, Dr. R. S. Bassler, of the National Museum, has often kindly helped him to further his studies, and during the past year has made a large number of photographs of the specimens under his ' charge. The many Mississippic speci- mens in the Frank Springer collection have not been studied. The author is greatly indebted to Dr. F. A. Bather, of the British Museum (Natural History), for a large number of gutta-percha and wax squeezes made by him of certain species in that great museum. These casts, which are all in the United States National Museum, have enabled the writer to understand several genera that otherwise he could not have worked out. He is also indebted to him for many bibliographic corrections and suggestions. The second largest American collection is at Harvard University, in the Museum of Comparative Zoology. Most of the material is in the Charles D. Walcott, Charles B. Dyer, and Charles Wachsmuth collections. In the first-named collection there is excellent material from the Middle Ordovicic; the second has many fine specimens 12 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. from the Cincinnatic formations; while the Wachsmuth material has species from the Mississippic. To all of this material the author has had access through the kindness of Director Samuel Henshaw and Profs. R. T. Jackson and P. E. Raymond. At the University of Chicago there is a great deal of excellent material in the W. F. E. Gurley, Charles Faber, and U. P. James collections, to which Weller has added other specimens. All of this material has been studied through the courtesy of Prof. Stuart Weller. Most of it is from the Cincinnatic formations. All of the specimens described by E. Billings and now in the Vic- toria Memorial Museum in Ottawa, Canada, have been seen at one tune or another. In this matter the writer was much aided by the late Dr. J. F. Whiteaves and more recently by Drs. P. E, Raymond and E. M. Kindle, to the latter of whom thanks are also due for several photographs reproduced in this memoir. Most of the speci- mens are from the Ordovicic. In the University of Toronto there is much material from the Ordovicic and Siluric in the Sir Edmund Walker collection, and there may be studied a great many specimens of Stenaster salteri. All of this has been seen through the great liberality of Prof. William A. Parks. The genotype of Pakeaster, from the museum of Cornell University, was loaned to the author at different times by Prof. H. S. Williams. It is one of the best preserved of Paleozoic specimens and has become very valuable because of the continual discussion that centers around the species, P. niagarensis. To all the specimens in the Peabody Museum of Yale University access was had through the kindness of the late Prof. Charles E. Beecher. Among them is the unique branching form, Eucladia beecfieri. Photographs have been made for this memoir by Dr. John M. Clarke of specimens in the New York State Museum, and by Prof. George H. Hudson, of Plattsburg, New York, of other material. Prof. Arthur M. Miller loaned two specimens that he collected and deposited in the State University, Lexington, Kentucky. Prof. George W. Harper, of Cincinnati, loaned a specimen of Palseaster darkana ( = Hudsonaster incomptus) . Nearly all the original drawings were traced from the specimens by aid of the camera lucida. These outlines were then inked and shaded by the late Dr. J. C. McConnell, a draftsman and artist of the first rank. To Miss Clara Mae LeVene the writer is especially thankful for the great interest she has taken hi getting the manuscript into order for the printer, keeping the bibliography up to date and correctly citing it, and for the careful reading of the proof. REVISION OF PALEOZOIC STELLEBOIDEA. 13 TERMINOLOGY AS APPLIED TO THE SKELETON OF PALEOZOIC ASTEROIDEA. 1 Abactinal. See Body-wall. The integumentary skeleton opposite to that having the ambu- lacral grooves, or that of the upper or dorsal side of the animal. Also known as the aboral surface. See Actinal. Aboral. See Abactinal. Accessory plates. This term refers to the abactinal longitudinal columns, or transverse rows, of small plates between the radials and supramarginals (=radial accessory plates or dorso- laterals), or between the inframarginals and supramarginals (=ambital plates or intermarginals), or where there are no radials, to all the ossicles between the two supramarginal columns. There may also be accessory actinal plates which are termed accessory interbrachials, accessory actinals, or inter actinals. The accessory plates form the secondary skeleton and are inserted between the columns of the primary skeleton. It is in the secondary skeleton that much of the specific and generic evolution takes place. See Interbrachial. Actinal. See Body-wall. The integumentary skeleton with the ambulacral grooves, or that of the under or ventral side of the animal. Also known as the oral surface. Adambulacrals. The single columns of prominent plates adj oining the ambulacrals . There are never more than two of these columns in a ray, one on each outer side of the depressed ambulacrals. They are generally large and thick and bear the larger, or adam- bulacral, spines . These spines may be inwardly directed and serve as a protection for the tube-feet. The oral armature in Paleozoic species usually consists of five pairs of modified adambulacrals, each pair being the oral projection of two united columns of adjoining rays. They are also called Adambulacralia. Also see Inter- brachial. Ambital. The abactinal space between the infra- and supramarginal columns, occupied by accessory plates. These are best developed in the axillary disk areas (disk ambi- tals), while those of the rays are referred to as ray ambitals. They are also known as intermarginals. Ambulacral. Ambulacral grooves (five or more), plates, or ossicles are situated actinally and per- radially along the center of the rays and disk. They are also called ambula- cralia. The ambulacral columns are in pairs, the interlocking movable plates of adjoining columns may be alternate or opposite (always so in recent forms), and the plates of a column may be superposed like the tiles on a roof. The podial openings through which the fleshy tube-feet protrude are situated be- tween the sutures of adjoining plates. In early forms there are but two columns of podial openings in each ray, one on each side of the axial line, but in later forms there may be as many as eight columns. The ambulacral grooves may be tapering or petaloid. Anal opening. The only Paleozoic form in which an anal opening may exist visually is Hudson- aster. Here it is on the disk between the central plate and the madreporite. The anal pore is nearly always obscured in fossil starfishes, and is absent in some recent species. 1 Also in the main applicable to the Auluroidea and somewhat to the Ophiuroidea. 14 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Arcs. See Axil. Axil. The angle formed by the junction of the rays. Also known as arcs. In primitive forms each axil is occupied by a single plate, the axillary inframarginal. In others this plate is crowded orally and its place occupied by two prominent proximal inframarginals. In forms with large disks having well-developed marginals these are termed disk marginals. Also see Ambital and Interbrachial. Bivium. The two rays of a starfish that inclose between them the madreporite. When oriented toward the student, the ray on the left is numbered 1 and the other on the right is numbered 5. A line drawn through the madreporite and the center of the upward directed ray (ray 3) divides the animal into symmetrical halves (=left and right sides). The upper left-hand ray is numbered 2 and the one opposite is 4. Rays 2, 3, and 4 form the trimum. Body-wall. The outer surface of both disk and rays, made up of movable calcareous plates which are either closely adjoining or form a more or less loose network of ossicles. These are held together in the living animal by connective tissue and muscular fibers, which after death soon decompose, freeing the plates. It is for this reason that fossil starfishes are so rarely found entire. Imbedded in the body-wall everywhere are blunt, short, calcareous spines, arranged in a more or less definite order, and often movable upon the underlying plates. Covering the whole surface of the body, including the spines and pedicellarise, is a delicate membrane or skin, clothed externally with closely placed vibratile cilia, which keep the integument clean. The calcareous skeleton of starfishes is therefore integumentary. Csecal pores. The openings between reticular plates for the protrusion of the respiratory csecal processes. These pores are either restricted to the abactinal area or may be distributed over the entire body. The processes are also known as papulx. Central disk. See Disk. Central disk plate. A prominent plate, most conspicuous in primitive forms, which often appears in the center of the disk. It holds the same position as the centro-dorsal plate in embryonic crinids of the genus Antedon. Also called the abactinal-central plate or simply central plate. Centro-dorsal plate. See Central disk plate. Disk. The central portion of a starfish. In this work, the term generally applies only to the abactinal central area formed either by the union of the rays, or by the rays plus the axillary areas. The mouth is in the center of the disk on the under or actinal side . Disk plates are the small plates over this area on the abactinal side . Central disk plate refers to a larger primary plate in the center of the disk. There may be other large primary plates, but these are more properly referred to the rays. See Radial, Interradial, Marginal, and Ambital. Inframarginal. See Marginal plates. Interbrachial. The actinal and marginal areas between the rays . These spaces have plates derived from various regions. In primitive forms, but a single plate is present, the axillary interbrachial, derived by crowding orally the axillary inframarginal. In other genera, inframarginals are crowded in pairs orally and form interbrachial REVISION OF PALEOZOIC STELLEBOIDEA. 15 areas, the interbrachial marginals. Or the adambulacrals may enter into the construction of these areas, forming interbrachial adambulacrals. When the adam- bulacrals and inframarginals are separated by small plates, the latter are called accessory interbrachials; they may extend almost to the distal ends of the rays. Interradial. A point half-way between the perradii. The interambulacral areas are inter- radial in position. See Perradial. Interradial plates. These abactinal plates are rarely retained in living mature starfishes. In many Paleozoic genera they are large, interradial in position, five in number, and lie inside the basal or proximal supramarginals and between the primary radials. They may be homologous with the basals of crinids. They are also known as Madreporite ormadreporic plate. A sieve-like or spongy plate, with many irregular openings for the circulation of sea water into the so-called "stone canal, " or better, into the water- vascular system. It is commonly granular or striated, and in Asteroidea is abactinal on the disk, basally situated between two rays, but in the Auluroidea is prob- ably always actinal in position. In recent multiradiate forms there are species with more than one madreporite. Marginal plates. In general the plates which make up the marginal columns bounding the rays, or disk, or the entire animal. They are parts of the primary skeleton. In the early Paleozoic genera the inframarginal and supramarginal columns are usually not directly superposed, making one column of superposed halves as in Meso- zoic and Recent starfishes, but the former column lies outside or laterally of the latter. The inframarginals are then the true marginals, and form a part of both the abactinal and actinal sides. In the primitive arrangement the inframarginals adjoin the adambulacrals, but in derived forms with large pen- tagonal disks they are separated from the latter by accessory interbrachial plates. The situation of the marginals is, however, not always at the margin of the disk and rays, but they may retain their primitive position beside the adam- bulacrals; this is more especially true in forms having a well-developed pen- tagonal disk. These plates are also referred to as marginalia, supramarginalia, and inframarginalia. In some forms the supramarginals are not recognizable as such, while the infra- marginals are well-developed; and in others none of the marginals are distinctly discernible. Sometimes the plates of the inframarginal columns adjoin one another, while those of the supramarginal series are separated by small acces- sory supramarginal pieces. Then the inframarginals may be separated from the supramarginals by ambital ossicles (see Ambital) . A proximal supramarginal is the proximal plate of a column and may not be homologous with the primary supramarginal in primitive forms. Proximal inframarginal is used in a similar way and refers to the proximal axillary plate, two of which of adjoining col- umns occupy an axil or lie on either side of the axillary marginal. The latter in primitive forms alone occupies the angle between the rays. The marginals in large disks without angles may be termed disk infra- and supramarginals. The large disk of some species is formed by the oral crowding of pairs of proximal inframarginals, the interbrachial marginals. Measurements. The size of starfishes is usually given in millimeters along the greater and lesser radii, that is, from the center of the mouth or disk to the tips of the rays. The symbol for this line is R. The smaller radius is from the center of the animal to the center of the interradii; the symbol is r. 50601 Bull. 8815 2 16 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Mouth angle plate. See Oral armature. Ocular plate. A single large, grooved eye-plate occupying the distal ends of rays and support- ing a sensory tentacle at the base of which occurs the eye-spot. They appear in the larval stage as the primary radials and with growth pass outward, remain- ing at the tip of the rays as the oculars. They are also known as terminals. These large plates are not present .as such in early Paleozoic genera, and are unknown before the Carboniferous. Oral. See Actinal. Oral angles. The interradial actinal areas around the mouth. Oral armature. The pairs of plates, usually five in number, composing the apices around the central actinal opening or mouth. In most Paleozoic forms the armature consists of the proximal modified adambulacral plates. These pieces are also known as Mouth angle plates. In some Paleozoic forms (Hudsonaster) there lies in front of each pair of mouth angle plates a single plate; this is known as the Torus. Gregory (1900, p. 241) writes that "the Oral Skeleton (or actinostomial ring) consists of a solid calcareous ring around the mouth. It is composed of thirty plates in a quinqueradiate starfish, there being always six times as many plates as there are rays. Each segment of the oral skeleton consists of two pairs of ambulacral, and of one pair of adambulacral ossicles. In Asterias [a cryp- tozonian] the ambulacral plates are more prominent than the adambulacrals, and project into the oral cavity." When the ambulacral elements are the more prominent, the oral skeleton is said to be of the ambulacral type. This is only present in cryptozonians. When the adambulacral ossicles are the most prominent, the oral skeleton is of the adambulacral type. This latter construction is the more primitive and occurs in all the Phanerozonia, but is also present in some of the cryptozonians. Ossicles. See Plates. Papulae. In living Asterias, a cryptozonian, from between the spicular plates there rise from all parts of the external surface short and small integumentary protuber- ances that are used for respiration. In the more heavily plated Phanerozo- nia they are limited to the abactinal surface enclosed between the supramar- ginal plates. It is probable that papulae were present in all Paleozoic forms having rounded or spicular ossicles, issuing in the open spaces between the plates. They probably were absent in the earliest closely plated starfishes, such as Hudsonaster. See Csecal pores. Paxillse. Minute calcareous processes arranged around large spines. None are known in Paleozoic genera. "Another type of spines occurs as part of the structures known as 'paxillse.' Each paxilla consists of a thick plate supporting a number of short, calcareous pillars, the summit of each of which is covered by a group of small spines. In some Phanerozonia * * * the paxillse occupy almost the whole abactinal surface of the Asteroid" (Gregory 1900, p. 247). Pedicellarids. Pincer or scissors-like calcareous appendages, attached to the spines, the plates, or the skin, which keep the body-wall clean of parasites. None have as yet KEVISION OF PALEOZOIC STELLEROIDEA. 17 been detected with certainty in Paleozoic genera. In the primitive Recent forms they are also absent, the integument being protected by an abundance of closely set spines. "The Pedicellarise of Asteroids are of four main types. The simplest form con- sists of a row of pairs of small, sessile, opposable spines; these are the 'pseudo- pedicellariae.* The members of the second set are ' sessile.' The next advance is the development of a stalk; of these pedunculate pedicellariae there are two varieties: (1) the 'forficiform,' in which the two hooks are attached to the near- est end of the basal plate nearest to them; (2) the 'forcipiform,' in which the two hooks cross one another and are attached to the end of the basal plate far- thest from them " (Gregory 1900, p. 247; also see Verrill 1914, pp. 25-34). Perradial. The area along the center of the rays. The ambulacra are perradial in position. See Interradial. Plates. Some writers restrict the term plates to the five primary radials, five interradials, and the central disk piece, and refer to all the other calcareous parts as ossicles. The word plates is here used in the widest sense, and as interchangeable with ossicles. For the sake of brevity the plates will be often referred to as margi- nalia, radialia, etc. Podial openings. The rounded openings between the ambulacral plates for the protrusion of the locomotor organs, the podia or water- tube-feet. See also Ambulacral. Primary radial. See Radial. Primary skeleton. The skeleton of a starfish is made up of primary and secondary ossicles. The primary skeleton consists of the longest inherited and the first appearing plates in the young, as the ambulacral, adambulacral, marginal and radial columns, plus the primary pieces of the disk. All the other ossicles inserted between these are of the secondary skeleton. Radial Refers only to the prominent abactinal primary plates holding a radial position. Where present, they form columns of continuous reticular or separated radialia, situated along the radial center of the rays. Proximal radials are those on or near the disk and should not be confounded with the primary radials, the probable homologues of the radials in crinids. These plates are also known as median dorsals, dorsals or carinals; also as radialia. Rays. The radial "arms " of starfishes. These are hollow and contain the digestive, reproductive, and water-vascular systems and other organs. The rays merge gradually and without sharp demarcation into the central disk. Respiratory pores. See Csecal pores. Secondary skeleton. See Primary skeleton and Accessory plates. Spines. See PaxillsR and Pedicellarids. Supramarginal. See Marginal plates. Terminals. See Ocular plate. 18 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Torus. See Oral armature. Trimum. See Bivium. Vertebral ossicles. In the Ophiuroidea each, pair of adjoining ambulacralia is united into a single plate, the vertebra of the arms. As the arms are very flexible, these ossicles articulate upon one another as do the bones in the vertebral column of verte- brates. FINDING LIST OF PALEOZOIC STELLEROIDEA. Page. Acroura (Ophiura) schlotheimii Minister. Known to the writer only from the list in Woodward, Geol. Mag., dec. 2, vol. 1, 1874, p. 7. Siluric. Agalmaster Schondorf 131 A. grandis Schondorf 132 A. intermedius Schondorf 133 A. miellensis Schondorf 132 Aganaster Miller and Gurley 263 A. gregarius (Meek and Worthen) 264 A. (?) ep. Miller and Gurley =Lapworthura (f ) ep . 251 Alepidaster Meek 228 A. flexuosus (Miller and Dyer) 231 A. granuliferus (Meek) 230 A. gregarius Meek= Aganaster gregarius 264 A. miamiensis (Miller) 233 A. f, new species -.- 230 Anorthaster, new genus 125 A. miamiensis (Miller) 127 Archaeasterias 'M.\JlleY=Xenaster 129 ? Argaster "0.2^1 Mesopalxaster 74 A. antiqua Gregory = Mesopalxaster (?) antiquus 86 Aspidosoma Goldi\i88=Encrinaster 241 A. arnoldi Goldf\i8s=Encrinaster arnoldi 243 A. eifelense SchondoTi=Encrinaster eifelensis 244 A. goldfussi Sch6ndori=Encrinaster goldfussi 243 A. grayss. Spencer=Encrinaster (?) grayse 245 A. petaloides SimonovitBch=Encrinaster petaloides 243 A. petaloides goslariensis Halfar=J!?icrmaser petaloides goslariensis 243 A.f pontis Cla,Tke=Encrinaster pontis 244 A. roemeri Schdndori=Encrinaster roemeri 244 A. schmidti Sch6ndoTi=Encrinaster schmidti 244 A. tischbeinianum Roemer=Jncrwaster tischbeinianus 244 Asterias Graham, Anthony, and 3ameB=Petraster (?) americanus 146 A. acuminatus Simonovitsch= Mesopalseaster (?) acuminatus 93 A. anthonii Dana=Pefr aster (?) americanus 146 A. antiqua ILi8mger=Lindstrdmaster antiquus 149 A. antiqua Troost= Mesopalseaster (?) antiquus 86 A. antiquata Ijocke=Promopalseaster spedosus 109 A. asperula Roemer= Urasterella asperula 188 A. asperula Roemer (pait)=Eoluidia decheni 262 A. constellata Thorent= Urasterella (?) constellata .' 187 A. matutina 'BsX\.=Hudsonaster matutinus '. 57 A. montanus Stschurowsky= Urasterella montana 189 A. primseva Salter and Soweiby=Stenaster (?) obtusus 167 REVISION OF PALEOZOIC STELLEROIDEA. 19 Page. Asterias primordialis Anon.=Promopalseaster, sp. undet 108 A. (Archseasterias) rhenana M.ulleT=Xenaster (?) rhenanus 131 A. spinosissima Roemer. An unsolved starfish from the Lower Devonic of Bundenbach, Germany. See Roemer, Palaeontographica, vol. 9, 1863, p. 147, pi. 29, fig. 4; and Schondorf, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 33. A. sp. undet. Hall 1847Ccelaster tenuiradiatus . Asteriscus antiquus Pictet=Lindstromaster antiquus 149 Astropecten Link . . 161 A. (?) schluteri Stiirtz 161 Ataxaster Jaekel 161 A. pygmxus Jaekel 161 Australaster, new genus 72 A. giganteus (Etheridge) - - 73 A. (?) stutchburii (Etheridge) 73 Bdellacoma Salter 254 B. vermiformis Salter 254 Bohemura Jaekel 223 B. jahni Jaekel 223 Bundenbachia Stiirtz 234 B. beneckei Stiirtz 234 B. grandis St\\rtz=Palaophiomyxa grandis 235 Calliaster Trautschold=CaZtoerea 190 C. mirus TT&\itschold=Calliasterella mira 190 Calliasterella, new name 190 C. mira (Trautschold) 190 Cheiropteraster Stiirtz 202 C. giganteus Stiirtz 202 Cholaster Worthen and Miller 265 C. peculiaris Worthen and Miller 266 Ccelaster Sandberger=$/>ara'aster 95 C. americanus D'Orbigny=Peraser (?) americanus 146 C. latiscutatus Sandberger=parmister latiscutatus 96 C. tenuiradiatus D'Orbigny, Prodr. de Paleont., vol. 1, 1849, p. 22. Based on an isolated madreporite described as Asterias, sp. undet., Hall, Pal. New York, vol. 1, 1847, p. 18, pi. 4, figs, lla, 116. Compsaster Worthen and Miller 192 C. formosus Worthen and Miller 193 C., new species 194 Cribellites carbonarius Tate 274 Devonaster, new genus 97 D. chemungensis, new species 101 D. eucharis (Hall) 98 E chinas terella Stiirtz 200 E. ? darwini Clarke 200 E. sladeni Stiirtz 200 Echinasterias Stiirtz 211 E. spinosus Stiirtz 211 EchJLnodiscaster Delage and Herouard 211 E. multidactylus (Stiirtz) 211 Echinodiscites SchucheTt=Echinodiscaster 211 Echinodiscus Stwrtz=Echinodiscaster 211 E. multidactylus St\ulz=Echinodiscaster multidactylus 211 20 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Page. Echinostella Stiirtz 212 E. traquairi Stiirtz 212 Eifelaster Schondorf 133 E. follmanni Schondorf 134 Encrinaster Haeckel 241 Encrinaster Haeckel (p&rt)=Protaster. . . : 224 E. arnoldi (Goldfuss) 243 E. eifelensis (Schondorf) 244 E. goldfussi (SchSndorf). . r 243 E. grayse, (Spencer) 245 E. petaloides (Simonovitsch) 243 E. petaloides goslariensis (Halfar) 243 E. pontis (Clarke) 244 E. roemeri (Schondorf) 244 E. schmidti (Schondorf) 244 E. tischbeinianus (Roemer) 244 Eoactis Spencer. See Urasterella 173 E. simplex Spencer. See Urasterella girvanensis 186 Eoluidia Stiirtz 262 E. decheni Stwtz 262 Eophiura Jaekel 222 E. bohemica, new species 222 Eophiurites Stiirtz =Eoluidia 262 E. decheni (Stiirtz) =Eoluidia decheni 262 Eospondylus Gregory 263 E. primigenius (Stiirtz) 263 Etheridgaster Gregory= Monaster 170 E. clarJcei Gregory = Monaster clarJcei 172 Eucladia Woodward 276 E. (?) beecheri, new species 278 E. johnsoni Woodward 276 E. woodwardi Sollas 277 Eugaster ILal\=EiLgasterella 237 E. continnus Ringueberg=Eugasterella ( ?) concinna 239 E. logani H&\\=Eugasterella logani 238 Eugasterella, new name 237 E. ( ?) concinna (Ringueberg) 239 E. logani (Hall) 238 Euryak annulatum DeKay. Apparently a lysophiurid. See Hall, Pal. N. Y., vol. 3, 1861, p. 134 (nomen nudum). Euthemon Sollas , 279 E. igerna Sollas 279 Furcaster Stiirtz 261 F. (?) daoulasensis (Davy) 261 F. palseozoicus Stiirtz 261 Gregoriura Chapman 233 G. spryi Chapman. . .- 234 HaUaster Stiirtz 254 #./or&m(Hall) 255 Helianthaster Clarke (part)==PaZa20$oZoster 209 Helianthaster Roemer 159 H. filiciformis Woodward 160 H. gyalum Clarke =PaZasoso Zoster ( ?) gyalum 210 REVISION OF PALEOZOIC STELLEKOIDEA. 21 Page. Helianthaster gyalum Clarke (paTtyLepidasterella babcocki 160 H. , new species Clarke Lepidasterella babcocki 160 H. rhenanus Roemer 159 H. roemeri Clarke PaZasoso Zoster roemeri 211 Hisingeraster Stiirtz =Lindstromaster 148 H. antiqua Sturtz=Lindstrdmaster antiquus 149 Hudsonaster Stiirtz 53 H. batheri, new species 65 H. incomptus (Meek) 61 H. matutinus (Hall) 57 E. milleri, new species 60 H. narrawayi (Hudson) 59 H. rugosus (Billings) 64 Jaekelaster Stiirtz 192 /. petaliformis Stiirtz 192 Lapworthura Gregory 250 L. cylindrica Parks= Tseniaster cylindricus 220 L. miltoni (Salter) 251 L. sollasi Spencer 251 L. (?)sp 251 Lepidaster Forbes 158 L. grayi Forbes 158 Lepidasterella, new genus 160 L. babcocki, new species 160 Lindstromaster Gregory 148 L. antiquus (Hisinger) 149 Loriolaster Stiirtz 201 L. mirabilis Stiirtz 201 Medusaster Stiirtz 212 M. rhenanus Stiirtz 212 Mesopaleeaster, new genus 74 M. ( ?} acuminatus (Simonovitsch) 93 M. (?) antiquus (Troost) 86 M. bellulus (Billings) 91 M. caractaci (Gregory) 92 M. ( ?) cataractensis, new species 89 M. (?) clarki (Clarke and Swartz) 94 M . ( ?) dubius (Miller and Dyer) 85 M.finei (Ulrich) 81 M. granti (Spencer) 89 M. intermedius, new species 79 M. ( ?) lanceolatus, new species 82 M. (?) parviusculus (Billings) 87 M. proavitus, new species 83 M. sha/eri (Hall) ; 77 M. ( ?) wilberanus (Meek and Worthen) 84 Miomaster Schondorf 96 M. drevermanni Schondorf 96 Miospondylus Gregory 263 M. rhenanus (Stiirtz) 263 Monaster Etheridge 170 Monaster Gregory= Australaster 72 M. darkei (De Koninck) 172 22 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Page. Monaster giganteus 'Ethendge=Australaster giganteus 73 M. stutchburii JZtheridge=Australaster stutchburii 73 Neopalseaster, new genus 134 N. crawJordsmlUnsis (Miller) 136 Onychaster Meek and Worthen 268 0. asper Miller 272 0. barrisi (Hall) 272 0. confragosus Miller 273 0. demissus Miller 274 0. flexilis Meek and Worthen 270 Ophiopege ~Bohm=Aganaster 263 Ophiura obtusa Eichwald= Pro taster (Ophiura) obtusus. 0. (?) ramosa Fahrenkohle. Unknown to writer. See Bull. Soc. Imp. Nat. Moscow, n. ser., vol. 17, 1844, p. 208, pi. 3, figs. 1-3; and Woodward, Geol. Mag., dec. 2, vol. 1, 1874, p. 7 (listed only). Coal Measures. 0. rhenana Sturtz= Miospondylus rJienanus 263 0. salteri Salter and Sowerby =Protaster ( ? ) salteri 227 0. schlotheimii Munster= Acroura (Ophiura) schloiheimii. Ophinrella Stwrtz=Eospondylus 263 0. primigenia St\irtz=Eospondylus primigenius 263 Ophiurina Stiirtz 246 0. lymani Stiirtz 247 Palseaster Hall 67 Paleeaster of authors =Hudsonaster } Mesopalseaster, Devonaster, Promopal- seaster, Anorihaster, Neopalseaster, Petraster, Palasterina, Monaster, Uras- terella. Palseaster (Monaster) Etheridge= Monaster 170 Paleeaster (Monaster) Etheridge (part) =Australaster 72 P. (Argaster) antiqua Hall= Mesopalseaster ( ?) antiquus 86 P. antiquatus H&ll=Promopaldeaster speciosus 109 P. antiquus Miller= Mesopalseaster ( f) antiquus 86 P. asperrima Salter= Urasterella (?) asperrima 187 P. caractaci Salter= Mesopalseaster caractaci 92 P. clarkana Miller =Hudsonaster incomptus 61 P. clarkei De Koninck= Monaster clarkei 172 P. clarkei Mi\ler=Hudsonaster i/icomptus 61 P. clarki Clarke and Swartz= Mesopalseaster ( f) clarki 94 P. coronella Salter=Stenaster (?) coronella 167 P. crawfordsvillensis Miller = Neopalseaster crawfordsmllensis 136 P. dubius Miller and Dyer= Mesopalseaster (?) dubius 85 P. dyeri 1&eek=Promopal3easter dyeri 120 P. eucharis Il f d,ll=Devonaster eucharis 98 P. exculptus M.i\\eT=Promopalseaster exculptus 117 P. find Ulrich= Mesopalseaster finei 81 P. ( Monaster) giganteus Etheridge =^Lttsratoer giganteus 73 P. granti Spencer= Mesopalseaster granti 89 P. granulosus ILall=Promopalseaster granulosus 112 P. granulosus MeekPromopalseaster speciosus 109 P. harrisi Miller= Urasterella grandis 180 P. hirudo Salter= Urasterella hirudo 188 P. imbricatus Salter= Tetraster (?) imbricatus 169 P. incomptus Meek.=Hudsonaster incomptus 61 P. jamesi Hall Petraster ( ?) americanus 146 REVISION OF PALEOZOIC STELLEROIDEA. 23 Page. Paldeaster longibrachiatus Miller =Promopalxaster spinulosus 115 P. magnificus Miller =Promopal9easter magnificus 122 P. matutinus Hall=Rudsonaster matutinus 57 P. miamiensis Miller =Anorthaster miamiensis 127 P. montanus Trautschold= Urasterella montana 189 P. niagarensis Hall 69 P. obtusus Salter=Stenaster (?) obtusus 167 P. parviusculus Billings= Mesopalseaster ( f ) parviusculus 87 P. pulchellus Billings= Urasterella pulchella 178 P. pygmsea Eichwald. Not known to writer. Middle Ordovicic, Pulkowa, Russia. P. ruthveni Salter= Urasterella ruthveni 187 P. shajferi Hall= Mesopalseaster shajferi 77 P. simplex Miller and Dyer=Hudsonaster incomptus 61 P. spedosus Meek=Promopalseaster spedosus 109 P. spinulosus Miller and ~Dyer=Promopalxaster spinulosus 115 P. squamatus Salter. Not known to writer. See Cat. Woodwardian Mus., 1873, p. 47; and Woodward, Geol. Mag., dec. 2, vol. 1, 1874, p. 7 (listed only). Bala of England. P. ( Monaster] stutchburii Etheridge=Australaster ( ?) stutchburii 73 P. wilberanus Hall= Mesopalseaster ( ?) wilberanus 84 P. ? wilsoni Raymond Promopalseaster wilsoni 106 P. wyJcoffi Miller and Gurley=Promopalseaster wykoffi 119 Palaeasterina Gregory =Palasterina 150 Paleeasterina of authors =Petraster 138 P. antiqua S<er=Lindstr6master antiquus 149 P. approximata Miller and Dyer=Petr aster spedosus 142 P. fimbriata Meek and WorthQn=Schcenaster fimbriatus 204 P. rugosa Billings Hudsonaster rugosus 64 P. spedosa Miller and Dyer=Petr aster spedosus 142 Palaenectria Stmtz=Palseostella 155 P. devonica StiiTtz=Palssostella solida 156 Palaeocoma Miller (p^Tf)=Ptilonaster 239 Palseocoma S<er=Sturtzaster 252 Palaeocoma (Bdellacoma) Salter =Bdellacoma 254 Palaeocoma (Bhopalocoma) SsilteT=Rhopalocoma 254 P. colvini SalteTStiirtzaster colvini 253 P. cygnipes Salter Sturtzaster cygnipes 253 P. cylindrica Billings= Txniaster cylindricus 220 P. marstoni Salter= Sturtzaster marstoni 253 P. princeps MilleT=Ptilonaster princeps 240 P. pyrotechnica S<er=Rhopalocoma pyrote'chnica 254 P. spinosa Billings= Tssniaster spinosus 219 P. vermiformis Salter = Bdellacoma vermi/ormis 254 Palseodiscus ferox Salter=an echinid. See Jackson, Mem. Boston Soc. Nat. Hist., vol. 7, 1912, p. 250. Palseophioinyxa Sturtz 234 P.grandis (Sturtz) 235 Palseophiura Sturtz 235 P. simplex Sturtz 235 Palaeosolaster Sturtz 209 P. gregoryi Sturtz 209 P. (?) gyalum (Clarke) 210 24 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Palxosolaster roemeri (Clarke) 211 Paleeospondylus Stuitz=Palastropecten 261 P. zitteli Sturtz=Palastropecten zitteli ". 262 Paleeostella Stiirtz 155 P. solida Stiirtz 156 Paleespondylus Stiirtz =Palastropecten 261 Palaeura Jaekel 223 P. neglecta, new species 223 Palasteracanthion Stiirtz= Urasterella 173 Palasterina Billings (p3tft)=Schuchertia 195 Palasterina McCoy 150 P. antiqua SsdteT=Lindstr omaster antiquus 149 P. bonneyi Gregory 153 P. follmanni Sturtz=Pseudopalasterina follmanni 157 P. (?) jamesi D&na l =Petraster (?) americanus 146 P. kinahani Baily= Uranaster kindhani 155 P. primseva (Forbes) 153 P. (?) ramseyensis Hicks 154 P. rigidus Billings =Petraster rigidus 141 P. rugosa Billings =Hudsonaster rugosus 64 P. stellata BillingB=Schuchertia stellata 196 Palasteriscus Stiirtz 200 P. devonicus Stiirtz 200 Palastropecten Stiirtz 261 P. zitteli Stiirtz 262 Palmipes antiqua Forbes= Lindstromaster antiquus 149 Petraster Billings 138 P. (?) americanus (D'Orbigny) 146 P. (?) antiqua Shumard= Mesopalxaster (?) antiquus 86 P. bellulus Billings = Mesopalseaster bellulus 91 P. rigidus (Billings) . % 141 P. rigidus (Billings) (p&Tt)=Hudsonaster matutinus 57 P. smyihi McCoy 147 P. spedosus (Miller and Dyer) 142 P. wilberianus Meek and Worthen= Mesopalseaster ( ?) wilberanus 84 Promopalaeaster, new genus 102 P. bellulus, new species 113 P. ^m'(Meek) 120 P. exculptus (Miller) 117 P. granulosus (Hall) 112 P. magnificus (Miller) 122 P. prenuntius, new species 107 P. spedosus (Meek) 109 P., sp. undet 108 P. spinulosus (Miller and Dyer) 115 P. wilsoni (Raymond) [ 106 P. wykoffi (Miller and Gurley) 119 Protaster Forbes 224 Protaster Kall=Hallaster 254 P. f barrisi TL2il\=Onychaster barrisi 272 P. biforis Gregory 226 P. brisingoides Gregory=Sturtzura brisingoides 236 P. daoulasensis Davy = Fur caster (?) daoulasensis 261 REVISION OF PALEOZOIC STELLEROIDEA. 25 Protaster decheni Dewalque= Tremataster (?) decheni 248 P. elegans Parks= Txniaster elegans 221 P. flexuosus Miller and DyeT=Alepidaster flexuosus 231 P. forbesi Hall= Hallaster forbesi 255 P. (?) granuliferus lAeek=Alepidaster granuliferus 230 P. (?) gregarius Meek and Worth en =Aganaster gregarius 264 P. groomi Sollas and Sollas 227 P. leptosoma Salter =Sturtzur a leptosoma 237 P. miamiensis Miller =Alepidaster miamiensis 233 P. miltoni SalterLapworthura miltoni 251 P. (Ophiura)obtususEichwald. Unknown to the writer. See Schichtensystem Esthland, p. 193; Lethsea Rossica, p. 661; and Woodward, Geol. Mag., dec. 2, vol. 1, 1874, p. 7 (listed only). Siluric of Russia. P. petri Salter. Unknown to the writer. See Salter, Cat. Mus. Pract. Geol., p. 30; and Woodward, Geol. Mag., dec. 2, vol. 1, 1874, p. 7 (listed only). Probably Bala, Church Stretton, England. P. ( f ) salteri (Salter) 227 P. sedgwickii Forbes 225 P. (?) stellifer Ringueberg 228 P. (?) whiteavesianus Parks 227 Protasteracanthion Stiirtz= Urasterella 173 P. primus Stiirtz= Urasterella asperula 188 ProtasterinaUlrich==.47epw/aser 228 P . fimbriata Ulrich=Alepidaster fiexuosus 231 P. flexuosa 3aio.eB=Alepidaster fiexuosus 231 Protoeuryale C. F. Roemer=a crinid. Protopalseaster Hudson= Hudsonaster 53 P. caractaci Spencer= Mesopalseaster caractaci 92 P. nanawayi Hudson = Hudsonaster narrawayi 59 Pseudopalasterina Sturtz 156 P.follmanni (Sturtz) 157 Ptilonaster Hall 239 P. princeps Hall 240 Bhenaster Schondorf 133 R. schwerdi Schondorf .,*. 133 Rhodostoma Sollas and Sollas =Sturtzura .7. 236 R. leptosoma Sollas and Sol\&s=Sturtzura leptosoma 237 Rhopalocoma Salter 254 R. pyrotechnica Salter 254 Roemeraster Stiirtz= Urasterella 173 R. ( ?) acuminatus Sturtz = Mesopalseaster ( ?) acuminatus 93 R. asperula Stiirtz= Urasterella asperula 188 Salteraster Stiirtz= Urasterella * 173 S. asperrimus Stiirtz= Urasterella (?) asperrima 187 Schoenaster Meek and Worthen 202 S. fimbriatus Meek and Worthen 204 S. (?) legrandensis Miller and Gurley 206 S. (?) montanus Raymond 207 S., new species 206 S. (?) wachsmuthi Meek and Worthen 205 Schuchertia Gregory 195 S. laxata, new species 198 S. ordinaria, new species 199 26 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Page. Schuchertia stellata (Billings) 196 Siluraster Jaekel 65 S. perfectus Jaekel 66 Spaniaster Schondorf 95 S. latiscutatus (Sandberger) 96 Squamaster Ringueberg 249 S. echinatus Ringueberg 249 Stenaster Billings 163 Stenaster Billings (part)= Urasterella 173 S. (?) confiuens Trautschold 167 S. (?) coronella (Salter) 167 S. grandis Meek= Urasterella grandis 180 S. huxleyi Billings= Urasterella huxleyi 182 S. (?) obtusus (Forbes) 167 8. pulchellus Billings= Urasterella pulchella 178 S. salteri Billings 165 Stiirtzaster Etheridge 252 S. colvini (Salter) 253 8. cygnipes (Salter) 253 8. marstoni (Salter) 253 S. (?) mitchelli Etheridge 254 Stiirtzura Gregory 236 S. brisingoides (Gregory) 236 S. leptosoma (Salter) 237 S. leptosomoides Chapman 237 Sympterura Bather 256 S. minveri Bather 256 Tseniaster Billings 216 T. australis McCoy =Sturtzura brisingoides 236 T. cylindricus (Billings) 220 T. elegans Miller 221 T. meafordensis Foerste 221 T. schoharise. Ruedemann 220 T. spinosus (Billings). 219 Taeniura Gregory= T&niaster 216 T. cylindrica Gregory = Tseniaster cylindricus 220 Tetraster Nicholson and Etheridge 167 T. asperrimus Nicholson and Etheridge = Urasterella (?) asperrima 187 T. (?) imbricatus (Salter) 169 T. sp. indet. Nicholson and Etheridge= Urasterella girvanensis 186 T. wyville-thomsoni Nicholson and Etheridge 168 T. wymlle-thomsoni Nicholson and Etheridge (part) = Hudsonaster batheri 65 Tremataster Worthen and Miller 247 T. (?) decheni (Dewalque) 248 T. diffidlis Worthen and Miller 247 Trentonaster Sturiz=Schuchertia 195 T. stellata Sturtz== Schuchertia stellata 196 Trichotaster plumiformis Wright. A 10-rayed asterid from the Wenlock lime- stone at Dudley, England, in the collection of Dr. Grindrod. Insufficiently described to be regarded as defined. See Quart. Journ. Geol. Soc., vol. 29, 1873, p. 421; also Woodward, Geol. Mag., dec. 2, vol. 1, 1874, p. 7 (listed as Trochitaster plumiformis). Trimerast er Schondorf ... 134 BEVISION OF PALEOZOIC STELLEKOIDEA. 27 Page. Trimeraster parvulus Schondorf 134 Trochitaster plumiformis WTight=Trichotaster plumiformis. Uranaster Gregory 154 U. elizse, Spencer 155 U. kinahani (Baily) 155 Uraster hirudo Forbes= Urasterella hirudo 188 U. obtusus ~FoTbes=Stenaster (?) obtusus 167 U. primsevus Forbes=PaZasmna primseva 153 U. ruthveni Forbes= Urasterella ruthveni 187 Urasterella McCoy 173 Urasterella Stiirtz I8$3=Stenaster 163 U. (?) asperrima (Salter) 187 U. asperula (Roemer) 188 U. (?) constellata (Thorent) 187 U. girvanensis, new species 186 U. grandis (Meek) 180 U. hirudo (Forbes) 188 U. huxleyi (Billings) 182 U. montana (Stschurowsky) 189 U., new species 189 U. pulchella (Billings) 178 U. ruthveni (Forbes) 187 U. selwyni McCoy 188 U. ulrichi, new species 183 Xenaster Simonovitsch 129 X. dispar Schondorf 131 X. elegans Schondorf 131 X. eucharis Schondori=Devonaster eucharis 98 X. margaritatus Follmann=^". elegans 131 X. margaritatus Simonovitsch 131 X. margaritatus Simonovitsch (part)=J5T. dispar and Agalmaster grandis 131, 132 X. (?) rhenanus (Miiller) 131 X. simplex Simonovitsch =Spaniaster latiscutatus 96 REMARKS ON THE DEVELOPMENT OF PALEOZOIC STELLEROIDEA. GEOLOGICAL. Distribution. This memoir treats of the following Paleozoic Stelleroidea : Species. North Ameri- can species. Euro- pean species. Southern Hemis- phere species. Asteroidea... Auluroidea. . Ophiuroidea. Total. . . 110 59 5 174 11 28 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Arranged geologically the distribution is as follows : Ordo- vicic. Siluric. Devonic. Lower Carbon- iferous. Upper Carbon- iferous. Permic. Asteroidea 46 16 36 6 6 Auluroidea 15 17 24 3 OphiuToidea 5 o o Total 61 33 60 14 6 Of all of the Paleozoic Stelleroidea (174 species), not a single one is common to North America and Europe. Of genera having specific representation in more than one continent there are the folio whig: Hudsonaster, 5 in America, 1 in Scotland; Mesopalseaster, 13 (7 doubt- ful) in America, 1 in England, ? 1 in Germany; Petraster, 3 in America (1 doubtful), 1 in Australia; Stenaster, 1 in America, ? 1 in Ireland and Wales, ? 1 in England, ? 1 in Kussia; Urasterella, 5 in America, 1 in Scotland, ? 1 in Wales, ? 1 in France, 2 in England, 1 in Australia, 1 in Germany, 1 in Russia; Paldeosolaster, 1 in Germany, ? 1 in America; Protaster, ? 2 in America, 1 in England, 2 (1 doubtful) in Wales; Tremataster, 1 in America, ? 1 in Belgium. Of Ophiuroidea, none have wide distribution. Nearly all of the species have very short geologic ranges, but very few occurring in more than a single formation. Mesopalxaster intermedius ranges from the ? Utica to the Maysvillian, M. sliafferi from the Maysvillian to the Richmondian, Alepidaster flexuosus from the basal Edenian to the upper Maysvillian, Hudsonaster incomp- tus from the Edenian into the Richmondian, Promopalxaster mag- nificus and Petraster spedosus from the Maysvillian to the Rich- mondian. These tables also show that the Ophiuroidea do not begin earlier than the Lower Carboniferous (Burlington) and that the Auluroidea cease to exist after the same period, for none are as yet known in the Upper Carboniferous or Permic. The latter attained their maximum of development in the Siluric and Devonic, with very rapid decline in the Lower Carboniferous. The great development of Asteroidea in the Ordovicic is due to the marked specific evolution of the genera Mesopalseaster and Promo- palseaster in North America. This knowledge is the result of care- ful work by local collectors in areas abounding in good fossils. On the other hand the marked rise hi specific representation in the De- vonic is due to the protracted collecting in the earlier beds of this period, the roofing slates of Bundenbach, Germany. Preservation and occurrence. Starfishes are most often preserved in sandstones and muds tones and least often in limestones. The REVISION OF PALEOZOIC STELLEROIDEA. 29 reason why they occur less frequently in limestones is partially accounted for by the fact that during the weathering process they are almost always ruined. Those found in such deposits nearly always occur in the thin shale partings between the bedded limestones. In coarse sandstones they are less often seen, probably because these sediments are so much moved by the storm waves action that destroys the skeleton by separating the plates. Unless a starfish is quickly covered by sediment it is sure to be broken up and jumbled into a mass of unrelated ossicles, because the skeletal parts are rarely coossified. Hence the best preserved specimens are found in mud- stones and especially in the fine-grained, somewhat muddy sand- stones. Here they usually occur as fine molds, since all of the calcium carbonate has been dissolved out by the atmospheric waters. Such material is apt to be fairly abundant in individuals, and although a little difficult to study is often well preserved. Its interpretation is dependent on artificial casts or squeezes in wax or gutta-percha. When found in black slates, as at Bundenbach, the skeleton is pre- served in iron pyrite, and even though these sediments have been subjected to mountain making, the specimens can be cleaned me- chanically of the adhering slate. The process is, however, a laborious one and has been successfully used only by Stiirtz and his two prep- arators. In the calcareous shales asterids are often well preserved, with the original skeleton more or less permineralized and the ossicles cemented together so that parts of the individuals weather out as free fossils. This is particularly the case in the Bichmondian deposits of Ohio and Indiana. As a rule, starfishes are obtained in single specimens and as acci- dental finds, and for this reason they are among the rarest of known Paleozoic animals; they are the " fancy fossils" of the local collectors and the " choice specimens" of the museums. All of this is, however, due to the accident of preservation plus their great destruction through weathering. That starfishes and ophiurids were common, though probably not so abundant as aulurids, is proven when they are located in their entombing sediment and then quarried for. This is best seen in the well-known Lower Devonic slates of Bundenbach, where the quarrying for roofing material has produced as many starfishes as all other localities put together. A great variety has also been secured here, so that it is the only locality and time of which we can say that we know the starfish fauna. The most remarkable starfish find, however, is that made near Saugerties, New York, where over 400 examples of the Middle Devonic Devonaster eucharis were found in a fine-grained, somewhat muddy sandstone, extending over an area of 200 square feet. They occur as natural molds, and as the animals are found closely associated with Grammy sia, it is thought that while feeding on these bivalves 30 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. they were suddenly entrapped by the engulfing sand. In the crinid bed of the Keokuk formation at Crawfordsville, Indiana, the ophiurid Ony chaster flexilis is also often met with. Aulurids, although as a rule not well preserved, are probably more often seen than starfishes, and certainly are far more abundant than ophiurids. The latter condition is, however, probably explained by the fact that no ophiurid is known older than the Carboniferous. Of the aulurid Tseniaster elegans 31 individuals occur on one small slab. It is probable that if aulurids and asterids were systematically dug for, a far greater harvest would result than that now at hand. Whenever such specimens are found in place, such prospects should be followed up by digging or by prolonged search on the part of the local workers. It is very seldom that a Paleozoic asterid is so well preserved that all of its parts can be made out, and this is especially true as regards the dorsal side. During the decay of the animals it is probably true that in most cases the individual lies with its ventral side down, and as the soft parts vanish the dorsal skeleton sinks in irregularly over the actinal skeleton. It is the central region of the disk that suffers dismemberment most. Then, as the two sides of the skeleton finally come to be fully compressed, the rays take on an unnaturally great width, so that in many of the fossils preserved in shales the ambulacral furrow is considerably wider than in nature. For these reasons it is often impossible to say whether the ambulacralia are of the opposite or the alternate arrangement, and what was the original structure of the disk and the interambulacral areas. ASTERID EVOLUTION. GENERAL OBSERVATIONS. The majority of the starfishes studied by the writer are from the Ordovicic strata of North America. Collectively these show considerable evolution, for at this early time in the history of the earth Phanerozonia, Cryptozonia, and primitive ophiurids ( = Aulu- roidea) are present. Even near the base of the Middle Ordovicic the two former groups appear to be equally common, while the aulurids are as a rule rare and small, though there are occasional large ones. Here the primitive phanerozonian asterids are also small, hardly ever exceeding 15 mm. in diameter, while the derived and far more complex Cryptozonia are large, some of them having rays exceeding 50 mm. In the Lower Ordovicic of America no starfishes are known, while the few that have been reported from Great Britain appear to the writer to be from the Middle Ordovicic. These facts show that in spite of their absence in Lower Ordovicio rooks tjfcere must have lived at that time various kinds of starfishes^ REVISION OF PALEOZOIC STELLEEOIDEA. 31 Not only this, but we may confidently add that the varied organi- zation of those of the Middle Ordovicic proves that there is a long previous history of which paleontology knows nothing. The origin of the Asteroidea goes back in all probability to the Proterozoic, as may be inferred from the complex metamorphosis of the starfish larva. In all of the Cambric, however, not a single starfish has been found, and this is the more surprising when one considers the abun- dance of life unearthed from these sediments in so many places throughout the world. This may mean that a preservable starfish skeleton was not evolved until Lower Ordovicic time, where they will surely be found. TJie most primitive fossil starfish. Of the Ordovicic starfishes, the genus Hudsonaster is the most primitive, but with the full quota of parts in the primary skeleton that an unmodified or unspecialized starfish must have. Among living starfishes this primitive skeleton is rarely seen, but occurs in Heterasterias volsellata, "and in a few other species, especially when young" (Verrill 1914:27). There are other genera with a simpler skeleton (Stenaster and Tetraster), but these are clearly cryptozonian forms that have originated in a phanerozonian stock not unlike Hudsonaster. The oldest Hudsonasters (see plates 1 to 6) are small animals about 10 mm. in diameter, with thick, highly con- vex, closely adjoining plates, all of which are devoid of spines ex- cepting the adambulacrals, which have simple, minute, articulating spines. On the ventral side there are in each radius double columns of rectangular ambulacralia, with the pieces of each ambulacrum arranged opposite to one another, or nearly so. Outside of these are single columns of adambulacralia of about the same number as the ambulacrals; they are, however, somewhat larger and more trans- verse. These are laterally bounded by much larger, highly convex, granulated, very prominent marginals, which, as they border the animals and are simple ossicles (there are not here two superposed marginalia as is so common in living forms), are the inframarginal columns. In the axils of the rays lies a single large marginal plate, the axillary, and these occupy the entire interbrachial areas. Around the inner sides of these axillaries the adambulacralia continue, the two basal pieces being the largest and the essential elements of the oral armature. All of the ventral ossicles are, therefore, seen to be of the primary skeleton. On the dorsal side of Hudsonaster, the rays have medially promi- nent columns of radials, while on each side of these are other columns of thick ossicles, alternating with the radialia, and these are the supramarginals. The former ossicles continue closely adjoining to near the center of the disk, where lies a single large plate usually called in this memoir the central disk plate, or, more rarely, the centro-dorsal. In the same way, the supramarginals abut in the axils upon a single 50601 Bull. 8815 3 32 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. large ossicle, the basal supramarginal plate. It is thought that the ancestors of Hudsonaster had a dorsal disk skeleton made up of a centro-dorsal, a first ring of 5 primordial radials, and a second ring of 10 plates, 5 of which are" the second radials, and 5 interradial pieces, the primordial supramarginals. In Hudsonaster, however, some progression has taken place in the disk in the way of increase of size. Therefore additional ossicles of the secondary skeleton have developed to take up this space, these accessory disk pieces being inserted between the centro-dorsal and the first ring of primordial radials. Such* a development of a small number of accessory pieces here is a hint of future complexity and the rise of the secondary skel- eton, and we shall see how in later and derived genera their number becomes multitudinous and their places of insertion nearly everywhere in the older portions of the skeleton between the columns of ossicles. The supramarginals do not margin the animals, but lie inside the inframarginals and axillaries which do margin Hudsonaster and most of the Paleozoic asterids. From Hudsonaster to all other progressive asterids of the Phane- rozonia type, the change lies mainly in the increasing number of the ossicles, relative decrease in the size of the plates, introduction of many new series of Accessory pieces, absorption and removal of others, with a marked general tendency to break up the stiff and ponderous inherited skeleton into one of small pieces, thus affording greater flexibility and greater podial strength through the endless duplication of ambulacral parts. These developmental tendencies take place more especially on the dorsal area and are further accent- uated through increase of body cavity, which demands an enlarged skeletal covering. The disk widens, and along with it the proximal parts of the rays, so that interbrachial areas are also affected, until finally in more than one phylum pentagonal asterids result. The interbrachial areas are dorsaUy increased by the insertion of acces- sory pieces between the infra- and supramarginalia, and ventrally by the crowding into these areas of, first, the single axillaries, and then more and more of the oldest inframarginals in pairs, assisted also here by the development of accessory ossicles. Origin of the wriggling type of starfish. From Hudsonaster to the other progressive asterids of the Cryptozonia type the evolutionary tendencies are in the same directions, but here even greater flexibility appears to be the main stimulus. Accordingly, the entire dorsal skel- eton tends to break up into small loosely adjoining pieces and finally even into a spicular spinose mesh. In these forms the ossicles of the primitive columns are no longer discernible as such, and this tendency is very apt to be likewise true of the primordial disk plates. On the ventral side the inframarginals are no longer wholly present as such, but may be here mixed up with the dorsal skeletal complex. How- EEVISION OF PALEOZOIC STELLEKOIDEA. 33 ever, some of these plates are at times retained in the interbrachial areas as large pieces, and in many forms they are still to be made out clearly on the sides of the growing distal ends of the rays. Here the inframarginals occur as columns of tiny closely adjoining ossicles situated directly beside the adambulacrals, and in a few cases the whole five primary dorsal columns (one radial, two supramarginal, and two inframarginal) can be made out without any accessory pieces be- tween them. In other words, a Paleozoic cryptozonian may retain the marginals throughout life, but because of their small size and isolation one from another by accessory pieces or because of intense spiculization, they are no longer recognizable as such. The classifies - tory value of the presence or absence of marginalia is discussed else- where (under Cryptozonia), and as the inframarginals are seemingly or actually lost independently in a number of phyla the term is here used as expressive of this condition, and not necessarily of relationship. Spencer (1914:9) takes up the origin of the wriggling habit from another point of view, that is, from a study of living Stelleroidea as described by MacBride. The former states that the living forms can be grouped into two divisions as follows : The graspers. Asteroid forms in which the tube-feet are used for walking, and for grasping and pulling open the bivalve shells of the mollusks upon which they usually feed. The ambulacralia form an arch to take the pull. The wrigglers. Ophiuroid forms in which the tube-feet have lost locomotory powers and become much reduced. The animals progress by wriggling movements of the arm, and the ossicles of the ambulacral groove are extensively modified for this pur- pose. The food is pushed into the mouth by the first two pairs of tube-feet, which become considerably enlarged and are known as buccal tentacles. If we trace the history of the forms backward we find that the difference between them tends to disappear. Both the ' ' graspers } ' and the ' ' wrigglers ' ' descended from a third group, which I call provisionally "the primitive Asterozoa." The postulated " primitive Asterozoa" above referred to are based upon the earliest stages of growth of starfishes, and had the following characters. The animals were attached to the ground by a well- developed, flexible stalk (seen in Asterina and Asterias); the disk was small compared with the five arms. In connection with the water-vascular system there were two series of ossicles, (1) a double column of flooring pieces forming the sides of the ambulacrum, and (2) a paired covering series as a protection to the soft structures under- neath. The podia emerged between the flooring pieces and these at first "were arranged not exactly opposite to each other, but slightly alternating, and in consequence we find that both the flooring and covering plates, which are in direct association with the podia, are not exactly opposite, but arranged alternately. " 34 BULLETIN 88. UNITED STATES NATIONAL MUSEUM. The rays terminated in a single ossicle with an unpaired tube-foot that later developed at its outer end a pigment spot and finally an "eye." At last the animal lost the stalk and moved about freely as the primitive asterozoon. The flooring plates changed into ambulacralia and the roofing plates into adambulacralia, while the tube-feet were used not for grasping and crawling but for passing small particles of food to the mouth. Inf Ax FIGS. 1 AND 2. VENTRAL AND DOKSAL VIEWS OF THEORETIC PHYLEMBRYO OF STELLEROIDEA, INDICATING HOW THE RAYS ARE INTRODUCED. BASED ON HUDSONASTER. A d, ADAMBULACRALS; Am, AMBULA- CKALS; AX, MARGINAL AXILLARIES; CZ>,CENTRO-DORSAL; /72/,INFRAMARGINALS; Rl , PRIMORDIAL RADIALS: R2 TO Rll, SUBSEQUENT RADIALSJ Sml , DORSAL INTERRADIALS OR PRIMORDIAL SUPRAMARGINALSJ Sm2 TO SmlO, SUBSEQUENT SUPRAMARGINALS. The asterid radicle. Hudsonaster is held to be very near the radicle that gave rise through modification and inheritance to all subsequent Stelleroidea. A study of the various species and specimens of Hud- sonaster, reinforced by the chronogenesis of this genus and the de- velopmental stages in the individuals, which are discernible in the youthful distal ends of the rays when contrasted with the mature proximal region of the same, has led the writer to speculate on the probable skeletal characters of the radicle of the Stelleroidea. An analysis of mature H. matutinus and of half-grown H. incomptus shows that their ancestors must have been devoid of all disk accessory pieces, or for that matter of all secondary ossicles, since it appears that it is on the disk that these plates first arise. Further, the same material indicates that the younger the individual specimen, or the older the species geologically, not only the smaller is the specimen, but the fewer plates has it in any column. As the REVISION OF PALEOZOIC STELLEROIDEA. 35 ossicles, other than those of the disk, the interbrachial areas, and the accessory pieces, are introduced in all the columns at the distal ends of the rays, we are permitted, in theorizing as to the characters of the radicle, and from what has been learned from Hudsonaster, to elimi- nate all of the rays with their ventral and dorsal ossicles down to the disk. If this is done, there finally results a tiny depressed and mammillated spheroid, made up dorsally of 11 primary plates, and 5 marginal axillaries, and ventrally of a limited number of adambu- lacrals and ambulacrals around the mouth. To make this matter clearer, two diagrams are presented of the probable characters of the direct ancestor of Hudsonaster (see figs. 1 and 2) . By taking away the rays down to near their bases, i. e., to R2, and correspondingly decreasing the size of the primordial supramarginals, we get the dorsal structure of a late stage of the probable typembryo of the Stelleroidea, as illustrated in fig. 3. If all of the ray plates are com- pletely removed, we apparently get the typembryo of the class Stelle- roidea, as iUustrated in figs. 4 and 5. It Should be Stated here, how- FlG - S.-DORSAL VIEW OF THEORETIC TYPEM- ,, t , , . ,, BRYO OF STELLEROIDEA AT THE BEGINNING ever, that the number of the adam- bulacrals and ambulacrals in figure 5 is entirely conjectural, though these plates and those of the mouth frame are among the first to develop in living forms. If the developmental views just stated are correct, it follows that the five branches of the water- vascular and nerve rings of the typ- embryo of Stelleroidea grow outward along the radii, developing as they grow the rays or arms with their dorsal and ventral skeletons. In other words, the growing tips of the radii spread outward beyond the disk, and each one gives rise to nine columns of ossicles (one radial, two supramarginal, two inframarginal, two adambulacral and two ambulacral) . The radial plates continue as simple columns, while upon the primordial supramarginals and the axillaries rest pairs of columns, the supramarginals and inframarginals, respectively. In Echini, however, the growing tips of the radii do not spread beyond the disk (apical disk of oculars and genitals) but remain internal and give rise to new ossicles, the ambulacrals (same ossicles in Stelleroidea) and interambulacrals ( = adambulacrals of Steller- OF RAY DEVELOPMENT. THE PRIMORDIAL SUP- RAMARGINALS HERE HAVE ON EACH SIDE THE BASAL PIECES OF TEN SUPRAMARGINAL COL- UMNS. CD, CENTRO-DORSAL; Rl, PRIMORDIAL RADIAL; R2, SECOND RADIAL; Sml, DORSAL IN- ^ERRADIAL OR PRIMORDIAL SUPRAMARGINAL; 82, BEGINNING OF SUPRAMARGINAL COLUMNS. 36 BULLETIN UNITED STATES NATIONAL MUSEUM. oidea), all of which appear from underneath the ventral edges of the oculars, each plate in turn being pushed orally 1 . In any event, whatever the pre-Hudsonaster evolution was, we are on safe ground in indicating what the developmental changes were, starting with this genus. We will not take up all of the skele- tal parts and discuss them in detail, but will attempt only the more essential portions. Ontogeny. Ontogenetic stages of development in the ossicles of asterids can readily be made out to some extent in any well pre- served specimen. The variations can best be seen in those forms with the least number of plates and therefore in the more primitive genera. Here all of the primary ossicles are introduced at the tips of the rays in the seven primary columns. The youngest pieces are also the smallest and the simplest. At first they are globular, FIGS. 4 AND 5. DORSAL AND VENTRAL VIEWS OF THEORETIC TYPEMBRYO OF STELLEROIDEA. Ax, VEN- TRAL OR MARGINAL AXILLARIES OR BASAL INFRAMARGINALS; CD, CENTRO-DORSALJ R, PRIMORDIAL RADIALSJ Sml, DORSAL INTERRADIALS OR PRIMORDIAL SUPRAMARGINALS. smooth, and loosely adjoining. Tracing them down the columns, the specific characters are seen to develop and finally the basal pieces are found to be the most modified of all. This is particularly true of the dorsal columns where the most ornate plates are the oldest. In the different species of Hudsonaster one can also trace the changes chronogenetically, as for instance in the basal radialia and supra- marginalia. In the oldest species these plates are not at all, or but very slightly stellate, more so in H. incomptus, and most so in the youngest and largest species, H. rugosus. In the stocks with more complex skeletons, the ontogenetic varia- tions in the plates can be traced in the same way. However, as these forms almost always have more or less of accessory plates that are introduced nearly throughout the entire skeleton during the juvenile growth and even at maturity, the study is complicated by the inter- mixture of ossicles of varying age. i See Jackson, Phytogeny of the Echini, Mem. Boston Soc. Nat. Hist., vol. 7, 1912. REVISION OF PALEOZOIC STELLEROIDEA. 37 Spinosity can also be studied ontogenetically because the youngest ossicles do not have spines, and when these appear they are smaller and less characterized than the older ones situated in the proximal portions of the columns. Many of the ontogenetic changes observed by the writer are re- corded throughout this work, and their significance in related species and genera has been made use of in discerning the phylogenetic lines of descent. Regeneration. Among living starfishes it is common to regrow arms that have been lost through accidental causes. From the base of the severed ray a new growing tip is established, forming a juvenile arm that gradually grows to full size and assumes mature characteristics. Schondorf (1909a: 96-97) states that this habit has been pronounced since the Jurassic but that he has failed to find marked regeneration in Paleozoic asterids. He did, how- ever, note partial replacement of minor losses among the Devonic species. Stiirtz, who has handled more Paleozoic asterids than any other paleontologist, also has not noted a single case of marked regeneration. The same is true for the 400 Devonaster eucJiaris found in a limited area of the Middle Devonic of New York. Clarke, in describing this find (1912: 44-45), however, does note a few specimens " which show the existence of only four instead of the normal five arms." These are the only examples of four-rayed Paleozoic starfishes so far recorded. The writer has also been unable to find a single case of regenera- tion, but in the Middle Ordovicic cryptozonian Urasterella ulrichi he describes a specimen with two normally developed rays and three short stumps. All of the arms are normal for the species, except for the length of three rays and their terminations, which are blunt. (See pi. 29, fig. 1.) It seems to him that this occurrence is not due to the accident of fossilization or weathering, but is an actual case of loss in life with subsequent healing of the wounds, but without regeneration of the lost parts. During most of the Paleozoic, the starfishes could have had no carnivorous enemies other than the cephalopods; as for marine fishes, the armored Arthrodires did not appear until the Middle Devonic, while the ancient sharks were not common until Lower Carboniferous (Mississippi) time. It is possible that regeneration among the starfishes is connected with the rise of carnivorous ene- mies, but as the habit is so common among living forms it is more probable that this power has always been inherent in the class. Regeneration among the crinids has been noted in several cases where lost distal ends of arms were being replaced by immature growths. Such have been seen in the Lower Carboniferous (Bur- lington and Keokuk formations) of America. 38 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Number and character of rays. In the geologically oldest and the most primitive asterids as well there are always 5 rays, no multirayed form being known back of the middle of Siluric time. In the most primitive genus of Phanerozonia, Hudsonaster, the rays are short and stout, there is a large disk without interbrachial arcs, and the skele- ton is simple but ponderous. From this type originate those with 5 slender rays and also forms with more or less large interbrachial arcs, the pentagonal starfishes. Long, slender, and flexible-rayed genera without marginalia are present as early as the most primitive Phanerozonia, i.e., as early as the older Middle Ordovicic, indicating that the two main types of asterid development arose still earlier. The same is true for the ophiurid ancestors (Auluroidea). In fact, slender-rayed Crypto- zonia are more commonly seen in the Middle Ordovicic than are the Phanerozonia. Multi-rayed Phanerozonia with primitive arm construction appear in the Middle Siluric (Wenlock) in Lepidaster. This genus has rays which are short and stout, reminding one in this respect strongly of the primitive 5-rayed forms. An insufficiently described 10- rayed form (Trichotaster plumiformis) is recorded from the Middle Siluric (Wenlock) of England. In Helianihaster of the Lower Devonic there are from 14 to 16 rays and here they are long and slender, while in the Upper Devonic there is a form with only 1 1 rays. Lepid- astereUa of the Upper Devonic has 24 arms. All of the multi-rayed Cryptozonia have slender rays and rather large disks. None are older than the Lower Devonic, where Medusaster has from 12 to 15 rays, Echinasterias 25, Palxosolaster 27 to 29 (an Upper Devonic species referred doubtfully to this genus has 24 or 25) , Echinostella 28 or 29, and Echinodiscaster 29. Curiously, no multi-rayed starfish is known hi the Paleozoic later than the Devonic. Further, the arms may be either odd or even in number in the same genus or even species. How the living multi- rayed starfishes introduce their rays is explained under Palseosolas- teridse (p. 207). Verrill (1914: 12-17) holds that the multi-rayed forms arose as " ' sports/ which have persisted by heredity and natural selection because they were advantageous." The advantage lay in "the increased number of ambulacral sucker-feet. " Most of these forms are now living in shallow water among the rocks exposed to the surf. He states: " This is true of the seven species of Heliaster, with very numerous rays; and of Pycnopodia, with 20 to 24 rays; and of the various shallow-water and littoral species of Solaster and Crossaster, which usually have 9 to 15 rays (rarely 8 or less). It is also true of the numerous 6-rayed species of Asterias, Pisaster, and allied genera." REVISION OF PALEOZOIC STELLEKOIDEA. 39 On the other hand, the increase of rays is more probably due " to the advantage gained in holding their food securely, and in opening bivalves, than for holding to the rocks, though both go together." "The ability to cling tenaciously to rocks may be perfected in other ways, involving equally an increased number of sucker-feet. This is often attained by lengthening the rays, * * * by crowding the suckers into more than four rows, * * * and by increasing the size and strength of the suckers." DEVELOPMENT or THE PARTS. Anus. In all living starfishes the anus apparently lies in the next interradius to the right of the madreporite, or in the direction as the clock goes, but in no Paleozoic asterid is there positive evidence of an anal opening. However, it is extremely rare in these forms to find the disk so well preserved as to make it possible to locate so small an orifice. In certain specimens of Hudsonaster, immediately adj acent to the centro-dorsal plate, there is an open space that either is the anal opening or where a plate has dropped out in fossilization. Schondorf has also failed to find this structure in the German Paleo- zoic asterids, but has indicated its probable position in Spaniaster (see fig. 8) and Calliasterella. Madreporite. In more than half of the Paleozoic asterids the madreporic plate has not been seen. In many cases this is due to the infallen condition of the disk skeleton and the jumbled state of the ossicles; in others it is due to the fact that it is so small as not to be readily distinguished among the mass of other small plates. Finally, a great number of asterids are known only from the actinal side. The madreporite in size varies from minute to very large and ponderous. It may be round, oval, or many-sided, concave, and depressed beneath the disk pieces, or protruding cone-shaped (Petras- ter) . As a rule, it is marked by striations that are coarse or very fine, straight or wavy lined, or it may be crenulostriate or with granular surface. In Petraster the surface canals enter into two spirals that carry the water into the stone canal. In living forms with more than six rays there may be two or more madreporites, but in no Paleozoic species has more than one been seen. In all of the Paleozoic phanerozonians the madreporite appears to be always dorsal in position. In the most primitive genus, Hud- sonaster, the plate may be large and conspicuous or at times can not be made out even when the disk is fairly well preserved. In no Mesopalseaster has this sieve plate been seen, although good speci- mens are at hand, a condition probably due to its small size among the small skeletal pieces of this genus. In Spaniaster it is very small. In Promopalseaster the madreporite is usually conspicuous and large, 40 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. and the same is true for Xenaster and Agalmaster. The plate is also known in Palseaster, Devonaster, Neopalseaster, and Palasterina. In multi-rayed phanerozonians the madreporite is unknown in Lepidaster and Lepidasterella. In Helianthaster it is large, conspic- uous, and situated on the edge of the disk, more actinal than abac- tinal. On the other hand, in the cryptozonian multi-rayed genera the madreporite is apparently always actinal and more or less large, as in Palseosolaster and Echinodiscaster. This plate is probably also always actinal on large-disked Cryptozonia such as Palasteriscus (plate enormously large), and OJieiropteraster (here small and situ- ated near the mouth) . Among the five-rayed Paleozoic cryptozonians the madreporite is known only in Monaster (large) , Urasterella (medium) , and Jaekel- aster (small) . In all it is always on the abactinal area. From this evidence we see that in the five-rayed Phanerozonia, forms that are more or less rigid, the madreporite is always on the dorsal or upper side of the animals, this being true for the deeply stellate as well as for the pentagonal species. In the multi-rayed genera, however, there is a marked tendency for this plate to move outward to the edge of the disk. In the five-rayed cryptozonians the madreporite so far as known is always dorsal, but in the large- disked or pentagonal forms and the multi-rayed species this plate is apparently always ventral. Spencer (1914:35-38) says that "many Paleozoic Asterozoa do not appear to have a madreporite" and he is "inclined to regard this, in these old forms, as usually a primitive feature." The present writer, it is true, states above that in more than half of the Paleo- zoic asterids the madreporic plate has not been seen. This is due in most cases, however, to poor preservation, and in others to the fact that the plate is so much like the other disk plates as not to be dis- tinguishable from chem. Spencer concludes further that it appears that the madreporite originated on the ventral side, or at least was marginal, in the primitive forms, and that it passed over in later Asteroidea to the dorsal side. In Eudsonaster we have the most primitive known starfish, and here in H. incomptus it is a large and conspicuous plate on the abactinal side. Further, the evidence of the Paleozoic starfishes (Phanerozonia), so far as the present writer knows them, is that this plate is always dorsal in position, and he is inclined to the view that all of them had the madreporite, that it originated on this side, and that in later cryptozonian multiradiate forms it moved outward to the margin of the disk or to the actinal surface. In regard to the Auluroidea, the writer has seen the madreporite so rarely that no deductions as to its original position in these forms can be made. REVISION OF PALEOZOIC STELLEROIDEA. 41 Spinosity. In general, it can be said that spines are least developed in the heavily plated Phanerozonia, more so in those forms where the dorsal skeleton is made up of small ossicles in abundance, and fully in most of the Cryptozonia stocks. In primitive Hudsonaster, all of the plates other than the am- bulacrals are more or less well tuberculated, and these granules do not now carry spines, if they ever did. In later forms, however, as Promopalseaster, they may all have borne spines, some of them of large size. Small club-shaped spines occur here in considerable abundance along the sides of the inframarginals and the adambula- crals. In Palseaster there are brushes of delicate spines along the inner edge of the adambulacrals. In Mesopalseaster the ventral spines are better developed, fully so in Promopalseaster, and probably equally so in most post-Ordovicic forms. Among the primitive cryptozonian genera Stenaster and Tetraster no spines are now seen, but undoubtedly such were present on the actinal side. In Ur aster ella the entire dorsal skeleton is more or less spinose, consisting of long, nonarticulating, slender rods, the dorsal extensions of the ossicles. In the multi-rayed Cryptozonia of the Devonic, spinosity is at its greatest development in the Paleozoic. Ambulacralia. Undoubtedly the most important skeletal parts of the Stelleroidea are the ambulacrals. In general they undergo the least alteration during geologic time of the entire asterid skeleton, and therefore any marked variation must be of broad classificatory value. Schondorf * is well aware of this fundamental value and has made full use of it in denning his three " classes, 57 Asteroidea, Aulu- roidea, and Ophiuroidea. In the Asteroidea the ambulacralia of ad- joining columns are nearly always placed directly opposite one another, with the podia issuing through openings that are not in the plates themselves, but laterally between two adjoining ambulacrals and the corresponding adambulacrals. Among the Paleozoic asterids one is not always certain whether the ambulacralia are arranged lt opposite" or "alternate," because the specimens in nearly all cases have suffered more or less from distortion. In many good specimens they are very slightly alternate, but in all such cases the arrange- ment is said to be alternate. In other fine fossils they are now alternate, but a close study seems to show that originally they were practically opposite in arrangement. On the other hand, certain of the species with wide ambulacral furrows and large ambulacralia, as in Promopalseaster, Anorthaster, and Urasterella, have an alternate arrangement, while in other species of the first and last named genera they are just as certainly opposite. The writer therefore does not i Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, pp. 206-256; Palaeontographica, vol. 57, 1910, pp. 1-56. 42 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. lay much stress upon this feature among the Ordovicic asterids, nor does he regard it even necessarily as of specific value, but in the later Asteroidea the arrangement is always opposite or nearly so. Ver- rill (1914:20) states that he has noticed the alternate arrangement in living Pycnopodia "as an abnormal variation in some of the rays, and also that it may occur from lateral bending." Evi- dently the alternate disposition is a fixed character among some Ordovicic forms, as in AnortJiaster, but more will certainly be dis- covered in these old strata and probably rather among the Crypto- zonia than the Phanerozonia. The writer is led to this view because of the alternate arrangement in many Auluroidea, a subclass of Stelleroidea that seems to have had its origin in some cryptozoniaii asterid. Out of the Auluroidea with opposite ambulacralia came the Ophiuroidea with their highly modified ambulacrals which are coossified into " vertebral ossicles." In the Auluroidea the arrange- ment of the ambulacrals is either alternate or opposite, but they are never coossified but are so modified as to suggest the Ophiuroidea rather than the Asteroidea. The structure of these plates is de- scribed elsewhere and need not be repeated here. The ambulacrals are very variable in number throughout the genera. In Hudsonaster, Stenaster, and Tetr aster there are as few as 10 in a column, while in Promopals&aster there are certainly 50 and possibly as many as 60. The cryptozonian genus Urasterella has certainly over 100 in a column. In general, these ossicles abut against each other, but in Urasterella and other genera they overlap proximally to a certain extent. As yet the writer has seen no Paleozoic asterid with more than 2 columns of ambulacrals, though in Promopalseaster magnificus the proximal areas of the rays for a short distance have 4 columns of podial openings. Here, however, the ossicles have not yet wedged themselves into 4 columns of ambulacralia. In shape and surface contour there is also considerable variation among the ambulacralia. In Hudsonaster, Tetraster, and several Devonic genera they are more or less rectangular, either squarish or drawn out laterally. In other forms of Hudsonaster, along with Stenaster and Urasterella, the ossicles are h- -shaped, with the lateral shaft thin, thus allowing for large, elongate podial openings. The more primitive genera have the outer surfaces plain or slightly ridged, while in derived forms like Promopalseaster the ridges are high and straight or have their medial terminations orally directed. In P. magnificus, the ridges in the proximal region are arranged in pairs having the shape of a tuning fork; in P. wyleqffi one sees that these ridges approach one another in pairs transitional to their transforma- tion into the tuning-fork form. REVISION OF PALEOZOIC STELLEROIDEA. 43 Adambulacralia. Adambulacralia are present in all Stelleroidea. They are best and typically developed in the Asteroidea, more or less modified in the Auluroidea, and completely altered into the thin scalelike side plates of the Ophiuroidea. In number and relative size they vary as do the ambulacralia, being least numerous and largest in the most primitive genera. Originally they probably bore no spines at all, the plates being practically smooth. In Hud- sonaster and most Ordovicic and Siluric, and some Devonic genera, their surfaces are granular to tubercular. These protuberances in the younger genera may have borne small spines, but long before these appeared each adambulacral bore two or more stout or slender, com- paratively long spines along its ambulacral edge. In most of the Paleozoic genera there are in each column as many adambulacrals as there are ambulacrals, and both series are as a rule arranged opposite one another. In Anorfhaster, however, there are fewer and therefore larger adambulacrals than ambulacrals, and these latter ossicles clearly alternate with one another. In the Phanerozonia the adambulacrals never margin the rays but always lie inside of the bordering inframarginals. In the Crypto- zonia, however, the adambulacrals margin the animals and here the ossicles are usually small, though at times they are relatively large and make a stout outer skeletal frame as in Stenaster, Tetraster, and Schcenaster (?} montanus. The same is true in Encrinaster of the Auluroidea. The adambulacrals in probably all the Paleozoic Phanerozonia continue into the oral region and each two adjoining columns meet here in a pair of modified, elongate, pointed pieces, the most promi- nent ossicles of the oral armature. No other skeletal parts lie in front of these oral ossicles except in the phanerozonian Hudsonaster narrawayi (pi. 1, fig. 1) and in the young of the cryptozonian Urasterella ulrichi (pi. 30, fig. 7). Both are primitive forms of their respective phyla and whether these five pieces or tori are to be interpreted as five spines or whether they represent five primi- tive orals is not yet determinable. Inframarginalia. In the great majority of Paleozoic Phanero- zonia the inframarginals alone margin the animals, and only in a few forms (Spaniaster, Miomaster, RJienaster, and Neopalseaster) have the supramarginals moved outward and completely covered the inframarginals, so that the two columns together equally bound the rays and disk. That condition is a peculiarity common to most Paleozoic phanerozonians, while the wholly superposed arrange- ment distinguishes nearly all the Mesozoic, Cenozoic, and Recent genera of the same kind of Asteroidea. The inframarginals are usually the most conspicuous ossicles of the Paleozoic Phanerozonia, and this is especially true where the 44 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. skeleton is strong and devoid, or nearly so, of accessory ossicles. They are ponderous in Hudsonaster, Palseaster, Australaster, Devon- aster, Xenaster, Trimeraster, Petraster, etc. When the smaller accessory plates are introduced, the inframarginals also reduce in size; this development is best seen when passing from the heavily plated Hudsonaster to the more abundantly plated Mesopalseaster and finally Promopalseaster. The next step in this phyletic line would be pronounced a cryptozonian because the infra- and supra- marginals could not be distinguished from the abundance of acces- sory ossicles. In fact, it is not always easy to make out the various categories of plates even in Promopalseaster, and here one must resort at times to a study of the tips of the rays where the seven pri- mordial columns are not yet disturbed by the introduction of accessory pieces. In primitive Hudsonaster, where the inframarginals are compara- tively very large, there are only half as many of these ossicles as there are of adambulacrals, but in the great majority of Paleozoic genera of Phanerozonia this marked difference tends to be less decided and toward equalization. In certain phyla, as, for instance, from Hudsonaster to Mesopalseaster and Promopalseaster, the equal- ization has gone on rapidly, hi fact, in but a short part of Ordovicic time. In more conservative stocks, as from Hudsonaster to Devonic Xenaster and Trimeraster, the equalization is far slower, as is seen in the following figures: Hudsonaster, 12 inframarginals and 24 adambulacrals; Trimeraster, 14 and 23; and Xenaster, 20 and 32, respectively. In nearly all the genera these columns of plates alter- nate with one another.' This of course must be so where the number of ossicles is different in the two columns, and complete alternation is retained even in certain species where the numbers of ossicles are alike, as for instance in Promopalseaster spinulosus. On the other hand, in Mesopalseaster sJiafferi, where there are 16 inframarginals and 18 adambulacrals, the plates alternate in the proximal half of the ambulacra and are opposite one another distally. A study of the growing ray tips shows that the various ossicles all appear practically of one size and where there are more adambulacrals than inframarginals, the former are crowded orally. The latter develop- ment appears to be the primitive type of growth, but this mode is rapidly changed to one in which the seven columns of primordial ossicles appear in cycles or rings. This type of growth is easily made out in primitive forms such as Hudsonaster and Mesopalseaster, and in some species of Promopalseaster. However, when accessory plates are developed in abundance, and especially in forms where the seven primordial columns of plates remain of one size, their introduction in cycles can only be made out in very young individuals or at the KE VISION OF PALEOZOIC STELLEKOIDEA. 45 extreme tips of the rays, conditions not often attained among these fossils. In the great majority of Paleozoic Phanerozonia the inframar- ginals and adambulacrals lie closely adjacent to one another. In typical asterids they are not separated by the introduction of acces- sory interbrachial pieces until in the Devonic. Here in Xenaster and Trimeraster isolated pieces, and finally a column of them, are inserted in the proximal half of the rays. In the large-disked asterids, this separation occurs much earlier, in the Middle Ordovicic in Petraster. In Paleozoic Phanerozonia where the skeleton is made up of an abundance of more or less equal-sized ossicles, one notices also on the distal portions of the rays a tendency for the inframarginals to appear on the dorsal side and not to remain on the ventral, as in primitive stocks. This alteration brings the adambulacrals to the margin of the rays, and is a hint as to how Phanerozonia gradually change into Cryptozonia. It is best seen in Promopalseaster, Petras- ter, and Mesopalxaster. In Australaster it is a marked feature, in fact, one can say that the distal parts of the rays in that genus are cryptozonian while the bulk of the arms are still phanerozonian. This ontogenetic appearance is in harmony with phylogenetic devel- opment and chronogenesis. The living Echinaster sepositus is in youth also a phanerozonian, but at maturity is a typical cryptozo- nian. In so typical a cryptozonian as Ur aster ella, one still finds in half -grown U. ulrichi a number of inframarginals in the interradial or axillary areas. The same retention of the first formed inframargin- als is also met with among the multi-rayed Cryptozonia in Helianih- aster, and even among the Auluroidea in Encrinaster. It has just been pointed out how the inframarginal columns are eliminated as marginal plates. They are not, however, removed from the skeleton in these early forms by resorption or by failure of development, but for want of special growth. In other words, the inframarginals are probably present in all Paleozoic Asteroidea, but because of lack of specialization through the developmental tendency to greater skeletal flexibility, remain small and are lost as such in the mass of the dorsal plates. This apparent eh'mination of the inframarginals has gone on independently in various stocks as pointed out elsewhere, and therefore the absence of large marginals, either infra- or supramarginals, or both series, is not of ordinal value. Supramarginalia. In Hudsonaster the prominent supramarginal plates of the dorsal side are placed decidedly inside of the inframar- ginals, though the former clearly overlap the latter. This primitive position is retained in many Paleozoic genera, and apparently not before the Devonic do these two columns of ossicles come to lie wholly upon one another, and then they together margin the animals. 46 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. It is apparently always a rare condition in the Paleozoic, but as the post-Devonic asterids are as yet little known, we can more accurately say that the modern type of Phanerozonia is unknown before Devonic tune. It should be added here that complete superposition of the supra- upon the inframarginalia undoubtedly took place in more than one stock and at different times. Accordingly, to group all asterids with large marginalia into the " order" Phanerozonia is further proof that these terms cannot be used in the phyletic sense. Phanerozonia and Cryptozonia are, however, of use in a descriptive way. The supramarginals, and the inframarginals as well, in primitive forms gradually attain larger size down the columns from the tips of the rays toward the disk and mouth. In other words, the older the ossicle in primitive genera, the larger is its size, and in all such forms the pairs of supra- and inframarginal columns terminate in single plates, the primary or basal supramarginals and the primary or basal inframarginals. The latter ossicles in the most primitive genera are situated in the axils of the rays and are therefore also known as axillary marginals, and in the more specialized genera the axillaries and even pairs of inframarginals are crowded orally into the interbrachial areas. The basal supramarginals in primitive forms always occur prominently on the disk, in derived genera they are, however, less conspicuous, and where the disk skeleton is made up of small ossicles, they are generally indistinguishable in the mass of plates. We have seen that accessory interbrachial plates in some genera are developed between the inframarginals and adambulacrals, and to a greater extent the same is true of the development of ambital accessory pieces between the former columns and the supramarginals. In probably all cases where the rays and disk grow wider the addi- tional space is largely covered in by the development of accessory plates. As it is in the disk area, and therefore necessarily in the basal portion of the rays as well, that the main increase of body extent takes place, here are developed the greatest number of these pieces. Rarely are the ambital accessory pieces crowded in between these columns beyond the proximal third of the rays. This is best seen in Devonaster and Xenaster, but in the large-disked forms, as Petraster, they extend farther out and are present to within the outer third of the rays. Ambital accessory plates may also be present in Mesopal&aster, but are more definitely so in Promopalse- aster (especially in P. magnificus) , and in both genera they occur throughout the greater part of the rays. Radialia. In many Paleozoic genera of Phanerozonia and even among some of the Cryptozonia, a more or less prominent column of ossicles is seen down the center of the dorsal side of the arms. These REVISION OF PALEOZOIC STELLEEOIDEA. 47 are the radial columns of ossicles, which are especially well seen in Hudsonaster, Spaniaster, Devonaster, Xenaster, CaUiasterella, and less prominently in Mesopalseaster, Promopalseaster, and Petraster. The prominence of the radials is lost in the same way as has been described for similar reduction in the supramarginals and inframar- ginals. In primitive genera the supramarginals and radials lie closely adjacent as prominent columns, but in the great majority of Paleo- zoic Asteroidea these columns are more or less extensively separated from one another by the introduction of radial accessory plates. The latter are always small pieces and in general one can say that their number increases with geologic time and with the evolutionary complexity of the various stocks. At first the accessory pieces appear singly between the plates in the proximal region of the rays, where, increasing in numbers, they separate the primary columns more and more and finally also crowd in between the individual ossicles, es- pecially the proximal radials, pushing these apart more or less com- pletely. This is true not only in chronogenesis but also in onto- genesis. It is best seen in a study of the dorsal areas beginning with Hudsonaster and Spaniaster, and proceeding to Mesopalseaster, Devonaster, Promopalseaster, and Urasterella. In the forms with great disks the development of accessory plates is extreme, as may be seen in Petraster speciosus. The columns of radial plates are less persistent in development than any other of the primary columns. Like all of the primary plates, the radials are apt to lose their ponderosity and in the course of geologic time become smaller. In this way they may lose their individuality in the mass of dorsal plaies, but in some cases radialia appear to fail to develop throughout the greater part of the rays, although one or more primordial ossicles of each radius are present on the disk. In certain species of Mesopalseaster, and more especially in Promopalseaster, the radialia are either not present or have been squeezed out of the mid line of each radius so as to be unrecognizable. In other forms of the latter genus they are well developed and arranged in linear columns. Promopalseaster as a rule has no radialia on the rays and the space between the supramarginals is occupied by many small accessory radial pieces. Among the Cryptozonia, radialia are usually not discernible as such on the rays, though present on the disk. Accessory plates. These small pieces of irregular form are space fillers between all of the primary columns of plates, with the exception of the ambulacrals and adambulacrals. As the body cavity grows larger, skeletal increase takes place either through enlargement of the seven primordial columns of ossicles or by the intercalation and formation of new pieces, the accessory plates. The former method 50601 Bull. 8815 4 48 BULLETIN is the one least often resorted to. The accessory pieces appear first chronogenetically on the disk, separating the primordial ossicles one from another in various ways, and as the body cavity grows larger in the course of geologic time, they wedge all of the primary columns apart proximally except the ambulacrals, which compensate for the widening of the rays by growing in size transversely. Some accessory plates also develop in the interbrachial areas separating the interbrachial marginals, as in Trimeraster and Xenaster, while in the large-disked forms, as Petraster, their number becomes excessive. Ocular plates. The large ocular plates seen in so many living Phanerozonia are very rarely developed in Paleozoic forms. In fact, the writer knows of but a single occurrence, in the Lower Car- boniferous genus Neopalseaster. It seems to the author that these ossicles should have considerable significance in phylogeny because their large growth indicates that they are much older plates than any of the small adjacent pieces. In other words, in all genera where the ocular or sensory tentacle does not lie in an enlarged and grooved plate, the newly developed ossicles of either the ventral or dorsal side appear distally to the pre- viously borne pieces, i. e., at the tips of the rays, whereas in the stocks having ocular plates the new skeletal arrivals appear not at the tips of the rays but on the proximal side or inside of the eye ossicles Therefore, where ocular plates are present, these pieces are not only enlarged in size with age, but progressively pushed farther and farther outward as well by the growing ray tips that lie on their inner sides. This mode of growth may have originated once or several times, there being in the latter case parallel developments of ocular plates. This is a study that can not be pursued far among Paleozoic genera, but in the later forms may be productive in the discerning of phylogenetic lines. After the above was written the writer read in Verrill (1914: 20-22) that Fewkes has shown the ocular plates to be the first to appear in the young, and that with growth they are pushed distaUy because of "the budding in of new plates between the apical plate and the one next to it. 7 ' These observations on living starfishes are therefore in harmony with the study of Neopalseaster, and go to show that the bulk of Paleozoic starfishes insert the plates of the primary columns in a different way, i. e., always distally. On this ground alone most of the Paleozoic starfishes should be grouped by themselves and apart from most of the subsequent Phanerozonia. Centro-dorsal. In the center of the disk of many Paleozoic genera of Asteroidea and of some Auluroidea is seen a more or less prominent plate, which is thought to be the equivalent of the centro-dorsal of the embryo of the living crinid Antedon (Comatula). It is a promi- nent though small plate in most of the forms of Hudsonaster and in REVISION OF PALEOZOIC STELLEEOIDEA. 49 Palxaster; it is also well developed though less conspicuous in Meso- palseaster, Spaniaster, Xenaster, Agalmaster, Devonaster, Neopalseaster, Palasterina, Monaster, CalliastereTla, and Urasterella. As the disk plates are rarely well preserved, and as the centro- dorsal has been seen in many Paleozoic species, the writer thinks- it will be found in most of the forms of this era. Primary disk plates. In the most primitive and oldest Paleozoic asterids the centro-dorsal is surrounded by a variable but small number of diminutive accessory disk pieces. In mature Hudsonaster they form a single ring, but in the young of the cryptozonian Uras- terella ( U. ulrichi) and in mature CalliastereTla there are none of these accessory pieces present. For these reasons it is thought that in the Ordovicic there will be found a small asterid, even more primitive than Hudsonaster, that will be devoid of accessory disk pieces. Around the centro-dorsal of young Urasterella and in mature Calliasterella, there follows directly a ring of five larger plates. The same five plates are also seen outside of the single ring of accessory pieces in Hudsonaster and Spaniaster, and directly beyond the two rings of accessories in Mesopalseaster. These more or less large plates are radial in position and are the basal radialia above which continue the columns of successively formed radials. Keturning to Hudsonaster, we see that the first ring of prominent plates has 10 pieces, 5 of which are radial in position, the remainder interradial. The same is true for Spaniaster and Xenaster, while Calliasterella has an equal number of plates in the second ring. An analysis of these disks shows that 5 of the plates are either primary or secondary radialia, while the remainder are situated interradially, and upon them rest the 10 columns of supramarginals. The first prominent ring in Mesopalseaster sJi'Offeri and Devonaster and the second ring in Urasterella ulriclii each have 15 plates, 5 of which are radialia, while the other 10 are supramarginals. Simplifying these statements and cutting out the accessory pieces, we learn that the disks of these genera and others could be added are composed of a centro-dorsal, a first ring of 5 radialia, and a second one with 5 radialia and 10 basal supramarginaiia. This, then, is the same struc- ture postulated for the hypothetic phylembryo of Asteroidea and deduced through reduction of ossicles, as gleaned from a partial ontogenesis of Hudsonaster. Axillary ossicles. In the most primitive asterid, Hudsonaster, one sees in each axilla a single large plate, on either side of which adjoin the 10 columns of inframarginals. The same arrangement obtains in Siluraster, Palseaster, and Australaster (here the genus occurs in the Permo-Carboniferous, and the axillaries are very large contrasting with the smaller ones of the Ordovicic Hudsonaster). In no other Paleozoic genera does this same development hold, though the single 50 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. axillaries occur in many other forms where they are, however, pushed orally into the skeleton of the interbrachial areas. Single, not double, axillaries, therefore, make the simplest and smallest type of inter- brachial areas. Further complexity and enlargement of these areas are produced by crowding into them orally, first the axillaries and then one or more pairs of proximal inframarginals. In Hudsonaster narrawayi one sees tiny single plates proximal to each axillary. These are clearly of no marked morphologic value and are to be interpreted as accessory plates. Similar single ossicles are also seen in genera with large interbrachial areas, as in Promopalseaster. They are regarded as space-fillers and without ancestral significance. Interbrachial areas. The simplest type of interbrachial area is that just described. Increase in size and complexity of interbrachial struc- ture is a natural result from an enlarged body cavity. The first change in this direction is seen in Mesopaldeaster, Spaniaster, Miomaster, and Devonaster, where the axillaries are crowded orally and surrounded distally by a pair of adjoining basal inframarginals, and proximally by a variable number of adambulacrals. The same development is seen but hi somewhat modified form in the cryptozonian genera Urasterella (U. ulrichi and U. girvanensis) and CaUiasterella. Here the axillaries are present only in the young, but may be retained to maturity; in either case they may or may not be margined distally by basal infra- marginals. As the inframarginals in these forms are vanishing (pri- mitive Cryptozonia), we should expect to find related genera devoid of even the axillaries, and such is the case in Stenaster and Tetraster. The next step in the enlargement of the interbrachial areas consists of the inward crowding of pairs of basal inframarginals, one or more at a time. In Trimeraster these areas have the single axillaries and a pair of basal inframarginals, with the second pair of proximal infra- marginals making the margin of the axils. In Xenaster the first and second pairs of proximal inframarginals are enclosed, while in Pro- mopalseaster there are from one to three pairs in each interbrachial area. As these areas enlarge, it follows that more and more of the adambulacrals must also take part in the expansion. These ossicles are therefore called interbrachial adambulacrals. Further enlarge- ment, but not to any marked extent, takes place through the insertion of a variable number of accessory disk interbrachials. This is best seen in Trimeraster and Xenaster. Abnormal enlargement, but to a small extent, of the interbrachial areas takes place through oral crowding entirely of adambulacrals in pairs and through the increase in size of some of these ossicles. This type of development is known only in Anorthaster. It has been shown how enlargement of the interbrachial areas comes about through oral crowding of the axillaries, proximal inframargi- nals and adambulacrals pieces of the primary skeleton but the REVISION OP PALEOZOIC STELLEROIDEA. 51 greatest expansion is brought about by the introduction of more and more accessory disk pieces between the adambulacrals and inframar- ginals. The beginning of this development of the secondary skeleton is seen in Petr aster and is variously progressive in the different species. They finally crowd more and more distally into the rays, as may be seen in Uranaster and Palseostella. Among the large-disked Crypto- zonia, as Schuchertia and the Palasteriscidse, it is the ambital and abactinal accessory pieces that go to enlarge the arcs and body cavity. CLASSIFICATION. A good history of the various attempts at classifying the fossil starfishes is given by Spencer (1914: 39-52). How recent starfishes are classified may be learned from Verrill (1914:24), and it will be seen that his principles are not always applicable to fossil forms be- cause of the loss of parts through fossilization. In subsequent pages the author will give his groupings up to superfamilies, but the time is not yet at hand for an ordinal classification applicable alike to fossil and recent starfishes. Catalogue of Paleozoic Stelleroidea. A catalogue of all Paleozoic starfishes was published by the writer in April, 1914, under the title "Fossilium Catalogus, Anirnalia, part 3, Stelleroidea palseozoica," pp. 53 (Junk, Berlin). It gives the entire bibliography of the genera and species up to that date, and an outline of the classification adopted. The actual date of several new genera is of the work above cited, although they are also printed in the present book as new. This Bulletin is there cited as of 1914 and is now to be corrected to 1915. The above-mentioned catalogue will always be useful for ready refer- ence to the literature, although the larger volume repeats all of the essential references, corrected to November, 1914. DESCRIPTION OF GENERA AND SPECIES. Class STELLEROIDEA. Subclass ASTEROIDEA. PHANEROZONIA Sladen. Asteroidea nearly always with large marginal plates. These consist in the Mesozoic and later asterids Qf completely superposed inframar- ginals and supramarginals; in the Paleozoic the rule is that the in- framarginals alone border the animal, while the supramarginals are always dorsal, generally alternate with the inframarginals, and are placed farther inward though resting more or less upon the latter. Ambulacrals always present and opposite in arrangement in post- Paleozoic forms. In the oldest and most primitive Paleozoic forms they are apt to be alternate throughout the rays, or only so distally in the young parts. By far the greater number of Paleozoic forms, however, have the opposite arrangement. 52 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Oral armature always with the adambulacral pieces most con- spicuous. This order is not known to appear before the Middle Ordovicic, and in the Paleozoic contains the following families: Hudsonasteridae. Xenasteridae. Palaeasteridae. Neopalaeasteridae. Promopalaeasteridae. Palasterinidae. Mesopalaeasterinae. Lepidasteridae. Promopalaeasterinae. Anorthasterinae. These seven families of Paleozoic asterids are all related to one another and seemingly had their origin more or less directly in Hudsonaster or a form very similar to it. Therefore the family Hudsonasteridae is the primordial stock out of which all asterids progressively developed. Hudsonaster is massively and simply built, with the spines rudimentary and restricted to the adambula- crals and inf ramarginals ; it is a small rigid asterid with few and comparatively large ossicles, nearly all of which are of the pri- mary skeleton. From it the evolutionary tendency was to get rid of the rigidity and ponderosity through the comparative reduction in size of the plates and the introduction of many smaller sec- ondary pieces, along with an abundance of spines. With this ten- dency once established, there arose several others: (1) elongation of rays, with greater flexibility and greater power of locomotion through the indefinite duplication of pairs of podia; (2) increase of body cavity in three different ways. This took place through the development of interbrachial areas (a) by the inward crowding of the single interbrachial axillaries, followed by more and more pairs of proximal inf ramarginals; (&) through the insertion of an indefinite number of accessory ossicles between the adambulacrals and the inf ramarginals, forcing the latter more and more outward; and (c) through the introduction of supernumerary rays. The latter development is probably abnormal, in that the tendency arose in the later larval life as is explained elsewhere. (See p. 207.) With these changes of the actinal side naturally must arise compensating growths of the abactinal surface and accordingly here is seen the introduction of accessory ossicles in various places. Starting with Hudsonaster, there is one phyletic line developed through the elongation of rays, flexibility, and the increase in size of the oral region by the (a) method from the primitive Hudson- asteridae into the derived Palaeasteridae, Promopalaeasteridae, Xen- asteridae, and Neopalaeasteridae. This is the most satisfactory line of evolution, resulting in the greatest variety of genera. Again with Hudsonaster as the radicle, another phyletic line is established through the increase of the body cavity by the (6) method, giving REVISION OF PALEOZOIC STELLEROIDEA. 53 rise to the Palasterinidse. This line of change is not at all so pro- ductive of genera as the former. Finally, a third phyletic line is established through larval adaptation, causing a sixth ray to develop and thereafter pairs of rays, resulting in the multi-rayed asterids of the family Lepidasteridae. These phyletic lines may be expressed taxonomically as follows: Superfamily Promopalseasteracea, new. Includes families Hudsonasteridae, Palseasteridse, Promopalseasteridae, and Xenasteridae. The Neopalseasteridae developed out of the Promopalaaasteridse. Superfamily Palasterinacea, new. Includes family Palasterinidse. Superfamily Lepidasteracea, new. Includes family Lepidasteridse. HUDSONASTERID.E, new family. Most primitive known Phanerozonia. Small, five-rayed, heavily plated asterids, with narrow ambulacral furrows and slightly alternate ambulacralia. The incipient inter- brachial arcs are occupied by single axillary marginal plates. Abac- tin ally the rays have five columns of ossicles, the radials in the center being bounded on either side by the supramarginals and inframarginals; the latter margin the rays and the ossicles of adjoining columns alter- nate with one another. There are no accessory ray plates of any kind. The disk has a prominent central disk piece separated by a small number of accessory disk plates from a ring of five large basal radials and five large basal interradial ossicles. Spines rudimentary and apparently restricted to the adambulacrals and inframarginals. This family contains the genera Hudsonaster Stiirtz and Siluraster Jaekel. Genus HUDSONASTER Sturtz. Plates 1-6. Palseaster (part) of authors. Hudsonaster STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 56, 1900, pp. 224, 225. Protopalseaster HUDSON, Ottawa Naturalist, vol. 26, 1912, p. 25; vol. 27, 1913, pp. 77-84. RAYMOND, Ottawa Naturalist, vol. 26, 1912, p. 105 (genoholo- type, P. narrawayi). SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palaeontgr. Soc. for 1913), 1914, p. 21. The name is derived from the " Hudson River formation/ 7 an Ordovicic series of strata of long endurance. If the stratigraphic term is used correctly, then these rocks have as yet furnished no starfishes of this genus; for this reason Hudsonaster is a misnomer and has no significance, but under the rules of nomenclature it must be accepted. Generic characters. Disk small and without actual interbrachial arcs. Rays five, short, tapering rapidly. 54 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Abactinal area of rays occupied by three columns of large, thick, more or less tumid or carinate, closely adjoining, tuberculate plates; more or less of the inframarginal plates are also to be seen, making five visible columns of plates on each ray abactinally. There is, therefore, a single radial column bounded on each side by two columns of supramarginal and two of inframarginal plates. The latter columns are the marginal plates both actinally and abactinally, but on the dorsal side are overlapped by the supramarginals. In the center of the disk there is a small but very tumid plate (central disk plate) around which is a nearly complete circle of smaller acces- sory plates and an open space which may represent an anal opening. Outside of this first circle of accessory plates there is a ring of 10 very large plates that in the geologically later species are stellate on the surface, 5 of which are radial and 5 interradial in position. The radial plates are the basal radial plates of the 5 medial columns, while the 5 interradial plates give rise to the 10 larger and conspicu- ous supramarginal ossicles. The basal supramarginals adjoin one another in the axillary areas, and in the geologically later forms these also have stellate surfaces. Between each of the 5 inter- radial plates there is inserted a single plate, making a second but disconnected ring of accessory plates. Madreporite abactinal, more or less large, adjoining one of the large interradial plates and between the basal plates of two supra- marginal columns of ossicles of adjoining rays. No accessory ambital or abactinal plates. Spines appear to be absent everywhere except on the inframarginal and adambulacral ossicles. Otherwise the surface is granular or tuberculate. Inframarginal plates well-developed, tuberculate, and common to both the actinal and abactinal areas. These proximally adjoin the five isolated and large axillary marginals, each one of which occupies .an axil of the rays. Each plate, except the axillary ones, bears on its outer edges one to three apparently smooth, small spines. Axillary plates quite conspicuous, marginal, and one in each actinal axil. In at least one of the species there is a single tiny acces- sory piece orad to each axillary (H. narrawayi), and it is only seen on the inside of the skeleton. Adambulacral plates numerous, one or two to each inframarginal plate and two on each side of each axillary marginal plate. All of these pieces bear at least two spines, one on the inner and another on the outer edges. Oral armature consisting of at least five pairs of pointed and elon- gate basal adambulacral plates; each bears spines like the other ambulacrals. In rare cases issuing from beneath these ossicles may be seen five single stout spines or plates (tori) that are connected REVISION OF PALEOZOIC STELLEKOIDEA. 55 with the pairs of ambulacral armature pieces. In H. narrawayi above or dorsal to the pair of large oral pieces there is another but smaller and modified pair of ambulacral ossicles clearly also a part of the oral armature. Ambulacral grooves originally narrow and regularly tapering. Am- bulacral plates rectangular, practically opposite or very slightly alternating, one to each adambulacral plate. The ambulacralia of each column may be so closely adjoining as to leave almost no space for the podial openings, or there may be laterally between adjoining ossicles very large openings. Genoholotype. The genotype selected by Sttirtz is Pdlasterina rugosa Billings. The above generic definition is, however, largely based upon Palxaster incomptus Meek and P. matutinus Hall. Distribution. Restricted to the middle and upper portion of the Ordovicic. The following are the known species : H. matutinus (Hall). Trenton. H. narrawayi (Hudson). Black River. H. milleri, new species. Trenton. H. incomptus (Meek). Maysvillian and Richmondian series. H. rugosus (Billings). Lower Richmondian, Charleton forma- tion. H. ~batheri, new species. Upper Ordovicic, Girvan, Scotland. Remarks. Sttirtz is unfortunate in selecting as the genotype of Hudsonaster, Palasterina rugosa Billings, a very poorly known starfish from the Richmondian of Anticosti Island. Therefore his original definition of Hudsonaster is not only very imperfect, but the specimen is wrongly interpreted. His characterization of the genus is as follows : Of Palasterina rugosa only the abactinal side is known. On the disk occur orna- mented, deeply crenulate, star-shaped plates. Four [the specimen is distorted and has forced into view one of the inframarginal columns, while the one on the other side of the ray is shoved out of sight] regular and decided columns mark the rays, and of these the outer rows are stronger than the inner columns. The outer columns consist of unmistakable marginal plates [one is inframarginal, the other is a supramarginal column]. According to Billings there are even traces of the actinal marginals to be seen from the abactinal side. Spines absent [the usual spines of this genus are also present here]. The arms terminate sharply in points. Sttirtz is correct in saying that this species can not be included in Palasterina and also that the genus has come to embrace a variety of unrelated genera. However, he does not say what are the actual generic characters of Hudsonaster, but as P. rugosa is not a Palasterina he concludes that as "new generic names are required I propose Hudsonaster for Palasterina rugosa Billings. " This species is undoubtedly congeneric with Palseaster incomptus, a form fully described in this work both from the actinal and abactinal sides, and is therefore taken as the genotype for the above generic defini- tion. 56 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. The adopted genotype of Protopalseaster (P. narrawayi) is described at length and in detail by Hudson, but, thinking the specimen showed the actinal side, he described the ambulacrals as epineural plates roofing over the ambulacral furrow, a character that he of course recognized as wholly anomalous for starfishes. Raymond later on showed that the actinal side of this specimen lies buried in the rock and that the entire abactinal plates are now gone, so that the view is of the actinal plates from their inner side, i. e., an inter- nal view of the ventral skeleton (later Hudson combats this view). Under these circumstances nearly all of Hudson's generalizations are fallacious and especially his statement that "we are dealing with an unrecognized and very archaic morphological type which links the Edrioasteroidea with the Stelleroidea " (1912 : 24) . A proper interpretation of the specimen shows it to be a primitive but other- wise a normal early Paleozoic starfish. Even though Hudson's genus is far better determined generically than that of Stiirtz, it must give way to the older name Hudsonaster. No one regrets this more than the writer, because the generic name Protopalseaster ex- presses the phyletic relationship of these primitive starfishes and is a name that he has had in manuscript for the past 15 years. Hudsonaster has its nearest known relationship in Palseaster, but differs at once from the latter in that the rays have abactinally five columns of plates, while Palseaster has but four. In other words, Palseaster has no radial columns of ray plates, but their place is oc- cupied by numerous small accessory pieces. The abactinal disk of Palseaster is also quite different hi having numerous small acces- sory ossicles, while in Hudsonaster this area is devoid of these pieces and instead there are disk plates that are larger, far less numerous, and with a definite arrangement. The plate arrangement of Hudsonaster is simple and primitive in that it distinctly retains to maturity much of the larval plate structure of recent species and does not develop accessory ambital or accessory abactinal plates as do its descendants Mesopalseaster and particularly Promopalseaster. In Hudsonaster the axils are occupied by single, large, axillary marginals, against which rest the columns of inframarginal plates of adjoining rays. In the other two mentioned genera, interbrachial arcs begin to develop by crowd- ing orally the axillary marginals, when, as in Mesopalseaster , the proximal plates of the inframarghial series of adjoining rays abut against each other. In Promopalseaster this crowding of the infra- marginal plates into the interbrachial areas is continued, always a pair of plates at a time passing orally. For other remarks on the development and distinction between these genera, see Mesopalse- aster and Promopalseaster. REVISION OF PALEOZOIC STELLEKOIDEA. 57 In the Lower Carboniferous genus Neopalseaster the primitive disk structure of Hudsonaster is also retained in its essential fea- tures. Otherwise these genera are easily distinguished, as is pointed out in the remarks on Neopalseaster. HUDSONASTER MATUTINUS (Hall). Plate 2, fig. 2; plate 3, fig. 2; plate 5, figs. 1, 2. Asterias matutina HALL, Nat. Hist. N. Y., Pal., vol. 1, 1847, p. 91, pi. 29, figs. 5o, 56. Palxaster matutinus HALL, Twentieth Eep. N. Y. State Cab. Nat. Hist., 1868, p. 283, pi. 9, fig. 2; rev. ed., 1868=1870, p. 325, pi. 9, fig. 2. Petraster rigidus (part) BILLINGS, Geol. Surv. Canada, Can. Org. Bern., dec. 3, 1858, pi. 10, fig. 36 (not fig. 3a). HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist.,1868, p. 294; rev. ed., 1868=1870, p. 337. Original description. "Body small, with five radiating arms, arms elongated (length twice and a half the width of the body), terete, acute, composed of three rows of plates [abactinally], which join above by their lateral margins. Beneath each upper lateral row of plates there is an inferior lateral range [inframarginal] vis- ible, leaving the sulcus beneath the middle row, which is often depressed. The dorsal places are somewhat hexagonal; those of the arms quadrangular; surface punctate or granulate, perhaps from the removal of the spines covering the surface. "The specimen is considerably crushed, and two of the arms broken off at the base; the madreporiform tubercle upon the back is not visible in our specimen. From its condition, the structure can not be entirely made out, but it is sufficiently clear to enable any one to recognize the species." Emended description. Specimen of the usual size (No. 3, Mus. Comp. Zool.) measures: R, = 14 mm., r = 3.5 mm., R = 4r. The larg- est specimen (Ottawa, Billings collection): 11 = 16 mm., r = 5 mm., R = 3.1r. The smallest New York specimen (No. 26, Mus. Comp. Zool.): R = 7 mm., r = 2.5 mm., R = 2.8r. Width of rays at base from 2.5 mm. to 5.5 mm. Rays short, tapering rapidly, convex abactinally and flat actinally. Disk comparatively large, probably convex abactinally during life. Abactinal area of rays consisting of 5 ranges of large, thick, and tumid granular plates. The 3 medial ranges, when not distorted by pressure, conspicuous, with the 11 to 12 plates of each column closely adjoining and considerably elevated above the inframarginal ranges. The proximal plates of each supra- marginal range lie against the large interradial plates and are more or less superposed on the large axillary marginals. In the center of the disk there is a distinct highly convex plate, around which is arranged a nearly complete ring of seven smaller pieces 58 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. and a vacant space which may be an anal opening. Five other small quadrangular plates are intercalated, one proximally to each of the large radial basal plates. Outside of these 12 intercalary plates there is a ring of 10 quite large and conspicuous plates, 5 of which are radial and 5 interradial in position. The radial plates are largest and are the basal pieces of 5 medial or radial columns, while against the 5 interradial plates rest the 10 inframarginal columns. Madreporite abactinal, quite small, situated immediately above a large interradial plate and between two basal supramarginal plates. Its position in relation to the anal opening is apparently as in P. incomptus. Inframarginal plates in mature examples about 14 to a col- umn, and when not displaced completely, inclosing the adarnbu- lacral plates and margining the rays. The axils are occupied by a very large, oval, axillary marginal plate. Each inframarginal bears upon its outer surface one to three short, slender, pointed spines, which are finely striated longitudinally. Adambulacral plates small, subquadangular in outline and about 19 in a column in specimens having 13 inframarginals. Each plate bears on both the outer and inner edges a single spine and is like those of the inframarginal ranges. The oral armature consists of 5 pairs of elongate adambulacrals, each pair being situated directly orad of the large axillary marginal. Ambulacral plates I -shaped, practically opposite or but very slightly alternating, one to each adambulacral plate. In the ex- panded medial portion the ambulacralia are slightly superposed one above the other proximally, while the lateral portion of each plate is very slender, leaving between adjoining pieces comparatively large podial openings. Locality and formation. From the Trenton limestone, as follows: The type-specimen was found in the " shelly layers" at Trenton Falls, New York (Amer. Mus. Nat. Hist., No. 199), and another specimen was secured there by Mr. Rust (New York State collection). Dr. Walcott collected a slab with three individuals in the Lower Trenton at Rathbone Brook, near Newport, Herkimer County, New York; and Mr. Taylor got two specimens at Deerfield, Oneida County, in the same State (Mus. Comp. Zool., Nos. 26 and 3). Mr. W. R. Billings found a specimen in the Trenton at Lachine, Quebec, and another from Government House Bay, Ottawa. Remarks. H. matutinus is intermediate in structure between the older E. narrawayi and the younger H. incomptus. It is larger than the former, has more plates in each of the columns, while the ambu- lacrals are deeply cut out for the podia, which is not the case in H. narrawayi. As H. matutinus is most closely related to H. incomptus more detailed comparisons must be made between them. Actinally KEVISIOISr OF PALEOZOIC STELLEEOIDEA. 59 the former differs in the wider ambulacral grooves. The difference of greatest value, however, is found on the abactinal area of the disk. In both species there is a central disk plate surrounded by a ring of accessory plates. In H. matutinus there are eight plates in this ring and in H. incomptus there are seven. Between all the basal radials in the latter species there is an accessory plate, or these are five in number, while in the former there are only two such plates. Again, the proximal plates of the supramarginal columns in H. incomptus are larger than any others of the abactinal area, while in H. matutinus they are smaller than the basal radials and their form and ornamentation are quite different. The madreporites are also quite different. In general, H. incomptus is larger, stouter, the rays broader and the disk has a greater number of accessory pieces. These features are still more pronounced in H. rugosus (Billings). Viewing these differences developmentally, the distinctions are slight, which proves the three forms to be genetically related, but specifically the differences are sufficiently marked to be regarded both morpholog- ically and geologically as of specific value. Apotype.Csii. No. 60603, U.S.JST.M. HUDSONASTER NARRAWAYI (Hudson). Plate 1, fig. 1; plate 2, fig. 1; plate 4, fig. 1. Protopalseaster narrawayi HUDSON, Ottawa Naturalist, vol. 26, 1912, p. 25, pis. 1-3; vol. 27, 1913, pp. 77-84, pis. 8, 9. RAYMOND, Ottawa Naturalist, vol. 26, 1912, pp. 105-107. HUDSON, Bull. N. Y. State Mus., No. 164, 1913, pi. 5 SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palseontogr. Soc. for 1913), 1914, p. 21, fig. 19. Description. Specimens small, an average individual from St. Paul, Minnesota; measuring: R = 5 mm., r = 3 mm., R = 1.6r. Width of ray at base 2.7 mm. The three good specimens show only the actinal side while the fourth free siliceous one is too poorly preserved to make out more than the form. Inframarginal columns have from 6 to 9 plates, while the adambu- lacrals and ambulacrals each have from 11 to 12. Ambulacral plates large and solid, rectangular, with slightly rounded ends. As these plates are not t- -shaped as in other species of this genus it may be shown that this character is of generic value, in which case Proto- pal&aster will have to be revived. Inside of each of the large axillary plates lies a pair of oral armature pieces that are in form very much like the adjoining adambulacrals. From beneath these ossicles projects a stout and long, rounded, arma- ture rod, the torus (a tooth or chisel), the 5 teeth meeting together in the center of the mouth. These armature rods and more especially the 10 pairs of ambulacral armature pieces are seldom seen in fossil starfishes, and none shows it more clearly than the Yale specimen 60 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. of H. narrawayi. In the holotype of this species, here illustrated, are seen additional pairs of oral armature ossicles. These are situated directly above or dorsal to the pairs of large armature plates, and are probably overlapping ambulacralia modified into the actinostomial ring. Locality and formation. In the Black Kiver formation, as follows: The holotype was found by Mr. J. E. Narraway at City View, near the Central Experimental Farm at Ottawa, Canada; it is now in his private collection. An excellent specimen was found by Mr. Townshend near Kirkfield, Ontario, and is now in the Peabody Museum of Yale University. Many years ago Dr. Ulrich found a fine specimen at St. Paul, Minnesota, and another good one in the Lower (Glade) Lebanon limestone of the Stones River series at Shelbyville, Tennessee. These two specimens are in the United States National Museum (Nos. 60602, 60619). Mr. Moritz Fischer secured a coarsely silicified individual of this genus at Curdsville, Kentucky, and this is now in the Beecher collection in the Yale Museum; it may, however, prove to be of another species, distin- guished by its stouter appearance, wider disk, and smaller axillary plates. Remarks. Originally the writer included specimens of this form under P. matutinus, regarding them as but young individuals of Hall's species. However, as the specimens are all geologically older, are always smaller and stouter, and with fewer plates in all of the columns, Hudson's species is retained as valid. It is the ancestral form and the most primitive species of Hudsonaster and hence the most primitive Paleozoic starfish. It is primitive because of its extreme simplicity of structure, in that the columns and plates are reduced to the smallest number and all of them are closely adjoining and without intercalary plates of any kind. Then the spines are developed only on the actinal side and as protection for the podia. For further information see generic remarks under Hudsonaster. Cat. Nos. 60602, 60619, U. S. N. M. HUDSONASTER MELLERI, new species. Plate 4, fig. 2. Description. Of this form there is at hand but a single poorly preserved individual showing the actinal side. It is a more slender species than H. matutinus, has a larger and more rectangular axillary plate, and the proximal inframarginals have each about 13 to 15 plates and the adambulacrals from 18 to 20. No ambulacrals are preserved. The specimen measures: R = about 12 mm., r = 3 mm. Width of ray at base 4 mm., at mid-length 3 mm. Locality and formation. Found by Prof. Arthur M. Miller, after whom the species is named, in the Lower Lexington (Wilmore) REVISION OF PALEOZOIC STELLEKOIDEA. 61 limestone of the Trenton, in Fayette County, Kentucky. The holotype is in the collection of the State University of Kentucky at Lexington. HUDSONASTER INCOMPTUS (Meek). Plate 6, figs. 1, 2. "Goniaster" of AGASSIZ, GRAHAM, ANTHONY, and JAMES, Amer. Journ. Sci., ser. 2, vol. 1, 1846, p. 441 (nomen nudum). Palseaster incomptus MEEK, Amer. Journ. Sci., ser. 3, vol. 4, 1872, p. 275; Geol. Surv. Ohio, Pal., vol. 1, 1873, p. 64, pi. 4, figs. 5a, 5&. Palseaster simplex MILLER and DYER, Journ. Cincinnati Soc. Nat. Hist., vol. 1, 1878, p. 29, pi. 1, fig. 6. MILLER, N. Amer. Geol. Pal., 1889, p. 266, fig. 380. Palseaster darlcei MILLER, Journ. Cincinnati Soc. Nat. Hist., vol. 1, 1878, p. 102, pi. 3, fig. 5. Palseaster clarkana MILLER, Journ. Cincinnati Soc. Nat. Hist., vol. 3, 1880, p. 236. Original description (P. incomptus). " Small; rays rather short, or only about once and a half as long as their breadth at their inner ends, and rapidly tapering to their outer extremities, which are somewhat obtusely angular. Disk equaling in breadth the length of the rays. Dorsal side of the rays composed each of three rows of pieces that are wider than long, about nine in each row, and increase rather rapidly in size inward to the margin of the disk, which is composed of smaller pieces; a few very minute pieces apparently some times occur between the rows on the dorsal side of the rays [none are apparent in our material]. Surface of the dorsal pieces a little granular, but apparently without spines. Madreporiform piece rather small, a little oval, or almost circular [or trapezoidal], nearly flat, and marked by fine, irregularly interrupted, radiating striae. Ventral side unknown." Emended description. B, = 11 mm., r = 6 mm., R = 1 .8r. Width of ray at base 6 mm. Another larger specimen measures: R = 16 mm., r = 7 mm., R = 2.3r. Width of ray at base 8 mm. Smallest speci- men: 11 = 4.5 mm., r = 2 mm. Rays short, stout, tapering rapidly, convex abactinally and longi- tudinally ridged, flat actinally. Disk large, probably very convex abactinally during life. Abactinal area of rays consisting of five columns of large, thick, tumid, alternating, granulose plates. In the center of the disk there is a distinct, highly convex plate around which is arranged a nearly complete ring of seven smaller pieces and a comparatively large open space which possibly represents an anal opening. This ring is fol- lowed by another composed of five large interradial plates between each of which laterally and somewhat centrally is intercalated a smaller plate radial in position. Laterally and distally to each of the five large interradial plates are five large radial plates which are the basal pieces of the five radial columns. Against the basal plates of the medial ranges there adjoin laterally two large stellate ossicles, 62 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. each of which is the proximal plate of a supramarginal column. The rays are margined by the inframarginal plates. There are 10 plates in each radial and 9 in each supramarginal column. Madreporite abactinal, situated interradially and dis tally against the large interradial plate of the second ring, which is in front of the anal opening, and between two of the proximal supramarginal plates. It is trapezoidal in form, flat, and marked by very fine, dichotomous, radial striae. Inframarginal plates completely inclosing all other plates and com- mon to both the abactinal and actinal surfaces. Usually there are 9 or 10 in a column, not counting the axillary plates. Spines are present at least along the inner edges of the inframarginals. Abactinally the axillary marginals are not prominent but actinally they are very conspicuous. No spines appear to be connected with these ossicles. Adambulacral plates small, subquadrate, usually 2 to each inframarginal or from 18 to 22 plates in each column. From 4 to 6 are situated around the inner edge of each axillary, of which the 2 central ones are largest, being one of the 5 pairs of plates making the oral armature. The adambulacral plates originally bore on their lateral and inner edges numerous, very short, thick spines. Ambulacral plates nearly entirely concealed by the adambulacral plates, apparently as numerous as the latter, and, so far as can be determined, like those in P. matutinus Hall. Locality and formation. In the lower, middle, and upper beds of the Cincinnatic (Ordovicic). Ten specimens have been studied: Four in the Harris collection in the United States National Museum (No. 40882) from the Richmond formation (Waynesville division) about Waynesville, Ohio; another from the same place and formation in the University of Chicago Museum (Gurley collection, No. 10977); one from the hill quarries in the Maysvillian series (Vaupel collection, U. S. Nat. Mus., No. 60616) back of Cincinnati; another, said to be from Cincinnati, in the University of Toronto (Walker collection, No. 691 H. R.); the type-specimen, now in the Museum of Com- parative Zoology, also from Cincinnati; one in the Ulrich collection of the United States National Museum (No. 60617), found back of Covington, Kentucky, in the Maysville formation, about 315 feet above low-water mark in the Ohio River; and the type of Palxaster simplex Miller, found near Raysville, Ohio (Faber collection, Uni- versity of Chicago Museum, No. 8830). The holotype of P. darkana Miller is in Professor Harper's collection at Cincinnati, while another specimen is in the Ulrich collection of the National Museum (No. 60618). Remarks. The writer has seen the type-specimen of Palseaster simplex Miller, which agrees with the description and figure given KEVISION OF PALEOZOIC STELLEEOIDEA. 63 by its author, and it is like other specimens in the Harris collection known to be H. incomptus (Meek) . Of the latter, Meek only described the abactinal area, while Miller's specimen shows the actinal side. The four specimens in the Harris collection unmistakably affirm that both forms are identical, and since Meek's name has priority it is here accepted. * For remarks as to the relationship of this species see H. matutinus. Through the kindness of Prof. George W. Harper, principal of the "Woodward High School, Cincinnati, Ohio, the writer was enabled to study the type of Palseaster darkana Miller." The specimen is a little abraded, but shows all the characters of H. incomptus except that it is much smaller in size. The abactinal area on a cursory examina- tion shows but three columns, but in several places can be seen pro- jecting small parts of the inframarginal columns. The disk also shows the same general arrangement and number of plates as in H. incomptus. There is therefore nothing except the smaller number of plates in the columns to distinguish P. darkana from H. incomptus. Bearing in mind that the former is but one-third the size of the latter, it is not remarkable that P. darkana should have but 6 or 7 plates while H. incomptus has 9 to 10 plates in the same columns. The entire difference is accounted for when one regards P. darkana as the young of H. incomptus. Another small but very poorly preserved specimen showing the actinal area is in the Ulrich collection of the National Museum and agrees with the type of P. darkana. Its measurements are: 11 = 4.5 mm., r = 2 mm., K = 2.2r. The only ontogenetic character observable in these small speci- mens of H. incomptus is that the columns have fewer ossicles, and the same is true of the abactinal area of the disk. It has a central disk plate, the first ring of small plates and the second ring of basal radial and interradial plates. But no intercalated small plates radial in position are present. The second ring of 10 plates and the central plate are the most prominent. If one were theoretically to reduce the size of the animal still more, the plates of the first ring would be the first to go and the rays would be simultaneously very much shortened, probably obsolete, resulting in a hemispheric nonstellate starfish. The abactinal area seemingly will then have but a central disk plate and 10 others, 5 radial and 5 interradial in position. On the actinal side there will be but the 5 axillary marginal plates separated from one another by the ambulacra! furrow and the ambulacralia. This hypothetic animal therefore closely resembles a very young Comatula without the stalk or column, and also the stage of living asterids following the brachiolarial. Cat. Nos. 40882, 60616, 60617, 60618 U.S.N.M. 50601 Bull. 8815 5 64 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. HUDSONASTER RUGOSUS (Billings). Plate 3, fig. 1. Palasterina rugosa BILLINGS, Geol. Surv. Canada, Rep. of Progress for 1853-1856, 1857, p. 291; Geol. Surv. Canada, Can. Org. Remains, dec. 3, 1858, p. 77, pi. 9, figs. 2a-c. Palasterina rugosa WRIGHT, Mon. British Foss. Echinod., Oolitic, vol. 2, pt. 1 (Paleeontogr. Soc. for 1861), 1862, p. 27. Hudsonaster rugosa STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 56, 1900, pp. 218, 224, 225. Original description. "Two inches [not over 47 mm.] in diameter, rays five, acute at their apices and rapidly enlarging to a breadth of four lines [about 11 mm.] at the disk, which is eight lines [about 17 mm.] in width. The specimen shows the upper [or abactinal] side of the fossil only; some of the plates are absent from the center of the disk, but those which remain are very prominent in their centers, and roughly ornamented by four or five deep crenulations or furrows from near the center to the edges, producing a star-like appearance resembling a half-worn plate of Glyptocrinus decadac- tylus; their diameter is from one to two lines [largest diameter 4 mm.]. "The rays are composed (at least the backs and sides of them) of four rows of plates [due to distortion, a careful analysis shows five columns] which are so very, prominent that they appear to be almost globular, and even pointed in their centers, the central [radial] rows are the smallest [as in H. incomptus]', the first four plates of the outer [supramarginal] row occupy three lines in length, and of the inner [radial] rows nearly as many. Toward the point of the arm all diminish rapidly in size. "Beneath the outer [or supramarginal] rows two others can be seen [the inframarginal], which are probably the outer marginal plates of the under [or actinal side], corresponding to those of Petrasier rigidus." Locality and formation. Richmondian series, Charleton formation, at Charleton Point, Anticosti Island. Two specimens collected by J. Richardson are now in the Victoria Memorial Museum at Ottawa, Canada (No. 1999). Remarks. The same description was republished in 1858 with the addition of three figures. These, and a direct examination of the two type-specimens, show that this species clearly belongs to Hudsonaster and that it is closely related to H. incomptus. The smaller free specimen preserves a portion of the disk and parts of two rays and shows clearly the presence of a single large axillary plate. H. rugosus, however, attained a considerably larger growth, has a greater number of plates in each column and these are described as "almost globular," while in H. incomptus the abactinal plates are very convex but not globular. Then the basal supramarginal and REVISION OF PALEOZOIC STELLEROIDEA. 65 radial plates are much more decidedly ornate. Each of the supra- and inframarginal columns has 14 plates against 9 in H. incomptus, and of adambulacrals there are not less than 24 in a column against 18 to 22 in the same form. All of the ossicles are as strongly papillose but not more so than in H. incomptus. The detail of the disk is not determinable. HUDSONASTER BATHERI, new species. Plate 3, fig. 3. Tetraster wyville-thomsoni NICHOLSON and ETHERIDGE (part), Mon. Silurian Foss. Girvan Dist., Ayrshire, fasc. 3, 1880, p. 324, pi. 21, figs. 1, 2 (not the other figs.). A small Hudsonaster. R = 6 mm., r = 2.7 mm. The largest specimen: R = 8 mm. Actinal side only known. Inframarginal columns the largest, with 8 or 9 distinctly tuberculate plates. Inside of these are the columns of narrower and slightly shorter adambulacrals that lie somewhat below the level of the inframarginals ; there are 10 or 11 of these plates bounding the ambulacral grooves, and the pieces of adjacent columns are opposite or nearly opposite one another. The ambulacral grooves are very narrow and deep and no ambulacralia are discernible. Formation and locality. Dr. F. A. Bather made wax squeezes for the writer from three natural molds in the collection of Mrs. Robert Gray, Edinburgh; the originals are from the Upper Ordovicic of Thraive, Girvan District, Scotland. The holotype is the specimen illustrated as above cited. Remarks. Nicholson and Etheridge confused at least one of these specimens with their Tetraster wyville-thomsoni , a species of totally different construction, a cryptozonian, described elsewhere in this work. Actinally the new species is in harmony with Hudsonaster and although the abactinal side is unknown, it is thought that it will be found to be like that in H. matutinus. Cat. No. 60601, U.S.N.M. Genus SILURASTER Jaekel. Text fig. 6. Siluraster JAEKEL, Zeits. geol. Ges., vol. 55, 1903, Protokol, p. 13 (106), figs. 2, 3, on p. 108. SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palaeontgr. Soc. for 1913), 1914, p. 18. Description. "A typical asterid with opposite ambulacralia and with well-developed marginalia, therefore a true form of Phanero- zonia. The strongly folded madreporite lies dorsally in an inter- radius." 66 BULLETIN UNITED STATES NATIONAL MUSEUM. l2 From the two excellent illustrations the following is made out: Inframarginals conspicuous and as large as the adambulacrals. Both columns finely tuberculate, with the ossicles of each alternating with one another. Proximally the inframarginals of adjacent rays rest on a somewhat larger axillary plate. The interbrachial areas are therefore made up of single axillary ossicles (Jaekel's marginal oral) and in front of these lies a pair of pointed adambulacrals, the conspicuous pieces of the oral armature. These are perforated for the continuation of the ring canal. The podial openings as usual lie in the corners of four adjacent adambulacrals and ambulacrals, but there is a de- cided tendency to restrict them to only one of the am- bulacral corners. Abactinal area unknown. Remarks. This genus appears to be closely related to Hudsonaster, but FIG. 6.-SILT7RASTER PERFECTUS (JAEKEL'S FIGURE 3). Oj TO Og, AMBU- , -1 o, -7 , LACRALIA OF ONE SIDE; ABOVE THESE ddi TO ddg, THE ADAMBULACRALIA, l l6 ' 1 adi AS MOUTH CORNER PIECE; dfg, PODIAL GROOVES; dpff, PODIAL OPEN- DlOrC fully WOrked INGS; m, INFRAMARGINALIA; mo, THE SINGLE AXILLARY INFRAMARGINAL; , , n RC, RING-CANAL IN MOUTH CORNER PLATES. Ut lt} ^ ^maUl aS an obscure genus. Genoholotype and only species. S. perfectus Jaekel. Ordovicic (D4), Zahorzan, Bohemia. Family PAL^ASTERID^E Gregory (emend.). Palaeasteridae (part) MILLER, N. Amer. Geol. Pal., 1889, p. 216. GREGORY, Geol Mag., dec. 4, vol. 6, 1899, p. 344 (contains Palseasterinae, Xenasterinse, and Lindstromasterinse); Treat. Zool., vol. 3, Echinoderma, 1900, p. 250. Palsecisterinas SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 244. Palxgoniasteridx (part) STURTZ, Palaeontographica, vol. 36, 1890, p. 246 (contains Aspidosomella and Pal&aster). Primitive, derived, five-rayed Phanerozonia with the ambulacrals slightly alternate or opposite. Interbrachial arcs incipient or some- what enlarged, and occupied by single axillary marginal plates. No accessory interbrachial or axillary interbrachial plates present, but there are developed accessory abactinal ray ossicles. Actinal plates consisting of ambulacrals, adambulacrals, and inframarginals. Mad- reporite abactinal. Contains : Palseaster Hall. Australaster, new genus. REVISION OF PALEOZOIC STELLEROIDEA. 67 Genus PALSEASTER Hall (emend.). Plate 7, figs. 1-4. Palseaster HALL, Nat. Hist. N. Y., Pal., vol. 2, 1852, p. 247; Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 282; rev. ed., 1868=1870, p. 324. ZITTEL, Handb. Pal., vol. 1, 1879, p. 452. Palseaster NICHOLSON and ETHERIDGE (part), Mon. Silurian Foss. Girvan Dist., Ayrshire, fasc. 3, 1880, p. 319. (Not Palseaster SALTER, Ann. Mag. Nat. Hist., ser. 2, vol. 20, 1857, p. 323= Uras- terella.) (Not Palseaster of most AUTHORS.) Generic characters. Disk small and without distinct interbrachial arcs. Rays five and slender. Abactinal area of disk centrally occupied by a large plate, then a ring of 5 separated, somewhat smaller, radially situated ossicles, and inside the axillary pieces there are 5 subtriangular interradial plates, against which lie 10 very large and conspicuous proximal supramarginal plates. Between the central, radial, and interradial plates are numerous small accessory pieces which separate the pri- mary ossicles one from another. These accessory plates also continue to crowd from the disk into the rays between the supramarginals, and thus take the place of the radial columns, but are absent in the distal third of the rays. Each ray has therefore 4 conspicuous columns of plates, the 2 supramarginals situated inwardly of the 2 marginal inframarginals. The former columns terminate in 10 very large proximal plates situated in the axils of the disk, while the inframarginals rest against the 5 very large actinal axillary margi- nal pieces. All of the large plates are centrally smooth, bordered by a pustulose area. Madreporite abactinal, of medium size, circular in outline, and radially striate. It lies on top of two proximal supramarginal plates and the adjoining interradial plate. Inframarginal plates thick and large, about as many in a column (13) as in those of the supramarginals (13 or 14). The latter do not completely overlie the inframarginals, since these bound the rays both actinally and abactinally. The inframarginal plates, like the plates of the abactinal area, have each a large, smooth, central space encircled by a pustulose area. From their inner sides articu- late short, thick, blunt spines, of which there are apparently not more than 2 to each inframarginal plate. Axillary marginal plates large, hexagonal, one in each axil and common to both sides of the disk axils. Adambulacral plates small, tumid, alternating in the main with the inframarginal plates throughout the distal two-thirds of the rays, but toward the mouth there are a few more adambulacrals than in- framarginals. (In the type-specimen there are 13 of the latter in a column to 16 of the former.) Oral armature consists of at least 5 pairs of drawn out and pointed basal adambulacrals. Along the inner 68 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. side of each adambulacral plate there is a linear brush of from 5 to 7 short, slender spines. Ambulacral grooves very narrow, tapering but very slowly. Am- bulacral plates obscured by the rock but apparently very short, one to each adambulacral and placed opposite one another. GenoJiolotype (type by monotypy) and only species. P. niagarensis Hall. Siluric (Rochester shale). Remarks. To this genus has been erroneously referred a multitude of Paleozoic starfishes. It is to the starfishes what Terebratula and Rhynchonella used to be among the brachiopods a general dumping ground for Paleozoic forms. This lack of generic conception regarding Palseaster lies somewhat in the fact that Hall's original diagnosis is very meagre and that he did not describe nor illustrate the abactinal surface, probably the most important area amongst Paleozoic aster- ids for specific and generic differentiation. Of the twenty-four Amer- ican species listed under Palseaster by S. A. Miller 1 only one now appears to belong there, the genotype. All others are here referred to Hudsonaster, Urasterella, Mesopalseaster, Promopalseaster, Anorth- aster, Devonaster, and Neopalseaster. No new material of this genus accessible to the writer appears to have been found since Colonel Jewett discovered the original two specimens of Palseaster niagarensis. The writer therefore had to make the most of this material, and to determine the structure of the abactinal area he worked away much of the blue shale from the back of the one good specimen. The finer detailed structure was then revealed by cleaning with caustic potash. In 1858 Billings proposed the genus Stenaster 2 and writes- that "as it has been suggested that the two species hereinafter described should be referred to Palseaster, I give the following figure of the genus in order to show the difference." This figure is a good reproduction of Hall's figure 27. Billings then points out that "if the large plates which border the grooves in Palseaster be adambulacral, then there are only five oral plates, whereas in Stenaster there are ten. But if they be not adambulacral but marginal plates, then Palseaster must have both marginal and adambulacral [the correct view], while Stenaster has only the latter." Hall's generic description is very meager and his figures give the impression that the rays bear but a single column of marginal spiniferous plates. With this evidence one can understand why Billings was in doubt as to whether these plates are inframarginals or adambulacrals and therefore the uncertainty as to the nature of the five large axillary marginal plates. If, however, he had read Hall's accurate specific description, Billings would have seen that an important discrepancy -existed between the description and figures. 1 North American Geology and Palaeontology, 1889, p. 265. 2 Geol. Surv. Canada, Canadian Organic Remains, dec. 3, 1858, pp. 77, 78. REVISION OF PALEOZOIC STELLEROIDEA. 69 Hall writes: " Lower side of arms showing two ranges of plates on each side of the avenue, the outer range composed of strong hex- agonal plates, with an inner range of smaller ones alternating, the latter usually covered by tufts of spines." The type-specimen clearly shows columns of adambulacral and inframarginals bordering the ambulacral grooves and that in each axil there is a single large axillary marginal plate. This matter need not here be followed further than to add that Stenaster originally included two generic types, both of which are widely different from Palseaster. On the other hand, Hall 1 is in error in regarding all of Stenaster as synonymous with Uraster- ella McCoy, a name never properly defined and finally abandoned by its author. Billings originally referred to Stenaster, S. salteri, and S. pulcTiellus. The latter is certainly congeneric with the type of Urasterella as illustrated by Salter, but the former is quite different, as is pointed out elsewhere in the remarks on Stenaster, which is a good genus. Palseaster in many respects reminds one of Hudsonaster, and the actinal generic characters may be regarded as alike in both. The abactinal area is also similar, but in Hudsonaster there are five columns with decidedly fewer plates, while in Palseaster there are no radial columns. The former genus is the more primitive one and the progression in differentiation toward Palseaster appears to be as fol- lows: In Hudsonaster, the rays throughout are composed of five columns of plates, two inframarginals, two supramarginals, and one radial. In Palseaster, there are but four columns in the outer third of the rays, two infra- and two supramarginals, but in the proximal two-thirds where the radials are also absent, their place is occupied by numerous small, irregular accessory plates. This introduction of numerous supplementary plates also takes place to a far greater extent on the disk, where they are crowded in between the larger central and the five basal radial plates. The primitive disk structure of Palseaster is in large part again repeated in Neopalseaster, but otherwise the two genera are different. PALSEASTER NIAGARENSIS Hall. Plate 7, figs. 1^. Palseaster niagarensis HALL, Nat. Hist. N. Y., Pal., vol. 2, 1852, p. 247, pi. 51, figs. 21-23 (not p. 352, pi. 85, figs. 8-10, possibly an undescribed species). BILLINGS, Geol. Surv. Canada, Can. Org. Rem., dec. 3, 1858, p. 78, fig. 1. HALL, Twentieth Hep. N. Y. State Cab. Nat. Hist., 1868, p. 282; rev. ed., 1868=1870, p. 324. QUENSTEDT, Petrefactenkunde Deutschlands, vol. 4, 1876, p. 73, pi. 92, figs. 32, 33. HUDSON, Bull. N. Y. State Mus., No. 164, 1913, pis. 9, 10, 12, 13. Original description. "Body stellate; disk small; arms short, terete with a deep avenue on the lower side, which is margined by strong short spines; centre of plates (in the fossil) nearly smooth, margins i Twentieth Rep. N. Y. State Cab. Nat. Hist., rev. ed., 1808=1870, p. 325. 70 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. strongly granulate; lower side of arms showing two ranges of plates on each side of the avenue, the outer range composed of strong hexagonal plates, with an inner range of smaller ones alternating, the latter usually covered by tufts of spines; a large pentagonal plate inserted at the base of the arms, on the lower side. "In this species, the arms are about twice as long as the width of the disk. There are about 15 or 16 plates in each range from the base to the apex of the arms. The range of the plates margining the avenue are usually not visible, the whole being covered by the short spines, which also partially fill the avenue. " In 1870 Hall added the following important observations: "I have distinctly recognized the two ranges, marginal and adambulacral plates; but the inner ones are not shown in the figure as they should have been, while the large plate at the axil of the ray (though the adjacent small oral plates of the inner range are not seen) is evidently part of an incomplete series, and clearly belongs to the marginal range. "The genus Palseaster has two ranges of plates on each side of the ambulacral groove; marginal and adambulacral plates on the lower side, besides ambulacral or poral plates. The upper or dorsal side has three or more ranges of plates." Emended description. Measurements: R=1S mm., r=5 mm, Width of ray at base 5 mm. Rays long, slender, tapering slowly; abactinally strongly convex and longitudinally ridged; flat actinally. Disk comparatively large, strongly convex and probably abruptly elevated above the rays abactinally. There are no interbrachial arcs, but the five axillary marginal plates occupy the axils and are also partially visible abactinally. Abactinal area of rays consisting of 4 columns of large, thick, convex plates. These are the supra- and inframarginal columns. The supramarginal plates of each ray are distally in contact with one another along the mid-line, but proximally they are forced more and more apart by a series of small, irregular, accessory plates. In the center of the disk there is a small, convex disk plate around which are situated radially 5 other small ossicles, all of which are sepa- rated from each other apparently by a considerable number of minute intercalary pieces. Just inside of the axils are 5 large, sub triangular, uiterradial plates and adjoining these directly are 10 large and con- spicuous ones, the proximal plates of. the supramarginal columns. In each column there are about 13 plates, not counting the proximal piece. Between all of the larger plates of the disk and for two- thirds the length of the rays between the supramarginals, there are numerous small, convex, irregularly shaped, accessory plates. All BEVISIOK OF PALEOZOIC STELLEROIDEA. 71 the larger ossicles of both sides are centrally smooth, surrounded by a finely pustulose border. Madreporite situated interradially and distally upon the three large plates just inside an axillary marginal plate, two of which are the proximal plates of supramarginal columns of adjoining rays. It is circular in outline, flat, and radially striated. Inframarginal plates occupying the outer edges of both the actinal and ab actinal areas; from 13 to 15 plates hi each column. The supramarginal plates are situated above and inside the inframarginals, which inwardly bear short, thick, club-shaped spines. Otherwise they are ornamented like the abactinal plates. Adambulacral ossicles small, subquadrate, each one alternating distally with an inframarginal plate, but proximally they are some- what more numerous, there being from 17 to probably 20 in a column. On these plates lie two series of spines the two short, thick, club- shaped ones of each mframarginal plate while on the inner side of each adambulacral there is inserted a linear brush of from five to seven slender, short spines that are outwardly and laterally directed. Ambulacral grooves very narrow, deep, and gradually tapering. Ambulacral plates not well seen but apparently very short, one to each adambulacral and directly opposite one another. Axillary marginal plates five in number, very large and conspicuous actmally, also showing slightly abactinally. Locality and formation. The only good specimen was found many years ago by Col. E. Jewett in the Siluric (Rochester shale) at Lock- port, New York. A fragment of the ray of a much larger individual, nearly twice the size of this type, was found by the same gentleman at the same place. These specimens are now in the paleontological museum of Cornell University (No. 7331) and were presented by Hon. Ezra Cornell. Long after the study of the form was com- pleted another specimen preserving two rays and showing the actinal side was learned of in the University of Toronto (Walker collection, No. 1008). It was found at Grimsby, Ontario. Remarks. This species has never been well described nor figured, and has accordingly led to endless generic confusion. Almost any Paleozoic starfish with large marginal plates has been referred to Pal&aster. This lack of generic conception was partially due, as stated above, to the fact that the abactinal side remained unknown. The specimen lies imbedded in blue shale showing the actinal area, and the abactinal side was revealed by the writer by cutting away the slab to the starfish and then developing the individual plates with caustic potash. In a general way Palseaster niagarensis resembles several species but all are distinguished by generic characters. The nearest one is Hudsonaster incomptus from the Cincinnatic strata, but it differs at 72 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. once in having five instead of four columns of large abactinal plates. There are also a far less number of the accessory plates so numerous on the disk and the proximal medial areas of the rays in Palseaster niagarensis. Hall 1 refers a fragment to this species which clearly is of quite another species. It is very far removed from P. niagarensis and for the present need not be considered. The specimen hus not been seen by the writer. AUSTRALASTER, new genus. Plate 4, fig. 4. Palseaster ( Monaster) ETHERIDGE, jr. (part), Mem. Geol. Surv. New South Wales, Pal., No. 5, pt. 2, 1892, p. 74. Monaster GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 345 (not of Etheridge, jr.). Australaster SCHUCHERT, Fossiliiim Catalogus, Animalia, pt. 3, April, 1914, p. 12. Austral=sout1n.eTn and aster. A Paleozoic starfish restricted to Australia. Generic diagnosis. Animal attaining a very large size, with five long; slender, almost straight-sided rays, terminating bluntly. Interbrachial areas comparatively small, widely concave, and occu- pied by ponderous, single, axillary marginals and inframarginal plates. Abactinal area unknown (see A. (?) stutcfiburii) . Actinally the animal is bounded by inframarginal plates which however do not appear to be present in the distal third of the rays. If such plates are present in these outer regions, they are either very small or gradually pass from the actinal to the abactinal area, the latter condition occurring ha related genera. The inframarginals in the outer third of the rays are small, subquadrangular, and rounded, thence they increase rapidly in size proximally, become decidedly transverse, and in the axils the columns abut against a very large axillary inf ramarginal. The largest plate is the axillary marginal, orad to which occur the pairs of enlarged oral armature ossicles. The inframarginals appear to be nearly smooth and devoid of spines. The adambulacrals in the distal third of the rays are numerous, very short, but quite wide that is, are decidedly transverse, and decrease rather rapidly hi breadth toward the tips of the rays. Proximally these plates also decrease quickly in width and rapidly pass inside of the inframarginals. Inside of the axillary marginals the plates are again prominent and especially the pairs of oral arma- ture pieces. Each adambulacral plate bears prominent tubercles for the articulation of stout but short spines. The columns have the pieces arranged with a decided slant outward and distally. i Palaeontology of New York, vol. 2, 1852, p. 352 and pi. 85, figs. 8-10. REVISION OF PALEOZOIC STELLEEOIDEA. 73 Ambulacral furrows deep, slender and tapering slowly. Ambu- lacra! plates unknown. Oculars not present. Genoholotype. Palseaster (Monaster) giganteus Etheridge, jr. The specific description and the bibliographic references are included in the generic diagnosis. Kes trie ted to the Lower Marine series of the " Permo-Carbon- iferous" of New South Wales. AUSTRALASTER(P) STUTCHBURH (Etheridge, jr.). Palxaster (Monaster) stutchburii ETHERIDGE, jr., Mem. Geol. Surv. New South Wales, Pal., No. 5, pt. 2, 1892, p. 73, pi. 13, fig. 1. Monaster stutchburii GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 346. This species is found associated with A. giganteus and its preserva- tion is such that the form is very imperfectly known. In some respects it resembles that species and may prove to be a smaller specimen of it, but until the disk marginals are known the differ- ences will remain obscure. If A. (?) stutchburii has the large disk marginals of Australaster, it then seemingly will have the essential structure of that genus. Regarding this the writer wrote Professor Etheridge in 1900 and he replied that " axillary plates are not visible in the specimen". Since the abactinal area of A. giganteus is unknown, it is desirable to give here the abactinal structure in A. (?) stutchlmrii, as it may prove to be congeneric with A. giganteus. " Abactinal surface moderately convex, bearing several (five or six) rows of small convex polygonal plates, inclusive of the margi- nal pieces, and becoming much crowded at the apices of the rays, where they form oblique rows." If these plates throughout the rays are as figured by Etheridge for the distal ends, the strongly quincuncial arrangement of the polygonal, closely adjoining plates is another good generic character for Australaster. PROMOPAL^ASTERID^E, new family. Progressive Phanerozonia with distinct columns of inframar- ginal plates. Interbrachial areas more or less complex, composed either of inframarginals, axillary interbrachials and ambulacrals, or of these with the addition of interbrachial marginals and acces- sory interbrachials, or entirely of adambulacral plates. Ambu- lacrals as a rule opposite, but they may also be slightly alternating. Podial openings through the sutures in the lateral corners of the ambulacral plates, but proximally a few alternate pores may grad- ually pass medially, when there are four columns of podial openings in each ambulacral furrow. Madrsporite abactinal. Abactinal plates very numerous, generally small, either in dis- tinct columns or rows or without either arrangement. The radial 74 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. and supramarginal columns may be very distinct or may be ob- scured as such. Accessory ossicles always more or less abundantly developed. Contains : Mesopalaeasterinae. Promopalseas fcerinse. Anorthasterinse. new subfamily. Palxocwnidde GREGORY, Treat. Zool., vol. 3, Echinoderma, 1900, p. 254. Primitive Promopalseasteridse with the interbrachial areas small, composed of one pair of inframarginals, single axillary infcerbra- chials, and the adambulacral plates. Contains : Mesopalszaster, new genus. Spaniaster Schondorf. Miomaster Schondorf. Devonaster, new genus. MESOPAL^EASTER, new genus. Plates 7-11. Palseaster (part) of AUTHORS. fArgaster HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 287: rev. ed., 1868=1870, p. 329. Mesopalseaster SCHUCHERT, Fossilium Catalogus, Animalia, pt. 3, April, 1914, p. 24. Meso = middle, and Palseaster. Not the largest of Palseasters of authors. Generic characters. Rays five, slender and tapering. Disk small, with incipient interbrachial arcs. Abactinal area with distinct columns of radial and supramar- ginal plates, between which are inserted rows of accessory pieces, one or two on each side of the medial columns. In mature specimens the radial columns are less prominent proximally than distally. The accessory columns proximally may also have here and there inserted on each side a smaller piece, when the area between the supramarginals has not only three or five columns of plates but also other additional small plates. The inframarginal columns margin the rays, while the supramarginals lie somewhat inside and above the former. Between the two columns of marginals there may be a row of accessory ambital plates. The arrangement of the plates of the disk appears to be as follows: In the center there is a small plate around which is a first ring of 7 pieces, followed by a sec- ond ring having not less than 14 plates, and these are nearly of one size. The medial ray columns terminate on the disk in 5 REVISION OF PALEOZOIC STELLEROIDEA. 75 large radial plates, but inside of the proximal supramarginals there is no basal interradial plate as in Hudsonaster. Madreporite unknown. Inframarginal and adambulacral plates as in Hudsonaster, except- ing that two of the former and a single axillary plate are in each interbrachial area. Oral armature so far as preserved adambulacral, consisting of five pairs of basal adambulacrals. A single, pentagonal, axillary interbrachial plate is present in each area, orad to the two axile inframarginal plates. Ambulacral grooves deep and almost closed by the adambulacral columns. Ambulacral plates carinate and apparently one to each adambulacral ossicle. Their relation to one another is not definitely determinable, but apparently they are slightly alternate or nearly opposite. Genoholotype. Palxaster shafferi Hall. Distribution. The species of Mesopalseaster are restricted to America in the upper portion of the Ordovicic and the lower portion of the Siluric. A species doubtfully referred to this genus occurs hi the Lower Devonic of Germany, and another in the Upper Devonic of Maryland. The following are the species of Mesopalseaster: M. shafferi (Hall). Middle and upper beds of the Cincinnatic series. Possibly also in the Utica formation of New York. M. intermedius, new species. Maysville formation. M. finei (Ulrich) . Eden formation. M. (?) lanceolatus, new species. Utica formation. M. proavitus, new species. Eden formation. M. (?) wilberanus (Meek and Worthen). Richmond series. M. (?) dubius (Miller and Dyer). Upper Trenton. M . (?) antiquus (Troost) . Upper Trenton. M. (?) parviusculus (Billings). Siluric (Lower Arisaig). M. (?) cataractensis, new species. Siluric (Cataract formation). M. granti (Spencer). Siluric (Cataract formation). M. bellulus (Billings). Siluric (Rochester shale). M. caractaci (Gregory). Caradoc of England. M. (?) acuminatus (Simonovitsch) . Lower Devonic of Germany. M. (?) clarki (Clarke and Swartz). Upper Devonic (Jennings formation) . Remarks. Mesopalseaster apparently developed directly out of Hudsonaster in that the single axillary marginal plates are in the former genus in nearly all the species pushed inward (interbrachial) and their former place occupied by the proximal plates of adjoining inframarginal columns. The same generic feature also distinguishes Mesopalseaster from Palseaster. Further, in Hudsonaster there are no abactinal accessory ray plates, but in Mesopalseaster one or two more or less completely developed columns are inserted on each side 76 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. of the radial columns or, in other words, between the radials and supramarginals. Between the inframarginal and supramarginal plates in Mesopalseaster there are also either a few incipient accessory plates, which in young specimens are wholly absent, or there is a complete column of these ambital plates. The accessory and axillary inter- brachial plates are additions to the generic structure of Hudsonaster and Palseaster, and further distinguish Mesopalseaster from both. This progression toward a greater number of columns of abactinal plates attains its maximum in Promopalseaster and is most marked in P. magnificus, which has not less than 28 columns at the base of a ray and 5 interbrachial marginal plates in each interbrachial area. The ambulacral furrows, which are very narrow in Hudsonaster, are likewise so in Mesopalseaster, but are wide in Promopalseaster. Spi- nosity, which is practically absent abactinally in Hudsonaster, is developed in Mesopalseaster and very pronounced in some forms of Promopalseaster (P. spinulosus and P. dyeri) . This is apparently also true for the spines of the actinal area. The youthful plate structure of recent species seems to be retained to maturity in Hudsonaster and somewhat so also in the genotype of Mesopaldeaster but in Promopalseaster the central area of the disk is occupied by numerous very small plates and no definite arrange- ment is discernible, though this is in large part due to their displace- ment through fossilization. However, it is certain that the basal radials and interradials do not increase nor maintain a relative size as in Hudsonaster and in the genotype of Mesopalseaster (here radials only), but must have diminished to that of the accessory plates from which they are now not distinguishable. From Hudson- aster through Mesopalseaster into Promopalseaster the animals are constantly increasing in size, and this also continues in the species of the latter genus, attaining culmination in P. magnijicus and P. dyeri. While the generic abactinal differences between Hudsonaster and Mesopalseaster are easily made out, this is not so readily accom- plished between the latter and Promopalseaster. All of the abactinal generic characters of Mesopalseaster are more pronounced in Pro- mopalseaster excepting that in the former the ambital accessory plates are almost always undeveloped as columns, while in the latter these plates are always well developed in one or more columns. When the actinal area is shown, Mesopalseaster is at once distin- guished from Promopalseaster in that it has but one axillary inter- brachial plate in each interbrachial area, while the latter has always two or more up to five ( = interbrachial marginals) in each area. Devonaster in some respects is closely related to Mesopalseaster and may be in the line of descent from the latter. However, the disk in Mesopalseaster has far fewer abactinal plates, and there are no distinct interbrachial arcs with numerous small accessory plates REVISION OF PALEOZOIC STELLEROIDEA. 77 as in Devonaster. The pieces of the radial columns in Mesopalseaster adjoin one another and are not separated by intercalary plates, nor are these nearly so numerous between the radial and supramarginal columns as in Devonaster. These accessory plates in Devonaster are also irregularly disposed and not in columns as in Mesopalseaster. The name Argaster has not been denned nor is it mentioned in such a way that anyone will recognize that Hall intended it for a new genus or subgenus. In addition, its genotype (Asterias antiqua = Mesopalse- aster (?) antiguus) is a poorly known species. Under these circum- stances, and the additional one that the name Argaster never has had any standing as a genus, it should not now be rehabilitated by paleontologists. MESOPALSEASTER SHAFFER! (Hall). Plate 8, figs. 1, 2. Palseaster sha/eri HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 284; rev. ed., 1868=1870, p. 326, pi. 9, fig. 1. MEEK, Geol. Surv. Ohio, Pal., vol. 1, 1873, p. 66, pi. 4, fig. 1. Original description. "Body of about medium size composed of five tapering, acutely pointed rays, which, in the specimen examined, measure seven-eights of an inch from the center of the disk to the extremity. The lower side of the ray is formed of two ranges of plates bordering the ambulacral groove. The marginal range consists of moderately convex plates which gradually decrease in size from the base to the extremity of the ray, 22 or 23 in number [not more than 20 exist in any ray of the type-specimen], besides a small terminal one at the angle of the range; each plate of the marginal range is marked on its outer surface by a comparatively large cicatrix for the attachment of a strong spine. The inner range of plates (adam- bulacral) are somewhat smaller, about the same in number, alternat- ing with those of the marginal range; the basal pair (oral plates) are elongate-triangular, and slightly constricted near the middle. Ambulacral a,reas narrow, composed of a double range of poral plates, which at the middle of the ray are about of equal length and breadth. Pores not observed. [The writer could not make out these plates in the type-specimen and believes those described to be the inner sides of the actinal plates.] Upper surface of the ray composed of three ranges of subnodose plates, the outer ranges bearing a strong spine on each plate; the central range apparently destitute of spines. No spines have been observed, but the plates are strongly carinate or pointed." Emended description. The largest specimen in the Harris collec- tion measures: B, = 19 mm., r 5.5 mm., R = 3.5r. Another specimen of the same collection: R = 16 mm., r = 5 mm., R = 3.1r. The smallest known specimen in the Vaupel collection: R = 6.5 mm., 78 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. r = 2.5 mm., R = 2.6r. The type-specimen, which is the largest knowii example, measures: R = 21 mm., r = 6 mm., R = 3.5r. Width of ray at base in various mature examples 5 mm. to 7 mm. Rays more or less long, slender, abactinally slightly convex, with marked longitudinal ridges, actinally flat or concave. Disk of medium size, abactinally convex and without interbrachial arcs. Abactinal area of rays occupied by a central range of radial plates, on each side of which there is a range of accessory ossicles which however do not attain the distal portion of the rays. In mature rays there are in the proximal region also inserted a few plates on each side of the intercalary columns. Outside of the latter are the ranges of well-defined supramarginal plates which slightly overlap the inframarginals. The marginal and radial plates are most promi- nent, and all are highly carinate or pointed. There is a single large spine on the apex of all radial and accessory plates and several smaller ones are also arranged around it. The supramarginal plates have numerous granules which bore articulating spines. The plates of the disk can not be clearly made out, but in a small specimen somewhat distorted there appears to be a small central plate sur- rounded by a first ring of 7 ossicles followed by a second ring of 14 plates. These pieces are all nearly of a size. The medial ray columns terminate in 5 large basal radial plates, while inside of the proximal inframarginals there is no large interradial plate as in Hudsonaster incomptus. Ambital area very illy defined, but in the angles between the large infra- and supramarginal plates there are here and there inserted tiny accessory pieces. In a young individual there are none of these accessory plates. Madreporite unknown. Inframarginal plates distinctly granulated, completely inclosing all other abactinal plates and common to both the abactinal and actinal areas. In mature specimens there are from 15 to 20 plates on each side of a ray. Laterally each plate has a Well-marked small pit, probably the place of attachment for rather strong spines, none of which however have been observed. Interbrachial areas with single, hexagonal, axillary interbrachial plates. Adambulacral plates slightly elongate-subquadrangular, of which there are from 18 to 20 on each side of a ray; distally they appear to pass over and to cover completely on the actinal side the inframar- ginal plates. Between the adambulacral columns there is a more or less wide ambulacral furrow. Each adambulacral plate bears on its inner edge a rather thick but short, slightly striate spine and several minute ones. Or ad to each axillary interbrachial there is a pair of prominent elongate-triangular adambulacral plates bearing spines, REVISION OF PALEOZOIC STELLEROIDEA. 79 one of the five pairs constituting a part of the oral armature. In other words, the adambulacral columns of adjoining rays are continued around the axillary interbrachials by two modified plates. Ambulacral grooves deep, nearly closed by the adambulacralia. Ambulacral plates not well shown but apparently carinate, one to each adambulacral ossicle, and whether alternate or opposite can not be stated. Locality and formation. -In the middle and .upper stages of the Cincinnatic group (Ordovicic). The type-specimen (No. 1195, Amer. Mus. Nat. Hist.) was found by Mr. D. H. Shaffer at Cincin- nati, Ohio, probably in the Maysville formation. Mr. Vaupel found a young individual showing the abactinal area (No. 60605, U.S.N.M.); another young one was secured by Mr. Faber (No. 9568, University of Chicago), and Mr. George Oeh found an excellent matured individual preserving the actinal area (now in the Yale Museum), all from the Maysville formation on the hills back of Cincinnati. Two good speci- mens and another small poor one are in the Harris collection of the United States National Museum (No. 59391) and were found in the Waynesville division of the Richmond formation near Waynesville, Ohio. A small asterid preserving two rays and a portion of the disk "is in the United States National Museum collection (No. 23540) and was found by Mr. W. P. Rust in the Utica shales one and a half miles east of Rome, Oneida County, New York. It is associated with Trinucleus concentricus, Dalmanella testudinaria multisecta, and other species. The abactinal area only is shown, and appears to be that of Meso- palseaster shafferi. It does not appear to be a Promopalseaster because of the fewer columns of abactinal accessory plates, and the axillary structure makes it almost certain that there is here but a single axil- lary interbrachial plate. For the present it may be provisionally referred to this species and attention is thus directed to it so that local collectors may be on the lookout for other material. Remarks. This species has actinally much the appearance of Hudsonaster incomptus, but the latter is at once distinguished by not having the five axillary interbrachial plates of Mesopal&aster shafferi. M. (?) parviusculus (Billings) is smaller, with far less plates in all the columns than in M. shafferi. M. (f) parviusculus and M. granti differ also abactinally and ambitally in the better developed accessory plates. Cat. Nos. 60605, 59391, 23540, U.S.N.M. MESOPAL^ASTER INTERMEDIUS, new species. Plate 9, fig. 4. This species is perplexing in its generic construction because of its intermediate or transitional characters on the one hand (but least) to Hudsonaster, and on the other (more decidedly) to Mesopalxaster 50601 Bull. 815 S 80 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. and prophetic of Promopalseaster in the conspicuous ambulacral furrows. Rays long and slender, elongate-petaloid, disk small. R = about 14 mm., r=4 mm. The interbrachial areas each have but a single, large, finely granu- lated, axillary plate that margins the axils, and in this agree best with Hudsonaster, but the rest of the known skeleton is that of Meso- palaeaster. In the former genus the ambulacral furrows are always very narrow and the ambulacralia are rarely seen, while in M. intermedius the furrows are wide and the ossicles are rather like those of Promopalseaster, though the same kind of ambulacralia is also known in Mesopalseaster. Orad to the axillary plates there were two small oral armature pieces that belong to the ad ambulacral columns. These are not now present in the specimen, but the distinct and large facet of the proximal side of the axillaries leaves no doubt on this construction. Inframarginal columns well-developed proximally, though never very prominent, and becoming rapidly smaller distally. The four proximal pieces are the largest, beyond which there are at least 14 other ossicles, and all are closely in contact and alternate with the adambulacrals. Another feature of these inframarginals is that not more than six can be seen from the actinal surface, while the rest pass on to the sides and distally even somewhat over to the abactinal area. All of the plates are finely granulated. Adambulacral columns the most prominent of the actinal skeleton, margining the distal half of the rays, and in the proximal portion pass more and more inside of the inframarginals, diminishing slowly in size and finally depressed beneath the axillaries. There are 22 plates in a column, all more coarsely granulated than the inframar- ginals, largest and most transverse near mid-length of the rays, and all are drawn out inwardly into blunt points that adjoin the high ridges of the ambulacralia. Therefore the adambulacral and the adjacent ambulacral columns have the ossicles directly opposite one another, but the adjacent ambulacral rows are slightly alternating. Ambulacral furrows conspicuous and wide, widest in the proximal third but converging somewhat quickly hi the mouth area and far more gradually distally. Therefore the furrows are elongate-petaloid in shape. There are about 20 ambulacralia in a column. Proximally the columns unite around and beneath the small adambulacral oral armature pieces. Ambulacralia square to transversely rectangular, each with a high L-shaped crest that laterally adjoins the points of the adambulacrals and medially bends orally. Podial openings large between the corners of pairs of ambulacrals and adambulacrals. There is a well-developed groove down the center of the ambulacral furrows. REVISION OF PALEOZOIC STELLEEOIDEA. 81 Abactinal area unknown, but probably that of Mesopalseaster and nearest to M. shafferi. Formation and locality. The holotype was found by Faber in the Maysville formation at Cincinnati, Ohio, about 350 feet above the Ohio River. The type is in the University of Chicago Museum (No. 9575). Remarks. This interesting species need be compared only with M. shafferi. It differs in having the axillaries in the axils of the rays and not inside the basal inframarginals as in the latter form. Then in M. shafferi the inframarginals are actinal in position throughout, while in M. intermedius they are more on the sides. Further, in the new form the ambulacral furrows are well developed with large ambulacralia, while in M. shafferi the furrows are narrow and the ossicles rarely seen. MESOPAL^SASTER FINEI (Ulrich). Plate 7, fig. 5; plate 9, fig. 5. Palseaster finei ULRICH, Journ. Cincinnati Soc. Nat. Hist., vol. 2, 1879, p. 19, pi. 7, figs. 15-156. Original description. "Small; rays five, of medium length, rather broad, pointed and narrower where they are attached to the much contracted body [probably due to distortion], than they are about the center of their length. " Dorsal side of rays composed of four [about seven, the supra- and inframarginals, radials, and two intercalary columns] rows of pieces, that are quite close [?] fitting, as wide as long, from 12 to 14 in each row, and increase in size inward to the disk, which is composed of irregularly shaped and prominent pieces some of which are smaller and others larger than those composing the rays; the pieces in the marginal rows [four columns infra- and supramarginals] are more prominent than the two [there are three, radials and two intercalary] rows between them, and have a small pit in the center, probably for the articulation of a spine [all of the plates originally bore several short slender spines]. Madreporiform body rather small, circular, very prominent, and marked by strong striae, which become more numerous toward the margin by intercalation. "Marginal [inframarginal] pieces on the ventral surface, convex, quite as long as wide, and numbering in different specimens on each side from 11 to 12 [probably not more than 8 or 9]; the piece at the junction of the rays is three times as large as any other of the series, subcircular and very convex. [It appears that this large plate is an axillary interbrachial since upon it proximally rest two basal plates of the inframarginal series.] 82 BULLETIN" 88, UNITED STATES NATIONAL MUSEUM. " Adambulacral plates more prominent [less prominent than the inframarginal plates], slightly wider than long, and numbering, on each side, from 9 to 10 [probably not less than 16 in each column]. Ambulacral pieces a little wider than long, not alternating with the adambulacral plates, and each provided with a rather sharp ridge across most of its width. "There are 10 oral plates [oral armature] formed by the junction of the adambulacral rows, which in form and size are scarcely dis- tinguishable from the other plates of those series. " Greatest breadth measuring between the opposite extremities of the rays ; 0.7 inch; breadth of rays at their inner ends, 0.1 inch; length of same, 0.3 inch; diameter of madreporiform piece, 0.02 inch." Formation and locality. This small species appears to be common at a very restricted horizon near the base of the Eden formation (Fulton beds), exposed at low water mark in the Ohio Biver in the eastern portion of Cincinnati, Ohio. Remarks. Not one of the twenty examples seen is well preserved and all the plates are more or less separated. The general structure seems to be that of Mesopalseaster. The composition of the axillary area is not positively ascertainable but one specimen shows what appears to be a large axillary interbrachial plate (see pi. 7, fig. 5), distally surmounted by two somewhat smaller inframarginals. This is probably the true structure since it is so in all the smaller Meso- palseasters, as M. lanceolatus, M. parviusculus, M. granti, and M. proavitus. The differences between M. finei and M. proavitus are not great. The former is a smaller species with less plates in all of the ranges, more prominent axillary inframarginal and axillary interbrachial plates, and is also found at a lower geological horizon. M. finei differs from the small species M. parviusculus of the earliest Siluric in having more plates in all of the columns. M. granti is a larger species and with smaller and more numerous plates than in M. finei. M. lanceolatus is probably the most closely related to M. Jinei but is distinguished abactinally by the diagonal arrangement of the plates on each side of the radial columns. Cat. No. 60604, U.S.N.M. MESOPALJEASTER(P) LANCEOLATUS, new species. Plate 4, fig. 3. Measurements: R = 4.5 mm., r=1.3 mm., R = 3r ) . Rays short, stout, distinctly lanceolate actinally and rapidly ta- pering abactuially. Disk comparatively large, abactinally convex. Interbrachial arcs distinct but small. Abactuially the disk has a ring of large, strongly stellate plates which are the basal plates of the radial and supramarginal columns. REVISION OF PALEOZOIC STELLEEOIDEA. 83 Inside the ring the plates can not be made out, due to the small size of the specimen. Rays each with a column of radials and two supra- marginals between which are inserted columns of diagonally arranged accessory plates. A few ambital accessory plates are probably also present. The arrangement of the ossicles on the abactinal areas of the rays is like that hi Promopalxaster magnificus. Madreporite small, subconical, situated near the edge of the disk, depressed between adjoining basal plates and marked by 10 or 11 sharp, somewhat spirally arranged ridges. Inframarginal plates large, about six in each column proximal to the very large single axillary interbrachial plate. These columns in the distal half can not be seen actinally as they pass over on the dorsal side. Axillary interbrachial plates large, one in each axil, and not com- pletely inclosed outwardly by the basal inframarginals. In this character the form is still in the Hudsonaster stage. Adambulacral plates small, subquadrangular, with 16 plates in each column. The plates of the adambulacral oral armature are larger and wedge-shaped, and are situated directly inside of the axillary interbrachial plates. Ambulacra! plates one to each adambulacral, apparently opposite and with proximally converging ridges. Podial openings apparently along the outer edge between the sutures of adjoining plates. Formation and locality. Two specimens were found by the late Prof. Charles E. Beecher near Rome, New York, in the Triarthrus bed of the Utica formation that preserves the entire ventral anatomy of these trilobites. The co types are in the Museum of Yale University. Mr. W. S. Valiant in a letter states that they are common in a layer 2 to 3 inches above the Triarthrus bed. Remarks. This species is most closely related to M. finei, from which it is readily distinguished abactinally. Here there are more and smaller plates than In M. lanceolatus, the interpolated plates are arranged diagonally and not radially as in M. finei, while the basal plates of the former are apparently much larger and are strongly stel- late. Actinally there are in M. lanceolatus more adambulacral plates in each column, although the specimens are only half the size of M. finei. MESOPAL^ASTER PROAVITITS, new species. R=18 mm., r^about 4.5 mm. The specimens are distorted, however, and no exact measurements can be given. On the distal abactinal portion of the rays the infra- and supra- marginal and radial plates closely adjoin. Proximally, however, between the two marginal series there is inserted a column of inter- calary plates gradually increasing in size and two other similar rows of plates are placed between the supramarginal and radial columns. 84 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. There are therefore at the base of the rays no less than 11 columns of plates. The radial columns proximally have a tendency to become less distinct and indistinguishable from the accessory pieces. The plates in general are highly convex, granular and abundantly spine- bearing. There are not less than 20 plates in each supramarginal series. Madreporite highly convex, broadly oval in outline and radially striated. Inframarginal plates finely granular, highly convex, subquad- rangular in outline and increasing in size slowly proximally. About 23 in each column. Adambulacral plates like the inframarginals but not increasing much in size proximally, with about 29 in each column. As in other species of this genus, each plate bears three prominent spines, two laterally and one ambulacrally. Ambulacral plates one to each adambulacral plate, and with a sharp ridge which medially bends abruptly orally. Podial openings as in other forms of Mesopalseaster. Interbrachial areas occupied by single axillary interbrachial plates upon each of which rest proximally two axillary inframarginals. The specimens are not figured, as an illustration can not readily be made. Formation and locality. Three more or less entire specimens and fragments of four rays were found by Dr. E. O. Ulrich in the Eden shale exposed back of Covington, Kentucky, at an horizon about 100 feet above low water in the Ohio River. All the specimens a^e now in the Herzer collection, a part of the late Prof. James HalFs private collection, recently purchased by the University of Chicago. Remarks. M. proavitus and M. granti are closely related, since both have 11 columns of abactinal plates, including the marginals, and they are nearly alike actinally. However, the former is twice as large and appears to have more conspicuous radial and supramarginal columns than the latter. These differences and the further fact that one occurs in the Utica and the other in the earliest Siluric will serve to distinguish the two species for the present. M. proavitus may prove to bo identical with M. (?) wilberanus, which see for further remarks. MESOPAL^ASTER (?) WILBERANUS (Meek and Worthen). Petraster wilberianus MEEK and WORTHEN, Proc. Acad. Nat. Sci., Philadelphia, for 1861, vol. 13, 1862, p. 142. Palseaster wilberanus HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 285; rev. ed., 1868=1870, p. 328. Original description. "This beautiful starfish resembles rather closely Petraster rigidus of Billings, 1 but is smaller, and has more 1 Decade 3, Org. Rem. Canada, pi. 9, fig. 3o. REVISION OF PALEOZOIC STELLEROIDEA. 85 slender rays, with more angular spaces between. It also differs in having but two ranges of plates on each side of the ambulacral grooves, on the under side, instead of three. These pieces are about of the same size in each row, and sometimes appear to alternate; they are all rather prominent, and those of the outer range project a little laterally in the form of small nodes. Toward the outer extremities of the rays, however, the lateral ranges are contracted behind the others, so as to be scarcely visible from below. About 23 of these pieces may be counted on each side of the ambulacral furrow in each ray. The ambulacral furrows are very narrow, and indeed seem to be closed toward the extremities of the rays, by the gradual approximation of the inner rows of pieces on each side, which alter- nate and appear to fit together. "Our specimen only shows the under side, but along the outer margins of two of the rays, there is some appearance either of the overlapping of some of the dorsal parts by pressure, or of a slightly developed disk. This part does not seem to have any distinct range of marginal plates, but appears to be made up of small pieces, covered with granules, or bases of small spines. "We take pleasure in dedicating this interesting species to Prof. C. D. Wilber, of the Illinois State Normal School, to whom we are indebted for the use of the only specimen we have seen. "Greatest diameter about 1 inch; smaller diameter 0.33 inch. "Locality and position. Oswego, Kendall County, Illinois, in rocks of the age of the Trenton or Hudson River Group of the New York series." Probably equivalent to the Richmond beds of the highest Ordovicic. The whereabouts of the type-specimen is not known. Remarks. Hall pointed out that this species is not a Petraster because P. wilberanus has but two columns of plates on each side of the ambulacral groove, while Petraster has in addition "a few disk plates on the ventral side." He therefore referred it to Pal&aster. However, Meek and Worthen describe the abactinal area as having no " distinct range of marginal plates, but appears to be made up of small pieces, covered with granules, or bases of small spines." The characters as far as mentioned appear to be those of Mesopal&aster. Should there prove to be in P. wilberanus single axillary interbrachial plates, then all of the characters will be in harmony with Mesopalse- aster and M. proavitus may prove to be a synonym of it. This can not be proven at present since we have not been able to locate the type-specimen. MESOPAL^ASTER (?) DUBITJS (Miller and Dyer). Palseaster dubius MILLER and DYER, Cont. to Pal., No. 2, 1878, p. 5, pi. 4, fig. 8. Original description. "Pentagonal; rays longer than the diameter of the body, and uniformly tapering. The ambulacral groove is sharply angular in the middle, formed by two series of plates having 86 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. their ends together in such manner as to make an angular gutter [the plates of adjoining columns are arranged practically opposite]. The length of these plates is three times as great as the diameter. The plates are placed with the length across the rays. There are 15 plates on each side of the groove in each ray, in the length of one quarter of an inch [this is an error as there are about 20 plates in a column]. The diameter of the body is three-tenths of an inch. "This species is founded upon a single specimen, in Mr. Dyer's collection, which shows only part of the ventral side. The ends of the rays [most of the adambulacrals] and marginal plates arc destroyed [each axillary area appears to bear one axillary interbrachial plate]. The parts preserved seem to distinguish it from any species hitherto described." Formation and locality. In the original description the locality is given as Cincinnati, Ohio, but the character of the rock shows that it is from the uppermost portion of the Trenton limestone, probably opposite Cincinnati, in the river quarries at Ludlow, Kentucky. The type (No. 25) is in the Museum of Comparative Zoology, Harvard University. Remarks. This specimen preserves well the very wide ambulacrals but almost no other plates. The adambulacrals are nearly all removed and but very little is to be seen of the marginals. In three of the axillary areas there are single interbrachial plates, one of the essential features of Mesopalseaster. It is referred to this genus provisionally, as nothing is known of the abactinal side. MESOPAL^ASTER (?) ANTIQUTJS (Troost). Asterias antiqua TROOST (not Hisinger 1837), Trans. Geol. Soc. Penn., vol. 1, 1835, p. 232, pi. 10, fig. 9; Proc. Amer. Assoc. Adv. Sci., vol. 2, 1850, p. 59 (cat. name). Petraster (?) antiqua SHUMARD, Trans. St. Louis Acad. Sci., vol. 2, 1866, p. 386 (catalogue name). Palxaster (Argaster) antiqua HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 287; rev. ed., 1868=1870, p. 329. Palxaster antiquus MILLER, N. Amer. Geol. Pal., 1889, p. 265 (cat. name). Argaster antiqua GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 345 (gen. ref.). Pal&aster antiqua WOOD, Bull. U. S. Nat. Mus., No. 64, 1909, p. 105, pi. 8, fig. 1. Although this is the first recorded American fossil starfish, very little is known about it and that little is mostly of a misleading nature. The specimen lies on a limestone slab and is very badly weather- worn or it may have been treated with hydrochloric acid so that now it is nothing more than a polished section of a starfish. An illustra- tion that will show its actual characters can not be made. Professor Hall errs in stating that Asterias antiqua has "ambu- lacral grooves occupied by a single row of subquadrate ossicula, which extend across and alternate with the adambulacral plates of each mar- gin. * * * It is possible that this character may prove to be of generic importance." It may have been this character on which REVISION" OF PALEOZOIC STELLEROIDEA. 87 Hall thought of basing the genus Argaster, but he does not character- ize it nor even mention that the name in parenthesis is intended as a new term with Asterias antiqua Troost as the genotype. This species, like all other Paleozoic starfishes, has double columns of ambulacral plates. This the specimen clearly demonstrates on the edge of the slab where the distal parts of the rays are broken away. Argaster should therefore be regarded as a nomen nudum, and should A. antiqua prove to be a Mesopal&aster, it should not be made to displace this genus. Asterias antiqua has about 15 inframarginal plates in each column and about 32 in each adambulacral column. Two of the latter plates meet as usual in a pair of triangular oral armature pieces. Each axil is occupied by two large, quadrangular, basa] inframar- marginal plates. Between these proximally there is a large, widely triangular, interbrachial plate the apex of which may or may not attain the margin. Proximal to each axillary interbrachial plate and between the four or five pairs of axillary adambulacral plates, there is in the specimen an open space in each of the five areas. What additional plates, if any, occupied this area is not de terminable. It may be that the axillary interbrachial plates occupied the entire interbrachial areas and that the present hiatus is due to the worn condition of the specimen. This appears to be the most natural interpretation as it is the normal inter brachial structure of Mesopa- Iseaster. In Promopalseaster there are always two, three, five, or seven interbrachial marginal plates hi each area, a fact which excludes Aste- rias antiqua from that genus. The abactinal area is not visible, but many of these plates are squeezed beyond the inframarginals, showing the presence of num- erous small plates recalling Mesopalseaster and Promopalseaster. Formation and locality. Troost's label reads: " Lower limestone on Harpeth Kiver, Davidson County, Tennessee." His manuscript reads: "It was found * * * on Harpeth Kiver, Davidson County, Tennessee. Associated with Spirifer lynx [PlatystropJiia l}iforata] } Cyathop7i[y]Ua [ Streptelasma] } Orthis [testudinaria], etc." This is apparently the same horizon as that about the city reservoir in Nashville, which is now regarded as of Upper Trenton (Catheys) age. The specimen is in the United States National Museum. Cat. No. 39914, U.S.N.M. MESOPAL^EASTER(?) PARVIUSCULUS (Billings). Plate 9, fig. 1. Palseaster parviusculus BILLINGS, Canadian Nat. and Geol., vol. 5, 1860, p. 69, figure. DAWSON, Acadian Geology, 2d ed., 1868, p. 594, fig. 197. HUDSON, Bull. N. Y. State Mus., No. 164, 1913, pis. 1-4. Original description. ' l The specimen is about six lines in diameter. The rays are two lines in length and one and a half in width at the 88 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. base, tapering at an angle of a little less than 45. The five oral plates are subpentagonal, about half a line in width. The first adambulacral plates of each pair of adjacent rays are in contact with each other outside of the oral plates, and not completely sep- arated as they are in P. niagarensis. There are six or seven adam- bulacral plates on each side of the ambulacral groove in each ray, and they gradually decrease in size from the oral plate outward to the point of the ray. The width of the ambulacral groove is equal to one-third the width of the ray and consequently the adambu- lacral rows of plates are also each equal to one-third the whole width of the ray. In each groove there are two rows of small and appar- ently nearly square ambulacral plates, 12 or 14 in each row, and they seem to be continued round on the inner margin of the oral plates; the mouth is about one line wide." Emended description. Measurements: R = 6 mm., r = 2.5 mm., R = 2.4r. Width of ray at base 3 mm. Rays short, stout, tapering rapidly. Disk comparatively large, and without interbrachial arcs. Abactinal area unknown. Inframarginal plates at the base of the rays very large and thick, highly convex, duninishing rapidly distally, usually quadrangular in form and seven on each side of a ray. Just within each axil, inside of the basal inframarginals, there is a single, large, very convex, pentagonal, axillary interbrachial plate. Adambulacral plates gradually sinking below the inframarginals proximally, convex and quadrangular in form. There are 13 plates in each column adjoining the very narrow ambulacral furrow, or nearly two plates to each inframarginal ossicle. Two of these plates (oral armature) are situated in front of each axillary inter- brachial, uniting the adambulacral columns of adjoining rays. Ambulacral plates unknown. Locality and formation. The type-specimen, a natural mold, was found by Rev. D. Honeyman in the Lower Arisaig of the Siluric ( = Clinton) rocks at Arisaig, Nova Scotia. The holotype is in Redpath Museum of McGill University, Montreal, Canada, and the illustration is taken from a gutta-percha squeeze kindly made for the United States National Museum by Sir William Dawson. Remarks. The general actinal appearance of this little asterid reminds one much of Palseaster and Hudsonaster. It differs, however, at once in a little detail of considerable importance, in that each axillary area has three plates, while in Palseaster and Hudsonaster there is but one. In other words, the large axillary plate of those genera has in M. (?) parviusculus been pushed orally and is here an interbrachial plate, while the axils of the rays are occupied by two basal inframarginal plates. For further remarks on this species see M. granti (Spencer). REVISION OF PALEOZOIC STELLEROIDEA. 89 So long as the abactinal area of M. (?) parviusculus remains unknown its generic position will be doubtful. For the present its relations appear to be nearest to Mesopalseaster and to M. (?) cata- ractensis. Cat. No. 60620, U.S.N.M. MESOPAL.EASTER (?) CATARACTENSIS, new species. Plate 9, fig. 2. The specimen of this species has been confused with M. (?) gmnti, from which it differs in being smaller and in having far fewer plates in the columns. The actinal side of M. (f) cataractensis alone is known, while of M. (?) granti only the abactinal, but even so the above comparison can still be made because the size of plates is relatively about the same in species of Mesopalseaster. Measurements: K = 9 mm., r=3 mm., R = 3r. The species is most closely related to M. (f) parviusculus, but differs in being larger with more plates in the columns. Of M. (?) cataractensis only the actinal side is known, with 12 to 13 plates in each inframarginal column (7 in M. (?) parviusculus) and about 19 in the adambulacrals (13 in M. (?) parviusculus). Then the inframarginals increase very rapidly in size proximally, while in M. (?) cataractensis there is but little enlargement. A single small, pentagonal, axillary interbrachial plate occurs in each area immediately beneath the basal inframarginals. These are much smaller relatively than those in M, (?) parviusculus. Ambulacral plates unknown. Ambulacral furrows deep, nearly completely closed by the adambulacral columns. Formation and locality. in the base of the Siluric (Cataract for- mation) at Hamilton, Ontario. The holotype was found by Mr. Elliott near the city reservoir and is in the collection of the Hamilton Natural History Society. It was loaned to the writer by the late Col. Charles Coote Grant. MESOPALJEASTER GRANTI (Spencer). Plate 9, fig. 3. Palseaster granti SPENCER, Bull. Mus. Univ. Missouri, No. 1, 1884, p. 53, pi. 7, fig. 1. Original description. "Body stellate and small with short arms, about 2 centimeters across; disk less than 1 centimeter wide, and apparently formed by the junction of the rays; rays 5 millimeters wide at base, tapering slowly, and terminating in rounded extrem- ities, at about 8 millimeters from junction of their base with the disk; upper surface of rays composed of 5 [not less than 9 in mature specimens at mid-length of rays] ranges of highly convex or tuberculi- form plates (the marginal [supramarginal] and ventral [inframarginal] 90 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. rows being the most conspicuous) and separated from each other by minute plates (becoming fewer on approaching the extremities of the rays) . "The ellipsoid [supra] marginal tuberculiform plates number about 12 [20 in mature examples] for each complete side of the rays, and the central range [radial] is composed of a similar number, but in form these plates are more circular. The disk between the ter- minal central [radial] row of plates of the rays is crushed and struct- ureless, except a slight elevation in the center. The terminal [proxi- mal] plates of the marginal series are larger than the others of the range. The madreporiform tubercle is of a spherical form and rela- tively large, being nearly 2 millimeters in diameter, and is situated at the axil of two rays. Both the tuberculiform plates and the madreporiform tubercle have a granulated surface. "The ventral side is unknown. Emended description. A mature specimen measures: R=16 mm., r=5 mm. ; R = 3.3r. Rays short, stout, tapering rapidly, and with very small inter- brachial arcs. Abactinal area of disk unknown. Along the center of each ray there is a radial column of small, highly convex plates bounded on each side, at about the mid-length of the rays, by two columns of smaller, strongly tumid, accessory plates. The radial column is, however, not distinctly differentiated from those adjoining. Out- side of these are the supramarginal columns, each with about 20 plates which are larger and more easily distinguished than the radial plates. Dis tally the rays have only the ossicles of the infra- and supramarginal and radial columns, with none of the accessory plates as yet developed. Ambital area with one column of very small accessory plates like the marginal plates on each side of them. These accessory pieces disappear before attaining the apex of the rays. Madreporite spherical in form and relatively large, nearly 2 milli- meters in diameter, and situated at the .axil of two rays between two adjoining supramarginal columns. Formation and locality. From the Cataract formation, the basal deposits of the Siluric system, near the reservoir in the city of Hamilton, Ontario. Col. Charles C. Grant, after whom the spe- cies is named, discovered most of the specimens, which are six in number, two in the Spencer collection, now destroyed by fire, one in the Redpath Museum of McGill University, Montreal, another in the Geological Survey of Canada at Ottawa, and two in the Grant collection, one of which has generously been donated to Yale University. Remarks. This species is apparently closely related to M. bel- lulus, which see for comparison. REVISION OF PALEOZOIC STELLEEOIDEA. 91 MESOPAL^EASTER BELLULUS (Billings). Text fig. 7. Petraster bellulus BILLINGS, Geol. Surv. Canada, Pal. Foss., vol. 1, 1865, p. 393, fig. 368. MILLER, N. Amer. Geol. Pal., 1889, p. 269, fig. 389. Original description. "Deeply stellate, about 18 lines across; disk 5 lines wide; width of rays, at the base, half the width of the disk, uniformly tapering to their extremities; ambulacral grooves, narrow and deep, with about 30 adambulacral plates on each side. These plates are strongly convex and nearly square [distally, but proximally are wider than long]. Outside of these there is a row of [about 21 infra] marginal plates, which appear to [actually do] extend to the extremities of the rays, but on this point there is some doubt, as the specimen is not perfect. [The four proximal pieces increase rapidly in size so that they are considerably larger than the others. All of the actinal plates are granular.] There appear to be one or two [there is but one axillary] small disk plates between the [basal adjoining infra] marginal and ad- ambulacral plates just outside of the oral angles. "Locality and formation. Township of Grimsby [Ontario]; in the Niagara shale [ = Rochester shale]. 1 i Collector. J ohnson Pettlt, Esq. , Fi G. ?. THE ORIGINAL FIGURE OF ME- Grimsby." The holotype is in the Vic- toria Memorial Museum, Ottawa, Canada (No. 2665). Remarks. The holotype and only known specimen of this species was seen by the writer at Ottawa, and has now been freed from the rock so that the abac tin al side can also be studied. The generic charac- ters are those of Mesopalseaster and one is impressed at once by the almost specific identity of the abactinal side of the specimen with that of M. granti. However, as the specimens show slight differ- ences and come from different Siluric formations, it is thought best to recognize them as distinct species. Moreover, the actinal side of M. granti is still unknown. Abactinally all of the skeletal pieces are of about the same size, small, each ossicle irregular in outline, and more or less stellate. There are about 17 in the length of one-half inch. This skeleton is more or less disturbed and it is impossible to clean off all the adhering shale, so that it is difficult to make out the detailed arrangement. The following, however, can be made out: There is no true disk as in Petraster , and the parts that appear as such are either adhering shale or displaced ossicles due to distortion during sedimentation, SOPAL.EASTER BELLULUS, AFTER BlL- LINGS. NATURAL SIZE. THE SPECI- MEN HAS STILL TOO MUCH SHALE ADHERING TO IT AND BESIDES IS CON- SIDERABLY DISTORTED, PREVENTING ITS BEING SATISFACTORILY PHOTO- GRAPHED. 92 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Distally the rays have columns of small radial pieces, and outside of these are the two supramarginal rows. Proximally between these primary columns there appear accessory pieces, one column of which wedge in between the radials and supramarginals and from one to three between the latter and the inframarginals. In this way finally as many as 13 columns can be counted at the base of a ray. The madreporite is situated near the margin between adjoining supramarginal columns, and is conspicuously protruded and large when compared with the small adjacent ossicles. It is a large, radially striate, irregularly circular plate about six times larger than the adjoining pieces. MESOPALJ2ASTER CARACTACI (Gregory). Plate 9, fig. 6; plate 11, fig. 1. Palseaster caractaci SALTER, Cat. Foss. Mus. Practical Geology, 1865, p. 30 (nomen nudum). NICHOLSON and ETHERIDGE, Mon. Silurian Foss. Girvan Disk, Ayrshire, fasc. 3, 1880, p. 321 (no description). GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 344. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 227. Protopalxaster caractaci SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palseontogr. Soc. for 1913), 1914, pp. 21, 30, pi. 1, fig. 5. Remarks. The writer was greatly pleased to receive from Dr. F. A. Bather of the British Museum wax squeezes of the type-specimens^ because they clearly show that the genus Mesopalseaster is present in Europe. M. caractaci has its closest relationship in M. sJiafferi, the genotype of Mesopald&aster. The differences are easily to be seen on the abactinal side, as illustrated in this work. Rays long and slender, abactinally flattened. R = 12.5 mm., r3.S mm., R, = 3.3r. Radialia small and no larger than the pieces of the accessory columns on either side of them ; the former alternate in position with the pieces of the latter and all of them are more or less spicular in outline. Basal radials slightly larger than the others. The ossicles of the supramarginal columns large and solid, the most prominent of the abactinal side, very large proximally and exceedingly small distally; there are about 14 in a column. Basal inframarginals most conspicuous of all abactinal plates, with each pair occupying the axils, and together with the smaller basal radialia form a ring bounding the disk. The disk has fallen in but still shows a few small pointed plates. " Madreporite is large, and placed close by one of the interradial angles" (Gregory). Inframarginal ossicles almost smooth, large and prominent proxi- mally, decreasing rapidly in size so that not more than 9 can be seen actinally, where the tiny pieces pass somewhat over to the abactinal surface, or better the lateral sides; Gregory states that there are 12 in a column. The basal pieces of adjacent columns do not as a rule touch one another in the axils. BEVISION OF PALEOZOIC STELLEROIDEA. 93 Interbrachial areas with single, large, hexagonal, axillary inter- brachial plates. Adambulacral pieces quadrangular, fairly uniform in size for two- thirds the length of the rays and then become gradually smaller distally; there are about 15 in a column. A peculiarity , of this species is that the ambulacralia do not continue around the axillary interbrachials, but cease with the basal inframarginals. There are, therefore, in M. caractaci no adambulacral oral armature pieces. Ambulacral furrows moderately wide, with as many rectangular ambulacralia as there are adambulacrals, plus probably four more in each column, and these continue around the axillary interbrachial. Therefore, the oral armature consists entirely of ambulacralia. All of these ossicles are directly opposite one another and have high medial crests that are continuous across the ambulacral furrow. Podial openings large laterally between the plates and the adambu- lacrals. Locality and formation. Caradoc sandstone, Soudley quarry, Church Stretton, England. The holotype, a split nodule, is in the British Museum (Natural History), No. 48206; wax squeezes from which the photographs have been made are in the United States National Museum. Other specimens, Bather states, are in the Museum of Practical Geology, London; these are from Marshbrook and occur at a somewhat higher geological level. Remarks. Gregory wrote in 1899 that "the nearest ally of this species is P. matutina (Hall)," but the illustrations here presented wiU show that the relationship is with Mesopaldeaster shafferi and not with Hudsonaster matutinus. Cat. No. 60606, U.S.N.M. MESOPAL^EASTER (?) ACUMINATUS (Simonovltsch). Asterias acuminatus SIMONOVITSCH, Sitzb. d. mat.-naturw. Classe Akad. Wiss., Wien, vol. 64, Abt. 1, 1871, p. 100, pi. 3, figs. 2-2d. Asterias acwninata SCHONDORF, Palaeontographica, vol. 56, 1909, pp. 72, 109, pi. 11, figs. 7, 8. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 32, pi. 5, figs. 4, 5. Roemeraster (?) acuminattis STURTZ, N. Jahrb. fiir Min., etc., 1886, vol. 2, p. 154. Actinally this small and slender species is nearly in harmony with the generic characters of Paldeaster and Hudsonaster, differing only in that the axillary interbrachial plates barely extend to the margin of the disk, while in those genera they are prominent and marginal in the axils. In Asterias acuminatus they are as yet not crowded so far orally as in most species of Mesopalseaster. It is a difference of degree, not of kind; therefore this species has more the actinal character of Mesopalseaster than Hudsonaster. Ambulacral plates unknown. Abactinally Simonovitsch's species is said to have six columns of tumid plates, all alike and persisting throughout the rays. If the 94 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. outermost columns of each ray are regarded as the inframarginals and the ones immediately inside as supramarginals, two medial columns remain to be homologized with those in Mesopalseaster. If Simonovitsch is correct, then by this interpretation there are two columns between the supramarginals. Both can not be radials and to regard one as such will leave a unilateral and apparently an unnatural development. If five or seven columns of abactinal plates were present, this species would be hi harmony with Mesopalxaster. Since Simono- vitsch's work is of the best, we have at present the only alternative of supposing that the radials in Asterias acuminatus are suppressed and that its two medial columns are homologous with the intercalary abactinal plates of other starfishes. Among Paleozoic asterids this development is rare and is found only in Encrinaster and Palseaster. In that event, A. acuminatus is not a Mesopalseaster. A single example of this species was found in the Lower Devonic (probably Upper Coblenzian) near Braubach, Germany. In his revision of the Lower Devonic starfishes of Germany, Schondorf was not able to discover the whereabouts of the type-speci- men and therefore had to leave this species as determined by Simono- vitsch. The former states that actinally the structure reminds one of Spaniaster latiscutatus, the genotype of Spaniaster, and closely related to Mesopalseaster. For further remarks see Spaniaster. MESOPAL^ASTER (?) CLARKI (Clarke and Swartz). Plate 10, figs. 1, 2. Pals&aster clarki CLARKE and SWARTZ, Maryland Geol. Surv., Upper Devonian, 1913, p. 543, pi. 46, figs. 3,4. Original description. "This species is represented in the collec- tions by a single specimen affording a pretty sharp cast of both sides of a very regular and complete individual. In general structure and appearance the species is quite similar, especially on the oral surface, to Palseaster eucharis Hall of the sandy Hamilton shales of central New York, but the latter is a much larger form. "The disk is small, rays long and slender, thecal plates all promi- nently developed. The ambulacral surfaces are represented only by a narrow linear depression beneath which the ambulacral plates are concealed. These depressions are bordered by thickened and somewhat elevated quadrangular or pentangular ad ambulacra. The marginal plates are in single rows, much thickened, with generally quadrangular outline and convex surface, each projecting on the margin of the ray. At each axilla is a single pear-shaped plate with its apex outward, these plates being the largest in the individual. The abactinal surface is tessellated by rows of strong convex plates of similar size to the mar- ginal plates. Of these there are three rows, a median row of narrow oval ones between the ends of which are interlocked the edges of the much larger plates of the lateral rows which are highly convex and REVISION OF PALEOZOIC STELLEROIDEA. 95 FIG. 8. SPANIASTER LATISCUTATUS, AFTER SCHONDORF. SCHEMATIC ARRANGEMENT OF THE ABACTINAL PLATES. C, PRESUMABLE POSITION OF ANUS; Ce, CENTRAL PLATE; Jn, BASAL INTERRADIALIA; mdp, PROBABLE POSITION OF MADREPORITE; mo, SUPRAMAR- GINALIA; RI, BASAL RADIALIA. thickened in the center and greatly depressed to the sutures . Thus each plate has a cushioned surface. Between the ends of each plate of the middle row are two minute accessory plates lying in the angles at which the lateral plates enter. At the base of each ray and upon the disk is a single large plate whose surface rises into a high clavate node. Between each two of these is one of less height.. The central portion of the aboral area is destroyed and no trace of madrepore is seen. "The width of this specimen from tip to tip is 33 mm. "Occurrence: Jennings formation, Chemung member. Yellow sandstone on the road northeast of Oakland, Garrett County [Maryland], where it is associated with Spirifer disjunctus. "Collection: Maryland Geological Survey." Remarks. This clearly determined species is a late survival of early Paleozoic primitive asterids. The characters so far as determinable in the natural mold are those of Meso- palseaster, but as the disk skeleton is not preserved, it is very prob- able that when this feature is known the form will be seen to belong to a new genus. Genus SPANIASTER Sehondorf. Plate 12, figs. 1, 2; text fig. 8. Calaster SANDBERGER (not Agassiz 1835), Verst. d. rheinischen Schicht. Nassau, 1855, p. 381. Spaniaster SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 60, 1907, p. 176; vol. 62, 1909, p. 30; Palaeontographica, vol. 56, 1909, pp. 73, 109. Remarks. This genus with its single very small species has its nearest relations with Mesopalseaster, in that it has a single axillary interbrachial plate in each actinal axillary area, but differs from it in that there are many more adambulacral and ambulacral plates than there are inframarginals. The marked and generic difference, however, is on the abactinal side, where there are but three columns of plates, one radial and two suuramarginal, of large and thick ossicles arranged in parallel rows, the pieces of which do not alternate with one another; further, the supramarginals are almost completely superposed upon the inframarginals, the two columns together, but more particularly the inframarginals, bounding the rays and not the inframarginals alone, and not so pronouncedly as in Mesopalseaster. The disk is also more primitive in its construction, in that the central 50601 Bull. 8815 7 96 BULLETIN disk plate is encircled by a ring of but 5 small accessory disk pieces, separating the former from the next ring of 10 much larger plates that are the primary pieces of the radial and supramarginal columns. Genoliolotype and only species. Ccelaster latiscutatus Sandberger. Restricted to the Lower Devonic of Germany. SPANIASTER LATISCUTATUS (Sandberger). Plate 12, figs. 1, 2; text fig. 8. Ccelaster latiscutatus SANDBERGER, Verst. d. rheinischen Schicht. Nassau, 1855, p. 381, pi. 35, figs. 1, la. Xenaster simplex SIMONOVITSCH, Sitz. d. mat.-naturw. Classe Akad. Wiss., Wien, vol. 64, Abt. I, 1871, p. 97, pi. 3, figs. 1, la. Spaniaster latiscutatus SCHONDORF, Palseontographica, vol. 56, 1909, pp. 73-82, 109, pi. 8, fig. 2; pi. 10, figs. 2, 9, 10; pi. 11, fig. 9 (complete synonymy given here); Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 31, pi. 5, figs. 1-3, text fig. 2. The Sandberger specimen, showing only the abactinal side, was found in the Lower Devonic Spiriferensandstein at Unkel, near Bonn, Germany. Another specimen from the Lower Coblenzian quartzite has been recently found at Bienhorntale, near Coblenz; this shows both sides. Still another is from the Lower Coblenzian at Oberstadtfeld in the Eifel. The holotype of X. simplex is from the Upper Coblenzian at Niederlahnstein on the Rhine. Genus MIOMASTER Sehondorf. Plate 8, fig. 3. Miomaster drevermanni SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 38, pi. 3, fig. 4; pi. 4, fig. 3. Remarks. This genus is most closely related to Spaniaster, in that both have the single axillary interbrachial plates and the infra- marginals and supramarginals are directly superposed. They differ from one another mainly on the abactinal side, though this area is so poorly preserved in Miomaster that almost nothing definite can be said. Sehondorf illustrates a section through .the rays that shows the presence of single columns of very small accessory pieces between the radialia and supramarginalia. In the description, however, he states: "One can not positively determine whether the supra- marginal plates are separated from the median pieces by small accessory ossicles, or whether the three dorsal columns were closely adjoining." The central disk area is also too much disturbed to make out the skeletal arrangement. There appear to have been small interbrachial abactinal areas. Under these circumstances, the genus Miomaster can not be said to be well established, and its final dispo- sition is dependent on finding better preserved material. Genoholotype and only species. M. drevermanni Sehondorf (as cited above). A single example was found by Drevermann in the Upper Coblenzian at Miellen-on-the-Lahn, Germany. The holotype is in the Senckenberg Museum, Frankfort-on-the-Main. REVISION OF PALEOZOIC STELLEROIDEA. 97 DEVONASTER, new genus. Plates 11 and 12. Palseaster (part) of AUTHORS. Xenaster (?) or new genus GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 346. Devonaster SCHUCHERT, Fossilium Catalogus, Animalia, pt. 3, April, 1914, p. 14. A characteristic starfish of the American Devonic. Generic characters. Rays five, stout, tapering rapidly. Disk large, with small interbrachial arcs. Abactinal area of rays with prominent columns of radial supra- marginals and less prominent inframarginal plates, all of which bear numerous small pustules. The radial columns are completely separated from the supramarginals by a great abundance of very small, conical, irregularly arranged, accessory plates which in the proximal region may also force apart single plates of these columns. The supramarginal columns of adjoining rays nearly meet in the axillary areas a short distance inside of the margin, at which level the radial columns also cease. Inside of the area bounded by the proximal plates of the radial and supramarginal columns, the disk has numerous small accessory plates like those of the rays. In the center of the disk is a plate somewhat more prominent than those surrounding it, and just inside of the adjoining proximal supramar- ginals is a small but distinct plate interradial in position. Ambital accessory plates abundantly developed between the marginal ossicles in the axillary areas. These plates are like the other accessory plates and do not extend beyond the third or fourth proximal inframarginal pieces, where the ambital areas cease and the supramarginal ossicles come to lie more and more completely over the inframarginals. Madreporite flat or concave, situated between two proximal supramarginal ossicles and outside of one of the small interradially situated plates. Inframarginal pieces lajge and thick, increasing rapidly in width proximally, so that the tw.o wedge-shaped proximal plates are nearly three times as wide as long. The plates are pustulose, and the proxi- mal ones bear several slender spines. Adambulacral plates more numerous than the inframarginals, continuing around the axillary interbrachial plates, in front of which the columns meet in two triangular and enlarged adambu- lacral oral armature plates. Each piece bears two short, obtusely pointed spines. Ambulacral grooves narrow and deep, either straight or very slightly petaloid. Ambulacral plates equaling in number the adambulacrals and arranged in directly opposed, closely adjoining columns. Podial openings in two columns through the sutures in the lateral part of the furrows. 98 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Axillary interbrachial plates large, a single one situated in each area between the adambulacral and inframarginal columns. Each bears short, thick, blunt spines. Mouth plates (tori) minute, one situated in front of each pair of oral armature plates. Genoholotype. Palseaster eucharis Hall. Distribution. Restricted to the Hamilton and Chemung of the American Devonic. The species are: D. eucharis (Hall). Hamilton. D. chemungensis, new species. Chemung. Remarks. This genus is readily distinguished from Hudsonaster hi having well-developed radial and ambital accessory plates, in the numerous very small plates of the disk and in the presence of interbrachial axillary plates. Devonaster differs from Palseaster in having radial columns of plates and in the presence of three plates in each interbrachial area against one in Palseaster. There are also other differences. From Neopalseaster it differs in having ambital accessory plates and no ocular plates; while the proximal supramarginals and radials are very large in the former, they are small in Devonaster, particularly the supramarginals, which are not readily distinguished from the accessory disk plates. In Mesopalseaster and Promopalseaster the accessory plates are always arranged in columns or rows and the ambital accessory plates are not restricted to the axillary areas as hi Devonaster. Further, in those genera (excepting in the smaller forms of Mesopalzeaster) there are always from two to five inter- brachial marginal plates, while Devonaster has but one axillary marginal in each interbrachial area. Devonaster greatly resembles abactinally Xenaster of the Lower Devonic of Germany, but is readily distinguished actinally in that the latter has more large interbrachial plates, and especially in the ad- dition of numerous accessory interbrachial pieces. These differences are pointed out in more detail in the remarks on Xenaster. DEVONASTER EUCHARIS (Hall). Plate 12, figs. 3-5. Palseaster eucharis HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 287, pi. 9, figs. 3-3a (74); rev. ed., 1868=1870, p. 330, pi. 9, figs. 3-3a (74). QUENSTEDT, Petrefactenkunde Deutschlands, vol. 4, 1876, p. 71, pi. 92, fig. 29. ZITTEL, Handb. Pal., vol. 1, 1879, p. 452, fig. 323. COLE, Bull. Geol. Soc. America, vol. 3, 1892, p. 512, pi. 15. CLARKE, Bull. N. Y. State Mus., No. 158, 1912, pp. 44-45, 6 pis. Xenaster eucharis SCHONDORP, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 66, 1913, pp. 87-96, pi. 3, figs. 1, 2, text figs. 1-3. Original description. " Body rather large; the largest individual being one inch and seven-eighths from the center of the body to the extremities of the rays; the whole having a robust aspect; rays REVISION OF PALEOZOIC STELLEROIDEA. 99 acutely pointed at the extremity. Upper [abactinal] surface of rays composed of three [radial and two supramarginal] ranges of large, highly convex or tuberculiform plates which are nearly cir- cular at the bases of the rays, becoming quadrate and widened to- wards the extremities; separated from each other in the lower part by numerous minute [accessory] plates or granules, which become fewer near the middle of the ray, and before reaching the extremity. The central portion of the disk is occupied by an elevated pentagon, the angles of which are formed by the abrupt termination of the central row of plates of each ray: the whole composed of very mi- nute, highly convex plates, which vary in size, the larger ones pen- tagonally arranged. The angles between the rays have a few [nu- merous] small [accessory ambital] plates outside of the outer ranges of tuberculose [supramarginal] plates on the upper side, uniting with the inframarginal plates below [and not extending beyond the third or fourth inframarginal proximal plate]. Madreporiform tubercle distinct, situated lateraUy at the bases of the outer [supra- marginal] range of large plates of two adjacent rays. Ventral sur- face having deep ambulacral grooves, bordered by two ranges of strongly tuberculose plates; the outer [infra] marginal range con- sisting of 27 or 28 plates, besides a large, round [or elliptical], ter- minal or axillary [interbrachial] plate; the others are wider than long in the basal portion of the ray, becoming gradually shorter towards the extremity where they are rounded. All the [infra] marginal plates are visible from the upper [abactinal] side, and usually appear as an additional range of plates on each margin of the ray, making five with the three properly belonging to the upper surface. Those of the inner range bordering the ambulacra (adam- bulacral plates) are smaller than the marginal plates, about 38 to 40 in number; the basal or oral plates [oral armature] are triangular, those of the adjacent rays uniting by their longer margins; and with a single minute plate situated at these points. The plates of the exterior surface, both upper and lower, present a granulose or stri- ato-granulose surface which appears to have been produced by short setae or spines; and at the angles of the rays the marginal plates are armed by a few spines, which are as long or longer than the transverse diameter of the plates. Ambulacra composed of a double range of short, broad poral plates (ossicula), equal in number to the adambulacral plates ; their outer ends excavated on the posterior border, forming a comparatively large pore, just within its junction with the adambulacral plate. There appears to have been [is] but one range of pores in each set of ossicula, but these are large, dis- tinct, and pass between [through and not between] the plates." The following additional description is that of Mr. Cole. The surface of the axillary interbrachial plate "is granulose and bears 100 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. three short, thick, blunt pointed spines. The [infra] marginal plates bordering each reentrant angle bear similar but more slender spines, which are not 'as long as the transverse diameter of the plates.' The spines are arranged in a row near the distal margin of the plates and number five on the plates at the angle, the number and size decreasing until they disappear at the sixth or seventh plate from the angle. All the [infra] marginal plates are nearly smooth on the free margin and become gradually more granulose toward the line of junction with the adambulacral plates. * * * "The adambulacral plates are apparently less numerous than stated in the original description, and l the single minute plate [mouth plates] ' at the points of the pairs of the oral plates [oral armature] is visible in this specimen and is armed with two relatively long, slender spines which are apparently but a part of the full armature. The adambulacral plates, including the triangular oral [armature] plates bear well-defined spines, which are shorter than the diameter of the plates to which they are attached. Each plate bears two spines so near to the distal margin that the impressions of the short and obtusely pointed spines frequently bridge the well-defined groove between the adjacent adambulacral plates and terminate near the proximal margin of the next plate. The spines decrease in size toward the end of the ray and a few plates show only one spine. The plates of this range are thick, equaling two- thirds to three-fourths the depth of the groove. The vertical angle of the faces forming the lateral walls of the groove are beveled, so that lateral extensions of the groove are formed between each two plates on the same side. These lateral expansions are narrow and shallow at the oral surface, deeper and wider inward; so that the faces of the adambulacral plates near their junction with the poral [ambulacral] plates are reduced to a narrow edge which projects inward and nearly touches the corresponding plate on the other side of the groove. The general appearance of the fossil as well as the outline of the rays at the points where the broken block presents a transverse section of them indicates that the plates have their normal position, not having suffered distortion by pressure. "The ambulacral plates are shown by a well-defined mold of their under or external surface. The soft matrix which filled the ambula- cral furrow pressed upon the membranes connecting the ambulacral plates and occupying their pores, and as these membranes decayed it was forced by gentle pressure into the pores and between the edges of the plates. The mold of the groove is less than one-eighth of an inch in width in a ray measuring five-eighths at its base. The upper surface of the mold bears a narrow longitudinal median ridge which marks the junction of the two ranges of ambulacral plates. Similar transverse ridges, which are continuous with the lines mark- REVISION OF PALEOZOIC STELLEROIDEA. 101 ing the junction of the inner faces of the adambulacral plates, mark the proximal and distal margins of the ambulacral plates. These ridges do not cross at right angles to the median line, but include between their proximal sides an angle of about 125. These ridges indicate that the ambulacral and adambulacral plates were equal in number, and that the former were united in pairs along a straight median line rather than in an alternate right and left arrangement along a zigzag line, as is shown in Dr. Hall's figures. The pores described as being ' excavated in the posterior border of the ambu- lacral plates and just within their junction with the adambulacral plates' are not clearly shown on this specimen, although there are irregular and inconstant markings at some of the points of the molds of the lateral extensions of the groove. A series of pores near the median line is indicated by a series of small rounded prominences on each side of the median ridge and very close to it. These promi- nences are opposite the lateral expansions of the groove, and one is found on the mold of each ambulacral plate. The pores appear to have perforations very near the edges of the plates, or excavations in their margins." Abnormal development. Among the 400 specimens of this species recently found near Saugerties, New York, there are a few indi- viduals each with but four rays, though otherwise they appear to be of normal development. This is the first discovery of a four- rayed starfish in the Paleozoic. Formation and locality. In the Hamilton of the Middle Devonic near Hamilton, Summit and Cooperstown, New York, Two sepa- rated rays of apparently this species were found by the writer in the lower third of the Hamilton near Bartletts Mills, south of Thedford, Ontario. Two other fragments from the same locality are in the University of Toronto (Walker collection, No. 1610H). The type and other specimens are in the American Museum of Natural History, one* is in Colgate University, another at Wesleyan University, and two in Yale University Museum. Recently the New York State Survey collected in the Hamilton sandstone at Mount Marion, near Saugerties, New York, over 400 examples of this fine starfish. They occur in a limited area and are often found in association with Grammysia and in such manner that Doctor Clarke believes the starfishes were feeding on the bivalves. This is probably the most remarkable find of Paleozoic starfishes, and is certainly so for America. DEVONASTER CHEMUNGENSIS, new species. Plate 11, fig. 2. Of this species only the actinal side is known, and its general structure so far as can be made out is that of D. eucharis (Hall). 102 BULLETIN It differs from the latter in that each axillary interbrachial plate bears one stout spine, while D. eucharis has three. Further, in D. eucharis six or seven of the proximal inframarginal plates of each column bear on their outer side a number of small, slender spines. In D. chemungensis these spines may also be present but in addition each plate bears centrally a prominent tubercle for the articulation of a large spine. Formation and locality. The type is in the Museum of Columbia University, New York City (No. 6228G). It is from the Spirifer disjunctus sandstone of the Chemung of " Central Pennsylvania." NEW GENUS. NEW SPECIES. In the University of Chicago collection there is a specimen (No. 14397) collected by Professor Weller in the St. Louis limestone on Fountain Creek, at Waterloo, Illinois. This five-rayed asterid is small, shows only the abactinal side, and has very long but narrow marginals and not over six or possibly seven in a column; the basal plates of adjoining columns make the axils. Otherwise the abactinal skeleton of the rays consists of many small, loosely adjoining pieces whose arrangement can not be made out. Nor can the disk struc- ture be determined. The relationship of this specimen seems to be with the Promo- palseasteridse, with possible nearest affinity to the genus Mesopa- Ixaster. Measurements: R = not less than 10 mm., r = 4 mm. I'liODVJIOI'-A.ILi^E^STElRIlSr^B, new subfamily. Progressive Promopalseasteridse with the interbrachial areas com- plex and composed of the single axillary interbrachials, more than one pair of interbrachial marginals, and adambulacral plates. Podial openings in each ambulacral furrow mainly in two columns, Jbut proximally there may be two additional ones in a length never more than one-third the rays. Contains : Promopalseaster, new genus. PROMOPAL^EASTER, new genus. Plates 13 to 23, 25. Palxaster (part) of AUTHORS. Promopal&aster SCHUCHERT, Fossilium Catalogus, Animalia, pt. 3, April, 1914, p. 24. Promos, chief, and aster. The largest and chief starfishes of the American Ordovic. Generic characters. Disk above medium size, with distinct and angular interbrachial arcs. Rays five, slender to stout, more or less long, and tapering. Some species attain a large growth. BEVISION OF PALEOZOIC STELLEROIDEA. 103 Abactinal area of rays with numerous columns of small, more or less tumid, closely adjoining or reticulated, spine-bearing plates. A medial radial and two lateral or supramarginal columns of plates are usually more conspicuous than the accessory columns; however, the radial plates may be also inconspicuous. The longitudinal arrange- ment in columns is usually most pronounced, but sometimes the plates on each side of the radial column may have a decided diagonal arrangement. Small accessory plates are usually inserted between the columns of ossicles or between the individual pieces both distally and proximally. Disk with numerous small accessory plates like those of the rays, with no apparent definite arrangement except the supramarginal columns, which continue over the disk and unite angularly in the axillary areas. The larger plates bear several short articulating spines of which the one on the apex is the largest. Ambital areas well developed, there being one or more columns of small plates like the radial accessory columns both in form and arrangement. Madreporite usually conspicuous, radially striated, and situated near the margin of the disk between the adjoining columns of supra- marginal plates. Inframarginal plates small and numerous distally like the adambu- lacrals, but proximally they usually increase rapidly in width and assist in forming the small interbrachial arcs. Each ossicle has numerous small granules which probably bore short, smooth, articu- lating, blunt spines. Adambulacral plates distally like the inframarginals, increasing in width proximally and continuing around the plates of the inter- brachial areas. The two proximal plates of adjoining columns are the paii^s of plates in the oral armature. Each adambulacral plate has a more or less well-developed narrow ambiilacral extension which unites with the carina of the ambulacral plates. These extensions belong to the adambulacral plates, apparently always so distally, but proximally they are often separated from the adambulacral plates by sutures ; throughout the greater portion of the rays they are situ- ated medially, but in the proximal region they are either on the distal or proximal edge, each alternate plate having the extension from the same corner a little longer. Each plate has more or less numerous spine-bearing granules in addition to the three larger articulating spines which are inserted one on the ambulacral and two on the lateral edge. Ambulacral furrows generally very wide and regularly tapering. In each ray there are two columns of these plates which are slightly alternating or opposite. The plates are as numerous as those of the adambulacral columns, highly carinated, about as wide as long in the extreme distal portion of a ray, increasing rapidly in width proximally 104 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. and at the base of the rays may be four times as wide as long. The podial openings, one to each plate and in single ranges, are at the extreme lateral edges excavated between the sutures of adjoining plates and beneath the adambulacrals. Proximally, however, every other podial opening progresses inwardly with the joined carinae and issues where these begin to fork. Here there are, therefore, four columns of podial openings in each ambulacral furrow. This change is indicated by the changed position of the adambulacral plate exten- sions and also b} 7 the pairs of forked carinse of the ambulacral plates. Throughout the greater portion of a ray the carinse are regular and alike on each plate, but toward the mouth they change rapidly in direction and soon they are arranged in forked parrs, one curving distally, the other proximally, with the lateral portions of each pair hi contact and uniting with the extensions of the adambulacral plates. The most proximal plate of each ambulacral column is usually considerably modified, longer than wide, and more or less triangular in outline, between which there is sometimes inserted a small quadrangular ossicle. These pieces belong to the oral armature. Interbrachial areas of medium size, with the interbrachial marginal plates usually arranged in pairs but in some forms the series may be terminated by single ossicles. The number of these plates in an area varies in different species, there being two, three, or five inside the marginal inframarginals, and all seem to be derived from the inframarginal series by inward crowding. Genoholotype. Palseaster granulosus Meek (not Hall = P. speciosus Meek). Distribution. Restricted to the Middle and Upper Ordovicic of America, chiefly within a radius of 50 miles about Cincinnati, Ohio. The following are the known species: P. wilsoni (Raymond). Black River (Lowville). P. prenuntius, new species. Trenton. P. sp. undet. ?Young of P. speciosus Meek. Maysvillian. P. speciosus (Meek). Maysvillian. P. granulosus (Hall). Lower Richmond. P. bettulus, new species. Richmond. P. spinulosus (Miller and Dyer). Richmond. P. exculptus (Miller). Richmond. P. wylcqffi (Miller and Gurley). Richmond. P. dyeri (Meek). Maysvillian. P. magnificus (Miller). Maysvillian and Richmond. Remarks. Eight of the ten species here referred to Promopalseaster have been described as Palxaster. They have little direct relationship with the latter genus in that the species are much larger, have well developed ambital areas, numerous accessory plates, well defined REVISION OF PALEOZOIC STELLEEOIDEA. 105 interbrachial arcs, and from two to five interbrachial marginal plates in each interbrachial area. Promopalseaster is a direct development from Mesopalseaster, in that it preserves the generic characters of the latter, which is in turn a descendant of Hudsonaster. In Hudsonaster there are no interbrachial plates, accessory or ambit al columns. In Mesopalseaster the single axillan^ marginal plate of Hudsonaster has been crowded orally and is no longer a marginal but an axillary interbrachial. Further, the radial and both marginal columns are more or less completely separated from one another by accessory columns of smaller plates. In Promopalseaster this addition of columns of small plates is carried to its greatest development. The character, however, which is relied upon to distinguish it from Meso- palseaster is that Promopalseaster always has two or more interbrachial marginal plates in each interbrachial area, while the former genus has one, the axillary interbrachial, and Hudsonaster has none. The position and number of axillary and interbrachial marginal plates will readily distinguish these genera, but if the abactinal area of a new form alone is seen it will be difficult to determine whether it is a Mesopalseaster or Promopalseaster. Small size and few accessory columns, especially ambital, will help somewhat to distinguish Mesopalseaster from Promopalseaster. Accepting Promopalseaster as a descendant of Mesopalseaster, it is seen that the axillary interbrachial plates of the latter have in P. exculptus been followed by two proximal pieces of adjoining infra- marginal columns. The same is probably true of P. spinulosus, but here no axillary interbrachial has been observed. In P. speciosus and P. bellulus, another pair of proximal inframarginals have become interbrachial plates, but the axillary interbrachial is also unknown in this species. In P. dyeri and P. magnificus the axillary inter- brachial is present along with two pairs of interbrachial marginals. In other words, there is in Promopalseaster a constant increase in the size of the interbrachial area produced by the crowding of pairs of proximal inframarginal plates. P. speciosus and P. bellulus are closely related species, and so far as can be determined have but two interbrachial marginal plates, while our theoretic development requires three. There is, however, in each area orad to the interbrachial plates a small hiatus which in some cases appears to be filled by extensions from the adambulacral plates. It seems therefore probable that the axillary interbrachial plate may be retained, but is not now discernible on account of its small size or adhering clay; it may also have dropped out or have been absorbed and its place occupied by the adjoining adambulacral plates. In a general way it can be stated that there is also a regular increase in the number of accessory plates from Mesopalseaster to Promopalse- aster through P. spinulosus, P. speciosus, P. bellulus, P. dyeri into 106 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. P. magnijicus. It seems probable from the mature material studied that the most differentiated species of Promopalseaster, P. magnificus, passed through ontogenetic stages comparable to Hudsonaster, Meso- palseaster granti, and Promopalxaster bellulus. P. magnificus at present stands alone in its beautiful diagonal and longitudinal arrangement of abactinal plates. PROMOPAL^ASTER WILSONI (Raymond). Plate 13, figs. 1, 2. Palseasterf wilsoni RAYMOND, Ottawa Naturalist, vol. 26, 1912, p, 77, pi. 5, figs. 1-1. Original description. " The specimen is exposed from the abactinal side, and preserves the greater part of one arm, the disk, and the stumps of the other four arms. The diameter of the specimen, when complete, must have been about 75 mm. (3 inches), and the diameter of the disk is 20 mm. This is large for a starfish from the lower Ordovician. The arms are quite convex, with a gentle taper, reminding one somewhat of the common recent starfish, Asterias vulgaris, and as in that species, the arms were probably somewhat flexible. The greater part of the abactinal side of the disk and arms is covered with small convex, overlapping, V-shaped plates, which are arranged with the point of the V directed toward the margins. Along the crest of each arm there is a single row of larger plates (radial column). These plates are quite large and hexagonal in out- line near the disk, but become smaller, triangular, and alternate in position farther out on the arm. There are two rows of marginals [supra and infra] these plates being larger and flatter than the other plates, and covered with minute tubercles, which may be spine bases. Close to the disk, the supramarginals and [infra] marginals seem to be of the same size, both rectangular, and the plates' of the supra- marginal row directly over those of the marginal series. Farther out on the arms, the plates are pentagonal, those of the two rows alternating in position, and dove-tailing, and the supramarginals are smaller than the marginals. One of the marginals, about half- way out on the arm, is 1.25 mm. high and of about the same breadth. The smaller triangular plates which cover the greater part of the arm average about 0.5 to 0.6 mm. in height. On one of the arms, the small triangular plates seem to be arranged in rows parallel to the axis of the arm, but the plates on the longer arm seem to be more irregular, although a general arrangement in rows can be seen. On this arm there are a number of very small plates scattered about, especially on the top of the arm, thus adding to the irregularity. The triangular shape of these plates gives the arm a neat pattern, the plates making diagonal rows backward and forward from the row of large plates along the top of the arm. The madreporite, which is nearly circular in outline, and 2 mm. in diameter, is in position, but slightly BEVTSION OF PALEOZOIC STELLEROIDEA. 107 tipped down at the inner side, in an interradius, and not far from the center of the abactinal side of the disk. The surface is probably worn, for it appears perfectly smooth." In places the abactinal ossicles are removed, "disclosing the ambu- lacral plates. These plates, which are long and rather thick, seem to be alternate in position. Two of the plates seem to be pierced by pores near their proximate ends, two pores piercing each plate ver- tically. Near the outer end of the more perfect arm there is a space where a few of the small triangular plates are missing, and here also the ambulacral plates can be seen from the upper side. Each plate has a narrow keel on that side. Other details of the plates of the actinal side are unknown." Formation and locality. In the Lowville limestone at City View, near Ottawa, Canada, where this interesting asterid was found by Miss A. E. Wilson, in whose collection the holotype now is. Remarks. It is interesting to note that this fully developed Promopalseaster, one closely related to P. magniiicus of the highest Ordovicic, occurs in much older rocks; that is, in the Black River formation. This occurrence again brings out the fact often noted by the writer, that Paleozoic asterids are slow to change in characters accessible to the paleontologist. P. magniiicus, like P. wilsoni, Raymond states, "has two rows of large marginals and a row of large plates along the top of each arm, while the greater part of the surface is covered with small convex triangular plates. Miller found the plates to be spine-bearing, a point which can not be definitely decided in the present species. From the photograph (submitted by Schuchert) it appears that there are spaces between the small triangular plates on the abactinal sur- face in Pals&aster f magniiicus, while in P. ? wilsoni these plates actually overlap each other." PROMOPAL^EASTER PRENUNTIUS, new species. Plate 13, fig. 3; plate 15, fig. 5. Remarks. This species is represented by a good specimen showing the actinal side, the other adhering to the limestone. R = somewhat less than 30 mm., r = 7 mm. Width of ray at base about 9 mm. P. prenuntius is closely related to P. speciosus, but is distinguished therefrom by the smaller size, less pustulose ornamentation of the plates, and the lower position in the geological column. The plates of the axial area are also more regular in size and there are in the proximal region nearly always two adambulacrals to each inframar- ginal; at about mid-length of the rays the plates of each of the two columns are about equal in number, but every now and then there is some duplication among the adambulacrals. P. prenuntius is equally as spinose as P. speciosus and the former is undoubtedly the ancestor of the latter. 108 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Formation and locality. The holotype was collected by Mr. Wendt near Frankfort, Kentucky, apparently in the Lower Trenton. The writer's attention was called to it by Prof. Arthur M. Miller, who has it in his charge at the State University of Kentucky at Lexington. PROMOPAL^ASTER, species undetermined. Plate 14, figs. 1, 2. Asterias primordialis ANONYMOUS, two lithographic figures sent out by the Western Academy of Sciences previous to 1872 without description or known author. Nothing more is known of this species than the two good original illustrations reproduced here. These figures were sent to F. B. Meek on May 18, 1872, by Mr. D. H. Shaffer, of Cincinnati, Ohio, with the following comment (they were found in the Meek correspondence kept in the United States National Museum): "The lithograph in- closed of this Asterias was found here, and in the possession of Joseph Clark (deceased). I saw it, and this is a faithful picture of it. I think the fossil is either in the possession of his nephew or in the cabinet of Maxwell's Female Seminary, which now is the owner of Mr. Clark's cabinet. I think it is worthy a place with Palseaster shafferi in the illustrations of the Ohio Geological Survey." From the illustration one sees that this form had on the abactinal side of the rays about nine columns of ossicles. Of these the radial, supra- and inframarginals are largest. Between the marginals there is a single column of tiny ambital pieces, while on each side of the radials are two columns of small accessory pieces. The structure of the disk pieces can not be made out. The interbrachial areas are small and do not appear to have more than two inwardly crowded inframarginals, but there may also be present single very small axillary pieces. The adambulacral and inframarginal columns appear to have not more than 20 ossicles hi a column. Of ambulacrals there are some- what more, about 25, and these alternate slightly with one another. "Asterias primordialis" appears to be the young of Promopalse- aster speciosus (Meek), and the differences can all be explained as due to incomplete development; that is, the former has fewer ossicles and fewer secondary columns of plates. Formation and locality. From the hills of Cincinnati, in the Mays- ville formation. The present whereabouts of this fine specimen is unknown. A still smaller and younger specimen is in the Faber collection of the University of Chicago (No. 9567), and was collected at 350 feet above the Ohio River at Cincinnati. It has but a single axillary plate in each interbrachial area, and about 15 adambulacrals in a column. It is interesting to note that in these young individuals of Promopalxaster, the smaller they are the more they approach Mesopalseaster and suggest Hudsonaster. EEVISIOST OF PALEOZOIC STELLEKOIDEA. 109 PROMOPAL^ASTER SPECIOSUS (Meek). Plate 14, figs. 3, 4; plate 15, figs. 1-4. Asterias antiquata LOCKE (not A. antiqua Troost), Proc. Acad. Nat. Sci., Phila- delphia, for 1846-47, vol. 3, 1848, p. 32, fig. on p. 33. Palseaster antiquata HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 286. Palssaster antiquatus HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., rev. ed., 1868=1870, p. 328. Palseaster granulosus MEEK (not Hall 1868), Amer. Journ. Sci., ser. 3, vol. 4, 1872, p. 276; Geol. Surv. Ohio, Pal., vol. 1, 1873, p. 60, pi. 4, figs. 3a-c. Palseaster speciosus MEEK, Amer. Journ. Sci., eer. 3, vol. 4, 1872, p. 277 (name at end of description); Geol. Surv. Ohio, Pal., vol. 1, 1873, p. 61 (name at end of description). Also see Promopalxaster, sp. undet., page 108. Original description by Meek (1873). " Attaining a large size; rays (as a little depressed by accident) slightly more than twice as long as their breadth at their inner ends ; and tapering regularly from the disk to their free ends, which are rather acutely angular; breadth of the disk a little less than that of the inner ends of the rays. Mar- ginal pieces convex [bearing numerous small spines], nearly or quite as long as wide, numbering on each side, in the whole length of each ray (1.40 inch), 31. Adambulacral pieces, a size smaller, as long as wide, convex, and numbering on each side 39 to 40; like those of the [infra] marginal series, each bearing a small spine [several spines], the largest of which are each about 0.14 inch in length and 0.02 inch in thickness [between each two adambulacral plates just above the podial opening is inserted a large and thick spine], Ambulacral pieces very short, or nearly three times as wide as long, apparently not alternating with the adambulacral pieces, and each provided with a rather sharp ridge across nearly its entire breadth. Dorsal side of disk and rays composed of small tuberculiform pieces, of which about 16 rows may be counted at about half way between the inner and outer ends of each ray, along the middle of which those of two [always two] or three of the rows are a little larger than the others; each dorsal piece bearing a minute, very short spine [they are nume- rous and like those of the marginal plates]. Surface of all the pieces minutely granular, one of the central granules always being somewhat larger than the others for the articulation of a spine [all of these granules bear spines]. Oral pieces and madreporiform body un- known." Emended description. Rays regularly tapering. R = 44 mm., r= 11 mm., R = 4r. Width of ray at base 14 mm. Abactinal area of rays covered by small, highly convex, tuber- culated plates, each bearing numerous short slender spines. These ossicles are arranged in longitudinal columns, of which there a?e 15 on each ray at about its mid-length. The plates of one or 110 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. two median columns are larger and more convex than the three columns immediately on each side, while the pieces of the fourth column are again more pronounced than those on each side of it, and probably represent the supramarginals. Between the supramarginals and the inframarginals are from one to three columns of ambital plates, and between these are inserted here and there some small extra pieces. The extra pieces are mainly developed beside the supra- marginal columns. Plates of the disk like those of the rays, but their arrangement is unknown. Madreporite unknown. Inframarginal plates very convex, about as wide as long except- ing near the base of the columns where they are about twice as wide as long. There are about 31 of these ossicles in a column, and they have numerous small, slender, sharp spines articulating on well- developed tubercles. Adambulacral plates greatly resembling the inframarginals. They are largest near the mid-length of each ray, diminish in size distally, but proximally they hold their width and decrease in length and therefore are more numerous. There are 43 plates in each column besides those in the oral armature. Ambulacrally each adambulacral plate has a short attenuate extension which articulates with the crests of the ambulacral plates. Between all the adam- bulacral ossicles there is inserted in the suture facing the ambulacral groove a single large, thick spine, excepting for a short distance proximally where such are inserted in every other suture. The lateral surface of the adambulacral plates in the distal half of the columns bears numerous small spines like those of the inframarginal series, but proximally these spines are gradually displaced by two or possibly more larger ones like those on the ambulacral face. Ambulacral furrows wide and deep. Ambulacral plates of adjoin- ing columns opposite, two or three times as wide as long, and slightly overlapping proximally. Upon each plate there is a high median ridge which arises at the inner end of the plate and abuts against the short extension of the adambulacral plate. The podial open- ings are between the plates and beneath the sutures of the adam- bulacral plates, excepting near the base of the rays, where they occur in every other suture. The ridges here are sigmoid and not straight, as they are more distally. There appear never to be more than two rows of podia in an ambulacrum. In each axillary area inside of two large inframarginal plates and between the converging adambulacral columns are inserted two pairs of wedge-shaped interbrachial marginal plates ornamented like the inframarginals. Formation and locality. The holotype and only known specimen was collected by Mr. Charles B. Dyer in the Maysvillian at Gin- REVISION OF PALEOZOIC STELLEROIDEA. Ill cinnati, Ohio. It is now in the Dyer collection, Museum of Com- parative Zoology (No. 22), Harvard University. The original of Astcrias antiquata Locke was found at Cincinnati, but the specimen appears to be lost. Remarks. Meek (1873) in describing this specimen, provisionally referred it to Palseaster granulosus Hall and commented as follows: "Although this fine Star-fish agrees pretty well in most of its characters with the description of Palseaster granulosus of Hall (which has not yet been figured) , I am far from being entirely satisfied that it is really the same, as it seems to differ in some important respects. For instance, the rays of P. granulosus are described as being 'obtusely rounded at the extremities'; while in the form under consideration they are rather acutely rounded, if not angular. Again, instead of having 25 of the marginal pieces on each side, in a space of one inch and a quarter from the apex of each ray, and 42 or 43 of the adam- bulacral, it shows in this space 28 marginal and only 32 adambula- cral pieces. The number of the latter being about 10 less on each side in the same space a rather decided difference, showing the inner row to consist of proportionally larger pieces. Its ambulacral ossicula seem to have the same proportional breadth and length as in the type of P. granulosus and also have each a similar ridge across the middle; but these ridges do not show the zigzag arrangement mentioned in the description of P. granulosus. Prof. Hall does not say how many rows of pieces are seen on the dorsal sides of the rays of his species; but he states that it is probably the same species that the Western Academy of Sciences sent out lithographs of under the name Asterias primordialis. One of these lithographs, now before me [reproduced here on plate 14, figs. 1, 2], represents from 8 to 10 rows of these dorsal pieces, which is 6 or 8 less than may be counted near the middle of the rays of our specimen. "In noticing this form in the Am. Jour. Sci., vol. 3 (3d series), p. 277, I proposed, in case it should be distinct from P. granulosus, to call it P. speciosus. J) With such marked differences between P. granulosus and P. speci- osus, and since the type-specimen of the former can not be located to determine the interbrachial structure, it is deemed advisable to adopt Meek's name P. speciosus for the specimen described above. In the upper portion of the Ordovicic in the Richmond formation occurs another closely related species, P. ~bellulus, which is dis- tinguished from P. speciosus in having more slender rays and usually a greater number of plates in the actinal columns. The greatest difference, however, is that it has seven pairs of adjoining adam- bulacral plates proximal to the interbrachial plates, while in P. speciosus there are but two or three pairs of adambulacrals. In other words, the adambulacral oral extensions are far longer in P. bellulus than in P. speciosus. 50601 Bull. 8815 8 112 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. The illustration of Asterias antiquata Locke referred to above by Meek is very poor, and yet it seems to be the same species which the latter described as P. granulosus Hall = P. speciosus Meek. This, however, can not now be established since the whereabouts of Locke's specimen is unknown. Both specimens are from Cin- cinnati and are of about the same size. Hall comments on this species (1870) as follows: "This species was noticed by Dr. Locke, as cited above, but with- out specific description, and expressing a doubt whether it was or was not identical with the Asterias antiqua of Troost. The figure would indicate a distinct species from that of Dr. Troost." PROMOPAL^EASTER GRANULOSUS (Hall). Palxaster granulosa HALL Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 285. Palseaster granulosus HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., rev. ed., 1868=1870, p. 327. Original description. "Body of medium size, five rayed; rays a little more than twice as long as their breadth at base; obtusely rounded at the extremities. Upper surface of rays composed of numerous very small tuberculose or subspinose plates; the madre- poric tubercle large, quite distinct, situated laterally at the base of two of the rays. Under surface of rays composed of a [infra]marginal range of small tuberculose plates, about 25 on each, side in a ray measuring one inch and a quarter from base to apex; and an inner (adambukicral) range of smaller plates, of which about 42 or 43 can be counted on the same ray; the terminal or oral [armature] plates are small, elongate, subtriangular, in pairs at the base of the adjacent rays. Ambulacral areas composed of a double series of short, broad, slightly curved poral plates (ossicula) , each plate marked by a sharply elevated ridge along its entire breadth, commencing on the one plate at the outer posterior angle and terminating on the anterior inner angle, and running in the opposite direction on the adjacent plate. When the outer ridged surface of the poral plate is ground away, the narrow openings or pores are visible between the plates, apparently in two rows in each series, making four ranges of pores in each ambu- lacral area.' (The marginal ranges of pores are obscure, and may only be apparent [there is probably an error here in regarding the inner openings as podial openings].) On the under surface, near the bases of the rays, the tubercles bear short spines some of which are still in place." ' * Some figures of a Palssaster, closely allied to or identical with this one, from Cincinnati, Ohio, have been circulated by the Natural History Society of that place, under the name of Asterias primordialis; but no description of it has ever been published, so far as I know, nor EEVISION OF PALEOZOIC STELLEKOIDEA. 113 do I find it at all recognized in the catalogues." These original figures are here reproduced as Promopalseaster, sp. undet. (pi. 14, figs. 1, 2). The specimen is probably an immature individual of P. speciosus. Formation and locality. The holotype was found by Mr. J. Kelly O'Neall in the Lower Richmond formation near Lebanon, Ohio, and was loaned to Hall for description. The whereabouts of the specimen can not now be located. Some years ago the writer asked the late Professor Whitfield about it and he answered that Halt had returned it to O'Neall, but the latter in turn wrote (Aug. 27, 1896) that he had not received it. Remarks. This species seems to be distinguished by the smaller number of plates in the inframarginal columns, as the following table shows. However, until the nature of the interbrachial areas is known, the relationship of P. granulosus to other Paleozoic forms can not be stated. Inframar- ginals. Adambu- lacrals. Inter- brachials. Adambu- lacrals orad to inter- brachials. P granulosus (Hall) 25 42-43 P speciosus (Meek) . * 31 43 4 4 P bellulus new species . 31-38 48-55 4 8 P. magnificus (Miller) . 45 50-52 5 5-6 PROMOPAL^ASTER BELLULUS, new species. Plate 15, figs. 6-8; plate 16, fig. 1; plate 18, figs. 4, 5. R = 36 mm., r= 10 mm., R = 3.6r. Width of ray at base, 11.5 mm. A first view of this species gives the impression of P. speciosus (Meek). Of the latter but a single specimen is known and it is now lost, while of P. bellulus there are numerous examples, all of which have the same specific characters. The specimens preserved in lime- stone have the rays more slender, while those from soft shale are as stout as in P. speciosus. In the holotype of P. bellulus from limestone there are 55 adam- bulacral and 38 inframarginal plates in each column, while in P. speciosus there are 43 and 31 plates, respectively, in the same columns. Shale specimens of P. lellulus have less plates in a column. One of these, a somewhat smaller specimen than the holotype, has 31 inframarginal plates, the same number as in P. speciosus, and about 48 adambulacrals in each column. P. bellulus has therefore from 31 to 38 inframarginal plates and from 48 to 55 adambulacral plates in each column. 114 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. A more important distinguishing character exists in the inter- brachial areas. In P. lellulus and P. speciosus there are two pairs or four inter brachial marginal plates, while the number of adjoining adambulacral plates is greater. In the latter there are four of these on each side, while in the former there are eight. In other words, the interbrachial areas in P. lellulus are much more elongated orally than in P. speciosus and the interbrachial arcs are also greater. Abactinally on the rays the arrangement and number of columns of plates appear to be very much like those in P. speciosus. The individual plates are, however, less convex, cruciform, more like those of P. spinulosuSj and have stronger and more abundant spines. The spines in P. bellulus are very numerous, short, and slender, and are inserted one on the apex of each plate and a number around the sides in the angles between the radial extensions. In P. speciosus the spines are grouped over the convex area of the plate and around the larger central spine. The arrangement of the plates of the disk appears to have no marked pattern and consists of very numerous small plates like those of the rays. The adambulacral plates each bear three spines as in P. speciosus and increase in size toward the base of the rays. One is inserted on the inner or ambulacral face of the plate and the two closely adjoining ones immediately above and laterally on the rounded surface. Ambulacral plates as in P. speciosus. Madreporite of medium size, subcircular in outline, with numerous straight, sharp, radiating ridges. It is inserted proximally between two adjoining columns of supramarginal plates. A development occurring only in this species of Promopalseaster is a single plate oral to the proximal adambulacrals. This plate has been seen in four axillary areas. It may be a modified ambu- lacral and may correspond to a similar plate, which is an ambulacral, hi P. exculptus. Formation and locality. Of this species quite a number of speci- mens were secured by Mr. Harris in the Waynesville beds of the Richmondian. They are all in the United States National Museum. One excellent specimen (the holotype) showing the actinal area, and two smaller poorer specimens are attached to limestone, while eleven fair specimens and fragments of eight other individuals are from a soft blue shale associated with Zygospira modesta and Ptilo- dictya sliafferi (upper form). All were secured in the vicinity of Waynesville, Ohio. The shale specimens were exhumed from one small area, showing that probably many starfishes could be similarly secured by collectors, since these soft specimens in weathering out are broken up and lost; more asterids will be found when they are dug or quarried for. Another good specimen is in the Gurley col- lection of the University of Chicago (No. 10983) and was found at Versailles, Indiana. REVISION OF PALEOZOIC STELLEEOIDEA. 115 Remarks. The similarity and difference between P. bellulus and P. speciosus are pointed out above. Another related species is P. spinu- losus, which has the same kind of abactinal plates and madreporite as P. bellulus. In the former the rays are far longer and narrower, with fewer ambital plates and columns between the supramarginals. Actinally P. spinulosus is at once distinguished from P. bellulus in the much smaller interbrachial areas, in that it does not have more than two interbrachial marginal plates while P. bellulus has four. In the latter species there are also far more adambulacral plates around the interbrachial pieces. It will always be difficult to distinguish fragments of P. bellulus, P. speciosus, and P. spinulosus from one another, even with good material, since at least two specimens of P. bellulus were identified as Palseaster granulosus Hall = Promopalseaster speciosus (Meek) . Cat. No. 40879, U.SJST.M. PROMOPAL^EASTER SPINULOSUS (Miller and Dyer). Plate 16, fig. 2; plate 17, figs. 1, 2; plate 18, figs. 1-3. Compare with Promopalseaster exculptus (Miller). Palseaster spinulosus MILLER and DYER, Journ. Cincinnati Soc. Nat. Hist., vol. 1, 1878, p. 32, pi. 2, figs. 12-126. Palxaster longibrachiatus MILDER, Journ. Cincinnati Soc. Nat. Hist., vol. 1, 1878, p. 102, pi. 3, fig. 4. Original description (of Palseaster spinulosus). "Pentagonal; rays longer than the diameter of the body; marginal plates globular near the apices of the rays, but lengthened toward the junction with the body. Six of these plates measure from the body on a ray three- tenths of an inch. Two wedge-shaped plates form the junc- tion of the marginal plates with the body. " Adambulacral plates a little smaller [in width] than the margi- nal pieces, and possessed of two or three [four or more] spines to each plate. These spines taper to a fine point, and are longer than the diameter of the plates. The oral pieces are not determined [they are like those in P. exculptus]. The ambulacral plates have their greatest length across the rays, and are possessed of a sharp ridge in the middle which seems to be [is] connected at one end with the adambulacral pieces. [The entire number of plates in each range can not be given, since the rays are imperfect.]- "The dorsal side is covered with plates united by angular exten- sions. They are of unequal size, and strongly tuberculated or spi- nous. "The madreporiform tubercle on its upper face is an oblate sphe- roid, much depressed, and marked by fine radiating stride, which become more numerous by intercalation, without bifurcation. [This plate is interpolated between the proximal plates of two ad- 116 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. joining columns of supramarginal ossicles and about medially be- tween the margin and the center of the disk.]" Supplementary description. II unknown, r=7 mm. Width of ray at base 9 mm. Abactinal area of rays margined by inframarginals, above and inside of which are columns of large, bipyriform, multispinous, su- pramarginal plates of which there are five in 5 mm. at the base of a ray. The columns of adjoining rays unite in the axillary areas. Inside of the supramarginal columns the rays and disk have numer- ous smaller accessory plates, usually triangular in outline or vari- ously stellate. They lie upon or against each other, each ossicle bearing at least one small spine, and leave between them numerous abactinal openings. Outside of the supramarginal plates in the ambital areas are spicular ambit al plates which cover the abactinal side of the inframarginals. Description of the type of Palseaster longibrachiatus. 11 = 38 mm., r = 7 mm., R = 5.4r. Width of ray at base 9 mm. Rays large, slender, tapering slowly; actinally somewhat convex at their outer ends, but elsewhere concave. Abactinal area unknown. Inframarginal plates granular, large, decreasing gradually in size dis tally, highly convex, in outline tetragonal or pentagonal and common to both the abactinal and actinal areas. From 28 to 30 plates in a column on each side of a ray. Axillary areas occupied by the proximal plates of adjoining infra- marginal columns, and in the interbrachial areas there are two sub- quadrangular or sub triangular interbrachial marginal plates. There may be an additional small single plate in each area. Adambulacral plates greatly resembling the inframarginal ossicles. Distally they progressively overlap the inframarginal plates con- siderably, while proximally they gradually increase in size and are entirely inside and depressed beneath the plane of the marginal columns. Orally the columns of adjoining rays unite in two wedge- shaped modified adambulacrals (oral armature pieces) . Ambulacral furrows deep and gradually tapering. Ambulacral plates unknown. Formation and locality. The type of Palseaster spinulosus (No. 16, Mus. Comp. Zool.) is said to have been found at Cincinnati, Ohio. This is probably an error, since il/s color and preservation indicate the Richmond formation and that it comes from some locality considerably to the north of Cincinnati. The only other known specimen, the type of P. longibrachiatus, was found in the Richmond formation near Clarksville, Ohio, and is in the Harris collection, United States National Museum. Remarks. The type-specimen of Palseaster spinulosus, which is free, preserves but the disk and a small proximal portion of the rays. REVISION OF PALEOZOIC STELLEROIDEA. 117 Comparing it with the actinal area of P. longibrachiatus Miller, it is evident that both forms are identical. In the latter the rays are complete, very long and slender, and there is nothing to show that they were dissimilar in P. spinulosus Miller and Dyer. Both types are nearly of the same size and the actinal plates are identical in form and arrangement. The abactinal area in P. longibrachiatus is not known, but since its actinal structure is that of Promopalseaster spinulosus it is safe to assume that the abactinal structure is alike in both. P. spinulosus actinally is also closely related to P. exculptus (Miller). The latter appears to have shorter and more rapidly tapering rays and in general seems to be a more robust species. Another character by which these species can be separated is that the interbrachial areas in P. spinulosus have two, while P. exculptus has three inter- brachial marginal plates. It is possible that the former species also has three interbrachial marginal plates, but as yet the small single piece of each area has not been observed, while in P. exculplus it is distinctly present in each of the five areas. Should P. spin- ulosus also prove to have three interbrachial plates, it would be diffi- cult to distinguish it actinally from P. exculptus. P. dyeri and P. magnificus are far larger species and have five interbrachial marginal plates in each area. Cat. No. 40881, U.S.N.M. PROMOPAL-EASTER EXCULPTUS (Miller). Plate 18, fig. 7; plate 20, fig. 2. Compare with Promopalasaster spinulosus (Miller and Dyer). Palseaster exculptus MILLER, Journ. Cincinnati Soc. Nat. Hist., vol.4, 1881, p. 69, pi. 1, fig. 1. Original description. " Pentagonal; rays a little longer than the diameter of the body; diameter of the body, about 0.93 inch; length of ray measuring to the center of the body or disk, about 1J inches; breadth of a ray at the junction with the body, about 0.57 inch; rays obtusely pointed. "The [infra-] marginal range consists of somewhat quadrangular plates, having a width a little greater than the length; the first 8 of these have a length of one-half inch, and there are about 18 in the length of an inch, and not far from 25 in each range, though the specimen does not permit us to make the count with certainty. The surface is strongly tubercular, and was probably spinous [originally covered with numerous small spines]. "The adambulacral range consists of about 28 [25] plates, on each side of a ray; they are narrower than the marginal plates, but have about the same length. Each plate [has a short ambulacral exten- sion conforming with the ridge of the ambulacral plates and] bore 118 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. strong spines, and some of them, preserved on our specimen, have a length greater than the length of a plate. * * * "The ambulacral plates have their greatest length across the rays, thus providing a wide ambulacral furrow. Each plate is furnished with a sharp ridge in the middle, that curves slightly outward, from the center toward the adambulacral range, increasing in height until it approaches or abuts against the adambulacral plate. * * * The appearance of having been carved out, which is presented by the ambulacral plates, suggested the specific name. [The plates of adjoining columns are opposite one another or slightly alternate.] "The dorsal side and madreporiform tubercle unknown." Supplementary description. R = 37 mm., *r=12 mm., R = 3.1r i . The specimen is very much flattened and the exact width of a ray can not be given but appears to have been about 10 mm. at the base. In the axils of the rays the proximal inframarginal plates of adjoin- ing rays rest against each other, inside of which are two wedge-shaped interbrachial marginal plates. Orad to these and adjoining the pairs of proximal adambulacrals (oral armature pieces) is situated a single subquadrate ossicle, making three interbrachial marginal plates in each interbrachial area. The podial openings in the ambulacral grooves are between the sutures of adjoining adambulacral and ambulacral plates along the line where these two columns adjoin. The proximal ambulacral plates orad to the axial adambulacrals are considerably modified and are to be considered as parts of the oral armature. A single rather large plate is here partially wedged between the pair of termi- nating adambulacrals, and on each end of it abuts a single broadly triangular plate, a modified ambulacral, completing the oral ends of the ambulacral columns (see drawing, pi. 20, fig. 2). Formation and locality. In the Richmond formation near Waynes- ville, Ohio. The holotype and only specimen is in the Harris collec- tion of the United States National Museum. Remarks. This species is near P. spinulosus, but apparently had wider, shorter, and more rapidly tapering rays. There appear to be but two interbrachial marginal plates in P. s-pinulosus, while P. exculptus has three in each area. Should the former species also prove to have three plates, a possibility, P. exculptus will be actinally identical with P. spinulosus. The wider and more rapidly tapering rays of P. exculptus can then be explained by the fact that the type- specimen is considerably distorted, since the adambulacrals and infra- marginals now lie wholly outside of the ambulacrals, which are com- pletely exposed, an unnatural condition. So long as the abacti- nal area of P. exculptus remains unknown the species had best be retained. Cat. No. 60608, U.S.N.M. REVISION OF PALEOZOIC STELLEROIDEA. 119 PROMOPAL^ASTER WYKOFFI (Miller and Gurley). Plate 18, fig. 6; plate 19, fig. 2. Palxaster wykoffi MILLER and GURLEY, Bull. No. 12, Illinois State Mus. Nat. Hist., 1897, p. 46, pi. 3, fig. 27. Original description. " Species medium size [R = 26 mm., r = 8 mm. Width of ray at base 10 mm.] Rays pateliform; length about one and a half tunes the diameter of the body; breadth of a ray at the place of junction with the body, about two-thirds the diameter of the body; obtusely pointed. "Marginal [inframarginal] plates wider than long and numbering about 15 in the length of half an inch from the body [or about 28 in a complete column]. The [infra] marginal plates curve regularly around the tips of the rays. The adambulacral range curves around the ends of the rays within the marginal plates and consists of sub- quadrate plates, wider than long. They are a little shorter than the marginal plates, so that there are about 18 in the length of half an inch [about 35 in a complete column]. There are 10 oral plates [oral armature] at the junction of the adambulacral rows, which present triangular extensions toward the center of the ventral cavity. A single irregular axillary plate rests between the terminal marginal plates and the angle formed at the junction of the adambulacral plates [this description is inadequate, see beyond for the detail]. The ambulacral plates have their greatest length across the rays, thus providing a wide ambulacral furrow. Each plate is furnished with a sharp ridge in the middle, extending from the middle furrow to the adambulacral plates. "Our specimen exposes only the ventral side and no spines are preserved. It is a well-marked and beautiful species." Formation and locality. Found in the Richmond formation "near Madison, Indiana, by Charles W. Wykoff, in whose honor we have proposed the specific name." The holotype is now in the Gurley coUection of the University of Chicago (No. 6066). Remarks. Each interbrachial area begins orally with a pair of elongate triangular basal adambulacrals (the oral armature pieces) dis tally to which there are two pairs of adjoining adambulacrals followed by a single axillary ossicle and usually one but also more or less of a second pair of interbrachial marginal plates. P. wylcoffi is therefore distinguished from P. exculptus in that it has orally to the single interbrachial plate two pairs of proximal adambulacrals instead of one as in the last-named species. These differences are not decided, and it may eventually be seen when more well-preserved material is at hand that they are either individual variations or due to differences in age. 120 BULLETIN" 88, UNITED STATES NATIONAL MUSEUM. P. wykoffi also greatly resembles Anorfhaster miamiensis, but the interbrachial areas in the latter are entirely composed of adambu- lacral plates, there being no interbrachial marginals present. While both species attained a similar size, another -difference is that the former has a greater number of plates in the adambulacral and inframarginal columns. PROMOPAL^EASTER DYERI (Meek). Plate 18, fig. 8; plate 20, figs. 3-6; plate 25, fig. 1. Palseaster dyeri MEEK, Amer. Journ. Sci., ser. 3, vol. 3, 1872, p. 257; Geol. Surv. Ohio, Pal., vol. 1, 1873, p. 58, pi. 4, figs. 2a-2/. Original description (1873). "Among the specimens loaned to me for study and description by Mr. Dyer, there is a very imperfect example of one of the largest known species of Silurian Starfishes. When entire, it could scarcely have measured less than 5 inches in diameter [R = probably 3 inches, or 75 mm.] across from end to end of the rays on opposite sides; and it presents a breadth of disk (as flattened by pressure) of about 2 inches [r = probably seven-eighths inch, or 22 mm.]. Its state of preservation is, unfortunately, such as not to admit of systematic description, but I think enough of its character can be given to enable the student to identify it, by the additional aid of figures. "The dorsal side of both disk and rays is composed of numerous small pieces [with stellate extensions], with the pores between them apparently so large that these pieces only seem to touch at three to four salient points of each, so as to form a kind of reticulated struc- ture; while they each bear a little central tubercle, with a minute pit in its top for the articulation of small, short [blunt] spines, generally about 0.07 to 0.09 inch in length, and about 0.02 to 0.03 inch in thickness. [Each plate bears but one of these spines and there appear to be no other smaller spines.] "In one of the axilla between two of the rays the so-called madri- poriform body can be seen near the margin of the disk. It is nearly flat, of an obtusely subtrilobate form [probably cruciform], with a breadth or transverse diameter of 0.30 inch, and a diameter at right angles to the same of 0.24 inch [not less than 10 mm.]. Its lobed edge is directed inward toward the middle of the disk, and its middle lobe is largest, and shows the little divisions between the slits or furrows, diverging and bifurcating inward like the nervation in the pinnules of some kind of ferns; while in the smaller lateral lobes these markings diverge outward. "The ventral side is also much obscured by the adhering matrix, and numerous detached and confusedly mingled spines. A row of comparatively small, tumid, nearly square [infra] marginal pieces, REVISION OF PALEOZOIC STELLEROIDEA. 121 however, can be seen in places, on each side of the ambulacral furrows, and alternating with a row of similar but slightly smaller adambulacral pieces, the number of pieces in each side of these being about the same. [In the proximal portion of a ray there are 19 marginal plates in 40 mm.]. The marginal pieces seem each to connect with the adjacent range of dorsal pieces, at places in the rays, by a little central salient point only, while those of both series are roughened by numerous comparatively coarse granules [all of which bear small spines], and each piece (especially [only] of the marginal row) also shows a small central pit for the articulation of a spine. These spines are larger than those connected with the dorsal pieces, being generally about 0.22 inch in length and 0.05 inch in thickness; they are smooth, straight, rounded, slightly thickened at the attached end, and tapering at the free end to a slightly blunted point. [The adambulacral plates also possess these large spines, which are inserted on the distal inner edge of each plate.]" In each interbrachial area there are two pairs of wedge-shaped pieces and a single diamond-shaped interbrachial marginal plate. There are at least 4 (and may be as many as 10) pairs of adam- bulacral plates orally or in the actinal interbrachial areas, the terminal pair being of the oral armature. Ambulacral plates much wider than long, increasing rapidly in width toward the mouth. Each plate is carinated as in P. sped- osus and proximally as in P. magnificus, with the podial openings confined to the extreme outer edge of the ambulacral furrows. Formation and locality. Found at Cincinnati, Ohio, in the Mays- ville beds, about 200 feet below the top of the hills. The ,holo type is in the Dyer collection of the Museum of Comparative Zoology, Harvard University (No. 13). Another specimen from near the top of the Maysville at Cincinnati, preserving only the disk and showing best the actinal area, is in the Gurley collection of the University of Chicago (No. 10989). Originally this individual appears to have been larger than the holotype. Remarks. The relationship of this species is with P. spinulosus, but the larger size, greater development of both dorsal and ventral spines, absence of columnar arrangement of the abactinal plates, and the presence of five instead of two or three interbrachial marginal plates will readily separate P. dyeri from the latter species. Even though this form appears to have attained a larger growth than P. magnificus, which it resembles in some characters, it differs in being more ponderous in its skeleton and in the far stouter and longer actinal spines. Regarding the generic position of Palseaster dyeri, Meek wrote: " Possibly I should have called it Petraster dyeri, for if the apparent presence of a few disk plates on the ventral side between the mar- 122 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. ginal pieces and the adambulacral, is not deceptive, it would, in that respect, conform to the structure of that group, and differ from Palseaster, as now understood; though I am inclined to think this appearance due to the accidental displacement of the parts at the point where there are some indications of a few disk pieces." These pieces are not to be correlated with the accessory interbrachial plates of Petraster, for in this genus accessory plates consist of two short columns situated between the adambulacral and marginal plates and uniting in the axillary areas. In P. dyeri, however, the four interbrachial marginal plates are continuous with the infra- marginal columns and evidently were derived from that series, and are not accessory interbrachials. It is true that all interbrachial plates increase the size of the disk, but in one they are derived from the inframarginal series, by crowding and in the other they are interpolated, newly developed plates that force apart the columns of plates in the rays sometimes almost to the distal ends. PROMOPAL.EASTER MAGNIFICUS (Miller). Plate 21, fig. 1; plate 22, fig. 1; plate 23, figs. 1-3. Palseaster magnificus MILLER, Journ. Cincinnati Soc. Nai. Hist., vol. 7, 1884, p. 16, pi. 4, figs. 3, 3a. Original description. "The diameter or breadth of the disk is one and one-fourth niches, and the distance from the point of one ray to the point of the opposite one, if the rays were wholly preserved in the specimen under examination, would be fully 6 inches. * * * "The plates upon the dorsal side are very convex, and part of them, at least, were spine-bearing, though it would seem that there was not more than one spine upon any single plate. The arrange- ment of the plates on the dorsal side of the rays is very ornamental. A single series of highly convex or conical plates, larger than the others, and each evidently bearing a central spine, occupies the middle of each ray; on either side near the margin of each ray there is a similar series, and the two intervening spaces are filled with smaller, convex plates arranged in rows which are directed diagonally forward from the plates of the side series to the plates of the central series, forming angles with each plate in the central series occupying an angle. This disposition of the plates on the dorsal side of the rays will, so far as known, serve to distinguish this species from any hitherto described. "The plates covering the dorsal side of the body or disk have been so much disturbed in our specimen that one can not correctly define them. "The ambulacral furrows are wide. The marginal plates are hexagonal, about the size of the larger plates on the dorsal side of KEVISION OF PALEOZOIC STELLEKOIDEA. 123 the rays, and separated from the side series by intervening smaller plates. Each bore several small spines, as shown by the small pits for their articulation. * * * "The adambulacral plates are hexagonal and much wider than long. They are more numerous than the marginal plates near the disk, but toward the apices of the rays they interlock alternately with the marginal plates. The number on each side of a furrow hi a complete ray would be 50 or more. The pits for the articulation of the spines are as numerous as they are on the marginal plates. These two series on the ventral side of the species, with numerous spines upon each plate, are in striking contrast with the plates on the dorsal side of the rays, where no plate bears more than a single spine. "The ambulacral plates have their greatest length across the rays, and seem to be about as numerous as the adambulacral ones. An angular depression marks the center of each ambulacral furrow, upon each side of which a sharp ridge arises upon each ambulacral plate, and curving forward and outward abuts against an adambula- cral plate. " Emended description. The largest specimen, the holotype, meas- ures R = (about) 67 mm., r = 17 mm., R = nearly 4r. Another speci- men: 11 = 58 mm., 3* =15 mm., R= nearly 4r. Abactinal surface very ornamental. Medially along each ray there is a radial column of highly convex plates, and two columns of similar plates, the supramarginals, are situated near the margin. Between the inframarginal, supramarginal, and radial columns are very nu- merous diagonal rows of smaller conical or angular plates, from 2 at the apex of the ray to 14 at base of same in each row on each side of the radial column. This may be stated in another way. Between the radial and supramarginal columns and again in the ambital area, or between the supramarginals and inframarginals, the diagonal rows have from one to seven plates in each area. These rows are directed diagonally and distally across the rays from the sides to the median column, the apex of the angle thus formed being occupied by the larger plates of the radial column. All of these smaller plates appar- ently bore a single short spine. Between all of the diagonal rows are inserted both proximally and distally numerous minute, conical, irregular-shaped plates. The supramarginal columns continue across the disk and those of adjoining rays unite in the axillary areas. The radial columns also extend over the disk but do not attain the center, and the diagonal arrangement of the small plates ceases near the proximal portion of the rays. The pattern arrangement of the disk plates can not be made out since their original position is disturbed. Madreporite quite large, quadri-lobate in outline, depressed, convex and very finely radiately striate. It is situated in one of the axillary 124 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. areas, less than 4 mm. from the margin. Against it rest two columns of supramarginal plates. The inframarginal plates, of which there are at least 45 in each column, are distally small and subquadrate but proximally increase rapidly in size and at the base of the columns are three to four times as wide as long. These plates bear numerous pitted pustules upon which articulated short, sharp, smooth spines. The columns of ad- joining rays meet in the axils and continue into the interbrachial areas. The adambulacral plates are like the inframarginals but do not increase so rapidly in width as those plates and are somewhat more numerous, there being in a complete ray about 52 in a column. Proximally some of the large plates are broken into two or three often very convex ossicles. In addition to the small spines like those of the inframarginals, each plate bears on its ambulacral edge a prom- inent socket in which articulated a long, thick, smooth spine. The adambulacral plates in the proximal third of the rays each have a small, more or less wedge-shaped, carinated, accessory adambulacral plate. In this region these ossicles separate the adambulacrals from the ambulacrals and beneath the accessory adambulacral plates are the podial openings. Five or six adambulacral plates of each column abut against the interbrachial marginal plates, while five or six pairs continue orally, the terminating pieces being of the oral armature. Ambulacral furrows wide, increasing in width proximally. An angular median sinus marks each ambulacral furrow. Ambulacral plates short but wide, slightly overlapping proximally, and one to each adambulacral plate. All of the plates are strongly carinated, the carina in the distal two-thirds converging medially and proximally. In the proximal third the carinse gradually are changed in form and are there arranged in forked pairs. In other words, the carinae adjoin the accessory adambulacrals, are arranged in pairs, continuing ad- joining and straight for a short distance, and then diverge in broad curves, terminating near the center of the ambulacral groove. This arrangement produces in the proximal portion of the ambulacral furrows four columns of ovoid pits, all terminating laterally into podial openings. Every other podial opening belongs to the same column, one series beneath the adambulacrals, the other gradually passing medially in the ambulacral grooves. In the distal two-thirds, however, the podial openings are arranged in single columns, one to each ambulacral plate and issuing from underneath the adambulacrals. Interbrachial areas with seven plates, each composed of two prox- imal inframarginals, beneath which are two pairs of interbrachial marginals and a single diamond-shaped axillary marginal plate. All bear similar spines and ornamentation on the inframarginals. Some- times one of the distal interbrachial marginals may be divided and REVISION OF PALEOZOIC STELLEEOIDEA. 125 the arrangement of the interbrachial plates made irregular. In a specimen from Jefferson County, Indiana, there are 3 pairs of inter- brachial marginals instead of 2 as in the Ohio individuals. There- fore there are also more adambulacrals here, 14 pairs against 10 in the typical specimens. Formation and locality. Two free specimens of this magnificent starfish were found in the Richmond formation near Waynesville, Ohio. Originally pieces of these two individuals were glued together as one specimen, which is the cause of our figures being less perfect than Mr. Miller's. Fragments of seven other individuals were found in the vicinity of Waynesville. In the University of Chicago col- lection (No. 10981) there is a specimen from Jefferson County, Indiana, apparently of this species. Mr. Vaupel secured the distal portion of a ray of an apparently young example of this species in the Maysville beds on Rons Hill, Cincinnati, Ohio, and Mr. Ulrich has fragments of four individuals from the Maysville strata about Cincinnati, and Covington, Kentucky. Remarks. This species is readily separated from the other species of Promopalxaster by the marked diagonal rows of abactinal plates. Actinally P. magnificus is readily distinguished from P. exculptus and P. spinosus in having five instead of three interbrachial marginal plates in each area, in the rapidly increasing size of the plates in the inframarginal and adambulacral columns, and in the proximal am- bulacral plates. P. speciosus differs at once from P. magnificus in that its abactinal plates are arranged in regular longitudinal columns, are far less numerous, larger and are all nearly of a size. This is one of the largest and best preserved of American Paleozoic starfishes. P. dyeri may be a larger species but is distinguished abac- tinally not only by the arrangement and difference in the forms of the plates but also by the short, thick, blunt spines which now ap- pear to have no definite arrangement, while P. magnificus has all of its very fine spines arranged in diagonal rows like the plates. Cat. Nos. 40883, 60621, 60622, U.S.N.M. new subfamily. Aberrant Promopalaeasteridae with the axillary and interbrachial areas composed entirely of adambulacral pieces. Contains : Anorihaster. ANORTHASTER, new genus. Plate 13, fig. 4; plate 20, fig. 1. Anorthaster SCHUCHERT, Fossilium Catalogus, Animalia, pt. 3, April, 1914, p. 11. A + orihos + aster = out of the regular, in reference to the completely adambulacral na-ture of the interbrachial areas. 126 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Generic characters. Disk small, with apparently small interbrachial arcs. Rays five, of moderate length and tapering at first slowly and then rapidly distally. Abactinai area unknown. Inframarginal plates small, the columns not attaining the axils, but resting upon the second enlarged axillary adambulacral plate. Adambulacral plates distally like the inframarginals, increasing in size slowly proximally until near the axillary region, where four plates suddenly attain great width, terminating in the axils of the rays and here suppressing the inframarginal plates. Eight inter- brachial adambulacrals of varying form are situated orally to the two large axillary adambulacral plates, and at first give one the impression of slender interbrachial areas of the same nature as in Promopalszaster. These plates, however, are not interbrachial marginals, as in other genera, and derived by the oral crowding of proximal inframarginals. They are distinctly adambulacral ossicles, forming slight interbrachial areas, and, with the axillary adambulacrals, form small arcs. For the shape and position of these plates see figure 1, plate 20. Ambulacral plates alternating, slightly carinated, and very short but wide. In the distal region they are as long as the adambulacral plates, but proximally decrease in length where there are nearly two of them to one adambulacral ossicle. Podial openings between every other plate through the sutures of adjoining pieces at their outer edge and immediately beneath the inner edge of the adambulacral ossicles. The proximal plates of each column are modified into oral armature pieces that are slightly tumid, as long as wide, and sub- triangular in outline. Between the proximal ambulacrals and orad to the proximal interbrachial adambulacrals are situated additional single, very small, quadrangular oral armature plates. Genoholotype and only species. Palseaster miamiensis Miller. Distribution. Ordovicic, Richmond formation, near Waynesville, Ohio. Remarks. The suppression of the inframarginal plates in the axils of the rays by two of the adambulacral plates of each column and the oral continuation beneath the latter of four pairs of adambulacral plates holding an interbrachial position distinguish this genus fromPromopalseaster, to which it otherwise seems very closely related. This character also distinguishes it from all other associated genera. The interbrachial structure of Anorthaster appears abnormal. In Mesopalseaster and Promopalseaster the proximal inframarginal plates are crowded orally and form interbrachial areas; in Anorthaster their formation is effected in the opposite way, not by the inframarginals, but by the adambulacrals. This character is seen in a single speci- men, but it is so regularly developed and well preserved as to indicate a normal and not a pathologic or abnormal condition. KEVISION OF PALEOZOIC STELLEROIDEA. 127 ANORTHASTER MIAMIENSIS (Miller). Plate 13, fig. 4; plate 20, fig. 1. Palseaster miamiemis MILLER, Journ. Cincinnati Soc. Nat. Hist., vol. 3, 1880, p. 143, pi. 4, fig. 3. Original description. " Pentagonal; rays about one and a half times the diameter of the body, or about nine-tenths of an inch; diameter of the body about six-tenths of an inch; breadth of a ray at the point of junction with the body a little more than half the diameter of the body, or about seven-twentieths of an inch; rays obtusely pointed. "Marginal plates wider than long, and numbering about 12 in the length of half an inch from the body. Two marginal plates form the junction of the rays. Ambulacral furrow wide, the plates being more than twice as long as wide. There are about 18 ambu- lacral plates in a length of one-half inch, and each one is provided with an angular ridge taperingfrom the marginal plates to the furrow." Emended description. R = 23 mm., r = 7 mm., R = 3.2r. Width of rays at base 8 mm. Rays tapering very slowly in the proximal half, but more rapidly distally. Abactinal area unknown. Inframarginal plates about 16 in each column and subquadr angular in outline. These ossicles do not continue into the axils, but rest upon the second large axillary adambulacral plate. Adambulacral plates about 18 in each column, like the infra- marginal pieces, increasing in size proximally, while the two axillary ones are suddenly enlarged, extend to the margin of the disk, and have the position of marginal ossicles. Orally to the latter the adambu- lacral plates continue as interbrachial ossicles. The distal pair is large and both are wedge-shaped, followed by two pairs of quite small quadrangular plates and a final pair of long and narrow plates; in all there are, therefore, 22 adambulacral plates in each column. The interbrachial adambulacral plates are terminated by a single quadrangular oral plate wedged in between the proximal ambu- a ILbulacral furrows very wide, broadly gutter-shaped. Ambu- lacral plates about 30 in a column, slightly alternating, decidedly wider than long proximally, increasing in length distally, and finally becoming about as wide as long near the apex of a ray. Each plate is slightly carinated, the ridges crossing the plates from the lateral distal edge to the medial proximal edge. The proximal ambulacral plates are modified, conspicuous, and subtriangular in outline Podial openings between two adjoining plates in alternate sutures at 50601 Bull. 8815 9 128 BULLETIN- 88, UNITED STATES NATIONAL MUSEUM. their extreme outer edges and immediately inside of the adambu- lacral plates. Formation and locality. In the Richmond formation, near Waynes- ville, in Montgomery County, Ohio. The holotype is in the Harris collection of the United States National Museum. Eemarks. -This species has a general resemblance to some of the Promopalseasters, particularly P. ~bellulus and P. wylcoffi, but the peculiar arrangement of the axillary and interbrachial adambulacral plates will distinguish it not only from them but from all other Ordovicic asterids. Cat. No. 40880, U.S.N.M. Family XENASTERID^E Schondorf. Xenasteridx GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 346; Treat. Zool., vol. 3, Echinoderma, 1900, p. 250. Palsegoniasteridse STURTZ, Palaeontographica, vol. 36, 1890, p. 247 (not opp. 246). Xenasteridse SCHONDORF, Palaeontographica, vol. 56, 1909, p. 105; Jahrb. nassau- isch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 25; vol. 63, 1910, pp. 244, 250. Progressive Phanerozonia having the general characters of the Promopalseasterinse, with the addition of accessory interbrachials. The ambulacral and adambulacral plates are directly opposite one another and do not waver between the slightly alternate and opposite arrangement in the Promopalaeasteringe. Podial openings in two columns in each ambulacral furrow, the apertures lying between each four adjoining plates, that is, between two adambulacrals and two ambulacrals. Contains : Xenaster Simonovitsch. Agalmaster Schondorf. Khenaster Schondorf. Eifelaster Schondorf. Trimeraster Schondorf. Schondorf defines the family as follows (leaving out much of his detail) : Lower Devonic starfishes with well-developed supramarginals and usually still larger inf ramarginals ; the latter are somewhat more numer- ous than the former, causing them to be unequally superposed. Ambu- lacra opposite one another and opposite the adambulacrals; podial openings in two rows. Actinally each interradius has either three or five interbrachial plates; abactinally this area is filled with small accessory plates and margined only by inframarginals. Disk skeleton consisting of small separated pieces, a central disk plate, five primary radialia, five primary interradialia, and more or less of accessory disk KEVISION OF PALEOZOIC STELLEEOIDEA. 129 ossicles. Eays with three columns of plates, one radial and two supramarginal, that may or may not be separated from one another by small accessory pieces. Madreporite dorsal and interradial. Genus XENASTER Simonovitsch emend. Sehondorf. Plate 24. Archseasterias MULLER, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 12, 1855, pp. 6, 8 (Archxasterias founded in error). Xenaster SIMONOVITSCH, Sitzb. mat.-naturw. Classe Akad. Wiss., Wien. vol. 64, Abt. 1, 1871, p. 88. SCHONDORF, Palseontographica, vol. 56, 1909, pp. 42, 105 (complete synonymy given here); Jahrb.nassauisch. Ver. Naturk., Wies- baden, vol. 62, 1909, p. 26. Generic characters. Xenaster has its nearest relationship in Devonaster. No detailed description need be given, since the splendid illustrations of Sehondorf give all the necessary characters. It differs from Devonaster as follows: In Devonaster there is but a single interbrachial plate in each actinal axil, the axillary interbrachial, while in Xenaster there are two pairs of interbrachial marginals and a single axillary inter- brachial plate; in other words, five interbrachial plates instead of one (1+2+2). In addition to these plates in Xenaster, the interbra- chials are more or less surrounded by small accessory pieces, and then other similar plates separate the inf ramarginal and adambulacral col- umns to near the outer third of the rays. The inf ramarginal plates in Xenaster do not increase so rapidly in width proximally as do these plates in Devonaster, and while the interbrachial areas are of about the same size in both genera, the space occupied by the additional interbrachial marginal and accessory interbrachial ossicles in the former is taken up in the latter genus by the increased size of the inf ramarginal pieces. Abactinally Xenaster has the generic characters of Devonaster. The plates in all the columns of the former genus are more prominent and rounder, particularly the accessory plates, of which there appear to be less than in Xenaster. The center of the disk in both genera is occupied by a small central disk plate, but there are many more accessory plates around it in Devonaster than in Xenaster. In both genera the supramarginal columns converge proximally, producing small disk areas between the infra- and supramarginal plates occu- pied by ambital accessory plates. Madreporite large and striate in Xenaster. Genotype. X. margaritatus Simonovitsch. 130 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Distribution. Restricted to the Lower Devonic of Germany. The species are: X. margaritatus Shnonovitsch. X. dispar Schondorf. X. elegans Schondorf. X. (?) rhenanus (Miiller). Remarks. Miiller in his description of Asterias rtienana, 1 the geno- type of Archseasterias , notes that the adambulacral plates of adjoining columns alternate and that the ambulacral columns ' ' have the appear- ance also of alternating/' which if established would be very remark- able, and would justify the creation of a distinct genus Archseasterias for this asterid of the Rhine grauwacke. Schondorf shows that these statements of Miiller' s are very faulty and further that the genus Archseasterias has not been denned so as to be recognizable. Under these circumstances it was best for Simonovitsch to disre- gard Archseasterias and to establish a new generic name for his species margaritatus. There is no Paleozoic genus of starfishes better described or illustrated than this one by Simonovitsch, and should it ever prove that Asterias rhenana Miiller is identical with Xenaster margaritatus, a possibility, the latter should not be made to give way to the former on the ground of priority. It is an injustice to extend the law of priority indiscriminately to all work, and to reject Archseasterias is no reflection on the work of Miiller, as his material was very poor and his work of early date in paleontology. He pub- lished at a time when it was very desirable to learn anything what- ever about Paleozoic starfishes. As pointed out above, Xenaster has much of the structure of Devonaster, yet it differs in the important character of having numerous accessory interbrachial plates. This feature seemingly relates it with forms like Palasterina which have a great abundance of accessory interbrachials, yet in no interbrachial area of this genus are there axillary interbrachials or interbrachial marginals. In other words, the Palasterina type of disk is made by the intercalation of accessory plates between the inframarginals and adambulacrals, whereas hi Devonaster it is accomplished by the greater width of the inframarginals and by the oral crowding of a plate which in primitive forms is an axillary marginal. Devonaster although from younger strata is more primitive than Xenaster because of its simpler inter- brachial actinal structure. If it were not for the accessory inter- brachials, Xenaster would be closely related to Promopalseaster which also has a number of proximal inframarginals crowded into the inter- brachial areas. It is the Promopalseaster stock out of which Xenaster probably developed, while Devonaster apparently came through Mesopalseaster. * Verb, naturh. Ver. preuss. Rheinl., etc., vol. 12, 1855, pp. 6, 8. REVISION" OF PALEOZOIC STELLEROIDEA. 131 XENASTER MARGARITATUS Simonovitsch emend. Schondorf. Plate 24, figs. 1,2. Xenaster margaritatus (part) SIMONOVITSCH, Sitzb. mat.-naturw. Classe Akad. Wiss., Wien, vol. 64, Abt. 1, 1871, p. 88, pis. 1, 2. SCHONDORF, Palseon- tographica, vol. 56, 1909, pp. 43, 106, pi. 7, figs. 1, 3, 4; pi. 8, fig. 1; pi. 11, figs. 2, 3, 10, 11 (complete synonymy given here); Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 26, pi. 2, figs. 1, 2. SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palaeontogr. Soc. for 1913), 1914, p. 30. This species is one of the best described and illustrated of Paleozoic starfishes and does credit to its author, and to its subsequent reviser, Schondorf. It occurs rarely in the Lower Devonic (Upper Coblen- zian) as natural molds at Niederlahnstein on the Rhine, Germany. XENASTER DISPAR Schondorf. Xenaster margaritatus (part) SIMONOVITSCH, Sitzb. mat.-naturw. Classe Akad. Wiss., Wien, vol. 64, 1871, Abt. 1, pi. 2, figs. 2, 2a, 26. Xenaster dispar SCHONDORF, Palaeontographica, vol. 56, 1909, pp. 50, 84, 106, pi. 9, fig. 1; pi. 11, fig. 13; Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 28, pi. 3, fig. 3. From the Upper Coblenzian at Niederlahnstein, and Miellen, Germany. XENASTER ELEGANS Schondorf. Xenaster margaritatus FOLLMANN (not Simonovitsch) , Jahrb. naturh. Ver. preuss. Rheinl., etc., vol. 48, 1891, p. 150. Xenaster elegans SCHONDORF, Palseontographica, vol. 56, 1909, pp. 53, 107, pi. 9, figs. 2, 3; pi. 11, figs. 4, 12; Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 40, pi. 4, figs. 6, 7. From the Lower Devonic (Coblenz quartzite) at Konigstuhl, near Rhens, Germany. XENASTER (?) RHENANUS (Muller). Asterias (Archseasterias} rhenana MULLER, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 12, 1855, p. 6, pi. 1, figs. 4-6. Xenaster (?) rhenana SCHONDORF, Palaeontographica, vol. 56, 1909, pp. 55, 107, text fig. 2 (complete synonymy given here); Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 28. This species is based upon incomplete material and can not be made out fully. It seems probable that it is a species of Xenaster and closely related to X. margaritatus Simonovitsch. The examples are in the Berlin Museum and were found in the Lower Devonic at Winningen, near Coblenz, and Kemmenau, near Ems, Germany. Genus AGALMASTER Schondorf. Text fig. 9. Agalmaster SCHONDORF, Palseontographica, vol. 56, 1909, pp. 58, 107; Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 29. Original remarks. Agalmaster differs from Xenaster in the following characters: Forms larger, broader rayed, with flat actinal but highly arched abactinal plates. Supramarginals much stronger than the large inframarginals and projecting considerably beyond these. 132 BULLETIN 88, UNITED STATES NATIONAL MUSEUM.. \ Abactinal interbrachial area small, and occupied in the main by two axillary inf ramarginals ; actinally this area has the same structure as Xenaster. Madreporite dorsal. Genoholotype (the writer here selects the first species). A. miel- lensis Schondorf. Other species are A. grandis Schondorf and A. intermedius Schondorf, both from the Lower Devonic of Germany. / \ Remarks. The only / \ character that may be of generic value and that at present distin- guishes Agalmaster from Xenaster is that in the former genus the supra- marginals margin the rays instead of the in- f ramarginals, as is the case in the latter genus. In a second species (A. grandis) the small ac- cessory interbrachials crowd in between all the pairs of interbrachial FIG. 9. AGALMASTER MIELLENSIS, AFTER SCHONDORF. SCHEMATIC marginals and further ARRANGEMENT OF ABACTINAL PLATES. R* AND -R 3 , SECOND AND , . ' THIRD PROXIMAL RADIAUA; Jn, PRIMARY INTERRADIALIAJ Cdp, tllCre &TQ J&T ID.OTG OI CENTRAL PLATE; mdp, MADREPORITE; mo, SUPRAMARGINALIA; ^ggQ pieces between mu, INFRAMARGINALIA; Pifm, PRIMARY INTERRADIALIA. the inframargmals and adambulacrals; in fact, there are three to four columns of these in- stead of one (as in A. mieUensis). AGALMASTER MIELLENSIS Schdndorf. Text fig. 9. Agalmaster mieUensis SCHONDORF, Palseontographica, vol. 56, 1909, pp. 58, 107, text fig. 3; pi. 8. fig. 4; pi. 9, fig. 4; pi. 11, figs. 1, 5; Jahrb. nassauisch. Ver.' Naturk., Wiesbaden, vol. 62, 1909, p. 29, pi. 3, fig. 2; pi. 4, figs. 1, 2. A single specimen with the rays about 43 mm. long was found in the Lower Devonic (Upper Coblenzian) at Miellen-on-the-Lahn, Germany. AGALMASTER GRANDIS Schdndorf. Xenaster margaritatus SIMONOVITSCH (part), Sitzb. mat.-naturw. Classe Akad. Wiss., Wien, vol. 64, Abt. 1, 1871, pi. 2, figs. 3, 3a. Agalmaster grandis SCHONDORF, Palseontographica, vol. 56, 1909, pp. 62, 108, pi. 9, fig. 5; Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 29, pi. 3, fig. 1. One specimen (Xenaster margaritatus Simonovitsch, part) from the Lower Devonic (Upper Coblenzian) at Niederlahnstein on the Rhine, Germany. REVISION OF PALEOZOIC STELLEROIDEA. 133 AGALMASTER INTERMEDIUS Schflndorf. Agalmaster intermedius SCHONDORF, Palaeontographica, vol. 56, 1909, pp. 82, 108, pi. 11, fig. 14; Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 30. From the Upper Coblenzian at Miellen-on-the-Lahn, Germany. Genus RHENASTER Sehondorf. Text fig. 10. Rhenaster SCHONDORF, Palseontographica, vol. 56, 1909, pp. 65, 108, pi. 8, fig. 3; pi. 10, figs. 1, 4. Remarks. Rhenaster is a small form and differs from Xenaster, Agalmaster, and Devonaster in the complete absence of all abactinal accessory plates, allow- ing the radial and supra- marginal pieces to ad- join one another closely. Dorsal axillary disk ~7=TP?^O g V\C\ T i I /I )s~*\-/ sr?~^ ^-' /\S. \V areas not completely but almost absent, supramarginals in the main lying wholly upon the inframarginals and the two columns of plates together margin- ing the rays. There also are no accessory plates of any kind on the actinal side. Other- wise the generic struc- ture is that of Xenaster, except that in Rhenaster the first pair of interbrachial marginals are not yet pushed so much orally and they almost take part in margining the axils. Madreporite unknown. Genoholotype and only species (one specimen). R. schwerdi Schon- dorf, from the Lower Devonic (Upper Coblenzian) at Dorrbachtal, near Coblenz, Germany. Genus EIFELASTER Sehondorf. Etfelaster SCHONDORF, Palaeontographica, vol. 56, 1909, pp. 70, 108, pi. 10, figs. 7,8. Remarks. This genus is very closely related to Rhenaster and differs only in that there are small dorsal interbrachial areas between the adjoining proximal portions of the supramarginal columns, and these are filled with a small number of accessory ambital pieces. The outer portion of these ambital areas is made up of the two basal inframarginals that occupy the axils of the ventral side. FIG. 10. REENASTER SCHWEKDI, AFTEK SCHONDORF. SCHEMATIC ARRANGEMENT OF ABACTINAL PLATES. SYMBOLS AS IN FIG. 9. 134 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. The genoholotype and only species, E. fottmanni, occurs in the Lower Devonic (Lower Coblenzian) at Landscheid, near Wittlich, in the Eifel, Germany. Plastotypes of the holotype were made for the writer by Professor Jaekel, of Greifswald, and these are now in the Yale University collection. Genus TRIMERASTER Schondorf. Plate 19, fig. 1. Trimeraster SCHONDORF, Palaeontographica, vol. 56, 1909, pp. 68, 108, pi. 7, fig. 2; pi. 10, figs. 5, 6. Remarks. Trimeraster is a small form, very closely related to Xenaster, and can be distinguished generically only in that the former has inside of the inframarginals but one. pair of interbrachial margi- nals and a single axillary interbrachial in each interbrachial field instead of the five plates of the latter genus. Genoholotype and only species (one specimen). T. parvulus Schon- dorf, from the Lower Devonic (Coblenz quartzite) at Konigstuhl, near Khens, Rhine Valley, Germany. NEOPAL^ASTERIDJE, new family. Progressive Phanerozonia without interbrachial arcs. The columns of supra- and inframarginal plates wholly superposed. Madreporite abactinal. Well-developed ocular plates present. Ambulacrals alter- nating. Disk with a ring of large plates of basal, radial, and inter- radial ossicles. Contains : Neopalseaster, new genus. The origin of this family is uncertain, but it seems to be related to the Promopalseasteridse, since its essential characters are those of the Mesopalaeasterinse. In the latter the supramarginals do not com- pletely overlie the inframarginals making one consolidated column as in Neopalseaster. It is, so far as known, the oldest family developing ocular plates. This fact is of considerable importance for it means that the oculars were originally basal radialia, and that they have progressively wandered from the disk and then remained at the tip of the rays. NEOPALSEASTER, new genus. Plate 13, fig. 5; pi. 23, fig. 4. Neopalseaster SCHUCHERT, Fossilium Catalogue, Animalia, pt. 3, April, 1914, p. 26. Generic characters. Disk of medium size, without interbrachial arcs. Eays five, short and tapering rapidly. Abactinally the disk has a small central plate surrounded by numerous smaller plates. Bounding the smaller accessory disk KEVISION OF PALEOZOIC STELLEEOIDEA. 135 plates is a ring of very large, thick, tumid, variously shaped, radial and interradial plates. Of the latter there may be one plate inside of the basal supramarginals of adjoining columns, or this plate may be divided, when the pieces appear as inwardly crowded supramar- ginals. Laterally and distally upon the interradials are situated the large basal radial plates. The rays are bounded laterally by columns of large, thick, subquadrate, tubercular supramarginals terminated dis- tally by single ocular plates. These plates appear to be common to both the actinal and abactinal areas and must therefore be the united infra- and supramarginal columns. Between the supramarginals dis- tally appear single small isolated plates which proximally become larger and continuous, forming an inconspicuous column of radial ossicles. On each side of this column are inserted a number of appar- ently irregularly arranged small accessory plates which are like those of the disk. All of the plates of the abactinal area are finely tuber- culose, these points being for the attachment of very fine, short spines. Madreporite small, rounded, not very convex, very finely radially striated and resting directly upon two basal supramarginal plates of adjoining rays. Ocular plates small but distinct, one terminating each ray and rest- ing against the distal marginal plates. Supra- and inframarginal plates large and conspicuous, apparently firmly united and indistinguishable laterally as separate columns. Abactinally the supramarginal plates are wider and overlie more or less the adambulacrals. Adambulacrals smaller than the inframarginals but otherwise resemble the latter. Ambulacral grooves narrow. Ambulacra! plates alternating, about as numerous as the adambulacrals, 1 -shaped, with the podial open- ings between adjoining plates. All other actinal characters unknown, Genoholotype and only Jcnown species. Palseaster crawfordsvillensis Miller. From the Keokuk crinid beds near Crawfordsville, Indiana. Remarks. Neopalseaster retains a very primitive abactinal plate structure and calls to mind Hudsonaster, Palseaster, and typical Mesopalseaster. It is distinguished from these early Paleozoic genera by the complete consolidation of the supra- and inframarginal plates into one column, and the presence of ocular plates. The inter- brachial area is unknown in Neopalseaster, but its abactinal structure seems to indicate a single interbrachial plate in each area bounded distally by two basal inframarginal plates. If this interpretation is correct, then this character will further distinguish Neopalseaster from Hudsonaster and Palseaster. 136 BULLETIN" 88, UNITED STATES NATIONAL MUSEUM. NEOPAL^EASTER CRAWFORD SVILLENSIS (Miller). Plate 13, fig. 5; plate 23, fig. 4. Palseaster crawfordsmllensis MILLER, Journ. Cincinnati Soc. Nat. Hist., vol. 2, 1880, p. 256, pi. 15, fig. 3; N. Amer. Geol. Pal., 1889, p. 265, fig. 379. Original description. "This species is founded upon the dorsal view of a single specimen. The rays are longer than the diameter of the body, and not of uniform size. They are flattened or depressed in the middle, as is also the central part of the body. Many of the plates possess a central tubercle or small spine, and probably all of them did. "The marginal plates are large ; somewhat elliptical in outline, and have their shorter diameters in the direction of the length of the rays. There are about 12 plates on each side of a ray, and they come together at about the eighth plate from the body, though in the ray opposite the madreporiform tubercle they come together at the seventh. The space between the marginal plates of each ray is filled with smaller plates; three of these unite the larger plates at the body, but they diminish in number toward the apex of the ray, and cease at the eighth plate. In addition to the two large plates which form the junction of the rays with the body, a few large plates cover the outer part while the central part is covered by smaller plates. The madreporiform tubercle is supported by three plates, two of them are large marginal plates, which form a junction between two rays, and the other is a large plate within, forming part of the covering of the body." Emended description. The type-specimen measures: R, = 22 mm., r=7 mm., K, = 3.1r > . Width of rays at base 7 mm. Another speci- men in the Museum of Comparative Zoology has rays 3 mm. longer. Rays short, tapering rapidly. Disk rather large for a small species. Abactinally the disk has a central circular plate surrounded by numerous small, conical, irregularly disposed, accessory plates, among which are a few larger pieces usually adjacent to the large interradial plates. Outside of the small disk plates is a ring of very large, irregu- larly shaped, radial and interradial plates. Five of these are inter- radial in position, and are situated just within the basal plates of adjoining supramarginal columns. These plates may remain normal as a single plate or may be divided into two plates, when they some- what resemble inwardly crowded supramarginals. Upon these laterally and distally rest five large radial pieces, the basal plates of the radial columns. It is very probable that the radial columns are continuous from the basal plate to near the distal portion of the ray, where a few of these plates are much reduced in size and are EEVISIOIST OF PALEOZOIC STELLEEOIDEA. 137 separated and wedged in between the adjoining four supramarginal ossicles. There appear to be 10 or 11 of these in a column. On each side of the radial columns in the proximal half of each ray are a number of small accessory plates arranged in irregular columns. All of the abactinal plates are more or less convex and bear numerous small pustules clustered around a larger central one, all of which bore small spines. Supramarginal plates very conspicuous, subquadr angular in out- line, generally a little wider than long, with numerous small pustules, among which toward each lateral edge is a large one upon which probably articulated prominent spines. There are from 11 to 13 of these plates in each column. In the distal third of a ray both columns approximate and adjoin each other closely. Each supra- marginal plate lies wholly and directly over an inframarginal and the pieces appear as one consolidated plate with the abactinal side convex and the actinal nearly flat. The supramarginals are the widest and extend in part over the adambulacrals. Madreporite small, circular in outline, slightly convex and very finely radially striated. It lies directly upon two basal supramar- ginal plates of adjoining rays and directly outside of one of the large interradial plates. An ocular plate terminates each ray. Inframarginals apparently like the supramarginals excepting distally, where they are nearly twice as long as wide. Adambulacral plates smaller than the inframarginals, nine in the space occupied by five of the latter, somewhat wider than long, convex, and covered with numerous small pustules. Ambulacral grooves very narrow in the distal region. The ambu- lacral plates appear to be as numerous as the adambulacrals, and are h- -shaped, with the podial openings between the plates in about the mid-width of the column. All other actinal characters are unknown, since the two specimens expose the abactinal area, though one has portions of the plates removed, revealing the actinal characters described. Formation and locality. The type-specimen is in the Harris collection of the United States National Museum, and another specimen is in the Dyer collection of the Museum of Comparative Zoology (No. 4). Both were found in the crinid beds of the Keokuk group of the Lower Carboniferous at Crawfordsyille, Indiana. Remarks. There is no Carboniferous asterid with which Neopalse- aster crawfordsvillensis can be confounded. Cat. No. 60607, U.S.N.M. 138 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Family PALASTERINID^E Gregory (emend.). Pateasterinidse GREGORY (part), Geol. Mag., dec. 4, vol. 6, 1899, p. 348 (contains Palasterina, Schaenaster, and Schuchertia)} Treat. Zool., vol. 3, Echinoderma, 1900, p. 250. Lindstromasterinss GREGORY (part), Geol. Mag., dec. 4, vol. 6, 1899, p. 346 (contains Lindstromaster and Uranaster). Palseasterinidse STURTZ, Palaeontographica, vol. 36, 1890, p. 246. Progressive Phanerozonia developing large interbrachial arcs. Ambulacral plates more or less alternate. Madreporite abactinal. Disk large, pentagonal, and the rays separated by well-developed interbrachial arcs. The inframarginal plates bound the animal and are separated from the adambulacrals more or less completely by a varying number of interbrachial plates. Abactinal surface with longitudinal columns of radial, supra- marginal and accessory columns of plates, or the radial columns may be reduced or even replaced by transverse rows of small accessory pieces. The central region of the disk may retain a ring of larger basal radial and supramarginal pieces. Contains: Pet raster Billings. Lindstromaster Gregory. Palasterina McCoy. Uranaster Gregory. Palxostella Stiirtz. Pseudopalasterina Stiirtz. Genus PETRASTER Billings. Plates 23, 26, 27. Petraster BILLINGS, Geol. Surv. Canada, Can. Org. Rem., dec. 3, 1858, p. 79. HALL, Twentieth Rep. N. Y. State Cab.. Nat. Hist., 1868, pp. 283, 294; rev. ed., 1868=1870, pp. 325, 337. Palseaster (part) ZITTEL, Handb. Pal., vol. 1, 1879, p. 452. PalsRosterina of some AUTHORS. Generic characters. Disk and interbrachial arches more or less large. General shape pentagonal. Rays five, short and slender or stout and tapering rapidly. Abactinal area of rays distally with one medial radial, two supra- marginal and two inframarginal columns of plates. Between the radial and supramarginal columns proximally are inserted columns of accessory plates' In P. speciosus there are five on each side of the radial column. Between the supramarginal and iinVaniarginal columns in the axillary and basal ray regions are inserted columns of ambital plates. In P. speciosus there are three. The plates are more or less stellate, with numerous subcircular openings in the angles between adjoining plates. REVISION OF PALEOZOIC STELLEKOIDEA. 139 Inframarginal plates bounding the entire outer edge of the animal, conspicuous, closely adjoining, not stellate and common to both the actinal and abactinal areas. Madreporite radially striate, abactinal. Adambulacrals closely adjoining the inframarginals in the distal portion of the rays but proximally the former are separated from the latter by more or less numerous, small, accessory interbrachial plates. ' The adambulacrals may be of nearly the same size, or considerably wider than long in the mid-length of the column. The adambula- crals of adjoining columns meet orally in a pair of larger triangular plates, the oral armature. Interbrachial areas more or less large and the space between the inframarginal and adambulacral columns filled with a variable number of accessory interbrachial plates. Ambulacrals slightly alternating, about as numerous as the adam- bulacrals, each with an L-shaped, proximally converging carina. Podial openings through the outer proximal corner of the plate and proximal to the carina. Grooves narrow, regularly tapering. GenoJiolotype (type by monotypy). Petraster rigidus Billings. The above diagnosis is based on the genotype and Pal&asterina spedosa Miller and Dyer. Distribution. Ordovicic of North America, Siluric of Australia. The species are: P. rigidus (Billings). Trenton. P. speciosus (Miller and Dyer). Maysville, Kichmond. P. (?) americanus (D'Orbigny). Maysville. P. smythi McCoy. Siluric of Australia. 'Remarks. While Billings has correctly pointed out the diagnostic characters of Petraster, yet the genus has been very imperfectly understood. This may partly be accounted for in that he later also included a specimen of Hudsonaster matutinus (Hall), which of course fails to have' the essential generic character of Petraster, namely, the interbrachial accessory plates. The following is Billings' original description: "This genus has both marginal and adambulacral plates, with a few disk-plates on the ventral side. The general form is deeply stellate, and the rays long and uniformly tapering. A single specimen has been collected, and as it shows the underside only, the characters of the dorsal surface can not be given. The structure of the mouth is also unknown. Generic name from petra, a stone. "It differs from Palasterina by the presence of large marginal plates outside of the disk-plates [that genus has small marginals], and still more from Stenaster, which has neither disk nor marginal plates. It is allied to Astropecten." 140 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Professor Hall at first recognized the distinctness of Petraster, since he wrote as follows (1870): "The character of Petrasier here described, and illustrated in the figure (3a, pi. 9), gives an intercalated partial range of disk-plates, between the adambulacral and marginal plates, which will separate these forms from any of the Palaeasters in my collection.' 7 On a later page, after an examination of the genotype, he added a "Note on the genus Petraster" in which he concludes that the specimen was injured during life, as shown by the unequal distribu- tion of the accessory interbrachial plates, and further that one ray is "without intercalated plates on either side.' 7 He adds: "This yiew is sustained by the fact that the other parts have the ordinary structure of Palseaster, and in all other respects the specimen agrees with the typical Palseaster matutina." The genotype was studied at Ottawa by the present writer and found to be as described by Billings. The accessory interbrachial plates, though somewhat irregular in development in different areas in the type species of Petraster, are normal in position and a similar but more complete development also appears in Lindstromaster and Palasterina. Petraster is distinguished from Palasterina as follows: The most important feature is that the inframarginal plates of the former are prominent, closely adjoining and short but wide, while in Palas- terina they are small, not conspicuous, and globular, or subquadrate. The next important difference is on the abactinal disk, where in Palasterina there is an irregular ring of prominent basal radial, and supramarginal plates. These are absent in Petraster. The acces- sory interbrachial ossicles in the latter genus never attain the distal region of the rays as in Palasterina where these pieces are also more numerous. Palasterina is also more abundantly spinose than Petraster. Schuchertia has no inframarginals as marginal plates, which at once distinguishes it from both Petraster and Palasterina. It is true that in Schuchertia inframarginals are also present but they are usually not well developed and remain adjoining the adambulacrals. These two columns are therefore not separated from one another by accessory interbrachial plates as in Petraster and Palasterina. The primary structure of Petraster is that of Hudsonaster. So many of the generic characters are common to both that the former genus appears almost certainly to have been derived through the latter. It is true that the central abactinal area of the disk in Petraster is devoid of the large basal plates, a marked character in Hudsonaster, but the other generic characters are all present in the former. Petras- ter adds more or less numerous columns of radial accessory ambital and accessory interbrachial plates, none of which are present in Hudsonaster. It is these accessory ossicles which differentiate these OF PALEOZOIC STELLEKOIDEA. 141 genera and show the line of development from the deeply stellate primary form to the pentagonal genera with well-developed inter- brachial arcs. PETRASTER RIGIDUS (Billings). Plate 27, fig. 5. Palasterina rigidus BILLINGS, Geol. Surv. Canada, Rep. of Progress, 1853-1856, 1857, p. 291. Petraster rigidus BILLINGS, Geol. Surv. Canada, Can. Org. Hem., dec. 3, 1858, p. 80, pi. 10, fig. 3a (not fig. 3bHudsonaster matutinus} . WRIGHT, Mon. British Foss. Echinod., Oolitic, vol. 2, pt. 1 (Palaeontogr. Soc. for 1861), 1862, p. 29. HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 294; rev. ed., 1868=1870," p. 337. Description of 1858. "This species has much the aspect of an Astropecten; the disk is one-fourth the whole diameter, the rays rather slender and uniformly tapering, the angles between the bases of the rays rounded. The plates [of the actlnal side] which appear to be adambulacral [increase very little in size from the tips of the rays toward the mouth], are quadrate and a little convex; [the adambula- cral columns terminated in the mouth area by a pair of pointed oral plates and not by a single plate as shown in the original illustra- tion], the marginal [inframarginal] plates oblong, and also convex [certainly not less than 16 and probably 20 in each column, increasing rapidly in size toward the axils, whsre there is a single large axillary plate] ; the disk plates [accessory interbrachials] consist of three at each angle [one orally and two distally], and a single row [of not more than seven plates] on each side of the ray, but extending only one- third or one-half of the length of the ray; they all lie between the [infra] marginal and adambulacral plates. [Abactinal side unknown.] The specimen figured was about 2 inches [or 50 mm.-] in diameter when perfect ; width of disk half an inch, and of rays at the base about three lines." Formation and locality. Trenton limestone, Ottawa, Canada. Holotype No. 140 la is in the Victoria Memorial Museum, Ottawa. The species has also been identified by Springer in the Lower Trenton (Kirkfield) at Kirkfield, Ontario. Remarks. Hall and Billings discussed their asterid genera and species at different times and finally the former examined Billings' s material. In this connection Hall showed that figure 3& of Petraster rigidus was based on the actinal side of Hudsonaster matutinus. Regarding figure 3a, which is the holotype of this species, he in 1870 wrote as follows: "The specimen illustrated in figure 3a has a few small intercalated plates between the marginal and ambulacral [adambulacral] ranges in two of the axils of the rays, and there are a smaller number of gran- ules in a similar position but unequally distributed on one side of 142 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. another axil, while two of the axils do not show any such intercalated plates or granules. In one of the rays, at least, the ambulacral, adam- bulacral, and marginal plates are distinctly visible, without inter- calated plates on either side. "The presence of these unequally distributed plates or granules is apparently an abnormal structure, probably the result of accident during the growth of the animal; and this view is sustained by the fact that the other parts have the ordinary structure of Palseaster, and in all other respects the specimen agrees with the typical Palse- aster matutina" The holotype was examined by the writer and although imperfect was found to have a normal development, with the plates in their proper places in at least two of the rays and axils. It certainly is not an abnormal development of "Palseaster matutina" but represents a normal individual of a distinct genus and species diverging from Eudsonaster matutinus toward the forms with large disks and great axillary areas, as in Lindstromaster and Palasterina. Hall seemingly was led to make the erroneous remarks above quoted through the fact that Billings had also included in his description a'nd illustration (36) an individual of Hudsonaster matutinus, an error due to the undoubtedly close relationship of this form with Petraster rigidus. PETRASTER SPECIOSUS (Miller and Dyer). Plate 23, figs. 5-7; plate 26, fig. 1; plate 27, figs. 1-4. Palssasterina spedosa MILLER and DYER, Journ. Cincinnati Soc. Nat. Hist., vol. 1, 1878, p. 30, pi. 1, fig. 7. MILLER, N. Amer. Geol. Pal., 1889, p. 266, fig. 381. Paldeasterina approximata MILLER and DYER, Journ. Cincinnati Soc. Nat. Hist., vol. 1, 1878, p. 30, pi. 1, fig. 8. Original description of Palseasterina spedosa. "Pentagonal; rays obtuse at their apices; greatest distance from point to point about 2J inches; breadth of body between rays about 1J inches, and dis- tance from tip of ray to next adjoining tip on either side about 1J inches. "The [infra] marginal plates are small and somewhat hemispher- ical near the termination of the rays, they gradually enlarge and become square, and then rectangular as they approach the disk, until at the narrowest part of the disk or body of the fossil they are twice as long as wide. There are about 50 marginal plates between the apex of one ray and the next one adjoining, or, in a perfect specimen of this size, about 250. "The back or dorsal side is covered with numerous plates (prob- ably in a complete specimen of this size there would be 1,000 or more), which are very prominent in the center or somewhat conical, and seem to have been joined together with deeply serrated edges. The plates have from three to eight of these indentations,, which give them a beautiful starlike appearance. KE VISION" OF PALEOZOIC STELLEROIDEA. 143 "The ambulacral grooves are narrow and deep, as shown by the sharp ridges on the back of the specimen. The small dorsal plates which cover the ambulacral pieces are exfoliated in some places, and show two rows of ambulacral plates coming evenly together and forming the sharp ridge.'' Original description of Palseasterina approximate, . " Pentagonal; rays more slender than in P. speciosa; greatest distance from point to point in the specimen figured 1.15 inches and breadth of the body one-half inch; another specimen from the collection of Mrs. M. P. Haines, of Richmond, Ind., measures from point to point 2.5 inches, and has a breadth of body of nine-tenths inch. "The [infra] marginal plates have about the same form, and are about as numerous as they are in P. speciosa. The adambulacral plates are arranged with their length across the rays. The oral plates are very prominent. [This statement is somewhat in error. The five prominent points around the mouth are each formed of one large interbrachial plate and a pair of proximal adambulacrals.] The space between the [infra] marginal plates and the oral ones [adambu- lacrals] is filled with numerous small [accessory interbrachial] pieces. The ambulacral grooves are narrow. Ambulacral ossicles unknown. "The dorsal side is covered with small conical pieces, which give it a coarsely granular appearance. The madreporiform tubercle is prominent, conical, and longitudinally striated. "This species is distinguished from P. speciosa by its narrower rays, more contracted body, and smaller dorsal plates." Emended description. The holotypo of P. speciosa measures: R = 33 mm., r 18 mm. The holotype of P. approximata measures: R = 15 mm., r = 6.5 mm. Rays in young specimens slender distally, but with age they become obtuse and very wide proximally and here merge into the large disk. Interbrachial arcs very large. The abactinal area of disk and rays is bounded by columns of very prominent inframarginal plates. Beginning at the apex of each ray they are very small and globular, soon they become quadrate and gradually shorter but wider, so that near the middle of each interbra- chial arc they have a crowded appearance. In a full-grown specimen there are about 50 and hi a half-grown example about 45 of these plates between the distal ends of adjoining rays. In the distal region these plates are rounded and nearly smooth but gradually become granular and strongly convex, each with one prominent point for an articulating spine, none of which, however, have been observed. On the actinal side these inframarginals are also very prominent and pustulose but here are flat. They are most prominent near the center of the interbrachial arcs. 50601 Bull. 88 15 10 144 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. The supramarginal columns begin at the very apex of the rays and for more than one- third of their length adjoin the inframarginals where ambital plates begin to appear. The supramarginal columns are continuous over the disk in a broad curve, with about 18 to 20 in a column, or from apex to apex of adjoining rays there are from 35 to 40 of these plates. In the axillary areas they are separated from the inframarginals by three or four rows of ambital plates. At the apex of the rays the supramarginals are highly convex and nearly circular in outline but soon pass into more and more dis- tinctly stellate plates. This is the form of all the abactinal plates inside of the inframarginals ; they are stellate, highly convex, each with a central node for an articulating spine and a few granules that are the bases for smaller spines. Each radial column appears immediately beneath the two terminal or rather distal supramarginal plates and then continues as a column to near the center of the disk. They are more prominent than the other columns excepting the marginals and have not less than 25 plates in each one. On each side of the radials are five columns of radial accessory plates which appear to continue as columns over the disk in broad curves joining those of the next ray. Between these columns in the axillary areas are additional accessory plates. These columns of accessory plates appear singly, first on one side and later on the other, and not in pairs simultaneously, one on each side of the radial columns. Ambital areas well developed in the axillary region, where there are about three columns of these plates. The columns pinch out rapidly dis tally and none are present in the outer third of the rays. The plate arrangement of the central part of the disk can not be made out. Madreporite of medium size, highly conical, and conspicuous, with numerous sharp single or bifurcating ridges; on the under side are seen two outwardly directed spiral cones (the white lines of the drawing representing the spiral tubes in the madreporite), reminding one of the brachia in Atrypa (pi. 27, fig. 4). Adambulacral plates depressed, convex, subquadrangular in out- line in young specimens, but in fully grown individuals a number of these in the central region of the column are much drawn out laterally and are here two or three times as wide as long. In a half-grown individual there are about 20 of these plates in a column, but in a mature specimen there appear to be not less than 26. These plates end inwardly in an apex, against which terminate the ridges of the ambulacral plates. The adambulacral and inframarginal columns are closely adjoining in the distal two-thirds of the rays but proxi- REVISION OF PALEOZOIC STELLEROIDEA. 145 mally they are separated from each other by numerous accessory interbrachial plates, of which there appear to be not less than 30 in each area. The apexes of these interbrachial areas form the oral armature and each consists of two adambulacral plates, back of which is a rather large interbrachial plate. Ambulacra! grooves narrow, tapering, and deeply V-shaped. Ambulacral plates slightly alternating, about as long as wide, and proximally superposed, with about 26 in each column. Each plate has an L-shaped ridge with the point proximally directed. The podial openings are situated in the outer proximal corner of the plates and between the inner angles of two adambulacrals. Formation and locality. Richmond and Maysville formations of the Upper Ordovicic. The holotype of P. speciosa was found by Mr. W. C. Barnhart, on Twin Creek, near Winchester, in Preble County, Ohio, and is now in the Dyer collection of the Museum of Comparative Zoology, Harvard University (No. 14). Four other less well-preserved specimens are in the Harris collection of the United States National Museum and were found in the vicinity of Waynesville, Ohio. The holotype of P. approximata and another similar-sized specimen but not so well preserved are also from the vicinity of Waynesville, Ohio. The first is in the Dyer collection of the Museum of Comparative Zoology (No. 15), and the other in the United States National Museum. There is a slab with two young individuals much etched with acid but otherwise excellent, showing the actinal side, in the Gurley collection of the University of Chicago (No. 10979). Finally there are two fine specimens, not fully grown, showing well the actinal and abactinal sides, in the Haines collection of the same university (Nos. 10839 and 10840), found near Hamilton, Ohio; these are apparently from the very top of the Maysville formation. Remarks. P. speciosa is described by Miller and Dyer from the abactinal side and P. approximata from the actinal side. These writers point out that the rays are more slender and the disk more contracted in the latter, but when one considers that it is less than half the size of the former and that during growth these forms con- tinually add accessory plates, this difference hi form is seen to be one of growth. The ' 'smaller dorsal plates" noted by these writers in Mrs. Haines's specimen of P. approximata are a character which the present writer does not regard as of value unless they be constant at this lower geologic horizon (Maysville at Hamilton), in which event this specimen can later on be distinguished by another specific name. Cat. No. 60609, U.S.N.M. 146 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. PETRASTER (?) AMERICANTJS (D'Orbigny). Plate 26, fig. 2. Asterias GRAHAM, ANTHONY, and JAMES, Amer. Journ. Sci., eer. 2, vol. 1, 1846, p. 441, figure but no description. Ccelaster americanus D'ORBIGNY, Prodr. de Paleont., vol. 1, 1849, p. 22 (name proposed). Asterias anikonii DANA, Manual Geol., ed. 1, 1863, p. 221, with figure but no description. Palasterina (tyjamesi DANA, Amer. Journ. Sci., ser. 2, vol. 35, 1863, p. 295; Manual Geol., ed. 2, 1864, p. 221, with figure but no description. Palseaster jamesi HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 286; rev. ed., 1868=1870, p. 329. MEEK, Geol. Surv. Ohio, Pal., vol. 1, 1873, p. 62, pi. 4, fig. 4. Meek's description. " Through the kindness of Mr. U. P. James, of Cincinnati, the discoverer of this fossil, I have now before me the original typical specimen, which I found not in a condition to be fully and systematically characterized, nor perhaps to afford the means of determining, beyond doubt, its generic affinities. It evidently attained a large size, and has a comparatively large disk, with its five rays rather broad at their inner ends, somewhat rapidly tapering, and apparently longer than the diameter of the disk. It evidently lies [does lie] in the matrix in such a manner as to show the ventral side [the inner dorsal aspect of the ventral plates], but pre- sents the appearance of having the marginal and disk plates of this side removed [they are present]. The five pairs of oral pieces [proxi- mal adambulacral plates] are seen in place, and proceeding outward from these can be seen a row of adambulacral pieces on each side of each of the rather wide [narrow] ambulacral furrows. These latter pieces are [proximally] nearly three times as long as their diameters in the direction of the length of the rays, and quite promi- nent at their inner ends, where they are articulated together by little processes and corresponding sockets or sinuses; while the outer two-thirds of each is flattened and distinctly more depressed than their inner ends. Outside of, and alternating with, the outer flat- tened ends of these adambulacral pieces, a row of [inframarginal] nearly square, or more or less oblong pieces, may be seen all along the margins of the rays, and connecting at their axilla in such a manner as to leave room for several disk [or interbrachial] plates between them and the oral and inner adambulacral pieces." In each interbrachial area there are inserted between the infra- marginals and adambulacrals interbrachial plates as follows: Proxi- mally one hexagonal plate upon which rest two similar plates mar- gined by the inframarginals. "At some places within the [narrow] ambulacral furrows, remains of ambulacral ossicula can be seen deeply interlocking with the inner ends of the adambulacral pieces, but their exact form, and the REVISION OF PALEOZOIC STELLEROIDEA. 147 position and arrangement of the ambulacra! pores, can not be clearly made out in the specimen studied. "The two oral pieces [proximal adambulacrals] of each of the five pairs are separated by deep sutures, and have an irregular form, being longer than wide, and, like the adambulacral pieces, which they somewhat exceed in size, flattened and depressed at their outer ends, and elevated into crestlike prominences farther in, with a lateral process or thickening on the outer side of each, in which a more or less distinct pit may be seen." Between the pairs of oral adam- bulacral plates are inserted small, single, sub quadrangular plates, the proximal plate of each pair of ambulacral ossicles. Orad to each pair of oral armature plates are situated single hemispheric plates. "The surface, and the arrangement of the dorsal parts, are un- known; but they were probably armed with short spines, as some remains of little spines are seen in the matrix at the margin of the disk, at one place. "As the extremities of none of the rays can "be clearly seen, the exact breadth of the whole fossil, between the ends of opposite rays, can not be exactly given, but it is evident that the length of one of the rays, measuring from the middle of the disk, must have been a little more than 2 inches, and consequently that the diam- eter across from the extremities of opposite rays could not be less than about 4 inches. The breadth of the disk is 1.20 inches; and that of the rays at their inner ends, 0.70 inch." Formation and locality. The single specimen was found in the Maysville formation (Ordovicic), by Mr. U. P. James, on or near Vine Street hill, Cincinnati, Ohio, and is now in the James collection of the Walker Museum of the University of Chicago. Remarks. This large species is quite distinct from any other star- fish in having three rows of interbrachial plates, one piece in the proxi- mal, two in the medial, and five or possibly more in the distal row. While the arrangement of these interbrachial plates is different than hi the other species of Petraster, their position is that of this genus. Therefore this species is for the present referred to Petraster. The preservation of this specimen is such that its generic and specific characters will remain hi obscurity until other material is found. It lies with the actinal side buried in the limestone while all of the abactinal plates have been worn away. What one sees is therefore the inner or dorsal aspect of the ventral skeleton. PETRASTER SMYTHI McCoy. Petraster smythi McCoy, Geol. Surv. Victoria, Prodr. Pal. Victoria, dec. ], 1874, p. 41, pi. 10, figs. 1-16. Original description. "Five broad semielliptical lobes meeting at slightly rounded reentering angles, leaving the length and the 148 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. width at base of the rays nearly equal and less than the width of the disk. The upper surface is covered with crowded irregularly polyg- onal tumid plates. Madreporiform tubercle very large (1J lines in diameter), irregularly porous, and rugged with branching vermicular ridges, excentric toward base of the two posterior rays. Ambulacral groove very narrow, bordered with a row of large transversely oblong adambulacral plates, wider than long, about 6 in 2 lines at middle of ray; margin of the rays bordered with a rather smaller row of similar marginal plates; between the row of adambulacral and mar- ginal plates an intercalary row of small irregular plates. Width of disk between the rays, 7 lines; from tip to tip of rays, about 1 inch 2 lines; length of ray, about 5 lines. " Very rare in the fine sandy Upper Silurian rocks of Moonee Ponds, Flemington, a little north of Melbourne [Australia]. ' ' Remarks. The holotype exposes the abactinal side, but the plates of two of the rays are absent, exposing the actinal skeleton, which is therefore seen from its inner side. So far as one can judge of the description and illustration, the species is a small but genuine Petras- ter. Abactinally the rays have conspicuous radial and supramar- ginal columns, each with about 17 ossicles, that are tumid and are all closely adjoining. Outside of these in the distal portion of the rays are the equally conspicuous inframarginals, and they margin not only these parts but the entire animal; proximally the inframar- ginals separate more and more from the supramarginals, forming small ambital areas, occupied by a number (can not be determined) of rounded plates that now appear not to have been closely adjoining. Madreporite large, radiately striate, situated in an interradius and well inside the ambital area. Ambulacralia large, about 12 in a column, and opposite one another. Adambulacrals smaller, about 14 in a column. Interbrachial struc- ture unknown. Genus LINDSTROMASTER Gregory. Lindstromaster GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, pp. 344, 346. Hisingeraster STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 56, 1900, pp. 224, 225 (same genoholotype as for Lindstromaster). Original description. "Palseasterinidae with flat pentagonal disk and five short blunt rays. ''The plates of the disk are large, polygonal, tuberculate, and crowded into a close tessellate pavement, which completely covers the interradial areas. REVISION OF PALEOZOIC STELLEROIDEA. 149 "Marginal [inframarginal] plates well developed and conspicuous, and but slightly smaller than the adambulacrals. Both supra- and inframarginals appear to be present. "The abactinal [actinal] furrows are large and subpetaloid. They are bounded by large tuberculate adambulacral plates. The ambu- lacral plates are boot-shaped [and slightly alternating]. The pores for the podia are large, and occur on the suture between the adam- bulacrals and ambulacrals. "The adambulacral elements in the oral armature are prominent, and consist of five pairs of subtriangular plates. "Type-species: Asterias antiqua, Hisinger 1837. Silurian: Got- land." No other species of this genus is known. Remarks. Of this genus only the actinal side is known and the published figure from a drawing by Mr. George Lilejevall is one of the best of Paleozoic starfishes. Lindstromaster differs from Petras- ter in but one generic character, if it be a generic one, in that the ambulacral furrows are large and subpetaloid in the former and narrow and slowly tapering in Petraster. However, since the abac- tinal area is unknown in Lindstromaster there probably will be dis- covered in this area differences distinguishing it from Petraster. It is also probable that the marginals in Lindstromaster which Gregory thinks are "both supra- and inframarginals" lying directly super- posed are only inframarginals. In Siluric starfishes these plates have not attained the superposed position so marked in recent genera. The supramarginals are here never true marginals and always lie more or less inside the inframarginals, which are the true marginals of these early genera. For other remarks see Palseostella. The name Hisingeraster was published after Gregory's name and is based on the same genotype. When more material is obtained and the genus is restudied, com- parisons should also be made with Palasterina primseva, as the two forms appear to have much in common. LINDSTROMASTER ANTIQUT7S (Hisinger). Asterias antiqua HISINGER, Lethsea Suecia, 1837, p. 89, pi. 26, fig. 6. Lindstromaster antiqua GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, pp. 343, 347, pi. 36, figs, la, 16 (complete synonymy given here). SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 225. A single specimen is known from "Bed C," Mount Hoburg, Got- land, corresponding to the Wenlock of England. It is in the Angelin collection of the Mineralogical Museum at Copenhagen. 150 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Genus PALASTERINA McCoy. Palasterina McCoY, British Pal. Foss., 1851, p. 59 (not defined). SALTER, Ann. Mag. Nat. Hist., ser. 2, vol. 20, 1857, pp. 324, 327. BILLINGS (part), Geol. Surv. Canada, Can. Org. Rem., dec. 3, 1858, p. 76. ZITTEL, Handb. Pal., vol. 1, 1879, p. 453. STURTZ (part), Palseontographica, vol. 36, 1890, p. 226; (part), Verb, naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, pp. 43, 60. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 220. Palxasterina GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 349 (complete synonymy here). The term Palasterina was used by McCoy in the following words: "Before I was aware Prof. Forbes had described them [the three species named below] it seemed to me that the U [rasterella] ruthveni and U. Jvirudo, as well as the similar American species, might be easily separated from the great starfishes forming the recent genus Uraster, by their small size and much more simple skeletons, and I had named the genus Urasterella in my manuscript. The U. primdevus I thought generically distinct from the other two, as the rays were not contracted at base, etc.; and I had named it Palasterina, from its resemblance to the recent genus Asterina." The three species mentioned by McCoy had been described by him in manuscript, but on learning that descriptions had been previously published by Forbes, the former, in his "British Palaeozoic Fossils/' adopted the names of Salter. In writing of these three forms, as above quoted, he uses the generic name Palasterina (not Palseasterina, the generally accepted form), and while a type-species is selected, yet McCoy did not point out a single generic character of present value. His statement that "the rays were not contracted at base" is a feature which at that time distinguished Palasterina from Urasterella, but now many Paleozoic genera are known having rays "not con- tracted at base." While McCoy does not strictly propose the generic name, Salter is the first to accept and define Palasterina, which he does in the following words : "Pentagonal, depressed, the arms a little produced, with three or five principal rows of tubercles above, combined with a plated disk which fills .up the angles ; ambulacra rather shallow, of subquadrate or slightly transverse ossicles, bordered by a single row of squarish large plates, the lowest of which (ad-oral adambulacral plates, Huxley; angle-ossicula, Forbes) are large and triangular, bearing combs of spines (Upper Silurian)." Recently Gregory, in a paper entitled "On Lindstromaster and the classification of the Palseasterids" (1899), redefines the genus, basing his definition largely on P. lonneyi. In this connection he writes: "The species [P. lonneyi] is of interest, as the structure is better preserved than in the type of the genus [P. primseva], and thus we are KEVISION OF PALEOZOIC STELLEKOIDEA. 151 largely dependent upon it for our knowledge of the generic charac- ters " (p. 350). Emended description. Disk large, with moderately developed inter- brachial arches. Rays five, short, and tapering more or less rapidly. U R about twice as great as r." Abactinal area of rays distally with one medial column of large, thick, and subquadrangular radial plates, on each side of which are similar columns of supramarginal and inframarginal pieces. In this region the columns adjoin and the ossicles are crowded, but proximally between the radial and supramarginal and between the latter and the inframarginal columns (ambital areas) are inserted numerous accessory smaller plates. The latter are most numerous in the ambital areas and at the base of the rays. The ossicles are packed into a tessellate pavement. In the center of the disk there seems to be a prominent central disk piece and at the bases of the rays there is a somewhat disconnected ring of ten most prominent plates, the five basal pieces of the radial and five interradial that give rise to the ten supramarginal columns. Inside of this ring there are probably numerous, very small pieces with no definite arrange- ment. A number of the proximal plates of both the radial and supramarginal columns may be separated from one another by small accessory pieces. Madreporite abactinal. Inframarginal plates bounding the entire outer edge and common to both sides of the animal. These plates are, however, small (smaller than the adambulacrals) and bear tufts of small spines. Adambulacrals subquadrangular, and closely adjoining, each plate bearing on its outer lateral edge a tuft of three spines. Oral armature consists of five pairs of triangular adambulacral plates which have slender spines on the inner edges. Interbrachial areas extending to near the distal ends of rays and occupied by numerous "isolated rounded or subangular [accessory interbrachial] plates bearing [single] small spines." Ambulacrals slightly alternating and about as numerous as the adambulacrals. Each plate has an L-shaped ridge, posteriorly directed. The podial openings are on the sutures between the ambulacrals and adambulacrals. Genoholotype. Uraster prims&vus Forbes. The above diagnosis is, however, based largely upon Palasterina bonneyi Gregory. Dr. Bather sent the writer a number of wax squeezes of the genoholo- type, and while the parts preserved are clear and sharp, yet the speci- mens are not complete enough to work out the generic characters. Therefore, for the present, P. bonneyi is taken as the genoholotype. Distribution. Ludlow shales (Upper Siluric) of Great Britain. 152 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Contains : P. primseva (Forbes). Ludlow of England. P. bonneyi Gregory. Ludlow of England. P. (?) ramseyensis Hicks. Tremadoc of Wales. For other remarks see Palseostella. Since the foregoing was written, Schondorf 's work cited above has come to hand. He studied wax squeezes of P. primseva and deter- mined that the genotype is devoid of marginalia, and therefore refers Palasterina to the Cryptozonia. He writes: " Special marginalia ap- pear not to be present, at least the marginal rows are in no way dis- tinguishable from the other intermediary [accessory interbrachial] plates. Accordingly Palasterina primseva Forbes sp. does not belong as stated by Gregory (p. 349) to the asterids with distinct marginal plates, the Phanerozonia, but with the asterids having inconspicuous or no marginalia, the Cryptozonia' 7 (p. 222). The present writer holds that there are inframarginalia present all along the edge of the interbrachial areas and the rays, and that they are most typical along the distal portion of the rays. Here the inframarginalia are ontogenetically youngest and less affected by subsequent growth displacement and alteration than are the older ones in the interbrachial areas. In the latter region it is true that the marginalia are not conspicuous, but so long as one can make them out to be inframarginalia, size alone can not be the criterion that will permit of the conclusion that no marginalia are present in Palasterina. However, the difference in interpretation between Schondorf and the writer is not a great one, as neither lays great taxonomic value on whether the marginalia are prominent or u not prominently developed" (p. 223). So long as the inframarginalia are more or less prominent, one of the generic characters of Palasterina, a phanerozonian, is present, and when they are so inconspicuous as not to be made out among the other ossicles, we have one of the generic characters of Schuchertia, a cryptozonian. Schondorf also states that the ambulacralia, and the adambulacralia as well, are all opposite and do not alternate with one another. The present writer states that the ambulacrals are " slightly alternating, " H and throughout this work he has given the arrangement of these ossicles his special attention. In a great many cases, and especially in the Ordovicic species, he has had difficulty in ascertaining what was their actual position in the living animals, for nearly all the fossils have been squeezed more or less. He has always given their arrangement as they impressed him in the fossil when no evident dis- tortion could be seen. The writer would not lay special stress upon slight alternation among the ambulacralia, for in many Ordovicic species there is some variation in each ray from opposite to slightly , alternate. Of course, when the alternation or opposite arrangement REVISION' OF PALEOZOIC STELLEROIDEA. 153 is complete and fixed throughout the rays, he would lay great value on the character. It may be said that in the earliest Asteroidea the chronogenetic tendency is from slight alternation to complete and fixed opposite arrangement among the amhulacral ossicles. PALASTERINA PRIM^EVA (Forbes). Uraster primxvus FORBES, Mem. Geol. Surv. Great Britain, vol. 2, pt. 2, 1848, p. 463; Mem. Geol. Surv. United Kingdom, dec. 1, 1849, p. 2, pi. 1, figs. 2a, 26; in McCoy, Brit. Pal. Foss., 1851, p. 60. MURCHISON, Siluria, 1854, p; 221, fig. 39. Palasterina primxvus McCoy, Brit. Pal. Foss., 1851, p. 59 (nomen nudum). SALTER, Ann. Mag. Nat. Hist., ser. 2, vol. 20, 1857, p. 327, pi. 9, figs. 2a-2c. WRIGHT, Mon. British Foss. Echinod., Oolitic, vol. 2, pt. 1 (Palseontogr. Soc. for 1861), 1862, p. 26, fig. 16a. QUENSTED, Petrefactenkunde Deutsch- lands, vol. 4, 1876, p. 74, pi. 92, fig. 35. STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, p. 44; vol. 56, 1900, pp. 214, 224. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 220. Palszasterina primseva GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 349. SPEN- CER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palseontogr. Soc. for 1913), 1914, pp. 37, 38, fig. 30. Formation and locality. A common species in the Ludlow rocks. Underbarrow, Kendal, Westmoreland, and Leintwardine, Shropshire, England. Remarks. Dr. Bather furnished the author with wax squeezes of specimens Nos. E4990-E4993 (E61 is quite another species and abactinally more closely related to Eudsonaster) in the British Museum (Natural History). They are from the Lower Ludlow at Kendal, Westmoreland. These show little of value abactinally, but actinally are well preserved along the ambulacra. Unfortunately, however, none of the marginals are present and but little of the interbrachial areas. When better material is at hand a comparison should be made also with Lindstrb'master antiquus (Hisinger), as the two look suspiciously alike. PALASTERINA BONNEYI Gregory. Palseasterina bonneyi GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, pp. 349, 350, text figs. 1-3, and pi. 16, figs. 2a, 26. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 223. "This species has hitherto been included in P. primseva, the type- species of the genus, from which it differs by having shorter and blunter rays, more numerous actinal interbrachial plates, and more adambulacral plates in each series ; the character of the abactinal ray plates is also different in the species, the longitudinal series being widely separated at the proximal end of the rays in P. bonneyi" (Gregory). Formation and locality. Ludlow shales. Leintwardine, Shrop- shire. Types in British Museum (Natural History). 154 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. PALASTERINA (?) RAMSEYENSIS Hicks. Palasterina ramseyensis HICKS, Quart. Journ. Geol. Soc. London, vol. 29, 1873, p. 51, pi. 4, figs. 21-23. WOODWARD, Geol. Mag., dec. 2, vol. 1, 1874, p. 96. This starfish needs to be restudied in the light of modern knowl- edge. The specimens represented by Hicks's figures 21 and 22 seem to be referable to Petraster rather than to Palasterina, while that of figure 23 appears to be nearer Mesopalseaster. Formation and locality. Said to be from the Tremadoc, but the development of the starfish seems to be in harmony with Upper Ordovicic time. The specimens are from Ramsey Island and St. Davids, Wales, and are in Mr. Lightbody's collection. Genus URANASTER Gregory. Uranaster GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 347. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 225. The name is "from Uranus, the grandfather of Pluto; the starfish resembles the living genus Plutonaster." Generic description. Animal deeply pentagonal, above medium size, with well-developed interbrachial arches, and five long slender rays. Inframarginal plates marginal, large, and elongated in the direction of the rays. These plates are apparently common to both the actinal and abactinal areas. Abactinally adjoining the inframarginals is another column of large submarginal plates ( ?supramarginals) alternating with the former, and sometimes there are smaller acces- sory pieces between these columns in the angles of the plates. The rays have no medial or radial columns of prominent plates, but the entire space between the submarginal columns is occupied by numer- ous closely fitting, convex, or tubercular plates which are arranged in alternate transverse rows, each having from three to four pieces on each side of a ray. Interbrachial accessory plates numerous, polygonal, and, according to Baily 's figure of U. Tcinahani, arranged in crescentic rows conforming to the shape of the disk margin. These plates do not appear to extend between the marginals and adambulacrals beyond the mid- length of the rays. Adambulacral plates tubercular, subquadrangular in outline, and widest transversely in the medial third of the column. Ambulacral furrows very narrow, deep, and tapering gradually. Ambulacra! plates traversed by strong transverse ridges. Madreporite unknown. GenoJiolotype. Palasterina Icinahani Baily. Upper Ordovicic. A second species is U. elizse, Spencer, also from the Ordovicic of Great Britain. Remarks. Uranaster differs from Petraster, Palasterina, and Lindstromaster in having the abactinal area of the rays between the REVISION OF PALEOZOIC STELLEROIDEA. 155 two pairs of marginal columns filled with small tuberculate plates arranged in transverse rows. The other genera have the plates in longitudinal columns. URANASTER KINAHANI (Baily). Palasterina kinahani BAILY, in KINAHAN, Man. Geol. Ireland, 1878, pp. xv, 26, pi. 2, fig. 1; Mem. Geol. Surv. Ireland, Expl. Mem. Sheets Nos. 169, 170, 180, 181, 1879, pp. 58, 59, fig. 6. Uranaster kinahani GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 348. SCHONDORP, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 225. This species occurs abundantly as natural molds in the Caradoc (Upper Ordovicic), near Ballymadder Point and Bannow, Wexford, Ireland. A specimen is in the British Museum (Natural History), No. E 194. URANASTER ELIZAS Spencer. Uranaster elizx SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palaeontogr. Soc. .for 1913), 1914, pp. 18, 20, 23, figs. 18, 23. From the Ordovicic of Great Britain. Genus PAL^OSTELLA Stiirtz. Palxostella STURTZ, Palseohtographica, vol. 36, 1890, p. 230, pi. 31, figs. 42, 43; Verb, naturb. Ver. preuss. Rheinl., etc., vol. 50, 1893, pp. 42, 58, pi. 1, figs.5^8, Palsenectria STURTZ, Verb, naturb. Ver. preuss. Rbeinl., etc., vol. 50, 1893. p. 12 (genoholotype the second specimen of Palseostella solida). Description. Animal attaining a large size, with large pentagonal interbrachial arcs. Rays five, stout. Ambulacral grooves wide and apparently regularly tapering. Abactinally the inframarginals are also the marginals for this area. Near the distal ends of the rays the supramarginals adjoin and lie either partially over or inside of the inframarginals and continue nearly to the axillae. There also seems to be a single radial column, confined, however, to the distal third of the rays. Other- wise the abactinal skeleton consists of very numerous, small, irregu- larly shaped, loosely interlocking plates without definite arrangement. Madreporite abactinal. Actinally the animal is bounded by subquadrate, numerously spine-bearing, inframarginal plates which increase rapidly in size towards the axillae, where they are ponderous and nearly twice as wide as long. Ambulacral plates alternating, L-shaped, with single columns of podial openings. Adambulacrals subquadrate, prominent, increasing regularly in width proximally and bearing spines like the marginals. Interbrachial areas large, reaching to near the distal ends of rays, widely separating the inframarginals from the adambulacrals in the axillary region, and the space occupied by very numerous, small, spine-bearing, apparently irregularly dis- posed plates. 156 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Genoholotype and only species. P. solida Stiirtz. Two speci- mens are known from the Lower Devonic roofing slates, Bundenbach, Germany. Remarks. The starfishes from the Bundenbach roofing slates are at best but poor fossils and were it not for the great skill of Mr. Stiirtz in cleaning them out of the matrix their systematic position could not well be determined. At best they are compressed to almost paper thickness and distorted during solidification and shortening or elongation of the slates. Under these circumstances it appears best to redescribe the detailed abactinal characters as given by Stiirtz. He mentions numerous paxillae and finds many characters common to both Palseostella and the living Nectria. Palseostella is closely related to Palasterina, also to Lindstromaster , and may be a descendant of the latter. The actinal structure in Lindstromaster and Palseostella is nearly alike, both having large inframarginals and adambulacrals, but the interbrachial areas in the former are occupied by large closely adjoining plates, while in the latter these are small and somewhat loosely interlocking. Abac- tinally they can not be compared, since that side is unknown in Lindstromaster. The greatest difference between Palasterina and Palseostella is that the former has small inframarginals while in the latter they are very large. Then the supramarginal and radial columns in Palasterina are better developed and comparatively larger, while in Palseostella they are shorter and most prominent in the distal half of the rays. In the latter genus the central region of the disk has no ring of fifteen basal radial and supramarginal plates as in Palasterina. From this evidence it seems probable that the line of descent is from Petraster to Lindstromaster to Palseostella in marntaining promi- nent columns of inframarginals, while the line from Petraster to Palasterina tends toward the reduction or elimination of these plates as a conspicuous skeleton element. Palsenectria is based on the second specimen, and the genus is described as follows: " Should the described starfish [the second specimen of Palseostella solida Stiirtz] prove not to be identical with Palseostella, which is possible, even though not probable the other specimen is in the British Museum I propose for it the new name Palsenectria devonica." Unless the author takes the time to work out his new forms, or decides the problem presented for himself, he should not expect some one else to do this for him. Under these circumstances the genus and species are regarded as not defined. Genus PSEUDOPALASTERINA Sturtz. Pseudopalasterina STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 56, 1900, pp. 219, 224. Genoholotype and only species. Palasterina follmanni Sturtz. REVISION OF PALEOZOIC STELLEROIDEA. 157 PSEUDOPALASTERINA FOLLMANNI (Sttirtz). Palasterina follmanni STURTZ, Paleeontographica, vol. 36, 1890, p. 226, pi. 29, figs. 29-31a. Pseudopalasterina follmanni STURTZ, Verb, naturh. Ver. preuss. Rheinl., etc.. vol. 56, 1900, pp. 219, 224. Actinally this species has ambulacrals and narrow adambulacrals and outside of the latter the interbrachial and abactinal areas consist of a membrane with small spine-bearing plates. It has no inframar- ginal nor radial and supramarginal columns, features characterizing Palasterina. Formation and locality. Lower Devonic of Bundenbach, Germany. NEW GENUS. NEW SPECIES. Prof. Weller got on Fountain Creek, at Waterloo, Illinois, in the St. Louis limestone, a small pentagonal asterid (No. 14396, Univer- sity of Chicago collection) showing only the actinal side, that differs from all other known forms of the large-disked genera. The five rays are completely within the pentagon, with the interbrachial areas large and straight-sided. These areas are made up of a few large tesselate plates, of which it seems that seven compose the interbrachial mar- gins and inside these there seem to be transverse rows, each with five, three, two, and one ossicles. It could not be made out whether the inframarginals and supramarginals are directly superposed, but this may be so. Ambulacral furrows exceedingly narrow, a mere line. These are bounded by columns of large, very ornate adambulacrals, each with seven ossicles. Those of opposite columns alternate decidedly with one another and on its outer side each bears two short but dis- tinct spines. Adambulacral oral armature in five very prominent pairs of pieces, long and pointed, and about twice as long as any other adambulacrals. Measurements: R = 8.5 mm., r = 7 mm. Family LEPIDASTERIME Gregory (emend.). Lepidasteridse GREGORY (part), Geol. Mag., dec. 4, vol. 6, 1899, p. 352 (includes Lepidaster&n&Etheridgaster)', Treat. Zool., vol. 3, Echinoderma, 1900, p. 255. Helutnihasterinse GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 258. Helianihasteridse STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 56, 1900, p. 204. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 235. Primitive multi-rayed Phanerozonia apparently with the general structure of Pabeasteridse, but with as many as 24 rays. Madreporite known only in Helianthaster, where it is interbrachial and on the 158 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. edge of the disk, more actinal than abactinal. Inframarginals or supramarginals bounding the rays. Contains : Lepidaster Forbes. Eelianthaster Koemer. Lepidaster ella, new genus. Remarks. For a discussion as to why there are multi-rayed that is, more than the normal five-rayed asterids, see the family Palse- osolasteridse. Genus LEPIDASTER Forbes. Lepidaster FORBES, Mem. Geol. Surv. United Kingdom, British Org. Hem., dec. 3, 1850, p. 1, pi. 1, figs. 1-3. WRIGHT, Mon. British Foss. Echinod., Oolitic, vol. 2, pt. 1 (Palseontogr. Soc. for 1861), 1862, p. 35. WOODWARD, Geol. Mag., dec. 2, vol. 1, 1874, p. 9. ZITTEL, Handb. Pal., vol. 1, 1879, p. 454. STURTZ, Palseontographica, vol. 36, 1890, p. 222, pi. 28, figs.19-20; Verb, naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, pp. 52, 72. Generic description. Disk very large, over 50 mm. in diameter, and apparently composed of heavy, closely set, polygonal plates. Rays 13 in number, fairly stout, not slender, and tapering fairly rapidly in the distal half. R = 52 mm. , r = 25 mm. A little more than half the length of the rays extends beyond the disk. Actinally, the rays are bounded on each side of the very narrow ambulacral furrow by two columns of strong, tumid, closely ad- joining, transversely oblong, adambulacral and inframarginal plates. The latter are also the marginals for the abactinal side. The adam- bulacrals are tuberculate and probably bore small spines; there are 25 ossicles in each column. Ambulacrals unknown, but the arrangement of the adambulacrals indicates that they were slightly alternate. Madreporite and interbrachial areas unknown. Abactinally little is known. The rays are described as having " numerous small polygonal nearly flat ossicula, closely set, and of various sizes." To judge by similar rays in other genera, it would appear that Lepidaster may have had radial and supramarginal columns of ossicles. Genoholotype and only species. L. grayi Forbes. The holotype was found in the quarries of Wenlock limestone in the Castle Hill at Dudley, England. Observations. Forbes's illustration leads the writer to believe that the abactinal plates have been deranged and that originally the larger pieces were arranged in supramarginal columns, while the smaller ones composed the radial row of ossicula. This type of structure generally goes with the actinal arrangement as here seen in Lepidaster , a duplicate of the same structure in Hudsonaster } Poise- aster, and Mesopalseaster. Stiirtz compares the abactinal skeleton of REVISION OF PALEOZOIC STELLEEOIDEA. 159 Lepidaster with the recent ScytJiaster, which is made up of plates of various sizes without regular arrangement. It seems more natural to compare Lepidaster with Siluric and Ordovicic genera than with recent forms, and particularly when nearly identical structures are found in the genera mentioned. Forbes also describes Lepidaster as having the ambulacral grooves " towards the extremity, in some instances partially filled up by small polygonal intervening plates." His figures show a few of these plates arranged in single columns. This anomalous structure is undoubtedly best explained by accepting Stiirtz's conclusion, that they are the separated ambulacral pieces, since otherwise none of these plates appear to be present. As a fifth actinal, or a single ambulacral column, the structure is out of all* harmony with star- fishes. Genus HELIANTHASTER Roemer. Eelianthaster ROEMER, Palseontographica, vol. 9, 1863, p. 147, pi. 28. QUENSTEDT, Petrefactenkunde Deutschlands, vol. 4, 1876, p. 82, pi. 93, fig. 32. ZITTEL, Handb. Pal., vol. 1, 1879, p. 453. STURTZ, N. Jahrb. fur Min., etc., 1886, vol. 2, p. 148; Palaeontographica, vol. 32, 1886, p. 81; vol. 36, 1890, p. 218, pi. 27; Verb, naturh. Ver. preuss. Eheinl., etc., vol. 50, 1893, pp. 12, 30. GREGORY, Treat. Zool., vol. 3, Echinoderma, 1900, p. 258. CLARKE, Bull. N. Y. State Mus., No. 121, 1908, p. 61, pi. 10. SCHONDORF, Jahrb. nas- sauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 34. Generic characters. Disk large; rays 14 to 16 in the genotype. Interbrachial actinal areas well developed but not over-large for a many-rayed asterid, either integumentary or granular, distally con- cavely cut out and margined by several small inframarginals, about seven on each side of a larger axillary piece. Madreporite marginal, large, more actinal than abactinal, radi- ately striate, interbrachial, and situated immediately outside of the inframarginals. In the fossils it appears to be of the actinal surface. Rays long and slender, protruding beyond the disk more than twice the length inside the disk. Adambulacrals opposite one another, with small lateral spines, prominent and large, wider than long, bounding the rays, and terminating proximally in pairs of modified pieces. These make the adambulacral parts of the oral armature. Ambulacral grooves narrow, ambulacral ossicles not seen but probably arranged opposite one another. Abactinally the disk appears to have a scale-like covering. The rays have median columns of radialia, outside of which are stout columns of supramarginal ossicles. GenoJiolotype (type by monotypy). H. rhenanus Roemer (citations as above). Lower Devonic of Bundenbach, Germany. The above definition is based on the illustrations of Stiirtz. There is a second species in Germany, and H. filiciformis occurs in England. 50601 Bull. 8815 11 160 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. HELIANTHASTER FILICIFORMIS Woodward. Helianthaster filiciformis WOODWARD, Geol. Mag., dec. 2, vol. 1, 1874, pp. 7-10, fig. on p. 8. A natural mold in sandstone measuring about 45 mm. in total diameter. Has 11 rays. The disk "must have measured 10 mm. in breadth. The arms are 17 mm. in length, and near their junction with the body are 5 mm. broad. " The specimen needs to be restudied and modernized. The holotype is from the Upper Devonic at Great Inglebourne, Harberton, South Devon, England, and is now in the collection of Mr. Champ ernowne. LEPBDASTERELL.A, new genus. Plate 25, fig. 2. Lepidasterella SCHUCHERT, Fossilium Catalogus, Animalia, pt. 3, April, 1914, p. 23. Much like Lepidaster, but with 24 rays instead of 13 as in that genus. Rays extending considerably beyond the disk. The abac- tinal side only is known, but it shows that the rays have slender radial columns, the ossicles of which are disposed alternatingly with the larger and more prominent supramarginals. Outside of these in some rays may be seen the edges of the inframarginalia, and these have well developed but slender, rather long spines. There appear to be no ambital areas, as the inframarginals of adjoining rays meet and at their junction begins the disk. The disk appears to have been made up of more or less closely adjoining polygonal plates whose arrangement is not discernible in the specimen. The madreporite and its position are unknown. Actinal area unknown. Genoholotype and only species. L. babcocki, new species. Remarks. The essential difference between Lepidaster and Lepid- asterella is that the former has 13 rays while the latter has 24. LEPIDASTERELLA BABCOCKI, new species. Plate 25, fig. 2. Helianthaster, new species CLARKE, Fifty-ninth Rep. N. Y. State Mus., 1906, p. 37, and plate. Helianthaster gyalum CLARKE (part), Bull. N. Y. State Mus., No. 121, 1908, p. 61 (not pis. 12, I3=Pal3eosolaster (?) gyalum}. Description. Rays, 24. Length of best preserved ray in the holotype, 28 mm. Width of ray at base, 3 mm. Width of disk, probably not less than 20 mm. Rays slender and regularly tapering. Their structure is only known abactinally, from a natural mold in a fine sandstone. Medi- ally there is a narrow but high column of radial plates, each one of KEVISION OF PALEOZOIC STELLEROIDEA. 161 which alternates with those of the supramarginal plates. The latter are slightly tumid, subquadrate in outline, regular in size, but decreasing gradually distally. Outside of these and margining the rays may be seen a little of the inframarginalia, which bear pairs of short, slender, and finely striated spines. There are from 31 to 33 closely adjoining plates in each one of the abactinal columns; there appear to be no accessory plates of any kind in the rays. Disk circular in outline, somewhat tumid and covered by small many-sided plates that originally appear to have been closely adjoin- ing. Remarks. As this is the only American starfish with more than five rays, it is readily distinguished from all others. The abactinal arm structure is practically that of Hudsonaster, with the difference that in L. labcocki the inframarginals do not project beyond the supramarginals and are therefore not seen as a rule from the dorsal side in the natural casts. Locality and formation. The holotype, an excellent specimen, was found in the lower part of the Upper Devonic, near Ithaca, New York, by Mrs. G. W. Babcock, after whom the species is named; it is in her possession. The writer's attention was called to this interesting specimen by Prof. H. S. Williams of Cornell University. Some years after this starfish was described in manuscript, Clarke directed attention as above cited to two others found in the Cashaqua shales of the Portage in Hunt's quarry, at Interlaken, New York. Cat. No. 60627 (cast), U.S.N.M. FAMILY, unknown. Genus ASTROPECTEN Link. The genus Astropecten is based upon living species, and it is not probable that the Devonic forms are identical. ASTROPECTEN (?) SCHLUTERI Sturtz. Astropecten schluteri STURTZ, Palaeontographica, vol. 32, 1886, p. 87, pi. 10, figs. 1, 16, 2, 2a; Verb, naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, pp. 51, 71. In the roofing slates of the Lower Devonic at Bundenbach, Ger- many. A single very much distorted and compressed example is the basis of this species. Genus ATAXASTER Jaekel. Ataxa.ster JAEKEL, Zeits. geol. Gesell., vol. 55, 1903, Protokol, p. 14 (106). Genoholotype and only species. A. pygmseus Jaekel. Ordovicic (DJ, St. Benigna, Bohemia. Description. "A small asterid whose ambulacrals in the distal parts are alternate, but in the proximal are opposite." Remarks. The above is all that is known about Ataxaster, and for the present the genus has no value. 162 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. CRYPTOZONIA Sladen. Specialized Asteroidea without true marginal plates. In other words, there are no large and conspicuous marginal plates bordering the animals, but in primitive forms large adambulacrals do margin the species. In other primitive forms more or less of the infra- marginals may be present, but if so they remain adjoining the adam- bulacrals. Ambulacrals alternating or directly opposite one another. Oral armature generally with the adambulacral elements the most conspicuous; more rarely the ambulacral elements are most promi- nent. Contains the following families: Stenasteridae. Schuchertiidse. Monasteridae. Palasteriscidae. Urasterellidae. Schcenasteridae. Calliasterellidae- Palaeosolasteridae. Compsasteridae. Remarks. This widely accepted order, which appears as if founded in nature on a single grand phylum, probably has no such value. In other words, asterids without either one or both of the marginal columns bounding the animal developed several times. Therefore, the absence of these ossicles is no indication of direct phyletic relationship, but such are simply cases of parallel development. As long as asterids retain the primitive characters of few columns of rather strong ossicles and the adambulacral type of oral armature, they are clearly Phanerozonia, but when the ponderous skeleton breaks up into a more flexible one of small pieces, then the marginal and abactinal columns vanish as such in the mass of small plates. On the other hand, ponderosity may even be retained with the loss of the inframarginals alone. In all of these Paleozoic forms, the adambulacral type of oral armature is the rule. Among the Paleozoic "Cryptozonia" it seems certain that at least the families Stenasteridae, Urasterellidae, Schuchertiidae, and Palaeo- solasteridae had independent origins in Phanerozonia stocks. The Urasterellidae and Calliasterellidae apparently had a single origin in the Phanerozonia subfamily Mesopalaeasterinae by the adoption of marked flexibility. On the other hand, the families Stenasteridae and Monasteridae retained rigidity and primitive characters, and more probably arose in the Hudsonasteridae through the loss of the iriframarginals alone. The Schuchertiidae also had their origin directly in some Phanerozonia stock and one naturally looks for it in forms developing large interbrachial areas as in the Palasterinidae. In this family, however, the accessory plates are introduced between the adambulacrals and inframarginals, while in the former family they KEVISION" OF PALEOZOIC STELLEROIDEA. 163 are introduced outside of the inframarginals and the entire abactinal skeleton is broken up into an abundance of tiny pieces. Therefore, one must go back for their origin at least as far as the Palaasteridse. In this case it was not primarily flexibility that led to this supposed ordinal change, but it was due to increase of body cavity and enforced enlarging of the skeleton in the interbrachial and ambital areas, making the more or less pentagonal disk. It is probable that the Palasteriscidas were their descendants, while the Schoenasteridse may have been the final Paleozoic representatives of the same phyletic line. The PalasosolasteridaB seemingly arose in the Phanerozonia family LepidasteridaB by the adoption of a flexible skeleton. It would, therefore, seem that the Paleozoic families of supposed Cryptozonia construction are to be grouped as follows: Superfamily Stenasteracea, new. Includes families Stenasteridse and Monasteridse. Superfamily Urasterellacea, new. Includes families Urasterellidse, Calliasterellidse, and Compsasteridse. Superfamily Schuchertiacea, new. Includes families Schuchertiidse, Palasteriscidee, and Schcenasteridae. Superfamily Palseosolasteracea, new. Includes family Palaeosolasteridse. STENASTERIDSE, new family. Rigid, small, primitive, five-rayed Cryptozonia, with rapidly taper- ing or petaloid rays. Ambulacrals few in number and opposite in arrangement. Interbrachial areas consisting of the single pairs of large basal adambulacral oral armature plates. Abactinal skeleton unknown, but in some forms seemingly integu- mentary. No spines as yet known. It is probable that this family will be referred to the Auluroidea. Contains: Stenaster Billings. Tetraster Nicholson and Etheridge. Genus STENASTER Billings (emend.). Plate 32, fig. 1. Stenaster BILLINGS, Geol. Surv. Canada, Can. Org. Bern., dec. 3, 1858, p. 77. STURTZ, Palseontographica, vol. 36, 1890, p. 220. GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 352. SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palseontogr. Soc. for 1913), 1914, p. 22. Urasterella STURTZ (not McCoy), Verb, naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, pp. 40, 41, 56. Generic characters. Small rigid asterids, the largest example having 11 = 20 mm., / = 4.5 mm. Disk small, without interbrachial arcs. Rays five, short, stout, lanceolate or petaloid. Abactinal area devoid of a preservable skeleton. For this reason when this side is at hand it is nearly always mistaken for the actinal, so nearly alike are the two. 164 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Madreporite unknown. Adambulacral plates with granular surfaces, few in number, never more than 17 hi a column, subquadr angular and thick, greatest width about the middle of the rays, rapidly diminishing in size distally and making the margin of the rays and disk. The great proximal adam- bulacrals are the orals, occupying the axils in pairs, acutely triangular, more or less pointed, and terminating orally. Ambulacral plates large, directly opposite one another and the adambulacrals as well, h- -shaped, with the widest portion in the per- radial center of the rays; there are as many as 16 or 17 ambulacrals in a column. Laterally each plate is more or less excavated, both distally and proximally, thus leaving very large, subcircular podial openings. Spines of any kind are so far unknown. Genoholotype (the first species and the one selected by subsequent workers). Stenaster salteri Billings. Remarks. This genus has not had a good standing, due largely to the fact that Palseaster Hall was poorly known when Billings described Stenaster and further because he included two distinct generic types in his genus. A careful reading of Stenaster, however, reveals that Billings actually based his genus on S. salteri, the first form following the generic description, and this selection has been followed since. Later on, McCoy applied the name Urasterella in a loose manner to the second species (Palseaster pulcJiellus) , and although it is a form of this genus it is not the genotype. Hall (1 868) directed attention to this work of McCoy, but erroneously assumed that all of Stenaster was synonymous with Urasterella. It is true, as stated by Hall, that Billings's second species is congeneric with Urasterella, but the first species and genotype is certainly quite dis- tinct from McCoy's genus. Sttirtz (1890) also points out the same thing and gives a short and emended diagnosis of " Urasterella (Stenaster)." In 1893 the same writer applied McCoy's name Urasterella to Stenaster salteri and made S. pulchellus the genotype of Urasterella, a proceeding that violates the rules of nomenclature. Stenaster as here restricted differs from Urasterella in being more decidedly rigid and not so flexible as the latter, while the rays are far shorter and petaloid in form. For these reasons, Stenaster has comparatively few adambulacrals and ambulacrals, while there are many more of these ossicles in Urasterella. Finally, the former reveals no abactinal plates or spines of any kind, while the latter has a skeleton of small plates that terminate in long inarticulate spines or rods. Actinally Stenaster is closely related to Tetraster, and the only marked difference is that the former has petaloid rays and more pronounced adambulacral oral jaws. REVISION OF PALEOZOIC STELLEROIDEA. 165 There would be no need to point out here the difference between such distinct genera as Stenaster and Palseaster were it not that Hall regarded the former as a synonym of the latter. In Stenaster the rays are petaloid and not slender and regularly tapering, as in Palseaster. The most marked distinction, however, is the fact that these two genera belong to different orders, for Palseaster has columns of iirframarginals that are completely absent in Stenaster. Further, in the last-named genus, the interbrachial areas are made up each of two pairs of oral armature pieces, while in Palxaster there are, distaUy to these, additional large single axillary interbrachials. Stenaster and the related genus Tetraster are very primitive in construction. Actinally they are seen to have the structure of Hudsonaster minus the inframarginal columns and the single axillary interbrachials. It is this difference that makes the former genera members of the order Cryptozonia, while Hudsonaster is the most primitive of Phanerozonia. Until the abactinal area is known ha Tetraster and Stenaster, positive phyletic relationship can not be stated, but in a number of specimens of the latter genus the evidence at hand indicates decided specialization through the complete removal of the heavily plated skeleton of Hudsonaster. Distribution. Ordovicic of America. The following are the known species: 8. salteri Billings. Black River and Trenton. 8. (?) oUusus (Forbes). Bala of Ireland and Wales. S. (?) coronella (Salter) . May Hill oandstone (Siluric) of England. S. (?) confluens Trautschold. Upper Carboniferous of Russia. STENASTER SALTERI Billings. Plate 32, fig 1. Stenaster salteri BILLINGS, Geol. Surv. Canada, Can. Org. Rem., dec. 3, 1858, p. 78, pi. 10, figs, la, 16. WRIGHT, Mon. British Foss. Echinod., Oolitic, vol. 2, pt. 1 (Palseontogr. Soc. for 1861), 1862, p. 28, fig. 15 (2) on p. 24. Original description. "This species has rather short, broad rays, which are narrower where they are attached to the very contracted body than they are at about the center of their length. In conse- quence of this form, the sides of the rays are not parallel, but a little curved outward. As however only two specimens have been col- lected, and both appear to be a little flattened by vertical pressure, it may be that this leaflike shape of the rays is accidental [an abundance of specimens show this form to be the natural one]. * * * The adambulacral plates are oblong [granular and from 7 to 17 in each column in the smallest and largest examples], and the sutures between them are nearly at right angles to the ambu- lacral grooves; those next the body are a little sloping outward. 166 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Their length is about twice their breadth, and they are so disposed that the greater dimension is transverse or at right angles to the groove; the extremities which lie next to the grooves are angular, and some of them appear to have the contiguous pores partly exca- vated in them. The oral plates [adambulacrals] are acutely triangu- lar, the sharpest angle being toward the mouth [and form the inter- brachial areas]. The plates are smooth [the adambulacrals are all distinctly granular]. The ambulacral pores are very large, and the ossicles are much contracted in the middle and greatly expanded along the median line of the bottom of the groove." These plates are directly opposite one another and the adambulacrals, and in number equal those of the latter. The type-specimen figured by Billings measures: R = 14 mm., r=3.5 mm., R = 4n Width of a ray at base about 4 mm., at about mid-length 5 mm. The largest specimen: K- = 20 mm., r 4.5 mm. In the University of Toronto there is a specimen that in every way, except one, has the characters of Stenaster salteri. It was found associated with many other individuals at Kirkfield. It differs from its associates in having what appears to be a distinct disk, rather large, with concave sides, filling in the spaces between the rays. One looks in vain, however, for plates or spines, as the inter- brachial areas are nothing more than an amorphous mass of cal- cium carbonate. These areas are very distinct and stand out prom- inently, but because they do not reveal any plated or spinif erous struc- ture the writer regards them as false characters, produced during the permineralization of the specimen. This conclusion is further sup- ported by the fact that otherwise the characters are those of S. salteri. Formation and locality. Three specimens (the type, No. 1398, and two fragments) are from the Trenton limestone at Belleville, Ontario, and are now in the Museum of the Canadian Survey. Another specimen in that museum is on the same slab with Petraster rigidus and has been labeled Palseaster matutina. In Mr. Ulrich's collection there are four isolated rays of this species obtained by him from the Black River limestone at Curdsville, Kentucky. A further specimen preserving two rays, from Government House Bay, Ottawa, was found by Mr. Walter R. Billings. In the Museum of Comparative Zoology there is still another individual which appears to be of this species; it was found by Mr. J. B. Perry at Pant on, Vermont. Finally, an abundance of material was obtained in the Black River limestones at Kirkfield, Ontario; there are more than twenty specimens from this locality in the Walker collection of the University of Toronto, gathered by Mr. Townshend. Cab. No. 60628, U.S.N.M. REVISION OF PALEOZOIC STELLEROIDEA. 167 STENASTER (?) OBTUSUS (Forbes). Asterias primseva SALTER and SOWERBY, Quart. Jour. Geol. Soc. London, vol. 1, 1845, p. 8, 20 (table) (rumen' nudum). Uraster obtusus FORBES, Mem. Geol. Surv. Gt. Britain, vol. 2, pt. 2, 1848, p. 463; Mem. Geol. Surv. United Kingdom, dec. 1, 1849, p. 2, pi. 1, fig. 3. MURCHI- SON, Siluria, 1854, p. 182, fig. 17. Palxaster obtusus SALTER, Ann. Mag. Nat. Hist., ser. 2, vol. 20, 1857, p. 326. WRIGHT, Mon. British Foss. Echinod., Oolitic, vol. 2, pt. 1 (Paleeontogr. Soc. for 1861), 1862, p. 24. SALTER, Mem. Geol. Surv. Gt. Britain, vol. 3, 1866, p. 289, pi. 23, fig. 1. Stenaster (?) obtusus STURTZ, N. Jahrb. fur Min., etc., 1886, vol. 2, p. 153; Verh. naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, pp. 41, 56. Stenaster obtusus SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palseontogr. Soc. for 1913), 1914, pp. 22, 23, 31, text fig. 21, pi. 1, figs. 6, 7. The little that is known of this species appears to be in harmony with the characters of Stenaster as here defined. It occurs in the Caradoc beds (Ordovicic) at Drumcannon, Waterford, Ireland, and in the Bala rocks west of Bala Lake (Moel-y-Garnedd) , North Wales. STENASTER (?) CORONELLA (Salter). Palxaster coronella SALTER, Ann. Mag. Nat. Hist., ser. 2, vol. 20, 1857, p. 326. WRIGHT, Mon. British Foss. Echinod., Oolitic, vol. 2, pt. 1 (Palseontogr. Soc. for 1861), 1862, p. 25. Stenaster (?) coronella STURTZ, N. Jahrb. fur Min., etc., 1886, vol. 2, p. 153. The description of this species is very short and almost nothing is known of its characters. It occurs in the May Hill sandstone (Siluric) at Gunwick Mill, Malvern, England. STENASTER (?) CONFLUENS Trautschold. Stenaster conftuens TRAUTSCHOLD, Nouv. Mem. Soc. Imp. Nat. Moscou, vol. 14, 1879, p. 11, pi. 2, fig. 5. SCHONDORF, Palseontographica, vol. 56, 1909, p. 337. Based on a fragment of the distal part of a ray. Appears to be a crj^ptozonian. Generic position unknown. From the Upper Car- boniferous (Moscovian) of Moscow, Russia. Genus TETRASTER Nicholson and Etheridge (emend.). Plate 33, fig. 4. Tetraster NICHOLSON and ETHERIDGE, Mon. Silurian Foss. Girvan Dist., Ayrshire, fasc. 3, 1880, p. 324, pi. 21, figs. 3-8 (not figs. 1-2= Eudsonaster batheri, and 9-10== Urasterella girvanensis) . The name has reference to the four columns of actinal plates of the rays. Remarks. The authors of Tetraster distinguished their genus from Palseaster on the ground that it had on the actinal side but four columns of plates, while that genus as typified by P. matutinus } now Hudsonaster matutinus, has six. In this definition they are correct, but they complicated their good intentions not only by referring to 168 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Tetraster another Cryptozonia form (Urasterella girvanensis, new species), but also by confusing with their genotype specimens that certainly have the character of Palseaster so far as six columns of actinal ossicles are concerned. This latter admixture the writer has taken out of the genotype and has described it in this work as Hudson- aster ~batheri. The adjusting of this matter has been made possible through the kindness of Dr. Bather, who made for this work wax squeezes of the various specimens. After seeing these and the other material originally referred to Tetraster by Nicholson and Etheridge, their various drawings became clear. Their figures are fairly good, but in places do not bring out 'all the characters, but it was the mixture of three generic forms under Tetraster that made it for a time impossible to determine on what the genus should stand. Emended description. Small, five-rayed asterids, E, = about 5 mm., r= about 2 mm. Abactinal side unknown. Actinally convex and in its general expression reminding much of living species with large marginals, but in Tetraster the comparatively massive margining plates are adambulacralia. Adambulacrals compose the entire outer actinal areas of Tetraster and margin the entire annual. Distally the pieces are small and the 12 ossicles of each column increase rapidly in size with the largest basal pieces of adjoining columns making the prominent interbrachial areas. Ambulacrally all are somewhat excavated, stand high above the furrow and each ossicle terminates in a ridge that joins those of the ambulacrals. All of these plates are abundantly granular. Ambulacral furrows not wide but rapidly becoming very deep proximally, with excavated adambulacral sides. The columns meet in front of the two largest interbrachial adambulacrals. Podial openings are large and issue deeply from beneath the adambulacrals and between the lateral sutures of the ambulacralia. The pieces of the ambulacral columns are as many as there are adambulacrals and are placed directly opposite one another. Medially the ambulacralia are elevated, roof-like, and without a central gutter. While the ab actinal area is unknown, it appears that the adambu- lacrals also margin this side of the animal. The nature of the skele- ton inside of these columns may be of small pieces as hi Urasterella, or, what seems more probable, consist of large radial and supramar- ginal ossicles, without accessory pieces. Spines of any kind are unknown. Genoholotype. Tetraster wyville-thomsoni Nicholson and Etheridge as above emended. T. (?) imbricatus (Salter) is also placed under this genus for the present. Tetraster as now defined is so distinct in its great adambulacral development that it will not be confused with any other Paleozoic genus except Stenaster. Both genera, so far as the actinal area is REVISION OF PALEOZOIC STELLEEOIDEA. 169 concerned, are closely related. The differences, however, are worthy of generic rank. In Stenaster the rays are petaloid, with the adambu- lacrals largest at the mid-length of the rays, while in Tetraster the rays are triangular, with the greatest ossicles in the axils. At first the genotype looks as if one had a direct representative of living Phanerozonia, but a little study quickly shows that the large actinal margining plates are not marginals or inframarginals, but adambulacrals, and therefore indicate a cryptozonian form. For a while it seemed that these large plates must be inframarginals, but then adambulacrals should be present, and they were looked for along the ver'tical ambulacral walls, but even here there are no diminutive pieces of these columns. Among Paleozoic asterids the writer has not seen a form without adambulacrals, and when there are but four actinal columns it is always the inframarginals that are diminutive or absent. This species may be identical with T. (?) inibricatus (Salter), which see for further remarks. Distribution. Restricted to the Upper Ordovicic at Thraive, Gir- van, Scotland. The specimens occur as very sharp natural molds in a micaceous sandstone, from which good wax squeezes can be made, revealing all the skeletal characters. The originals are in the collec- tion of Mrs. Robert Gray, Edinburgh, Scotland. TETRASTER (?) IMBRICATUS (Salter). Palseaster imbricatus SALTER, Mem. Geol Surv. Gt. Britain, vol. 3, 1866, p. 289, pi. 23, fig. 8. Original description. "P. rigidus, uncialis. Radii breves conici, nee ad basin valde expansi, subcarinati, tuberculis seu jugis obliquis remotiusculis asperi. "Ray conical, its length about five lines, its breadth at base less than two lines. Only two rays are preserved; these are convex, slowly tapering, somewhat carinate above (we do not see the lower side), and ornamented by tubercles gathered into curved oblique rows, which run forward from the avenues toward the carina above, and are separated from each other by a space little more than their own breadth. " Locality: Caradoc rocks, Llanfyllin, Montgomeryshire [Wales]. The specimen is in the collection of Mr. Prosser." Remarks. As the writer has not seen the holotype, his conclusions are based on the lithographic figure, which adds but little knowledge to that in the description. As there are but two columns of dorsal ray plates, it hardly seems that the species is a phanerozonian. In general it looks most like Tetraster, but as the dorsal side of this form is unknown, no direct comparisons can be made between them. On the other hand, the imbrication of the plates, their large size, 170 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. and completely opposite arrangement suggest that T. (?) imbricatus may be an auluroid. Until the original specimen is restudied no definite assignment can be given it. MONASTEKID^E, new family. Progressive but primitive five-rayed Cryptozonia, with massive plates, especially the adambulacrals, which are common to the actinal and abactinal areas. Ambulacrals opposite. Interbrachial areas con- sist of a number of pah's of adambulacrals crowded orally Abacti- nally the radial and supramarginal columns consist of large, closely adjoining plates, and are similar to those of Hudsonaster. Contains : Monaster Etheridge. Genus MONASTER Etheridge. Palseaster (Monaster} ETHERIDGE, jr. (part), Mem. Geol. Surv. New South Wales, Pal., No. 5, pt. 2, 1892, pp. 70, 71. Etheridgaster GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 353 (genoholotype, Palxaster clarlcd). Emended description. Animal large, very stout, with five thick petaloid rays. Disk very large, with marked ray angles. Abactinal area of rays with prominent radial, and supramarginal columns of large, contiguous, strongly convex, hexagonal plates, densely covered with small granules radially arranged. The ossicles of the radial columns alternate with those on each side and are largest in the mid-length of the rays, decreasing in size both distally and proximally, but most rapidly distally. This growth gives the ray columns a decided petaloid form. Outside of the supramarginals the animal is bounded by single columns of prominent, short, but very wide plates, which are the abactinal aspect of the actinal adam- bulacrals. These plates bear numerous spiniferous tubercles. The ossicles of the axillary disk region are not determinable. In the center of the disk there appear to be five distinct radials and a central disk plate, and around these probably were numerous small accessory pieces. Madreporite abactinal, large, oval, radially striate, and situated near the axillary margin. Actinally the rays consist almost entirely of the adambulacrals, which are numerous, abundantly tuberculate, short, and very wide. These columns increase rapidly in width, join and form angular axillse, and leave between themselves deep petaloid ambulacral grooves. The interbrachial areas consist of the oral extensions of adjoining columns of adambulacrals, of which there are a great num- ber of pairs, terminating in the mouth in five small pairs, the oral armature. KEVISION OF PALEOZOIC STELLEEOIDEA. 171 Ambulacrals in two columns, opposite, short but wide, and about as numerous as the adambulacrals. Podia! openings unknown. Genoholotype. Palseaster darlcei De Koninck. This is also the only species of Monaster and occurs in the "Upper and Lower Marine series" of the " Permo-Carbonif erous " of New South Wales. Remarks. Mr. Etheridge proposed Monaster as a subgenus of Palseaster, from which it differs so radically as to give it easily the rank of a genus. He defines the group as follows : In Palseaster " the adambulacral plates, bordering the ambulacral avenues, are small and quadrangular, followed by large transverse marginal plates. In our Permo-Carboniferous species, on the contrary the adambulacral plates are transversely elongated, and occupy nearly the whole of the actinal surface on each side the avenues. The marginal plates [having reference to the plates along the margins which consist of adambulacrals, and also form part of the abactinal skeleton], in con- tradistinction to those of Hall's Silurian species, are here smaller and subdorsal in position. The question now presents itself, of what value in a classificatory sense is this character ? Hall lays particular stress on the position of these plates on the actinal side of Palseaster. He says it 'has two ranges of plates on each side of the ambulacral groove; marginal [inframarginal] and adambulacral plates on the lower side, besides ambulacral or poral plates. The upper or dorsal side has three or more ranges of plates/ In the case of our specimens, only one set of plates, excepting those of the ambulacral grooves, are, as before stated, absolutely actinal; the marginals [same usage as before] are strictly so, or, at the least sub-dorsal. Under these cir- cumstances, I purpose distinguishing our Australian species under the subgeneric name of Monaster (from the one or single row of adambulacral plates on each side of an ambulacral avenue) ." The original definition is here given at length to show that it can only apply to the structure in Palseaster darkei De Koninck, which is also the first species following the generic discussion. The author in the same work describes a new species which he also refers to Monaster, Palseaster (Monaster) giganteus, but its structure is radi- cally different and does not conform with the subgeneric definition by Etheridge. Elsewhere in this work it is taken as the genotype of the new genus Australaster . Gregory in the serial above cited takes as the genotype of Monaster the last named species, the third of Etheridge, and transposes the former as the genotype to his new genus Eiheridgaster. This the writer holds can not be done even though Etheridge selected no genotype for Monaster. That author's genotype, as the name implies, refers to a species with but a single column of plates on each side of the ambulacrum. The species selected by a subsequent author should have the structure implied in the 172 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. original generic definition and when none actually applies, then the first species following the generic diagnosis. That Gregory fully appreciated the structural difference between Palseaster clarkei and P. (M.) giganteus is shown in his reference of the former to the order Phanerozonia and the latter to the Cryptozonia. Our disagreement is not as to structure but in the interpretation of the rules of nomen- clature. Under these circumstances P. clarkei is retained as the geno- type of Monaster, since it appears that P. giganteus has both infra- marginals and adambulacrals on the actinal side, a structure quite different from that described for and present in Monaster Etheridge. After the above was written in 1900, the writer explained by letter to Professor Etheridge what Gregory had done. Etheridge replied under date of June 6, 1900: "The laws of zoological nomenclature should have guided Doctor Gregory to have selected the first species described, as the subgeneric type, in the absence of any statement to the contrary on my part. As you say, Eiheridgaster is certainly a synonym of Monaster." Monaster is most like Hudsonaster. It retains the primitive pon- derous skeleton of the latter, but curiously has lost the inframarginal columns of ossicles, although the supramarginals are present. MONASTER CLARKEI (De Koninck). Palseaster clarkei DE KONINCK, Mem. Soc. roy. Sciences, Liege, ser. 2, vol. 2, 1877, p. 166, pi. 7, figs. 6, 6a; (David), Mem. Geol. Surv. New South Wales, Pal. No. 6, 1898, p. 127, pi. 7, figs. 6a, 66. Palseaster (Monaster) clarkei ETHERIDGE, jr., Mem. Geol. Surv. New South Wales, Pal. No. 5, pt. 2, 1892, p. 71, pi. 14, figs. 1, 2; pi. 15, fig. 4. Etheridgaster clarkei GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 353. Of this species three natural molds have been found in the "Upper Marine " and "Lower Marine' 7 series (Carboniferous) of Northum- berland County, New South Wales. The holotype in the Clarke collection was destroyed by fire. The species attained a large size, the rays from the center of the disk to their extremities are not less than 5 cm., and in another specimen are 6.5 cm. URASTERELLID^E, new family. Txniasteridse GREGORY (part), Geol. Mag., dec. 4, vol. 6, 1899, p. 351 (includes Tseniaster, fStenaster, Urasterella, Protasteracanthion, Salteraster). Roemerasterinx GREGORY, Treat. Zool., vol. 3, Echinoderma, 1900, p. 255. Specialized Cryptozonia, with alternate ambulacra, and with adam- bulacral type of oral armature. Rays five, rather flexible, long and gently tapering, proximally united without forming interbrachial areas. No inframarginals discernible as such at maturity, the actinal margin being occupied by well-developed adambulacrals. Abactinal area composed of numerous small plates arranged in columns and KEVISION OF PALEOZOIC STELLEKOIDEA. 173 quincunx. The radial and supramarginal columns may be dis- cernible in somewhat larger plates. Adambulacrals many, like coins set on edge. * Contains: Urasterella McCoy Genus URASTERELLA McCoy (emend.). Plates 27 to 30. Urasterella McCoY, British Pal. FOBS., 1851, p. 59 (not defined). HALL, Twentieth Kept. N. Y. State Cab. Nat. Hist., 1868, p. 289; rev. ed., 1868=1870, p. 332. McCoY, Geol. Surv. Victoria, dec. 1, 1874, p. 42. ZITTEL, Handb. Pal., vol. 1, 1879, p. 453. STURTZ, N. Jahrb. fur Min., etc., 1886, vol. 2, p. 152; Palseontographica, vol. 36, 1890, p. 219. GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 352. Stenaster BILLINGS (part), Geol. Surv. Canada, Can. Org. Rem., dec. 3, 1858, p. 77. STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, pp. 40, 56. Eoemeraster STURTZ, Palaeontographica, vol. 32, 1886, p. 85 (genoholotype, Asterias asperula Roemer); vol. 36, 1890, p. 220; Verh. naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, pp. 52, 73. Protasteracanthion STURTZ, Palseontographica, vol. 32, 1886, p. 90 (genoholotype, P. primus Stiirtz= Asterias asperula Roemer); Verh. naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, pp. 54, 75. Palasteracanthion STURTZ, N. Jahrb. fur Min., etc., 1886, vol. 2, p. 153 (genoholo- type, "A, primus "= Asterias asperula Roemer). STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, p. 75. Salteraster STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, pp. 43, 60 (genoholotype, Palseaster asperrima oalter). fEoactis SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palseontogr. Soc. for 1913), 1914, p. 30 (genoholotype, E. simplex Spencer=? Urasterdla girvanensis). Generic characters. Disk very small, without interbrachial arcs. Rays five, very long and slender at maturity, but shorter and rapidly tapering when immature. Abactinal side with very numerous small stellate plates, all of which, excepting one or three medial columns, are drawn out into more or less long, blunt, stout, erect, nonarticulating rods. The disk is occupied by a central plate and a first ring with five plates, which in some species are seen to be the basal radial plates, followed by a second ring with 1 5 plates. Five of the latter are the second proximal radial plates, and 10, basal plates of the supramarginal columns. Between the central plate and the first ring there may be inserted a few accessory, much smaller pieces. These disk ossicles may remain more prominent throughout life or may have the size of the other abactinal plates, in which case their arrangement is not plainly made out. In the center of the ray the adial column may be the more conspicuous throughout life or it may be more or less completely suppressed when the supramarginals adjoin and these may then 174 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. be the more prominent columns. Or the entire abactinal area of the rays may be composed of ossicles arranged in columns and in quincunx when the plates are of nearly equal size. The number of columns varies with age and in different species, the greatest number observed being 13 at the base of a ray. Ambital plates often not readily distinguishable at maturity from the radial and supramarginal columns. The growing distal tips of the rays of Urasterella have no ambital plates, but shortly after their introduction they begin to develop the rodlike extension so charac- teristic of this genus. Of these columns there may be as many as four at the base of a ray on each side of the supramarginals. Inframarginals, as such, usually not distinguishable at maturity from the ambital plates. Their position is rather abactinal than actinal. At the distal end of rays, however, these plates lose their ambital character, passing over somewhat to the actinal side, and closely adjoin the adambulacrals. In the young of U. ulrichi the inframarginals are well developed in the axillary regions and here there is also a single interbrachial marginal plate. More dis tally, however, the inframarginals have the character of the ambital plates. In none of the mature specimens have been seen well-developed or larger axillary inframarginals or interbrachial marginal plates. In U. asperula the axils have actinally single plates which may be remnant inframarginals. Madreporite abactinal, of medium size, slightly convex, depressed and granules triate. Adambulacral plates very numerous, coin-shaped, and arranged on edge with the actinal surface pustulose. Each plate on its actinal surface bears two or three short, thick spines, and on its ambulacral side there is another similar spine. Along the outer edge of these plates toward the abactinal side there is another row of spines, in pairs, which are long and slender, flat, and longitudinally grooved on two sides. The adambulacral columns terminate in small tri- angular plates of the oral armature. In the young of U. ulrichi five very stout, short, pointed spines (tori) are inserted inside of the plates of the oral armature. Ambulacral columns with the ossicles generally slightly alternating, but they may also be opposed, one plate to each adambulacral piece, more or less wedge- or club-shaped, and those of each column par- tially overlapping. Medially the columns loosely adjoin. Each plate is excavated laterally, along the proximal edge, leaving a more or less long, slender podial opening between adjoining plates. Genoholotype. Uraster ruihveni Forbes. The above diagnosis, however, is based on Stenaster grandis Meek, a form thought to be in generic harmony with the genotype. KEVISION OF PALEOZOIC STELLEKOIDEA. 175 Distribution. The species occur in North America and Europe and range from the Ordovicic to the Upper Carboniferous. The following are the known species: U. pulcJiella (Billings). Trenton. U. grandis (Meek). Richmond. U. fiuxleyi (Billings). ?Chazy. U. ulrichi, new species. Lowville. U. girvanensis, new species. Upper Ordovicic of Scotland. U. (?) asperrima (Salter). Bala of Wales. U. (?) constellata (Thorent). Siluric of France. U. ruthveni (Forbes). Ludlow of England. U. Jiirudo (Forbes). Ludlow of England. U. selwyni McCoy. Siluric of Australia. U. asperula (Roemer). Lower Devonic of Germany. U. } new species. Ithaca beds. Upper Devonic. U. montana (Stschurowsky). Upper Carboniferous of Moscow, Russia. Remarks.- McCoy in suggesting the name Urasterella did it in the following words: 11 Before I was aware Prof. E. Forbes had described them it seemed to me that the U. [raster] RutJiveni and U. Hirudo, as well the similar American species, might be easily separated from the great star- fishes forming the recent genus Uraster, by their small size and much more simple skeletons, and I had named the genus Urasterella in my manuscript." This is all that McCoy did to establish the genus Urasterella, abandoning his manuscript definition on learning that Forbes had described the species as those of Uraster. Salter 1 in 1857 paid no attention to this name, although he was the first to recognize and define Palasterina, which McCoy proposed at the same time and in the same manner. It seems that Hall (1868, p. 332) was the first to revive the name Urasterella and regarded Billings's Stenaster as a synonym by taking his second species, 8. pulchella, as the genotype. Hall's remarks are as follows: "Mr. Salter, apparently overlooking my description of Pal&aster niagarensis, has placed both the above- cited European species [ Uraster rutTiveni and U. Jiirudo] under Palae- aster, as defined by himself; but as it has been shown that this genus has two ranges of plates on each side of the ambulacral groove, these forms can not with propriety be arranged with typical Palseasters. " There was the same reason for adopting the generic term Uras- terella as for adopting Palasterina, both of which were proposed by McCoy at the same time, and the typical forms of each were specified. Mr. Salter has for some reason recognized the genus Pabea,ster pro- iAnn. Mag. Nat. Hist., ser. 2, vol. 20, 1857, p. 326. 50601 Bull. 88 15 12 176 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. posed by me a little earlier, but probably published after that of McCoy; and has given the two species of Urasterella to that genus. Now that there seems no reason for continuing them under that generic term, they should be restored to their proper position as indicated above. The two species of Forbes must either be placed under Urasterella as proposed by McCoy, or fall under the later genus Stenaster of Billings. Were the latter genus of my own proposing I should nevertheless return to Urasterella, and I have no doubt that Mr. Billings will accord with this view." Palseaster pulcJiellus is a true Urasterella, but this does not make the genus Stenaster a synonym of Urasterella, since S. salteri Billings has quite another structure and is here accepted as the genotype of Stenaster. Meek 1 also recognizes McCoy's name, but because of certain pecul- iarities in S. salteri, which he recognizes as the genotype, and also since he had not seen the dorsal side of McCoy's S. grandis, he prefers to retain Stenaster for the latter. McCoy returned to his genus in 1874, after going to Australia, and described a new form, U. selwyni. It is important to quote here his remarks : " Genus Urasterella (McCoy )= Stenaster (Billings). "Gen. Char.: Small starfishes, with five moderate rays, narrowed at the base, and without disk. Ambulacral grooves narrow, bor- dered on the under side, with only one row of large (adambulacral) plates; no marginal plates. Upper surface with numerous rows of small tubercular plates. Confined to Silurian rocks. "The late Mr. Salter and Mr. Billings refer the starfishes of this type to the subsequently published genus Pateaster of Hall; but, as Prof. Hall objects that his genus Palseaster has ambulacral, adam- bulacral, and marginal plates, and the types of my genus U. ruthveni and U. Jiirudo of the English Ludlow rock, like our Australian species and the American Pals&aster or Stenaster pulchella, have only one row of plates on each side of the ambulacral groove, I return to the use of my old generic name." Zittel 2 says that Stenaster is a synonym for Urasterella and defines the latter genus. Stiirtz 3 points out the errors in the diagnosis of Billings and that of Zittel, redefines the genus, and correctly states that Stenaster Billings contains species of two genera, that is, Sten- aster salteri, the genotype of Stenaster, and S. pulchella, which he refers to Urasterella. In this way the genus Urasterella has come to be recognized. The figure and description of Uraster ruthveni given by Forbes* appears to be of a species generically identical with the American 1 Pal. Ohio, vol. 1, 1873, p. 67. s Palseontographica, vol. 36, 1890, p. 219. 2 Handb. Pal., vol. 1, 1879, p. 453. * British Org. Rem., dec. 1, 1849, p. 1, pi. 1, fig. 1 . REVISION OF PALEOZOIC STELLEROIDEA. 177 Stenaster pulchellus Billings and this is the general consensus of opinion. The writer unfortunately had no material of the genotype of Uras- terella for study. However, since U. grandis is closely related to U. pulchella and shows so much more detail, it is here accepted as the genotype for the time being. Professor Hall (1870) wrote: " In Stenaster = Urasterella, however, we have the absence of adambulacral plates." This is nevertheless not the case, for the corn-shaped marginal series in S. pulchellus are the ambulacral plates, since weU-developed inframarginal plates, as in Palseaster, are absent in mature Urasterella. The inframarginals are, however, seen to be present in U. pulchella or in the distal ends of the rays in other species, though proximally they are so modified as not to be readily distinguishable from the ambital plates. In the axils of the young of U. ulrichi the inframarginals are well developed, along with single interbrachial marginal plates, but in mature speci- mens of other species none have been seen. Of Roemeraster the writer has seen three excellent specimens skil- fully prepared by Sttirtz which clearly show that this genus has the generic structure of Urasterella. Stiirtz 1 describes this genus as having "strong marginal plates" inside of the "small, cornered adam- bulacrals." Our specimens do not show such plates, and if present they can not be adambulacrals. This fact can be demonstrated in a well-preserved distal end of a ray, where two columns are present on each side of the ambulacrals. The innermost column of closely adjoining plates increase in size, retain iheir position beside the ambu- lacrals, and become the most prominent of the entire animal, while the outermost column soon change in aspect, the plates become separated and assume the characters of the ambital pieces. This is likewise true in the young of U. ulrichi where the axillary areas also have distinct interbrachial marginal and two larger proximal inframarginal plates. Passing distally, the latter also soon change then- form, are separated one from another and are not distinguish- able from the ambital plates. Stiirtz further describes and figures five interbrachial ossicles in mature Urasterella asperula which are also well developed in the young of U. ulrichi. Urasterella is an easily recognized genus and the triangular ambital plates drawn out into short blunt erect rods are peculiar to it, a character again re- peated in Roemeraster. Under these circumstances the latter appears to be indistinguishable from Urasterella. For other remarks see U. asperula. On the other hand, Gregory regards Roemeraster as a good genus and places it erroneously among the ophiuridlike asterids in the family Tseniasteridse. i Verb, naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, p. 52. 178 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. In regard to Protasteracanthion Stiirtz, see remarks under Uras- terella asperula, and for Salteraster Sturtz see U. asperrima. Urasterella differs from Stenaster as redefined in this work in having long, slender, flexible rays which are never short and petaloid as in the latter genus. Again, the adambulacral pieces of Urasterella are very numerous, thin, and coin-shaped, while in Stenaster they are few in number and comparatively thick; the ambulacral plates are wholly different. The long, slender and flexible rays, the apparent absence of in- framarginal columns, and the absence of distinct interbrachial plates (these are not the same as the single axillary pieces seen in U. asperula) at maturity readily distinguish this genus from all other Paleozoic genera. For the origin of Urasterella see U. ulrichi. Eoactis. After this book was completed appeared the work of Spencer, with the new genus Eoactis. It is probable that the genus is good and can be distinguished from Urasterella, but the necessary comparisons can not now be made. URASTERELLA PULCHELLA (Billings). Plate 28, figs. 3, 4; plate 30, fig. 5. Palseaster pulchellus BILLINGS, Geol. Surv. Canada, Rep. Progress for 1853-1856, 1857, p. 292. Stenaster pulchellus BILLINGS, Geol. Surv. Canada, Can. Org. Hem., dec. 3, 1858, p. 79, pi. 10, fig. 2. WRIGHT, Mon. British Foss. Echinod., Oolitic, vol. 2, pt. 1 (Palaeontogr. Soc. for 1861), 1862, p. 28. Urasterella (Stenaster) pulchellus HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 289; rev. ed., 1868=1870, p. 332. The largest specimen from Middleville, New York (Hall collection), measures: 11 = 52 mm., r=ll mm., R = 4.8r. An individual of the usual size from Trenton Falls, New York (Mus. Comp. Zool., No. 31), measures: R = 30 mm., r = 4. mm., R = 7.2r. One from Ottawa (Geol. Surv. Canada collection) measures: R = 25 mm., r = 4 mm., R = 6.1r. The smallest specimen seen (Mus. Comp. Zool., No. 36) measures: R = 7 mm., r = 1.5 mm., R = 4.7r. Width of rays at base in mature examples from 4.5 mm. to 7 mm. Original description. " Diameter, 2J inches; rays subcylindrical, 2J lines in width at the base, with a length of 1 inch; disk 3J lines in diameter; grooves narrow, bordered throughout by narrow oblong plates, nine in the length of two lines; the length of these plates in a direction transverse to the rays is about 1 line; near the disk there appears to be but one row of marginal plates." Formation and locality. Restricted to the Trenton limestone. Most abundant at Trenton Falls, New York. From this locality REVISION OF PALEOZOIC STELLEROIDEA. 179 there are two specimens in the New York State Museum and six in the Museum of Comparative Zoology, Harvard University, col- lected by Dr. C. D. Walcott (Nos. 31-36). The American Museum of Natural History has one specimen from Canajoharie, New York (No. 660) ; in the Hall collection, Albany, there is a specimen from Middle viUe, New York. At Hull, Canada, Mr. W. R. Billings col- lected two specimens; the type and another specimen were found by E. Billings at Ottawa, Canada; these are now in the Museum of the Geological Survey of Canada at Ottawa. Remarks. This graceful species has the essential structure of U. grandis and will therefore not require a detailed description. U. pulchella is from a lower geological horizon, is probably the an- cestor of U. grandis and has far less numerous abactinal plates, which on the disk reveal a primitive arrangement reminding one of Hudsonaster. In the center of the disk there is a single plate around which is a first ring of 5 basal radial plates followed by another with 15 plates. The plates of the second ring are a little more conspicu- ous than any other abactinal plates and are disposed as follows: Five are radial in position and are the second basal plates of the 5 radial columns, while the other 10 are disposed interradially in pairs immediately on each side of the second basal radials. The plates interradial in position are the basals of the supramarginal columns. Between the central ossicle and first ring there appear to be a few small accessory plates. The arrangement of the disk pieces above described is plainly preserved in a very young example (No. 36, MILS. Comp. Zool.). The radial columns remain the most prominent while the supra- marginals a little distally from the base of the rays are not readily distinguishable from the other columns. Here and there an accessory piece may be inserted between the radials and supramarginals but the greater width of the rays is attained by increasing the size of these columns and by the introduction of ambital columns outside of the supramarginals. In a mature ray on each side of the radials proximally, there appear to be not more than four columns of plates but it is possible that a few pieces of a fifth column may be developed. The form and spinosity of the abactinal plates is as in U. grandis, except that the nonarticulating spines are less long in U. pulchella. Inframarginal columns at first view do not appear to be present but a careful examination near the distal end of a ray reveals two columns of nearly equal-sized plates on each side of the ambulacral furrow. Here these plates closely adjoin, the adambulacrals are elongate subquadrangular and the inframarginals quadrangular in shape. Proceeding proximally the former become more prominent and wider but remain throughout closely adjoining. The inframar- ginals, however, gradually pass over to the abactinal side although 180 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. they remain beside the adambulacrals, become more and more spinose, longer than wide, and eventually stellate in form and then are inseparable from the abactinal plates. There are about 42 adam- bulacrals in each column. U. pulchella is readily distinguished from U. ulricJii and U. grandis in being smaller, with more slender and graceful rays, and in its fewer columns of plates. From U. Jiuxleyi it is distinguished by the less convex rays and the far smaller number of plates. Cat. No. 60610, U.S.N.M. URASTERELLA GRANDIS (Meek). Plate 27, figs. 6-8; plate 28, figs. 1, 2; plate 30, figs. 1-4. Stenaster grandis MEEK, Amer. Journ. Sci., ser. 3, vol. 3, 1872, p. 258; Geol. Surv. Ohio, Pal., vol. 1, 1873, p. 66, pi. 3 bis, figs. 7a-7c. MILLER, N. Amer. Geol. Pal., 1889, p. 283, fig. 432. Urasterella grandis MEEK, Geol. Surv. Ohio, Pal., vol. 1, 1873, p. 67. Palseaster harrisi MILLER, Journ. Cincinnati Soc. Nat. Hist., vol. 2, 1879, p. 117, pi. 10, figs. 2, 2, LOWER TRANS- VERSE MUSCLE; Mv, VERTICAL MUSCLE ; Bo, UPPER MARGINALIA; Ru, LOWER MARGINALIA; Wr, RADIAL WATER- VESSEL. FlG. 13. A SECTION THROUGH THE ARM OF A TRUE OPHIURID, AFTER SCHONDORF. Bs, VENTRAL SHIELD; C^CUTIS; F, PODIA; L, BODY CAV- ITY; N, NERVE-RING; oZm, uZm, UPPER AND LOWER INTERMEDIARY VERTEBRAL MUSCLES; Rs, DORSAL SHIELD; Ss, LATERAL SHIELD; W, VERTEBRA; Wr, RADIAL WATER- VESSEL. 214 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Ad. FIG. 14. A SECTION THBOUGH THE ARM OF AN AULUBID, WITH ALTERNATING AMBULACEALIA, AFTEE SCHON- DOEF. A, AMBULACEALIA OF THE DOESAL AND VEN- TRAL SKELETON; A 1 , VENTEAL EXTENSION OF THE AMBULACEALIA; Ad, AD AMBULACEALIA = LATERAL SHIELDS OF OPHIUEIDS; F, PODIA, ON LEFT THE CANAL PASSES THBOUGH THE PLATE, ON EIGHT THE PLATE IS DISSECTED DOWN TO THE CANAL; Wr, INTEESKELETAL EADIAL WATEE- VESSEL. are covered with spines and tubercles.] Disk with concave, or con- vex ( ?) margins, with or without marginal ossicles. When the latter are present, they never extend along the rays, but are wholly re- stricted to the disk. The rays on both sides are margined by the adambulacrals. Ventrally the rays have broad open ambulacral fur- rows, bounded laterally by the adambulacrals. A typical mad- reporite [probably always] lies in one of the ventral interradii. "The group [subclass] is re- stricted to the older Paleozoic." Remarks. This subclass of Stelleroidea appears to be an in- dependent development whose structure partakes of that of the Asteroidea and the Ophiu- roidea, though more like the latter. The Auluroidea agree with the asterids in having open ambulacral furrows, and a true madreporite, which is, however, ven- trally situated, but otherwise the class is more like the ophiurids, since the body cavity does not extend into the rays as in true starfishes. However, the water-vascular canal in the Auluroidea does not lie outside of the ambulacrals as in the Asteroidea, but within these ossicles along the mid-line of the ambulacrum in a canal that is cut out of the sides of adjoining ambulacral columns. On the other hand, the Ophiu- roidea have the am- AA A A i* A4 A A Ad bulacralia coossified and modified into vertebral ossicles; there are no open ambulacral furrows, and the entire rays are covered by four columns of shields or by an integument more or less studded with calcareous plates and granules. It appears that the Auluroidea all have dorsally four columns of plates, of which the two medial ones are the most prominent. These are undoubtedly the dorsal aspect of the thickened ambulacrals. Outside of the disk they are usually convex and more or less orna- mented with granules, but within the disk are less prominent and take on other expressions, commonly with concave surfaces. The plates outside of these margin the rays and are the adambulacrals; while in some forms they are prominent and tuberculate, usually they are made up of narrow vertical pieces that bear spines along the ventral or only the distal ventral edge. FIGS. 15 AND 16. VENTEAL AND DOESAL AEM STBUCTUEE OF AULUBOIDEA, WITH THE AMBULACEA ALTEENATING. SCHEMATIC, AFTEE SCHSNDOEF. A, AMBULACEALIA; A 1 , VENTEAL EXTENSIONS OF AMBULACEALIA; A 3, AD AMBULACEALIA; FS, PODIAL CAVITIES. KEVISION OF PALEOZOIC STELLEKOIDEA. 215 The Auluroidea can be divided into two orders on the basis of whether the ambulacrals on each side of the median line are arranged alternately (the Lysophiurae), or are opposite to one another (the Streptophiurae). The following classification is a modification of the one offered by Gregory in 1897, and kis work was in turn a modification of that of Bell in 1892. After the present work was written, its author first became aware of the works of Sollas and Sollas, and Spencer. The reader should consult those papers for important studies on the Auluroidea. Order LYSOPHIUR^E: Gregory. Encrinasterise BRONN (part), Klassen u. Ordnungen d. Tierreichs, vol. 1, 1859, p. 287. SCHONDORF, Palaeontographica, vol. 57, 1910, p. 61. Ophio-Encrinasterise STURTZ, Verb, naturh. Ver. preuss. Rheinl., etc., vol. 56, 1900, p. 198. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, pp. 234, 237. Lysophiurse GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1030 (contains Protasteridse and Palaeophiuridae) ; Treat. Zool., vol. 3, Echinoderma, 1900, p. 274. SOLLAS and SOLLAS, Philos. Trans. Roy. Soc. London, ser. B, vol. 202, 1912, p. 223. Ophiurasterix SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 251. Auluroidea in which the ambulacral ossicles of adjoining columns are alternate to one another. The individual plates are separate and not coossified into vertebrae. Remarks. The Lysophiurae are Paleozoic or primitive Auluroidea, that is, ophiurid-like animals, with wide and open ambulacral furrows in which the ambulacralia are free and alternately arranged. The ambulacralia are common to the actinal and abactinal sides. This order contains the following families: Protasteridse. Have boot-shaped ambulacralia ventrally. Palseophiuridae. Have subquadrate ambulacralia ventrally. Encrinasteridae. Have boot-shaped ambulacralia ventrally and large disks that are bordered by marginalia. Family PROTASTERIDSE Miller. Protasteridse MILLER, N. Amer. Geol. Pal., 1889, p. 216. GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1031; Treat. Zool., vol. 3, Echinoderma, 1900, p. 274. PARKS, Trans. Canadian Inst., vol. 8, 1908, p. 366. Paldeophiomyxidse STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 56, 1900, p. 202. "Lysophiurae with boot-shaped ambulacral ossicles, each composed of a 'body ' in the median line of the arm, and a lateral 'wing' at right angles to it" (Gregory). There is either a well-marked scale-covered or granular disk, and five flexible arms. The adambulacral plates are rather stout, with many lateral spines. 216 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Schondorf states that Gregory's family Palseopliiuridae has no value and is to be merged into the one under discussion. The reasons for this are given beyond under the former family heading. The family contains: Tseniaster Billings. Alepidaster Meek. EopTiiura Jaekel. Gregoriura Chapman. Bohemura Jaekel. Bundenbachia Stiirtz. Palseura Jaekel. PalseopJiiomyxa Stiirtz. Protaster Forbes. Genus T^ENIASTER Billings. Plate 36, figs. 1-3. Tseniaster BILLINGS, Geol. Surv. Canada, Can. Org. Rem., dec. 3, 1858, p. 80. CHAPMAN, Canadian Journ., n. ser., vol. 6, 1861, p. 517. WRIGHT, Mon. British Foss. Echinod., Oolitic, vol. 2, pt. 1 (Paleeontogr. Soc. for 1861), 1862, pp. 24, 34. HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 300; rev. ed., 1868=1870, p. 338. ZITTEL, Handb. Pal., vol. 1, 1879, p. 445. STURTZ, N. Jahrb. fur Min., etc., 1886, vol. 2, p. 150; Palseontographica, vol. 32, 1886, pp. 78, 83. MILLER, N. Amer. Geol. Pal., 1889, p. 285. GREGORY, Geol. Mag., dec. 3, vol. 6, 1889, p. 26. STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, p. 20. J. F. JAMES, Journ. Cincinnati Soc. Nat. Hist., vol. 18, 1896, p. 138. GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1035; Treat. Zool., vol. 3, Echinoderma, 1900, p. 250. SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Paleeontogr. Soc. for 1913), 1914, pp. 8, 19. Taeniura GREGORY, (not Tseniura Miiller and Heule 1837, sting-ray), Proc. Zool. Soc. London for 1896, 1897, p. 1035. Original description of Tseniaster. " Generic Characters. Body deeply stellate; no disk [there is a disk] or marginal plates; rays long, slender, flexible, and covered with small spines; two rows of large ambulacral pores; adambulacral plates elongated and sloping outward so that they partly overlap each other; ambulacral ossicles contracted in the middle, dilated at each end. Generic name from tainia, a riband." Genolectotype (the first species of Billings is here selected as the genolectotype) . Palxocoma spinosa Billings. Remarks. Hall restudied the type-material of T&niaster and records his observations as follows : "In reviewing the characters of Protaster and Eugaster, I became satisfied that there was an intimate relation between these and Tseniaster of Billings, and in order to satisfy myself on this point, I have * * * received from Sir William E. Logan, permission to examine specimens of T&niaster spinosus and T. cylindricus. "An examination of the specimen illustrated in fig. 3 (ut sup.) reveals what I conceive to be a disk not at all unlike the disk of Prot- aster, but less extended than in the Lower Helderberg species. The structure of the ray is precisely of the same character as the ray of REVISION OF PALEOZOIC STELLEKOIDEA. 217 that species which I have named Protaster forbesi, the proportions of plates and relations of parts showing specific differences/' "Mr. Billings remarks, under the description of T. spinosus, that ' the ambulacral ossicles appear in some places to alternate with each other, but this is owing to a distortion; those on one side of the furrow are opposite to those upon the other. 7 "Now the specimen of this species which I have examined, and which I suppose to be the one figured upon Plate X, figures 3d, Z>, has the ambulacral plates alternating * * *. "In the specimen of T. cylindricus examined the ambulacral plates are less distinctly alternate, but the relation of the adam- bulacral plates and the pores are the same as in the other form. Not having seen the specimen showing the dorsal view, figure 4o- of plate X, I can only remark that the structure of the rays is very similar to that of Protaster." Without restudying the type-material of Tseniaster, or paying attention to Hall's statements that T. spinosus has what he con- ceived "to be a disk not at all unlike the disk of Protaster," Gregory (1897) regarded Tseniaster as a Cryptozonia starfish and made of Billings's second species, T. cyliiidricus, a new genus, Tseniura. The latter he correctly refers to the primitive ophiurids. Under these circumstances it is desirable to repeat here what Gregory wrote about Tseniaster. Tseniaster "was described as diskless, which in respect to the type species T. spinosus is correct. In that species the oral skeleton consists of five pah's of large adambulacrals, as in ordinary Asterids. The affinities of the true Tseniaster appear to me to be with such forms as Palssaster ruthveni [= Urasterella ruthveni]. It is asteroid in the oral armature, in its alternately arranged ambulacral ossicles, and in the absence of a disk. I therefore consider Tseniaster a genus of Asteroidea. The second species placed by Billings in this genus has, however, a well-marked disk, and has the oral armature com- posed of five pairs of Ophiuroid syngnaths. It must, therefore be included among the Ophiuroidea." To adjust this matter, the writer restudied the genotype of Tseniaster at Ottawa (two specimens, No. 1404, the originals of figures 3a-3d of Billings's Plate X). There can be no doubt that we have here a genuine lysophiurid. There is a small disk present in the individual figured as 3a and 3Z>, but it is not one with a round outline, as in typical ophiurids, but is concave along the margin between the rays. Actinally this disk is abundantly covered with prominent but slender spines. The ambulacrals are distinctly alternate throughout the rays and ap- pear as in the sketch presented on page 218; this condition is not due to distortion or other causes. The side plates are narrow, and each 218 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. bears three spines (see fig. 17B); toward the mouth the two columns diverge far more than one would assume from the original figure, and in consequence there are long and slender syngnaths and not the asterid-like mouthpieces figured in the illustration. (See fig. 17C.) 3 FIG. 17. DIAGRAMS OF T^NIASTEB SPINOSUS (BILLINGS). A, To SHOW THE SPINOSE DISK WITH CONCAVE SIDES. B, ACTINAL PLATE ARRANGEMENT: C, AMBULACRALS; ad, ADAMBULACRALS OR SIDE PLATES, WITH THEIR SPINES. C, MOUTH PLATES: a, AMBULACRALS; S, SYNGNATH. The specimen of figures 3c and 3d is too poorly preserved to add anything further. Associated with these types there is another and much larger specimen with a well-marked disk, of which Billings said nothing in his original description. This form is clearly of another species (probably T. cylindricus) , and is mentioned here so that other workers will not confound it with the original material of T. spinosus. In 1900 Gregory referred Tseniaster to the asterid family Tsenias teridse, and placed here as well the unrelated genera Stenaster, S alter- aster, and Urasterella. From what has been stated above it is plain that T&niaster can not be closely related to the cryptozonian genus Urasterella , since the latter has an abundance of narrow ambulacralia, asterid in type, while these ossicles in the former are far fewer in number and of the form seen in other and related Lysophiurse. It is very probable that Tseniaster -arose in the same stock that gave rise to Urasterella and Stenaster, but the former genus is clearly one of the aulurids. In 1896 Gregory took out of T&niaster the second species of Billings (T. cylindricus) and based on it the new genus Tseniura. As will be seen, this name can have no standing, as the generic char- acters are those of Tseniaster, and, further, the name is preoccupied since 1837. Thinking T. cylindricus most closely related to the primitive ophiurids PalseopTiiura and St&rteura, Gregory defined T&niura as follows : "Diagnosis: Palseophiuridse with a small pentagonal disk, not bounded by marginal ossicles. The ambulacral furrow is broad. The oral skeleton is conspicuous and the syngnaths each composed of two separate pieces. The two jaws of each oral angle are closely attached; the mouth frames are separated, and each of them is a short, thick, slightly bent bar." REVISION OF PALEOZOIC STELLEEOIDEA. 219 Genotype and only species. Palxocoma cylindrica Billings. 11 Remarks: This genus differs from PalseopTiiura by the absence of marginal ossicles from the disk, and from Sturtzura by the smaller size of the disk and the form of the syngnaths. In Sturtzura the jaws end bluntly against a jaw plate, whereas in Tseniura they appear to taper to a point and have no jaw plate." The two cotypes of T. cylindricus (1405a) were lost for many years, but have recently been rediscovered in the collections at Ottawa. In addition, there is another specimen in that Museum under this name, and it is probable that it was so identified by Billings (No. 1405). This was the only specimen studied, and its generic characters agree very well with the originals of the cotypes (figs. 4b and 4c of pi. X), and appear to be in harmony with the genotype of Tseniaster. The only important difference is that here the ambulacralia are directly opposite one another, and not slightly alternating, as in T. spinosus (see fig. 18), a difference, it seems to the writer, not in itself alone worthy of being the base of a new genus distinct from Tseniaster. Tseniaster has the following species: T. spinosu^s (Billings). Lower Trenton. T. cylindricus (Billings). Trenton. T. schofiarize Ruedemann. Upp^r Trenton. T. elegans Miller. Richmondian (Waynesville). T. meafoadensis Foerste. Richmondian. TXENIASTER SPINOSUS (Billings). Plate 36, fig. 1; text fig. 17. Palseocoma spinosa BILLINGS, Geol. Surv. Canada, Rep. Progress for 1853-1856, 1857, p. 292. Tseniaster spinosus BILLINGS, Geol. Surv. Canada, Can. Org. Hem., dec. 3, 1858, p. 81, pi. 10, figs. 3a-3d. MILLER, N. Amer. Geol. Pal., 1889, p. 285, fig. 439. Revised description ly BiUings. "The specimens collected are about seven lines in diameter; the rays linear-lanceolate, one line in width at the base, and covered at the sides with numerous small spines. "In the view of the enlarged specimen (pi. X, 3s), the ambulacral ossicles appear in some places to alternate with each, other, but this is owing to a distortion. Those on one side of the furrow are oppo- site those upon the other. The adambulacral plates are elongated, and so placed that the outer extremity of the one lies upon the inner extremity to the next. The rays are flexible." For remarks on this species, see the discussion under Tseniaster. Formation and locality. Two specimens from the Lower Trenton, Falls of Montmorency, east of Quebec, Canada. The original of figures 3a to 3d is in the Victoria Memorial Museum, Ottawa (No. 1404). 220 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. T^NIASTER CYLINDRICUS (Billings). Plate 36, figs. 2, 3; text fig. 18. Palseocoma cylindrica BILLINGS, Geol. Surv. Canada, Rep. Progress for 1853- 1856, 1857, p. 292. Txniaster cylindricus BILLINGS, Geol. Surv. Canada, Can. Org. Rem., dec. 3, 1858, p. 81, pi. 10, figs. 4a-4c. WRIGHT. Mon. British Foss. Echinod., Oolitic, vol. 2, pt. 1 (Palseontogr. Soc. for 1861), 1862, p. 34. PARKS, Trans. Cana- dian Inst., vol. 8, 1908, p. 371. Txniura cylindrica GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1035. Lapworthura cylindrica PARKS, Trans. Canadian Inst., vol. 8, 1908, p. 371. Original description. "One inch and a half in diameter; rays five [very flexible], covered with spines, subcylindrical, regularly rounded on the upper side, flattened on the lower, about one line in width at base, and regularly tapering to an acute point." Formation and locality. In the Tren- ton limestone at Ottawa, Canada. The . i8.-DiAGRAMOp7cTmAL BAY-PLATES cotypes (14050), along with a third OF TJSKIASTKR CYLINDKICUS BILLINGS, specimen (1405), are in the Victoria a, AMBTJLACKAL; ad, SIDE PLATES. _* . , v , Memorial Museum at Ottawa. T^ENIASTER SCHOHARL Ruedemann. Tseniaster schoharise RUEDEMANN, Bull. N. Y. State Mus., No. 162, 1912, p. 88, pl. 3, fig. 1. Original description. The holotype "is small, the arms about 15 mm. long, and it is without a disk. The arms are slender and flexible, about 1.3 mm. wide at the base, and almost as high in lateral view, originally probably cylindrical. The ventral view shows a straight or slightly zigzagged ambulacral canal, and on both sides of this squarish depressions surrounded by the ambulacral and adambulacral ossicles, the covering lower arm plates not being retained if they existed. The inner (ambulacral or vertebral) ossicles appear as narrow and outwardly curved ridges, the outer or adambulacral ossicles as ridges bent in the opposite direction with a projection in the middle of the outer arch. The ambulacral ossicles are not directly opposite nor regularly alternating, but those of the right side on the ventral view advanced about one-third the length of the ossicles beyond those of the other side. "The lateral view of one other arm exhibits the ossicles as ver- tical bars, that are thickened at both ends resembling vertebrae and terminating at the dorsal side with a flat surface and forming there an apparently continuous layer. They appear almost twice as wide as the intervening spaces. The dorsal surface bears bundles of obliquely forward directed spines, one bundle corresponding to each ossicle. These spines appear in the right arm to proceed from REVISION OF PALEOZOIC STELLEROIDEA. 221 small spine-bearing plates attached to the adambulacrals, as in the closely related Bundenbachia. On one arm two bundles are seen to proceed from each segment, one oblique, the other vertical to the arm. The oral skeleton consists of stout, bifid pieces, much resembling the oral pairs of adambulacrals of the genotype." Formation and locality. Upper Schenectady formation, the eastern shale phase of the Upper Trenton limestones. From near Schoharie Junction, New York. The holotype is in the New York State collection at Albany. T^ENIASTER ELEGANS Miller. Tseniaster elegans MILLER, Journ. Cincinnati Soc. Nat. Hist., vol. 5, 1882, p. 41, pi. 1, figs. 6-6c. J. F. JAMES, Journ. Cincinnati Soc. Nat. Hist., vol. 18, 1896, p. 139. PARKS, Trans. Canadian Inst., vol. 8, 1908, pp. 365, 371. Protaster elegans PARKS, Trans. Canadian Inst., vol. 8, 1908, p. 368. Original description. "This species is founded upon more than 30 specimens occurring on a single slab, but showing only the ventral side, with the exception of the ends of some of the rays. "The body is deeply stellate, and the rays long, slender, and flexible, and margined on either side with a row of spines. "From the length of some of the rays, a complete specimen, properly spread, would probably have a diameter of an inch and a quarter, or more. The rays taper to a sharp point, are rounded on the dorsal side, and margined on the ventral side by a row of spines, one spine arising from each plate. Only two rows of plates seem to form the dorsal side of a ray. The ventral side of each ray is marked by a furrow in the middle, separating two series of plates or ossicles. These plates are a little longer in the direction of the ray than wide, and alternately break joints at the middle, where they are slightly contracted." Formation and locality. This extraordinary slab is in the Harris collection of the United States National Museum. It was found near Waynesville, Ohio, in the Richmond formation of the Ordovicic period. Cat. No. 40878, U.S.N.M. TffiNIASTER MEAFORDENSIS Foerste. Txniaster meafordensis FOERSTE, Bull. Sci. Lab. Denison Univ., vol. 17, 1914, p. 326, pi. 4, figs. 5-7. The reference to this species is inserted as this book is going through the press. The form appears to be common in the lowest Richmon- dian on Workman Creek, three miles southeast of Meaford, Ontario, Canada. The types are probably in the Victoria Memorial Museum, Ottawa, Canada. 222 BULLETIN 88, UNITED STATES NATIONAL MUSEUM, Genus EOPHIURA Jaekel. Text figs. 19, 20. EopUura JAEKEL, Zeits. geol. Gesell., vol. 55, 1903, Protokol, p. 14 (107), figs. 1, 4. GenoJwlotype. EopMura, species not named. To give this genus proper standing according to the rules of nomenclature, the form is here named E. bohemica, new species. It is the only species of the genus, and occurs in the Ordovicic (Dj) at Osek, Bohemia. The genus is not defined in the regular manner, but from the general discussion and the illustrations the following has been made out: Kays outside of the disk, apparently petaloid i n form, with wide-open am- bulacral furrows, having narrow alternating ambu- lacralia that are \ shaped, with the podial openings excavated on each side of the lateral extensions. The adambulacrals are very large, drawn out lat- erally, rectangular, lie hori- zontally, and in no way resemble the side plates of typical ophiurids, or, for that matter, even the Lys- ophiurse. The lateral edges of these ossicles are slightly elevated, tubercu- lated, and each bears about four long and very slender spines. Inside of the disk the adambulacrals are nar- row and stand vertically as in Paleozoic ophiurids ; the columns first diverge The two proximal pieces of mouth extensions, with the FIG. 19. END OF A KAY OF EOPHIUKA BOHEMICA, MUCH EN- LARGED, AFTER JAEKEL. a, AMBULACRALIA; ad, ADAMBULA- CRALIA; Afg, PIT OF PODIA; ads, LATERAL EXTENSION OF AD AMBULACRALIA; Rr, RADIAL FURROW; St, SPINE BASES. and then converge in the mouth area. each column make the adambulacral terminal ossicles of adjacent columns side by side but not coalesced, forming the syngnaths. The two most proximal ambulacrals rapidly decrease in size, enlarging the mouth area, with none of these ossicles on each side of the two final adambulacrals. The disk seems to be small, with concave interbrachial sides. Actinally these areas are covered with scale ossicles. Madreporite ventral, in the outer and lateral part of an interradius. Immediately outside of the syngnaths occur interradially two small ossicles, reminding one of vestigial inframarginalia. REVISION OF PALEOZOIC STELLEROIDEA. Genus BOHEMURA Jaekel. Text fig* 21. 223 Bohemura JAEKEL, Zeits. geol. Gesell., vol. 55, 1903, Protokol, p. 18 (111), fig. 6. This genus, like the foregoing, is not defined in the regulation way. Jaekel's fine drawing shows that it is closely related to Eophiura, differing mainly in that the rays do not end bluntly, but have slowly tapering whip-like terminations. These ends are of course ontoge- netically young and reveal the progressive ray development. Here FIG. 20. ORAL SKELETON OB- EOPHIUEA BOHEMICA, AFTER JAEKEL. a, AMBULACRALIA; ad, ADAMBU- LACRALIA; Ma, MADREPORITE AMONG THE SCALE PLATES. the rays are made up of overlapping and alternating adambulacrals that distally have no ambulacral grooves. Gradually an ambulacral groove is developed and it becomes rapidly wider as one examines the rays more and more proximally, and the adambulacrals turn over and flatten out, finally making the great open ambulacral furrows. The illustration here reproduced, figure 21 on page 224, will show the generic characters as drawn by Jaekel. Oenoholotype and only species. B. jahni Jaekel. Ordovicic (DJ at Zahorzan, Bohemia. Genus PALvEURA Jaekel. Text fig. 22. Palseura JAEKEL, Zeits. geol. Gesell., vol. 55, 1903,Protokol,p.l6(109)-17(110),fig. 5. Oenoholotype. Palseura, species not named. No species is again mentioned, nor is it stated that the generic name is new. To give the name standing, the form figured is here named P. neglecta, new genus, new species. It occurs in the Ordovicic (D t ) of Bohemia; the exact locality is not given, but may be Osek. 50601 Bull. 88 15 15 224 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Jaekel here again fails to diagnose his genus, but some of the characters can be clearly made out from his excellent drawing. The disk is small and tends to have convex interbrachial areas. So far as the oral disk area is concerned, the general structure is that of Eophiura, with this marked difference the opposite and asterid- like ambulacralia of the rays do not vanish proximally as in that genus but crowd over and finally rest wholly upon the adambulacrals, and end against the distal edges of the syngnaths . These are made up of the coossified terminal adambu- lacral ossicles. Genus PROTASTER Forbes. Text fig. 23. Protester FORBES, Mem. Geol. Surv. United Kingdom, dec. 1, 1849, pi. 4, pp. 1-2; in McCor, British Pal. Foss., 1851, p. 60. MURCHISON, Siluria, 1854, p. 221, fig. 39-4. SALTER, Rep. Twenty-sixth Meet., Brit. Assoc. Adv. Sci., Notices and abstracts, 1857, p. 76; Ann. Mag. Nat. Hist., ser. 2, vol. 20, 1857, pp. 325, 330 HALL, Nat. Hist. N. Y., Pal., vol. 3, 1859, p. 134. SALTER, Ann. Mag. Nat. Hist., ser. 3, vol. 8, 1861, p. 484. WRIGHT, Mon. Brit- ish Foss. Echinod., Oolitic, vol. 2, pt. 1 (Palseontogr. Soc. for 1861), 1862, pp. 23, 31. HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 293; rev. ed., 1868=1870, p. 335. QUENSTEDT, Petrefacten- kunde Deutschlands, vol. 4, 1876, p. 133, pi. 95, fig. 11. STURTZ, N. Jahrb. fiir Min., etc., 1886, vol. 2, pp. 148, 150. GREGORY, Geol. Mag., dec. 3, vol. 6, 1889, p. 26. ST^RTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, p. 18. J. F. JAMES, Journ. Cincinnati Soc. Nat. Hist., vol. 18, 1896, p. 137. GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1031, fig. 1. WHIDBORNE, Palseontogr. Soc., 1898, p. 206. PARKS, Trans. Canadian Inst., vol. 8, 1908, pp. 363, 366. SCHONDORF, Jahrb. nas- sauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 238. SOLLAS and SOLLAS, PMlos. Trans. Roy. Soc. London, ser. B, vol. 202, 1912, p. 223. SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palseontogr. Soc. for 1913), 1914, p. 35. Encrinaster (part) HAECKEL, Gen. Morphologie, vol. 2, 1866, p. Ixvii (no definition). Also Bee Sturtzura. FiG.21. VENTRAL VIEW OF A RAY OF BOHEMURA JAHNI, ENLARGED ABOUT X 2, AFTER JAEKEL. a, AMBULACRALIA; ad, AD AMBULACRALIA WITH THEIR LATERAL EXTENSIONS; ddm, SYNGNATHS. REVISION OF PALEOZOIC STELLEBOIDEA. 225 Original description. " (Order Ophiuridse. Family Euryales.) Body circular, covered with squamiform plates; genital openings in the angles of junction of the arms beneath; arms (simple) formed of alternating ossicula." Genoholotype. " Species Unica. Protaster sedgwickii. Forbes." Original description of P. sedgwickii. "The disk is circular, and shaped like that of an OpTiiura. The arms are five in number, very narrow, equidistant, and similar. "The upper and under surfaces of the disk were covered by small, similar, more or less regular, polygonal or crescentic plates, imbri- cated in scale-fashion, and having punctated surfaces. Those of the under side of the body are smaller and more regular than those of the upper. The mouth is cen- tral, and rather small in proportion to the disk. The buccal ap- paratus is composed of ten parts or proc- esses, arranged in pairs; half of each springs from the origin of each arm in a diverging manner, and meets the corre- sponding half to form a lanceolate tooth- 1 i lr <* rkrrn'antirkn FlG> 22. ORAL SKELETON OF PAI^EUBA NEGLECTA, MUCH ENLARGED, u j t? i o II, AFTER JAEKEL> at AMB ULACBALIA; ad, ADAMBULACEALIA. deeply indenting the cavity of the mouth. Of how many separate ossicles each of the buccal processes was composed the specimen affords but very indis- tinct indication. They seem to affect a slightly falcate form at their extremities. The arms were composed of alternating, somewhat quadrate ossicula, the sides of which were deeply indented supe- riorly, in order to form spiniferous crests. The spines were appar- ently short, and not equal in length to the length of an ossicle, obtuse, and few in a row. The under surfaces of the brachial ossicula were not indented laterally. The central portions of the upper surfaces of the brachial ossicles were hollowed out slightly, and the sutures between them deeply impressed. About 12 of the brachial ossicles were imbedded in the disk, and the parts of the dorsal surface of the latter, corresponding to the arms and central skeleton, do not present traces of scales, a feature seen in existing Euryales. 226 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. "The dimensions of Prof. Sedgwick's specimens are as follows: "Breadth of disk, 8/12ths inch. "Average breadth between arms, 5/12ths inch. Breadth of an arm, at its junction with the body, 1/1 2th inch. "Locality and, geological horizon. Silurian, from Ludlow rocks at Docker Park, near Kendal, Westmoreland (specimen figured), associated with Encrinites, and at Benson Knot, also near Kendal, hi hard sandstones, full of characteristic Ludlow fossils." Gregory restudied this genus and in 1897 defined it as follows: "Protasteridse with a well-marked disk; long, tapering, very flexible arms. Some of the adambulacral ossicles are Y-shaped. Scales of the disk fairly large. Type species. Protaster sedgwickii, Forbes." Gregory remarks that "Forbes's figures of the arm-structures are not satisfactory," figure 23 as drawn by the former. This illustration a b shows each set of FIG. 23. AMBULACHALIA AND ADAMBULACRALIA OF PROTASTEB SEDGWICKII, AFTER GREGORY, a, NEAR THE DISTAL END; 6, IN THE MIDDLE; c, AT ambulacralia dif- THE PROXIMAL END. THESE DIFFERENCES ARE DUE TO PRESERVATION. tion thought to be due to preservation and adhering rock rather than to structure (see fig. 28, on p. 242). Madreporite probably abactinal in position. In regard to the term Encrinaster, see Encrinaster (Aspidosoma) , page 242. Protaster has the following species : P. sedgwickii Forbes. Ludlow of England. P. biforis Gregory. Wenlock of Wales. P. groomi Sollas and Sollas. Ordovicic of England. ' P. (?) whiteavesianus Parks. Black River. P. (t) salteri (Salter). Ordovicic of Wales. . P. (?) stellifer Ringueberg. Rochester. PROTASTER BIFORIS Gregory. Text figs. 24, 25. Protaster biforis GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1032, figs. 2, 3 on p. 1033. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 238. Original description. "Disk fairly large; interbrachial outlines concave. The syngnaths are simple, prominent, and stout. The ambulacral ossicles consist of a thick body and a stout curved wing. The distal margin of the ossicles is notched by a depression for a ventral muscle-field, which also cuts into the proximal margin of the adjoining ossicle. Owing to these muscular depressions the arm has apparently two series of pores. REVISION OF PALEOZOIC STELLEROIDEA. 227 "The adambulacral ossicles are massive and taper slightly to their distal ends; they are closely attached and form a regular series of marginal plates. "Arms very flexible. "Dimensions: Length of longest arm 18 mm.+a; Diameter of arm at the base 2.5 mm. Diameter of arm near the distal end 1 mm. Diameter of mouth 1.5 mm. Length of syngnath 1 mm. Width of ambulacral furrow at edge of disk 1.25 mm. "Distribution. Wenlock Shale, Castell Dinas, Bran, near Llan- gollen," Wales. Museum Practical Geology, London. if 24 25 FIGS. 24 AND 25. PROTASTER BIFOEIS. AFTEE GEEGOEY. 24, AMBULACRALIA AND ADAMBULACEAUA. 25, O, OUTLINE OF THE DISK AND AEMS; 6, A PAIE OF SYNGNATHS. PROTASTER GROOMI Sollas and Sollas. Protaster groomi SOLLAS and SOLLAS, Philos. Trans. Roy. Soc. London, ser. B vol. 202, 1912, pp. 223, 224. Occurs in the Middle Ordovicic of Shropshire, England. PROTASTER (?) WHITEAVESIANUS Parks. Protaster whiteavesianus PARKS, Trans. Canadian Inst., vol. 8, 1908, p. 368, figs. 1-6. This is a large form with a, disk 15 mm. across and with rays 60 mm. long. It is described in too much detail to quote here. The ambula- cralia are distinctly alternate and boot-shaped. Formation and locality. In the Lower Trenton (Kirkfield), at Kirkfield, Ontario, Canada. The cotypes (five) are in the University of Toronto (No. 638 T.). PROTASTER (?) SALTERI (Salter). Ophiura salteri SALTER and SOWERBY, Quart. Journ. Geol. Soc. London, vol. 1. 1845, pp. 9, 20 (table) (nomen nudum). Protaster salteri WRIGHT, Mon. British foss. Echinod., Oolitic, vol. 2, pt. 1 (Pala>on- togr. Soc. for 1861), 1862, p. 33 (nomen nudum). SALTER, Mem. Geol. Surv. Great Britain, vol. 3, 1866, p. 289, pi. 23, fig. 3 (is inclined to refer it to Tseniaster}. PARKS, Trans. Canadian Inst., vol. 8, 1908, pp. 364, 367. Formation and locality. The holotype is from the Ordovicic, near Cerrig-y-Druidion on the Holyhead road, Wales. 228 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. PROTASTER (?) STELLIFER Ringueberg. Protaster stellifer RINGUEBERG, Bull. Buffalo Soc. Nat. Sci., vol. 5, 1886, p. 7, pi. 1, fig. 2. Generic relations not established. Original description. "Disk of medium size, flat, circular, slightly flexible, distinct from the rays on the dorsal side, and has on that side a quinque-dentate, stellate, central elevation, which again has a central, stellate depression of about one-half the lateral extension of the stellate elevation upon which it is impressed. The points of the star-like figure are opposite the several rays and extend about two-thirds across the disk. "Surface finely granulose. Rays slender, almost imperceptibly tapering in the upper half of their known length; rounding on the dorsal side, with two rows of regular, quadrilateral, alternating [ambulacral] plates which can with difficulty be made out, and which have a fine granulose [integumentary] surface resembling the surface of the disk; opposite each transverse suture there is a corresponding linear depression across the surface of the opposite plate on the other side of the median suture, which at first sight gives the impression that the ray is composed of opposite plates of only one-half the real length. "Ventral side with ten short, oral plates which are slightly sepa- rated below, and meet at the discal surface, where they are rounded. "Ambulacral series long, regularly quadrilateral; alternately arranged. "Adambulacrals indistinct in the only specimens found showing the ventral surface. "Marginal series slightly imbricating; spiniferous; spines rather short." Formation and locality. Three specimens were collected in the Rochester shale of the Siluric at Lockport, New York. They are in the author's collection. Two other fine specimens were found by C. J. Sarle at Lockport and are now at Yale University. Another good specimen is in the Walker collection of the University of Toronto (No. 1007) ; it was found at Grimsby, Ontario. Genus ALEPIDASTER Meek. Plate 36, fig. 4; text fig. 26. AlepidasterM-EEK, Amer. Journ. Sci., ser. 3, vol. 4, 1872, p. 275; Geol. Surv. Ohio, Pal., vol. 1, 1873, p. 68, at end of specific description. Protasterina ULRICH, Journ. Cincinnati Soc. Nat. Hist., vol. 1, 1878, p. 95. J. F. JAMES, Journ. Cincinnati Soc. Nat. Hist., vol. 18, 1896, p. 139 (genoholotype, P. fimbriata Ulrich). Original description. "The only specimen I have seen that is certainly known to belong to this species [P. ? granuliferus] is very REVISION OF PALEOZOIC STELLEROIDEA. imperfect, being merely an incomplete disk, and the inner ends of the rays. It does not conform to the characters of Protaster given in Prof. Forbes' s diagnosis, in all respects, since its disk, especially on the upper side, is covered by an integument composed of a vast number of very minute grains of calcareous matter, instead of dis- tinct imbricating scales. It is therefore not improbable that perfect specimens would show other characters that would warrant the establishment of a new genus or subgenus for such forms, in which case the name Alepidaster might be applied to the group." Genoholotype. Protaster (?) granuliferus Meek. Original description of Protasterina. "Rays five, slender, flexible, and ex- tending much beyond a cir- cular and minutely granu- lar disk, which is provided with short, slender, and outwardly directed spines; inner ray pieces [ambu- lacra] regularly alternating, of an hour-glass shape, and interlocking along the me- dian line, which is there- fore not straight but zig- zag ; outer ray pieces elon- gated [adambulacrals or side plates], directed ob- liquely outwards, so as to partly overlap each other; two rows of large pores between the inner [ambulacral] and outer [side plates] ray pieces; in the type-species these pores appear to have been occupied by loosely-fitting subpyramidal plates, some of which have a deep depression in the top, as though they were perforated; their 1 true nature, however, is very uncertain. Oral pieces ten, each pair being formed by two of the outer ray pieces. Type, P. fimbriata." When the original material of Alepidaster and Protasterina is re- studied, it may be shown that the latter is not a synonym of the former, but at present the writer does not see any generic differences. Alepidaster is easily distinguished from Protaster, not only by its greater geologic age, but also in that its disk is minutely granular and spiniferous while that of the latter is scale-bearing. Further, when the exact shape of the ambulacrals is determined, it is probable that other differences 'will be found there. Particular attention should be paid to Ulrich's statement that in Protasterina the podial openings " appear to have been occupied by loosely fitting, sub- FIG. 26. DIAGRAMS OF ALEPIDASTER GRANULIFERTJS (MEEK). DRAWN BY E. O. ULRICH FROM A SPECIMEN IN THE UNITED STATES NATIONAL MUSEUM. A, ACTINAL VIEW OF RAY PLATES. B, SAME FROM ABACTINAL SIDE. C, SECTION THROUGH RAY. D, SlDE-PLATE FROM THE INSIDE AND IN SECTION. 230 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. pyramidal plates/' If these extra plates existed, Protasterina will be a good genus, differing not only from Protaster, but also from Alepidaster. Apparently no one has actually seen the abactinal side of the rays. The writer also fails to see it in the few specimens studied. It would, therefore, seem that there are no dorsal plates and that the dorsal covering is integumentary or finely granular. This genus contains the folio whig species : A. (?), new species. Trenton. A. granuliferus (Meek). Bichmondian. A.flexuosus (Miller and Dyer). Eden and Maysville. A. miamiensis (Miller). Bichmondian (Waynesville) . ALEPIDASTER (?), new species. In the Walcott collection at Harvard University there are three specimens (Nos. 28, 29, 30) of a "Protaster" that probably will prove to be new. They are from the Trenton limestone at Trenton Falls, New York. ALEPIDASTER GRANULIFERUS (Meek). Text fig. 26. Protaster ? granuliferus (Alepidaster at end of description) MEEK, Amer. Journ. Sci., ser. 3, vol. 4, 1872, p. 274; Geol. Surv. Ohio, Pal., vol. 1, 1873, p. 68, pi. 3 bis, figs. 8a, 86. J. F. JAMES, Journ. Cincinnati Soc. Nat. Hist., vol. 18, 1896, p. 138. PARKS, Trans. Canadian Inst., vol. 8, 1908, p. 368. Original description. "Disk small, apparently circular; rays rather slender, and of unknown length. Dorsal surface of disk and rays covered by an integument composed of innumerable minute grains of calcareous matter. Ventral side of disk not well exposed in the specimen but apparently provided, in the interradial spaces, with minute spines directed outward. Oral pieces not well exposed in the specimen. Arm pieces [ambulacrals] regularly alternating, but apparently rectangular at their inner ends, and not interlocking along the minute mesial impressed line, longer transversely than in the direction of the length of the rays; each largely excavated at its anterior outer end so as to form a large pore, or porelike depression, and divided transversely by a furrow into two parts, the anterior of which is very short, and the posterior longer, and marked by a minute [muscle] pit at its inner end; about eight or nine of these pieces in each range of each ray included within the margin of the disk. Outer arm [or ray] pieces (adambulacral of some) smaller than those of the the inner ranges, and placed edge upward, with an oblique outward direction so as to imbricate outward or toward the extremities of the rays, each bearing one or more minute articulating spines. " Breadth of disk, about 0.43 inch; breadth of arms at their inner ends, 0.10 inch." REVISION OF PALEOZOIC STELLEROIDEA. 231 Formation and locality. In the lower part of the Richmondian at Moore's Hill, Indiana. It was found by C. B. Dyer, and is now in the Museum of Comparative Zoology, Harvard University (No. 21). ALEPIDASTER FLEXUOSUS (Miller and Dyer). Plate 36, fig. 4. Protaster flexuosus MILLER and DYER, Journ. Cincinnati Soc. Nat. Hist., vol. 1, 1878, p. 31, pi. 2, figs. 1, la. MILLER, N. Amer. Geol. Pal., 1889, p. 276, fig. 409. PARKS, Trans. Canadian Inst., vol. 8, 1908, p. 368. Protasterina ("Protaster" lapsus) fimbriata ULRICH, Journ. Cincinnati Soc. Nat. Hist., vol. 1, 1878, p. 95, pi. 4, figs. 9-9c. Protasterina flexuosa and P. fimbriata J. F. JAMES, Journ. Cincinnati Soc. Nat. Hist., vol. 18, 1896, pp. 139, 140. Original description of P. flexuosus. " The disk is composed of very thin, small plates, the order of arrangement of which is not deter- mined. The specimens examined differ in size, and show the disk varying in diameter from one-quarter to one-half an inch. The rays were very flexuous when living, and are found winding and turning and thrown in different directions in different specimens. "Four series of plates are seen upon the dorsal side of each ray near the disk. The two inner series [the ambulacrals seen from the dorsal side] form an angular ridge, each plate is concave at the uniting surface, the arrangement is alternate, and the appearance, therefore, of the top of the ray, is something like the alternate arrangement of two series of hourglasses. The outer series, or marginal [side] plates, are spinous, the spines directed toward the point of the ray. A weathered ray shows three series of pores one row between the inner series of plates, and one between the marginal plates and inner series, upon each side of the ray [probably the plates are worn through by weathering; there should be but two rows of podial openings, the others being either muscle pits or weathering holes]. The plates are a little the longest in the direction of the length of the ray, and in this direction there are about four to the line. The rays cross the disk on the dorsal side and unite near the center of it, but the speci- men figured, Plate II, figure 1, has this part of the rays removed. Plate II, figure la, illustrates the ventral side of a specimen, which is too indistinct to show the arrangement of the plates. Small pieces, showing the ventral side of the rays, from other specimens, have two rows of spines springing from the marginal plates on each side of the rays." Original description of P. fimbriata. 1 "Disk of medium size, circu- lar. Dorsal side of disk, and rays to margin of disk, covered with a granular integument. Ventral surface of disk provided with a large number of outwardly directed, short and slender spines. Oral i Ulrich still maintains that this species is a good one and distinct from P. flexuosus. 232 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. pieces 10, subrhomboidal in outline, arranged in pairs, each pair being formed of two of the marginal series of ray plates ; on the lower inner edge there are five spines, and extending from each pair, over the mouth, is a bundle of rather long ones. "Rays apparently very flexible, contracted toward the mouth; six series of plates are exposed on the ventral side of each ray. The two middle series [ambulacrals] alternate regularly, and interlock along the mesial line; they are about twice as long in the direction of the ray, than measuring transversely; each piece is contracted in the middle on the inner side to admit the wide ends of the two pieces immediately opposite, and on the outer side to form a pore- like impression, between them and the outer series of plates; four [five or possibly six] of these plates in each range of each ray are included within the disk, and the series terminate abruptly at a distance of one and a half lines from the inner end of the oral plates; from the margin of the disk to the extremity of the ray there are about 16 pieces in each range. The pores mentioned above, are occupied by a series of loosely fitting, obtusely conical or pyramidal plates, some of which distinctly show a deep depression in the top, and may have been perforated. Outer or marginal pieces [side plates] flat, placed on edge, and directed obliquely outward so as to overlap each other; the two free edges, that is, the one toward the point of the rays and the one seen on the ventral surface, are lined with from 10 to 12 short club-shaped spines, varying somewhat in length. The marginal [side plates] and middle [ambulacral] series of plates articulate by means of corresponding prolongations from the sides of the plates. These prolongations arise from near that end of the plate which is directed toward the mouth. " Dorsal side of rays composed of two rows of alternating and interlocking plates, which near the disk are very deeply sculptured and about as wide as long, becoming gradually less excavated, and longer in proportion to the width, toward the point of the rays; on each side they articulate with the upper edge of the oblique mar- ginal series of plates [these are the ambulacral plates seen from the dorsal side, either because there were no dorsal plates or because there was an integument]. " Breadth of disk, 0.60 inch; breadth of arm at margin of disk, 0.16 inch; length of same from oral plates, 0.88 inch. This species is related to Protaster flexuosus" Formation and locality. "This species has been found at differ- ent elevations from near low-water mark in the Ohio River [that is, in basal Eden] to the top of the hills at Cincinnati," in the upper Maysville beds. The type-specimens of P. flexuosus were found by Mr. C. B. Dyer and are now in the collection of Harvard Uni- versity (Nos. 17, 18). Those of P. fimbriata were obtained by REVISIOH OF PALEOZOIC STELLEEOIDEA. 233 Ulrich in the lower Eden at about 100 feet above low-water mark in the Ohio River at Covlngton, Kentucky. This material is now in the Ulrich collection of the United States National Museum. Other poor specimens from a still lower level in the Eden and asso- ciated with Triarihrus are from the first ward, eastern Cincinnati. Two specimens are at Yale University. Cat. No. 60615, U.S.N.M. ALEPIDASTER MIAMIENSIS (Miller). Protaster miamiensis MILLER, Journ. Cincinnati Soc. Nat. Hist., vol. 5, 1882, p. 116, pi. 5, figs. 6-66. J. F. JAMES, Journ. Cincinnati Soc. Nat. Hist., vol. 18, 1896, p. 138. PARKS, Trans. Canadian Inst., vol. 8, 1908, p. 368. Original description. "This species is large, with a proportion- ally small disk. A specimen having a disk four-tenths of an inch in diameter, has rays an inch in length. Five specimens have been examined, all showing the ventral side. The dorsal side is unknown. Only that part of the disk between the rays is visible in any of the specimens, and the plates are so anchylosed together that no special definition of them can be given. The rays are long and coarser and stronger than usual in this genus, though they were quite as pliable and flexuous when living as others. Two series of subquadrangular plates, or ambulacral ossicles, alternating with each other, constitute the bottom of each ambulacral furrow; these are bordered by spinous adambulacral plates, which terminate at the angles of the mouth in only five oral plates." Formation and locality. In the Richmond formation, about Waynesville, Ohio. The specimens are in the Harris collection of the United States National Museum. Cat. No. 40886, U.S.N.M. Genus GREGORIURA Chapman. Gregoriura CHAPMAN, Proc. Royal Soc. Victoria, n. ser., vol. 19, 1907, p. 24, pi. 6, fig. 1; pi. 8, figs. 1, 3. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wies- baden, vol. 63, 1910, p. 238. Original definition. "A Protasterid in which the usual boot- shaped ambulacrals are laterally developed, and modified into a subtriangular form. Ossicles on each side of the ambulacral canal subalternate, excepting at the junction with the mouth frames, where they are parallel. Adambulacral ossicles narrow, slender, extending laterally in a line with the proximal border of the ambu- lacral ossicle. Spine-bearing plates, slender, at right angles to the adambulacrals, carrying (in the genotype) two conspicuous spines. Oral skeleton having jaw plates three-fourths the length of the mouth frames; teeth thick and prominent. No traces of a disk preserved in the specimen on which the genus is founded. Arms very slender and very flexible." 234 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Genoholotype and only species. G. spryi Chapman (same references as above). Found in the Siluric (Melbournian) shale at South Yarra, Vic- toria, Australia. Genus BUNDENBACHIA Sturtz. Text fig. 27. Bundenbachia STURTZ, Palaeontographica, vol. 32, 1886, p. 83; vol. 36, 1890, p. 216. GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1033, fig. 4 on p. 1034. PARKS, Trans. Canadian Inst., vol. 8, 1908, p. 367. SCHON- DORF, Palseontographica, vol. 57, 1910, p. 57. Rays five, short, basally wide, terminally pointed or blunt. Disk large, soft and delicate, with concave interradial margins. Abactinally the disk is covered with symmetrical scales, and medially there is a pentamerous rosette (the oral armature squeezed through), from the points of which the rays originate. Rays covered with a granular integument. Gregory states : "Ambulacral ossicles with a dumbbell-shaped body and thin tapering wing. The body of the ossicle is apparently divided into two pieces by a transverse de- pression. The adambulacral plates are small an( * narrow > an d support a triangular spine- bearing plate. [There is only one plate here, FIG. 27.- VENTRAL ABM STRTTC- L J > TURK OF BUJNDENBACHIA. EN- the adambulacral. I here are no lateral LARGED. AFTER GREGORY. pieces as in Ophiuroidea.] "The syngnaths are curved, narrow bars." Genoholotype and only species. B. beneckei Sturtz. Remarks. " Bundenbachia differs from Protaster by the irregular nature and soft plating of the disk, by the presence of spine-bearing plates attached to the adambulacral ossicles, and by the different form of the ambulacral ossicles." (Gregory.) BUNDENBACHIA BENECKEI Sturtz. Bundenbachia beneckei STURTZ, Palseontographica, vol. 32, 1886, p. 83, pi. 8, figs. 7, 7a; pi. 9, figs. 1, la. GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1034, figs. 4. Formation and locality. Occurs in the Lower Devonic roofing slates of Bundenbach, Germany. Genus PAL^EOPHIOMYXA Sturtz. Palseophiomyxa STURTZ, Verb, naturh. Ver. preuss. Kheinl., etc., vol. 56, 1900, p. 202. Palxophwmyxidx STURTZ, Verh. naturh. Ver. preuss. Kheinl., etc., vol. 56, 1900, p. 202. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 235. Genoholotype and only species. P. grandis (Sturtz). REVISION OF PALEOZOIC STELLEROIDEA. 235 PAL^OPHIOMYXA GRANDIS (Sttirtz). Bundenbachia grandis STURTZ, Palaeontographica, vol. 32, 1886, p. 84, pi. 9, figs. 2, 2a, 3, 3a. Palseophiomyxa grandis STURTZ, Verh. naturh. Ver. preuss. Rhei.nl., etc., vol. 56, 1900, pp. 196, 202. Formation and locality. Lower Devonic roofing slates of Bun- denbach, Germany. Cat. No. 59388, U.S.N.M. Family PAL^OPHIURID^E Gregory. Palseophiuridx GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1034; Treat. Zool., vol. 3, Echinoderma, 1900, p. 274. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 237. "Lysophiuras in which the [alternate] ambulacral ossicles consist of a bar-shaped or subquadrate 'body' without wings. " (Gregory, 1900.) Remarks. This family differs from the Protasteridse in "having the ambulacral ossicles longer than wide, and never divided trans- versely by muscular depressions. The ambulacral ossicles are either bar-shaped or thickened to a subquadrate form. They are never boot-shaped." (Gregory, 1897.) Schondorf states that Gregory established the families Protas- terida6 and Palaeophiuridae on the form of the ambulacralia. He says: "The first family has boot-shaped, and the second family bar-shaped or subquadrate ambulacrals. The form of the latter is nothing more than a poor development of the boot-shaped type of ambulacrals. Accordingly the second family has no value and is to be withdrawn." In addition the forms of Sturtzura have boot- shaped ambulacrals, and these "are so little alternate in arrangement that their alternation appears to me more than questionable." (Schondorf, 1910.) Contains the following genera: Pal&ophiura Stiirtz. Sturtzura Gregory. Eugasterella, new name. Ptilonaster Hall. Genus PAL^EOPHIURA Stlirtz. Palxophiura STURTZ, Palseontographica, vol. 36, 1890, p. 233. GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1034. " Palaeophiuridse with the disk surrounded by rod-shaped marginal ossicles. The ambulacral ossicles are rods lying parallel to the arm. " (Gregory.) Genoholotype and only species. P. simplex Sturtz (citations as above). Lower Devonic, Bundenbach, Germany. 236 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Genus STURTZURA Gregory. Sturtzura GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1034. SCHON- DORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, pp. 215, 237. fRhodostoma SOLLAS and SOLLAS, Philos. Trans. Roy. Soc. London, ser. B, vol. 202, 1912, p. 225 (genoholotype, Protaster leptosoma Salter.) " Palseophiuridse having thick, subquadrate, ambulacral ossicles and narrow adambulacral plates. The disk is fragile, and its plates are small and thin: it has no marginal plates. The mouth-frames are narrow and separate. "Distribution. Silurian, England and Australia." Genoholotype. Protaster brisingoides Gregory. "Remarks. This genus differs from PalseopTiiura, as the ambu- lacral ossicles are thick and subquadrate, instead of being in the form of narrow bars; also by the absence of the strong marginal plates round the disk." (Gregory.) Chapman in 1907 refers the genotype back to Protaster, adding: "In consequence of this determination Sturtzura leptosoma may now be considered as the type of the genus." This is in violation of the rules of nomenclature and the generic name will stand or fall on the basis of Protaster brisingoides. Schondorf states: "The ambulacra of S. brisingoides are not at all bar-shaped or subquadrate, if the figures of Chapman ( 1907, pi. 8, fig. 2) are correct, but are also essentially boot-shaped and alternate in arrangement. S. leptosoma also has alternating ambu- lacra. Their identity with LapwortTiura miltoni ranges them with the Auluroidea, where S. brisingoides undoubtedly also belongs. That this form and with it also the genus Sturtzura is to be retained as valid is very questionable, but the final answer may as well wait until the originals are restudied. In any event it is plain that both forms can not belong to the Lysophiurse." Contains the following species: S. brisingoides (Gregory). Siluric of Australia. S. leptosoma (Salter). Ludlow of England. S. leptosomoides Chapman. Siluric of Australia. STtJRTZURA BRISINGOIDES (Gregory). Tseniaster australis McCoY, see CHAPMAN 1907, p. 23 (nomen nudum). Protaster brisingoides GREGORY, Geol. Mag., dec. 3, vol. 6, 1889, pp. 24-27, fig. 1. CHAPMAN, Proc. Roy. Soc. Victoria, n. ser., vol. 19, 1907, p. 22, pi. 6, fig. 2; pi. g; fig. 2. Sturtzura brisingoides GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1034. Formation and locality. From Siluric sandstones at Moonee Ponds Creek, Flemington, near Melbourne, Australia. The type-material is in the British Museum (Natural History), London. Other speci- mens from the same locality and from the parish of Yering, Upper Yarra district. Victoria, are in the National Museum at Melbourne. REVISION" OF PALEOZOIC STELLEEOIDEA. 237 STtfRTZURA LEPTOSOMA (Salter). Protaster leptosoma SALTER, Ann. Mag. Nat. Hist., ser. 2, vol. 20, 1857, p. 331, pi. 9, fig. 5. WRIGHT, Mon. British Foss. Echinod., Oolitic, vol. 2, pt. 1 (Palaeontogr. Soc. for 1861), 1862, p. 33. QUENSTEDT, Petrefactenktmde Deutschlands, vol. 4, 1876, p. 133, pi. 95, fig. 12. STURTZ, Palseonto- graphica, vol. 32, 1886, p. 79; N. Jahrb. fiirMin., etc., 1886, vol. 2, p. 148. Sturtzura leptosoma GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1035. CHAPMAN, Proc. Roy. Soc. Victoria, n. ser., vol. 19, 1907, p. 26, pi. 8, fig. 5 . SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 215. Rhodostoma leptosoma SOLLAS and SOLLAS, Philos. Trans. Roy. Soc. London, ser. B, vol. 202, 1912, p. 223, text fig. 40, pi. 9, figs. 3, 4. Formation and locality. Common in the Siluric, Ludlow forma- tion, at Leintwardine, Shropshire, England. Remarks. Schondorf says S. leptosoma is but a miniature edition of Lapworthura miltoni, both forms being completely identical. The disk margin is not convex but concave between the radii. STURTZURA LEPTOSOMOIDES Chapman. Sturtzura leptosomoides CHAPMAN, Proc. Roy. Soc. Victoria, n. ser., vol. 19, 1907, p. 26, pi. 7, figs. 1, 2; pi. 8, fig. 4. Formation and locality. In the Siluric (Melbournian), Moonee Ponds Creek, Flemington, near Melbourne, Australia. Genus ETJGASTERELLA, new name. Eugaster HALL (not Eugaster Seville 1839, Orthoptera), Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 290; rev. ed., 1868=1870, p. 332. ZITTEL, Handb. Pal., vol. 1, 1879, p. 444. STURTZ, N. Jahrb. fur. Min., etc., 1886, vol. 2, p. 151. MILLER, N. Amer. Geol. Pal., 1889, p. 244. STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, p. 20. GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1035. Eugasterella SCHUCHERT, Fossilium Catalogus, Animalia, pt. 3, April, 1914, p. 19. Original description. "Body stellate, consisting of a central alated disk, and five long, slender, somewhat flexuous rays. Disk composed of small, polygonal tuberculose or subspinose plates on the ventral side. Rays consisting of a double series of alternating subquadrate ambulacral ossicles, and a series of curved adambulacral plates bordering the grooves, and forming the margin of the ray; the outer ends overlap the edge of the next plate in advance. Oral plates ten, arranged in pairs, forming the terminal plates of the adambulacral ranges. Pores large, arranged in two rows in each ray; penetrating the ray at the junction of the ambulacral and adambulacral plates hi such a manner that four different plates border each perforation. Adambulacral and disk-plates bearing spines. Dorsal surface unknown." Genoholotype. E. logani (Hall). 238 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Gregory comments on this genus as follows : "Palseophiuridae in which the ambulacral ossicles are subhep- tagonal in form, the central suture along the arm being zigzag, while the outer angles of the ossicles are cut away for the reception of the podia. The adambulacral ossicles have a flat base, and thence bend forward crescentically. The mouth frames are massive, and those of each pair meet along the middle line of the oral angle. "This genus I only know from Prof. James Hall's figures, and, in spite of their clearness, I feel much doubt as to the wisdom of diag- nosing it from these alone. Its affinities are clearly with the Palse- ophiuridae, but it approaches the Protasteridse in one respect, for, owing to the deep depressions in the outer angles of the ambulacral ossicles, the outer side forms a short rudimentary wing. The genus differs from all the Protasteridae by the absence of a muscular groove across the ambulacral ossicles. " Among the Palseophiuridse it differs from Palseophiura by the absence of marginal ossicles, and from Sturtzura and T&niura by the massive nature of the mouth frames." Contains : E. logani (Hall). Hamilton. E. (?) concinna (Ringueberg) . Rochester. EUGASTERELLA LOGANI (Hall). Eugaster logani HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, p. 290; rev. ed., 1868=1870, p. 333, pi. 9, figs. 7, 8. Original description. "Body stellate, with a small disk and long attenuate fiexuose rays. Disk composed of numerous small poly- gonal plates with radiated surfaces; the diameter, measured from the sinus to its extension on the opposite ray, is about half as great as the length of the ray measured from the center of the disk. Rays, as seen from the ventral side, narrow, attenuate, with the ambulacral plates curving, and near the base of the rays a little wider than long, and toward the extremities longer than wide, with a strong elevated transverse ridge. There are about eight pairs of plates inclosed within the limits of the disk. Pores penetrating the interstices near the outer extremities of the plates, while near the inner end there is a depression or pit resembling a partially excavated pore. The adambulacral plates as seen from below are extremely narrow and very convex on their outer surfaces, forming the margins of the ray. Oral plates in pairs, narrow and elongate. "This specimen measures, from the center of the disk to the extrem- ities of the rays, about one inch and a half; the widest part of the ray, which is near the margin of the disk, is about an eighth of an inch in diameter. The rays toward their extremities have the ambula- cral field covered by the curving of the marginal plates over the REVISION OF PALEOZOIC STELLEROIDEA. 239 groove, and which, closely uniting by their margins, give a tereti- form termination of nearly one-third the entire length, each one having the aspect of a slender crinoidean proboscis. " Geological formation and locality. In the Hamilton group, near Fenner, in Madison County, New York. The specimen was collected during the geological survey, and is now in the State Museum." EUGASTERELLA (?) CONCINNA (Ringueberg). Eugaster condnnus RINGUEBERG, Bull. Buffalo Soc. Nat. Sci., vol. 5, 1886, p. 8, pi. 1, fig. 3. This species does not belong to EugastereUa, but until the holo- type is restudied it can remain in this genus. It is from the Kochester shale at Lockport, New York, and is in Dr. Bingueberg's collection. Original description. "Disk flat, thin, alated, composed of very fine granulose plates. Rays broad at their base; flattened, rapidly tapering for two-thirds of their length; terminal third attenuate, rounded; plates, if any, undefinable. " Dorsal side with four series of plates; two median and two lateral; the transverse sutures dividing the outer series are continuous with every other transverse suture dividing the central series. Medial series with twice the number of plates of the outer, and are raised above them. " Sutures depressed. Surface of the plates rounded and finely granular. "The lateral series of plates decrease regularly in size after reach- ing the disk until they end in a pronounced elevation formed by the last pair of the medial series at a point about halfway from the border toward the center of the disk measuring at the narrowest part where both series end. Between these elevations the disk has a shallow, stellate depression with the points opposite the several rays. The two outer rows of plates apparently disappear toward the attenuate tip before the median does so. The imperfect cast of the upper part of one of the rays is all that is known of the ventral side; this shows traces of an alternating series of ambulacral plates. "Marginal plates spiniferous. "Length of ray, one-half inch." Genus PTILONASTER Hall. Ptilonaster HALL, Twentieth Rep. N. Y. State Cab. Nat. Hist., 1868, pp. 291-292; rev. ed., 1868=1870, p. 334. LUTKEN, Dansk. Vid. Selsk. Skrift., ser. 5, vol. 8, pt. 3, 1869, p. 82. STURTZ, Palseontographica, vol. 32, 1886, p. 83; N. Jahrb. fur Min., etc., 1886, vol. 2, p. 152. GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1036. STURTZ, Verb, naturb. Ver. preuss. Rbeinl., etc., vol. 56, 1900, p. 188. Palseocoma MILLER (part), N. Amer. Geol. Pal., 1889, p. 266. Original description. "Form and general features as in Eugaster, but differing in the plates of the rays. Rays composed of an ambu- Bull. 88 15 16 240 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. lacral, adambulacral and marginal series, which are united by their edges, and apparently not imbricating, the projecting or oblique anterior face of the marginal plates bearing spines which are inclined toward the extremity of the ray. Margins of the rays alated by the extension of the disk. "I had originally united this form with Eugaster, but further ex- amination has shown the ray to have a range of marginal plates outside of the adambulacral plates, and I am unable to discover any such feature in Eugaster logani. "This and the preceding genus [EugastereTla] belong to the Ophiu- ridae, to which may also be referred the genus Protaster." " Note. It is only as these pages are going to press that my atten- tion has been directed to the similarity of structure in the ray of Palseocoma of Salter with that of the proposed genus Ptilonaster. In this illustrated species of the former genus (Palseocoma marstoni) the disk is proportionately larger, and the rays much shorter, while the outer range of plates is represented as imbricating; but being characterized by a double row of plates bordering the ambulacral area, it must be regarded as very nearly related if not generically identical with Ptilonaster. 11 Gregory comments on Ptilonaster as follows: "This genus is an ally of Eugaster; * * * it is, however, generically distinct. I only know it from Hall's figures, and therefore prefer to leave the prep- aration of a formal diagnosis to an American palaeontologist." GenoJwlotype and only species. P. princeps Hall. PTILONASTER PRINCEPS Hall. Ptilonaster princeps HAT.?., Twentieth Hep. N. Y. State Cab. Nat. Hist., 1868, p. 292; rev. ed., 1868=1870, p. 334, pi. 9, fig. 9. Palseocoma princeps MILLER, N. Amer. Geol. Pal., 1889, p. 267. Original description. "The specimen consists of the impression of the greater part of one ray, with parts of two others, and intervening portions of the disk. The disk has extended along the ray nearly an inch from the center of the body; the plates are small, and have been furnished with slender spinules. The ray is strong and extremely elongate, having been at least 4 inches in length; its greatest width is outside of the disk, where it measures seven-sixteenths of an inch. The ventral side of the ray shows three series of plates ambulacral, adambulacral, and marginal on each side of the center; of these at least seven ranges have been included within the disk. The ambula- cral plates are a little wider than long, arranged in alternating order. The pores penetrate the interstices near the outer extremity of the plates, while the partial or obsolete pores are obscure, becoming deeper and more conspicuous toward the extremity of the ray. Near the base of the ray there are 12 plates in the length of an inch, REVISION OF PALEOZOIC STELLEROIDEA. 241 and 14 plates in the same distance in the central portion, while toward the extremities there are 20 or more in the same space. The marginal plates are subhexagonal, a little longer than wide, and ornamented by long slender spines on their outer margins. "This is a much larger and more robust species than the Eugaster logani, and in its entire condition it is the largest star-fish known in our palaeozoic rocks. " Geological formation and locality. In the Chemung group. The specimen is in a brownish-gray sandstone, which weathers to a very light ashen color. It was received from Henry S. Randall, Esq., of Cortlandville [New York], in the vicinity of which place it was ob- tained." The holotype, an imprint, is in the American Museum of Natural History, No. 4472. Family ENCRINASTERID^E, new name. Aspidosomatidse GREGORY, Geol. Mag., dec. 4, vol. 6, 1899, p. 351; Treat. Zool., vol. 3, Echinoderma, 1900, p. 250. SCHONDORF, Palseontographica, vol. 57, 1910, pp. 55-61. Palxgoniasteridce STURTZ (part), Palseontographica, vol. 36, 1890, p. 246 (contains Aspidosomella and Palxaster). Lysophiurae with boot-shaped ambulacralia ventrally, and large disks that are bordered with simple marginalia. Apparently only the inframarginals are present. As Aspidosoma Goldfuss is preoccupied, this name can no longer be used for the family name. Encrinaster Haeckel replaces it and therefore becomes the base of the family designation. The family has the single genus Encrinaster. Genus ENCRINASTER Haeckel. Plate 35, text fig. 28, 29. Aspidosoma GOLDFUSS (not Aspidosoma Fitzinger 1843, Keptilia), Verh. naturh. Ver. preuss. Rheinl., etc., vol. 5, 1848, pp. 145-146. SIMONOVITSCH, Sitzb. d. mat.-naturw. Classe Akad. Wiss., Wien., vol. 64, Abt. 1, 1871, p. 103. ZITTEL, Handb. Pal., vol. 1, 1879, p. 451, fig. 322. STURTZ, Palseonto- graphica, vol. 32, 1886, p. 92; N. Jahrb. fiir Min., etc., 1886, vol. 2, p. 151; Verh. naturh. Ver. preuss. Rheinl., etc., vol. 50, 1893, pp. 38, 55. SCHON- DORP, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 35; Palseontographica, vol. 57, 1910, pp. 1-63. SPENCER, Mon. Brit. Pal. Aste- rozoa, pt. 1 (Palseontogr. Soc. for 1913), 1914, pp. 23, 24, 49. Encrinaster HAECKEL, Gen. Morphologie, vol. 2, 1866, p. Ixvii (genolectotype Aspidosoma arnoldi Goldfuss). Diagnosis. Animals attaining large size, with a comparatively large disk. Rays five, more or less petaloid, slender and long, with wide and open ambulacral furrows. 242 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. "Ambulacrals alternating in adjoining columns, while the neigh- boring ambulacrals and adambulacrals are directly opposite each other. The margin of the disk is concave and limited by a row of more or less conspicuous simple [not two rows of superposed] marginalia restricted to the disk. The sides of the free rays are bounded by adambulacrals. These and the ambulacrals are [common to the two sides of the animals, and abactinally are] roughly sculp- tured, usually naked, and only in rare cases were adambulacral spines observed. Madreporite actinal, situated near the peri- stome in one of the interradial ~ I " c ' areas nearest to the left adam- FIG. 28. AMBULACKALIA AND ADAMBULACRALIA OF bulacrals of a radius. 7 ' (Sch6n- ENCRINASTER ROEMERI, AFTER SCHONDORF. a, AP- A^^t lain "\ - ' ' PEARANCE OF THE PLATES WHEN CRUSHED OR OB- SCTTRED BY CLAY. EACH AMBULACRAL HERE APPEARS AS IF OF TWO PIECES; b, SAME BETTER PRESERVED -,-j- (^ O l r lf 11 ^ G. prm q OR MORE CLEANED; c, THE ACTUAL FORM OF THE ossi- noLai ^ ss - ^ e CLES WHEN WELL PRESERVED AND COMPLETELY to the Lower Devonic of Rhenish Germany. Remarks. This well-known genus has been referred to the asterids and to the ophiurids, and though it has most often been placed among the true starfishes, it now appears to belong to neither of these orders. Schondorf has restudied the material with great care and has made out the detail of the ambulacralia, with the result that he erects for Aspidosoma the order Auluroidea. In the work of 1910 cited above the matter is set forth in great detail, and all that can be done here is to direct atten- tion to the more important of his illustrations, with their descriptions, reproduced in part here and on pages 213 and 214. As Aspidosoma was first used for a genus of reptiles, it can not, under the rules of nomen- clature, be used again for another group of animals. It is possible, however, under the FIG. 29. ORAL SKELETON OF rules to make use of Enerinaster, but in doing ^7*^"! SO the proceeding should first be explained. MUCH ENLARGED. AFTER Encrinaster was proposed by Haeckel as fol- SCHONDOEF - lows : In defining his sixth order of asterids, the Crinastra or lily stars, forms that he said already had crinoid characters, he cited as exam- ples of it Protaster sedgwickii and P. = Aspidosoma arnoldi. These forms, he added, are so different from Protaster miltoni ( i that we here separate the former as Encrinaster (E. sedgwickii, E. arnoldi, etc.)." Under ordinary circumstances it would be best to disregard the name because of the inadequate definition, but as Aspidosoma is preoccupied, we are obliged to propose another name to take its EEVISION OF PALEOZOIC STELLEKOIDEA. 243 place. The first mentioned genotype of Encrinaster is Protaster sedgwickii, but as this is also the type-species of Protaster, a genus in good standing, we may accept the second named species, Aspido- soma arnoldi. Therefore it is here proposed to replace Aspidosoma Goldfuss (not Fitzinger) by Encrinaster Haeckel, with Aspidosoma arnoldi Goldfuss as the genolectotype. ENCRINASTER ARNOLDI (Goldfuss). Aspidosoma arnoldi GOLDFUSS, Verb, naturh. Ver. preuss. Rheinl., etc., vol. 5, 1848, pp. 145-146, pi. 5. MULLER, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 12, 1855, p. 4, pi. 1, figs. 1-3. STURTZ, Palaeontographica, vol. 32, 1886, p. 93. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. 38; Palseontographica, vol. 57, 1910, pp. 4, 62, pi. 1, fig. 2; pi. 2, figs. 6, 7; pi. 3, figs. 7-9; Jahrb. preuss. geol. Landesanst., Berlin, vol. 29, 1908, p. 706. SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palseontogr. Soc. for 1913), 1914, p. 40. Encrinaster arnoldi HAECKEL, Gen. Morphologie, vol. 2, 1866, p. Ixvii. Formation and locality. This species is common in the grauwacke of the Lower Coblenzian at Winningen, on the Mosel, Germany. ENCRINASTER GOLDFUSSI (Schbndori). Text fig. 29. Aspidosoma goldfussi SCHONDORF, Palseontographica, vol. 57, 1910, pp. 11, 62, pi. 3, figs. 1, 6. Formation and locality. A common form found associated with E. arnoldi about Winningen, Germany. ENCRINASTER PETALOIDES (Simonovitsch). Plate 35, figs. 5, 6. Aspidosoma petaloides SIMONOVITSCH, Sitzb. mat.-naturw. Classe Akad. Wiss., Wien., vol. 64, Abt. 1, 1871, p. 106, pi. 4, figs. 1-3. ZITTEL, Handb. Pal., vol. 1, 1879, p. 453, text fig. 322. STURTZ, Palseontographica, vol. 32, 1886, p. 92. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 62, 1909, p. -37, pi. 4, figs. 4, 5; Palaeontographica, vol. 57, 1910, pp. 14, 63, pi. 1, figs. 3, 4; pi. 2, fig. 4; pi. 3, figs. 10-16 (complete synonymy given here). This species is described in great detail by its author and by Schondorf, and is well illustrated. Here as elsewhere the abactinal skeleton is that of the actinal area seen from the dorsal side. Four specimens of this species were found in grauwacke of Upper Coblenzian age, in the old quarry back of the Hohenrheiner furnace at Niederlahnstein, Rhenish Germany. ENCRINASTER PETALOIDES GOSLARIENSIS (Halfar). Aspidosoma petaloides, var. goslariensis HALFAR, Jahrb. preuss. geol. Lande- sanst., Berlin, vol. 13, 1893, p. 186, pi. 10, figs. 1, la. SCHONDORF, Palaeontographica, vol. 57, 1910, pp. 21, 63, pi. 2, fig. 5; Jahrb. preuss. geol. Landesanst., Berlin, vol. 29, 1908, p. 707. Formation and locality. The holotype is from the Lower Devonic of the Harz. Germany. 244 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. ENCRINASTER TISCHBEINIANUS (Roemer) emend. Sttirtz. Aspidosoma tischbeinianum ROEMER, Palaeontographica, vol. 9, 1863, pp. 144-146, pi. 23. QUENSTEDT, Petrefactenkunde Deutschlands, vol. 4, 1876, p. 76, pi. 92, fig. 39. STURTZ, Palaeontographica, vol. 32, 1886, p. 92, pi. 12, figs. 1, 2. SCHONDORF, Palaeontographica, vol. 57, 1910, pp. 23, 62, pi. 1, fig. 1; pi. 3, figs. 4, 5 (complete synonymy given here). Formation and locality. Common in the roofing slates of the Rhenish Lower Devonic at Bundenbach, near Birkenfeld, Germany. As many as 60 individuals have been found on a slab. Cat. No. 35117, U.S.N.M. ENCRINASTER ROEMERI (SchSndorf). Text fig. 28. Aspidosoma tischbeinianum ROEMER (part), Palaeontographica, vol. 9, 1863, pi. 25, fig. 11. Aspidosoma roemeri SCHONDORF, Palaeontographica, vol. 57, 1910, pp. 28, 62, pi. 2, fig. 3; pi. 3, figs. 2, 3. Formation and locality. Found associated with E. tiscTibeinianus at Bundenbach, Germany. ENCRINASTER SCHMIDTI (SchSndorf). Aspidosoma schmidti SCHONDORF, Palaeontographica, vol. 57, 1910, pp. 31, 63, pi. 1, fig. 6; pi. 3, figs. 17, 18 (complete synonymy given here). Formation and locality. A slab with six specimens from the Lower Devonic (Herdorfer zone of the Siegen) at Oberdielfen ; Germany. ENCRINASTER EIFELENSIS (SchOndorf). Aspidosoma eifelense SCHONDORF, Palaeontographica, vol. 57, 1910, pp. 34, 63, pi. 1, fig. 5. Formation and locality. Known by several examples from the Lower Coblenzian at Oberstadtfeld, Eifel, Germany. ENCRINASTER PONTIS (Clarke). Aspidosoma ? pontis CLARKE, Monog., I, Serv. Geol. Min. Brasil, 1913, p. 312, pi. 27, figs. 1-8; Bull. N. Y. State Mus., No. 164, 1913, pi. 35, figs. 1-4. Original description. "This very graceful starfish occurs in abund- ance in the soft gray shales of Ponta Grossa, sometimes in clustered groups with its long and slender arms curled and intertwined, more often as single individuals lying in beds with shells of Orbiculoidea and various lamellibranchs. I have counted 20 individuals among the specimens at my command and, singularly, every one of these is preserved with its ambulacra! face up. The removal of all calcareous matter has left only internal casts of this surface from which it has been possible to gain an idea of external structures from carefully REVISION OF PALEOZOIC STELLEEOI0EA. 245 made impressions. Na specimen gives any clue to the dorsal exterior save one arm which has been turned sidewise showing a part of that surface. "The striking character of this starfish is the unusual delicacy of its arms, which are very slender and long for a true asteroid and present an appearance very suggestive of an ophiuran. These slender arms are usually so exposed as to show only the alternating am- bulacral plates which increases the effect of slenderness; but the adambulacrals are occasionally present and give the arms increased width. These arms are all clearly sulcate, the median groove being sharply marked in all casts by a thin elevated ridge running the entire length of the arm. The soft shale holding these specimens is not a very satisfactory matrix from which to take the special struc- tures of the animal but in good part these can be made out. The ambulacral plates are quadrate in outline, alternate in position along the arm groove, and together are arched so that normally the am- bulacral surface is an arched ridge rising above the rest of the surface. These plates, in their arched position, are slightly spaced or dis- placed along adjoining edges leaving passages from the ambulacra. They are usually broken down or slipped over each other in com- pression so that the arm seldom* shows them in the true position. The surface of these plates is finely granular. The adambulacrals constitute one (or more ?) lateral rows and are so seldom well defined that I am not able to represent them with precision but each of them seems to carry one short spine and all to be granulated on the surface. The body of the animal is deeply incut by the arms, which radiate from a narrow center, sharply quinquepartite by the buccal cavity. The oral apparatus is pretty well made out. The oral frames, one for each arm, are very conspicuous, though varying in size in different individuals. Each is broadly saddle-shaped, split medially, the distal parts joined by a suture, the proximal ends spreading apart at a sharp angle which opens into the oral cavity. Their surface is rounded and granulate, the distal ends elevated, and often each half is inclined downward toward the other along the median suture. At each proximal angle lies a sharp jaw or tooth. The whole of the body surface is occupied by the oral rosette composed of the five Jbuccal divisions." Formation and locality. From the Devonic of Ponta Grossa and Jaguariahyva, Brazil. ENCRINASTER (?) GRAY^E (Spencer). Aspidosoma grayse. SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palseontogr. Soc. for 1913), 1914, pp. 23, 24, 33, 38, text figs. 22, 24, pi. 1, fig. 8. Occurs in the Caradoc of Great Britain. 246 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. Order STREPTOPHIUR^ Bell. Streptophiurse BELL, Proc. Zool. Soc. London for 1891, 1892, p. 179. GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1036; Treat. Zool., vol. 3, Echino- derma, 1900, p. 274. Protophiurex STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 56, 1900, p. 200. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, pp. 234, 239, 251. Ophiurasterise SCHONDORF, Palseontographica, vol. 57, 1910, p. 61. Diagnosis by Gregory (1897). Auluroidea "in which the ambu- lacral ossicles are opposite and are generally fus.ed into vertebral ossicles. In such cases the vertebral ossicles articulate by a more or less simple ball-and-socket joint." "Remarks. The main character of this order is that the ambu- lacral ossicles are paired, but primitive. The order differs from the preceding [Lysophiurse] by having the vertebral ossicles always opposite instead of alternate. In some of the simplest members of the order, as Ophiurina, the ossicles are not fused, there are no ventral arm shields, and an ambulacral furrow runs along the ven- tral side of the arm. In the next higher family, as in the genus Lapworihura, the ambulacral ossicles are fused, but have plain articulating surfaces, and there is* an ambulacral furrow." Contains the f amilies : Ophiurinidae. Lapworthuridas. Eoluidiidae. Aganasteridae. Cholasteridae. Family OPHIUKINIIXE Gregory. Ophiurinidse GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1037; Treat. Zool., vol. 3, Echinoderma, 1900, p. 275. STURTZ, Verh. naturh. Ver. preuss. Bheinl., etc., vol. 56, 1900, p. 201. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 234. Original diagnosis. " Strep tophiurae with ambulacral ossicles, only slightly united, and without ventral arm plates." Contains the genera: Ophiurina Sttirtz. Tremataster Worthen and Miller. Genus OPHIURINA Sturtz. Ophiurina STURTZ, Palaeontographica, vol. 36, 1890, p. 232. GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1037. Gregory's diagnosis reads as follows: "Disk circular, with mar- ginal plates. Ambulacral ossicles long, narrow bars. Syngnaths rod-shaped. Adambulacral plates absent or altogether lost from the fossil." KEVISION OF PALEOZOIC STELLEEOIDEA. 247 Genoliolotype and only species. 0. lymani Stiirtz (citations as above). Lower Devonic, Bundenbach, Germany. Genus TREMATASTER Worthen and Miller. Tremataster WORTHEN and MILLER, Geol. Surv. Illinois, vol. 7, 1883, p. 330. MILLER, N. Amer. Geol. Pal., 1889, p. 286. GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1037. Original description. "Body stellate, central part discoid, rays long, flexuous, and gradually tapering. Rays consisting of a double series of ambulacral plates, forming, together, a subcuneiform se- ries, with the tapering ends directed toward the apices of the rays and uniting with the larger ends of the succeeding double plates, upon each side of which there is a series of curved adambulacral plates, which form the margins of the rays. The pores are large and situated between the contracting sides of the ambulacral plates and the concave sides of the curving adambulacrals. Four plates border upon each pore though nearly surrounded by two of them. Plates 10. Dorsal surface unknown. "This genus is distinguished by the parallel arrangement of the ambulacrals and adambulacrals, instead of the alternate order, and by the large pores." Gregory's diagnosis is as follows: "Disk with concave inter- brachial outlines. No marginal plates. Ambulacral ossicles short. Adambulacrals present." Genoholotype. T. difficilis Worthen and Miller. T. (?) decheni (Dewalque) may also belong here. TREMATASTER DIFFICILIS Worthen and Miller. Tremataster diffidlis WORTHEN and MILLER, Geol. Surv. Illinois, vol. 7, 1883, p. 330, pi. 31, figs. 3fl, 36. Original description. "Body stellate, disk small, rays long, flexuous, and gradually tapering. Ambulacral plates subtrigonal, elongated, and united upon their straight faces in parallel order, so as to form a series of subcuneiform sections at the bottom of the ambulacral groove, with the tapering end of each directed toward the apex of the ray, and uniting with the larger end of the next succeeding double plate. "Adambulacral plates parallel with the ambulacrals, and each forms a curve from the larger end of an ambulacral toward the apex of the ray, to unite with the next succeeding adambulacral as it curves away from its attachment to the ambulacral. By this order of construction the adambulacrals are fixed, at one end, to the ambulacrals, while the other end moves upon the curved surface of an adambulacral so as to allow the same lateral flexibility of the rays, secured in other genera by the alternate arrangement of 248 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. the plates. The pores are large and situated between the contract- ing sides of the ambulacral plates, the concave sides of the adam- bulacrals, and the attaching ends of the succeeding plates, so that they are bounded by four plates while nearly surrounded by two. "The oral plates form, in pairs, elongated triangles, with the acute angle directed toward the center of the visceral cavity. The apices are joined by a suture with the basal portions, which are partly separated by a dividing furrow. A single genital pore per- forates each basal section of the oral plates. Dorsal side unknown. "Locality and position: Prairie du Long Creek, Monroe County [Illinois], Chester limestone. "No. 2479, Illinois State collection of 1878." .' TREMATASTER (?) DECHENI (Dewalque). Protaster decheni DEWALQUE, Ann. Soc. ge"ol. Belgique, vol. 8, 1880, pp. 52-54, pi. 3, figs. 1, 2. Tremataster (?) decheni GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1037, footnote. Formation and locality. Upper Devonic at Walcourt (Namur) and Tohogne (Luxemburg), Belgium. Family LAPWORTHUKID^E Gregory. Lapworthuridse GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1037. STURTZ, Verb, naturh. Ver. preuss. RheinL, etc., vol. 56, 1900, p. 203. GREGORY, Treat. Zool., vol. 3, Echinoderma, 1900, p. 275. SCHONDORF, Palseontographica, vol. 57, 1910, p. 58; Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, pp. 235, 239. Furcasteridx STURTZ, Verb, naturh. Ver. preuss. RheinL, etc., vol. 56, 1900, p. 203. SCHONDORF, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 235. Palxospondylidsd STURTZ, Verh. naturh. Ver. preuss. Rheinl., etc., vol. 56, 1900, p. 201. SCHONDORP, Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 235. Original diagnosis. "Streptophiuroida without ventral armplates or buccal shields; ambulacral ossicles fused, but their articulating surfaces are plain. Madreporite dorsal." Schondorf states that this plate is on the ventral side, and that the ambulacrals are not fused (p. 239). Contains the genera: Squamaster Ringueberg. Lapworihura Gregory. Sturtzaster Etherrdge. Bdellacoma Salter. RTiopalocoma Salter. Hallaster Stiirtz. Sympterura Bather. Fur caster Stiirtz. Palastropecten Sttirtz. REVISION OF PALEOZOIC STELLEROIDEA. 249 Sturtz proposes that the family Lapworthuridse be divided into three families. In the original family will remain Squamaster, Lap- worthura, HaTLaster, and Sympterura. The family Furcasteridse has Furcaster, while the Palaeospondylidae embraces Palseospondylus, the latter name being later changed to Palastropecten by Sturtz, as it was preoccupied; the family name will, therefore, have to be changed to Palastropectinidse, as was done by Bather in 1905 (p. 168). % Genus SQUAMASTER Ringueberg. Squamaster RINGUEBERG, Bull. Buffalo Soc. Nat. Sci., vol. 5, 1886, p. 5. Original description. "Body small. Rays long, slender, tapering, flexible, and divided into annular segments. "Ventral side of ray with two series of paired, opposite ambulacral plates, divided by a straight median line; they are irregularly sub- quadrilateral, with one short side, and are arranged in pairs in each series, so that two long and two short sides are joined alternately; consequently there is a long transverse suture regularly alternating with a short one. This disposition gives dentate lateral margins to the ambulacral series. The lateral ends of the long transverse sutures curve away from each other, leaving a small notch between the inferior plates of each pair and the superior plates of the adjoining pairs. This notch receives the inner lateral corners of the squamose dorsal plates. Between the dentate elevations of the ambulacral plates and the lateral margins of the dorsal plates there are large triangular openings, which constituted, or in which were placed, the pores. "Dorsal side composed of large, quadrangular, squamose plates, that encircle the arm as far as the ambulacral groove of which they form the margin. They are divided in the upper portion of the ray by a median suture. "Dorsal, ray plates fimbriated on the free margins. "All plates are perforated by minute, closely arranged pores." Genoholotype and only species. 8. echinatus Ringueberg. SQUAMASTER ECHINATUS Ringueberg. Squamaster echinatus RINGUEBERG, Bull. Buffalo Soc. Nat. Sci., vol. 5, 1886, p. 6, pi. 1, fig. 1. Original description. ' ' Body small. Rays long, slender, not widen- ing at the base ; regularly tapering to a subobtuse point. Dorsal side covered by thin imbricating plates, with free external margins which are edged on the outer and lateral sides with closely arranged, long, slender, filiform spinules, that articulate into minute rounded sockets, which are scalloped out of the margin of the plate; they are slightly longer than the exposed portion of the plate, and are placed about their own diameter apart. 250 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. ''Ray divided into about 40 segments or articulations by the encircling, dorsal plates; these grow rapidly narrower near the body, and are divided by a dorsal suture running along a raised median ridge which extends about two-fifths of the length of the ray from the body out, where it gradually disappears with the last divided plate; the rest are entire. These plates are quadrangular, and have then- inner corners articulated into the socket formed in the dentate lateral -extensions of the ambulacral series. They are ornamented on each side by two or three closely arranged, subparallel, very fine ridges, that commence near the dorsal line at the free margins and extend diagonally backward across the plate toward the inner lateral articulated corner. Ambulacral groove furnished with opposite, irregularly quadrilateral, ambulacral plates, divided by a median suture which is crossed alternately by one long and one short suture. "The superior plate of each pair has the greatest extension later- ally, while in the inferior it is the direction of the median line, it being about twice as long as the superior, which, however, is but slightly broader. "Between the lateral margins of the dorsal plates and the dentate elevations of the ambulacral plates are large sub trigonal open spaces. "The plates are punctured by very closely arranged minute perforations." Formation and locality. In the Rochester shale, at Lockport, New York. The holotype is in the collection of Dr. Ringueberg. At least one (No. 1110) and probably two other specimens (No. 465) are in the Walker collection of the University of Toronto. They are from the Rochester shale, at Grimsby, Ontario. Genus LAPWORTHURA Gregory. Text fig. 30. Lapworthura GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1037, fig. 5. SCHONDORF, Palseontographica, vol. 57, 1910, p. 58; Jahrb. nassauisch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 208. SOLLAS and SOLLAS, Philos. Trans. Roy. Soc. London, ser. B, vol. 202, 1912, p. 213. SPENCER, Mon. Brit. Pal. Asterozoa, pt> l (Pa 186011 * ^- Soc. for 1913), 1914, pp. 25, 32, 40, 49. Original diagnosis. "Disk circular, well marked. Arms very flexible, broad; at first FIG. SO.-VENTRAL STRUCTURE uniform in width and then tapering slowly. OP ARM OF LAPWORTHURA, Ambulacral ossicles with the distal and prox- AFTER GREGORY. . . . imal margins parallel; with lateral wings curv- ing round the podial pores. Madreporite large," dorsal. Genoholotype. Protaster miltoni Salter. Schondorf remarks as follows (1910&:58): The ossicles of Lapworthura have been described several times, but never correctly. The ambulacrals are undoubtedly opposite, REVISION OF PALEOZOIC STELLEROIDEA. 251 and both dorsally and ventrally have the structure of Encrinaster. The adambulacrals are also opposite to the ambulacrals, and have long forward directed spines. In regard to the dorsal position of the madreporite, he states that the abactinal area has been mistaken for the actinal, and that therefore the madreporite lies upon the ventral side. The disk is not circular in outline, but concave between the inter- radii. LAPWORTHURA MILTONI (Salter). Protaster miltoni SALTER, Ann. Mag. Nat. Hist., ser. 2, vol. 20, 1857, p. 330, pi. 9, fig. 4. WRIGHT, Mon. British Foss. Echinod., Oolitic, vol. 2, pt. 1 (Palse- ontogr. Soc. for 1861), 1862, p. 32, fig. 18. QUENSTEDT, Petrefactenkunde Deutschlands, vol. 4, 1876, p. 75, pi. 92, fig. 36. STURTZ, N. Jahrb. fur Min., etc., 1886, vol. 2, p. 148; Palseontographica, vol. 32, 1886, pp. 79, 94, pi. 8, fig. 6. Lapworthura miltoni GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1037, fig. 5. SCHONDORF, Palseontographica, vol. 57, 1910, p. 58; Jahrb. nassau- isch. Ver. Naturk., Wiesbaden, vol. 63, 1910, p. 208. SOLLAS and SOLLAS, Philos. Trans. Roy. Soc. London, ser. B, vol. 202, 1912, p. 214, text figs. 1, 2, 4 D, pi. 9, figs. 1, 2; pi. 10, figs. 1-4. SPENCER, Mon. Brit. Pal. Aster- ozoa, pt. 1 (Palseontogr. Soc. for 1913), 1914, pp. 25, 32, 35, 36, pi. 1, fig. 9. Formation and locality.- Abundant in the Lower Ludlow of the Siluric at Leintwardine, England. See elsewhere in this work (p. 237) under Sturtzura leptosoma, which Schondorf states is but the young of L. miltoni. The same author states further that the latter species either embraces several species or the form has a long geologic range, for a number of species and genera appear to be identical with L. miltoni. . Cat. No. 59394, U.S.N.M. LAPWORTHURA SOLLASI Spencer. Lapworthura sollasi SPENCER, Mon. Brit. Pal. Asterozoa, pt. 1 (Palseontogr. Soc. for 1913), 1914, pi. 1, fig. 10. Occurs in the Siluric of Great Britain. LAPWORTHURA (? "probably n. gen.") species (Miller and Gurley). Aganaster (?) sp., MILLER and GURLEY, Sixteenth Rep. Geol. Surv. Indiana, 1891, p. 372, pi. 9, figs. 12, 13; authors' extracts, 1890, p. 58. GREGORY, Proc. Zool. Soc. London for 1896, 1897, p. 1038. ' Original description. "We have figured this fragment to show the anchylosis of the oral plates, and the pores passing through the ambulacral plates, and also those passing between them. It will be observed that one pore passes through each plate near the marginal end, while a double row of pores passes between the plates in the central part of the arm furrow. We believe this is an Aganaster, from the from the actinal area, X 2. The central furrow in the ambulacrum is decidedly V-shaped. Upper Ordovicic (Utica). Near Rome, New York. Collection of Peabody Museum, Yale University. FIG. 4. Australaster giganteus (Etheridge). An axillary area and part of two rays somewhat reconstructed from the original figure. Natural size. Note the single very large axillary plate, the enlarging adambulacrals, and the diminishing inframarginals. "Permo-Carboniferous" (Lower Marine). Parley, Northumberland County, New South Wales. Collection of the Mining and Geological Museum, Sydney. PLATE 5. FIGS. 1 and 2. Hudsonaster matutinus (Hall) (also see plates 2 and 3). 1. Camera lucida drawing X 3.5 of the actinal side of one of the speci- mens photographed on plate 2, fig. 2. Middle Ordovicic (Trenton). Near Newport, New York. Collection of Museum of Comparative Zoology, Harvard University (No. 26). 2. Camera lucida drawing X 4.5 of the abactinal area of a specimen in which the ossicles are somewhat displaced. Middle Ordovicic (Trenton). Lachine, near Montreal, Canada. Collection of the United States National Museum (Cat. No. 60603). Collected by W. R. Billings. PLATE 6. FIGS. 1 and 2. Hudsonaster incomptus (Meek). 1. Camera lucida drawing X 4 of the actinal side. Upper Ordovicic (Richmondian). Near Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 40882). 2. Camera lucida drawing X 4 of the abactinal area in a well preserved specimen. The anal opening may have been in the first circle of small plates adjacent to the centro-dorsal piece and in the same interradius as the madreporite. Upper Ordovicic (Richmondian). Near Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 40882). REVISION OF PALEOZOIC STELLEEOIDEA. 291 PLATE 7. FIGS. 1-4. Palseaster niagarensis Hall. 1. Camera lucida drawing X 4 of the abactinal side in its present preservation. Holotype. 2. Abactinal side of theholotype restored, X 4. The drawing probably has too many accessory disk pieces. 3. Camera lucida drawing X 4 of the a'ctinal side of one of the rays of the holotype. Since this drawing was made the rock in the ambulacral furrow has been dug away down to the ambulacral plates, which are small and deep-seated. 4. A few of the inframarginals and supramarginals from the abactinal side to show the smooth central areas surrounded by granular borders, X 4. Siluric (Rochester shale). Lockport, New York. Collection of Cornell University (No. 7331). FIG. 5. Mesopalseaster find (Ulrich) (also see plate 9). Camera lucida drawing X 16 of one of the rays. The ossicles are always more or less displaced in this species. Upper Ordovicic (Eden shales). Cincinnati, Ohio. Ulrich collection of the United States National Museum (Cat. No. 60604). PLATE 8. FIGS. 1 and 2. Mesopalseaster shafferi (Hall). 1. Camera lucida drawing X 8 of the abactinal side of a one-third adult-sized specimen. As the medial disk pieces are displaced in the individual, they are here drawn in in their probable natural arrangement all should be pointed like the central disk plate. Upper Ordovicic (Maysvillian, Corryville member). Cincinnati, Ohio. Vaupel collection of the United States National Museum (Cat. No. 60605). 2. Camera lucida drawing X 3 of the actinal area of a mature indi- vidual. Upper Ordovicic (Kichmondian). Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 59391). FIG. 3. Miomaster drevermanni Schondorf. Actinal side as reconstructed by Schondorf. Natural size. Lower Devonic (Upper Coblenzian). Miellen, Germany. PLATE 9. FIG. I. Mesopalseaster (?) parviusculus (Billings). Camera lucida drawing X 5 of a gutta-percha squeeze of the holotype, actinal side. Siluric (Lower Arisaig). Arisaig, Nova Scotia. Original at McGill University; gutta-percha squeeze in United States National Museum (Cat. No. 60620). FIG. 2. Mesopalseaster (?) cataractensis, new species. Photograph X 3 of the well-preserved holotype. Siluric (Cataract formation). Hamilton, Ontario. Collection of the Hamilton Natural History Society. FIG. 3. Mesopalseaster (?) granti (Spencer). Photograph X 2 of a specimen showing the dorsal side. Siluric (Cataract formation). Hamilton, Ontario. Collection of Peabody Museum, Yale University. 292 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. FIG. 4. Mesopalseaster intermedius, new species. Photograph X 3 of the holotype. Between the two rays lies an arm of a crinid, giving the impression that this form has a large disk. Upper Ordovicic (Maysvillian). Cincinnati, Ohio. Collection of the University of Chicago (No. 9575). FIG. 5. Mesopalseaster finei (Ulrich) (also see plate 7). Ulrich's original figures. Fig. 5, one of the cotypes from the abactinal side X 2; fig. 5a, the madreporite X 6; fig. 56, a ray from the actinal side X 3. Upper Ordovicic (Eden). Cincinnati, Ohio. Ulrich collection of the United States National Museum (Cat. No. 60604). FIG. 6. Mesopalseaster caractad (Gregory) (also see plate 11). Retouched photograph X 3 of a wax squeeze by Bather, now in the United States National Museum (Cat. No. 60606). Actinal area. Ordovicic (Caradoc). Church Stretton, England. PLATE 10. FIGS. 1 and 2. Mesopalseaster clarki (Clarke and Swartz). 1. Drawing of the actinal side of the holotype X 1.5. 2. Abactinal area X 1.5. Note the few accessory pieces in the angles between the radials. Upper Devonic (Chemung). Near Oakland, Maryland. Collection of the Maryland Geological Survey. PLATE 11. FIG. 1. Mesopalxaster caractad (Gregory) (also see plate 9). Retouched photograph X 3 of a wax squeeze by Bather, now in the United States National Museum (Cat. No. 60606). Abactinal area. Ordovicic (Caradoc). Church Stretton, England. FIG. 2. Devonaster chemungensis, new species. Photograph of the natural mold of the actinal side of the holotype, natural size. Upper Devonic (Chemung). "Central Pennsylvania." CoUection of Columbia University (No. 6228G). PLATE 12. FIGS. 1 and 2. Spaniaster latiscutatus (Sandberger). 1. Abactinal side. 2. Actinal side. After Schondorf . Lower Devonic. Germany. FIGS. 3-5. Devonaster eucharis (Hall). 3. Abactinal side, natural size, a, the madreporite. After Hall. 4. Actinal side, natural size. After Hall. The tiny ambulacral plates should have been drawn as opposite one another and not alternate. 5. Hall's diagram (3a) of the actinal plate arrangement. The ambu- lacrals are incorrectly drawn, as their arrangement is opposite and not alternate, a, ambulacralia; aa, adambulacralia; m, inframarginalia; o, pairs of oral armature ossicles; p, podial openings, but drawn a little too large; tm, the single interbrachial axillaries. Middle Devonic (Hamilton). Near Hamilton, New York. REVISION OF PALEOZOIC STELLEROIDEA. 293 PLATE 13. FIGS. 1 and 2. Promopalseaster wilsoni (Raymond). 1. Photograph X 1.5 of the holotype, showing the abactinal side. 2. One of the rays of the holotype X 3. Middle Ordovicic (Lowville). City View, Ottawa, Canada. Collection of Miss A. E. Wilson. FIG. 3. Promopalseaster prenuntius, new species (also see plate 15). Photograph, natural size, of the actinal side of the holotype. Middle Ordovicic (Lower Trenton). Frankfort, Kentucky. Collection of the State University of Kentucky. FIG. 4. Anorthaster miamiensis (Miller) (see also plate 20). Retouched photograph, natural size, of the actinal side of the holotype. Upper Ordovicic (Richmondian). Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 40880). FIG. 5. Neopalseaster crawfordsvillensis (Miller) (see also plate 23). Retouched photograph, natural size, of the holotype, showing the abac- tinal area. Mississippic (Keokuk). Crawfordsville, Indiana. Harris collection of the United States National Museum (Cat. No. 60607). PLATE 14. FIGS. 1 and 2. Promopalseaster speciosus (Meek)? Reproduction, natural size, of the original lithograph of "Asterias prir mordialis." Probably the young of P. speciosus. Upper Ordovicic (Maysvillian). Cincinnati, Ohio. The specimen is now lost. FIGS. 3 and 4. Promopalseaster speciosus (Meek) (also see plate 15). Photographs, natural size, of the abactinal and actinal sides of the holotype. Upper Ordovicic (Maysvillian). Cincinnati, Ohio. Dyer collection of the Museum of Comparative Zoology, Harvard University (No. 22). PLATE 15. FIGS. 1-4. Promopal&easter speciosus (Meek) (also see plate 14). 1. Camera lucida drawing X 4 of an interbrachial area. 2. Camera lucida drawing X 4 across a ray at about its mid-length. From the actinal side, showing the ambulacrals, podial openings, adambulacrals, inframarginals, and supramarginals. 3. Two adambulacral spines near oral region, X 7. 4. A part of the abactinal area of a ray near its mid-length, X 2.5. All drawn from the holotype. Upper Ordovicic (Maysvillian). Cincinnati, Ohio. Dyer collection of the Museum of Comparative Zoology, Harvard Uni- versity (No. 22). FIG. 5. Promopalseaster prenuntius, new species (also see plate 13). Diagram of the actinal interbrachial plate arrangement. 294 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. FIGS. 6-8. Promopalseaster bellulus, new species (also see plates 16 and 18). 6. Camera lucida drawing X 4 of an interbrachial area. The adambu- lacrals are in natural position, but the interbrachial axillaries and the inframarginals are somewhat displaced. Taken from the specimen on plate 16, fig. 1. 7. Camera lucida drawing X 7 of the abactinal area of a ray. The ossicles are all displaced, but originally they were united in a spicular and partially overlapping mesh. 8. Madreporite X 7. Upper Ordovicic (Richmondian). Waynesville, Ohio. Harris collection of the United States National .Museum (Cat. No. 40879). PLATE 16. FIG. 1. Promopalseaster bellulus, new species (also see plates 15 and 18). Retouched photograph, natural size, of the actinal side. The specimen is in limestone. Upper Ordovicic (Richmondian). Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 40879). FIG. 2. Promopalseaster spinulosus (Miller and Dyer) (also see plates 17 and 18). Retouched photograph, natural size, of the actinal side of the holotype of Palseaster longibrachiatus Miller. The well-preserved columns are the inframarginals, but in places some of the adambulacrals are also shown. Upper Ordovicic (Richmondian). Near Clarksville, Ohio. Harris collection of the United States National Museum (Cat. No. 40881). PLATE 17. FIGS. 1 and 2. Promopalseaster spinulosus (Miller and Dyer) (also see plates 16 and 18). Retouched photographs X 3 of the actinal and abactinal sides of the holotype. Upper Ordovicic (Richmondian). Probably near Waynesville, Ohio. Dyer collection of the Museum of Comparative Zoology, Harvard University (No. 16). PLATE 18. FIGS. 1-3. Promopalseaster spinulosus (Miller and Dyer) (also see plates 16 and 17). 1. Camera lucida drawing X 2, somewhat restored, of the holotype of Palseaster longibrachiatus Miller. 2. Camera lucida drawing X 7 of the abactinal area of a ray near its base. Shows the large infra- and supramarginals and ambital pieces. Center of ray has spicular ossicles. 3. Madreporite X 7. Upper Ordovicic (Richmondian). Near Clarksville, Ohio. Harris collection of the United States National Museum (Cat. No. 40881). REVISION OF PALEOZOIC STELLEEOIDEA. 295 FIGS. 4 and 5~Promopalseaster bellulus, new species (also see plates 15 and 16). 4. Adambulacralia and their spines, X 7. The ambulacrum lies to the left. 5. A few abactinal spines, X 7. Upper Ordovicic (Richmondian). Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 40879). FIG. G.Promopalseaster wyloffi, (Miller and Gurley) (also see plate 19). Diagram X 3.5 of one of the interbrachial areas. From the holotype. Upper Ordovicic (Richmondian). Near Madison, Indiana. Collection of the University of Chicago (No. 6066). FIG. 7. Promo palaeaster exculptus (Miller) (also see plate 20). Adambulacral and ambulacral plates at about mid-length of a ray, X 4. From the holotype. Upper Ordovicic (Richmondian). Near Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 60608). FIG. S.Promopalssaster dyeri (Meek) (also see plates 20 and 25). Five inframarginal plates X 4. To show the articulations for the stout spines that lie on the right. Also note the smaller spines. The plate extensions to the left connect with the vertical ridges on the ambulacral plates. From the holotype. Upper Ordovicic (Maysvillian). Cincinnati, Ohio. Dyer collection of the Museum of Comparative Zoology, Harvard Univer- sity (No. 13). PLATE 19. FIG. 1. Trimeraster parvulus Schondorf. Reconstruction X 1.75 of the actinal side. After Schondorf. Lower Devonic (Coblenzian). Near Rhens, Germany. FIG. 2. Promo palseaster wykoffi (Miller and Gurley) (also see plate 18). Photograph X 2 of the holotype, showing the actinal side. One of the finest of Paleozoic specimens. Upper Ordovicic (Richmondian). Near Madison, Indiana. Collection of the University of Chicago (No. 6066). PLATE 20. FIG. 1. Anorthaster miamiensis (Miller) (also see plate 13). Camera lucida drawing X 5.5 of one of the interbrachial areas of the holotype. The only Paleozoic form known with the area wholly made up of adambulacrals. Upper Ordovicic (Richmondian). Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 40880). FIG. 2.Promopalseaster exculptus (Miller) (also see plate 18). Camera lucida drawing X 4 of an interbracbial area. Note the modified ambulacral plates in connection with the oral armature. Holotype. Upper Ordovicic (Richmondian). Near Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 296 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. FIGS. 3-6. Promopalseaster dyeri (Meek) (also see plates 18 and 25). 3. One of the interbrachial areas with the ossicles displaced and the adjoining adambulacral plates with their spines, X 4. 4. Madreporite X 4. The rest of the plate is covered. 5. Two abactinal spines, X 7. 6. Five different abactinal ossicles, X 7. All drawn from the holotype. Upper Ordovicic (Maysvillian). Cincinnati, Ohio. Dyer collection of the Museum of Comparative Zoology, Harvard Uni- versity (No. 13). PLATE 21. FIG. 1. Promopalseaster magnificus (Miller) (also see plates 22 and 23). Photograph, natural size, of the actinal side of one of the cotypes. This speci nen is probably the best preserved of Paleozoic starfishes and is one of the largest species. Upper Ordovicic (Richmondian). Near Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 40883). PLATE 22. FIG. 1. Promopalseaster magnificus (Miller) (also see plates 21 and 23). Abactinal side of the same specimen illustrated on plate 21. PLATE 23. FIGS. 1-3. Promopalseaster magnificus (Miller) (also see plates 21 and 22). 1. Camera lucida drawing X 3.5 of an interbrachial area. Additional tube-feet probably also came out at the junction of the forked crests. 2. Camera lucida drawing X 4 of the actinal area of one of the rays toward the distal ends. Large adambulacrals bound the rays distally, while the smaller pieces outside are the inframarginals. 3. A distal portion of the abactinal area, X 3.5. All drawn from the cotypes. Upper Ordovicic (Richmondian). Near Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 40883). FIG. 4. Neopalseaster crawfordsvillensis (Miller) (also see plate 13). Camera lucida drawing X 4 of the abactinal area of the holotype. Ac- cessory disk pieces are drawn in; in the specimen they are all displaced. The madreporite should be more finely striate. In places the dorsal skeleton is lost, exposing the ambulacrals. Mississippic (Keokuk). Crawfordsville, Indiana. Harris collection of the United States National Museum (Cat. No. 60607). FIGS. 5-7. Petraster spedosus (Miller and Dyer) (also see plates 26 and 27). 5. The ambulacral, adambulacral and inframarginals of a ray at its mid-length, X 4. Seen from the actinal side. 6. Inframarginal, accessory interbrachial, and adambulacral pieces, X 4. Seen from the dorsal side. 7. Ambulacral and adambulacrale, X 4. Seen from the dorsal area. REVISION OF PALEOZOIC STELLEROIDEA. 297 PLATE 24. FIGS. 1-2. Xenaster margaritatus Simonovitsch. 1. Reconstruction, somewhat enlarged, of the abactinal side. The structure of the medial portion of the disk is based upon Agalmaster miellensis. 2. Reconstruction, somewhat enlarged, of the actinal area. After Schondorf. * Lower Devonic (Upper Coblenzian). Niederlahnstein, Germany. PLATE 25. FIG. 1. Promopalssaster dyeri (Meek) (also see plates 18 and 20). Retouched photograph, natural size, of the actinal side of the holotype. Upper Ordovicic (Maysvillian). Cincinnati, Ohio. Dyer collection, Museum of Comparative Zoology, Harvard University (No. 13). FIG. 2. Lepidasterella babcocki, new species. Photograph X 1.5 of the holotype. A natural mold of the abactinal area in sandstone. Upper Devonic (Lower Chemung). Near Ithaca, New York. Collection of Mrs. G. W. Babcock. PLATE 26. FIG. 1. Petraster spedosus (Miller and Dyer) (also see plates 23 and 27). Retouched photograph, natural size, of the abactinal area of the holotype. Upper Ordovicic (Richmondian). Near Winchester, Ohio. Dyer collection of the Museum of Comparative Zoology, Harvard Uni- versity (No. 14). FIG. 2. Petraster (?) americanus (D'Orbigny). Reproduction, natural size, of Meek's original figure. The specimen is seen from the abactinal side, exposing, however, the inner surface of the actinal skeleton, which is deeply embedded in limestone. Upper Ordovicic (Maysvillian). Cincinnati, Ohio. Collection of the University of Chicago. PLATE 27. FIGS. 1-4. Petraster spedosus (Miller and Dyer) (also see plates 23 and 26). 1. Camera lucida drawing X 4 of the abactinal side of a ray. Holotype The inframarginals border the animal, with the supramarginals imme- diately inside of them. The radial column is readily distinguished along the center of the rays. 2. Camera lucida drawing X 6 of an interbrachial area in a young individual (Harris collection, Cat. No. 60609, U.S.N.M.). The inframar- marginals are above and the adambulacrals below to the right. The rest rest of the plates are accessory interbrachials. 3-4. Madreporite from the side and from the lower or under surface, X 8. FIG. 5. Petraster rigidus (Billings). Outline tracing of the actinal skeleton from a photograph X 2 of the holotype. Middle Ordovicic (Trenton). Ottawa, Canada. Collection of the Geological Survey of Canada (No. 1401a). 298 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. FIGS. 6-8. Urasterella grandis (Meek) (also see plates 28 and 30). 6. Camera lucida drawing X 8 of part of the abactinal area of a ray proxi. mally. Note how the ossicles are drawn out into nonarticulating blunt spines. 7. Camera lucida drawing X 4 of the actinal side about the mouth. The complete oral armature is preserved. 8. Madreporite in outline, X 7. From specimens in the Harris collection of the United States National Museum (Cat. No. 40885). PLATE 28. FIGS. 1 and 2. Urasterella grandis (Meek) (also see plates 27 and 30). 1. Retouched photograph, natural size, of a large and folded specimen. Both ventral and dorsal areas are shown. 2. Retouched photograph, natural size, of an average individual seen from .the actinal side. Upper Ordovicic (Richmondian). Near Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 40885). FIGS. 3 and 4. Urasterella pulchella (Billings) (also see plate 30). 3. Photograph, natural size, of an electrotype, seen from the actinal side. Middle Ordovicic (Trenton). Hull, Ottawa, Canada. Original in collection of the Geological Survey of Canada; electrotype in United States National Museum (Cat. No. 60610). 4. Retouched photograph, natural size, of the abactinal area. Middle Ordovicic (Trenton). Trenton Falls, New York. Collection of the Museum of Comparative Zoology, Harvard University (No. 31). FIG. 5.' Urasterella girvanensis, new species. Reproduction X 2 of the original figure by Nicholson and Etheridge, somewhat altered. From a wax squeeze made by Bather, now in the United States National Museum (Cat. No. 60611). Upper Ordovicic. Thraive, Girvan, Scotland. PLATE 29. FIG. 1. Urasterella ulrichi, new species (also see plate 30). Retouched photograph, natural size, of the abactinal side of a cotype. The three short blunt rays are interpreted as healed stumps. Middle Ordovicic (Black River). Minneapolis, Minnesota. Ulrich collection of the United States National Museum (Cat. No. 60612). FIG. 2. Urasterella huxleyi (Billings). Photograph X 2 of the holotype. Abactinal view. Middle Ordovicic (about Chazy). Point Rich, Newfoundland. . Collection of the Geological Survey of Canada (No. 554.) PLATE 30. FIGS. 1-4. Urasterella grandis (Meek) (also see plates 27 and 28.) 1-2. Camera lucida drawing X 16 of the adambulacrals with their articu- lar spines and probable paxillse. 3. Camera lucida drawing X 8 of the two columns of ambulacra! plates a little displaced. The podial openings are situated laterally between the thinner ends of the ossicles. 4. Camera lucida drawing X 7 of the adambulacrals with their spines. From specimens in the Harris collection of the United States National REVISION OF PALEOZOIC STELLEROIDEA. 299 FIG. 5. Urasterella pulchella (Billings) (also see plate 28). Camera lucida drawing X 8. The large plates to the left are the coin- shaped adambulacrals (some have tipped over and show their articular faces), next are the small elongate inframarginals, and outside of these are the spinose ambitals. FIGS. 6 and 7. Urasterella ulrichi, new species (also see plate 29). 6. Camera lucida drawing X 4 of the abactinal side of a young individual. The medial ray plates should be more tumid than here represented. Note the primitive structure of the disk. 7. Camera lucida drawing X 6 of the actinal side of a young indi- vidual. The outer dark margin is adhering rock, through which some of the abactinal ambital nonarticulating spines project. Note the phanerozonian interbrachial characters. Middle Ordovicic (Black River). Minneapolis, Minnesota. Ulrich collection of the United States National Museum (Cat No. 60612). PLATE 31. FIGS. 1 and 2. Compsaster formosus Worthen and Miller. 1. Actinal view, natural size, of the holotype. Reproduction of the original illustration. 2. Part of a ray, X 2. After Worthen and Miller. Upper Mississippic (Chester). Okaw Bluffs, Illinois. " Illinois State collection of 1880, No. 2476." FIG. S.Calliasterella mira (Trautschold). Reproduction, natural size, of Schondorf's reconstruction. Upper Carboniferous (Moscovian). Mjatschkowa, near Moscow, Russia. PLATE 32. FIG. 1. Stenaster salteri Billings. Retouched photograph X 3 of a cotype. Actinal view. Middle Ordovicic (Trenton). Belleville, Ontario, Canada. Collection of the Geological Survey of Canada. FIG. 2. Schuchertia stellata (Billings) (also see plate 33). Photograph X 4 of the actinal side of the holotype. Middle Ordovicic (Trenton). Ottawa, Canada. Collection of the Geological Survey of Canada (No. 1399). FIG. 3. Schuchertia laxata, new species (also see plate 33). Retouched photograph X 2 of the holotype from the abactinal side. Upper Ordovicic (Richmondian). Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 60613). FIGS. 4 and 6. Schcenaster (?) legrandensis Miller and Gurley. 4-5. Actinal and abactinal views, natural size. 6. Part of actinal side X 6.5. After Miller and Gurley. Lower Mississippic (Kinderhookian). Le Grand, Iowa. ?Gurley collection of the University of Chicago. PLATE 33. FIG. 1. Schuchertia stellata (Billings) (also see plate 32). Camera lucida drawing X 8 of the abactinal surface. Somewhat restored. Middle Ordovicic (Trenton). Near New Edinburgh, Canada. Collection of the Geological Survey of Canada. FIGS. 2 and 3. Schuchertia laxata, new species (also see plate 32). 2. Some of the outer ambital plates, all disjointed, X 8. 3. An interbrachial area and ambulacrum, X 4. Upper Ordovicic (Richmondian). Waynesville, Ohio. Harris collection of the United States National Museum (Cat. No. 300 BULLETIN 88, UNITED STATES NATIONAL MUSEUM. FIG. 4. Tetraster wyville-thoTtisoni Nicholson and Etheridge. Diagram X 4 of the actinal skeleton. From a wax squeeze by Bather, in the United States National Museum (Cat. No. 60614). Upper Ordovicic. Thraive, Girvan, Scotland. FIG. 5. Schcenaster (?) montanus Raymond. Reproduction X 2 of the original figure. Actinal side. Mississippi (Madison). Spring Canyon, near Alder, Montana. Collection of the Carnegie Museum. FIG. 6. SJicenaster (?) wachsmuihi Meek and Worthen. Reproduction, natural size, of the original figure. Actinal view. Lower Mississippi (Burlington). Burlington, Iowa. Collection of the Museum of Comparative Zoology, Harvard University (No. 7). PLATE 34. FIG. 1. Palseosolaster (?) gyalum (Clarke). Actinal view of one of the cotypes, natural size. Upper Devonic (Portage). Ithaca, New York. Collection of Cornell University. PLATE 35. FIGS. 1-4. Schcenaster fimbriatus Meek and Worthen. 1. View of the abactinal side, natural size, of one of the cotypes. 2. Enlarged view of one of the rays, showing the arrangement of the ossicles and the pores, as seen when the surface is ground down. 3. Actinal side of a ray, enlarged, showing only the adambulacrals. 4. Actinal side of another cotype, natural size. The original figures of plate 19, figs. 7 a to 7d. Upper Mississippi (St. Louis). St. Clair County, Illinois. FIGS. 5 and 6 Encrinaster petaloides (Simonovitsch). 5. Abactinal view, about X 2. 6. An arm from the actinal side, about X 2. After Schondorf. Lower Devonic (Upper Coblenzian). Niederlahnstein, Germany. PLATE 36. FIG. 1. Tseniaster spinosus (Billings). Photograph X 2 showing the actinal side of one of the cotypes. Middle Ordovicic (Lower Trenton). Montmorency Falls, Quebec, Canada. Collection of the Geological Survey of Canada (No. 1404). FIGS. 2 and 3.Taeniaster cylindricus (Billings). 2. Abactinal view X 2 of one of the cotypes. 3. Actinal view X 2 of another cotype. Both specimens are on the same slab. Middle Ordovicic (Trenton). Ottawa, Canada. Collection of the Geological Survey of Canada (No. 1405a). FIG. 4. Alepidaster flexuosus (Miller and Dyer). An unpublished photograph X 3.5, made many years ago, of the holo- type of Protasterina fimbriata Ulrich. Upper Ordovicic (Lower Edenian). Covington, Kentucky. Ulrich collection of the United States National Museum (Cat. No. 60615). REVISION OF PALEOZOIC STELLEKOIDEA. 301 PLATE 37. FIGS. 1 and 2. Eucladia woodwardi Sollas. Dorsal and ventral reconstructions X 5/3 of this remarkable fossil. On it is based the order Ophiocistia Sollas. "One arm, the most proximal of the radius on the lower right-hand corner, is omitted, so as to show the aperture from which it proceeds." After Sollas. Siluric (Lower Ludlow). Leintwardine, England. PLATE 38. FIG. l.Eucladm (?) beecheri, new species. Photograph X 2 of the holotype from the actinal side. The specimen itself is difficult to make out, hence the indistinctness of the photograph. Lower Devonic (Coeymans). Jerusalem Hill, Litchfield, Herkimer County, New York. Collection of Peabody Museum, Yale University. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 1 HUDSON ASTER NARRAWAYI. FOR EXPLANATION OF PLATE SEE PAGE 289. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 2 1, HUDSONASTER NARRAWAYI; 2, H. MATUTINUS. FOR EXPLANATION OF PLATE SEE PAGE 289. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 3 1, HUDSONASTER RUGOSUS; 2, H. MATUTINUS; 3, H. BATHERI. FOR EXPLANATION OF PLATE SEE PAGE 289. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 4 1, HUDSONASTER NARRAWAYI; 2, H. MILLERI; 3, MESOPAL/EASTER (?) LANCEOLATUS; 4, AUSTRALASTER GIQANTEUS. FOR EXPLANATION OF PLATE SEE PAGE 290. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 5 HUDSONASTER MATUTINUS. FOR EXPLANATION OF PLATE SEE PAGE 290. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 6 HUDSONASTER INCOMPTUS. FOR EXPLANATION OF PLATE SEE PAGE 290. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 7 1-4, PAL/EASTER NIAGARENSIS; 5, MESOPAL/EASTER FINEI FOR EXPLANATION OF PLATE SEE PAGE 291. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 8 1, 2, MESOPAL/EASTER SHAFFERI; 3, MIOMASTER DREVERMANNI. FOR EXPLANATION OF PLATE SEE PAGE 291. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 9 Six SPECIES OF MESOPALXE ASTER. FOR EXPLANATION OF PLATE SEE PAGES 291 AND 292. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 10 MESOPAL/E ASTER CLARKI. FOR EXPLANATION OF PLATE SEE PAGE 292. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 11 , MESOPAL/EASTER CARACTACI; 2, DEVONASTER CHEMUNGENSIS. FOR EXPLANATION OF PLATE SEE PAGE 292. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 12 1 , 2, SPANIASTER LATISCUTATUS; 3-5, DEVONASTER EUCHARIS. FOR EXPLANATION OF PLATE SEE PAGE 292. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 13 1-3, PROMOPAMEASTER; 4, ANORTHASTER; 5, NEOPAL/EASTER. FOR EXPLANATION OF PLATE SEE PAGE 293. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 14 Asterlas prtmorci tails PROMOPALXEASTER SPECIOSUS. FOR EXPLANATION OF PLATE SEE PAGE 293. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 15 1-4, PROMOPAL/EASTER SPECIOSUS; 5, P. PRENUNTIUS; 6-8, P. BELLULUS. FOR EXPLANATION OF PLATE SEE PAGES 293 AND 294. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 16 1, PROMOPAL/EASTER BELLULUS; 2, P. SPINULOSUS. FOR EXPLANATION OF PLATE SEE PAGE 294. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 17 PROMOPAL/CASTER SPINULOSUS. FOR EXPLANATION OF PLATE SEE PAGE 294. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 18 4 DETAIL OF FIVE SPECIES OF PROMOPAL/EASTER. FOR EXPLANATION OF PLATE SEE PAGES 294 AND 295. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 19 1 , TRIMERASTER PARVULUS; 2, PROMOPAL/EASTER WYKOFFI. FOR EXPLANATION OF PLATE SEE PAGE 295. S. NATIONAL MUSEUM BULLETIN 88 PL. 20 1, DETAIL OF ANORTHASTER; 2-6, PROMOPAL/EASTER. FOR EXPLANATION OF PLATE SEE PAGES 295 AND 296. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 21 PROMOPALXEASTER MAGNIFICUS. FOR EXPLANATION OF PLATE SEE PAGE 296. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 22 PROMOPAL/EASTER MAQNIFICUS. FOR EXPLANATION OF PLATE SEE PAGE 296. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 23 1-3, DETAIL OF PROMOPALXEASTER; 4, NEOPAL/EASTER; 5-7, PETRASTER. FOR EXPLANATION OF PLATE SEE PAGE 296. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 24 XENASTER MARQARITATUS. FOR EXPLANATION OF PLATE SEE PAGE 297. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 25 1, PROMOPAL/EASTER DYERI; 2, LEPIDASTERELLA BABCOCKI. FOR EXPLANATION OF PLATE SEE PAGE 297. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 26 i j:| | .$ - t: r1-% M/7-? 4- > y . ' - l -4- >r r^T > - -^f^^f 1 : fc# V : '>'" wtt. m?' ^%| "'" 1, PETRASTER SPECIOSUS; 2, P. (?) AMERICANUS. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 27 1-5, DETAIL OF PETRASTER; 6-8, URASTERELLA. FOR EXPLANATION OF PLATE SEE PAGES 297 AND 298. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 28 1, 2, URASTERELLA GRANDIS; 3, 4, U. PULCHELLA; 5, U. QIRVANENSIS. FOR EXPLANATION OF PLATE SEE PAGE 298. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 29 1, URASTERELLA ULRICHI; 2, U. HUXLEYI. FOR EXPLANATION OF PLATE SEE PAGE 298. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 30 DETAIL OF URASTERELLA. FOR EXPLANATION OF PLATE SEE PAGES 298 AND 299. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 31 1, COMPSASTER FORMOSUS; 2, CALLIASTERELLA MIRA. FOR EXPLANATION OF PLATE SEE PAGE 299. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 32 1, STENASTER; 2, 3, SCHUCHERTIA; 4-6, SCHCENASTER (?). U. S. NATIONAL MUSEUM BULLETIN 88 PL. 33 ^K^fflK^ 1-3, DETAIL OF SCHUCHERTIA; 4, TETRASTER; 5, 6, SCHCENASTER (?). U. S. NATIONAL MUSEUM BULLETIN 88 PL. 34 .~~ Kg^K, ji ^ .^ "^l^te'^ PAL/EOSOLASTER (?) GYALUM. FOR EXPLANATION OF PLATE SEE PAGE 300. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 35 1-4, SCHCENASTER FIMBRIATUS; 5, 6, ENCRINASTER PETALOIDES. FOR EXPLANATION OF PLATE SEE PAGE 300. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 36 1, T>ENIASTER SPINOSUS; 2, 3, T. CYLINDRICUS; 4, ALEPIDASTER FLEXUOSUS. FOR EXPLANATION OF PLATE SEE PAGE 300. U. S. NATIONAL MUSEUM BULLETIN 88 PL. 37 EUCLADIA WOODWARDI. FOR EXPLANATION OF PLATE SEE PAGE 301. U. S. NATIONAL MUSEUM BULLETIN 88 PL. EUCLADIA (?) BEECHERI. FOR EXPLANATION OF PLATE SEE PAGE 301. INDEX. [The important references In each case are in heavy-faced type. Synonyms are in italics.] Page. abactinal 13 abnormal development 101 aboral, see abactinal. accessory actinals 13 accessory interbrachials 13, 15,45,46 accessory plates 13, 47, 51 Acroura ( Ophiura) schlotheimii 18 actinal 13 acuminatus, Asterias. ., 93 acuminatus, Mesopalseaster (?) 75, 93 acuminatus, Roemeraster (?) 93 adambulacralia, see adambnlacrals. adambulacral oral skeleton 16 adambulacral spines 13, 43 adambulacrals (see also lateral shields) 13, 34,43,162,213,214,222 Agalmaster 40, 49, 128, 13 1 , 133 A.grandis 132 A. intermedium 133 A.miellensis 132 Aganaster 263 A . gregarius 264 A. (?) sp. Miller and Gurley 251 Aganasteridse 246, 263 Alepidaster 216,228 A. flexuosus 230, 231 A. granuliferus 229, 230 A. gregarius 264 A. miamiensis '. 230, 238 A. n. sp 230 ambital and ambital plates 13, 46, 174 ambulacralia, see ambulacrals. ambulacral oral skeleton 16 ambulacrals 13,33,34,41,213,235,258 alternate arrangement 13, 33, 41 , 152, 172, 174, 213, 214, 215, 222, 226 opposite arrangement 13, 41,163,213,214,224,246 americanus, Cadaster 146 americanus, Petraster (?) 139, 146 ampullae 213 anal opening 13,89,54,61 Anorthaster 41,42,43,50, 125 A. miamiensis 127 Anorthasterinse 52, 74, 126 anthonii, A sterlas 146 antiqua, Argaster 86 antiqua, Asterias, Hisinger 149 antiqua, Asterias, Troost 86 antiqua, Hisingeraster 149 antiqua, Palseaster (Argaster) 80 antiqua, Palxasterina 149 50601 Bull. 8815 20 Page. antiqua, Palmipes 149 antiqua, Petraster (?) 86 antiquata, Asterias 109 antiquata, Palxaster 109 antiquus, Asteriscus 149 antiquus, Lindstromaster 149, 153 antiquus, Mesopalseaster (?) 75, 86 antiquus, Palxaster 86 anus, see anal opening. approximata, Palxasterina 142 A rchxasterias 129, 130 arcs, see axil. ? Argaster 74, 77, 87 A . antiqua 86 arnoldi, Aspidosoma 243 arnoldi, Encrinaster 241, 248 asper, Onychaster 270, 2 72 asperrima, Palxaster. 187 asperrima, Urasterella (?) 175, 18 7 asperrimus, Salteraster 187 asperrimus, Tetraster 187 asperula, A sterias 188, 262 asperula, Roemeraster 188 asperula, Urasterella 175, 188 Aspidosoma 241 A . arnoldi 243 A.eifelense 244 A. goldfussi 243 A . gray se 245 A. petaloides 243 A. petaloides goslariensis 243 A . ? pontis 244 A . roemeri 244 A. sehmidti 244 A. tischbeinianum 244 Aspidosomatidx 241 Asterias 33 A sterias Graham, Anthony and James 146 A . acuminatus 93 A . anthonii 146 A . antiqua Hisinger. 149 A . antiqua Troost 86 A . antiquata 109 A, asperula 188,262 A.constellata 187 A. matutina 57 A. montanus 189 A . primxva 167 A . primordialis 108 A. (Archxasterias) rhenana 131 A. spinosissima 19 asterid radicle 34 ysterina 33 303 304 INDEX. Page. Asteriscus antiquut 149 Asteroidea 41, 42, 43, 51 evolution of 11, 30, 48, 52, 106 number of, in Paleozoic 2 7, 28 terminology 13 Astropecten 161 A. (?) schliiteri 161 Ataxaster 161 A. pygmseus 161 Auluroidea 80,38,40,41,42,43,163,212,213 geological distribution of 28 number of, in Paleozoic 27,28 Australaster 44,45,49,66,72, 171 A.giganteus 73, 171 A. (?) stutchburii 73 australis, Txniaster 236 axil 14 axillary inframarginal 14 axillary interbrachial 14 axillary marginal 15 axillary ossicles 34, 49 babcocki, Lepidasterella 160 barrisi, Onychaster 270, 272 barrisi, Protester (?) 272 bathed, Hudsonaster J 55, 65, 167 Bdellacoma 248,252,254 B. vermiformis 254 beecheri, Eucladia (?) 278 bellulus, Mesopalseaster 75, 91 bellulus, Petraster 91 bellulus, Promopalseaster 104, 105, 1 13 beneckei, Bundenbachia 234 bibliography 281-288 biforis, Protaster 226, 227 bivium 14 body-wall 14 bohemica, Eophiura 222 Bohemura 216, 22 8 B.jahni 223 bonneyi, Palxasterina 153 bonneyi, Palasterina 151, 152, 168 brisingoides, Protaster 236 brisingoides, Stiirtzura 236 buccal processes 225 Bundenbach 29 Bundenbachia 216, 284 B . beneckei 234 B. grandis 235 caecal pores 14 Calliaster 190 C. mirus 190 Calliasterella 39, 47, 49,50, 190 C. mira 191 Calliasterellidse 162, 163, 190 Calliasteridx 190 caractaci, Mesopalseaster 75, 92 caractaci, Palseaster 92 caractaci, Protopalxaster 92 carbonarius, Cribellites 274 carinals, see radials. catalogue of Paleozoic Stelleroidea 51 cataractensis, Mesopalaeaster (?) 75, 89 central disk, see disk. central disk plate, see centrodorsal plate. central plate, see centrodorsal plate. centrodorsal plate 14, 34,35,48 Page. Cheiropteraster 40, 199, 202, 252 C. giganteus 202 chemungensis, Devonaster 101 Cholaster 265 C. peculiaris 266 Cholasteridse 246, 265 ciliaris, Ophiura 268 clarkana, Palseaster 61,63 clarkei, Etheridgaster 172 clarkei, Monaster 171, 1 7 2 clarkei, Palseaster, De Koninck 172 clarkei, Palseaster, Miller 61 clarki, Mesopalaeaster (?) 75, 94 clarki, Palxaster 94 Classification 51 Cadaster 95 C. americanus 146 C. latiscutatus 96 C. tenuiradiatus 19 colvini, Palzocoma 253 colvini, Stiirtzaster 253 Compsaster 192 C. formosus 193 C. n.sp 194 Compsasteridse 162, 163, 191 concinna, Eugasterella (?) 238, 289 concinnus, Eugaster 239 confluens, Stenaster (?) 165,167 confragosus, Onychaster 270, 273 constellata, Asterias 187 constellata, Urasterella (?) 175, 187 coronella, Palxaster 167 coronella, Stenaster (?) 165, 167 crawfordsvillensis, Neopalseaster 135, 186 crawfordsvillensis, Palaeaster 136 Cribellites carbonarius 274 Crossaster 38 Cryptozonia.... 80,33,38,40,41,43,44,45,46,47,162 cygnipes, Palseocoma 253 cygnipes, Sturtzaster. 253 cylindrica, Lapworthura 220 cylindrica, Palseocoma 220 cylindrica, Txniura. 220 cylindricus, Tseniaster 216,217,218,219,220 daoulasensis, Furcaster (?) 261 daoulasensis , Protaster 261 darwini, Echinasterella (?) 200 decheni, Eoluidia 262 decheni, Eophiurites 262 decheni, Protaster 248 decheni, Tremataster (?) 247, 248 demissus, Onychaster 270,274 development, abnormal 101 geological, of Stelleroidea 27, 28 Devonaster 40, 44, 46, 47, 49, 50, 74 , 76, 77, 9 7 , 129, 130, 133 D. chemungensis 101 D.eucharis 29,37,98 devonica, Palxnectria 155 devonicus, Palasteriscus 200 difflcilis, Tremataster 247 disk 14,213,225 how enlarged 32 disk accessory plates 34, 46 disk ambitals 13 disk inframarginals 15 INDEX. 305 disk marginals 14 disk plates 14 primary 49,179,184,190,191 disk supramarginals 15 dispar, Xenaster 130, 131 dorsal, see abactinal. dorsals, see radials. dorsal shield 213 dorso-laterals, see accessory plates. drevermanni, Miomaster 96 dubius, Mesopalseaster (?) 75, 85 dubius, Palxaster 85 dyeri, Palxaster 120 dyeri, Petraster 121 dyeri, Promopalaeaster 104, 105, 120 Echinaster 45 Echinasterella 199, 200 E. (?) darwini 200 E . sladeni 200 EcMnasterias 38,207,209, 211 E. spinosus 211 echinatus, Squamaster 249 Echini, radii of 35 Echinodiscaster 38,40,207,209,211 E. multidactylus 211 Echinodiscites 211 Echinodiscus 211 E. multidactylus 211 Echinostella 38,207,209,212 E. traquairi 212 Eifelaster 128, 133 E. follmanni 134 eifelense, Aspidosoma 244 eifelensis, Encrinaster 244 elegans, Protaster 221 elegans, Tseniaster 219,221 elegans, Xenaster 130, 131 elizae, Uranaster 155 Encrinaster (part) 224 Encrinaster 43, 45, 226, 241 E. arnoldi....; 241,243 E. eifelensis 244 E.goldfussi 243 E. (?)grayae 245 E.petaloides 243 E. petaloides goslariensis 243 E.pontis 244 E. roemeri 242, 244 E. schmidti 244 E. tischbeinianus 244 Encrinasterix (part) 215 Encrinasteridse 215, 241 Eoactis 173,178 E. simplex 186 Eoluidia 262 E. decheni 262 Eoluidiidse 246,262 Eophiura 216,222,223,259 E . bohemica 222 Eophiuridx 262 Eophiurites 262 E. decheni 262 Eospondylus 262, 2 63 E . primigenius 263 Etheridgaster 170 Page. E. clarkei 172 eucharis, Devonaster 29,37, 98 eucharis, Palxaster 93 eucharis, Xenaster 93 Eucladia 275, 276 E. (?)beecheri 278 E. johnsoni 276 E. woodwardi 276, 277 Eucladiidge 276 Eugaster 237 E. concinnus 239 E. logani 238 Eugasterella 235, 237, 259 E. (?) concinna 238, 239 E . logani 237, 2 3 8 Euophiuroidea 213 EuryalQ annulatum 20 Euthemon.....' 275,276, 279 E. igerna 279 evolution of Asteroidea .% 30 exculptus, Palxaster 117 exculptus, Promopalseaster 104, 105, 117 eye-plate, see ocular plate. ferox, Palaeodiscus 23 filiciformis, Helianthaster 159, 1 60 fimbriata, Palxasterina 204 fimbriaid, Protaster 231 fimbriata, Protasterina 231 fimbriatus, Schoenaster 204 flnei, Mesopalseaster 75, 81 fine i Palxaster 81 flexilis, Onychaster 269, 270 flexuosus, Alepidaster 230, 231 flexuosus, Protaster 231 flexuosa, Protasterina 231 follmanni, Eifelaster 134 follmanni, Palasterina 157 follmanni, Pseudopalasterina 156, 157 forbesi, Hallaster 254, 2 66 forbesi, Protaster 255 formosus, Compsaster 193 Furcaster 248,259, 261 F. (?) daoulasensis 261 F. palseozoicus 259, 261 Furcasteridx 248 genitals, see interradial plates. genitals in Echini 35 geological development of Stelleroidea 27, 28 geological distribution of Stelleroidea 28 giganteus, Australaster 73, 171 giganteus, Cheiropteraster 202 giganteus, Monaster 73 giganteus, Palxaster ( Monaster) 73 girvanensis, Urasterella 167, 175, 186 goldfussi, Aspidosoma 243 goldfussi, Encrinaster 243 Goniaster 61 grandis, Agalmaster 132 grandis, Sundenbachia 235 grandis, Palseophiomyxa 234, 286 grandis, Stenaster 180 grandis, Urasterella 174, 175, 180 granti, Mesopalaeaster 75, 89 granti, Palxaster 89 granuliferus, Alepidaster 229, 280 306 INDEX. Page. granuliferus, Protester (f) 230 granulosus, Palxaster, Hall 112 granulosus, Palxaster, Meek 109 granulosus, Promopalaeaster 104, 1 1 2, 113 graspers 33 grayx, Aspidosoma 245 grayse, Encrinaster (?) 245 grayi, Lepidast*r 158 gregarius, Aganaster 264 gregarius, Alepidaster 264 gregarius, Protester (/) 264 Gregoriura 216,233 G. spryi 234 gregoryi, Palseosolaster 209 groomi, Protaster 226, 227 gyalum, Helianthaster 160,210 gyalum, Palseosolaster (?) 210 Hallaster 248, 254 H. forbesi 254, 255 Harris, I. H., gift oietarflshes by 9 Jiarrisi, Palxaster 180 Helianthaster Clarke (part) 209 Helianthaster Roemer 38, 40,45,157,158,159,209,211 H. flliciformis 159,160 H. gyalum (part) 160, 210 H. rhenanus 159 H.roemeri 209,211 H. n. sp . Clarke 160 Helianthasteridx 157 Helianthasterinse 157 Heliaster 38,208 hirudo, Palxaster 188 Urudo, Uraster 188 hirudo, Urasterella 175,188 Hisingeraster 148 H, antiqua 149 Hudsonaster 31,32,34,36, 38, 39, 40, 41, 42, 44,45, 47, 48, 49, 52, 68, 69, 75, 105, 135, 140, 158, 165, 179, 194, 196 H.batheri 55,65,167 H. incomptus. 34,36,40,55,61 H. matutinus 34,55, 57, 139, 141, 166 H. milleri ..... 55,60 H.narrawayi 43,50,55,59 H. rugosus 36, 55, 64 Hudsonasteridse] 52, 53,162 huxleyi, Stenaster 182 huxleyi, Urasterella 175, 182 igerna, Euthemon 279 imbricatus, Palxaster 169 imbricatus, Tetraster (?) 168, 169 incomptus, Hudsonaster 34,36,40,55, 61 incomptus, Palxaster. 61 inframarginalia 16, 34, 48, 152, 194 inframarginal plates 16, 33, 174, 179, 185 Interact] nals 13 interbrachial adambulacrals 1 5, 126 interbrachial areas and arcs.. . . 14, 60, 138, 185, 194 interbrachial marginals 15 intennarginal plates, see ambital plates, intermarginals, see ambital. intermedius, Agalmaster 133 intermedius, Mesopalseaster 75, 79 interadial plates 15, 32 Jaekelaster 40, 192 Page. J. petaliformis 192 jahni, Bohemura 223 jamesi, Palxaster 146 jamesi, Palasterina (t) 146 johnsoni, Eucladia 276 kinahani, Palasterina 155 kinahani, Uranaster 154, 155 Labidaster 208 lanceolatus, Mesopalaeaster (?) 75, 82 Lapworhtura 248, 250, 254 L. cylindrica 220 L. miltonl 250,261,253 L. sollasi. 251 L. (?)sp 251 Lapworthuridae 246, 248 lateral shields 214 latiscutetus, Coelaster 96 latiscutatus, Spaniaster 96 laxata, Schuchertia 196, 198 legrandensis, Schoenaster (?) 206 Lepidaster 38,40,158,160 L . grayi 158 Lepidasteracea 53 Lepidasterella 38, 40, 1 60 L.babcocki 160 Lepidasteridae 52,53,157 leptosoma, Protaster 237 leptosoma, Rhodostoma 237 leptosoma, Sturtzura 236, 237 leptosomoides, Sturtzura 236, 237 Lindstromaster 138, 140, 148, 154, 156 L. antiquus 1 4 9 , 153 Lindstromasterinx 138 logani, Eugaster 238 logani, Eugasterella 237, 288 longibrachiatus, Palxaster 115 Loriolaster 199,200, 201, 202,252 L. mirabilis 201 lymani, Ophiurina 247 Lysophiurae 215 madreporite 15,39,40,209,211,214,222,223,226 magniflcus, Promopalaeaster 42, 104,105,106,113,122 margaritetus, Xenaster 131,132 margaritatus, Xenaster 129, 130, 181 marginalia, see marginal plates. marginal plates 1 6,33, 43, 135, 152, 213 marstoni, Palxocoma 253 marstoni, Sturtzaster 253 matutina, Asterias 57 matutinus, Hudsonaster 34, 55, 6 7,139, 141, 166 matutinus, Palxaster 67 meafordensis, Tseniaster 221 measurements 15 median dorsals, see radials. Medusaster 38,207,212 M. rhenanus 212 Mesopaleeaster 28, 39, 41, 44, 45, 46, 47, 49, 50, 56, 74,95,98, 105, 126, 130, 135, 158 M.(?) acuminatus 75, 93 M. (? ) antiquus 75, 8 6 M.beUulus 75,91 M. caractaci 75, 92 M.(?) cataractensis 75,89 M.(?)clarki 75,94 M.(?)dubius 75,85 INDEX. 307 Page. M.finei 75,81 M. grant! 75,89 M. intermedius 75, 79 M.(?) lanceolatus 75, 82 M.(?) parviusculus 75, 87 M. proavitus 75, 83,85 M.shafferi 44,75,77 M.(?) wilberanus 75, 84 Mesopalseasterinae 52, 74,162 miamiensis, Alepidaster 230, 238 miamiensis, Anorthaster 127 miamiensis, Palseaster 127 miamiensis, Protaster 233 miellensis, Agalmaster 132 milleri, Hudsonaster 55, 60 miltoni, Lapworthura 250, 251, 253 mittoni, Protaster 251 minveri, Sympterura. . : 256 Miomaster 43,50, 74, 96 M. drevermanni 96 Miospondylus 262, 263 M. rhenanus 263 mira, Calliasterella 191 mirabilis, Loriolaster 201 mirus, Calliaster 190 mitchelli, Sturtzaster (?) 254 Monaster Gregory 49, 72 Monaster Etheridge 40, 1 70 M. Clarkei 171, 172 M. giganteus 73 M . stutchburii 73 Monasteridae 162, 163, 1 70 montana, Urasterella 175, 189 montanus, Asterias 189 montanus, Palseaster 189 montanus, Schoenaster (?) 43, 207 mouth angle plate, see oral armature. multidactylus, Echlnodiscaster 211 multidactylus, Echinodiscus 211 multlrayed starfishes 38, 40, 157, 207 narrawayl, Hudsonaster 43, 50, 55, 69 narrawayi, Protopalseaster 59 neglecta, Palseura 223 Neopalseaster 40, 43, 48, 49, 57, 69, 98, 134 N. crawfordsvillensis 135, 136 Neopalseasteridse 52,53, 184 niagarensis, Palaeaster 68, 69 obtusus, Ophiura 25 obtusus, Palssaster 167 obtusus, Protaster (Ophiura) 25 obtusus, Stenaster (?) 165, 167 obtusus, Uraster 167 occurrence of starfishes 28 ocularplate 16,34,48,134,135 oculars in Echini 35 ontogeny of Stelleroidea 36 Onychaster 268 O.asper 270,272 O. barrisi 270, 272 O. confragosus 270, 273 O. demissus 270, 274 O.flexllis 269,270 Onychasteridse 268 Ophiocistia 275 Ophio-Encrinasterix .' 215 OpJiiopege 263 Page. Ophiura ciliaris 268 O. obtusus 25 O. (?)ramosa 22 O . rhenana 263 O . salteri 227 O. schlotheimii 13 Ophiurasterix 215,246 Ophiurella 263 0- primigenia 263 Ophiurina 246, 260 O. lymani 247 Ophiurinidse 246 Ophiuroidea 41, 42, 43, 2 1 2 , 258, 267 geological distribution of 28 number of, in Paleozoic 27, 28 oral, see actinal. oral angles 16 oral armature 16, 43, 162, 172, 210, 242 oral skeleton, see oral armature. ordinaria, Schuchertia 196, 199 ossicles, see plates. Palseaster 10, 40, 41, 44, 49, 56, 66, 67 , 75, 98, 104, 135, 158 Palseaster (Monaster) Etheridge (part) 72,170 P. (Argaster) antiqua 86 P. antiquata 109 P. aniiquus 86 P. asperrima 187 P. caractad 92 P. clarkana 61,63 P. clarke i De Koninck .- 172 P. clarkei Miller 61 P.darki : 94 P. coronella 167 P. crawfordsvillensis 136 P.dubius 85 P.dyeri 120 P. eucharis 98 P.exculptus 117 P. fin* 81 P. (Monaster) giganteus 73 P. grant i 89 P. granulosus Hall 112 P. granulosus Meek 109 P.Mrrisi 180 P.Mrudo 188 P. imbricatus 169 P. incomptus 61 P.jamesi 146 P. longibrachiatus 115 P. matutinus 57 P. miamiensis 127 P. montanus 189 P. niagarensis 68, 69 P . obtusus 167 P. parviusculus 87 P.pukJiellus 178 P. pygmaea. 23 P.ruthveni 187 P.sJiafferi - 77 P. simplex. 61, 62 P.speciosus 109 P. spinulosus - ^ P. squamatus 2 s P. (Monaster) stutchburii 73 P. wilberanus 84 308 INDEX. P. ? wilsoni 106 P.wykoffi 119 Palaeasteridse 52, 53, 66 Palxasterina 138, 150 P. antiqua 149 P. approximata 142 P. bonneyi. 153 P. fimbriata 204 P. primxva 153 P. speciosa 142 Palxasterinx 66 Palxasterinidx 138 Palxbrisingidx 199 Palxchinasteridx 199 Palxgoniasteridx (part) 66,241 Palxgoniasteridx 128 Palxnectria 155 P. devonica 155 Palxocoma Miller (part) 239 Palxocoma Salter 252 Palxocoma (Bdellacoma) 254 Palxocoma (Rhopalocoma) 254 P. colvin i 253 P. cygnipes 253 P. cylindrica 220 P. marstoni , 253 P. princeps 240 P. pyrotechnica 254 P. spinosa 219 P. vermiformis 254 Palaeodiscus ferox. . 23 Palseophiomyxa. 216, 234 P. grandis. 234, 235 Palxophiomyxidx 215, 234 Paiseophiura 235 P. simplex 235 Palxophiuridx 215,216, 235 Palaeosolaster 38, 40,207, 209, 211 P. gregoryi 209 P. (.)gyalum 210 P.rcemeri 211 Palseosolasteracea 163 Palseosolasteridse 162, 163, 20 7 Palxospondylidx 248 Palxospondylus 261 Palaeostella 51,138,155 P.solida 156 palaeozoicus, Furcaster 259, 261 Palxspondylus 261 P.zitteli 262 Palaeura 216,223 P.neglecta 223 Palasteracanthion 173 Palasterina Billings (part) 195 Palasterina Salter 40, 49, 130, 138, 140, 1 50, 154, 156, 196 P. bonneyi 151,152,153 P.follmanni. , 157 P.(f)jamesi. 146 P. kinahani. 155 P. primaeva 149, 151, 152, 153 P. (?) ramseyensis 152, 154 P. rigidus 141 P. rugosa 64 P. stellata 196 Palasterinacea. . . 53 Page. Palasterinidae 52,53, 138, 196 Palasteriscidae 162, 163, 199 Palasteriscus 40, 199, 200, 252 P. devonicus 200 Palastropecten 248, 261 P. zitteli 262 Palmipes antiqua 149 papulae 16 parviusculus, Mesopalseaster (?) 75, 87 parviusculus , Palxaster 87 parvulus, Trimeraster 134 paxillae 116,156 peculiaris, Cholaster 266 pedicellariae 16 perfectus, Siluraster 66 perradial 17 petaliformis, Jaekelaster 192 petaloides, A spidosoma 243 petaloides, Encrinaster 243 petaloides goslariensis, A spidosoma 243 petaloides goslariensis, Encrinaster 243 Petraster 39, 44,45,46,47,51,138,149,154,156,196 P. (?) americanus 139,146 P. (?) antiqua 86 P. bellulus 91 P.dyeri 121 P. rigidus (part) 57 P.rigidus 139,141 P. smythi 139,147 P. speciosus 47, 139, 142 P. wilberianus 84 petri, Protaster 25 Phanerozonia 80, 38,39, 40, 41, 43, 44, 45, 46, 48, 51 phylembryo of Stelleroidea 34 Pisaster 38 plates 17 plumiformis, Trichotaster 26, 38 podia and podial openings 18, 17,33,38,39,209,213 pontis, Aspidosoma? 244 pontis, Encrinaster 244 prenuntius, Promopalseaster 104, 1 7 preservation of starfishes 28 primxva, Asterias 167 primxva, Palxasterina 153 primaeva, Palasterina 149, 151, 152, 153 primxvus, Uraster 153 primary inframarginal 15 primary radials 1 7, 32 primary skeleton 13, 15, 1 7 , 31, 32, 34, 56, 191 primary supramarginal 15 primigenia, Ophiurella 263 primigenius, Eospondylus 263 primitive starfish, most 31, 33 primordialis , Asterias 108 primus, Protasteracanthion 188 princeps, Palxocoma 240 princeps, Ptilonaster 240, 259 proavitus, Mesopalaeaster 75, 83, 85 Promopalaeaster 28, 39,41,42,44,45,46,47,50,56,76,98, 102, 126 P. bellulus 104,105,118 P.dyeri 104,105,120 P.exculptus 104,105,117 INDEX. 309 Page. P. granulosus 104, 112,113 P. magnificus 42, 104, 105, 106, 113, 122 P. prenuntius 104, 107 P. speciosus 104, 105, 109, 113 P. sp. undet 104, 108, 109 P, spinulosus 44, 104, 105, 116 P. wilsoni 104, 106 P. wykoffi 42, 104, 119 Promopalaeasteracea 53 Promopalsesteridse 52, 53, 73 Promopalaeasterinse 52, 74, 102 Protaster Hall (part) 254 Protaster Forbes 216,217, 224, 229 P. (?) barrisi 272 P. biforis 226, 722 P. brisingoides 236 P. daoulasensis 261 P.decheni 248 P.elegans * 221 P. fimbriata 231 P. flexuosus 231 P.forbesi 255 P. (?) granuliferus 230 P. (?) gregarius '264 P.groomi 226,227 P. leptosoma 237 P. miamiensis 233 P. miltoni 251 P. (Ophiura) obtusus 25 P.petri 25 P. (?) salted 226,227 P.sedgwicki 225,226,243 P. (?) steUifer 226, 228 P. (?) whiteavesianus 226, 227 Protasteracanfhion 173, 178 P. primus 188 Protasteridae 215 Protasterina 228 P. fimbriata 231 P.flexuosa 231 Protoeuryale 25 Protopalxaster 53, 56 P. caractad 92 P. narrawayi 59 Protophiurex 246 Protophiuroidea 213 proximal inframarginals 14, 16 proximal radials 17 proximal supramarginal 15 Pseudopalasterina 138, 156 P.foUmanni 156,157 Ptilonaster 235,289 P. princeps 240, 259 pulchella, Urasterella. 175, 178 pulchellus, Palxaster 178 pulchellus, Stenaster 178 Pycnopodia 38, 42, 207, 208 pygmaea, Palaeaster 23 pygmaeus, Ataxaster 161 pyrotechnica, Palxocoma 257 pyrotechnica, Rhopalocoma 254 radial accessory plates 13 radialia 17,34,35,46 radicle of asterids 34 ramosa, Ophiura (?) 22 s, Palasterina (?) 152,154 Page. ray ambitals 13 rays 17,38,44,48,207,213,223 regeneration of. 185 regeneration 87, 185 respiration, see papulae and csecal pores. rhenana, Asterias (Archxasterias) 131 rhenana, Ophiura 263 rhenanus, Helianthaster 159 rhenanus, Medusaster 212 rhenanus, Miospondylus 263 rhenanus, Xenaster (?) 130, 1 8 1 Rhenaster 43, 128, 188 R. schwerdi 133 Rhodostoma 236 R. leptosoma 237 Rhopalocoma 248,252, 254 R. pyrotechnica 254 rigidus, Palasterina 141 rigidus, Petraster 57 rigidus, Petraster 139, 141 Roemeraster 173, 177 R. (?) acuminatus 93 R. asperula 188 Roemerasterinx 172 roemeri, Aspidosoma 244 roemeri, Encrinaster 242,244 roer eri, Helianthaster 209, 211 roemeri, Palseosolaster 211 rugosa, Palasterina 64 rugosus, Hudsonaster 36,55, 64 ruthveni, Palxaster 187 ruthveni, Uraster. 187 ruthveni, Urasterella 174, 175, 18 7 SaUeraster 173,178 S. asperrimus 187 salted, Ophiura 227 salted, Protaster (?) 226,227 salted, Stenaster 164, 166 schlotheimii, Acroura (Ophiura) 18 schlotheimii, Ophiura 18 schluteri, Astropecten (?.) 161 schmidti, A spidosoma 244 schmidti, Encrinaster 244 Schcenaster 202 S.fimbriatus 204 S. (?) legrandensis 206 S. (?) montanus 43,207 S. (?)n.sp 206 S. ( ?) wachsmuthi 205 Schcenasteridse 162,163,202 schohariae, Taeniaster 219,220 Schondorf, on revision of German starfishes. . 11 Schuchertia. 51, 140, 152, 194, 195, 252 S. laxata 196, 198 S.ordinaria 196,199 S.stellata 195,196 Schuchertiacea 163 Schuchertiidae 162,163,194 schwerdi, Rhenaster 133 secondary skeleton, see secondary plates. sedgwicki, Protaster 225, 226, 243 selwyni, Urasterella 175,188 shafferi, Mesopalajaster 44, 75, 77 shafferi, Palxaster 77 Siluraster 49,53 65 S.perfectus 66 310 INDEX. simplex, Eoactis . -. 186 simplex, Palxaster 61, 62 simplex, Palaeophiura 235 simplex, Xenaster 69 skeleton, secondary, see accessory plates. sladenl, Echinasterella 200 smythi, Petraster 139, 147 Solaster 38 solida , Palseostella 156 sollasi, Lapworthura 251 Spanlaster 39,43,47,49,50,74,95 S. latiscutatus 96 speciosa, Palxasterina 142 speciosus Palxaster 109 speciosus, Petraster 47, 139, 142 speciosus, Promopalaeaster 104, 105, 109,113 spines, see paxillae and pedicellarise. spinosa, Palxocoma 219 spinosissima, Asterias 19 spinosity 37,41 spinosus, Echinasterias 211 spinosus, Tgeniaster 216, 217,218, 219 spinulosus, Palxaster 115 spinulosus, Promopalaeaster 44, 104, 105, 115 spryi, Gregoriura 234 Squamaster 248,249 S. echinatus 249 squamatus, Palaeaster 23 stalk of primitive starfishes 83, 34 starfish, most primitive 31 , 33 origin of graspers 33 origin of wriggling type 32 radicle 34 stellata, Palasterina 196 stellata, Schuchertia 195, 1 96 stellata, Trentonaster 196 SteUeroidea geological development 27,28 geological distribution 28 ontogeny 36 phylembryo 34 typembryo 8 5, 36 stellifer, Protaster (?) 226, 228 Stenaster (part) 173 Stenaster 31,41,42,43,50,68,163,168,212 S. (?) confluens 165,167 S. (?) coronella 165,167 S.grandis 180 S.huxleyi 182 S. (?) obtusus 165, 167 S.pulcJiellus 178 S.salteri 164,165 Stenasteracea 163 Stenasteridae 162, 168 stone-canal, see madreporite. Strep tophiurse 215, 249 Stiirtzaster 248,252 S.colvini 253 S.cygnipes 253 S. marstoni 253 S. (?) mitchelli 254 Stiirtzura 235,286 S. brisingoides 236 S.leptosoma 236,237 S. leptosomoides 236,237 Page. stutchburii, Australaster (?) 73 stutchburii, Monaster 73 stutcJiburi i, Palaeaster ( Monaster} 73 supramarginalia 15, 34,35, 46 Sympterura ; 248, 256 S. minveri 256 syngnaths 218,222,224,242,256,267 Tseniaster 216,218 T. australis 236 T. cylindricus 216,217,218,219,220 T.elegans 219,221 T. meafordensis 221 T. schohariae 219, 220 T. spinosus 216,217,218, 219 Txniasteridx 172 Txniura 216 T.cylindrica 220 tenuiradiatus, Coelaster 19 terminals, see ocular plates. terminology of Asteroidea 13 Tetraster 31,41,42,43,50,164,165,167,212 T. asperrimus 187 T. (?) imbricatus 168,169 T . sp . ind . Nicholson and E theridge 186 T. wyville-thomsoni (part) 65 T. wyville-thomsoni 168 t ischbeinianum, Aspidosoma 244 tischbeinianus, Encrinaster 244 torus 1 6, 43, 54, 174 traquairi, Echinostella 212 Tremataster 246, 247 T. (?) decheni 247, 248 T. difficilis 247 Trentonaster 195 T. stellata 196 Trichotaster plumiformis 26, 38 Trimeraster 44,45,50, 128, 184 T. parvulus 134 trivium 14 Trochitaster, see Trichotaster. typembryo of Stelleroidea 85, 36 ulrichi, Urasterella 37, 43, 45, 49, 175, 183 Uranaster 61,138,154 U. elizse 155 U. kinahani 154, 155 Uraster hirudo 188 U. obtusus 167 U. primxvus 153 U.ruthveni 187 Urasterella Sturtz 163 Urasterella McCoy 40, 41,42,45,47,49,50,69,164,178,194,212 U. (?)asperrima 175,187 U. asperula 175, 188 U. (?) constellata 175, 187 U. girvanensis 167, 175, 186 U. grandis 174,175,180 U. hirudo 175,188 U.huxleyi 175,182 U. montana 175,189 U. n. sp 189 U.pulchella 175,178 U . ruthveni 174, 175, 1 8 7 U.selwynl 175,188 U. ulrichi 37,43,45,49,175,188 INDEX. 811 Page. Urasterellacea 163 Urasterellidse 162,163,172 ventral, see actinal. ventral shield 213 vermiformis, Bdellacoma 254 vermiformis, Palxocoma 254 vertebral ossicles 18,213,257,267,269 wachsmuthi, Schoenaster (?) 205 whiteavesianus, Protaster (?) 226, 22 7 wilberanus, Mesopalseaster (?) 75, 84 wilberanus, Palseaster 84 Wilberianus, Petraster 84 wilsoni, Palseaster (?) 106 wilsoni, Promopalseaster 104, 106 woodward!, Eucladla 276, 27 7 wrigglers 32,33,38 Page. wykoffi, Palseaster 119 wykoffi, Promopalseaster 42, 104, 1 1 9 wyville-thomsoni, Tetraster 65 wyville-thomsoni, Tetraster 168 Xenaster 40,44,45,46,47,49,50,98,128,129 X. dispar 130,181 X.elegans 130,181 X. eucharis 98 X. margaritatus Follmann 131 X. margaritatus Simonovitsch (part) 131, 132 X. margaritatus Simonovitsch 129,130, 181 X. (?)rhenanus 130,181 X. simplex 96 Xenasteridse 52,53, 128 zitteli, Palseos'pondylus 262 zitteli, Palastropecten 262 O 14 DAY USE RETURN TO DESK FROM WHICH BORROWED EARTH SCIENCES LIBRARY This book is due on the last date stamped below, or on the date to which renewed. Renewed books are subject to immediate recall. LD 21-50m-4,'63 (D6471slO)476 General Library University of California Berkeley