The Coal Measures Amphibia of North America

The Carnegie Institution of Washington has already published several monographs upon paleobiological subjects, written by its research associates, Hay, Wieland, and Case. Each author has dealt with the subject-matter of his particular field, but each has brought to bear upon his work common factors which have placed his labors upon a broader basis than the mere morphological descriptions of fossil forms of life. Case has published four monographs upon the morphology and taxonomy of the Permo-Carboniferous vertebrates of North America, and has followed these by a fifth, in which all the known factors bearing upon the development of the life were assembled in an effort to discuss the paleogeography of the period. In his conception paleogeography is a very broad term, involving not only a study of the distribution of land, water, and life in any one interval of time, but a consideration of all the factors in the extremely complex inter-relations of organic and inorganic matter and causes which influence the development of each part.

Geologists and paleobiologists have alike suffered in their interpretation of past conditions, because of their lack of knowledge of the work done by others. Stratigraphy may not be interpreted from the preserved fossils without a knowledge of biological laws, and the formations of the earth may not safely be rearranged to account for the present or past distribution of life without a knowledge of geological processes.

It is obvious that such work is beyond the possibilities of any one man; it is rather the work of a group of men, each broadly trained and each master of his own field and able to contribute to and criticize the work of his fellows. Nowhere could close cooperation of this kind be better accomplished than under a system such as the Carnegie Institution of Washington has developed, whereby the research associates of the Institution and others of its staff may call in the assistance of men in related fields. Already the value of this procedure is apparent in the results accomplished by cooperation.

The following monograph, by Dr. Roy L. Moodie, adds an important link to the series of paleobiological publications of the Institution and is closely connected with the work already done upon the Permo-Carboniferous vertebrates, since it supplies a description of the life of the period immediately preceding. It is hoped that the volume will contribute in no small measure to an understanding of the broader problems of paleogeography and the recognition of the mutual problems of the paleobiologists and the geologists.

E. C. Case.

University of Michigan, March 15, 1916.

The question of the origin of land vertebrates, which has appealed so strongly to students of fossil Amphibia, is by no means solved from the material furnished by the Coal Measures of North America. The Amphibia are, however, well known from several localities in the Coal Measures of this continent, where skeletons have been recovered which are sufficiently well preserved to afford a fair knowledge of their anatomy. The specimens rescued from the dumps of the old mines are regrettably few in comparison with the number that must have been burned as fuel, or carried down the slopes as silt. Yet scanty as is the material thus collected, it is of great importance, because it represents such an early period in the recorded history of the air-breathing vertebrates.

The amphibian fauna in the Coal Measures of North America is represented by several hundred individual specimens, preserved in various museums. All of the collections have been available in the preparation of this memoir, with the exception of those species from Nova Scotia which are preserved in the Peter Redpath Museum of McGill University and in the British Museum of Natural History. The European material, which has been used in comparisons with the American forms, has been studied chiefly from the literature, although there have been available a series of specimens of Branchiosaurus amblystomus Credner, from Saxony, presented by the late Professor Credner, and a single specimen of Archegosaurus from Dr. von Huene, of Tübingen.

The collection which has been of the greatest value is that at the American Museum of Natural History, chiefly assembled by Dr. J. S. Newberry from the dumps of the coal mines at Linton, Ohio, while he was in charge of the Ohio Geological Survey (1869-1884). This collection, a part of which is at Columbia University, furnished Cope with the most of his type material for the "Synopsis of the Extinct Batrachia from the Coal Measures" (123).[A] This entire collection, including all of Professor Cope's types and representing many new and hitherto undescribed forms, was generously placed at the writer's disposal for a period of five years through the kindness of Dr. Bashford Dean and Dr. Louis Hussakof. Dr. Hussakof made a trip through the Linton region and his description of the place occupied by the "Old Diamond Mine" is given on page 16.

An interesting collection of air-breathing vertebrates from the Coal Measures, representing 19 species, is in the U. S. National Museum (464). This is chiefly the collection of Mr. R. D. Lacoe and includes specimens from Mazon Creek, Illinois, from Kansas, and from Linton, Ohio. It is especially important in that it contains the skeleton (plate 20, fig. 3) of the oldest known reptile, Eosaurus copei Williston (Jour. Geol., XVI, 295). It contains also, besides many of Cope's types, new forms which have been described by the writer (464, 470, 471, 472, 473, 474, 478, 479). Dr. Stuart Weller first secured the use of this collection for me, and its continued use has been granted by Dr. C. D. Walcott. Mr. Charles Gilmore has called my attention to several interesting specimens and has kindly loaned them for description.

A small but interesting collection of Mazon Creek Amphibia is that of the Peabody Museum of Yale University. Through the courtesy of the officers of this museum the writer was permitted to study these specimens and was given a grant for their illustration. The results of that study are contained in a previous paper (478) and in the present memoir. Dr. Schuchert has offered suggestions as to the environmental conditions of the ancient Amphibia.

A few specimens of Coal Measures Amphibia are at the Walker Museum, University of Chicago. This collection includes the type of Micrerpeton caudatum Moodie, the first branchiosaur discovered in the western hemisphere, and a few specimens from Linton, Ohio.

A single specimen of Amphibamus grandiceps Cope, very beautifully preserved, is in the possession of Mr. L. E. Daniels, of Rolling Prairie, Indiana. This specimen has been studied and described by Hay (316) and by the writer (462, 469, 478).

The works of Cope and Dawson, published between 1860 and 1897, on the Amphibia from the Coal Measures, have been indispensable in the present study. It has been necessary to rely on the published descriptions and photographs of the interesting fauna from Nova Scotia, since it has not been possible for me to visit and examine the types preserved in the Peter Redpath Museum of McGill University and in the British Museum of Natural History. It has been possible to check Dawson's work, to a certain extent, by a study of a series of excellent photographs of the types of Coal Measures Amphibia collected by Dawson and Lyell and described by Dawson and Owen. The descriptions of these authors have been drawn on for the discussion of the Canadian forms.

The descriptions given below have been made full and complete in the belief that in this way our knowledge of these interesting vertebrates may be advanced. Many of the species have been described elsewhere in scattered papers by various authors. These descriptions have been revised and verified and are collected here in monographic form. The work is a morphologic and taxonomic revision of the Amphibia from the Coal Measures of North America. Especial attention has been paid to the factors which have been most active in the evolution of the group, so far as these factors may be interpreted. It is the author's hope that this review may open up the field for many more workers, since we are just beginning to learn about the evolution of this group of vertebrates.

The trustees of the Elizabeth Thompson Science Fund allotted a grant for the present investigation. This aid has enabled the writer to present his work in much better form than would have been possible otherwise. Dr. S. W. Williston has offered many suggestions and criticisms which have been gratefully adopted. It is with the greatest sense of pleasure that the author dedicates this memoir to his teacher and friend. After the manuscript was completed the author enjoyed a visit from Mr. D. M. S. Watson, of King's College, London, whose knowledge of the European and African forms enabled him to offer several very valuable suggestions.

It is fitting also to express my indebtedness to the Carnegie Institution of Washington for the privilege of publishing my work in the series of monographs contributed by Dr. E. C. Case, dealing with the anatomy and relationships of the early land vertebrates of North America.

Roy Lee Moodie.

PLATES.

TEXT-FIGURES.

THE PROBLEM OF THE AMPHIBIA FROM THE COAL MEASURES.

The Amphibia from the Coal Measures of North America present the problem of the origin of the land vertebrates, since the air-breathing vertebrates in the Coal Measures of this continent are the earliest known in the western hemisphere. The difference in age between the chief amphibian-bearing deposits of North America and Europe is not great, although it has been asserted that Pholidogaster and its allied fauna, described by Huxley from Scotland (331), is much older, probably Mississippian. It is interesting to note that these earliest representatives of the Amphibia in Scotland are all temnospondyles, of which there are very few representatives in the Coal Measures of North America.

The forms so far described from the North American Coal Measures present a very high degree of development and differentiation, the earliest known species being already specialized and well adapted for various modes of life. As far back in geological time as the middle Coal Measures, when the first well-defined forms are known, environmental conditions had effected a wide diversity of structure within the group. Thus, early in the geological history of the land vertebrates, we have, among the Coal Measures Amphibia, various forms which had specialized into strictly aquatic, terrestrial, subterrestrial, and arboreal, or at least partly arboreal. Specialization had extended to the loss of limbs, ribs, and ventral armature in a few species, and to the acquirement of claws, running legs, or a long propelling tail with expanded neural and hæmal arches in others. The forms range in size from small creatures less than an inch in length to large species which must have attained a length of several feet. A rather interesting parallel, though of no phylogenetic significance, can be drawn between the Amphibia of the North American Coal Measures and the reptiles of to-day. The snakes are represented by the limbless, snake-like forms, such as Ptyonius and Phlegethontia. The lizards find their counterpart in the Hylonomidæ and the Tuditanidæ. No known characters of these animals tend to ally them directly with any known group of fishes, except in the most general way. These facts all indicate a long antecedent history for the amphibian group or else a preceding period of greatly accelerated development of which we now know nothing.

The Amphibia whose remains have been brought to light from the Coal Measures have hitherto been regarded as pertaining to a single order, the Stegocephalia, characterized by the completely roofed-over cranium and a large parasphenoid. The writer (469) had previously assigned 5 suborders to the group: the Branchiosauria, Microsauria, Aistopoda, Temnospondylia, and Stereospondylia. All of these groups are represented in the Coal Measures of North America. It has seemed inadvisable, in the light of our present knowledge of the Amphibia, to retain these 5 groups as suborders, and, in the revised scheme of classification which has been published elsewhere (469), they are given the rank of orders all excepting the Aistopoda, which are now regarded by the writer as specialized Microsauria.

The recent Caudata are possibly represented in the North American Coal Measures by forms which may be assigned tentatively to the Proteida. Such forms as Cocytinus gyrinoides, Hyphasma lævis, and Erierpeton branchialis possibly represent this group in the Pennsylvanian. This relationship is based chiefly on the structure of the hyobranchial apparatus and on the general structure of the species. The three above-mentioned species are, however, very insufficiently known, and the relationship can hardly be regarded as more than suggested by the characters which are at hand.

The Salientia, or frogs, may possibly have their ancestral type in Pelion lyelli, the first known species from the Linton, Ohio, Coal Measures. Oddly enough, among the hundreds of specimens collected later from this horizon, not a fragment can be identified with this species. The type specimen is unique, and although incomplete its characters are suggestive.

The Branchiosauria are represented in North America by four species: Micrerpeton caudatum, Eumicrerpeton parvum, Mazonerpeton longicaudatum, and M. costatum. Three other genera which occur in North America have been placed (642) in this group, but they do not belong there, for reasons given below. The branchiosaurs were salamander-like in appearance. They were naked, with the exception of small ovoid scales on the back and the chevron-shaped armature of the ventral surface, the latter being almost universally present among the Paleozoic Amphibia. They were adapted for life in the water for at least the early part of their existence, as is shown by the possession of gills on many of the late Carboniferous and early Permian forms of Europe. The group is, without doubt, ancestral to the modern Caudata. No branchiosaurians have been described elsewhere from so low in the geological series as those here given and they are the first and only evidence of the occurrence of the group in the western hemisphere.

The Microsauria are represented in the Coal Measures by numerous forms which are usually characterized as lizard-like animals with a well-developed ventral scutellation. Other characters, such as the possession of lateral-line grooves on the cranium, the arrangement of the cranial elements, and the condition of the ribs, will be discussed further on. The pectoral arch is well developed and is composed of five dermal bones plus the regular skeletal elements. The skeletal membrane bones are sculptured after the manner of those of the cranium. The bodies of the animals were, in a few cases, covered with scales; but most of them appear to have been completely naked, even the ventral armature being absent in some cases. The ventral scutellation was especially strong and highly developed in some of the forms; e.g., in the genera Saurerpeton and Sauropleura. The vertebræ are uniformly of the hour-glass or notochordal type. This is so generally the case that the characters of the vertebræ and ribs are taken as the chief diagnostic characters of the major groups. Various peculiarities are seen among the Microsauria, such as the development of horns in various genera which are, apparently, related. The order seems to have gone completely out of existence during the early Permian, and if their descendants continued on as reptiles, as has been suggested (469), we do not know the intermediate stages.

The Aistopoda are without doubt specialized microsaurs, and, in the opinion of the writer, are not entitled to separate rank. Some of these forms reached a high degree of specialization. One American species has the skeleton reduced to a long, slender head and a slender series of elongate vertebræ, all other parts of the skeleton, even the ventral armature, being absent. The proportions attained by this species, Phlegethontia linearis Cope, recall those of the coach-whip snake, Zamenis flagellum Shaw, of the western plains. Some of the so-called Aistopoda have been credited by Fritsch with the possession of peculiar clasping organs, "Kammplatten." Newberry has written of the discovery of similar structures in the Ohio Coal Measures (498), but the statement of the actual association of these "Kammplatten" needs confirmation. Dr. R. H. Traquair wrote to the author under date of April 28, 1909:

"I maintain that the association of a bundle of 'Kammplatten' with a specimen of Ophiderpeton in the Bohemian gas coal was entirely accidental. Of such pitfalls the paleontologist has to beware or serious mistakes may be the consequence, as has happened more than once. I must, however, publish a short paper on the Kammplatten, for I think I know what they are now."

Fritsch, however, has very clearly figured a nearly complete specimen of Ophiderpeton (251, Bd. IV) as possessing the Kammplatten in place near the cloaca, where he suggests they may have served the function of accessory copulatory organs or claspers.

The Temnospondylia are represented by scanty remains of species from Illinois, Pennsylvania, and Nova Scotia. The forms belonging to this group are all relatively large, and they had a wide geographical distribution during the Permian. This group contains two types of vertebræ, known as the embolomerous and the rachitomous, both of which are present in the Coal Measures. Such forms as Eosaurus, Baphetes, Eobaphetes kansensis, Macrerpeton, and Dendrerpeton are regarded tentatively as temnospondyles, but there is no definite assurance that they are such. It is possible that Eosaurus is a stereospondyle, but the species is too incompletely known for a definite statement to be made. The close resemblance between the vertebræ of Eosaurus and Anthracosaurus has been noted by Huxley (332).

The Stereospondylia are very scantily represented in the Coal Measures, if at all. Eosaurus may belong here as indicated above. The tooth and cranial fragments discovered and described by Williston from the Coal Measures of Kansas may represent a stereospondyle as he states (608), but the evidence is incomplete. A fragment of a large rib (plate 22, fig. 4) of a species from Linton, Ohio, otherwise unknown, may be a stereospondyle. We would expect an early development for this group, but it is' an interesting fact that no stereospondyles are known definitely before the Triassic, during which period they had an extensive distribution.

HISTORY OF THE DISCOVERY OF AMPHIBIA IN THE COAL MEASURES.

Sir William Logan, in 1841, discovered in the Coal Measures of Horton's Bluff, Nova Scotia, some tracks of Amphibia which he carried to London and which Sir Richard Owen pronounced to be undoubted "reptilian" tracks. This fact was published in 1842 (380) and was the first recorded evidence of the occurrence of land vertebrates in the Carboniferous rocks of the world. To these tracks Sir William Dawson later gave the name of Hylopus logani.

Two years later Dr. Gergens (291) wrote a letter to Professor Bronn, the founder and one of the editors of the "Neues Jahrbuch für Mineralogie, Geologic und Paleontologie," in regard to an important discovery in the Carboniferous rocks of Germany. The letter is of such exceptional interest in connection with the history of the fossil Amphibia that it is given here:

The form discovered by Dr. Gergens and described by Hermann von Meyer as an amphibian is a little puzzling as to its characters. Miall (449, p. 183) says that the remains are too imperfect for close definition. The form, as figured, resembles an immature branchiosaurian, as one is at once reminded, from an examination of Von Ammon's Branchiosaurus caducus (7, Taf. IV, fig. 1). In 1844 Dr. Alfred King (356) announced the discovery of "reptilian" footprints in the Carboniferous of Pennsylvania.

The next announcement of fossil Amphibia was made by Goldfuss (296), who in 1847 described the famous Archegosaurus from the upper Carboniferous of Germany, from the remains which had as long ago as 1777 been regarded as a fish. Two years later Isaac Lea (371) announced to the British Association for the Advancement of Science, through Buckland, the discovery of footprints in the old Red Sandstone (Mauch Chunk) of Pennsylvania. These objects occur not rarely in the Mauch Chunk shales, which are of upper Mississippian age. Barrell (21, p. 460) records the finding of imperfect tracks in the same beds, and Rogers (Geology of Pennsylvania, pt. II, 1856, p. 831) records three unnamed varieties from 2,200 feet below the top of the Mauch Chunk. Branson (50) has recorded the finding of other amphibian footprints from the Mississippian of Giles County, Virginia.

Lyell and Dawson (396), in 1853, read a paper before the Geological Society of London, in which they announced the discovery of remains of Amphibia in the Coal Measures of North America, although Dawson had previously, in 1850, discovered the skull of Baphetes planiceps Owen, which was not described until the latter part of 1853 (509). The specimen had lain unnoticed in the collection of the Geological Society for more than two years. When, however, the announcement was made by Lyell and Dawson of the discovery of Amphibia in the Coal Measures of Nova Scotia, so much interest was excited that the skull, now known as Baphetes planiceps, was brought to light by the president or secretary and was described (509) by Sir Richard Owen. The only other known evidences of land vertebrates in the Paleozoic of North America, up to this time, had been the footprints described by Lea and King from the Mississippian (Mauch Chunk) and Pennsylvanian of Pennsylvania. The specimens presented to the Geological Society of London by Lyell and Dawson were found at the South Joggins, Nova Scotia, and consisted of scutes, a few limb bones, a fragment of a jaw, and a few vertebræ, a part of which were associated. The remains were found quite accidentally and unexpectedly by them in the petrified trunks of ancient Sigillariæ which were exposed on the coast. Dr. Jeffries Wyman, of Harvard College, had examined these remains in the United States and had pronounced (638) them to be amphibian, comparing them with similar elements in Menobranchus. On the arrival of the specimens in England they were submitted to Sir Richard Owen, who suggested the name (514) Dendrerpeton acadianum and compared the remains with Archegosaurus. At the same meeting of the London Geological Society, Owen read a paper on a small amphibian (508) from the British Carboniferous which he named Parabatrachus. Subsequent discoveries have shown, however, that this form belongs among the fishes. At the meeting of the Geological Society held in the latter part of the same year Owen announced (509) further discoveries in the Nova Scotia coal beds.

Hermann von Meyer (436), in 1857, described numerous stegocephalian remains from the upper Carboniferous of Germany. Dr. Jeffries Wyman, in the same year, described (639) a new form of amphibian from Linton, Ohio. This form he called Raniceps lyelli, but as the name Raniceps had been preoccupied by Cuvier for a genus of gadid fishes, Wyman later (1868) changed the name to Pelion. This was the first form to be described from the locality at Linton, which has since yielded the remains of half a hundred species.

Dawson (204), in 1859, made a further contribution to the fauna of Nova Scotia by the description of Hylonomus and other species of Dendrerpeton from the South Joggins deposits. Huxley (331), in 1862, described the genera Loxomma and Pholidogaster from the Carboniferous of Scotland. The same year Owen made a further contribution (514) to the fauna of the Nova Scotia beds, and Huxley (332) discussed the anatomy of Anthracosaurus from Scotland. Marsh (404), in the next year, described, as an enaliosaurian, the interesting Eosaurus acadianus from the Nova Scotia Coal Measures, basing the species on two vertebræ, apparently from the dorsal region. The vertebræ resemble the stereospondylous type, and Huxley (332) called attention to the similarity of these vertebræ to those of Anthracosaurus.

Cope (105), in 1865, began his researches among the Coal Measures Amphibia of North America by the description of Amphibamus grandiceps from the Mazon Creek shales of Illinois. Ten years later (123) he published a complete synopsis of the Carboniferous Amphibia of North America, with especial reference to the Linton, Ohio, species, illustrating many of the forms now known from Linton. Between the years 1865 and 1897, Cope published numerous papers (105-177) on the Amphibia of the Paleozoic, and to his researches is due a large part of our knowledge of these forms.

Great credit is due Dr. J. S. Newberry (495, 498) for the enthusiasm and interest which his collections of Coal Measures Amphibia exhibit. He furnished Cope with the majority of the type material described by him, and it was through Dr. Newberry's instrumentality that the "Synopsis of the Extinct Batrachia from the Coal Measures" (123) was published. The material which Dr. Newberry had collected he took with him from Ohio to Columbia University, New York, and a part of his collection still remains in the geological collection of that institution, although the greater portion has been transferred to the American Museum of Natural History. The Newberry collection forms the basis for the larger part of this memoir.

Between the year 1853 and the early nineties, Dawson continued (200-223) his researches on the Amphibia of the Coal Measures of Nova Scotia. His most notable single work (208) is "The Air-Breathers of the Coal Period," published in Montreal in 1863, in which he gives a complete account of the forms then known from Canada, attempting some restorations. Since his death there have been no new species described from Canada, and, so far as I can learn, there has been no further collecting at the South Joggins.

Recently G. F. Matthew (409) has rearranged the classification of amphibian footprints from Nova Scotia. Jaekel (347) has described very fully the remains of Diceratosaurus punctolineatus (Cope) from Linton, Ohio, basing the new genus on a species described by Cope as a member of Ceraterpeton. Hay (316) has added to the knowledge of the anatomy of Amphibamus, his most interesting contribution being the detection of long, curved ribs in this form. This character excludes the species from the order Branchiosauria and shows the relationship of the form to the Hylonomidæ and the Microsauria. Schwarz (540) has described the characters of the vertebræ and ribs of several genera of the Coal Measures Amphibia and has (541) offered his views as to the descent of the Amphibia, based entirely on his work on the vertebræ of species from North America and Europe.

Since 1908 the writer has published several contributions (457-489) on the Amphibia from the Coal Measures of North America. The results of these investigations are given in this work.

STRATIGRAPHIC AND GEOGRAPHIC DISTRIBUTION OF AMPHIBIA IN THE COAL MEASURES OF NORTH AMERICA.

There are but four localities in North America which have furnished any notable remains of Amphibia in the Coal Measures. These are, in the order of their discovery, the deposits at the South Joggins, Nova Scotia; the Linton, Ohio, Coal Measures; the Mazon Creek, Illinois, shales; and the Cannelton slates near Cannelton, Pennsylvania. There are, however, several other localities on the continent which have furnished evidences of Amphibia in the Coal Measures. The principal one of the latter localities is doubtfully of Coal Measures age, although recent discoveries would tend to show it is such. The deposits in question, those of the Clepsydrops shales of Vermilion County, Illinois, have, heretofore, been regarded as Permian, but the discovery of similar remains in rocks of undoubted Pennsylvanian age in Pennsylvania would seem to indicate that the Illinois deposits were contemporaneous with them.

Fig. 1. Map of Upper Pennsylvanian showing land and water conditions under which the Coal Measures amphibian fauna lived. It will be noted that the chief deposits which have furnished amphibian remains are on the margins of the heavily shaded areas. (After Schuchert.)

Explanation of symbols: Lands are white. Water areas are lined. Formation outcrops are black or dotted. Known shore-lines are solid lines; probable ones broken. Vertical lines in middle of continent indicate Gulf marine.

(a) The deposits in Vermilion County, Illinois, lie along the north bank of Salt Fork Creek, at the tip of the "Horseshoe Bend," about 2 miles south of Oakwood, Illinois. They were discovered by Dr. J. C. Winslow, of Danville, in 1875. The remains discovered by him were forwarded to Professor Cope for identification. Later the deposits were thoroughly explored by W. F. E. Gurley, and the specimens collected by him are now preserved (86) in Walker Museum, University of Chicago. In 1907, the writer, while working for the University of Chicago, in exploring the same locality, exhausted the beds so far as they could at that time be uncovered from the landslide which had overwhelmed them. The formation in which the bones occur is a soft gray or reddish shale, and it lies without any apparent stratigraphic break on shales of Pennsylvanian age. Below these shales are several feet of limestone containing invertebrates of typical Pennsylvanian facies. There are indications of at least 3 species of Amphibia in the deposits. Case (86) has indicated with doubt a fourth species. The species are: Cricotus heteroclitus Cope, Cricotus gibsonii Cope, Diplocaulus salamandroides Cope. The remains are very fragmentary, and consist for the most part of incomplete vertebræ, with a few small skull fragments.

(b) In 1897 Dr. Williston (607) described some fragments of Cricotus from a deposit in Cowley County, near Winfield, Kansas. There has been some dispute as to the age of the deposit, but the consensus of opinion seems to be that the beds are of approximately the same age as those of Illinois and Pennsylvania in which similar remains are found, and those deposits are looked upon as Upper Pennsylvania (Case (94), pp. 239-240). No new forms were described from Winfield, since only a few fragments were obtained. Williston referred the phalange, the fragment of a jaw, and the tooth to Cricotus heteroclitus Cope.

(c) Later in the same year Williston (608) announced the discovery of a tooth of typical labyrinthodont structure from near Louisville, Kansas (plate 21, fig. 6). The tooth was accompanied by fragments of bone and was probably not far from the bed in which it was fossilized. Williston states that the remains were from the shales which are "nearly at the upper part of the Carboniferous, probably within one hundred feet of the Manhattan Limestone."

(d) In 1894 Marsh (406) and earlier (1873) Mudge (490) described footprints of vertebrates from the stone-quarries near Osage City, Kansas. The stone in which they were found was a fine-grained limestone which occurs near the middle of the Kansas Coal Measures.

(e) Two years later Marsh (407) announced the discovery of traces of the oldest known (Devonian) air-breathing vertebrate. The footprints of Thinopus antiquus were regarded by Marsh as "apparently amphibian." This still remains the oldest geological evidence of air-breathing vertebrates, although Lohest some years ago (381) called attention to remains from the Devonian of France which he thought might be amphibian. The footprint described by Professor Marsh was "found in the town of Pleasant, one mile south of the Allegheny River, Warren County, Pennsylvania, by Dr. Charles E. Beecher, who presented it to Yale Museum, and also furnished the information in regard to its geological position.... The geological horizon is near the top of the Chemung, in the upper Devonian. In the same beds are ripple marks, mud cracks, and impressions of rain drops, indicating shallow water and shore deposits."

(f) Among the collections of the American Museum there is an impression of a small amphibian foot obtained from Phoenix Tunnel, Pennsylvania. The impression is in hard black slate very similar to the slate of the Cannelton region. It is possible that the specimen may have been obtained from the Cannelton beds, since they would be expected to occur at Phoenix Tunnel. The impression is rather small. It is the footprint of a 5-toed animal, probably of the right foot, since no amphibian (465) so far is known from the Coal Measures with 5 digits on the hand. The first digit is short and thick, with a large ball at its base. The foot measures from the posterior edge of the palm to the tip of the longest digit 12 mm. The length of the first digit is 7 mm. The impression differs in some respects from the impressions so far known from the Coal Measures, but no attempt will be made to assign it to a species. It may have been made by either a branchiosaurian or a microsaurian, but more probably the latter, since we do not know of any of the former animals from the Cannelton beds, or in fact from any of the Coal Measures beds excepting the Mazon Creek shales. The specimen is No. 2872 of the American Museum.

Fig. 2. Distribution of Coal Measures Amphibia in North America.

1. Linton, Ohio, near Yellow Creek P. O., Jefferson County, Ohio, on the banks of Yellow Creek, near the Ohio River, 16 miles north of Steubenville.

2. Mazon Creek shales, Grundy County, Illinois, near Morris.

3. "Clepsydrops shales," Salt Fork Creek, Vermilion County, Illinois, near Oakwood, on Tate farm, 8 miles west of Danville, Illinois.

4. Danville, Illinois, coal where the type of Proterpeton gurleyi Moodie was found.

5. Breeze, Illinois, where Dr. J. A. Udden, in 1907, found a fragment of an amphibian phalange on the dump of the Cooperative Coal Company.

6. Pitcairn, Pennsylvania, 15 miles east of Pittsburgh.

7. Cannelton, Pennsylvania, Beaver County, Cannelton slates, Kittanning formation, 45 miles northwest of Pittsburgh.

8. Fairfield, Iowa, where Dr. J. A. Udden found remains attributed by Dr. Eastman to Pleuroptyx clavatus Cope.

9. Louisville, Pottawatomie County, Kansas, where Dr. S. W. Williston discovered remains of Mastodonsaurus in the Coal Measures.

10. Washington County, Kansas, source of type of Eobaphetes kansensis Moodie, from the Coal Measures.

11. Osage City, Osage County, Kansas, amphibian footprints from the Coal Measures.

12. Winfield, Kansas, source of Cricotus material.

13. Lander, Wyoming, in Wind River Carboniferous.

14. Pictou, Pictou County, Nova Scotia, 84 miles northeast of Halifax. Source of Baphetes planiceps Owen.

15. Joggins (Joggins Mines), Cumberland County, Nova Scotia, 4 miles from River Hebert. Source of Hylerpeton and Dendrerpeton faunas.

16. South Joggins, Nova Scotia, source of the Eosaurus acadianus Marsh.

(g) Dr. J. A. Udden, in 1907, discovered a fragment of a phalanx of some amphibian (plate 22, fig. 3) on the dump of the Cooperative Coal Company, a mile east of Breeze, Illinois. It was obtained from below the Shoal Creek limestone and somewhere above the (Illinois) Coal No. 6, according to Dr. Udden's notes. The maximum width of the phalanx is 10 mm. and it probably had a length of 16 mm.

(h) Mr. N. H. Brown, in 1914, discovered in the Carboniferous rocks to the east of the Wind River Mountains, near Lander, Wyoming, a single fragment of an amphibian. The writer was accompanying Mr. Brown at the time of the discovery and there can be no doubt that the fragment was amphibian; the location of the beds was such that no later age than the Coal Measures can be assigned to them.

(i) Dr. J. A. Udden (577), in 1912, announced the discovery of an amphibian in the Des Moines formation of Iowa. These remains were identified by Dr. Charles Eastman as Pleuroptyx clavatus Cope. Since the Des Moines is probably nearly contemporaneous with the Mazon Creek shales of Illinois, the discovery does not extend the geological range to any extent, but is of interest as it adds another note to our knowledge of the geographical distribution of the Amphibia in the Coal Measures.

(j) The Gurley collection of the University of Chicago possesses a single cervical vertebra of some amphibian (?). The vertebra is unlike anything previously described and represents a new form (plate 22, fig. 2) which may be designated Proterpeton gurleyi, new genus and species. The material was collected near Danville, Illinois.

(k) Deposits have been discovered in Pennsylvania in which are found the remains of amphibians and reptiles, very similar to those from Vermilion County, Illinois, Cowley County, Kansas, and the Texas Permian. The remains (plate 18, fig. 2) were found in a thin stratum below the "Ames" limestone, and are therefore in the Coal Measures, fairly well below the top. The fossils, as described by Case (94), consist of fragments which he ascribes to pelycosaurian reptiles and to temnospondylous amphibians. The genus Eryops (94) is recognized in several fragments and a nearly complete dorsal vertebral centrum. Other types of Amphibia are likewise represented.

(l) The ironstone nodules, in which the Mazon Creek fossils (plate 1) occur, are found in the shale which forms the roof of the Morris or "No. 2" Coal of Illinois, which "lies probably somewhat lower than the horizon of the Lower Kittanning Coal of Pennsylvania" (599). "The nodules of iron contained in the Coal shales on the banks of Mazon Creek near Morris, Illinois, generally contain organic nuclei, and thousands of beautiful specimens have been obtained there. They are usually fragments of fern fronds, but are sometimes shells, crustaceans, myriapods, scorpions, spiders, cockroaches, ... fishes" (498, p. 214), and amphibians, of which 10 species are at present known.

MOODIE

PLATE 1

VIEWS ALONG MAZON CREEK, ILLINOIS.

1. A nodule weathering out of the shale, at the head of the hammer. Most of the nodules at the so-called "lower beds" contain specimens of Neuropteris.

2. The nodules in the creek bed at the "upper beds." Many of them have been cracked open by the frost.

3. Looking south at the "upper beds." The nodules found in the background are non-fossiliferous.

4. Nodules may be seen through the clear water embedded in the shale. Neuropterid insects in the water.

5. Looking for nodules at the "upper beds." The uppermost reaches of the fossiliferous beds correspond with the extreme background of the picture.

6. Nodules in the stream bed at the "lower beds." Many of these are cracked open by the frost and good specimens are sometimes found in the nodules.

These species have been arranged zoologically according to the following plan:

Class Amphibia Linné, 1758.
Subclass Euamphibia Moodie, 1909.
Order Branchiosauria Lydekker, 1889.
Family Branchiosauridæ Fritsch, 1879.
Micrerpeton caudatum Moodie, 1909.
Eumicrerpeton parvum Moodie, 1910.
Mazonerpeton longicaudatum Moodie, 1912.
Mazonerpeton costatum Moodie, 1912.
Order Caudata Duméril, 1806.
Suborder Proteida Cope.
Family Cocytinidæ Moodie, 1912.
Erierpeton branchialis Moodie, 1912.
Subclass Lepospondylia Zittel, 1887.
Order Microsauria Dawson, 1863.
Family Amphibamidæ Cope, 1875.
Amphibamus grandiceps Cope, 1865.
Amphibamus thoracatus Moodie, 1911.
Cephalerpeton ventriarmatum Moodie, 1912.
Family Molgophidæ Cope, 1875.
Erpetobrachium mazonensis Moodie, 1912.
Subclass Stegocephala Cope, 1868.
Order Temnospondylia Zittel, 1887.
Suborder Embolomeri Cope, 1885.
Family Cricotidæ Cope, 1884.
Spondylerpeton spinatum Moodie, 1912.

It will be seen from the above arrangement that nearly all of the orders of Amphibia are represented in the Mazon Creek fauna. These animals are the oldest known land vertebrates of North America.

Fig. 3. Portion of the "Morris sheet" of the U. S. Geological Survey, to show topography and situations of the exposures of fossil-bearing shales at Mazon River, a, the "Bartlett place," the so-called "upper beds"; b, "lower beds."

The writer was able, during July 1911, to spend a week studying the fossil beds (479) at Mazon Creek. The object of the visit was primarily to collect Amphibia, but although several thousand nodules were examined, not one contained an amphibian nor a fragment of one. Mr. J. C. Carr, of Morris, Illinois, who has collected at Mazon Creek for more than 30 years, has never collected an amphibian. These facts interested me in making the following comparison: If we take 100,000 nodules as a basis for computation of the rarity of the various forms, something like the following will be the approximate result of the investigation:

Of 100,000 nodules, 20,000 will be barren or contain only indeterminate fragments; 68,500 will contain plants; 7,500 will contain insects, Crustacea, myriapods, scorpions, spiders, and other arthropods; 3,900 will contain fish coprolites or scales; 95 may contain fish or fragments of fish; 4 may contain mollusks; and 1 may contain an amphibian or a fragment of one.

Perhaps even 100,000 is low as a basis of estimate. Mr. Carr was of the opinion that 1 nodule in every 500,000 might contain an amphibian.

The beds from which the nodules are usually collected occur along both banks and in the bottom of the creek, in two localities. One locality known as the Bartlett place is situated 8 miles southeast of Morris, in Grundy County, Illinois, Wauponsee Township, N.W. quarter, section 30, Township 33, Range 8, the land being now owned by Mrs. Emma Akerly, of Wilmington, Illinois.

The fossil-bearing nodules occur throughout 6 to 8 feet of shale along both banks of the creek at the "upper beds" (plate 1, fig. 3), as the Bartlett place is called. They may also be seen in the bed of the creek, when the water is low (plate 1, fig. 4), still embedded in the shale. With a potato fork the shale is easily turned and the nodules come out like potatoes. One sometimes finds a "pocket" of nodules from which as many as a peck may be secured. Nearly every nodule has a fossil at the "upper beds," but all of the fossils are not well preserved, possibly only 1 or 2 out of every 10 being worth carrying to the museum. The nodules crack best when wet, and it requires some skill to crack them evenly. They seem quite light and, in one place where the stream curves, are piled in a long windrow. On this were found, in nodules cracked by the frost, several good crustaceans and many good plants.

Table of Pennsylvanic Formations.

Series Northern Appalachian. Bituminous.
Pennsylvania-Ohio.           Illinois.
Pennsylvanic
{
{  Coal Measures
{    {     {(*) (Salt Fork).
{    { Monongahela or Upper Productive Coal Measures {Break.
{    {
{    { Conemaugh or Lower Barren Ames limestone near middle:
{    {    (*) (Pitcairn)       Coal No. 6.
{    {
{    {             { Freeport
{    { Allegheny or Lower Coal Measures    {  (*) (Cannelton) Coal No. 2 (*).
{    {     { Kittanning (*).  (Morris?) (Mazon Creek).
{    {     {  (Linton).
{    {     { Clarion.
{
{  Pottsville
{    {
{    { Home wood.
{    { Mercer.
{    { Conoquenessing.
{    { Sharon.

(*) marks the position of the Amphibian-producing horizons in these regions. (After Schuchert.)

The fossils at the "upper beds" are localized into special strata. At one place in the upper part of the deposit, in a reddish shale, one finds that insects are more abundant than they are lower down. The Crustacea seem to come from apparently the same shale. At the lower end of the deposit certain definite species of Pecopteris are localized. It is an interesting fact that one seldom finds a Neuropteris at the "upper beds." The most abundant fossils are the various species of Pecopteris and Annularia. When specimens of Neuropteris are found they are usually discovered at the lower end of the exposures. In one place behind the "island" very blue nodules, hard and flinty and with sometimes well-preserved specimens of Pecopteris, are found quite definitely localized. These nodules are apt to assume an irregular shape. These localizations of the fossils are, of course, what we would expect from our knowledge of the recent fauna and flora. There is, to be sure, more or less intermingling of the species. The myriapods, so far as they have been found, are also localized. Mr. Carr found 3 within a space of a few feet, but again these are found widely scattered. The exposures at the "upper beds" are about a quarter of a mile long. They disappear under a heavy ledge of sandy limestone.

At the "lower beds" (plate 1, fig. 6) , those further down the creek, conditions are quite different from those just described, although of the same horizon: the banks of the creek are higher and almost perpendicular, so that the chances of collection from the shales are fewer. The bed of the creek, however, is wider and there are more nodules washed out. The most abundant fossil at this place is Neuropteris. The nodules at the upper end of the exposure are all, almost without exception, barren of fossils. The exposures here are of about the same extent as the "upper beds," though the species are not so varied. Judging from the collections made while there, Arthropoda are more abundant at the "lower beds."

Bradley (Geol. Surv. Illinois, IV, 196, 1870) mentions the occurrence of these nodules at or near Morris. Other than these places the nodules have been thrown out of a coal mine near Braidwood, Illinois. Doubtless close search would reveal other localities where the shale is cut through in mining. The beds at both places are slightly folded. This is true especially of the "upper beds," where a conspicuous fold caused the beds to disappear in the bed of the creek and to reappear farther down stream. This is directly across the large "ox-bow" bend of the creek.

The beds at Mazon Creek were first explored in 1857 by Mr. Joseph Evans, who sent his specimens to Berlin, Germany, where they excited great interest. It was he who collected the type specimen of Amphibamus grandiceps Cope. Since the time of Mr. Evans many have collected at Mazon Creek, and without doubt the fossil-bearing nodules from this locality are more widely scattered in the museums of the world than are organic remains from any other one horizon.

So far as we know there was no upland vertebrate life at that time. The forms at present known were confined to the water or the margins of the water. The absence of knowledge of upland and terrestrial deposits of this time doubtless accounts for the absence of known vertebrates. It is, however, especially interesting to speculate on the ancestral types of the land vertebrates, and it must be admitted that the Coal Measures Amphibia as at present known throw the ancestry of land-living vertebrates far back into geological time.

(m) The Cannelton slates of Beaver County, Pennsylvania, have furnished 3 species of Amphibia and fragments of other species are represented in the U. S. National Museum (462). The species so far known are: Tuditanus minimus Moodie, Erpetosaurus sculptilis Moodie, Erpetosaurus minutus Moodie.

They are the first evidence of the occurrence of amphibian remains in these deposits. The Cannelton specimens are found in a thin stratum of slate which forms part of the Middle Cannelton Coal. The Cannelton slate, in which the fossils occur, forms the roof of the Middle Kittanning Coal, which is only 20 to 30 feet above the Lower Kittanning bed (I. C. White), so it becomes evident that the deposition of the Cannelton slates was at only a slightly later period than that of the shales in which the Mazon Creek nodules occur, since the Mazon Creek shales form the roof of the Morris, which "is probably somewhat lower than the Lower Kittanning of Pennsylvania." From the Cannelton slates are known the remains of plants, insects, crustacea, especially "Eurypterids found in shale immediately below the Darlington (Upper Kittanning) Cannel Coal, near Cannelton, Darlington Township, Beaver County, Pennsylvania, Horizon, Allegheny River Series" (Hall, 1884). In these shales occur also, fishes and the 3 species of amphibians referred to above. The Amphibia known from this region are small, the largest of them not exceeding 6 inches in length.

Fig. 4. Portion of "West Virginia-Ohio-Pennsylvania, Wellsville Quadrangle" of the U.S. Geological Survey, to show topography and situation of Linton coal mines. Some fossil amphibians doubtless came from across the line in Columbiana County.

(n) The Linton, Ohio, beds outcropped near Linton post-office, which was formerly located at the mouth of Yellow Creek, a few hundred yards from the present station, Yellow Creek, Salem Township, Jefferson County, in the valley of Yellow Creek, near the Ohio River, and thus near the Pennsylvania state line.

In regard to the exact location of the town of Linton, which has long since been abandoned, I quote from a letter from Dr. Louis Hussakof, who visited the locality:

Newberry says, in regard to the fauna of the Linton Coal:

On page 18 is a list of the Amphibia which are thus far described from the Linton deposits. They all belong, so far as known, to the Microsauria, the reference of any of the species to other orders being doubtful. The larger Amphibia seem to be indicated by a large rib which resembles very much that described by Huxley in 1863 for Anthracosaurus.

Amphibia from the Linton Beds (51 SPECIES).

Brachydectes newberryi Cope. Fragment of a skull.
Cercariomorphus parvisquamis Cope. Impression of body.
Cocytinus gyrinoides Cope. A skull and anterior dorsal vertebræ.
Ctenerpeton alveolatum Cope. Large portion of skeleton, no skull.
Diceratosaurus lævis Moodie. Complete skull.
Diceratosaurus punctolineatus Cope. Anterior vertebræ, part of skull, with ribs and portion
of ventral armature.
Diceratosaurus robustus Moodie. Incomplete cranium.
Eoserpeton (Ceraterpeton) tenuicorne Cope. Incomplete skull.
Erpetosaurus acutirostris Moodie. Complete skull.
Erpetosaurus obtusus Cope. Incomplete skull.
Erpetosaurus radiatus Cope. Incomplete skull.
Erpetosaurus tabulatus Cope. Incomplete skull, with clavicles.
Erpetosaurus tuberculatus Moodie. Incomplete skull.
Eurythorax sublævis Cope. A single interclavicle. (Operculum of lung fish, Sagenodus.)
Hyphasma lævis Cope. Incomplete skull and anterior vertebræ.
Ichthycanthus ohiensis Cope. Portion of dorsal region.
Ichthycanthus platypus Cope. Posterior portion of body.
Leptophractus dentatus Moodie. Mandible.
Leptophractus lineolatus Cope. Incomplete skull.
Leptophractus obsoletus Cope. Portions of skull.
Macrerpeton deani Moodie. Mandible and part of skull.
Macrerpeton huxleyi Cope. Part of cranium.
Molgophis brevicostatus Cope. Part of vertebral column with ribs.
Molgophis macrurus Cope. Vertebral column.
Molgophis wheatleyi Cope. Part of skull with 25 vertebræ.
Odonterpeton triangularis Moodie. Skull and anterior part of body.
Œstocephalus rectidens Cope. Part of mandible.
Œstocephalus remex Cope. Skull and anterior part of body.
Pelion lyelli Wyman. Cranium, fore part of body, hind limb.
Phlegethontia linearis Cope. Skull and anterior part of body.
Phlegethontia serpens Cope. Series of 22 dorsal vertebræ.
Pleuroptyx clavatus Cope. Part of vertebral column and limbs.
Ptyonius marshii Cope. Part of skull and anterior vertebræ.
Ptyonius nummifer Cope. Skull and greater part of vertebral column.
Ptyonius pectinatus Cope. Many specimens, some nearly perfect.
Ptyonius serrula Cope. Nearly complete skeleton.
Ptyonius vinchellianus Cope. Skull and anterior vertebræ.
Saurerpeton latithorax Cope. Skull and fore part of body.
Sauropleura digitata Cope. Greater part of body minus skull.
Sauropleura (Anisodexis) enchodus Cope. Part of jaw.
Sauropleura foveata Cope. A single interclavicle with impression.
Sauropleura longidentata Moodie. Incomplete skull with mandible.
Sauropleura newberryi Cope. Two incomplete skulls with vertebræ.
Sauropleura pauciradiata Cope. Elements of a pectoral arch.
Sauropleura scutellata Newberry. Imperfect skeleton.
Stegops divaricata Cope. Nearly complete skull.
Thyrsidium fasciculare Cope. Dorsal vertebræ.
Tuditanus brevirostris Cope. Skull and anterior vertebræ.
Tuditanus longipes Cope. Part of vertebral column with limbs.
Tuditanus punctulatus Cope. Skull and anterior part of body.
Tuditanus walcotti Moodie. Skull and portions of body.

Besides the above-listed species there are others indicated by fragments too poorly preserved to be worthy of specific designation. The Linton Amphibia are all apparently confined exclusively to that locality. Species from the Cannelton slates have been assigned, however, to genera which occur at Linton, i.e., Erpetosaurus and Tuditanus. This reference may be due to lack of knowledge, as the forms are insufficiently known. A single Linton species has been assigned to Ichthyerpeton, a genus known otherwise only from the Coal Measures of Kilkenny, Ireland. Cope referred species from Linton to the genus Ceraterpeton of Huxley, from Kilkenny, Ireland, but Jaekel (347) and the writer (462) have shown that the species were incorrectly assigned to the genus Ceraterpeton, and that in fact they represent widely distinct genera. A single species has been identified by Eastman from the Des Moines limestone of Iowa as identical with one from Linton, Pleuroptyx clavatus Cope. The Linton fauna is distinct from that of the Mazon Creek beds, and also from that of South Joggins, Nova Scotia.

(o) The deposits in Nova Scotia have been correlated with the Coal Measures strata of the United States (Bell, Summ. Rpt. Geol. Surv. Canada, 1912, 1914, 360-371). They are very near the same age as the Linton beds and come in near the base of the Allegheny River series. The exposures are at the South Joggins, along the sea-coast. Here in strata of clay interstratified with coal are found the erect stumps of the Sigillariæ, and it was in the rock within these stumps that Lyell and Dawson, in 1853, discovered the remains of the amphibians which they termed "reptiles."

Fifteen species have been described from the Joggins deposits. Two are known from the Albion mines, south Nova Scotia, where were obtained the remains of Baphetes planiceps Owen and B. minor Dawson.

The following 17 species of Amphibia are known from the Carboniferous of Canada:

Amblyodon problematicum Dawson. Teeth and fragments.
Baphetes minor Dawson. An incomplete mandible.
Baphetes planiceps Owen. An incomplete cranium from Albion.
Dendrerpeton acadianum Owen. A jaw, limb bones, and fragments.
Dendrerpeton oweni Dawson. Phalangeal bone and fragments.
Eosaurus acadianus Marsh. Two dorsal vertebræ.
Fritschia curtidentata Dawson. A mandible, vertebræ, ribs.
Hylerpeton dawsoni Owen. Mandible, teeth and incomplete maxilla.
Hylerpeton intermedium Dawson. Mandible and portions of skull.
Hylerpeton longidentatum Dawson. Fragments of mandible and skull.
Hylonomus latidens Dawson. Mandible and teeth.
Hylonomus lyelli Dawson. Incomplete skeleton and part of skull.
Hylonomus multidens Dawson. Fragments of skull.
Hylonomus wymani Dawson. Mandible and vertebræ.
Platystegos loricatum Dawson. Incomplete skull, vertebræ.
Smilerpeton aciedentatum Dawson. Teeth, ribs, fragments.
Sparodus sp. indet. Teeth, scales.

(p) All the remains representing the above species were collected by Sir J. William Dawson at the South Joggins and at the mines of Albion, with the exception of Eosaurus, which was collected by O. C. Marsh. The collections of Dawson are now in the Peter Redpath Museum of McGill University in Montreal and in the British Museum of Natural History at South Kensington, London. The history of the discovery of the deposits and their amphibian fossils at the South Joggins is so interesting that it was thought worth while to reproduce in large part Dawson's paper "On the Mode of Occurrence of Remains of Land Animals in Erect Trees at the South Joggins, Nova Scotia," published in 1891 in the Transactions of the Royal Society of Canada, section IV, p. 127:

Fig. 5. Dawson's tree No. 13 at the South Joggins, Nova Scotia. Upper part, in situ, in the reef after it had been exposed by blasting. (After Dawson, based on a photograph.)

After the discovery of these entombed amphibians Sir William Dawson was given a grant of £50 from the Government Fund by the council of the Royal Society of London, to aid in the extraction of these trees and the collection of their contents. The trees were carefully taken out and their contents examined; the portions containing the animal remains were carefully boxed to be taken to Montreal for final cleaning and study. Erosion goes on rapidly at the South Joggins, but no one has paid any attention to the occurrence of Amphibia along the coast of Nova Scotia within recent years.

(q) A deposit which will be of undoubted interest in connection with the occurrence of Amphibia in the Coal Measures is that which outcrops along the banks of Rock Creek in the South western part of Douglas County, Kansas, in Marion Township (Township 14 south, Range 18 east, SW. and SE. quarters of section 7), about 2 miles from the now-abandoned post-office of Twin Mounds, so called from the two flat-topped, elongated mounds of Oread limestone to the west of the town.

The interest in these beds is not due to the discovery of Amphibia in them, but the possibilities of such discoveries. This is indicated by the occurrence of fossils, in nodules similar to those obtained from Mazon Creek, which are identical generically, nd in most cases specifically, with the Mazon Creek animals and plants.

The fossils so far collected from this interesting locality are:

The fossils occur in definite strata of nodules immediately above a 10-inch coal seam which is worked for local consumption. The coal lies near the base of the exposure in the more western portion of the outcrop, but it is raised by an anticlinal fold to near the top of the creek-banks by the bridge across Rock Creek, along the banks of which the shales are exposed. Nodules containing fossils are found most abundantly at the western exposure on the McKinzie place, only a few having been found near the bridge.

Below the coal-seam, nodules of various shapes and sizes occur, but they seldom contain fossils and never good ones. Occasionally, as at Mazon Creek, fragments of plants adhere to the outside of the incrusting shale. A single nodule may have adhering to it fragments of 4 genera of plants. The fossiliferous nodules all occur above the coal and are most prolific and abundant immediately above the seam, within the first 10 inches. In the same stratum of shale with the nodules are found abundant impressions of plants in the shale, often perfect fronds being uncovered. (See, in this connection, Twenhofel and Dunbar, 1914, "Nodules with Fishes from the Coal Measures of Kansas," Amer. Journ. Sci., XXXVIII, pp. 157-163.)

G. F. Matthew (408-413) has described numerous genera and species of footprints, presumably amphibian, from the Carboniferous of Canada. The impressions indicate small creatures for the most part. Other imprints have been described by Logan, Dawson, Lyell, Marsh, Mudge, and Lea. Since the present work is intended largely for a morphological review, only passing notice can be given to the ichnites. The literature on the "Ichnites" has been brought together in Hay's "Bibliography and Catalogue of Fossil Vertebrata of North America," pp. 538-553. References since the publication of Hay's catalogue (317) will be found in the bibliography at the end of this work. Footprints are of interest in that they are the only evidence we have of the occurrence of land vertebrates in the Devonian and Mississippian of North America.

THE MORPHOLOGY OF THE COAL MEASURES AMPHIBIA.

The anatomy of the Coal Measures Amphibia presents many primitive types of structure. Their organization represents a stage passed through in the ontogeny of higher vertebrates. The animals are similar in a general way, yet so diverse are the modifications which they have suffered under different environmental conditions, that close scrutiny is needed to discern the exact relationship of the forms. Our knowledge of this relationship is based on the structures preserved, which are largely skeletal, since little is known of the soft anatomy (471) of the air-breathing vertebrates of the Coal Measures. The pubis is ossified in the Paleozoic Amphibia later than the ischium and ilium; the carpus and tarsus are cartilaginous; the vertebræ consist of a pleurocentrum and two neurocentra, thus paralleling conditions in modern mammalian embryos.

Fig. 6.—Generalized figure of dorsum of an early amphibian skull to show position of elements and terminology adopted in this work. The outline is based on that of Eryops, but is in no way intended to indicate that form. a. com, anterior commissure of lateral-line canals; com, commissural communication between infra- and supra-orbital lateral-line canals; fr, frontal; inf, interfrontal; inn, internasal; info, infraorbital lateral-line canal; it, intertemporal; jl, jugal lateral-line canal; j, jugal; lar, lacrimal; mx, maxilla; n, nasal; oc, occiput; occ, occipital cross-commissure of the lateral-line system; or, orbit; par, parietal; pof, postfrontal; pmx, premaxilla; pf, prefrontal; po, postorbital; pp, postparietal; g, quadrate; qj, quadratojugal; spo, supraorbital lateral-line canal; sq, squamosal; spt, supratemporal; t, temporal lateral-line canal; tab, tabulare.

(a) The skull of the Coal Measures Amphibia has (fig. 6) essentially the same structure in the different groups. It is largely formed of bones of intramembranous origin, representing the face bones of the mammalian skull. The skull in life was doubtless a chondrocranium with the membrane bones laid down upon the cartilaginous box containing the sense-organs, as in the sturgeon (Acipenser), where the surface bones of the face were probably originally scales, which later became consolidated into large bony scutes. The membrane bones of the early Amphibia may have been originally derived from scales, but at present nothing is known of this origin; doubtless the elements had an intramembranous origin in the ancestors of the group. Judging from Credner's studies on the series of specimens of Branchiosaurus amblystomus Credner (187), the skull bones do not ossify completely until relatively late in the life of the individual. The skull in the youngest individual figured by Credner (op. cit., Taf. XVI, fig. 1) seems to be largely cartilaginous, with beginnings of separation into the skeletal elements. The manner and time of development and ossification of the skull seems to proceed much as it does in modern amphibians. The condition found in the skull of Cryptobranchus allegheniensis or Necturus maculosus will represent pretty accurately the condition of most of the Coal Measures Amphibia. The face bones in certain forms were sculptured and cut by lateral-line canals.

A median suture divides the skull into two equal regions dorsally. On the sides of this median suture lie pairs of elements which are common to all higher vertebrates. These elements are: the premaxillæ, nasals, frontals, parietals, and post-parietals. All of these elements vary somewhat in shape and slightly in arrangement, but always occupy the same relative positions. To the side of these elements lie the prefrontal, the postfrontal, the supratemporal, the squamosal, and tabulare, and occupying the margin of the skull are the maxilla, the jugal, the quadratojugal, and possibly the quadrate in a few forms. The parietal foramen occurs usually within the parietal bone, but its position is subject to slight variations and it may occur on the suture between the frontal and the parietal, or even far posterior near the postparietal. The nostrils often lie well forward and are included by the premaxillæ, nasals, and prefrontals. The orbit is usually well posterior, but it may occur far forward. It is bounded by the prefrontal, the frontal, the postfrontal, the post-orbital, and the jugal. Sometimes the lacrimal is present and has been clearly identified on the anterior margin of the orbit in a few cases.

(b) Sclerotic plates often occur within the orbits, and are not confined to any particular group, though they are quite constant among the Branchiosauria. They are usually delicate, thin, broad plates which evidently overlap and operate as in modern animals. The number varies, as many as 30 occurring within the orbit of one branchiosaur. Between the margin of the orbit and the sclerotic plates there often occur, in the Branchiosauria (186) particularly, small scale-like particles which were doubtless embedded in the heavy skin above the orbit during life.

(c) The palate of the skull is very incompletely known, being indicated in a very few cases. These specimens, however, show that the characters of the palate were quite similar, if not identical, in essential respects with the palate among the European species of the same or slightly later time.

A large cultriform parasphenoid occupies the posterior portion of the palate, on either side of which in some species lies the posterior palatine foramen. On the sides of the anterior prolongation of the parasphenoid lie the vomers (186), membranous bones often bearing minute tubercular teeth, apparently adapted for crushing. The vomers and the maxillæ, with sometimes the palatine, surround the anterior palatine foramen, which is almost always present; sometimes, however, quite small. The transverse or ectopterygoid unites the pterygoid, a broad plate of thin bone, with the maxilla and jugal.

(d) The teeth of the Coal Measures Amphibia (194) are remarkably similar in the various forms. They are always slender, pleurodont denticles arranged in a single row on the jaws or as tubercular eruptions on the palate bones, with a large pulp-cavity and the enamel often striated. The food of the creatures must have been small Crustacea, worms, insects, and succulent vegetation, such as is the food of the modern Amphibia.

(e) The occiput is formed of partially ossified (465) ex- and basi-occipitals, though these elements are never firmly united by ossific union. Often a pair of condyles occur, one on either exoccipital. The occiput was usually, however, cartilaginous and no trace of its structure is preserved.

(f) The mandible is usually as long as the skull and is slender. It is composed of 6 elements so far as known (465); these are the articular, the surangular, the angular, the coronoid, the dentary, and the splenial. Other elements may be present, but the anatomy of this portion of the animals is not very completely known. The bones are sculptured and cut by lateral-line canals (458) in a few forms. Whether the articular operated on an osseous or cartilaginous quadrate is unknown, though certain specimens seem to indicate an osseous condition for that element. The anterior symphysis was doubtless ligamentous, the halves always separating before fossilization. The dentary always bears a single row of pleurodont teeth, which may vary greatly in size and number.

(g) The hyoid apparatus is well preserved in a few forms (123). Doubtless it was present in all of them, though it has seldom been preserved. The condition of the hyobranchial apparatus in Cocytinus gyrinoides (text-fig. 16) from the Coal Measures of Linton, Ohio, seems to indicate that the species was a perennibranchiate salamander (123). It is well known from the studies of Credner that the European Branchiosauria, in the young, possessed external branchiæ (187) supported by lateral basibranchials. The gill-arches seem to have been slightly calcified or ossified in a few cases, and they supported denticle-like projections which bore the gill-filaments. When the Branchiosauria had attained a length of 100 mm. or more they lost their gills (187). This change was accompanied by the reduction of the tail, expansion of the pelvis, and increase in ossification of the skull and skeletal elements. Gills have not yet been detected among the American Branchiosauria.

(h) The eye in a few species had a large amount of black pigment, as indicated by the blackening of the stone in the Mazon Creek nodules. Professor Cope (107) thought that this would indicate a nocturnal and crepuscular habit for these vertebrates, since the pigmentum nigrum of the choroid is largely developed. Other than this suggestion nothing is known of the soft parts of the head.

(i) The alimentary canal (text-fig. 7) is beautifully preserved as a cast in three specimens of the American branchiosaur species Eumicrerpeton parvum Moodie (471) from the Mazon Creek beds. The nodules which contain these interesting little fossils measure less than 3 inches in long diameter. The fossil salamanders, about 30 mm. in length, are preserved on their backs and occur as nearly as is possible in the center of the nodule.

If it were not for the fact that the œsophagus became loosened and dropped from its place shortly after death, the alimentary canal would be in place and would immediately recall a freshly dissected specimen of a recent salamander. The anterior end of the œsophagus lies obliquely across the chest region with its tip pointing slightly downward. The length of the œsophagus proper, in one specimen, is only about 3 mm. As it is preserved, the œsophagus lies in a semi-sigmoid curve with the convexity anterior, and enters the cardiac portion of the stomach by a gradual constriction. The stomach is clearly preserved as a distinct sac-like organ with two lobes which correspond to the cardiac and pyloric limbs. It measures about 7 mm. in length by 2 mm. in its greatest diameter. The muscular constriction which divides the organ into pyloric and cardiac divisions occurs at a distance of 4 mm. from the upper end. The pylorus is designated by a rather pronounced constriction which may be partly accidental, although it recalls the pylorus of modern frogs. From this constriction, which lies on the left side of the fossil, as it is preserved, the duodenal portion of the intestine makes a straight course posteriorly to near the anal region, where it takes a sharp bend and curves back to run parallel with itself for the distance of 4 mm. In its upward course the intestine enlarges, and practically the same enlargement continues throughout the remainder of the course to the anus. At a distance of 1 mm. from the anal end, the rectum dilates probably 0.125 mm. to form the cloaca. After the intestine has continued its parallel course for the 4 mm., as above stated, it turns abruptly to the right for a distance of 2 mm. It then runs posteriorly for a short distance, then bends back and under itself to again make a double sigmoid curve, when at a distance of 6 mm. from the anus it assumes a straight course, which it continues to the end.

Fig. 7.—Alimentary canal of Coal Measures salamander as illustrated by the smaller specimen of Eumicrerpeton parvum Moodie, from the Mazon Creek shales. × 3. Original in Yale University Museum. a, anus; dd, duodenum; in, intestine; l, impression of liver(?); oes, œsophagus; st, stomach.

The anus lies at a level which is approximately that of the lower end of the femur, which is preserved as an impression on the left side of the fossil, thus agreeing in its position with that found in modern Caudata. Lying inside the curve of the stomach and partly inclosed by the œsophagus is a smooth area which may possibly represent the impression of some of the accessory digestive glands, such as the liver. Occurring in this smooth area are numerous fine lines which possibly represent the impressions of blood-vessels; but they are so imperfectly preserved that one can not be sure.

Representatives of several genera of the modern Caudata have been dissected in order to make a direct comparison of the fossil alimentary canal with that of the recent forms. The alimentary tract of Desmognathus fuscus Raf. from the vicinity of Ithaca, New York, resembles in a marked degree that of the fossil form. The nearest approach to the condition there represented is found, however, in an immature branchiate individual, some 47 mm. in length, of Diemyctylits torosus Esch., from a fresh-water pond on Orcas Island in Puget Sound. The presence of this species on the island is very suggestive. It is of extreme interest, too, that the condition represented in the alimentary tract of the fossil branchiosaurs should resemble so closely that of an immature rather than a mature form.

(j) The vertebral column is clearly and readily separable into cervical, dorsal, sacral, and caudal regions. The neck is always short, with from 5 to 10 vertebræ, cervical ribs often present. The dorsal region is not long, but varies from 20 to 30 in the constituent vertebræ. There is a single sacral vertebra not always to be readily distinguished from those of the dorsal and anterior caudal series. The tail may be very short or extremely long, with neural and hæmal spines elongate and flattened into an oar-like appendage. The distal caudals are in some species cartilaginous, apparently always so in the Branchiosauria.

(k) The atlas and axis are unknown among the American specimens, but we are able to infer from the structure of the other vertebræ what this must have been; and our inferences are partly confirmed by the conditions existing in the European forms (187). The atlas, apparently, consisted of a pair of neurocentral plates which are partly ossified, partly embedded in cartilage, judging from the edges of the plates which have been preserved. The centrum seems not to have been present in the atlas, or if present it was only very slightly developed and quite free from the neural pieces and largely embedded in cartilage. A fairly accurate picture of the condition of the atlas and axis may be seen on examining a cow, pig, or chick embryo (378) in the early stages of vertebral development, which has been cleared by the Schultze method (Amer. Journ. Anat., VII, No. 4, 1908).

(l) The dorsal vertebræ, as well as those of the other series, present a primitive character (fig. 8) in the persistence of the notochord (540). Among the Branchiosauria the notochord was not at all or but slightly constricted intravertebrally, but among the Microsauria it was constricted so far that the notochordal remnants in each centrum resemble an hour-glass.

The structure of the vertebræ among American forms agrees fully with that outlined by Credner, Fritsch, and others for the European species. The details of structure are so fully given by Zittel (642, pp. 346-353) and by Schwarz (540, 541) that it will not be necessary to state more here as to their structure, since there is nothing new to add concerning the American species.

The temnospondylous vertebra of the same nature and type as exhibited by the Permian forms has its representatives (94, 478) in the Coal Measures. Spondylerpeton spinatum Moodie (478) (plate 4, figs. 1, 2) and Eryops sp. (plate 18, fig. 2) indicate the embolomerous and rachitomous types of vertebral structure. The occurrence of these widely different types of vertebral structure indicates a long history for the group prior to the Coal Measures. This history is further indicated by footprints in the Mississippian and Devonian of this continent.

Fig. 8.—Vertebræ and ribs of Amphibia from the Coal Measures of Linton, Ohio. Originals in Geol. Inst. Berlin. (All after Schwarz.)

A. Caudal vertebra of Œstocephalus remex Cope. Lateral view. × 4.

B. Caudal vertebra of Ptyonius vinchellianus (?) Cope. Lateral view. × 6.

C. Dorsal vertebra of Ptyonius pectinatus Cope. Lateral view. × 9.

D. Notochordal cones and spinal canal of Thyrsidium fasciculare Cope. × 3.

E. Rib of Molgophis sp. Cope. × 1.75. c=capitulum; t=tuberculum.

F. Vertebra of Molgophis sp. Cope, from ventral side. × 2.

G. Dorsal vertebra of Ptyonius pectinatus Cope. From above. × 8.

H. Dorsal vertebra of Thyrsidium fasciculare Cope, from below. × 2.5.

I. Vertebra of Phlegethontia linearis Cope, from above. × 5.5

J. Rib of Œstocephalus remex Cope, from posterior dorsal region. × 5.

K. Dorsal vertebra of Thyrsidium fasciculare Cope, from above. × 1.5

L. Anterior dorsal vertebra (cervical?) of Thyrsidium fasciculare Cope. Lateral view. × 1.5.

M. Vertebra of Phlegethontia linearis Cope, from side. × 5.

(m) The ribs (fig. 8) are very diverse in structure and in their mode of articulation (541) with the vertebral column. The characters of the ribs and vertebræ constitute the best means of classification of these animals so far discovered. In the Branchiosauria the ribs are always straight, heavy, and short, and articulate intravertebrally upon a large and strong transverse process. They occur throughout the vertebral column. There is a single pair of sacral ribs which are not to be clearly distinguished from the pre-sacral and post-sacral series. The cervical and caudal ribs are shorter than the dorsal series. The branchiosaurian rib is composed almost entirely of perichondral ossification. It presents the same condition as does the humerus of the cow embryo of 2 to 3 inches in length. The ribs of the branchiosaurs are identical in every way with the ribs of modern salamanders and form one of the strong arguments in favor of the relationship of the Branchiosauria to the Caudata. Among the Microsauria the ribs are always long, slender, curved, and intercentral. They may be either single or double headed, but usually the former. They resemble in their characters the ribs of some of the early reptiles and an attempt has been made to relate the Microsauria (469) to the primitive reptiles on this basis. The ribs of the other groups are still unknown. Indeed, representatives of the Temnospondylia and the Stereospondylia are very scanty in the American Coal Measures. One large rib (plate 22, fig. 4) may represent a labyrinthodont, but nothing is known of the species to which the rib belonged.

(n) The pectoral girdle (187) is a very simple and uniform structure, although the details of the association of the elements still remain to be determined. A single, median, usually large and elongate interclavicle occupies the ventral line of the chest. This is morphologically the same element which occurs in the middle line of the chest of the lizards. It is a dermal bone and is usually, especially among the Microsauria (462), highly sculptured. It varies considerably in size and shape, but is remarkably uniform throughout the various groups. Lying anterior to the interclavicle and overlapping its antero-lateral margins lie the two clavicles, which are usually diamond-shaped and are sculptured, dermal bones. The position of the coracoid is still uncertain, and in fact its clear association in the pectoral girdle of these species is still a question. It seems to be constant in the European (186, 251) species and is usually represented by a small rounded plate of bone, which in life no doubt had a large amount of cartilage to form its borders. A cleithrum (285) has been ascribed to one of the Linton, Ohio, species (plate 15, fig. 3) by Jaekel (347), but this needs confirmation. An osseous scapula is usually present, resembling the scapula of modern salamanders, in that it was largely embedded in cartilage. The position of the pectoral girdle is largely a matter of doubt, especially for the American species. After death and before fossilization the girdle was always moved by post-mortem shifting, so that its exact relation to the ribs and vertebral column is still in doubt. Credner (186) has restored the pectoral girdle close behind the head, which would cause an amount of rigidity in the body which probably did not exist.

(o) The arm consists of the humerus, radius, ulna, and 4 digits. The characters of the arm-bones are such as is constant among primitive animals and developing mammals. The osseous portion is perichondral. Epiphyses are totally lacking and it is doubtful if the endochondrium was at all ossified. The digits are often terminated by ungual phalanges, although usually the terminal phalanx was merely embedded in the web of the foot; and among the terrestrial forms a claw was well developed. An osseous carpus is not known in the species from the Coal Measures. Its impression indicates a broad hand, well adapted for swimming.

(p) The pelvic girdle consists uniformly (462) of the ilium and ischium. A small rounded pubis is present in some of the later forms of Amphibia; it is, however, totally absent from the Coal Measures species. The condition of the pelvis is paralleled by the partially grown pelvis of mammalian embryos in which the elements ossify in the order of ilium, ischium, and pubis. The ilium is always the larger of the elements. It supported or was attached to the sacral rib by means of a ligamentous union. The ischium did not ossify completely until the animal was nearly mature. The union between the elements of the pelvis was probably of a loose, membranous sort or else the whole mass was embedded in cartilage; of the two hypotheses the former is the more probable.

The pubis is indicated as a calcified quadrangular plate in a specimen of Amphibamus grandiceps Cope (478) from the Mazon Creek shales, and it is present as a rounded osseous element among some of the Permian forms.

(q) The leg (fig. 21, B) is composed of the femur, tibia, fibula, and 5 digits. The tarsus is usually cartilaginous, a single osseous tarsus (483, 484) being known (plate 23, fig. 1) from America. The distal phalanges may or may not be clawed, depending on the habits of life of the animal. The elements of the leg are ossified in a similar manner to those of the arm.

(r) The ventral scutellation (fig. 9), so commonly present among all groups of Amphibia in the Coal Measures, consists of a series of ossifications or calcifications in the myocommata. Among modern amphibians they occur as thin perpendicular planes of connective-tissue which are sometimes cartilaginous, especially in Necturus, regarded by Wilder (Memoirs of the Boston Society of Natural History, vol. V, No. 9, p. 400, fig. 6, 1903) and by Wiedersheim (605, p. 58) as a homologue or predecessor of the sternum, although Wiedersheim says:

This supposition is fully sustained by the anatomy of the Branchiosauria (459), which must be looked upon as the actual ancestors of the Caudata. Wilder says of these structures in Necturus (op. cit., p. 400):

Fig. 9.—Ventral scutellæ of Micrerpeton caudatum, a Coal Measures salamander from Mazon Creek. × 5. f, femur; h, humerus; ls, lines of scutes; v, vertebral column.

The homologue of the ventral scutellæ is found in plesiosaurs, crocodiles, Sphenodon, and other reptiles in the "abdominal ribs," and the same myocommatous ossifications undoubtedly go to the formation of the chelonian plastron. What the causes were which produced the development of the ventral scutellæ to such a high degree among the primitive land vertebrates is uncertain, but they are certainly more highly developed among the primitive reptiles and amphibians than among the later members of those classes. Among the Amphibia of the Coal Measures they attained, in some forms, a high degree of development and differentiation. They are present in all families so far known, except the Tuditanidæ, in which the myocommata may have been cartilaginous. The Sauropleuridæ present the highest development of these structures among the American forms, in which the scutes are large and osseous. Among the Branchiosauria they are calcified or partially ossified and are always arranged en chevron on the belly, chest, arms, and throat, their arrangement and direction of the chevron being modified according to the myomeres of the various regions. The ventral scutellæ of the European Branchiosauria are figured and described fully by Credner (192, p. 21, figs. 4 to 11).

(s) Scales (fig. 10 and plate 24, figs. 2 and 3) are present on the body of (462, 485) several species. It is a matter of regret that their preservation is so imperfect that nothing can be found out as to their structure. The Linton species, which possess scales, are, of course, carbonized and hence impracticable for microscopic study, and in the Mazon Creek species of Amphibamus and Micrerpeton the scales have been replaced by kaolin. The bodies of two species (Cercariomorphus parvisquamis and Ichthyerpeton squamosum) of the Linton Coal Measures Amphibia were completely scaled. The scales in the Branchiosauria (462), so far as they have been observed, are slightly imbricated; rounded, with concentric markings after the manner of the modern cyprinoid fish-scale. They are extremely minute, and whether or not they covered the entire body of the animal is unknown. On the body of Cercariomorphus the scales have the appearance of being tubercular without imbrication, and they apparently covered the entire bodily surface of the animal.

Fig. 10.—Horny armor of back of Hylonomus. a, imbricated scales; b, horny plates; c, horny spines or tubercles; d, small imbricated scales. (After Dawson, based on a photograph.)

Among the Paleozoic Amphibia from Nova Scotia as described by Dawson and Owen (193, 201, 485) scales are well developed and frequent, although the details as to their occurrence on the bodies of the animals are still unknown, since the Nova Scotian species are all based on very fragmentary remains. Dawson (208, p. 34) has given a detailed discussion of the discovery and anatomy of the various types of scales possessed by the species from the Coal Measures of Nova Scotia. Suffice it to say here that none of the scales appear to be bony, but have a cuticular appearance with concentric markings. Some of them are tubercular, and Dawson thought that a few specimens indicated that some of the species possessed scaly lappets and a dorsal nuchal fringe of scaly skin along the back. He has indicated these facts in his restorations of the forms. The scales were all carbonized and burned readily with a strong flame. A section of the scale shows a thick upper corium with a vascular body (208, pl. IV, fig. 29) much like a fish-scale. Fragments of the skin were also preserved with the scales. Dawson says of the skin:

(t) Muscle tissue (fig. 21) is preserved in a single specimen, previously described by the writer (464, p. 17, pl. 7, fig. 1). The carbonized muscles show a myomeric arrangement and the portions preserved probably represent one of the recti muscles of the abdominal wall.

(u) The lateral-line system in the Coal Measures Amphibia will be best understood from a comparative review of the occurrence of these organs among all extinct Amphibia. Since all the orders of Amphibia are represented in the Coal Measures, such a review will not be out of place here.

The preservation of the lateral-line system among ancient Amphibia is due to the fact that the skull of many forms (especially the later and larger) are grooved and marked by a regular series of furrows and pits, in which the sense-organs of the lateral-line system were contained (see fig. 6), as well as by the preservation of a series of clearly marked scales on the sides of the tails and bodies of others. The grooves are never arched over as in the Macropetalichthyidæ, where "in favorable specimens each is shown to be covered by a delicate roof perforated by two lines of minute openings" (Dean, N. Y. Acad. Sci. Mem., vol. II, pl. III, p. 115). They are always widely opened canals, either with perfectly smooth bottoms and sides or roughened with large pits, or even becoming a series of well-marked pits. An attempt has been made (458) to homologize the organs with those of fishes.

The nomenclature adopted here for the canals does not depart from that used by Allis for Amia (Journ. of Morphology, vol. II, 1889). The supraorbital and infraorbital canals are readily correlated with those of the same name in fishes, where they are very clearly marked. The anterior commissure is also homologous with that of the fishes, as is also the canal here called the "antorbital commissure." The others are not so readily homologized. The upper canal (see fig. 6) in the posterior part of the cranium is here designated the temporal canal. It is, however, clearly a part of the infraorbital of the fishes. Its relations in the Stegocephala are such that a new name is deemed necessary. The jugal canal is, I believe, a new formation in Amphibia. The transverse canal of the amphibian skull is homologous with the "occipital cross-commissure."

The figure (see fig. 6) is a composite picture of the lateral-line system of the higher or truly stegocephalous Amphibia. The outline of the skull is based on that of Eryops. All of the canals do not exist on any one skull or in any one order, but all are found somewhere in the group.

Fig. 11.

A. Skull of Eoserpeton tenuicorne Cope, showing arrangement of cranial elements. × 2. fr, frontal; j, jugal; mx, maxilla; n, nasal; or, orbit; par, parietal; pof, postfrontal; pmx, premaxilla; po, postorbital; pp, postparietal; qj, quadratojugal; sq, squamosal; spt, supratemporal; tab, tabulare.

B. Outline of skull of Ceraterpeton galvani Huxley from the Carboniferous of England. Heavy broken lines show the distribution of lateral-line canals. × 1. (After Andrews.) fr, frontal; par, parietal; or, orbit; po, postorbital; pp, postparietal; spt, supratemporal; tab, tabulare.

The canals have been described in all known orders of fossil Amphibia and the system is found likewise in all the living orders, including the Gymnophiona, which have "a strong line of lateral sense-organs" (Gadow). In the Branchiosauria, the earliest of the true Amphibia (Euamphibia) and ancestral to the modern Caudata, the lateral-line system is known on the tails of two genera (462, 478) from the Mazon Creek, Illinois, shales—Micrerpeton and Eumicrerpeton. The system as there defined has been fully discussed in the description of the anatomical details of the species, to which reference may be made for further data (pp. 52-60). Suffice it to say here that the system of sense-organs there preserved is identical with that of the larval Necturus; the lines arising as a median from the tip of the tail and a dorsal springing from the median at a distance of a few millimeters from the tip of the tail. The lines are more evident on account of the fact that the lateral-line sense-organs were located under specialized pigmented scales. The significance of the close similarity between the arrangement of the lateral-line systems on the tail of Necturus, Micrerpeton, and Eumicrerpeton is doubtless of genetic (459) importance, indicating the origin of the caudate Amphibia from the Branchiosauria by a degenerative or recessive evolution in other structural characters. This system of sense-organs has been described in no other branchiosaurian.

The Microsauria (458) are exceedingly interesting in possessing a very peculiar type of lateral-line system. It is known in a few forms and in one specimen especially well (Erpetosaurus tabulatus) (fig. 22, G). In this species, which is represented by a single imperfect skull, there are evidences of a nearly complete lateral-line system of canals and pits. The occipital cross-commissure is represented on the posterior border of the skull by a row of elongate pits such as Andrews described for Ceraterpeton (8). I fail to find in American species the pores described by Andrews. The temporal canal forms with the jugal canal a complete ring, much as it is in Trematosaurus, only in Erpetosaurus tabulatus the temporal canal does not touch the tabulare. I think there are indications of a connection of the temporal canal with the supraorbital. The temporal canal cuts the supratemporal, the squamosal, and jugal. The jugal canal lies for the most part on the supratemporal and quadratojugal, and joins the infraorbital on the jugal. A portion only of the infraorbital canal is preserved. There is also a portion of the supraorbital canal. It seems not to be connected with the temporal canal, although there is a possible indication of this connection. The supraorbital crosses the frontal, prefrontal, and a part of the nasal. The squamosal element is peculiar in Erpetosaurus tabulatus in that it is excluded from the parietal by the extension of the tabulare and postorbital. This condition is found in several other species of the Microsauria. It will be noticed that with the changed condition of the squamosal the temporal canal has changed also, and this is further proof of the close connection between the cranial elements and the lateral-line canals, as Allis has maintained for Amia. (See in this connection C. J. Herrick, Journ. Comp. Neurol., vol. II, p. 224, 1901.)

The Diplocaulia, an amphibian order allied to the Branchiosauria (477) and through them to the Caudata, have the lateral-line system apparently well-developed. The skulls are always crushed flat, so that the canals are nearly obliterated. On the mandible, however, the canals are clearly distinct and apparently run the entire course around the mandibular rami. On a well-preserved skull of Diplocaulus magnicornis Cope there are indications of three lateral-line canals (477, pl. 1). The infraorbital is clearly marked as a well-defined groove just below the orbit. The supraorbital is indicated only for a short distance, and there are indications of the temporal canal. The operculo-mandibular canal has its course, for the most part, near the middle of the rami, but as it approaches the posterior angle of the mandible it suddenly changes its course and drops down to the lower edge, only to rise again and to come out strongly marked near the median plane on the posterior angle of the mandible.

The Temnospondylia, as represented by Eryops, Cricotus, and Archegosaurus, possess well-developed lateral-line canals (458). H. von Meyer (428) many years ago made out the course of the canals in Archegosaurus. The greater part of the following description is based on Eryops megacephalus Cope from the Texas Permian. The entire surface of the cranial elements in Eryops, as in other of the Stegocephala (458), is covered with coarse pits. The fossæ are present even in the bottoms of the grooves which represent the lateral-line system, and are more marked in the members of the Temnospondylia than in the Stereospondylia.

The occipital cross-commissure occurs in a well-developed form in Eryops. It is short and ends abruptly within the limits of the tabulare. Its ends are occupied by large pits. The commissure, as in Amia, grooves the postparietal and the tabulare elements. There is no evidence of an anterior commissure. I think there is evidence of a temporal canal on the left side of the skull, but am not sure. The jugal and infraorbital canals are well developed and strongly connected. The jugal canal starts far back on the supratemporal, and after curving around over the quadratojugal joins the infraorbital, or rather becomes a part of that canal, somewhere on the jugal. There is nothing unusual in the infraorbital. The antorbital commissure is well developed. It is longer and better developed than in any other known form. The supraorbital canal starts in the region of the orbit, and after curving downwards to meet the antorbital commissure, ends abruptly anterior to the nostril. There are faint traces of a lateral-line canal, the operculo-mandibular, on a poorly preserved mandible of Eryops. It does not differ greatly from that described below for Anaschisma.

Although Archegosaurus has been known for more than a century, we have had no adequate discussion of the manner of occurrence of the lateral-line canals. Burmeister (80) gave a figure of the canals as he thought they occurred on the cranium, but H. von Meyer (428) states that the representation is inaccurate, and they seem to be based largely on Trematosaurus.

The lateral-line canals occur in well-developed form on the skulls of the Stereospondylia. The sutures between the elements of the skull are usually clearly marked by smooth, narrow grooves. The lateral-line canals can always be distinguished from the sutural grooves by the shape of the bottom, being U-shaped in the former and V-shaped in the latter. The lateral-line canals also at times have their bottoms roughened by pits occurring in them; the sutural grooves always have smooth bottoms. The lateral-line canals are usually rather shallow and sometimes broad, with the edges of the grooves more or less perpendicular, but in Metoposaurus the canals are deep and the borders are sharply incised.

The temporal canals in Anaschisma from the Triassic (49) of Wyoming are represented by broken furrows. The portions preserved exhibit the usual downward tendency to unite with the infraorbital on the postorbital element. In its course forward from the epiotic the temporal canal cuts the squamosal. The supraorbital canal has an unusually deviating course in Anaschisma, but aside from the minor twists and curves it does not differ essentially from the same canal in other forms. It ends abruptly on the anterior end of the muzzle. In its course it gives off the vestige of an antorbital commissure which tends to join a vestige from the infraorbital canal. The jugal canal begins broadly at the very posterior edge of the skull as though it were continued, as it undoubtedly was, to the body of the animal. In its course forward it joins the infraorbital canal on the jugal. The course of the infraorbital is not unusual in any respect. There is no anterior commissure on the skull, nor is the occipital cross-commissure developed on either skull of the genus at hand.

There are distinct evidences of an operculo-mandibular lateral-line canal on the mandibles. The canal enters the mandible on the surangular and passes forward around the mandible as described for Diplocaulus (477).

Other members of the Stereospondylia, such as Mastodonsaurus, Metoposaurus, and Trematosaurus possess well-developed lateral-line canals, but the above description fits, in a general way, the condition in all genera; and for our present purposes that will suffice.

THE AMPHIBIA OF THE DEVONIAN AND MISSISSIPPIAN OF NORTH AMERICA.

Evidences of the earliest land vertebrates are exceedingly scanty in the strata between the close of the Silurian and the opening of the Coal Measures, being represented solely by footprints. In the Devonian our knowledge of the group is confined to a single footprint, and in the Mississippian to series of footprints from several localities. These have been described by Lea (371), Rogers (Geology of Pennsylvania, pt. II, 1856, p. 831), Barrell (21), Dawson (223), and Branson (50). The last-named author has described a new species from the Mississippian of Giles County, Virginia. His description of the footprints, with a photograph of one of the series, are published herewith (plate 18, fig. 3). Branson (50) has given a résumé of the knowledge of Mississippian Amphibia in North America.

Thinopus antiquus Marsh, 1896.

Type: Specimen No. 784, Yale University Museum.

Horizon: Near top of Chemung, in the upper Devonian.

Fig. 12. Copy of Marsh's drawing of footprint of Thinopus antiquus, from the Devonian of Pennsylvania. × 1/3.

This still remains after nearly 20 years the only evidence of air-breathing vertebrates in the Devonian of the world.

Dromopus aduncus Branson.

Type and other specimens in Oberlin College Museum.

Horizon and type locality: Near the bottom of the Hinton formation in Giles County, Virginia. (Plate 18, fig. 3.)

The following description of the shales and footprints are from Dr. Branson's paper cited above:

A HISTORY OF THE CLASSIFICATION OF THE AMPHIBIA, WITH ESPECIAL REFERENCE TO THE SPECIES FROM THE COAL MEASURES.

It has been necessary, in the course of the present study, to review thoroughly the classifications which have been proposed for the group. A classification of some sort is necessary for the proper grouping of the species which have been recovered from the Coal Measures deposits of this continent, and my reason for publishing this relatively dry material is that the classifications formerly proposed (469), as well as the one here given, may have a proper historical background.

The review of the proposed systems of classification has been much facilitated by the discovery, in the University of Chicago, of some notes by the late Dr. George Baur on the "Stegocephali." The notes were not discovered until after the literature had been pretty thoroughly covered, and it was a source of some gratification, on comparing notes with those of Dr. Baur, to find but few omissions. Whether Dr. Baur had ever contemplated a work on the Stegocephala or not I have been unable to learn, but it is certain that he carefully and laboriously went through the literature on the subject and copied by hand the classifications of each author from 1842 to 1895, together with other notes of interest on the structure, distribution, and phylogeny, including many tracings. The classifications given below are taken, in part, from his notes, although all references have been verified with the original sources.

The first attempt to combine in classification the knowledge of the extinct and recent amphibians was made by Johannes Jacob von Tschudi in 1839 (574). Previous to that time Goldfuss (295) and von Meyer (418) had described various species of salamanders and frogs from the Tertiary deposits of Switzerland, and these Tschudi considered in his following classification:

A. Ranæ.
a. Hylæ.
b. Cystignathi.
c. Ranæ.
d. Ceratophrydes.
e. Bombinatores.
f. Bufones.
g. Pipæ.
B. Cœciliæ.
a. Cœciliæ.
C. Salamandrinæ.
a. Pleurodeles.
b. Salamandræ.
c. Tritones.
d. Tritonides.
D. Protoideæ.

Although the remains of Mastodonsaurus had been known and widely commented on for several years before Tschudi proposed this scheme, he does not include this genus in his classification of the Amphibia, for the reason that for nearly a quarter of a century after the discovery of the labyrinthodonts they were regarded as reptiles, even so eminent an authority as von Meyer (423) including them in his "System der fossilen Saurier." The view that the labyrinthodonts were reptiles was at times disputed, but no one seemed to pay any attention to the argument of Quenstedt in 1850 that "Die Mastodonsaurier im grünen Keupersandstein Würtemburgs sind Batrachier" (527), nor to the contention of Vogt (581) in 1854 that "Archegosaurus und alle Labyrinthodonten sind Amphibien, nicht Reptilien."

In 1842 von Meyer (420) proposed to include all the early forms allied to the Mastodonsaurus in the "Labyrinthodontes." His definition of the group follows:

Labyrinthodontes: Saurier deren Zahn-Struktur jener ähnlich ist, welche in den nach prismatischer Art gebauten Säugethier-Zähnen wahrgenommen wird, u.s.w.

Three years later von Meyer (423) proposed his "System der fossilen Saurier," where the extinct Amphibia are treated as follows:

No other classification was proposed for the extinct Amphibia for 15 years, when Owen (512) in 1859 proposed the new order Ganocephala and retained von Meyer's Labyrinthodontes under Labyrinthodontia. Owen's classification is as follows:

Class Reptilia.
Order I. Ganocephala.
Genera: Archegosaurus, Dendrerpeton, Raniceps.
Order II. Labyrinthodontia.
Genera: Mastodonsaurus, Anisopus, Trematosaurus , Metopias, Capitosaurus,
Zygosaurus, Xestorrhytias.

In his Paleontology published in 1861, Owen gives the same classification, but adds new genera.

Huxley in 1863 (332) did not accept Owen's Ganocephala, but instead proposed the following:

Labyrinthodontia.
A. Archegosauria. Archegosaurus, Pholidogaster.
B. Mastodonsauria. Mastodonsaurus, Labyrinthodon, Capitosaurus, Trematosaurus.

In the same year Dawson proposed (208) the term Microsauria to include the genera Hylonomus, Dendrerpeton, and Hylerpeton, all known from the Carboniferous rocks of Nova Scotia. Two years later Cope proposed the new order Xenorhachia (105) for the reception of the form Amphibamus grandiceps from the Coal Measures of Illinois. He gave as the characters of this order cartilaginous vertebræ and the absence of ribs.

In 1866 Owen proposed (516) the most elaborate and comprehensive scheme of classification which had thus far been offered. His classification is as follows:

Subclass Dipnoa.
Order Ganocephala (extinct). Genera: Archegosaurus, Dendrerpeton, etc.
Order Labyrinthodontia. Genera: Labyrinthodon, Rhombopholis , etc.
Order Batrachia.
Suborder Ophiomorpha. Family: Cœcilidæ.
Suborder Ichthyomorpha. Family: Proteidæ, Salamandridæ.
Suborder Theriomorpha (Anura).
Family 1. Aglossa.
Family 2. Ranidæ.
Family 3. Hylidæ.
Family 4. Bufonidæ.

Haeckel the same year proposed (312) an entirely different scheme of classification nd in some respects more acceptable than Owen's. Haeckel's classification is as follows:

Class—Amphibia.
Subclass I. Phractamphibia.
Ordnung i. Ganocephala.
Genera: Archegosaurus, Dendrerpeton, Raniceps.
Ordnung 2. Labyrinthodonta.
Genera: Baphetes, Zygosaurus, Mastodonsaurus, Trematosaurus, Capitosaurus.
Ordnung 3. Peromela.
Subclass II. Lissamphibia.
Ordnung i. Socobranchia.
Genera: Siren, Proteus, Menobranchus, etc.
Ordnung 2. Sozura (Caudata).
Genera: Cryptobranchus, Triton, Salamandra.
Ordnung 3. Anura (Ecaudata).
Families: Aglossa, Bufonidæ, Ranidæ.

This classification is further elaborated in the edition of 1895.

Cope in 1868 proposed (110) the scheme of classification which was in use for some time, although it has since suffered some change. His classification follows:

Batrachia.
Order 1. Trachystoma.
Order 2. Proteida.
Order 3. Urodela.
Order 4. Gymnophiona.
Order 5. Stegocephali.
Suborder Xenorhachia.
Amphibamus grandiceps Cope.
Suborder Microsauria.
Genera: Pelion Wyman, Hylonomus Dawson, Pariostegus Cope, Dendrerpeton
   Owen, Hylerpeton Owen, Brachydectes Cope, Sauropleura Cope,
   Œstocephalus Cope, Molgophis Cope.
Suborder Labyrinthodontia.
Genera: Dictyocephalus Leidy, Centemodon Lea, Baphetes Owen, Eupelor Cope.

Huxley in 1869 proposed (335) the following classification, which does not differ essentially from that proposed in 1863:

Amphibia.
Order 1. Urodela.
Order 2. Batrachia.
Order 3. Gymnophiona.
Order 4. Labyrinthodontia.
Suborder Archegosauria.
Suborder Mastodonsauria.

The next classification of the extinct Amphibia of any importance was that devised by the committee (450) for the British Association in 1874. This committee was formed of Huxley, Harkness, Henry Woodward, Thompson, and Brigg, with Miall as secretary. This classification is, however, too cumbersome and has never come into general use, and indeed none but English authors have paid it a great deal of attention, although the contribution was a valuable one. The group Aistopoda, which was the ninth group proposed by the committee, has generally been accepted as the group represented by the snake-like forms. The committee's classification follows:

Labyrinthodontia.
Section I. Euglypta.
Genera: Mastodonsaurus, Jaeger; Capitosaurus, Münst.; Pachygonia, Huxley; Trematosaurus,
Braun; Gonioglyptus, Huxley; Metopias, von Meyer; Labyrinthodon, Owen;
Diadetognathus, Miall; Dasyceps, Huxley; Anthracosaurus, Huxley.

Section II. Brachyopina.
Genera: Brachyops, Owen; Micropholis, Huxley; Rhinosaurus, Waldheim; Bothriceps, Huxley.

Section III. Chauliodonta.
Genera: Loxomma, Huxley; Zygosaurus, Eich.; Melosaurus, Meyer.

Section IV. Arthroödonta.
Genera: Batrachiderpeton, Hancock and Atthey; Pteroplax, Hancock and Atthey.

Section V. An uncharacterized group.
Genera: Pholidogaster, Huxley; Ichthyerpeton, Huxley; Pholiderpeton, Huxley.

Section VI. Archegosauria, von Meyer.
Genera: Archegosaurus, Goldfuss.

Section VII. Heleothrepta.
Genera: Lepterpeton, Huxley.

Section VIII. Nectridea.
Genera: Urocordylus, Huxley; Keraterpeton, Huxley.

Section IX. Aistopoda.
Genera: Ophiderpeton, Huxley; Dolichosoma, Huxley.

Section X. Microsauria, Dawson.
Genera: Dendrerpeton, Owen; Hylonomus, Dawson; Hylerpeton, Owen.

The schemes of classification used for the next 10 years did not depart in any appreciable degree from those already given.

In 1875 Cope, in his Check-list of North American Batrachia and Reptilia (120), with a systematic list of the higher groups, published the following classification as being "adopted provisionally by the Smithsonian Institution":

Class Batrachia.
Order Anura. (Anura, Duméril; Salientia, Merrem, Gray.)
Raniformia. (Raniformia, Cope, Nat. Hist. Rev., V, 114, 1865.)
Families: Ranidæ, Colostheidæ.
Firmisternia. (Bufonoid Raniformia, Cope, Jour. Acad. Nat. Sci., Philadelphia, n.s., VI,
190, 1867.)
Families: Dendrobatidæ, Phryniscidæ, Engystomidæ, Breviceptidæ.
Gastrechmia. (Gastrechmia, Cope, Jour. Acad. Nat. Sci., Philadelphia, n.s., VI, 198, 1867.)
Family: Hemisidæ.
Bufoniformia. (Bufoniformia, Duméril et Bibron, partim; Cope, partim.)
Families: Rhinophrynidæ, Bufonidæ, Batrachophrynidæ.
Aglossa.
Family: Pipidæ.
Odontaglossa.
Family: Dactylethridæ.
Arcifera. (Arcifera, Cope, N. H. Rev., V, 104, 1865.)
Families: Cystignathidæ, Hemiphractidæ, Hylidæ, Scaphiopidæ, Pelodytidæ, Asterophrydidæ, Discoglossidæ.
Order Stegocephali. (Stegocephali Cope, Proc. Acad. Nat. Sci., Philadelphia, 1868, 209.)
Labyrinthodontia.
Families: Baphetidæ Cope, Anthracosauridæ Cope.
Ganocephala.
Family: Colosteidæ Cope.
Microsauria.
Families: Phlegethontiidæ Cope, Molgophidæ Cope, Ptyoniidæ Cope, Tuditanidæ Cope,
Peliontidæ Cope.
Order Gymnophiona. (Gymnophiona Müller.)
Family: Cœciliidæ Gray, 1850.
Order Urodela.
Families: Pleurodelidæ Gray, 1858; Salamandridæ: Gray, 1858; Hynobiidæ Cope, 1866;
Desmognathidæ Cope, 1866; Thoriidæ Cope, 1869; Plethodontidæ Cope,
1866; Amblystomidæ Cope, 1866; Menopomidæ (Protonopsidæ Gray,
1850), Amphiumidæ Cope, 1866; Cocytinidæ Cope.
Order Proteida.
Family: Proteidæ Gray, 1850.
Order Trachystomata.
Family: Sirenidæ Gray, 1850.

In 1885 Cope proposed 2 new orders, which he arranges with the other orders already known, as follows:

Batrachia.		Order	I.	Rachitomi.	Order	V.	Urodela.
		Order	II.	Embolomeri.	Order	VI.	Trachystomata.
		Order	III.	Stegocephali.	Order	VII.	Anura.
		Order	IV.	Proteida.

The new order Rachitomi was to include forms like Eryops and the new order Embolomeri was to include forms like Cricotus; the other orders were as they had been given before.

Zittel in 1888 proposed (642) the next classification of any note in his "Handbuch der Paleontologie," where it stands as follows:

Classe Amphibia.
Ordnung 1. Stegocephali.
Unterordnung 1. Lepospondyli.
Familie 1. Branchiosauridæ Fritsch.
"    2. Microsauria Dawson.
"    3. Aistopoda Miall.
Unterordnung 2. Temnospondyli.
Genera: Archegosaurus, Eryops, etc.
Unterordnung 3. Stereospondyli.
Familie 1. Gastrolepidoti.
"    2. Labyrinthodonta.
Ordnung 2. Cœciliæ.
Ordnung 3. Urodela.
Unterordnung 1. Ichthyoidea.
Familie 1. Phanerobranchia.
"    2. Cryptobranchia.
Unterordnung 2. Salamandrina.
Ordnung 4. Anura.
Unterordnung 1. Phaneroglossa.
Familie 1. Ranidæ.
"    2. Bufonidæ.
"    3. Cystignathidæ Cope.
"    4. Pelobatidæ Boul.
"    5. Discoglossidæ Cope.
"    6. Palæobatrachidæ Cope.

Lydekker (393) in the next year proposed a system of classification which did not depart widely from that proposed (450) by the committee of the British Association for 1874. Lydekker's classification is as follows:

Class Amphibia.
Order  I. Labyrinthodontia.
Suborder 1. Branchiosauria.
Family Protritonidæ.
"  Apateonidæ.
Suborder 2. Aistopoda.
Family Dolichosomatidæ.
Suborder 3. Microsauria.
Family Urocordylidæ.
"  Limnerpetidæ.
"  Hyloplesionidæ.
"  Microbrachidæ.
Suborder 4. Labyrinthodontia vera.
Family Archegosauridæ.
"  Diplospondylidæ.
"  Nyraniidæ.
"  Dendrerpetidæ.
"  Anthracosauridæ.
"  Mastodonsauridæ.
Uncertain family, Eosaurus.
Order  II. Apoda.
Order III. Caudata.
Family Hylæobatrachidæ.
"  Sirenidæ.
"  Proteidæ.
"  Amphiumidæ.
"  Salamandridæ.
Order IV. Ecaudata.
Family Discoglossidæ.
"  Pelobatidæ.
"  Palæobatrachidæ.
"  Cystignathidæ.
"  Ranidæ.

In 1890 Döederlein proposed a scheme of classification which is notable on account of the peculiar relations which it expresses between the groups relations which, in reality, do not exist. His classification is as follows:

Class Amphibia.
Ordnung  I. Stegocephali.
A. Microsauria.
Unterordnung 1. Branchiosauri.
Genera: Branchiosaurus, Dawsonia, Melanerpeton, Pelosaurus.
Unterordnung 2. Sauromorphi.
Familie 1. Hylonomidæ.
"    2. Nectridæ.
"    3. Aistopodidæ.

B. Ganocephala.
Unterordnung 1. Rhachitomi.
"      2. Embolomeri.
"      3. Labyrinthodontia.
Ordnung  II. Urodela.
Ordnung III. Gymnophiona.
Ordnung  IV. Anura.

In 1890 Lydekker used the same classification, with minor changes, which he had used in his Paleontology. Credner, who wrote at about the same time (193), followed Zittel's classification. Zittel in 1895 merely repeated his former classification. In 1898 appeared Smith Woodward's Paleontology, where the following scheme is adopted:

Class Batrachia.
Order  I. Stegocephalia.          Order  II. Gymnophiona.
Suborder 1. Branchiosauria.    Order III. Caudata.
"    2. Aistopoda.          Order  IV. Ecaudata.
"    3. Microsauria.
"    4. Labyrinthodontia.

Hay's Catalogue of Fossil Vertebrata of North America contains the next scheme for the classification of the Amphibia which pays especial attention to the extinct forms. His classification is as follows:

Class Batrachia Macartney, 1802.
Order Stegocephali Cope, 1868.
Suborder Microsauria Dawson, 1863.
Family Protritonidæ Lydekker, 1889.
Genera: Amphibamus Cope, Pelion Wym.
Family Molgophidæ Cope, 1875.
Genera: Phlegethontia Cope, Molgophis Cope.
Family Hylonomidæ Fritsch, 1883.
Genera: Hylonomus Dawson, Smilerpeton Dawson, Hylerpeton Owen, Fritschia
Dawson; Brachydectes Cope.
Family Ptyoniidæ Cope, 1875.
Genera: Keraterpeton Huxley, Œstocephalus Cope, Ptyonius Cope,
Ctenerpeton Cope.
Family Tuditanidæ Cope, 1875.
Genera: Tuditanus Cope, Cocytinus Cope.
Family Diplocaulidæ Cope, 1881.
Genus: Diplocaulus Cope.
Lepospondylous Genera of uncertain position: Amblyodon Dawson, Hyphasma
Cope, Eurythorax Cope, Thyrsidium Cope, Pleuroptyx Cope,
Cercariomorphus Cope.
Suborder Apœcospondyli Hay, 1902.
Family Dendrerpetontidæ Fritsch, 1889.
Genera: Dendrerpeton Owen, Baphetes Owen, Platystegos Dawson.
Family Sauropleuridæ Hay, 1902.
Genera: Sauropleura Cope, Leptophractus Cope.
Family Archegosauridæ.
Genera: Trimerorhachis Cope, Dissorophus Cope.
Family Cricotidæ Cope, 1884.
Genera: Cricotus Cope.
Family Anthracosauridæ.
Genus: Eosaurus Marsh.
Family Eryopidæ Cope, 1882.
Genera: Eryops Cope, Ichthycanthus Cope, Zatrachys Cope, Anisodexis Cope,
Acheloma Cope.
Family Mastodonsauridæ Huxley, 1863.
Genera: Mastodonsaurus Jaeger, Eupelor Cope, Pariostegus Cope, Dictyocephalus
Leidy.
Order Urodela.
Genera: Scapherpeton Cope, Hemitrypus Cope.
Order Salientia Laurenti, 1768.
Family Ranidæ.
Genera: Rana Linné, Eobatrachus Marsh.

This classification given by Hay is only for the forms which occur in North America.

CLASSIFICATION OF AMPHIBIA ADOPTED IN THIS WORK, AND A LIST OF THE COAL MEASURES AMPHIBIA FROM NORTH AMERICA.

A few words of explanation will be necessary for an understanding of the following classification. The term Amphibia Linné, 1758, is, according to Stejneger (550), the correct term for the entire group, and this term is adopted. The term Stegocephala, formerly used as a group name for the entire Carboniferous, Permian, and Triassic Amphibia, regardless of structure, has been retained as a third subclass. The choice, so far as priority is concerned, between Labyrinthodontia and Stegocephala, is not easy. The terms, however, imply different structures. The labyrinthodonts proper have stereospondylous vertebræ, while the Stegocephala have either temnospondylous or stereospondylous vertebræ; so the latter term has been adopted.

The ordinal terms are those which have been used previously as subordinal, sectional, or family names, with the exception of the new ordinal term "Diplocaulia" (477). The term Branchiosauria is well-established, and it is here retained with the definitions previously given. The same may be said for the Microsauria, although Dawson first (208) regarded it as a family, though giving the term an ordinal form. The Aistopoda are not entitled to consideration as a group for reasons which are given subsequently. The Temnospondylia and the Stereospondylia, the Embolomeri and the Rachitomi may or may not be good group names, but they have priority, so far as our knowledge of structure goes. They have been retained in their original meanings. They have been variously regarded as sections, superfamilies, groups, and subfamilies.

The following list of species is arranged according to the proposed scheme of classification:

Class Amphibia Linné, 1758. Devonian to Recent.
Subclass Euamphibia Moodie, 1909. Coal Measures to Recent.
Order Branchiosauria Lydekker, 1889. Coal Measures to Permian.
Family Branchiosauridæ Fritsch, 1879.
Micrerpeton caudatum Moodie, Mazon Creek.
Eumicrerpeton parvum Moodie, Mazon Creek.
Mazonerpeton longicaudatum Moodie, Mazon Creek.
Mazonerpeton costatum Moodie, Mazon Creek.
(?) Sparodus sp. indet. Dawson, Nova Scotia.
Order Caudata Duméril, 1806. Coal Measures to Recent.
Suborder Proteida Cope, 1868. Coal Measures, Eocene, and Recent.
Family Cocytinidæ Cope, 1875.
Cocytinus gyrinoides Cope, Linton, Ohio.
Erierpeton branchialis Moodie, Mazon Creek.
Hyphasma lævis Cope, Linton, Ohio.
Order Diplocaulia Moodie, 1912. Coal Measures to Permian.
Family Diplocaulidæ Cope, 1881.
Diplocaulus salamandroides Cope, Salt Fork, Illinois.
(?) Order Salientia Laurenti, 1768. Coal Measures (?) to Recent.
Family Peliontidæ Cope, 1875.
Pelion lyelli Wyman, Linton, Ohio.
Subclass Lepospondylia Zittel, 1887. Coal Measures.
Order Microsauria Dawson, 1863. Coal Measures.
Family Hylonomidæ Fritsch, 1883.
Hylonomus latidens Dawson, Nova Scotia.
Hylonomus lyelli Dawson, Nova Scotia.
Hylonomus multidens Dawson, Nova Scotia.
Hylonomus wymani Dawson, Nova Scotia.
Smilerpeton aciedentatum Dawson, Nova Scotia.
Hylerpeton dawsonii Owen, Nova Scotia.
Hylerpeton intermedium Dawson, Nova Scotia.
Hylerpeton longidentatum Dawson, Nova Scotia.
Fritschia curtidentata Dawson, Nova Scotia.
(?) Amblyodon problematicum Dawson, Nova Scotia.
Family Tuditanidæ Cope, 1875.
Tuditanus punctulatus Cope, Linton, Ohio.
Tuditanus brevirostris Cope, Linton, Ohio.
Tuditanus minimus Moodie, Cannelton, Pa.
Tuditanus longipes Cope, Linton, Ohio.
Tuditanus walcotti Moodie, Linton, Ohio.
Erpetosaurus acutirostris Moodie, Linton, Ohio.
Erpetosaurus minutus Moodie, Cannelton, Pa.
Erpetosaurus obtusus Cope, Linton, Ohio.
Erpetosaurus radiatus Cope, Linton, Ohio.
Erpetosaurus sculptilis Moodie, Cannelton, Pa.
Erpetosaurus tabulatus Cope, Linton, Ohio.
Erpetosaurus tuberculatus Moodie, Linton, Ohio.
Odonterpeton triangularis Moodie, Linton, Ohio.
Family Stegopidæ Moodie, 1909.
Stegops divaricata Cope, Linton, Ohio.
Family Urocordylidæ Lydekker, 1890.
Diceratosaurus punctolineatus Cope, Linton, Ohio.
Diceratosaurus lævis Moodie, Linton, Ohio.
Diceratosaurus robustus Moodie, Linton, Ohio.
Eoserpeton tenuicorne Cope, Linton, Ohio.
Family Amphibamidæ Moodie.
Amphibamus grandiceps Cope, Mazon Creek.
Amphibamus thoracatus Moodie, Mazon Creek.
Cephalerpeton ventriarmatum Moodie, Mazon Creek.
Family Nyraniidæ Lydekker, 1889.
Ichthyerpeton squamosum Moodie. Linton, Ohio.
Cercariomorphus parvisquamis Cope, Linton, Ohio.
Family Ptyoniidæ Cope, 1875.
Ptyonius nummifer Cope, Linton, Ohio.
Ptyonius marshii Cope, Linton, Ohio.
Ptyonius vinchellianus Cope, Linton, Ohio.
Ptyonius pectinatus Cope, Linton, Ohio.
Ptyonius serrula Cope, Linton, Ohio.
Œstocephalus remex Cope, Linton, Ohio.
Œstocephalus rectidens Cope, Linton, Ohio.
Thyrsidium fasciculare Cope, Linton, Ohio.
Family Ichthycanthidæ Moodie.
Ichthycanthus ohiensis Cope, Linton, Ohio.
Ichthycanthus platypus Cope, Linton, Ohio.
Family Molgophidæ Cope, 1875.
Molgophis macrurus Cope, Linton, Ohio.
Molgophis brevicostatus Cope, Linton, Ohio.
Molgophis wheatleyi Cope, Linton, Ohio.
Erpetobrachium mazonensis Moodie, Mazon Creek.
Pleuroptyx clavatus Cope, Linton, Ohio.
Phlegethontia linearis Cope, Linton, Ohio.
Phlegethontia serpens Cope, Linton, Ohio.
Family Sauropleuridæ Hay, 1902.
Sauropleura digitata Cope, Linton, Ohio.
Sauropleura newberryi Cope, Linton, Ohio.
Sauropleura scutellata Newberry, Linton, Ohio.
Sauropleura pauciradiata Cope, Linton, Ohio.
Sauropleura longidentata Moodie, Linton, Ohio.
Sauropleura foveata Cope, Linton, Ohio.
Sauropleura (Anisodexis) enchodus Cope, Linton, Ohio.
Sauropleura sp., Linton, Ohio.
Ctenerpeton alveolatum Cope, Linton, Ohio.
Saurerpeton latithorax Cope, Linton, Ohio.
Leptophractus obsoletus Cope, Linton, Ohio.
Leptophractus dentatus Moodie, Linton, Ohio.
Leptophractus lineolatus Cope, Linton, Ohio.
Order Temnospondylia Zittel, 1887. Coal Measures and Permian.
Family Cricotidæ Cope, 1884.
Spondylerpeton spinatum Moodie, Mazon Creek, Ill.
Family Eryopidæ Cope, 1882.
Eryops sp. indet., Pitcairn, Pa.
Family Macrerpetidæ Moodie, 1909.
Macrerpeton huxleyi (Cope), Linton, Ohio.
Macrerpeton deani Moodie, Linton, Ohio.
Family Anthracosauridæ Cope, 1875.
Eosaurus acadianus Marsh, Nova Scotia.
Eobaphetes kansensis Moodie, Kansas.
Baphetes planiceps Owen, Nova Scotia.
Baphetes minor Dawson, Nova Scotia.
Dendrerpeton acadianum Owen, Nova Scotia.
Dendrerpeton oweni Dawson, Nova Scotia.
(?)Platystegos loricatum Dawson, Nova Scotia.
Genera and species of uncertain position:
Amblyodon problematicum Dawson, Nova Scotia.
Proterpeton gurleyi Moodie, Illinois.
Brachydectes newberryi Cope, Linton, Ohio.
Order Stereospondylia Zittel, 1887. Coal Measures (?) and Triassic.
Family Mastodonsauridæ Huxley, 1863.
Mastodonsaurus sp. indet., Kansas.

The above list of species is interesting in that it shows the diversity of structure among the Coal Measures Amphibia. There are at present 88 species known, many of them incompletely, divided among 17 families and 5 orders. The majority of the species are from the Linton Coal Measures, there being 50 species described or indicated from these beds. The Mazon Creek shales have furnished 10 species; Nova Scotia 18 species; the remainder of the species are from various localities.

DEFINITION OF THE CLASS AMPHIBIA LINNÉ, 1758, DEVONIAN TO RECENT.

(World-wide distribution.)

DEFINITION OF SUBCLASS EUAMPHIBIA, MOODIE, 1909. COAL MEASURES TO RECENT.

(World-wide distribution.)

The present group was established for the reception of the Branchiosauria and their descendants, the Caudata, with the related forms, the Apoda. The Salientia are included provisionally, since there is no evidence of the origin or relationship of this group of animals to other Euamphibia save that they have attained the same stage of evolution. They are in no way closely related to any known group of Amphibia, recent or extinct, but they stand on the same plane of development as the Caudata and present similar structures, i.e., a single ventricle in the heart, external branchiæ in the young, a glandular skin, perichondral bone, and a large parasphenoid. The origin of the Salientia is a puzzle and must remain so until further paleontological evidence is forthcoming. Wyman, Cope, and the writer have all remarked on the similarity of structure between the Salientia and the single known specimen of Pelion lyelli Wyman from the Linton, Ohio, Coal Measures.

The subclass Euamphibia may be defined as follows: Aquatic or terrestrial Amphibia; development by metamorphosis; external branchiæ present in the young; bones almost entirely perichondral; carpus and tarsus never ossified; osseous pubis absent; vertebræ usually amphicœlous with persistent notochord; ribs short and straight or flat and slightly curved, or absent; digits 4 in the hand and 5 in the foot; skull never grooved or pitted, nor cut by the lateral-line canals; lateral-line organs present in the skin; sclerotic plates present or absent; tail long and flattened or absent. Ribs in Triton walthi are secondarily long and curved.

DEFINITION OF THE ORDER BRANCHIOSAURIA LYDEKKER, 1890. COAL MEASURES AND PERMIAN OF NORTH AMERICA AND EUROPE.

Extinct, salamander-like amphibians, with broad, obtusely rounded cranium; external branchiæ present in young; sclerotic plates present; bones of the cranium not ornamented with deep pits and grooves nor cut by the lateral-line canals, though sometimes ornamented with slight scorbiculations; notochord always persistent; vertebræ cartilaginous (in caudal region) or but partially ossified, the ossification being entirely perichondral; a single sacral vertebra; transverse process of vertebræ large in dorsal region; ribs always short, straight, and heavy, present throughout the length of the vertebral column and borne on the transverse processes centrally; caudal region of moderate length with elongate fleshy tail; usually 20 presacrals, of which 4 or 5 may be considered cervicals; limbs natatory and always present, well developed; elements of the appendicular skeleton composed entirely of perichondrium; carpus and tarsus cartilaginous; digits 4 in the hand and 5 in the foot; phalangeal formula for the hand usually 2-2-3-2 and for the foot 2-3-4-3-2; distal phalanges not clawed; abdomen covered with closely packed corneous scutes or scales; body naked or covered with minute horny scales; median and dorsal lateral lines present on the posterior part of body and on tail.

THE AMERICAN COAL MEASURES BRANCHIOSAURIDÆ.

Definition Op the Family Branchiosauridæ Fritsch, 1879. Coal Measures and Permian Op North America and Europe.

Stegocephalic, salamander-like animals, with broad, anteriorly truncate skull. Teeth smooth with large pulp-cavity. The parasphenoid narrowed anteriorly, posteriorly expanded to a shield-shaped plate. Vertebra; with the notochord persistent and intravertebrally expanded. Pelvis well developed, the ilium and ischium osseous with large cartilaginous margins, the pubis unknown, possibly hyaline cartilage. Ribs short, straight, present on almost all vertebræ. Skin with delicately ornamented scales. Eyes with sclerotic plates. Palatal elements toothless or with small tooth-like tubercles on pterygoids and palatines. Ventral armature on throat, chest, and abdomen, extending on to the limbs, consisting of small delicate scutellæ arranged in a chevron pattern.

The above definition is modified from Fritsch (Fauna der Gaskohle, Bd. I, p. 69, 1879).

The North American species are: (?) Sparodus sp. indet. Dawson, Micrerpeton caudatum Moodie, Mazonerpeton longicaudatum Moodie, Mazonerpeton costatum Moodie, Eumicrerpeton parvum Moodie.

Genus MICRERPETON Moodie.

Type Micrerpeton caudatum Moodie.

The genus Micrerpeton, of which the single species is described below, was the first evidence of the occurrence of the Branchiosauria in America. There have been three other genera referred to the Branchiosauria from North American deposits, but there is good evidence that none of them belong there. The genus Amphibamus was originally referred to the Xenorhachia by Cope (105, pp. 134-137) on account of the supposed cartilaginous condition of the vertebræ and the absence of ribs. Later he abandoned this order and placed the form in the Branchiosauria, where it was retained by Zittel (642). Recently Hay has shown (316), and I am able to corroborate his statement, that ribs are present in the species, and that they are long and curved, not at all like the short ribs of the true Branchiosauria. These long, curved ribs undoubtedly exclude Amphibamus from the Branchiosauria and indicate its close affinities with the Microsauria. The genus Pelion has also been referred to this order on purely negative evidence (642, p. 375). This genus is excluded from the Branchiosauria by the well-ossified condition of the limb bones, in which the endochondral ossification is seen to be well developed, a condition not found, so far, among the true Branchiosauria. The form of the head and the elongate hind limb would also tend to exclude the genus from the Branchiosauria. In the Branchiosauria the fore limb is usually larger than the hind limb, the reverse of which is the case in Pelion. The genus Sparodus, as it occurs in North America, is uncertain. It is indicated by remains which are almost impossible of determination.

The genus Micrerpeton may be distinguished from other known Branchiosauria by the large size and anterior position of the orbits, absence of a posterior table to the skull, the short, heavy limb bones, the slender ilium, and the expanded, elongate, and laterally compressed tail. The genus may be defined as follows: Small forms, the known representative attaining a length of less than 2 inches; head broad and short; sclerotic plates present; interorbital space less than the least diameter of the orbit; occiput concave; pineal foramen in the line which cuts the posterior edge of the orbits; teeth small, pleurodont denticles; presacral vertebræ 20 or 21, of which probably 5 are cervical; 1 sacral; ribs short, straight, and heavy, central; scapula ovoid; limbs stumpy and heavy, fore limb exceeding the hind in size; endochondral ossifications distinctly absent; tail long, expanded, and flattened, probably provided with a thin expanded membrane; body covered with minute, ovoid or rounded scales which are ornamented with concentric lines; color markings vertical to the long axis of the body and abundantly present on the tail; lateral-line organs represented by the dorsal and median lateral lines on the tail, the sensory pits probably occurring in specialized darkened scales. Coal Measures of Mazon Creek shales near Morris, Grundy County, Illinois.

Micrerpeton caudatum Moodie.

Type: Specimen No. 12,313, Walker Museum, University of Chicago.

Horizon and locality: Mazon Creek shales, near Morris, Illinois.

The species is represented by very complete remains (plate 2), which are preserved on opposite halves of a nodule. The specimen was collected many years ago by Mr. W. F. E. Gurley at Mazon Creek, but it has never before been studied, although Dr. Newberry examined it and said in a note that Professor Cope should see it. Unfortunately Cope did not see it and it lay unknown for more than a quarter of a century. I am indebted to Dr. Stuart Weller for calling my attention to the specimen, as well as for the privilege of describing it.

The specimen is exceptionally perfect (plate 25, fig. 4). Nearly all the skeletal elements are present, and the general contour of the body, the character of the dermal covering, the color-markings, the lateral-line system, and many other features of interest have been detected. Such completeness of preservation is very uncommon even among the remains obtained from this locality. In this case the entire form was preserved, but the collector, in cracking the nodule, lost the chips containing the feet, so that only portions of the limbs remain. It is thus impossible to determine the phalangeal formula, but the feet were probably like those of Branchiosaurus amblystomus Credner, as given by Credner, to which species the present form is closely allied and indeed must be placed in the same family with Branchiosaurus, Pelosaurus, and Melanerpeton.

MOODIE

PLATE 2

Drawing × 3.5. of type specimen of Micrerpeton caudatum Moodie, from the Coal Measures of Mazon Creek, Illinois, showing skeletal elements, form of head and tail, the lateral-line sense-organs, banded color-markings, and ventral scutellæ. On the edges of the tail impression are indications that in life the tail had a thin fold of skin above and below the fleshy portion, much as in the larvæ of Amblystoma at the present day.

The remains here described represent a small, salamander-like form, and they are among the earliest geological evidence of the group, which, without doubt, gave rise to the modern salamanders. The parts preserved in the specimen are: the complete outline of the head with most of the cranial elements clearly distinguishable, as well as the black pigment of the choroid; the entire vertebral column, including pits in the tail region, where the vertebræ were without doubt entirely cartilaginous; parts of the pectoral girdle; the ilium; the left humerus; the ventral scutellation; the ribs of one side of the body and indications of ribs on the other; portions of both hind limbs; and a complete impression of the fleshy tail. On this impression of the tail are preserved small, horny scales, transverse color-markings, and the distinct impressions of the lateral-line system.

The bones of Micrerpeton caudatum, as in so many of the fossils from this locality, have been replaced by a white, friable mineral which is probably kaolin. The animal is preserved on its back and it is thus illustrated from the ventral side. The entire length of the animal is only 49 mm., of which the tail occupies nearly half.

Fig. 13. Restoration of Micrerpeton caudatum, a branchiosaur from the Coal Measures of Illinois. × 2.

The head has much the same shape as in the species of Branchiosaurus described and figured by Fritsch (251), Credner (181), and Thevenin (568). The eyes occupy relatively the same position as in that genus. The orbits are very large and broadly oval. Within the borders of the rim the stone is blackened as though by the black pigment of the iris, such as Cope has described in Amphibamus. Under a rather high power of magnification the cranial bones are seen to be represented by mere flakes of white mineral matter. The sutures separating the cranial elements are distinctly preserved on the main half of the nodule.

The openings of the skull are five the two orbits, the two minute nostrils, and the pineal foramen. A median suture separates the skull into halves and the pineal foramen lies slightly anterior to the posterior third of its length. The boundaries of the premaxillæ are not distinct, but they are very small elements and form the inner border of the nostrils, which are clearly indicated by bosses of stone. The nasal element is nearly square and lies anterior to the frontal, which it borders broadly. The parietal is about the same size as the frontal and it apparently forms a portion of the inner border of the orbit, although this is not assured. The parietal is elongate and unites posteriorly with the postparietal. The postparietal, with the tabulare and the squamosal, form the posterior boundary of the skull, and they are hence not unlike the same elements in other Stegocephala. The pref rental forms the anterior border of the orbit. The lacrimal has not been detected. The maxilla is elongate and forms the antero-lateral border of the skull. The jugal forms an important element in the lateral border of the cranium and joins the quadratojugal posteriorly. The postfrontal is triangular and with the postorbital forms the posterior border of the orbit. Both of the elements are acuminate posteriorly, although the suture between them is indistinct, and they inclose between their posterior acuminations an anterior projection of the supratemporal. The squamosal has the usual relations and borders the supratemporal laterally. The latter element forms the quadrate angle of the cranium.

The entire length of the vertebral column is preserved, although the nature and structure of its elements can not be determined. The impressions of a few of the vertebræ show that some of the centra were amphicœlous, but other than this nothing is definite. The cavities which the centra occupied were filled by the white mineral matter and the force of the blow which cracked the nodule destroyed the form of the mold. It is possible that where the mineral matter has filled the cavities the centra were osseous or partly cartilaginous, and where the cavities were unfilled the centra were entirely cartilaginous. The length of the vertebral column from the base of the skull to the last impression of a cartilaginous centrum is 33 mm.

The number of centra between the sacral vertebra and the skull is 20 as they are preserved, but there may have been one more, the atlas. Fritsch has represented 21 in his restoration of Branchiosaurus salamandroides, and this is a further indication of an affinity between the two genera, although Credner has represented 26 presacral vertebra; in Branchiosaurus amblystomus. The presacral vertebræ are thus seen to vary within narrow limits, but the number of presacrals is near 20, and this may be taken as typical. It is interesting to notice that in modern forms of the salamanders the presacral vertebræ number about 20. There is but a single sacral centrum in Micrerpeton. The sacral rib has not been detected, but it is restored after the condition found in Branchiosaurus. The right femur partially covers the sacral vertebra, and its structure can not be determined. I count impressions of 17 caudal centra, of which at least 12 may have been partially ossified. In the cervical region there are distinct impressions of transverse processes on at least 5 vertebræ, and this number is assigned to the neck, although it is by no means certain that this is the correct number. The neck was at least short, if we may judge from the position of the remains of the pectoral girdle. No cervical ribs are definitely determined. There is a short rib lying between the fifth and sixth vertebræ, but to which it belongs is uncertain.

There are impressions of 10 ribs preserved on one side of the vertebral column and one on the other side. They are short, straight, and heavy, as are the same elements in Branchiosaurus. This character alone is sufficient to place Micrerpeton among the Branchiosauria, since no such ribs are known in other groups of the extinct amphibians. The ribs preserved lie next the seventh to the seventeenth vertebræ on the left side, and there is one on the right side which may belong to either the fifth or sixth vertebra. They are central in their attachment, and in this they agree well with the mode of rib attachment in the modern salamanders. All of the ribs are single-headed and are composed, for the most part, of perichondral tissue. The position of the ribs in the matrix inclined backwards, and, making a small angle with the vertebral column, is very suggestive of the condition in Branchiosaurus.

The pectoral girdle is represented by three distinct elements of the left side, which are identified as scapula, clavicle, and coracoid, following the nomenclature given by Woodward (631), although Credner (186) would name them otherwise. The scapula is represented by an ovoid fragment lying next to the vertebral column. The clavicle was probably spatulate, as in Melanerpeton, but the inner end of the element is not visible. The coracoid is represented by its outer end only, and its inner pointed extremity is not visible. The interclavicle has not been detected.

The humerus lies somewhat to one side of the pectoral girdle, as if there had been a large amount of epiphyseal cartilage. Its position may be due to post-mortem shifting, but there is little other evidence of any movement after deposition. The humerus is a large, heavy bone in comparison with the rest of the skeleton. It is expanded at each end, and its ends show concavities, proving that the bone is formed principally of perichondral tissue, as would be expected from such an early Branchiosaurian. The endochondrium has not yet developed in this form, which is evidently adult. There is no other element of the arm present.

There is but a single element of the pelvis preserved, a slender elongate rod which is undoubtedly the ilium, since it has the usual position for that element and is much too large for a sacral rib. It has much the same shape as the ilium in the modern Salamandra, and is not expanded as is the ilium of Branchiosaurus. This element, like the humerus, seems to have been but a hollow cylinder of bone and undoubtedly had cartilaginous ends, as in the ilium of the recent Salamandra. The two femora are preserved nearly entire, the right one lying upon and partly obscuring the sacral vertebra. The femur is much more slender than is the humerus, with slightly expanded ends, and, like the humerus, shows the concavities at the ends, indicative of the perichondral character of the tissue composing it. There are two elements of the leg preserved more or less entire. The larger bone may represent the tibia and the smaller the fibula. They both present characters similar to those of the femur and humerus. They are simple rods of bone tapering at the distal end. The feet have been lost, though doubtless they were present at one time.

The ventral surface of the body, as in other members of the Branchiosauria, was covered and protected by a series of small scutes arranged in the regular chevron pattern. The form of the scutes and their number can not be determined. The lines which represent them are, however, distinct. Some of the scutes are missing and some of them are obscured by lying over the vertebral column. They are all somewhat shifted to the left. The lines are very small and close together. I count 16 of them in a distance of 2 mm. In length the longest line preserved is a little more than 4 mm., measuring from the point of the chevron. The lines representing the scutes come to a point in a median ridge which is now represented by a line. The dermal scutes on the abdomen were probably the forerunners of the abdominal ribs of the reptiles (fig. 9).

The impression of the tail contains some of the most interesting features in the entire specimen. Scattered over it and in places laid in mosaic are impressions of small dermal scales, which may have covered the entire body. In form the scales are ovoid, being half as wide as long, and the markings on the scales partake of the nature of radiating lines, much after the pattern of the sculpturing of the cranial bones in the Microsauria. The scales are less than 0.5 mm. in diameter and their character can only be ascertained under high magnification. Near the middle of the tail there are preserved distinct transverse bands of dark color, which are more or less evident throughout the entire tail impression, but they are elsewhere not so distinct as in the central region. The lines are evidently due to rows of pigmented scales, and in all probability the animal's body was transversely striped.

The most interesting and important single structure discovered on the specimen is the impression of the lateral-line system, which is clearly evident as two dark lines on the impression of the fleshy part of the tail. The sense-organs are represented by two longitudinal rows of pigmented scales, one beginning at the tip of the tail, the other taking its origin from the median line somewhat further forward. I am indebted to Dr. Katashi Takahashi for calling my attention to the similarity of this arrangement to that found in the recent Necturus. The arrangement and disposition of the lines containing the sense-organs is practically the same in the two forms. The median lateral-line takes its origin from the extreme tip of the tail and is continued to the base, where the impression is broken. The dorsal lateral-line has its origin rather abruptly from the median lateral-line at a distance of 6 mm. from the tip of the tail. The sense-organs were undoubtedly located beneath specialized pigmented scales on the surface, and to this pigment is due the preservation of the lines.

The fact that the arrangement of the sense-organs of Micrerpeton corresponds so exactly to the condition found in Necturus is of considerable interest. Necturus alone among the modern tailed Amphibia has the arrangement described for the lateral-line system of Micrerpeton. All other forms of the Caudata, as also the larval forms of the Salientia, have an arrangement of the lateral-line system which is perfectly distinct from that found in Necturus, although the basis of the same arrangement is found in all. In Amblystoma, for instance, the median lateral-line is not present on the tail, and the dorsal line is incompletely developed. The close similarity of the arrangement of the systems of sense-organs in the two forms, Micrerpeton and Necturus, may be of genetic significance with regard to the latter form. The lateral-line sense-organs are of a very fundamental significance, and it is not at all improbable that the same arrangement of the lines has existed from the Carboniferous period or earlier. We know that such has been the case in a great many of the fishes. The ancestors of the modern Caudata must have originated somewhere in the basal Carboniferous or earlier periods, and, in the writer's opinion, the Branchiosauria represent the ancestral group of the Caudata. This suggestion is by no means new, since Baur and others have held the same view. This topic has been discussed at length elsewhere (459) by the writer.

The relations of the form Micrerpeton caudatum are readily determined. The number of the presacral vertebræ, the form and position of the ribs, the shape of the skull, the arrangement of the cranial elements, the structure of the pectoral girdle and the character of the ventral armature all clearly bespeak a close relationship with Branchiosaurus, Melanerpeton, Pelosaurus, and other European branchiosaurian forms from the Upper Carboniferous and Lower Permian.

The above-described species, with others given below, is the earliest geological evidence of the Branchiosauria, since the oldest European forms are from the Stephanian (Upper Carboniferous) , which probably lies somewhat above the horizon of the Allegheny series of North America. The presence of the Branchiosauria in America is of considerable interest in the bearing it has on the distribution and migration of the Paleozoic animals. Knowledge of how the group came to occur in such widely separated localities in approximately contemporary geological strata is an unsolved problem of paleontology. It is possible that the piscian ancestors of the Amphibia migrated across or along the borders of the seas and began the amphibian phase of development independently in the two continents. That evolution should, in this case, have followed almost exactly parallel lines seems incredible.

Genus EUMICRERPETON Moodie.

Type: Eumicrerpeton parvum Moodie.

The genus is established on three well-preserved specimens representing nearly the entire anatomy. The generic characters are found in the very broad posterior table of the skull, with its short length, reduction of tympanic notch, and shortness of body. The body-length of Eumicrerpeton (plate 5, fig. 1) is less proportionately than that of other closely allied genera. Other generic characters are found in the sharp postero-lateral angle of the skull, and it is to be distinguished from Micrerpeton, especially, by the short, stumpy limb bones. The narrow, elongate eye, placed close to the edge of the skull, is a character not observed hitherto in the Branchiosauria. The genus is closely allied to Branchiosaurus of Europe.

Eumicrerpeton parvum Moodie.

Type: Specimen No. 803, Yale University Museum. Other specimens, No. 802, Yale University Museum, and No. 4400, U. S. National Museum.

Horizon and locality: Mazon Creek shales, near Morris, Illinois.

The impression of the outline of the entire body is preserved (plate 3, figs. 1 and 2) in three specimens, and in all are found molds and impressions of the alimentary canal, which, in one specimen (471), are remarkably complete and instructive. The three specimens will be discussed separately, since they show different features.

The impression of the larger animal (No. 803, Yale University Museum), which is probably an adult, presents the following elements: the entire skull, both humeri, impressions of posterior and anterior ventral armature, portions of the alimentary canal, one femur, portions of a fibula and tibia, and the entire impression of the tail, on which, as in Micrerpeton caudatum, there occur two definite dark lines, one beginning at the tip of the tail and running obliquely along the tail to where the impression is broken at the anal region; the other beginning at a distance of 4.5 mm. from the tip and running almost parallel with the median line. These two lines undoubtedly represent the lateral-line system.

The skull is especially noted for its shortness and the great posterior width, as well as for the almost entire absence of the tympanic notch. The pineal foramen is located on a line with the posterior border of the orbits. The eyes themselves are narrow and acuminate at each end, with a pronounced convexity inwards and a flattening on the outer margin. They are located on the very border of the skull, but relatively more posterior than in Micrerpeton. No sclerotic plates are evident. The median suture can be indistinctly observed running the entire length of the skull. The sutures bounding the outside of the f rentals and the squamosals are partially evident, but not satisfactorily preserved. The mandible is represented by a mold which in wax impression shows short, stumpy teeth.

Posterior to the skull at a distance of a millimeter there are two sharp impressions which may represent the anterior edges of the interclavicle or they may be branchial elements. They are distinctly curved, however, and probably represent portions of the interclavicle. A wax impression does not show a discrete structure, but the boundaries of some larger element. No other remains of the pectoral girdle can be discerned. The humeri are short and relatively thick. Wax impressions show them to have had truncate or slightly concave ends, thus indicating the absence or slight development of the endochondrium. No other elements of the arm are preserved.

The impression of an elongate femur and the heads of the tibia and fibula of the left side are preserved.

The ventral armature is preserved in two small patches, and these show the chevron-shaped rods to have been very fine much more delicate than in Micrerpeton.

The body impression is very instructive and interesting, both in showing the form of the body and because in it are preserved the larger portions of the alimentary canal. The form of the body can best be discerned by reference to the figures (plate 3, figs. 1 and 2; plate 5, fig. 1).

The portions of the alimentary canal preserved consist of the greater portion of the stomach, three coils or loops of the small intestine, the rectum, and a pit which undoubtedly represents the anal opening. The anus is found at a distance of 16 mm. from the tip of the tail and is somewhat removed from the body portion, as in modern salamanders. On each side of the posterior end of the rectum there occur a pair of enlargements which probably represent the oviducts at their posterior ends (fig. 15, C).

MOODIE

PLATE 3

1. The larger specimen of Eumicrerpeton parvum Moodie. × 1.

2. The smaller specimen of Eumicrerpeton parvum Moodie. × 1.

3. Type specimen of Erpetobrachium mazonensis Moodie. × 1.

4. Type specimen of Erierpeton branchialis Moodie. × 1.

5 and 6. Type specimen of Mazonerpeton longicaudatum Moodie. × 1.

7. A copy of Cope's drawing of the type specimen of Amphibamus grandiceps Cope. The original was destroyed by fire. × 3.

The tail impression is more acuminate than in Micrerpeton, but shows the same structures as in that form, i.e., the lateral lines which have already been mentioned. Micrerpeton was a more rapid swimmer than the present form on account of the greater development of the tail.

The second specimen of the species (No. 802, Yale Museum) shows much the same character as the specimen already described, except that there are impressions of small, blunt teeth on the mandible. The two humeri and the femur of the left side are preserved and the interclavicle is represented by an identical impression, as in the first-described specimen. The tail impression, though similar in form, does not exhibit so much of the structure of the lateral lines (fig. 15, B).

The matter of especial interest in connection with this second specimen is the remarkably perfect preservation of the alimentary canal, which is entire, except for the very anterior end of the œsophagus. The posterior portion of the œsophagus, which measures 3.5 mm., is clearly preserved. Its anterior end is thrown around posteriorly and indicates that this end was loosened after death and became displaced before fossilization. The length preserved may represent the entire œsophagus. The œsophagus is constricted before it enters the stomach, which shows the usual curvature found in modern salamanders. The stomach measures 6 mm. in length by 2 mm. in breadth, and consists of a single enlargement as in the modern Amblystoma punctatum. It increases in size somewhat toward the pyloric end and then very gradually constricts to the pylorus. Three divisions of the small intestine can be seen. The most anterior one, corresponding to the duodenum, is segmented, as though the intestine had been filled with food before interment. The remainder of the intestine, corresponding to the ilium, is looped in the form of two figures 8 which are superimposed, with the upper portions of the 8 at right angles to each other. The rectum is clearly discernible, though its lower portion is somewhat obscured by having the lower part of the upper loop of the intestine lying over it. The anus lies at a distance of 1.5 mm. posterior to the transverse line from, the upper end of the femur, and is quite well back on the tail, as in modern salamanders. In this specimen also occur two oval bodies which may be identified as the lower ends of the oviducts, thus indicating, in all probability, that the animal was a female.

Fig. 14. Mazon Creek Amphibia.

A. Third specimen of Eumicrerpeton parvum Moodie, which exhibits the alimentary canal well preserved. × 2. Original in United States National Museum. a, anus; f, femur; h, humerus; ic, interclavicle; in, intestine; m, mandible; or, orbit; st, stomach; t, tibia and fibula.

B. Type specimen of Amphibamus thoracatus Moodie. Original in United States National Museum. × 2 (For description see P. 132.)

A dissection of several species of modern caudates has resulted in the discovery that the adult condition of the alimentary canal of all the species dissected (Amblystoma punctatum, Necturus maculosus, Diemyctylits torosus, D. viridescens, etc.) is much more complex than that exhibited by the specimen under discussion. A very near approach to the condition found in Eumicrerpeton parvum is found in an immature branchiate individual of Diemyctylits torosus, 56 mm. in length, from a fresh-water pond on Mount Constitution, on Orcas Island, Puget Sound, Washington.

The similarity of the intestinal structure is of considerable importance to our understanding of the relationship existing between the Carboniferous Branchiosauria and the modern Caudata, and only confirms other arguments, offered in another place (459), concerning their immediate relationship.

The branchiosaurian affinities of the present species are almost too evident to need discussion. The entire structure is essentially similar to that of other genera of the order.

The third specimen of this species (No. 4400 of the U. S. National Museum) is almost as perfectly preserved as were the other two specimens. The skull structure, the intermediate position of the pineal foramen, the epiotic notch, and the shape of the skull are essentially similar to the described specimens of the species. The present specimen is more developed than the other two and probably represents an adult. The alimentary canal is perfectly preserved.

It is nearly half again as long as the smallest of the above-described specimens, and the skull is proportionately longer and wider. There is preserved also an impression of the anterior ends of both clavicles. The right humerus is imperfectly preserved, as is also the right femur and tibia; other than these the fossil is merely an impression.

The skull is so similar to those described above that additional description is unnecessary. The pineal foramen is quite large and lies on a line which cuts the orbits into equal longitudinal parts. The interorbital space is about equal to the long diameter of the orbit. Traces of sclerotic plates are observed in the left orbit, but they are quite imperfect.

The alimentary canal (fig. 14a) is unlike the previously described structures, in that the intestine is longer and more convoluted. It lies in five longitudinal folds and ends in an enlarged cloaca, near which there are impressions of two glands, or the posterior ends of the oviducts, as was suggested for the Yale specimens. The creatures undoubtedly fed on small plants and animals much as do the recent salamanders.

MOODIE

PLATE 4

1 and 2. Vertebræ of Spondylerpeton spinatum Moodie. × 1.

3. Type specimen of Mazonerpeton costatum Moodie. × 1.

4. Type skeleton of Cephalerpeton ventriarmatum Moodie. × 0.9.

5 and 6. The halves of the nodule containing a practically complete skeleton of Amphibians grandiceps Cope. × 1.

Originals of above figures in the Yale University Museum.

Genus MAZONERPETON Moodie.

Type: M. longicaudatum Moodie.

The genus is distinguished from other known branchiosaurian genera by the great length of the dorsal region, the elongate tail (plate 5, fig. 2), with its well-developed caudal ribs, the reduction of the tympanic notch, the broad nature of the scapula, the elongate interclavicle, and the slender ilium. The number of dorsal vertebræ is identical with that of Branchiosaurus of Saxony.

Mazonerpeton longicaudatum Moodie.

Type: Specimen No. 795 (1234), Yale University Museum.

Horizon and locality: Mazon Creek shales, near Morris, Illinois. (Plate 3, figs. 5 and 6.)

The remains consist of the following elements: an incomplete skull; nearly the entire vertebral column, consisting of cervical, dorsal, sacral, and caudal vertebræ, 36 in number; several ribs preserved on each side of the vertebral column; a portion of the ventral armature; the scapulæ; a clavicle; the interclavicle; both humeri; the radius and ulna of one side and the ulna of the other; portions of both hands; the ilium of the right side; both femora, and a partial impression of the left tibia.

The skull is, unfortunately, very poorly preserved. Enough remains, however, to determine the essential characters. The skull bones, unlike any other American branchiosaurian, have an ornamentation consisting of sharp pits and elevations which in places have a quincuncial arrangement and in others take the form of definite lines of pits or tubercles similar to the condition found in many of the Microsauria. The orbits are large and are situated back of the median transverse line of the skull. They are almost circular in form and contain 6 elongated sclerotic plates very closely arranged around the borders of the right orbit. The plates are twice as long as wide. The interorbital width is 1.25 times the transverse diameter of the orbit.

Not many of the sutures of the skull are discernible. Portions of the frontals, the nasals, the prefrontals, the parietals, and the supratemporals can be identified. Their arrangement is shown in figure 14a. There is a decided posterior table to the skull, with truncate posterior border. The tympanic notch is shallow, with its outer border not so well protected as in Branchiosaurus.

The cervical vertebræ are incomplete, but their number was 4 or 5, as in Micrerpeton. The structure of the dorsal vertebræ is also uncertain, although the shape can be discerned. The vertebræ are short and thick, very unlike the long, cylindrical vertebræ of Cephalerpeton. The heavy transverse process is quite evident on the best preserved vertebræ. This process recalls that described by Credner for the Saxony Branchiosauria. Several of the vertebræ show the articulation of the ribs with this process. The ribs of the caudal region recall very strongly those of Branchiosaurus. They are quite heavy in the anterior caudal region and then diminish rather rapidly to the point where the tail is broken and lost.

The ventral armature is represented by a patch of chevron rods 21 mm. in length. The rods take a very peculiar form, being short crescentic bundles of fine rods, hair-like in appearance. In one of these bundles I count 5 smaller rods. The bundles are arranged in rows similar to the pattern so characteristic of the Carboniferous Amphibia, as described elsewhere. The patch of ventral armature preserved belongs to the abdominal region. A single row of the crescentic bundles measures 11 mm.

Both scapulæ are preserved in their entire form. They are quite different from those of any other genus, being broadly crescentic with a posterior concavity and an anterior protuberance. The anterior surface of both scapulæ is obscured. Vascular foramina occur near the base of both scapulæ; there being three of them in the right scapula, arranged in the form of an isosceles triangle. The morphology of these foramina is uncertain. They have never been observed among the Carboniferous Amphibia, and, so far as I am aware, they are entirely unknown among higher vertebrates.

The temnospondylous Amphibia of the Carboniferous and Permian possess, in the coössified scapula-coracoid, three foramina, very similar to the ones in the present form, but they are confined to the coracoidal region and, in the Branchiosauria, as identified by Credner, the coracoid is a free element, although I have never been sure of its identity among American forms. Williston has called these foramina the glenoid, the supraglenoid, and the supracoracoid foramina (Journal Geology, XVII, No. 7). They are not, however, to be correlated with the three foramina above mentioned, since in the Temnospondylia the foramina belong with the coracoid and not with the scapula. The condition of the Temnospondylia occurs in the bony fish Xiphactinus audax Leidy, and an analogous condition obtains in the reptiles, as in the Mosasaurs and Dinosaurs.

Fig. 14a. Skeleton of Mazonerpeton longicaudatum Moodie. c, carpus; cl, clavicle; cr, caudal ribs; cv, caudal vertebræ; h, humerus; f, femur; or, orbit; r, radius; sp, sclerotic plates; sc, scapula; u, ulna: vs, ventral scutellæ. From Mazon Creek. Original in Yale University Museum.

Near the outer edge of the right scapula there is a large fragment preserved, which, I think, must be the misplaced clavicle. It is obscurely triangular or, more exactly, spatulate. The interclavicle is represented by fragments only, and seems to have had a narrow form.

The humeri recall those of Micrerpeton. They are somewhat elongate and apparently cylindrical in their normal condition, although somewhat flattened in the fossil. The shaft is considerably constricted at the middle and the ends are expanded, in which expansion the lower end exceeds. The ends are abruptly truncate, indicating a small amount of endochondral ossification or its entire absence.

The mesopodial elements, unlike what is described for Cephalerpeton, are quite dissimilar in form, recalling the condition in Mesosaurus brasiliensis McGregor. The larger element is, apparently, the ulna. It has the lower end greatly expanded and the shaft is curved outward, resembling very much a reptilian ulna. The radius is much smaller than the ulna, lacks the lower expansion, and is shorter by 1 mm.

The carpus is represented merely by a blank space, with evidences of impressions of cartilage in the sandstone. The hand of the right side contains 4 phalanges. There are 2 phalanges preserved in the first digit, including a sharp-pointed terminal phalanx, and the second digit has only the metacarpal. The third has the metacarpal and the first phalanx, which does not differ in form, but only in size, from the metacarpal. The fourth digit contains only the metacarpal. Of the left hand there are portions of 3 digits preserved, including 3 metacarpals and a phalanx, which in structure are not different from those of the right hand.

The ilium of the left side is preserved, apparently entire. It is elongate and cylindrical, its upper end adjoining the twenty-eighth vertebra. The head of the femur lies close to the lower end of the ilium, so that that element must have been suspended in the flesh, much as in modern salamanders. It could not have been of much use as a support for the body. The form of the femur is not unlike that described for the humerus, save that its lower end is smaller than the upper, while in the humerus the extremities are of equal diameter. A portion of the right femur is preserved, extending in an opposite direction to the left.

Mazonerpeton costatum Moodie.

The remains on which the present species is based are inclosed in a much fractured nodule. The parts of the animal which have been identified are: a part of the skull and left mandible, two clavicles, a humerus, impressions of several vertebræ, a portion of the dorsal region of the body with several ribs, two portions of the caudal region with several ribs and unidentified fragments. (Plate 4, fig. 3.)

The animal, from the shape and form of the ribs, is undoubtedly a branchiosaurian, since short, heavy, straight ribs have not yet been found to be associated with other than branchiosaurian structures. It is placed in the genus Mazonerpeton on account of the structure of the pectoral elements, the form of the humerus, and the length of the tail, all of which agree in structure with Mazonerpeton longicaudatum. The animal attained, perhaps, a length of 4.5 inches, while that of M. longicaudatum was about 3 inches. The tail in the present species is very long and slender, more elongate than in any other known branchiosaurian.

Fig. 14b. Skeleton of Mazonerpeton costatum Moodie. × 1.5. Original in Yale University Museum, dv, dorsal vertebra; ch, neutral spines; cl, clavicle; cv, caudal vertebræ; f, femur: h, humerus; m, mandible; rb, rib; sk, skull; v, vertebræ.

The part of the skull preserved is very unsatisfactory and, aside from the fact that it seems to represent the under side of the left half of the skull, little can be said. Three sutures can be observed, but what sutures they are is undetermined. The left mandible lies crushed on the edge of the skull and partially obscures what little there is of that structure. The slightly curved impression, from which the bone has been either broken or weathered, measures 13 mm. in length by 3 mm. in posterior diameter by 1 mm. in anterior diameter. These measurements show the element to have been slender and pointed anteriorly.

Very little accurate information can be derived from the study of the vertebral column of the specimen, nor can the dorsal vertebral formula be made out, since only a portion of the length of that region is preserved and only a few rather indefinite impressions can be discerned. These impressions show the vertebræ to be short and higher than in most Branchiosauria.

The caudal series is represented by two sections, one of which is, apparently, from near the base of the tail, judging from the size of the caudal ribs preserved; the other is from near the tip of the tail, and shows the constituents to have been long and slender. Ribs are apparently absent on this section. The position of the two caudal sections shows that when the animal died it was coiled up much like a snake, so that in the fractured nodule three sections of the body are visible. The tail was probably half as long again as the body.

The ribs throughout the body are short, heavy, and straight, with, in the dorsal series, a lateral and a distal expansion which is taken as a distinctive specific character. Judging from imperfect impressions in the dorsal series, the ribs were attached to a transverse process of the centrum, thus agreeing with other branchiosaurians in this respect. The ribs show a progressive decrease in length from the cervical region to the point of their disappearance on the tail.

MOODIE

PLATE 5

1. A reconstruction of the possible appearance in life of the Coal Measures branchiosaurian, Eumicrerpeton parvum Moodie, a small, primitive salamander, less than 2 inches in length, based on three specimens from the Mazon Creek Shales. The lateral-line organs are represented as dark bands on the tail, the sense-organs being, apparently, situated beneath specialized pigmented scales, to which is due the preservation of the lines.

2. A restoration, natural size, of the branchiosaurian, Mazonerpeton, based on two specimens. The form of the animal is quite salamander-like. It is shown when about to feed on a specimen of Acanthotelson stimpsoni, which is said to be a brackish-water crustacean. The branchiosaur and crustacean may possibly have inhabited the same body of water.

The pectoral girdle is represented by two elements, one of which is certainly the right clavicle, and the other is possibly the left clavicle, though its form is somewhat distorted by pressure. Both elements are in the form of an elongate spatula with the dorsal surface greatly concave and the inner end acuminate.

The right humerus is imperfectly preserved, though the impression allows one to gain an exact idea of its form. It lies under the right clavicle. Its ends are truncate with a contracted shaft and expanded extremities; the bone was apparently hollow.

Fig. 15.

A. Impression of Erierpeton branchialis Moodie. bb, basibranchial; hyp, hypohyals; m, mandible; d, body impression. × 3.

B. Eumicrerpeton parvum Moodie. a, anus; f, femur; h, humerus; in, intestine; l, liver; st, stomach; r, radius; u, ulna. × 3.3.

C. Larger specimen of Eumicrerpeton parvum Moodie. a, anus; d, dorsal lateral-line; h, humerus; in, intestine; ml, median lateral-line; st, stomach. × 2.6.

D. Skeleton of Erpetobrachium mazonensis Moodie. cl, clavicle; h, humerus; r, radius; sc, scapula; u, ulna. × 2.

E. Rib of Mazonerpeton costatum Moodie. × 2.5. Originals in the Yale University Museum.

In another nodule (No. 804, Yale Museum) there is a single bone preserved which resembles, to a great extent, a rib of the present species (fig. 15, E), although somewhat larger, and it has been provisionally identified as such. The element is very slightly curved, but shows the expanded head of the rib of this species.

Sparodus sp. (?).

Type: Specimen in the Peter Redpath Museum of McGill University.

Horizon and locality: Coal formation at the South Joggins, Nova Scotia.

The material on which the above determination is based was collected in 1878 by Sir J. W. Dawson in the coal formation at the South Joggins, Nova Scotia. Nothing has been collected since that date that would give additional information as to the nature of the form represented. I give here Dawson's description of the remains:

Fig. 15a. Type material of Sparodus, consisting of a, a tooth (× 25); b, four of the smaller teeth (in maxilla?) (× 25); c, three bony scales (× 5); d, fragment of a limb bone (× 2); e, a vertebra (× 2). (After Dawson.)

"In the coaly matter or mineral charcoal at the base of tree No. 10 appeared a few fragments of an animal which may possibly belong to the above-named genus of Fritsch, though I am by no means certain of this identification or of the real nature of the animal.

"The skull is represented by a fragment of a maxillary or intermaxillary bone, with blunt conical teeth. It is smooth or marked merely with microscopic dots. There is also a fragment which may be a palatal bone studded with minute teeth.

"A few vertebræ associated with the above bones are long and narrow, with large zygapophyses and long neural spines. Length of body (i.e., of the vertebra) about 3 millimeters.

"With these remains are a few bony scales different from those of any other species found in these trees, and more resembling scales of Ganoid Fishes. They are somewhat rectangular in form, enameled on the surface and beautifully sculptured with waving lines.

"In the same trunk were found some teeth and bones referable to Hylerpeton dawsoni, and it is not impossible that the remains above referred to may have belonged to some creature devoured by that animal, and which would not otherwise have obtained admission to the interior of an erect tree. The tree itself had been removed by the sea, all but a little of the base, and this was in a very unsatisfactory state, so that doubt might even exist as to the limit between the deposit in the interior of the tree and that under its base."

ORDER CAUDATA DUMÉRIL, 1806. COAL MEASURES TO RECENT.

Suborder PROTEIDA Cope, 1868.

This order agrees generally with the Caudata, but presents one most important feature of difference in the presence of the opisthotic. It is this point which gives the Proteida its intermediate position between the extinct amphibians and the recent species, and seems to indicate a connecting line from the Coal Measures down to the present. The structure of the hyobranchial arches sustains this view.

The hyoid apparatus differs from that of other adult Caudata and resembles that of their larvæ in having three epibranchials, instead of one only. The second basibranchial is also connected with the first, which is not the case with the other Caudata. Three extinct genera are placed tentatively in this suborder.

Family COCYTINIDÆ Cope, 1875.

The present family, as here defined, includes the forms whose structure seems to ally them with the modern salamanders. The character on which most dependence is placed is that of the branchial apparatus, lacking in Hyphasma. The forms are all incompletely known and the family will doubtless require revision on acquisition of additional material. Three genera, each with a single species, are:

Genus COCYTINUS Cope, 1871.

Type: Cocytinus gyrinoides Cope.

Vertebræ and ribs osseous; teeth on the premaxillary bone, none on the maxillary; hyoid elements largely developed, an axialhyal with basihyal on each side, closely united with the corresponding ceratohyal, at the end of which is an element in the position of a stylohyal; hæmal or basibranchials 3, the anterior 2, each supporting i pleural branchihyal, and the third supporting one also, the first hæmal branchihyal on the inner side of the ceratohyal, approaching the median line, and with elongate pleural element.

The reference by Cope of this genus to the Caudata is one of the most interesting facts connected with the Paleozoic Amphibia. He says: "The present genus is, then, to be referred to the neighborhood of Amphiuma and Protonopsis, but forming the type of another family" (123). He regards the branchial apparatus as being more fish-like than that of any of the modern genera. It is possible that Cocytinus gyrinoides was a larval branchiate and consequently aquatic form. It should be more fully compared with Erierpeton branchialis from the Mazon Creek shales when better known, as well as with Hyphasma lævis from the Linton locality.

All three of these forms are included, provisionally, under the Cocytinidæ.

Fig. 16. Obverse of Cocytinus gyrinoides Cope, from the Coal Measures of Ohio. × 2. pmx, premaxillæ: mx, maxilla; m, mandible; ah, axialhyal; h, basal branchihyal; ch, ceratohyal; bv, hæmal branchihyal; b, bII, bIII, bIIII, pleural branchihyals.

Cocytinus gyrinoides Cope.

Type: Specimen No. 8613 G, American Museum of Natural History.

Horizon and locality: Linton, Ohio, Coal Measures.

Two specimens of this interesting amphibian are known, one of them fairly complete (No. 2564, Am. Mus. Nat. Hist.). The type specimen consists of the inferior bones of the cranium in a fairly complete state of preservation, with the muzzle and its teeth; also 8 anterior vertebræ, with their short recurved ribs.

The condition of the hyal elements in the type specimen is as follows: the hæmal elements of the first branchial arch are partially concealed on both sides of the ceratohyal. An expanded truncate face for attachment to the axial element is visible on both sides, but the body of the bone is flat and presents the edge of the specimen.

The first pleural element proceeds from just behind the axialhyal; it is longer than the other pleural elements. A slender bone is visible extending from the space between the ceratohyal and mandibular angle; it may, therefore, pertain to the suspensorium of the jaw as well as to that of the hyiod arch, or be squamosal as well as stylohyal. The second hæmal bone is slender, but with an enlarged axial extremity; that of the right side is not so well preserved as to be safely determined. The third hæmal elements are the smallest, and originate immediately in front of the occipital condyles and diverge outwards and backwards. They are little curved, subcylindric, and slightly expanded at the extremities.

Of the pleural elements the first and second are little curved and the first is marked by a pit or foramen on the under side near the distal end, which is clearly visible on both sides of the specimen. The third and fourth pleurals are more curved and the outer ends slightly expanded and directed backwards.

MOODIE

PLATE 6

Dendrerpeton acadianum Owen. Mandibles, parts of anterior extremities, humerus, etc. Nearly natural size. Erect tree, Coal formation South Joggins, Nova Scotia. Photograph by Dawson, published through the courtesy of Dr. Arthur Willey. Original specimen in the British Museum of Natural History.

The obverse of the specimen (fig. 16) shows that the anterior axialhyal is wedge-shaped. The lateral basihyals are massive. The second hæmal branchihyal is dilated, fan-shaped distally, and supports two pleural elements. The muzzle projects over the lower jaw and was rather broadly truncate. The premaxillary teeth are cylindric and 6 in number on each side. The maxillary bone is represented by a lamina at each lateral extremity of the premaxillary. The mandibular rami are very stout, as are also the ceratohyals. The vertebræ have possessed some apophyses, apparently keel-shaped diapophyses. The ribs are slightly curved.

The other specimen of this species (fig. 16a) is interesting in having 40 consecutive vertebræ preserved, and 19 pairs of ribs attached in their natural relations to the skull and hyal elements. There are a few hyal elements preserved, but nothing is added to our previous knowledge. The ribs are quite as in the type specimen, as are also the vertebræ. The animal was apparently a slender, eel-shaped amphibian comparing favorably with the modern Amphiuma in this respect. There are no indications of limbs or limb girdles.

Fig. 16a. Nearly complete specimen of Cocytinus gyrinoides Cope, from the Coal Measures of Linton, Ohio. Original in the American Museum of Natural History. × 0.95.

Genus ERIERPETON Moodie.

Type: Erierpeton branchialis Moodie.

The generic characters are found, first of all, in the presence of hyobranchial arches which indicate its relationship to the formerly described Cocytinus gyrinoides Cope, from Ohio. The only other known extinct genera of Caudata which possess, or at least have preserved, the hyobranchial arches are the Jurassic Hylæobatrachus from Belgium and Lysorophus from the Permian of Texas. The present form is widely distinct from both of these genera in the shape of the mandible and the form and arrangement of the hyobranchial bars. The genus Erierpeton finds its closest ally in Cocytinus, in the family Cocytinidæ, which possibly belongs in the order Caudata and the suborder Proteida of Cope.

Erierpeton branchialis Moodie.

Type: Specimen No. 801 (222) 5, Yale University, Museum.

Horizon and locality: Mazon Creek shales, near Morris, Illinois.

The amphibian remains designated by the above name consist of a distinct mandible and some rather indefinite body impressions (plate 3, fig. 4). Three elongate impressions occur between the rami of the mandibles (fig. 15, A), which, I suppose, must represent hyoid bones belonging to the hyobranchial arches. The lateral elements are paired and the median impression is straight and lies between the paired impressions of the hyoids. The paired portions probably represent the hypohyals or hypohyals plus the ceratohyals, and the unpaired portion of the first basibranchial, according to the nomenclature of Wiedersheim (Comparative Anatomy of Vertebrates, 1897, p. 86). If the impressions have been correctly interpreted the present specimen is of very great interest, since it is the first evidence we have of the hyobranchial arches in the Amphibia of Mazon Creek, and the second in the Carboniferous of North America. Dawson doubtfully identified (216) some elements of the Joggins Amphibia as hyoids, but was uncertain as to their position. Cope described fully the well-developed hyobranchial apparatus of Cocytinus gyrinoides (123) from the Coal Measures of Ohio. Among other Paleozoic Amphibia Williston (614) has described branchial arches in the peculiar form Lysorophus tricarinatus Cope, from the Permian of Texas.

The form of the impression of the mandible in the present specimen is unlike anything known to the writer among other Carboniferous or later Amphibia. The rami are long, slender, deep, slightly curved, and pointed anteriorly. The anterior symphysis was not a complete sutural union, but was occupied partly by cartilage or other connective tissue.

There are no definite traces of the appendicular skeleton. The traces of the body (fig. 15, A) indicate an elongated, rather slender animal, but further than this nothing can be said in regard to its structure.

The occurrence of a typically caudate form in the Carboniferous is unusual and complicates still further our understanding of the origin and relationships of the early Amphibia.

Genus HYPHASMA Cope, 1875.

Type: Hyphasma lævis Cope.

"Vertebræ osseous, the posterior dorsals, and probably the caudals, furnished with fan-like neural spines; limbs unknown—(?) wanting. Thoracic shields present. Ventral armature, consisting of rhomboidal scuta, forming packed rows arranged in chevrons, directed backwards, on top of which are the usual rod-like scales arranged in packed chevrons, with the angle directed forward.

"The general appearance of the type of this genus is that of a Ptyonius, but the ventral armature is different from anything observed in the known genera of this group. The larger external scuta are like those of the species of Colosteus (Sauropleura) , but their series have a different direction. The inner chevrons are those of many other genera" (123).

Hyphasma lævis Cope.

Type: Specimen No. 9023 (in counterpart), American Museum of Natural History.

Horizon and locality: Linton, Ohio, Coal Measures.

The specimen is very indistinctly preserved and the characters given by Cope can not all be made out. It is puzzling to see just on what he bases his conclusion. It is possible that the specimen is a poorly preserved Ptyonius. The outlines of the vertebræ are so indistinct that I am uncertain about them. In certain lights there appear to be regular impressions which resemble the spines of the vertebræ of Ptyonius, but they are doubtful. The skull appears totally distinct from any known species of Ptyonius, but it is very imperfect. The condition of the pectoral elements is very uncertain and I can not be sure that what Cope described as thoracic "scuta" are such. The interclavicle, however, is clearly preserved as a diamond-shaped structure. It is almost smooth, with a few faint radiating lines near the base. It measures 5 mm. in greatest breadth by 8 mm. in length.

DEFINITION OF THE ORDER SALIENTIA, LAURENTI, 1768.

(World-wide distribution.)

The order is suggested in the Coal Measures by a single species, known from a single poorly preserved specimen (plate 24, fig. 1). The form Pelion lyelli Wyman was the first known of the Linton Amphibia, and its striking frog-like (123, 639) appearance was early noticed. There is no assurance that the species belongs with this order, but since a well-developed and highly specialized frog (480, 481) occurs in the Como Beds (405) of Wyoming, it is not impossible that we may have in the Pelion lyelli a suggestion (460), at least, of the ancestral structure. It is certain that the frogs have, in past ages, had a much greater length of vertebral column than they possess at present, as is witnessed by the coalescence of several vertebræ to form the urostyle. It is suggested that the ancestral vertebral column is represented in Pelion.

Family PELIONTIDÆ Cope, 1875.

The present family includes but a single species, that of Pelion lyelli Wyman, first described in 1858 (640), from Linton, Ohio.

The family characters are to be found in the broad and obtusely rounded cranium, in the frog-like scapular arch, the frog-like hind limb, and in the form of the palate, so far as these structures have been preserved.

It has been suggested that the present form shows decided affinities with the frogs of to-day and it may possibly be looked upon as the actual ancestor of the living frogs. The length of the vertebral column would seem to militate against such a relationship, since it is well known that frogs have had a short vertebral column since the Jurassic (480, 481). But this is not a good argument, since the developing urostyles of modern tadpoles show metameric fenestrations in the developing bone which doubtless correspond to openings between the vertebræ. The notochord of the tail is segmented, apparently through the influence of former vertebral structure. At any rate, the suggestion is an interesting one and, whether sustained or disproven, the present discussion is based on the probabilities of the case.

Genus PELION Wyman, 1868.

Type: Pelion lyelli Wyman.

MOODIE

PLATE 7

Hylerpeton dawsoni Owen. Above: Mandible, teeth, rib, and bones of extremity. Below: Bones of pelvis and posterior limb and bony scales. Nearly natural size. Erect tree, Coal formation, South Joggins, Nova Scotia. Photograph by Dawson, published through the courtesy of Dr. Arthur Willey. Original specimens in Peter Redpath Museum of McGill University.

Pelion lyelli Wyman.

Type: Specimen No. 7909 G, American Museum of Natural History.

Horizon and locality: Linton, Ohio, Coal Measures. (Plate 24, fig. 1.)

This was the first species described from the Linton, Ohio, deposits. It was made known by Dr. Wyman in 1857 at the meeting of the American Association for the Advancement of Science for that year. The species was subsequently studied by Cope. He merely confirmed Wyman's observations. The following description is based on the descriptions of Cope and Wyman and on my own study of the type specimen.

This is the most frog-like, in appearance at least, of all the Amphibia which have so far been discovered in the Carboniferous. The skull especially has a shape which is strikingly frog-like, and the long hind limbs lend further likeness to the tailless forms. Pelion may have been a jumping creature, if we may judge from its long hind legs. Wyman and Cope have both called attention to the frog-like appearance of the specimen, and this is apparent at the first glance. It is probable that the resemblance has some significance as to the ancestry of the Salientia, and it may indicate the first step in the origin of the tailless Amphibia. It is possible that the frogs began to be separated from the other Amphibia during the Carboniferous. The first frogs we know are from the Jurassic, where they are well-developed ranids. If Pelion be a frog ancestor, then the history of the group from the Coal Measures to Jurassic is an unknown story.

The specimen is preserved on its back and it is thus impossible to tell as to the structure of the skull. Cope was of the opinion that the depressed areas on the sides of the elongate parasphenoid were the orbits, and if so the resemblance to the frogs is much more striking. In the frogs there is a strong process from the pterygoid which projects inward to meet a corresponding process from the parasphenoid. This forms a heavy rod behind the palatine vacuity. There is a heavy rod represented in the specimen and a part of it is certainly the external process of the parasphenoid, but whether it is to be interpreted as in the frog is an open question. The outline of the cranium is partially obscured by the mandibles, but the anterior part is represented by a raised line, as shown in figure 17. In the anterior part of this space there are two ridges which may be tooth ridges. If they are teeth there is a great similarity to the premaxillary and vomerine teeth of Necturus, since the ridges are widely separated at the median line and approximated distally, as they are in Necturus. The mandible is preserved entire and its form is strikingly frog-like. Its posterior angles project over the quadrate area and seem to have had an upturned projection such as is found in the mandibles of the Crocodilia.

There are impressions of 20 vertebræ preserved, and they cover a little more than half of the presacral region. There may have been 28 to 30 presacrals. The vertebræ, as preserved, are somewhat quadrate in outline and constricted at the middle, as though they were of the typical microsaurian type. No ribs are preserved.

Fig. 17. Pelion lyelli Wyman, an amphibian from the Coal Measures of Ohio, the supposed ancestral salientian. (After Wyman.) × 0.75. pmx, premaxilla; pv, palatine vacuity; m, mandible; sc, scapula-coracoid; h, humerus; r-u, radius and ulna; f, femur; t, tibia.

There is an impression anterior to the right humerus which may represent a part of the pectoral girdle, but its form is so obscure that it can not be determined. The pelvic girdle is entirely wanting in the specimen. Remains of the fore and hind limbs are preserved. The arms are especially well preserved and consist of a strong humerus, a separate radius and ulna, and phalanges, the carpus having undoubtedly been cartilaginous, since there are no traces of carpal bones. Wyman has figured a small ossicle (fig. 17) which might be interpreted as a carpal, but it is further removed from the carpal region than his figure shows and I would interpret it as a fragment of a phalange, since the first digit seems to be turned aside over the vertebral column. The right hand is but imperfectly preserved, but the left hand is nearly entire. There are evidences of 4 digits, possibly 5. The metacarpals are elongate and rather stout. The phalanges of the distal series have been lost, so the phalangeal formula can not be determined. On the whole, the hand has a very broad aspect and is not at all slender, as in the majority of the microsaurians from the Coal Measures of Ohio. It resembles in a great measure the broad hand of a toad and may thus be indicative of a terrestrial life. The humerus is well developed and has pronounced swellings, as though for the attachment of strong muscles. These indications would favor the view of the animal being a land dweller.

The femur and a part of the tibia (?) of the right side are all there is preserved of the hind limb. These elements show the leg to have been quite long, though weaker than the fore limb. The femur has a large distal articular surface. The fibula is, apparently, absent, though it may simply be lost.

The genus Pelion stands alone among the Carboniferous Amphibia. The form can not be placed in the order Branchiosauria on account of the well-developed limb bones and the large mandible. It may belong with the Microsauria. I have placed it under the Salientia in the hope of learning more about the early relatives of the tailless forms. There is no assurance at all that it is even ancestral to the Salientia, but the resemblances are striking.

The following gives the measurements of the type specimens:

SUBCLASS LEPOSPONDYLIA ZITTEL, 1887. COAL MEASURES TO PERMIAN.

(Europe and North America.)

The group is here defined according to the English edition of Zittel's Text Book of Paleontology, 1902, p. 125. "Notochord persistent and enclosed in constricted bony cylinders, hour-glass-shaped in longitudinal section. Teeth simple, conical, hollow." According to Zittel there are two families, the Microsauridæ and the Aistopodidæ. The latter family is dealt with under Aistopoda (p. 76) and it is there shown that the group is in no wise a valid one. The former family is regarded as an order and is fully entitled to that rank. As defined here the subclass Lepospondylia contains but a single order, the Microsauria.

Extinct, terrestrial, aquatic, or semi-aquatic amphibians; skull pitted and grooved by lateral-line canals and by sculpturing marks, or the skull may be smooth; teeth present on most of the palate bones; exoccipitals cartilaginous or calcified, never completely osseous; sclerotic plates sometimes present; skull of various shapes. Vertebræ with notochord largely persistent, hour-glass-shaped; neural spines low or high, or absent; ribs intercentral and single-headed, with an incipient tubercle in some forms; vertebral column differentiated into dorsal and caudal series; cervical series not clearly defined. Limb bones with well-ossified perichondrium, endochondrium partly ossified; epiphyses absent; carpus and tarsus (tarsus osseous in two species) cartilaginous; phalanges clawed or not; digits 4 in hand and 5 in foot. Pubis sometimes calcified but never osseous.

The order Microsauria was established by Sir William Dawson in 1863 (208) as a family of "reptiles" for the reception of the genera Hylonomus, Hylerpeton, Smilerpeton, and Fritschia. Hylonomus lyelli is the type species of the order. Dawson (216, p. 635) says of the species Hylonomus lyelli Dawson: "It is the type of the genus Hylonomus and of the family Microsauria." The forms which have been referred to the genus Hylonomus, and hence to the order Microsauria, from the deposits of Europe are discussed under Hylonomus.

The Microsauria have been regarded by the writer and others as being ancestral in a sense to some of the later reptiles (469), but there seem to be insuperable obstacles in the way of a direct derivation of the reptiles from the Microsauria. One of these obstacles seems to be found in the structure of the hand. In all Microsauria, so far as is known, there is no evidence of more than 4 digits in the hand, while no true primitive reptile possessed less than 5. The carpus of all true reptiles is osseous, while that of the Microsauria is merely cartilaginous. It is possible that the Microsauria stand in some such ancestral relation to the later reptiles as the Crossopterygia (489b) do to the Amphibia. The Microsauria had undergone adaptive modifications as to structure and habit, so that they have paralleled many of the groups of reptiles, but their structure is quite different. The evidence, as far as we can see now, points to a close genetic relation between the reptiles and the Microsauria, but that this relation is ancestral I, for one, am not ready to say.

The Group Aistopoda Miall, 1873, is untenable.

The group Aistopoda was established in 1873 as section IX by the Committee of the British Association for the Advancement of Science, in their "Tabular View of the Classification of the Labyrinthodonts." L. C. Miall (449, 450) was the secretary of the committee, and the report was published in two parts. Two genera were at that time attributed to the Aistopoda, Ophiderpeton and Dolichosoma, both described by Huxley from the Coal Measures of Kilkenny, Ireland.

Fritsch (251, pp. 107-126) in 1883 refers to the group as "Familie" and describes 4 genera and 9 species as belonging to the group. Zittel (642, p. 383) refers to the group as "Familie" and places 5 genera in it. Smith-Woodward (Vertebrate Paleontology, 1898, p. 129) refers the Aistopoda to a suborder. In Eastman's translation of Zittel's Paleontology the group is called a family, "Aistopodidæ." Lydekker (393, p. 205) regards the group as a suborder. The writer (469) refers the Aistopoda to an order. The group Aistopoda has been adopted by practically all paleontologists and zoologists who have had occasion to refer to these animals.

Lydekker (393) in 1890 defined the group as follows:

MOODIE

PLATE 8

Fritschia curtidentata Dawson. Above: Bones of skull and anterior extremity, and bony rods of belly. Below: Bones of pelvis and posterior extremity. Nearly natural size. Erect tree, Coal formation, South Joggins, Nova Scotia. Photograph by Dawson, published through the courtesy of Dr. Arthur Willey. Original specimen in the Peter Redpath Museum of McGill University.

If now we take up a consideration of each of the characters mentioned by Lydekker we find that the first one holds good for all examples of the group. The second character, "without limbs," is not good. Species of Œstocephalus, Ptyonius, Molgophis all possess limbs; and doubtless Ophiderpeton will be found to possess limbs also, since it has a well-developed pectoral girdle. The limbs in all of these genera are small. The third character, "apparently without pectoral and pelvic girdles," is not at all a good character, since nearly every specimen of some species and almost all species show evidences of pectoral girdles and a few exhibit pelvic girdles. The fourth character, "vertebræ with elongated centra and aborted neural spines," is not a good distinguishing character, since Amphibamus, an undoubted microsaurian, possesses the same vertebral characters. "Ribs slender, and barbed like those of fishes" is a character which is common to several widely distinct genera. All Microsauria possess long, slender ribs, and the barbed condition is one which is possessed by only a few, Thyrsidium, Ophiderpeton, etc., the so-called "barb" being merely a highly exaggerated tuberculum. The teeth of nearly all Microsauria are smooth, so that the character "teeth smooth" is not a good one for a group definition. It has not been possible to examine any of the American specimens for the plication of the dentine, since the forms are so rare and the fossils very fragile. The last character, "external gills probably persistent," is certainly not true for the American species, and the evidence for the European species is negative. Fritsch described and figured (Fauna der Gaskohle, Bd. I, 1883, p. 114, Tafeln 18 and 23) structures which he regarded as supporting structures for the external branchiæ. He says in regard to these structures:

John Samuel Budgett (79, p. 162), in his discussion of the "Structure of the Larval Polypterus," refers to the above-described specimen of "aistopodous Stegocephali," i.e., Dolichosoma longissimum Fritsch, and calls especial attention to the similarity of the external rod of segmented cartilage on the hyomandibular of Polypterus to this structure, to which Fritsch has assigned a branchiate nature in Dolichosoma. There is no doubt in the mind of the present writer, however, that the rod of cartilage, referred by Fritsch to the gills, can be other than scutellate rods of the ventral armature, these rods belonging to the armature of the breast or throat. The evidence for this conclusion is furnished by Fritsch himself (Fauna der Gaskohle, Bd. I, plate 18, fig. 11), where all may read in the figure of the specimen the facts of the case. There is quite evidently no justification for Fritsch's conclusion of the branchiate nature of Dolichosoma. There is no evidence of any gill-like structure in the American snake-like amphibians of the Coal Measures.

Reviewing, then, the characters of the group which have been assigned by various observers, it will be seen that there is but a single character which holds good: "body long and snake-like." This is totally insufficient for the retention of the group. I therefore propose to abolish the group entirely from zoological classification. It is not even a family. It will, however, be convenient to refer to the snake-like forms as "aistopodous."

THE MICROSAURIAN FAMILY HYLONOMIDÆ, FROM THE COAL MEASURES OF NOVA SCOTIA.

Family HYLONOMIDÆ Fritsch, 1883.

The following characterizations of the family are those given by Lydekker (393, p. 201) based on Fritsch (251): Body slender and lizard-like; skull narrow, with smooth or faintly sculptured bones; neural spines of vertebræ well developed, and long, slender ribs. Teeth smooth, or with grooved summits. The whole body covered with sculptured scutes. Internal gills may be developed.

Fritsch (251, Bd. I, p. 159) gives the following in his original description:

Fritsch includes the following genera in the family:

Hylonomus, Hyloplesion, Smilerpeton, Seeleya, Orthocosta, and Ricnodon. The family includes the following species:

Hylonomus lyelli Dawson, Nova Scotia.
latidens Dawson, Nova Scotia.
multidens Dawson, Nova Scotia.
wymani Dawson, Nova Scotia.
fritschii Gein. and Deichm., Saxony.
geinitzi Credner, Saxony.
(?) pictus Fritsch, Bohemia.
wildi Woodward, England.
Smilerpeton aciedentatum Dawson, Nova Scotia.
Hylerpeton dawsonii Owen, Nova Scotia.
intermedium Dawson, Nova Scotia.
longidentatum Dawson, Nova Scotia.
Fritschia curtidentata Dawson, Nova Scotia.

Genus HYLONOMUS Dawson.

Type: Hylonomus lyelli Dawson.

The genus Hylonomus is a very important one from a taxonomic viewpoint, since it was regarded by Dawson (216, p. 635) as the typical genus of the order Microsauria, the most abundant group in the Carboniferous. Unfortunately the species of the genus Hylonomus are known only from fragmentary remains. I have reproduced in plate 9 Dawson's figures of the remains of Hylonomus as published by him in 1891.

Dawson (216) gave, in 1882, the following definition of the genus Hylonomus: "Form lizard-like, with the posterior limbs somewhat large in proportion to the anterior. Size, small. Mandibular and maxillary teeth numerous, small, conical, pointed. Palatal teeth minute. Abdominal scales oval."

MOODIE

PLATE 9

Hylonomus lyelli Dawson. 1, maxillæ and skull bones; 1a, sternal bones; 2, mandible; 3, humerus, ribs, and vertebræ; 4, posterior limb; 5, pelvis; 6, caudal vertebræ. Nearly natural size. Erect tree, Coal formation, South Joggins, Nova Scotia. Photograph by Dawson, published through the courtesy of Dr. Arthur Willey. Original in the British Museum.

Credner (186), Fritsch (251, Bd. 1, p. 89, Taf. 12, figs. 1, 4, 15), and Woodward (629) have referred remains of Microsauria discovered in the Coal Measures or lower Permian deposits of Saxony, Bohemia, and Lancashire, England, to the genus Hylonomus. There is much uncertainty as to the validity of these references, due to the uncertain nature of the type of Hylonomus. There are 4 American species of the genus: Hylonomus latidens Dawson, H. lyelli Dawson, H. multidens Dawson, and H. wymani Dawson. All the species are from the Coal Beds at the South Joggins, Nova Scotia.

Hylonomus lyelli Dawson.

Type: Specimens Nos. R 443 to 445 in the British Museum (393, pt. IV, p. 223). Horizon and locality: Coal formation of the South Joggins, Nova Scotia. This species is by far the most abundant (plate 9) in the erect trees examined by Dawson. Its characters Dawson (216) defines as follows:

Interclavicle and numerous scutellæ are present. Upper surface protected with imbricated horny scales. In front two rows of horny tubercles and plates, with epaulettes composed of bristle-like fibers projecting from the skin.

The animal possibly fed on insects, as is indicated by the coprolitic matter associated with the remains of the species.

The following measurements are given by Dawson for the largest individual discovered:

Hylonomus latidens Dawson.

Type: Specimen No. 3061-1, Peter Redpath Museum, McGill University. The British Museum (393, pt. IV, p. 224) also has a specimen, No. R 447.

Horizon and locality: Coal formation at the South Joggins, Nova Scotia.

Fragments of 3 specimens from 3 trees represent this species (plate 10). It seems to have been of stouter build than H. lyelli, with the limbs shorter in proportion. Its generic affinities are somewhat doubtful, as it presents in some respects characters intermediate between Hylonomus and Hylerpeton.

Mandibular and maxillary teeth broadly conical, about 20 in each mandible 3 in 1 mm.; anterior mandibular teeth somewhat larger than the others, and bent or hooked. Vomer or palate with minute teeth. Thoracic plate large. Scales of abdomen oval, but somewhat narrow, and tending to be oat-shaped.

Hylonomus multidens Dawson.

Type: Specimen No. 3061-2, Peter Redpath Museum, McGill University.

Horizon and locality: Coal formation at the South Joggins, Nova Scotia.

This animal is known only by portions of bones of the head and a few other fragments. The scattered bones of the extremities are inseparable from those of H. lyelli occurring with it. As compared with that species, the bones of this are smoother and more delicate. The teeth are more numerous and slender. The crushed distal end of a femur or humerus found near the skull indicates that the limbs were well developed.

Hylonomus wymani Dawson.

Type: Specimen No. 3061, Peter Redpath Museum, McGill University. There is also specimen No. R 446 in the British Museum (393, pt. IV, p. 224).

Horizon and locality: Coal formation at the South Joggins, Nova Scotia.

As compared with the H. lyelli the present species is smaller in size, more elongated in form, had the teeth less numerous (about 22 in the mandible), and shorter and more obtuse in form. There are 6 to 7 in I mm.

This species is much more rare than H. lyelli, but quantities of minute bones, probably belonging to it, occur in the coprolitic matter. In one specimen 38 vertebræ of this species were found partially associated, indicating along, slender body. The body is covered with scales and ventral scutellæ are present. Dawson questions whether this species may not be the young of H. lyelli.

MOODIE

PLATE 10

Hylonomus latidens Dawson. Skull, portion of skeleton, foot, scapula, sternal bones, humerus, and rib, believed to belong to this species. Erect tree, Coal formation. Nova Scotia. Nearly natural size. Photograph by Dawson, published through the courtesy of Dr. Arthur Willey. Original in the Peter Redpath Museum of McGill University.

Genus SMILERPETON Dawson.

Type: Smilerpeton aciedentatum Dawson.

The type species was originally referred to Hylonomus, but further study induced Dawson to refer it to a new genus. Dawson gives (216) the following characteristics of the genus:

Smilerpeton aciedentatum Dawson.

Type: Specimen No. 3061-3, Peter Redpath Museum, McGill University. The British Museum (393, pt. IV, p. 224) also has a specimen, No. R 433.

The important characteristic (plate 12) is found in the form of the mandibular and maxillary teeth, which are of a peculiar wedge-shape, being broad and oval at the base and narrowed to a longitudinal edge at top. Thus, when viewed from the side they appear narrow and blunt, but when the jaw is broken across, and they are viewed from the rear or front, they appear broad and sharp-edged. The effect of this arrangement is that the jaw is armed with a closely placed series of chisels or wedges, giving an almost continuous edge. At the end of the mandible some of the teeth are longer and more conical.

Another important character is that the palatal and vomerine bones seem to have bristled with teeth, mostly of very small size; but there are also some larger palatal teeth, of which some are sharply pointed and others blunt with furrowed points.

The vertebræ are of the same type as those of Hylonomus; but some which appear to be caudal have a pointed spine above, indicating perhaps a flattened tail. The ribs are short and stout.

The body seems to have possessed an interclavicle and ventral scutellæ. Above it was, apparently, clothed with small tubercles and horny scales, and to have had cuticular pendants like those of Dendrerpeton.

An additional species of this genus was apparently indicated by some fragmentary remains, but Dawson thought best not to describe them as such, since they might indicate only a young individual of the present species.

Fig. 18. Skeletal elements of Smilerpeton aciedentatum Dawson, from the Coal Measures of Nova Scotia. (After Dawson.) a, shaft of femur, × 2; b, intermaxillary and teeth, × 25; c, sections of teeth, × 25; d and e, palatal teeth, × 25; f, femur, × 2; g, rib, × 2; h, palate; i, caudal vertebræ; j, long palatal tooth, × 25; k, bony scale.

Genus HYLERPETON Owen.

Body stout, with strong limbs. Mandibular and maxillary teeth strong, not numerous, grooved at apex. Palatal teeth numerous, and some of them large. Thoracic plate broad. Abdominal scales pointed or oat-shaped.

Type: Hylerpeton dawsoni Owen.

Hylerpeton dawsoni Owen.

Type: Specimen No. 3061-4, Peter Red path Museum, McGill University. There are also specimens, Nos. R 441 and 442, in the British Museum (393, pt. IV, p. 225). (Plate 7.)

Horizon and locality: Coal formation at the South Joggins, Nova Scotia.

Bones of skull slightly striated, but not sculptured as in Dendrerpeton. Lower jaw with distinct ascending ramus or coronoid process, a feature not known in any other of the Nova Scotia fauna, but observed by Cope in Brachydectes. Teeth, 12 in each ramus of the mandible, bluntly conical, slightly striated at the apex. Pulp-cavities large and longitudinally striated at the sides, though the teeth are not folded. Maxilla furnished with similar teeth, one of which near the front is larger than the others. Palatal teeth numerous, small, conical, with a few large teeth at the sides.

MOODIE

PLATE 11

Hylerpeton longidentatum Dawson. Mandible and other bones. Nearly natural size. Erect tree, Coal formation, South Joggins, Nova Scotia. Photograph by Dawson, published through the courtesy of Dr. Arthur Willey. Original specimen in the Peter Redpath Museum of McGill University.

MOODIE

PLATE 12

Smilerpeton aciedentatum Dawson. Mandible, portions of skull, scales, and various bones. Nearly natural size. Erect tree, Coal formation, South Joggins, Nova Scotia, photograph by Dawson, published through the courtesy of Dr. Arthur Willey. Original specimen in the Peter Redpath Museum of McGill University.

Vertebræ short, cylindrical, well-ossified, with well-developed zygapophyses and neural spines; ribs strong and much curved, with well-developed division of the proximal ends; pelvis imperfect, but apparently large, with broad ilium.

Humerus half the length of the mandible; radius half as long as humerus; femur very large and stout, nearly as long as the mandible; leg bones and phalanges correspondingly stout.

The thoracic plate (plate 7) is indicated only by some fragments. The abdominal scales are narrow and pointed (oat-shaped), smooth externally and with a ridge at one side within. The following are the dimensions of the largest specimen:

Hylerpeton longidentatum Dawson.

Type: Specimen No. 3061-6, Peter Redpath Museum, McGill University. There is also a specimen, No. R 440, in the British Museum (393, pt. IV, p. 225). (Plate 11.)

Horizon and locality: Coal formation at the South Joggins, Nova Scotia.

Measurements of Hylerpeton longidentatum Dawson.

Hylerpeton intermedium Dawson.

Type: Specimen No. 3061-5, Peter Redpath Museum, McGill University. Horizon and locality: Coal formation at the South Joggins, Nova Scotia. This species is known only by the mandibles and portions of the skull, which are rather shorter than those of adult individuals of the last species. The extremity of the mandible and the cranial bones have the same slightly waved surface as in the other species. Mandibles 3 cm. long and the teeth, which are about 15 in each ramus of the lower jaw, are simple, with large pulp cavities, those of the maxillary bone slightly enlarging upwards, and intermediate in form between the long, slender teeth of H. longidentatum and the thick, obtuse teeth of H. dawsoni.

Coal formations, South Joggins, Nova Scotia, in erect tree, discovered by P. W. McNaughton, 1893.

Genus FRITSCHIA Dawson, 1882.

Type: Fritschia curtidentata Dawson.

Fritschia curtidentata Dawson.

Type: Specimen No. 3061-7, Peter Redpath Museum, McGill University. There is also specimen No. R 449, in the British Museum (393, pt. IV, p. 225).

Horizon and locality: Coal formation at the South Joggins, Nova Scotia.

Represented by 2 specimens (plate 8). Bones of the head very smooth, only a few microscopic punctures. Teeth conical, somewhat obtuse, striated at the inner side of the apices; there are about 30 in each ramus of the mandible, and about 27 in the maxillary bone. Teeth implanted in a furrow. Vertebræ short and well ossified, 3 in 1 cm. Ribs strong, curved, about 1 cm. in length. Limbs robust, the bones better ossified than in any other known species from Nova Scotia. Toes of foot probably 5, central ones long and slender. Interclavicle of moderate size and somewhat rounded. Ventral scutellæ needle-like.

THE MICROSAURIAN FAMILY TUDITANIDÆ, FROM THE COAL MEASURES OF OHIO AND PENNSYLVANIA.

Family TUDITANIDÆ Cope, 1875.

Tuditanus punctulatus Cope, Linton, Ohio.
brevirostris Cope, Linton, Ohio.
longipes Cope, Linton, Ohio.
minimus Moodie, Cannelton, Pennsylvania.
walcotti Moodie, Linton, Ohio.
Erpetosaurus radiatus Cope, Linton, Ohio.
obtusus Cope, Linton, Ohio.
tabulatus Cope, Linton, Ohio.
minutus Moodie, Cannelton, Pennsylvania.
sculptilis Moodie, Cannelton, Pennsylvania.
acutirostris Moodie, Linton, Ohio.
tuberculatus Moodie, Linton, Ohio.
Odonterpeton triangularis Moodie, Linton, Ohio.

The association of these species in the one family is provisional and will need revision on the acquisition of new and more complete material.

Genus TUDITANUS Cope, 1874.

Type: Tuditanus punctulatus Cope.

The genus as here defined is a somewhat composite group and it is quite probable that some of the species here included will have to be removed to another genus when the anatomy of the forms is better known. The species of the genus are all moderately small, the largest barely attaining a length of 8 inches.

There are 5 species of Tuditanus thus far known. All of the species are characterized by the possession of a peculiar triangular-shaped clavicle with radiating grooves, and this has been taken as one of the distinctive characters of the genus, as well as of the family. The structure of the cranium where known is quite uniform among the different species. The squamosal is quite large and the supratemporal is not always closely joined to the parietal. The species are:

Tuditanus punctulatus Cope, Linton, Ohio.
brevirostris Cope, Linton, Ohio.
longipes Cope, Linton, Ohio.
minimus Moodie, Cannelton, Pennsylvania.
walcotti Moodie, Linton, Ohio.

Tuditanus punctulatus Cope.

Type: Specimen No. 110, American Museum of Natural History, where there is also specimen No. 111.

Horizon and locality: Linton, Ohio, Coal Measures.

This species, together with the form, T. brevirostris, described on p. 88, was used by Cope as the type of the genus Tuditanus. Cope subsequently associated some reptilian remains from the Linton mines with the type of T. punctulatus and changed the generic term to Isodectes, which was known by another species, I. megalops Cope, from the Permian of Texas. The remains associated by Cope with I. megalops undoubtedly represent a reptilian species and which has been described elsewhere under the name Eosauravus copei Williston. The species is of exceeding interest because it is the oldest known reptile and places the range of the Reptilia down towards the base of the Pennsylvanian.

The species Tuditanus punctulatus Cope was founded on well-preserved remains of nearly the entire skeleton of a single individual. The bones are represented by shining carbonaceous matter, and since both of the slabs containing the impression were preserved, a great many characters have been determined. The head, 1 fore-limb, and 23 consecutive vertebræ with ribs are well defined, but of the pelvis and hind limbs nothing is visible.

The cranium (fig. 19) is very similar to that of T. minimus, from the Cannelton, Pennsylvania, slates. It is triangular in shape, with a narrowed obtuse muzzle. The orbit of the left side is well defined and lies well forward. It is oval in outline and its width is about two-thirds of its length. The nostrils are small and are located well toward the tip of the muzzle. The parietal foramen lies behind the median transverse line which divides the skull equally.

The cranial elements are for the most part destroyed, but the outlines of a few can be determined. Those elements which are preserved are ornamented with a sculpturing of minute punctulations which, on the postfrontal, assumes a radiating arrangement. The ornamentation of the other elements consists of inosculating pits, but they seldom assume the form of ridges or grooves. The bones of the premaxillary region of the cranium are lacking. The first element which can be detected is the pref rental, which occupies a position in front of the orbit. There seems to be space for a lacrimal, but its outline is not distinct. The frontal can be readily separated and is seen to be an elongate element occupying the median region of the skull between the orbits. The parietal is apparently the largest element of the cranial roof and the pineal foramen is located in the anterior fourth of the median suture separating the parietal elements. The form of the postparietal and the tabulare can not be determined, as the greater part of this region is lacking. The squamosal seems to be located well forward and is rather small, but has the usual relation of this element. Only fragments of the other elements remain and nothing can be said of their form. The mandibles of both sides are represented by depressions, and they are ornamented with longitudinal grooves and ridges. The teeth are not preserved, but there are evidences of the maxillary teeth. These are minute and sharply conical. Just posterior to the skull there is preserved the impression of a short, round rod which is not definitely determined. It may be an element of the hyoid apparatus, although it is rather stout for such. It does not have the relations indicated by Cope in his figure (123, pl. XXXIV, fig. 1).

Fig. 19. Drawing of skull and skeletal elements of Tuditanus punctulatus Cope from the Coal Measures of Linton, Ohio. × 1.5. fr, frontal; ic, interclavicle; cl, clavicle; h, humerus; ph, phalanges; par, parietal; pp, postparietal; tab, tabulare, supratemporal and squamosal; u, ulna; r, radius.

There are three elements of the pectoral girdle preserved. These undoubtedly represent the interclavicle and the clavicles. The interclavicle is rhomboid in shape and is attenuated posteriorly. The attenuation is abruptly truncate posteriorly and it is thus of quite a different character from the acutely pointed interclavicle of T. minimus. The clavicle has a somewhat semicircular form, but is not attenuated at either end. It seems to be uniformly broad.

The forearm of the right side is preserved in part. The humerus is seen to be a heavy, somewhat expanded element lying displaced with relation to the pectoral girdle. It is greatly expanded at the ends. The ulna presents characters similar to the humerus and only differs from it in being shorter and less stout. The radius is not preserved. The carpus is unossified and its position is occupied by a blank space. The digits are represented by 4 metacarpals, and this may have constituted the entire number of the fingers. The phalangeal bones preserved are a little scattered. They are elongate with expanded ends.

Evidences of 23 consecutive osseous vertebræ are preserved. Their character can not be determined, although Cope (123) describes them as amphicœlous. This may be inferred to be the case, but I am unable to verify his observation. In form the vertebra are subquadrate. The neural spines are not evident. The osseous ribs articulate, apparently, between the bodies of the vertebra. Cope figured them as intercentral. There are 22 or 23 pairs preserved. They are single-headed and the extremities are attenuated. No traces of ventral scutellæ are present.

The entire length of the animal probably did not exceed 5 or 6 inches. Its form was quite lizard-like and it was probably of an ambulatory type, though it may have spent a part of its time in the old lagoon in which its remains were finally buried. No traces of external gills have been detected in this or any other Linton species.

Tuditanus brevirostris Cope.

Type: Specimen No. 8609 G, American Museum of Natural History.

Horizon and locality: Linton, Ohio, Coal Measures.

This species was associated by Cope with the type T. punctulatus in the description of the genus. Represented by a portion of the skeleton of one individual, the skull is preserved on one block, with a considerable part of the anterior ribs, pectoral girdle, and vertebral column, although this last is not clearly represented, but as in so many of the coal specimens the bones are covered with a thin layer of carbonaceous matter which makes it impossible to definitely determine the form.

The cranium is large in proportion to the size of the body. The skull is in the form of a wide oval and is wider than it is long. The elements of the skull were ornamented with a coarse sculpturing which partakes of the nature of incomplete radiations on the squamosal region. The different elements of the cranium can not be distinguished, although I think the outlines of the parietal are indicated. The position of the nostrils is well forward and they are slightly elongate transversely. The pineal foramen can not be determined. The orbits are oval in shape and their width is about equal to two-thirds of their length. The interorbital space is greater than the length of the orbit. Cope's figure of this specimen is not accurate, since he has the orbits drawn too far to the side. They are located near the central line of the skull and resemble in some respect those of the preceding species. Cope has described teeth in the maxillary region, but I am unable to detect them. There are portions of two pectoral elements which may represent a clavicle and a portion of the interclavicle.

The clavicle has much the same shape and practically the same ornamentations as has the clavicle in Tuditanus minimus. The clavicle preserved shows a somewhat triangular shape and is slightly acuminate at the anterior end, as preserved, and obtuse at the posterior end. The nature of the interclavicle can not be determined.

The vertebral column is represented by a line which Cope suggests (123) may be the chorda dorsalis (notochord). Osseous vertebræ were probably present, but their nature is obscured by the carbonaceous matter covering them.

The ribs as preserved are long and curved. They are slender and attenuated at the distal ends. They were probably single-headed, but whether their articulation was intercentral or not can not be determined.

The other specimens which are referred to this species show nothing of importance in the way of structure. They consist for the most part of fragments which may or may not represent the species.

The species differs from the type of the genus (T. punctulatus) in the possession of a broadly rounded muzzle. This character will also separate it from other species of the genus. The sculpturing of the bones of the cranium is coarser in the present species than in the type. The form of the clavicle is different in the two species. The above-described species seems to be more closely allied to the form described as Tuditanus walcotti than to other species of the genus. I have been unable to detect the presence of limbs, although Cope says they are present.

The material consists of the type specimen with counterpart and two fragments which probably are to be associated with this species. Collected by Doctor J. S. Newberry.

Tuditanus longipes Cope.

Cope, Trans. Amer. Phil. Soc., XV, p. 210, 1874 (Sauropleura).

Cope, Geol. Surv. Ohio, II, pt. II, 398, pl. XXVI, fig. 2, 1875.

Type: Specimen No. 1099 G, American Museum of Natural History.

Horizon and locality: Linton, Ohio, Coal Measures.

The cranium of this species is quite unknown. The only genus with which the specimen can be compared in the structure of the skeleton is Tuditanus. From the other species of the genus the present form differs in the presence of ventral chevron rods and the elongate character of the limbs, as well as in the possession of large iliac bones, which, in the only other species in which the ilia are known, are small and slender. It seems best to locate the species in this genus for the present, although it may eventually have to be removed to another group. Very little is known of the main portion of the skeleton of the species of Tuditanus, other than in T. longipes, so an exact comparison is impossible. From all the species of Tuditanus thus far known the present species differs in the elongate character of the limbs and in the presence of ventral scutellation. There are three other species of Tuditanus in which the limbs are known. These are: T. walcotti Moodie, T. minimus Moodie, and T. punctulatus Cope. In these three species the limbs are short and weakly developed. From the other species of Tuditanus the present species maybe separated by its size principally, since nothing of the bodies of the other species is known.

The body of the present species is elongate and slender, with a long neck and probably a long tail. Ribs, as preserved, are 19 to 22, though there may possibly have been more. They are moderately curved backwards, have intercentral articulation, are attenuated at the distal extremity, and are single-headed. The anterior ribs are stouter, with a widened upper portion and attenuated distal part. The posterior ribs are more slender.

Fig. 20. Cope's drawing of Tuditanus longipes, from the Linton, Ohio, Coal Measures. × 1.

There are evidences of 28 vertebræ present. All regions of the vertebral column are present and the dorsal region is preserved entire. The cervical series is represented by the posterior vertebræ only. These are very indistinctly preserved. The dorsal vertebræ are elongate and were probably amphicœlous, although this has not been definitely determined. They are expanded at each end, thus ending in a slightly raised rim. The single-headed ribs articulate between the vertebræ. The exact number of the dorsal series can not be ascertained, although this may have been 25. The spines of the vertebræ are not determinable, since the animal is preserved on its back. The caudal vertebræ are represented by two patches of the remains of what was once probably the entire series. Cope ascribes 70 mm. to the tail, but I do not find that much. The specimen may have been mutilated since he studied it. The caudals are slender and, like the dorsals, are expanded at the extremities.

The scapular arch is not preserved, but the pelvic arch is represented by the two iliac bones in good state of preservation. These are short, flat bones expanded at the anterior extremity, as preserved. They lie turned a little to each side of the vertebral column and partially obscure the femora. The iliac bones are quite characteristic of this form, since similar-shaped elements have not been observed in any of the other Carboniferous forms from the same deposit.

The greater part of the forelimb is preserved, although much of the hand is missing. The humerus is an unusually elongate bone and lies somewhat across the vertebral column. It is crushed flat and the ends are partly destroyed. It shows evidences of expansion at the ends, although not a great deal. It is much longer than the radius and ulna, which are of about equal length. The ulna is larger than the radius and has expanded ends, with the upper end more expanded than the lower and both ends slightly truncate. The radius is a simple rod of bone and is but slightly expanded. The carpus was evidently cartilaginous, since there is no evidence of osseous material in its place. There is but one phalangeal bone preserved, and, since this is displaced with reference to the ulna and radius, its position can not be determined. It may have been a metacarpal. It is short and expanded at the ends.

The hind limbs are represented by the two femora and the upper portion of the tibia. The femur is almost as elongate as the humerus and is more slender. It is not so much expanded as the humerus. Its ends appear to have been cartilaginous and do not represent the well-formed articular surfaces preserved in the T. minimus. The upper part of the tibia is preserved, and appears to have been truncate.

If this species belongs with Tuditanus it is of interest in that the ventral chevrons are present. The species is particularly characterized by the elongate limbs.

Tuditanus minimus Moodie.

Type: Specimen No. 4555, U. S. National Museum.

Horizon and locality: Cannelton slates of Pennsylvania (Upper Freeport).

The species is represented by a nearly complete skeleton preserved on a slab of slate from the Cannelton shales of Pennsylvania. The obverse slab has been lost, which is very unfortunate, since there is no doubt that the entire skeleton was originally present. The species is placed in the genus Tuditanus on account of the close resemblance to the type form T. punctulatus Cope, although it is much smaller than that species.

The type specimen of the species did not attain a length of more than 3.5 inches. Its form is very lizard-like, but its structure is typically amphibian. The form of the skull is especially similar to that of the type species T. punctulatus, which it resembles in the narrow posterior truncation of the skull, as well as in the anterior position of the orbits.

The skull is in the form of a narrow oval, sharply narrowed posteriorly and truncate. The orbits are located well forward and their posterior border lies in front of the line dividing the skull transversely into equal parts. The interorbital space is greater than the diameter of the orbit. Impressions of teeth are preserved on the premaxillæ and maxillæ; there are 8 of them in a distance of 3 mm. The teeth appear to be mere blunt denticles and were possibly pleurodont.

The elements of the cranium are very poorly preserved. It has been impossible to determine all of the sutures. The bones of the premaxillary region have been destroyed, but the arrangement of them was probably not far different from that which obtains in other members of the genus. The posterior boundaries of the nasals are preserved and prove this element to have had an obtuse posterior border. The sutures bounding the frontals are clear and show that they were small and that they formed a part of the inner boundary of the orbits. The parietal is recognized as a large element, apparently the largest in the skull. Together the parietals form a wide oval inclosing, on the median suture, the circular pineal foramen. The parietals are sculptured with coarse radiating grooves and ridges, much after the manner of Erpetosaurus radiatus Cope. The pittings present on that form are, however, absent in T. minimus. The sutures bounding the postparietal are tolerably well defined and these show that element to have been rather large and quadrate, with the usual relations. The tabulare is distinct, triangular, and small. It is produced into an angle on the posterior border strongly recalling a similar condition in T. punctulatus. The boundaries of the prefrontals and the upper borders of the maxillæ are not clearly ascertained. The lacrimal has not been detected. The post-frontal and postorbital form the posterior boundary of the orbit, although all of the limits of the latter element have not been definitely determined. The position of the supratemporal is well assured, although its entire boundaries are not determined. It has the usual relations and joins the parietal broadly. The jugal is broad and widens posteriorly to join the squamosal, which, as usual, forms the quadrate angle of the skull. The sutures bounding the quadratojugal and the posterior end of the maxilla are not determined.

There are but two fragmentary vertebræ preserved and an estimate based on the length of these remains gives about 30 presacral vertebræ. The structure of the vertebræ preserved can not be ascertained, but the neural spines appear to have been low and stout.

There are six elements of the pectoral girdle preserved. These are: the six clavicles, the interclavicle, the coracoid of one side, and the two scapulæ. The interclavicle is rhomboid in form and acuminate posteriorly. It is sculptured with radiating grooves and ridges. It is quite different from the same element in T. punctulatus, in that the base is acuminate, not truncate. The clavicle presents much the same shape as does that element in Erpetosaurus tabulatus. It is ornamented by a sculpturing of radiating lines which take their origin from the lower external angle as the bone lies in the matrix. The clavicle is somewhat triangular in shape and lies close to the skull, but this close approximation of the pectoral elements to the cranium is due probably to post-mortem shifting, since the scapulæ are shifted backward. There can be little doubt, however, that the pectoral arch was close to the cranium. There is an oval fragment preserved on the left of the specimen which I take to be a portion of the coracoid. The scapula is preserved entire on the left side and is represented by fragments on the right side. It is almost semicircular in form and narrows externally until it is somewhat fan-shaped. There appears to be an ornamentation of lines on the surface of the bone. These lines follow the contour of the anterior border.

The arm is preserved nearly complete on the left side, and the right side shows the humerus and the forearm. The humeri are unusual in having well-developed articular ends as though the endochondral tissue was well developed. The humerus is expanded at the ends and it is larger at the upper than at the lower end. The ulna is expanded at the proximal extremity, but is more attenuated at the distal portion. It is shorter than the humerus by about one-third of its own length. The radius is a mere slender rod of bone and presents well-developed articular ends. It is slightly shorter than the ulna. The carpus is unossified and its position is represented by a blank space. There are phalanges of 4 digits preserved and they are 4 in number. The phalangeal elements, like the other bones of the extremity, have the articular surfaces prominent, with the terminal phalanx claw-like.

There are no ribs nor traces of them preserved, and a conjecture as to their character can not be hazarded, since they are known in but two other species, in which they are slender and curved. There is no evidence of a ventral scutellation, and so far as is at present known this structure is absent from all of the species of the genus, or at least it is weakly developed.

The ilium is all that is preserved of the pelvis. The bone itself has disappeared and has left a depression which shows this element to have been an elongate rod very similar to that described for Micrerpeton. The sacral vertebra seems to be indicated by a depression between the iliac depressions.

One hind limb is preserved nearly entire and the greater part of the other is also preserved, although some of the phalanges are disturbed. The femur is slender and more elongate than the humerus. It has well-formed, rounded, articular ends. The tibia presents unusual characters in that its ends are truncate, as though the cartilage composing its articular surfaces was not so highly calcified as in the other limb bones. It is somewhat expanded at the ends and is throughout its length broader than the femur. The fibula, like the tibia, is a slender rod of bone, although it is somewhat shorter than the tibia. The tarsus is unossified and its position is occupied by a blank space. Portions of both feet are preserved, but only one digit in the right foot is complete. The metatarsals are elongate and -slightly expanded at the ends. There are 4 phalanges present in the complete digit, which possibly represents the fourth. The first digit is wanting, with the only terminal phalanx preserved claw-like.

Tuditanus walcotti Moodie.

Type: Specimen No. 4474, U. S. National Museum.

Horizon and locality: Linton, Ohio, Coal Measures.

A small species of Microsauria is preserved as a smooth impression on a block of soft coal from Linton, Ohio. Nearly the entire form of the body is discernible. The specimen is especially interesting and valuable as exhibiting for the first time among the Linton forms the shape of the body of the small microsaurians of the Tuditanus type. It differs so markedly in the form of the skull from other species of the genus that it is regarded as a distinct form, and the name Tuditanus walcotti was proposed for it as an expression of the writer's indebtedness to the Secretary of the Smithsonian Institution for the vise of the material among which the present form was included.

Fig. 21. A. Outline drawing of type of Tuditanus walcotti Moodie, from the Coal Measures of Linton, Ohio, showing impression of body and muscle at M. × 25. cl, clavicle; fr, frontal; f, femur; h, humerus; nos, nostril; or, orbit; par, parietal; rb, rib; pp, postparietal; v, vertebra; pfo, pineal foramen.
B. Left leg of second specimen of Tuditanus walcotti. × 3.

The specimen includes, besides the body impression, the complete skull, a right clavicle, with portions of the left, a left humerus, 12 cervical and dorsal vertebræ, 10 pairs of ribs somewhat disturbed as to position, and a portion of the mandible. There are no traces of ventral scutellæ nor body scales in the smooth impression of the carbonized skin. One would expect to find impressions of the ventral scutæ in this specimen if they were present. Cope remarked on the apparent absence of scutellæ from members of the genus Tuditanus as they were known to him, and no contrary evidence has since been brought to light. Until such evidence is forthcoming the absence of scutes will be taken as one of the generic characters of the genus Tuditanus. Under a magnification of 50 diameters the carbonized skin shows as folds and wrinkles, like muscle fibers, in some places; in others no traces of the muscular structure can be detected. The wrinkles may be impressions of the internal musculature of the body-wall of the abdomen. It is especially well preserved in the pelvic and pygal regions. Sections of the coal were made, but nothing definite could be determined as to the character of the impressions, as they were too poorly preserved and the coal was too soft to bear much handling.

The specimen is preserved on the belly, with the dorsum of the skull uppermost. It has been practically impossible to determine the arrangement of any of the cranial elements except the f rentals, parietals, and postparietals, which have the relations indicated in figure 21, A. A median suture is clearly evident, with the pineal foramen well back in this suture. The bones of the skull are marked with faint radiating lines. It is in the form of the skull and the position of the orbits that the specific characters are found, as follows: the backward position of the eyes and the oval, pointed shape of the skull. The species is closely related to Tuditanus minimus Moodie, from the Cannelton slates of Pennsylvania, and serves further to connect the forms from the Ohio and Pennsylvania localities. It differs from the last-named species in the position and form of the orbits, these structures being more oval in the present form and placed further back. The shape of the skull differs also in the almost entire absence of the posterior table. The median points of the orbits occupy the line which bisects the skull, and the interorbital width is less than the width of the orbit. The mandible is heavy and appears to have borne sharp, pleurodont teeth.

The vertebral column is represented by little more than a mold of the form of the vertebræ, so that little can be said of its character. The individual vertebræ are short and hour-glass-shaped. The ribs are borne intercentrally, as in all the microsaurians which have been studied from the Linton deposits. The ribs are rather long and somewhat heavy, slightly curved and expanded at the proximal end, as though an incipient bicipital condition were present.

The right clavicle, which is preserved as an impression, is entire. Its impression shows this element to have been ornamented on its ventral surface with radiating grooves and ridges which started at the lower angle of the bone. The element is distinctly triangular, which is characteristic of the genus Tuditanus, so far as known. The fragment of the left clavicle adds nothing to our knowledge of the element.

The left humerus recalls in a striking way that of Tuditanus longipes Cope, and it was once entertained as a possibility that the present form might be a member of that species, since the skull is lacking in T. longipes. Sufficient specific differences were found, however, in the ribs, which, in T. longipes, are very long, slightly curved, and delicate, but which, in the present form, are comparatively heavy. Other characters sufficiently diagnostic are found in the form assumed by the vertebræ in the two forms.

The above-described specimen was collected by Mr. R. D. Lacoe, of Pittston, Pennsylvania, from Linton, Ohio.

A second individual (No. 4481, U. S. National Museum) of this species is indicated by a rather poorly preserved specimen on a slab of soft coal from the Linton mines. The following portions of the animal have been detected and will be discussed: partial impression of the skull, with a fragment of a minute jaw, in which are minute teeth; right clavicle; part of the impression of the body; nearly entire left hind limb; impressions of about a dozen vertebræ, very indistinct.

The impression of the skull is distinct only in a favorable light, and even then the boundaries of the cranium are a little uncertain. For this reason no representation of the form will be attempted. The sculpturing on the parietals is, however, distinct enough to show relationship with the previously described specimen, and the form of the body impression, the absence of abdominal scutes, the shape of the clavicle and its sculpture, and the proportions of the hind limbs all agree with the characters which have been assigned to the genus Tuditanus. The fragment of the jaw is interesting as giving the first information as to the character of the mandible in the genus Tuditanus. It is very slender and of uniform width as far as preserved. The teeth are short, blunt cones, apparently pleurodont.

The clavicle is of the typical Tuditanus form, with the sculpturing lines radiating out from the angle. The impression of the body adds nothing to that already described for the type specimen. The nearly entire hind limb is of great interest as adding another example of the phalangeal formula. The foot is almost perfectly preserved, and the formula was probably 2-2-3-3-2. The endochondrium of the limb bones is not highly developed. About a dozen vertebræ are represented by molds in the soft coal, but nothing of their structure can be determined.

The sharp, reptile-like claws in which the toes end (fig. 21 , B) recall those of Eosauravus and of Tuditanus minimus Moodie. It is another link in the chain of the suggested relationship between the microsaurians and the early reptiles.

Genus ERPETOSAURUS Moodie, 1909.

Type: Erpetosaurus radiatus Cope.

Skull stout, elements sculptured with radiating grooves, ridges, and pits; orbits large and usually placed far forward; occiput sometimes with posterior table; skull more or less rounded; lateral-line canals consisting of supraorbital, suborbital, jugal, and temporal canals, the last two uniting to form a circular canal in one species; clavicle triangular, sculptured like the skull.

Our knowledge of the genus is confined to the skull. The genus was established for certain members of the genus Tuditanus and other forms which have been recently described. The species of the genus are: E. radiatus Cope type, E. tabulatus Cope, E. tuberculatus Moodie, E. obtusus Cope, E. minutus Moodie, E. acutirostris Moodie, E. sculptilis Moodie. All of the species are from the Linton, Ohio, Coal Measures, with the exception of E. sculptilis and E. minutus, which are from the Cannelton, Pennsylvania, slates.

The position of the genus as to family is a little uncertain, since family characters are not yet well understood among the Carboniferous forms on account of the lack of information as to the structure of the animals. If we take the absence of the ventral scutellæ as a family character, the genus will be in the family Tuditanidæ, but the evidence on this point is negative. For the present we may place the genus only provisionally in the family Tuditanidæ. The arrangement will undoubtedly require revision later.

Erpetosaurus radiatus Cope, 1874.

Type: Specimen No. 8600 G, American Museum of Natural History. Locality and horizon: Linton, Ohio, Coal Measures. (Plate 25, fig. 1.) Cope originally described this species from a portion of a skull. He (123) characterized the form as follows:

The base of the specimen is broken and there is no place for the occipital condyles. Unless the specimen has been mutilated since Cope studied it, the occipital condyles are not present.

The restoration of the skull given in figure 22, B varies but little from that given by Cope in 1875. The elements are practically as he represented them.

The premaxillæ are small and lie in the usual relations to the other elements. Minute conical teeth are present as impressions. They are quite similar to the teeth found in other Microsauria. The nasals are nearly square and form the inner boundary of the somewhat oval nostril, which is represented by a depression in the coal. The frontal is elongate. It is about twice as long as wide. It forms a portion of the inner border of the orbit, the remainder being made up by the prefrontals and the postfrontal. The parietals are the largest elements of the skull, but they do not greatly exceed the jugals. Together the parietals form a somewhat obtuse oval in the median region of the skull and they contain between them, in their posterior third, the small circular pineal foramen. The postparietal forms the posterior boundary of the skull. The pref rental forms the anterior border of the orbit and is triangular in shape. The lacrimal is not identified. The maxilla is an elongate element the boundaries of which are uncertain, though probably somewhat as given. The postfrontal and the postorbital form the posterior boundary of the orbit, inclosing between them the anterior projection of the squamosal. The squamosal is an elongate element and is acuminate at each end. The tabulare is a large element lying lateral to the postparietal. The jugal is a very elongate element, apparently acuminate anteriorly. The quadratojugal is small and elongate. The supratemporal is definitely bounded and its limits are as indicated (fig. 22, B), being a large element which forms the quadrate angle.

There are two other specimens of this species in the collections and a fragment of a fourth which it is difficult to make out. Cope identified and figured one of these as E. radiatus (Geol. Surv. Ohio, II, pt. II, pl. 34, fig. 3), but the identification is doubtful and the figure shows structures which I am unable to identify in the specimen. Nothing of importance is to be learned from the other two specimens, except that they show a diversity of size. They consist of incomplete skulls, concerning which Cope (123) remarks:

The specimens of this species were collected by Dr. J. S. Newberry.

Erpetosaurus obtusus Cope, 1868.

Type: Specimen No. 8601 G, American Museum of Natural History.

Horizon and locality: Linton, Ohio, Coal Measures.

The species, Tuditanus obtusus, was first described by Professor Cope as Dendrerpeton obtusum, but he subsequently referred the species to the genus Tuditanus. It was removed by the writer to the genus Erpetosaurus in 1909. The species is known from two partially preserved crania. The skull elements seem to have disappeared and left only the impressions. The sutures are, for the most part, clearly represented, but the skull shows no sculpturing. On the posterior third of one cranium there is a small space which seems to be slightly sculptured, as Cope indicated in his drawing. The general form of the skull is that of a broad oval, truncate posteriorly. The orbits lie in the anterior third of the skull, and the pineal foramen in the posterior third. Cope compared the skull to that of Huxley's Erpetocephalus, to which it has some resemblance. The nostrils are elongate and are situated at an obtuse angle with relation to the main axis of the skull.

The premaxilla is small and forms the inner border of the nostril. There seem to be impressions of small teeth, but no large ones are evident. The nasals are separated by a zigzag suture, and are nearly square. They have the usual relations. The frontals form a portion of the inner boundary of the orbits and unite behind with the parietals, the anterior extensions of which they inclose. The parietals are large elements and together form a broad oval, truncate posteriorly. They inclose between them the pineal foramen in the median suture. It lies in the posterior fourth of the parietals. The postparietals are elongate transversely, and have the usual relations. The pref rental is somewhat triangular and forms the anterior boundary of the orbit. The lacrimal has not been detected. The maxilla is elongate, and with the quadrate jugal forms the exterior border of the cranium. No teeth are observed on the maxilla. The postfrontal and the postorbital form the posterior border of the orbit, and between them inclose the anterior extension of the squamosal. The supratemporal is pointed anteriorly and has the usual relations of that element, joining the postorbital, the postfrontal, the parietal, the tabulare, and the squamosal. The tabulare is larger than the postparietal and is acuminate, the point being inclosed by the squamosal and the supratemporal. The jugal widens fan-shaped posteriorly. It forms a portion of the border of the orbit. The supratemporal, as usual, forms the quadrate angle of the skull. In front of it lies the elongate quadrate jugal.

Fig. 22.

A. Outline of skull and cranial elements of Erpetosaurus minutus Moodie, from the Cannelton slates of Pennsylvania. Original in U. S. National Museum. × 2. fr, frontal; mx, maxilla; par, parietal; po, postorbital; pp, postparietal; pr, postfrontal; sq, squamosal; spt, supratemporal; tab, tabulare.

B. Outline of skull and cranial elements of Erpetosaurus radiatus Cope, from the Coal Measures of Linton, Ohio; partially restored. × 0.75. Original in American Museum of Natural History, fr, frontal; j, jugal; mx, maxilla; n, nasal; or, orbit; par, parietal; pf, postfrontal; po, postorbital; pr, pref rental; pp, postparietal; pmx, premaxilla; qj, quadratojugal; sq, squamosal; spt, supratemporal; tab, tabulare.

C. Palate of Erpetosaurus (tabulatus?), from the Coal Measures of Linton, Ohio. × 1. American Museum of Natural History, ex, exoccipital; m, mandible; mx, maxilla; pal, palatine; pt, pterygoid; prv, prevomer; pd, parasphenoid; x, anterior and posterior palatine vacuities.

D. Outline of skull and cranial elements of Erpetosaurus acutirostris Moodie, from the Coal Measures of Linton, Ohio. Original in American Museum of Natural History. × 1. ex, exoccipital; fr, frontal; j, jugal; mx, maxilla; n, nasal; or, orbit; po, postorbital; par, parietal; pp, postparietal; pf, postfrontal; sq, squamosal; spt, supratemporal; tab, tabulare.

E. Outline of larger part of skeleton of Odonterpeton triangularis Moodie, from the Coal Measures of Linton, Ohio. Original in U. S. National Museum. × 4. h, humerus; il, ilium?; v, vertebræ, all of which are not represented.

F. Right mandible of Erpetosaurus tabulatus Cope, from the Linton, Ohio, Coal Measures. Original in American Museum of Natural History. × 2. art, articular; cor, coronoid; d, dentary; ang, angular; sa, surangular.

G. Skull elements and lateral-line canals of Erpetosaurus tabulatus Cope, from the Coal Measures of Linton, Ohio. × 1.5. fr, frontal; j, jugal; jl, jugal lateral-line canal; mx, maxilla: n, nasal; or, orbit; par, parietal; po, postorbital; pf, lacrimal or prefrontal; so, suborbital lateral-line canal; pp, postparietal; pmx, premaxilla; qj, quadratojugal; spo, supraorbital lateral-line canal; sq, squamosal; spt, supratemporal; tm, temporal lateral-line canal; tab, tabulare; so, supraoccipital cross-commissure of lateral-line canal.

Dendrerpeton is not well enough known for an exact comparison with Erpetosaurus obtusus, but Cope separated the latter from Dendrerpeton on the position of the orbits and the broadly rounded muzzle. This species differs from the other species of the genus in the form of the cranium, as well as in the characters which separate it from Dendrerpeton.

Two other specimens, Nos. 8602 G and 8608 G of the American Museum, are associated in this species.

Erpetosaurus tabulatus Cope.

Type: Specimen in the Zoological Collection of Columbia University.

Horizon and locality: Linton, Ohio, Coal Measures. (Plate 25, fig. 2.)

The species is known from a single well-preserved skull and its obverse in the collection of Columbia University in New York City. I am indebted to Dr. Bashford Dean for the privilege of studying this interesting form. It is from the Linton deposits of Ohio. The remains include a nearly complete cranium and a complete clavicle of the right side. The species agrees in all essential respects with the characters of the genus Erpetosaurus, presenting a broad, flat head and a triangular clavicle.

MOODIE

PLATE 13

Dendrerpeton oweni Dawson. Above: Skull, mandible, and bones of anterior limbs. Below: Posterior limb, pelvis, and bony scales. Nearly natural size. Erect tree, South Joggins, Coal formation, Nova Scotia. Photograph by Dawson, published through the courtesy of Dr. Arthur Willey. Original specimen in the Peter Redpath Museum of McGill University.

The cranium is wider than long, the muzzle broadly rounded. The orbits are wide ovals, and their posterior borders fall little behind the transverse line dividing the skull equally. The interorbital width equals the long diameter of the orbit. The posterior outline of the cranium is truncate in a straight transverse line between the prominent tabulare angles. The composition of the cranium is different from that of any other species of Erpetosaurus in the large size of the tabulare and the fact that the supratemporal is excluded from the parietal by the extension of the postorbitals and the tabulare. This may be a generic character and entitles the species to be placed in a new genus, but it will be retained here until more of the anatomy of the species is known. The elements of the anterior part of the skull are not preserved, but they are indicated by the broken lines in the drawing (fig. 22, G). The nostrils are, however, clearly indicated as bosses of shale. There is a mere fragment of the nasal preserved posterior to the crack indicated by the transverse line in the drawing. The frontal is elongate as in other species of the genus and forms the inner border of the orbit. The parietal, as usual, is one of the larger bones of the skull roof and the pineal foramen is inclosed in the median suture by the two parietal elements. The pineal opening lies in the posterior half of the parietal. The postparietal is almost square, being slightly elongate transversely, uniting laterally with the tabulare, with which it forms the truncate table of the skull. The suture separating the tabulare from the supratemporal is clearly distinct. Although such a position for the supratemporal is unusual it is not unique, since the same character has been observed in Diceratosaurus lævis Moodie, described elsewhere (p. 120) in this paper. The postfrontal is rather small and it, together with the postorbital, forms the posterior boundary of the orbit. The postorbital is truncate posteriorly and joins the tabulare broadly. The supratemporal lies posterior to the postorbital and jugal and borders the quadratojugal, which is an unusual condition, but what significance the condition has remains to be determined. Posterior to the supratemporal lies the squamosal, which forms the quadrate angle of the skull. The quadratojugal is a small element and forms part of the lateral boundary of the skull. The jugal is a large element and forms the entire lateral border of the orbit. There are no teeth preserved on the fragment of the maxilla, but there are some impressions farther forward which resemble the pleurodont denticles of the modern Amphibia.

The sculpture of the surface of the cranial bones consists of parallel ridges which are separated by grooves equal to them in width. The ridges radiate inward on the squamosals and f rentals and outward on the supratemporals. They are somewhat interrupted on the other skull elements. The right clavicle is ornamented with a sculpture of similar radiating grooves and ridges.

Cope described an atlas in connection with this skull, but I do not find it. The slender impressions to the right of the clavicle may possibly represent ribs. They are gently curved and truncate at the inner end.

A nearly complete system of lateral-line canals has been detected on this skull. The canals preserved are: the temporal, the jugal, the infraorbital, the occipital cross-commissure, and the supraorbital. These terms were used for the first time for the Amphibia by the writer (458) in a discussion of the organs and their significance in the correlation of the skull elements. The occipital cross-commissure in the present skull is represented by a row of elongate pits, such as Andrews (8) has described for Ceraterpeton galvani Huxley from the Coal Measures of England. The cross-commissure is contained within the tabulare. The jugal and temporal canals form a complete ring, much as the same canals do in Trematosaurus. The supratemporal in Erpetosaurus tabulatus Cope is excluded from the parietal by the extension of the tabulare and the postorbital, and it is to be noticed that the temporal canal has a changed position to correspond with the changed condition of the squamosal. This is of considerable interest in connection with the correlation of the supratemporal in fishes and amphibians. This subject has been fully treated in another place (458) and it will only be necessary to state here that on the basis of the lateral-line canals and their arrangement in fishes and the Amphibia the true correlation of the supratemporal elements in amphibians and fishes has been made. The temporal canal in the present specimen has, apparently, an indication of a connection with the supraorbital canal, but of this I am not sure. The jugal canal occurs on the supratemporal and quadratojugal, and it joins the infraorbital on the jugal. The infraorbital is indicated by a short portion a few millimeters long under the orbit and the remainder, i.e., its connection with the jugal canal, is restored (fig. 22, G). There is, nothing unusual to be observed in that portion of the infraorbital canal which is preserved. The supraorbital canal is indicated by a curved, broad, shallow groove on the inner side of each orbit. As stated above, there seems to be a connection between this canal and the temporal, but I am not sure. The primitive conditions shown in the lateral-line canals in Erpetosaurus tabulatus Cope are the presence of the occipital cross-commissure and the ring-like formation of the temporal and jugal canals, which is too clearly indicated to be overlooked.

Mandible Provisionally Associated With Erpetosaurus tabulatus Cope.

This specimen is preserved almost completely on a slab of soft coal. It is impossible to determine positively to what microsaurian species the mandible belongs, but it may, for the present, be associated with Erpetosaurus tabulatus Cope on account of its size and the character of its sculpture. This is the first and almost the only known example of a mandible of an American microsaurian. The form of the jaw is perfectly preserved, although the condition of the articular surface can not be determined.

The proportions of the mandible, as maybe judged from the table of measurements, are rather stout and the teeth are strong and numerous. There are evidences of 19 teeth preserved. The sutures separating the articular (art), angular (ang), surangular (sa), coronoid (cor), and the dentary (d) (fig. 22, F) are clear for at least the greater part of their length and they may be easily restored for the remainder of their course. The surangular is thus seen to rival the dentary in size and on it occurs the peculiar sculpturing which approximates so closely that on the skull of Erpetosaurus tabulatus Cope. The presence of the long anterior tooth is strikingly characteristic of many Microsauria. It is well developed in Sauropleura longidentata Moodie and Sauropleura (Anisodexis) enchodus Cope. It is also present in well-developed form in the later labyrinthodonts. The teeth are all, with the exception of the fourth from the anterior end, rather short, curved, and sharply pointed, with an indication of longitudinal fluting. The arrangement of the mandibular elements recalls in a striking way the mandible of Eryops megacephalus Cope as figured by Branson (49).

Another specimen of this same species is 8550 G, of the American Museum. Linton, Ohio, Coal Measures.

PALATE OF ERPETOSAURUS TABULATUS COPE.

The specimen is half a cranium with its impression. It is referred to Erpetosaurus on the basis of the sculpturing of the mandible and the posterior table of the skull. On the surangular there is seen the rugosity which is common to other members of the genus. The characters presented are those of Erpetosaurus tabulatus Cope, but the reference is rather uncertain.

The specimen is unique among the known American material in showing the structure of the palate. Jaekel (347) has figured and described the palate of Diceratosaurus punctolineatus Cope from the Linton beds. The present palate differs from that species only in the enlarged ectopterygoid and the smaller palatine.

The parasphenoid, in the present form, does not differ from that element in other Paleozoic Amphibia. Its form is slender, arising from an enlarged base and separating the pterygoids by its own width. The exoccipitals are probably represented in the specimen and they have been indicated in the drawing (fig. 22 C). They are rather large and extend some distance under the base of the skull to unite anteriorly with the pterygoids, a very unusual arrangement. The pterygoids are elongate elements and are bounded anteriorly by the vomer and laterally by the ectopterygoid. The vomer shows no evidence of being toothed, although it may have been so anteriorly. The same may also be said for the palatines. The relations of the ectopterygoids are rather unusual for the Amphibia, especially in the posterior extension of the element. The bone lies all along the side of the pterygoid and anteriorly projects forward between the pterygoid and the palatine. In this unusual posterior projection the ectopterygoid has almost obliterated the infratemporal foramen, which possibly may be still represented by the triangular space between the bases of the pterygoid and the ectopterygoid. The anterior palatine foramen (internal nares) lies between the anterior ends of the palatine and the vomer, its usual relations in the labyrinthodonts. The foramen may be recognized as the rounded depression slightly anterior to the palatine.

The mandible is rather heavy and is coarsely sculptured with radiating grooves and ridges. The character of the teeth can not be determined, save to say that they were present. The posterior end of the mandible projects somewhat beyond the quadrate angle of the skull.

The interest in the present specimen is heightened by the light it throws on the characters for the separation of the Amphibia and Reptilia. The wide separation of the pterygoids by the parasphenoid is an amphibian character of undoubted value. The reduction of the parasphenoid in this specimen is noteworthy.

Erpetosaurus minutus Moodie.

Type: Specimen No. 4545, U. S. National Museum.

Horizon and locality: Cannelton slates, Pennsylvania (Upper Freeport). (PLATE 20 D.)

The specimen on which the species is based is composed of the greater portion of a small skull preserved in the hard shale from Cannelton, Pennsylvania, and was collected by Mr. R. D. Lacoe, of Pittston, Pennsylvania. The characters of the specimen had not previously been determined, since the museum label and number had partially obscured the snout of the skull. The skull is very small, but has the form assumed by other members of the genus. The specimen may belong to a young individual, but even though it does, it is, nevertheless, quite distinct from the other species of Erpetosaurus. At first sight the specimen looks like a broken scute of some larger form. Close inspection, however, revealed the two impressions representing the orbits, and a Zeiss binocular revealed the characters. The large size and anterior position of the orbits, the character of the sculpturing, the presence of the posterior table of the skull, as in Erpetosaurus (Tuditanus) tabulatus Cope, are the characters on which a specific diagnosis is possible. The specific characters which distinguish this form from E. tabulatus Cope, its nearest ally, are the slight development of the posterior table, the more delicate form of the sculpturing, the more posterior position of the orbits, and the varying shape assumed by the parietals in the two species. Any one of these characters would be valid as a specific character.

The pineal eye is indistinct, but is observed to lie in the broken tract in the median line of the skull, in the middle of the portion posterior to the orbits. The interorbital space is equal to the width of the orbit. The orbits themselves are slightly oval and not round, as in the case of E. tabulatus Cope.

The skull elements are sculptured with radiating grooves and ridges, and on the postparietals and tabulare the grooves take the form of pits in a row, which undoubtedly represent the occipital cross-commissure of the lateral-line system first observed in a microsaurian by Andrews (8) in the skull of Ceraterpeton galvani Huxley. The supraorbital canal is represented by a slight elongate depression observable over each orbit and extending, in one case, for about 5 mm. The presence of the circular arrangement of the lateral-line canals in the jugal region is suggested by a depression on the posterior edge of the squamosal.

The portion of the skull anterior to the orbits is wanting, curiously enough, just as in Erpetosaurus tabulatus Cope. In the remainder of the skull, the post-parietals, the tabulare, the parietals, the supratemporal, and a portion of the right frontal can be detected, although the boundaries of but three can be accurately defined. The depression bounding the anterior outline of the skull is taken to be the impress of the mandible, in which case this structure would be of some depth, as in the case of the mandible associated with E. tabulatus Cope, described below.

This specimen is of interest in respect to the presence of the lateral-line canals, its small size, and its generic identity with forms from Ohio. There is still another form known from the Cannelton slates, described below as Erpetosaurus (Tuditanus) sculptilis Moodie.

Erpetosaurus sculptilis Moodie.

Type: Specimen No. 12,315, University of Chicago, Walker Museum.

Horizon and locality: Cannelton slates, Pennsylvania (Upper Freeport). (PLATE 18.)

There is preserved in the collections of Walker Museum of the University of Chicago a small amphibian skull pressed flat on a slab of slate from Cannelton, Pennsylvania. It formed part of the Hall collection acquired by the University of Chicago in 1908.

The specimen presents only a portion of the skull and fragmentary pectoral plates. The skull is wider than long and the muzzle is broadly rounded. The orbits are narrow ovals and their posterior border falls on the transverse line dividing the skull equally. The interorbital width is slightly greater than the width of the orbits and about equal to their length. The posterior outline of the skull is somewhat truncate, as in E. tabulatus Cope and other species of the genus. The distal extremities of the quadrates do not project as far backward as do the supraoccipitals. The skull roof is formed of the regular elements, except that a quadrate seems to be indicated by a scale of bone on the posterior angle. The nostrils are oval and the pineal opening is small.

The premaxillæ are apparently relatively large elements, though their boundaries are not definite. The nasal is of an oblong shape and borders the frontal anteriorly. The frontal forms the whole of the interior border of the orbit and borders the parietal broadly behind. The parietal is a large element and the pineal foramen is inclosed in the median suture about midway of the parietals. The postparietal is wider than long and with the tabulare forms the greater part of the posterior border of the skull. The prefrontal (plate 18, fig. 1) apparently forms the entire anterior border of the orbit and sends an acuminate projection to the side of it. The maxilla is excluded from the orbit and is an elongate element with sharp conical teeth, of which there are 4 preserved. These measure about 1 mm. in length. The jugal lies along the lateral border of the orbit and it is acuminate both anteriorly and posteriorly. It borders the supratemporal broadly. The postfrontal forms the greater part of the posterior boundary of the orbit. It is triangular and acuminate behind, and is bordered broadly by the parietal and supratemporal. The supratemporal is also triangular and it borders the parietal broadly. The squamosal is evidently the largest element in the skull and on its posterior corner there is a flake of bone which may represent the quadrate, though this is by no means certain. The quadrate has not been detected in any of the Carboniferous Microsauria so far studied. The tabulare is an elongate element in the transverse line of the skull. Its entire boundary is uncertain, though part of the sutures are present. The quadratojugal is elongate and lies posterior to the maxilla and with that element forms the lateral boundary of the skull.

The canals of the lateral-line system have not been detected on the skull. The sculpturing of the cranial elements consists of grooves and ridges which radiate from a center. They are more prominent on the parietals than elsewhere, although the other skull elements present a strong sculpturing.

There are also preserved on the slab of slate, about 10 mm. posterior to the skull, fragments of the pectoral plates, probably representing the clavicles and the interclavicle. They are so badly fractured that their form can not be determined. No limbs or vertebræ have been observed.

Pectoral Girdle Provisionally Associated
With Erpetosaurus sculptilis Moodie.

The present specimen is preserved on a block of slate from Cannelton, Pennsylvania. It is associated with the previously described Erpetosaurus sculptilis Moodie on account of its size, the geological and geographical distribution, and the character of the sculpture. It may pertain to an unknown species. Other remains besides the 3 elements of the pectoral girdle are preserved on the block of slate, but they are, for the most part, too imperfectly preserved for recognition. Some of them are phalanges, and I believe I detect a scapula in the rounded curved plate lying near the right clavicle. The 3 pectoral elements, the interclavicle and the 2 clavicles, are preserved intact, with the ventral surface uppermost.

The specimen is particularly important in that it furnishes further evidence of the simplicity of the microsaurian pectoral girdle, which Jaekel regarded (347) as being extremely complex, in one species at least, Diceratosaurus punctolineatus Cope. The 3 elements are broken, but either the elements or their impressions are present, so that identification is possible. The elements are sculptured with radiating grooves and ridges, as in so many of the Microsauria. The interclavicle is spatulate and bears a general resemblance to the same element of Metoposaurus fraasi Lucas from the Triassic (383) of Arizona. The clavicles are triangular, with rounded angles and the hypothenuse on the interior border.

MOODIE

PLATE 14

1 and 2. Photograph of the specimen of Amphibamus grandiceps Cope, from the Mazon Creek shales. × 1.5. Original in possession of Mr. L. E. Daniels, Rolling Prairie, Indiana.

3. Specimen of Sauropleura (Colosteus) scutellata Newberry, from the Linton Coal Measures. The species was the first known of the Ohio Coal Measures Amphibia; at first ascribed by Newberry to the fishes, but later correctly identified by Cope. × 1. Original in American Museum of Natural History.

4. The type of Diceratosaurus (Ceraterpeton) punctolineatus Cope, from the Linton Coal Measures. × 1. Original in American Museum of Natural History.

Erpetosaurus acutirostris Moodie.

Type: Specimen No. 8598 G, American Museum of Natural History.

Horizon and locality: Linton, Ohio, Coal Measures.

The present species adds another form to the diversity of structure presented by the Carboniferous Microsauria. It is closely allied to Erpetosaurus (Tuditanus) obtusus Cope, from the same beds (Linton, Ohio), but differs from it especially in the position and shape of the orbits and the acute form of the skull. Other characters which amount almost to generic significance are found in the posterior prolongation of the frontal and in the triangular form of the skull. Only the skull of the animal is preserved. The character which is common to all members of the genus Erpetosaurus, the cranial rugosity, is present in this species on the squamosal and supratemporal. This character alone would not, however, suffice to separate the form generically, but the general morphology and arrangement of the cranial elements is such that reference to any other genus save Erpetosaurus would not be possible.

The skull of E. acutirostris takes the form of a rounded triangle. Its base is some 50 mm. in extent, and this width gradually narrows to 31 mm. across the orbits and still more towards the snout. The form of the skull is not widely different from that of the type species, E. radiatus Cope, but the differences are sufficiently apparent.

Nearly all the elements of the cranium can be detected (fig. 22, D). The bony portion of the cranium has nearly all been lost, leaving only the impression; and the matrix in which the skull was embedded has been forced up into the sutures between the cranial elements, thus forming ragged ridges where the bones of the skull joined.

The position of the nostrils can not be determined accurately. The orbits are placed well forward, a character common to several species of the genus. The interorbital space is equal to the long diameter of the eye. The orbits are separated by narrow prolongations of the postfrontals and by the anterior portion of the frontals. The frontals are remarkable in their great backward extension. In E. obtusus the frontals are nearly confined to the interorbital space. The parietals, which, on the median suture, inclose the parietal foramen, lie well posterior, and the parietals and the tabulare are small. A portion of the sculpturing of these elements has been preserved and it is seen to be made up of pits and elevations much as we find in the skull of Saurerpeton latithorax Cope. The left squamosal also shows sculpturing, which here tends to take the form of grooves and ridges, and also of pits and elevations. It is quite probable that the anterior portion of the skull was ornamented with grooves and ridges and undoubtedly the lateral-line canals were well developed. The postfrontals and the postorbitals are both large and elongated. The postorbital is especially large. The supratemporal apparently separates the tabulare and the squamosal in their posterior extremities. The squamosal projects posteriorly to the tabulare and apparently even goes beyond the limits of the exoccipital. The outlines of the jugal are fairly definite, as are also the limits of the maxilla.

An additional specimen shows that the skull, of which the anterior half is preserved, is practically of the same size as the type and shows much the same characters, though more extensively. The sculpture of the squamosal region is not confined to that portion of the skull, but extends throughout the cranial elements, apparently including the premaxillæ.

The sutures, which are clearly distinct, are of the same type as has been described for the type specimen. Perhaps one of the most interesting characters discovered on the present specimen is that of the greater part of the left supraorbital lateral-line canal, which is exhibited as a rather deep and broad canal running in a slight curve across the lower edges of the postorbital and the parietals and partly cutting the jugal.

On the left of the fossil, as it is preserved, there is an indistinct impression of the clavicle, with the sculpture in radiating lines from the apex as a center, such as have been found in other species of the Tuditanidæ.

Still a third specimen of this species is possibly represented by a nearly complete skull of a small individual which exposes the mandible and the ventral portion of the skull. The remains are crushed flat, though not at all distorted. It is quite possible that the present specimen represents a distinct species, but only a small portion of the dorsum is present and the shape of the cranium is indistinct, so it is retained in this species. The portion of the skull shows the sculpture to be quite similar to that of Erpetosaurus acutirostris, so far as the species is known; and the skull apparently tapers to a point anteriorly. It may be a dwarfed or immature form. The sculpturing of the jaw is such as we would expect of this species.

The specimen is about half the size of the type. The palate of the skull is well preserved and is extremely interesting. The sutures separating the various palatal elements are not distinct. The parasphenoid is especially large and the exoccipitals are partly ossified, if we may judge by the projecting condyles. Anteriorly the parasphenoid contracts and then expands and on each side of the expanded part lie fragments of the palatines. To the right of the posterior end of the parasphenoid lies a portion of the dorsal element showing the cranial sculpture.

The left mandible is somewhat displaced to the right of the skull, and crushed and weathered to such an extent that the sutures are entirely obliterated. There are 3 teeth, with indications of others. They are typically pleurodont and sharp and slender. The mandible tapers somewhat anteriorly and at the tip bears an elongate enlarged tooth.

Erpetosaurus tuberculatus Moodie.

Type: Specimen Nos. 8693 G and 8610 G, American Museum of Natural History.

Horizon and locality: Linton, Ohio, Coal Measures.

This species is based on a fragmentary cranium (plate 26, fig. 1) consisting of the posterior part of the right side of the skull. Its association in the genus is solely on the character of the sculpturing of the cranial elements. It is most closely related, in the characters preserved, to the form described by Cope as Tuditanus radiatus, from which it differs especially in the character of the sculpture and in the position of the orbits, as well as the arrangement and size of the various cranial elements, so far as these elements can be detected in the present specimen. In Erpetosaurus radiatus the skull is sculptured by radiating grooves and ridges which did not arise from a definite center. In E. tuberculatus this center of radiation is marked by an elevation or tubercle on each cranial element exposed, from which the grooves and ridges radiate outward. These tubercles have an elevation of 4 mm. above the cranial element proper. The orbit is located near the median line of the skull, so far as can be determined. In E. radiatus Cope the orbits are located well forward. In that species also the postparietal is smaller than in the present species and the squamosal is longer and more slender. (Plate 25, fig. 1.)

The fragment of a skull on which the above comparison has been made consists of the right postparietal, a portion of the tabulare, the parietal, the frontal, and a portion of the squamosal. The other elements are not clear. The elements in the median line are elongate, as in Erpetosaurus radiatus. The pineal foramen is located well back on the median line and lies posterior to two-thirds of the length of the parietals. The sutures separating the frontal and parietal elements from each other in the median line are of the zigzag form so characteristic of the labyrinthodonts.

Genus ODONTERPETON Moodie.

Type: Odonterpeton triangularis Moodie.

The generic characters may be found in the triangular shape of the skull, the large size of the teeth, the shape of the vertebræ, the small size of the orbits, and their anterior position as shown in the type specimen (fig. 22, E). The name of the genus is derived from the remarkable size of the teeth as compared with the size of the skull.

Odonterpeton triangularis Moodie.

Type: Specimen No. 4465, U. S. National Museum.

Horizon and locality: Linton, Ohio, Coal Measures.

The specimen described under the above name is a representative of the smallest species of the Microsauria so far described from North America. Orthocosta microscopica Fritsch, from the Carboniferous of Bohemia, is a rival of the present form in size, but the form described by Fritsch is an entirely different animal and was formerly included among the so-called Aistopoda, which are regarded by the writer as merely specialized microsaurians. The present form shows clear affinities with the Microsauria.

As may be seen by referring to the list of measurements, the skull of this form measures only 6.5 mm. in length. The form may possibly be larval, though I do not think so, if I may judge from the well-developed condition of the skull bones and the complete ossification of the vertebræ. The sides of the skull are equal and the occiput is a straight table, so that the skull forms almost an exact equilateral triangle. The orbits are very small and are placed well forward. The interorbital space is four times that of the diameter of the orbit, a very unusual character and in itself worthy of ranking as a generic character. The median suture of the skull is zigzag and incloses the minute parietal foramen near the posterior end of the skull. The relations of the elements of the skull, with the exceptions of those of the frontals and parietals, can not be determined with accuracy, although there are here and there indications of sutures. The characters of the cranial elements, so far as they can be determined, are those of the family Tuditanidæ, and the form may, for the present, be regarded as a member of that group. The teeth are very long, slender, and sharp, and are placed close together. There is no indication of fluting on the teeth. They are slightly curved inward.

There are 13 vertebræ present. The centra are hour-glass shaped, and are apparently phyllospondylous, with the notochord largely persistent. The vertebral centra are unusually long and slender, with the ends rounded. The humerus of the right side is preserved. It is a long, slender bone with expanded extremities. There is no evidence of abdominal armature nor of ribs (fig. 22, E).

The discovery of this form in the Linton deposits is of considerable interest in that it indicates a wide range in size and character of the fauna of the time. The forms now known from the Linton beds range from Odonterpeton, which possibly had a total length of 2 inches in life, to the form designated Macrerpeton huxleyi Cope, with a skull of at least 8 inches in length and whose body may have attained a length of some feet. The large rib described below undoubtedly indicates a large form of the ancient Amphibia from Linton, as do also the vertebræ described by Marsh in 1863 from Nova Scotia.

THE MICROSAURIAN FAMILY STEGOPIDÆ, FROM THE COAL MEASURES OF OHIO.

Family STEGOPIDÆ Moodie, 1909.

The chief family characters are the large lacrimal, unknown in other species of Coal Measures Amphibia, the central position of the orbits, the general form of the skull, and the peculiar, short, divaricate horns from the squamosal. If an intertemporal element is present in the skull, which is suggested as a possibility, the family is further distinct. The type species is Stegops divaricata Cope from the Linton, Ohio, Coal Measures.

The group seems to be distinct and has no immediate allies, being confined to the American Coal Measures.

Genus STEGOPS Moodie, 1909.

Type: Stegops divaricata Cope.

This genus has been erected for the reception of the peculiar form described by Cope as Ceraterpeton divaricatum, but there are good reasons why the form can not be retained in the genus. The position of the orbits in Stegops (plate 25, fig. 3) is different from Ceraterpeton and the muzzle is rounded, not truncate as in the latter form. The horns are of a different type and there is no indication of the tabulare protuberance which is present in Ceraterpeton. The sculpturing of the cranial elements is also distinctive in the present form, consisting of radiating grooves and ridges; the cranium of Ceraterpeton appears to be but slightly sculptured. There is no lateral projection from the border of the skull in Stegops, as there is in the other genus. The structure of the skull of Ceraterpeton is practically unknown, except in a very general way, although Andrews (8) was able to make out some of the elements and to trace the lateral-line canals. A structural comparison is thus impossible, but on the basis of form alone there are good generic distinctions. The present genus is apparently distinct from other genera in the presence of an intertemporal, but additional material will be required before a satisfactory determination is possible. The genus Diceratosaurus of Jaekel (347) is distinct in the arrangement of the elements of the cranium, the general form of the skull, and in the two known species of Diceratosaurus the orbits are located well anteriorly, but in Stegops they are in the median transverse line of the cranium. The genus Stegops is distinct from Eoserpeton in the smaller size of the prosquamosal, in the broadly rounded muzzle, in the larger and more posteriorly placed orbits, and in the presence of an intertemporal bone, or at least in the elongate character of the postorbital if the intertemporal is not present. The species on which the genus Eoserpeton is based was first described by Cope as Ceraterpeton ten n iconic. The form is quite distinct, in spite of Jacket's protestations to the contrary. The genus Stegops stands alone among the Carboniferous Amphibia of North America, so far as I am aware, in the possession of a well-defined lacrimal of the labyrinthodont type.

Stegops divaricata Cope.

Type: Specimen No. 2559 G, American Museum of Natural History. The obverse of this is No. 12,311, Walker Museum, University of Chicago.

Horizon and locality: Linton, Ohio, Coal Measures.

The skull on which this species is based consists of the impressions on two slabs of coal, one belonging to the Newberry Collection of the American Museum of Natural History, No. 2559 G, and the other to the Gurley Collection of the University of Chicago, No. 12,311. The slab belonging to the University of Chicago contains the better-preserved remains, so that the description is based largely on that portion (plate 25, fig. 3). Nearly all of the elements of the skull are determined with considerable certainty and many important characters in the morphology of the Microsauria are thus brought out.

Fig. 23.—Skull elements of Stegops divaricata Cope., fr, frontal; it, intertemporal; j, jugal; la, lacrimal; mx, maxilla; n, nasal; par, parietal; pf, prefrontal; po, postorbital; pof, postfrontal; pmx, premaxilla; sq, squamosal; spt, supratemporal; qj, quadratojugal; tab, tabulare; pp, postparietal. × 1.3.

The skull is oval, elongate, truncate behind, and the quadrate angles project into sharp horns. The orbits are elongate ovals and their center lies in the median line which divides the skull transversely. The nostrils are elongate and have an oblique position. The pineal foramen lies in the posterior third of the skull.

Teeth are preserved on both maxillæ and premaxillæ. They are simply sharp pleurodont denticles, and seem to have been fairly abundant. The bones have been completely carbonized and nothing of the original texture is preserved, although the details of the structure are beautifully preserved. (PLATE 25, fig. 3.)

The skull is somewhat triangular in its general form. The premaxilla lies on the anterior border of the cranium, and forms the median border of the nostril. The suture which separates the maxilla and the premaxilla is not evident, and it may not be correctly defined in the figure (fig. 23). The nasal is a very large element and is elongate. It unites with the premaxilla, the lacrimal, the prefrontal, and the parietal. It is acuminate behind and the point is inclosed by the pref rental and the parietal. The frontal is quite narrow and elongate, and does not border the orbit; its posterior boundary is not accurately represented. The radiations on the surface indicate the extent of the element. The parietals are remarkable in being smaller than the frontal and nasal. The pineal foramen is inclosed in the median suture in the anterior third of the parietals. The parietal has the usual relations of that element. The postparietal lies posterior to the parietal, but a portion of its bounding sutures are destroyed. The prefrontal forms the antero-interior border of the orbit and borders the postfrontal posteriorly, in an unusual manner. The lacrimal is a large element and is clearly separable from the other cranial elements. It, with the prefrontal, forms the anterior border of the orbit. The maxilla is very elongate and forms the larger part of the lateral border of the skull. Sharp-pointed teeth are present on it and they may have been pleurodont. The lateral border of the orbit probably received a portion of the maxilla. The postfrontal and the postorbital form the greater part of the posterior boundary of the orbit. The postorbital seems to be divided by a median suture which would indicate an intertemporal bone, but this is not certain. The appearance may be due to a fracture. The supratemporal is a large element bordering the parietal and lies in front of the tabulare. The squamosal is elongate to form the posterior horn-like extension of the skull. The tabulare is transversely elongate and has the usual relations. The jugal widens to a fan-shape backwards, and helps to form the lateral border of the orbit. Its lateral and anterior boundaries are not assured. The quadratojugal seems to lie as indicated, although the anterior part of the suture is not distinct. It is apparently an elongate element and with the maxilla forms the lateral border of the skull. The base of the skull as restored (fig. 23) is irregular and may have had a slightly different form.

The genus Stegops is exceptional in the elongate character of the cranial elements of the single species known. In this respect it recalls the species Diceratosaurus lævis described below. The large size of the nasals, frontals, and lacrimals and the small size of the parietals are, so far as I am aware, unparalleled among the other Coal Measures Amphibia of North America.

MOODIE

PLATE 15

1. Dorsum of skull of Diceratosaurus punctolineatus (Cope), from the Coal Measures of Linton. Ohio. Original in the Museum at Berlin University. × 2. After Jaekel. po=postorbital: fr= frontal; pof=postfrontal; j=jugal; la=lacrimal; mx = maxilla; ps="perisquamosal;" pp=post-parietal; pmx= premaxilla.

2. Ventral surface of the skull of Diceratosaurus punctolineatus (Cope ), from the Coal Measures of Linton, Ohio. × 2. After Jaekel. a=anterior palatine vacuity; ex=exoccipital; j=jugal; mx=maxilla; tr=transverse; pal=palatine; pt=pterygoid; pr=prevomer; ps="perisquamosal;" pv=posterior or suborbital palatine vacuity; ph=parasphenoid.

3. Pectoral girdle of Diceratosaurus punctolineatus (Cope), from the Coal Measures of Linton, Ohio. Original in the paleontological Museum at Berlin. × 2. After Jaekel. cli=inner side of clavicle; ic=interclavicle; cl=clavicle (lower side); cle=cleithrum.

4. (e) Cervical or anterior dorsal vertebra of Diceratosaurus punctolineatus (Cope), from the Linton, Ohio, Coal Measures. cv=caudal vertebra shown with its cap of dermal bone; dv=dorsal vertebra with ribs; x=dermal plate on neural spine; h=humerus; u=ulna; r=radius.

THE MICROSAURIAN FAMILY UROCORDYLIDÆ, FROM THE COAL MEASURES OF OHIO.

Family UROCORDYLIDÆ Lydekker, 1890.

Type of the family: Urocordylus wandesfordii Huxley.

Locality and horizon: Coal Measures of Kilkenny, Ireland.

There are 4 genera which constitute this family: Urocordylus, from the Coal Measures of Ireland; Ceraterpeton, from the Coal Measures of Ireland and England; Diceratosaurus, from the Coal Measures of Linton, Ohio; Eoserpeton, from the Coal Measures of Linton, Ohio.

These may be distinguished by the following characters:

The relationships of the family are not far to seek. They fall in immediately with the Amphibamidæ and Hylonomidæ in being among the most reptile-like of the Paleozoic Amphibia. The group is, however, distinctly amphibian in the possession of 4 fingers, with the usual microsaurian phalangeal formula.

Genus DICERATOSAURUS Jaekel, 1903.

Type: Diceratosaurus punctolineatus Jaekel.

The most important differences between Diceratosaurus and Ceraterpeton, the most nearly allied genus, is (in Diceratosaurus) in the more anterior position and small size of the orbits, the backward extension of the quadrate region, and the dorsal expansion of the vertebral spines. A further, and more important, difference between the genera is in the location of the backwardly directed processes from the skull. In Ceraterpeton they project backward from and are a portion of the tabulare element, while in Diceratosaurus the projection consists almost entirely of squamosal and supratemporal.

Diceratosaurus punctolineatus Cope.

Type: Specimen No. 8606, American Museum of Natural History.

Horizon and locality: Linton, Ohio, Coal Measures.

The species was first described by Cope as Ceraterpeton punctolineatum (122). It was redescribed on more abundant material (plate 19) by Jaekel (347), and the following is taken from the discussions of these two authors, checked by my own observations on the type specimen. This shows a portion of the skull, consisting of the squamosal and supratemporal with a projecting, convergent horn. The sculpturing on the skull is similar to that on the pectoral plates, of which there are three preserved (plate 14, fig. 4). The bones of the fore limbs are stout and short. The ribs are only slightly curved. The character of the vertebræ can not be ascertained. The sculpturing of the bones consists of radiating ridges, grooves, and pittings.

Jaekel (347) described from the museum collection at Berlin 3 specimens of this species, among which were 2 skulls. There were associated with these remains some pectoral plates and limb bones with a nearly complete series of vertebræ. A modified translation of Jaekel's description follows:

The skull of the largest specimen has, including the horns, a length of 35 mm. and a width of 30 mm. on the occipital border. From the anterior end to the posterior border of the skull (exclusive of the horns) there is a length of 25 mm. The pineal foramen lies about midway of this length. The orbits are rather large, almost circular, and lie about midway between the pineal foramen and the anterior border of the skull. The nostrils, which lie anterior to the orbits, are a rather oblique oval and narrowed on the lateral ends. The distance between them is about the same as that between the orbits, which measures 7 mm.

MOODIE

PLATE 16

1. Type specimen of Diceratosaurus punctolineatus Cope, from Linton, Ohio, beds. × 1. Drawn from photograph. cl=clavicle; h=humerus; ic=interclavicle; mc=metacarpals; r-u=radius-ulna; rb=rib; sc=scapula; sq=squamosal; vs=ventral scutellæ.

2. Skull of Sauropleura longidentata, Moodie, from the Coal Measures of Linton, Ohio. × 1. Drawn from photograph. fr=frontal; m=maxilla; n=nasal; or=orbit; par=parietal; pp=postparietal; tab=tabulare.

3. Mandible of Sauropleura longidentata, Moodie, from Coal Measures of Linton, Ohio. × 1.5. Drawn from photograph.

4. Type specimen of Sauropleura enchodus Cope, from the Coal Measures of Linton, Ohio. × 1.

5. Additional specimen of Diceratosaurus punctolineatus Cope, from the Coal Measures of Linton Ohio. × 1. f=femur; il=ilium; mc=metacarpal; mt=metatarsal; ph=phalanx; r=radius; t=tibia.

The skull roof is sculptured with pits much like those of Archegosaurus. The larger the bones, the rougher the sculpture. The bones of the middle of the skull (that is, the parietals, frontals, premaxillæ, nasals, and postparietals) are of the normal form and only show an unusual difference in that they are large in the inverse order. The pineal foramen lies in the anterior third of the parietals; that is the primitive condition which occurs in the young forms of the Branchiosauridæ. The nostrils are inclosed by the premaxillæ in front, in the median line by the nasals and laterally by the maxillæ. The jugals, by their backward prolongation, form, behind the maxillæ, the border of the skull, and only attain to some size on the upper side of the cranium.

Jaekel's "perisquamosal" (see plate 15), which is of a doubtful nature, is not indicated in the type specimen nor in the specimens of the other two species (462) assigned to this genus. In D. robustus and D. lævis the "perisquamosal" region is easily separable into its component elements. The sutures between the elements may have been indistinct in his specimen, but it is hardly conceivable that a union of the skull bones would occur in one species and not in another of the same genus. Jaekel's suggestion (347) that "Etwas mehr Wahrscheinlichkeit möchte ich der Vorstellung beimessen, dass diese Ausbreitungen zum Schutz freier Kiemen dienten, wie sie z. B. bei den Perennibranchiaten als baumförmige Organe weit am Halse herausragen" can hardly find acceptance with students of the Paleozoic Amphibia, since there is not the slightest evidence that the Microsauria ever possessed external gills and considerable presumptive evidence that they did not. His comparison of the "perisquamosal" to the "Kiemendeckel" of the fishes is also very unhappy on morphological grounds, since the elements of his "perisquamosal" form constituent parts of the skull roof, which the operculum never does.

The palate of the skull (plate 15, fig. 2) has been determined by removing the skull bones of one specimen. Anteriorly the premaxillæ and maxillæ are clearly recognizable as large dentiferous elements. The premaxillæ have 4 to 7 teeth, the short maxilla has 3 to 4. All the teeth are of nearly equal size. Smaller teeth seem to be indicated by impressions found between the larger ones. The vomers, which are tolerably large, unite with the premaxillæ behind and inclose at least half of the palatine foramen on the inner side. They are furnished with small teeth, which in the anterior part are very irregularly placed, but they are more regular posteriorly.

The palatines and transverse bones are questionably identified. They seem to lie posterior and lateral to the vomers, but the sutures are indistinct. The large parasphenoid seems well displayed and is more or less heart-shaped. There would seem to be a slight indication of double occipital condyles. The pterygoids are broad plates which inclose the parasphenoid and form the lateral boundary of the palate. The cotyli are very indistinct, but appear as elongate grooves.

The pectoral girdle (plate 15, fig. 3) consists, apparently, of seven elements, three paired and one unpaired. The unpaired element (the interclavicle) is truncate posteriorly and acuminate in front, with its surface radially grooved and the anterior borders beveled for articulation with the clavicles. The clavicles are triangular, as is usual with the Microsauria. They are sculptured with radiate grooves and ridges, with decided inosculations at the ossific center. The coracoids have only part of their surface ornamented; most of their surface is smooth for articulation with the interclavicle and scapula. A long spine projects from the inner surface of the coracoid. A pair of small elements lie one on either side of the clavicles and Jaekel (347) interprets them as the cleithra. If they are cleithra they are unique among the Microsauria. The pectoral girdle does not, however, indicate that Diceratosaurus is "unique among all known quadrupeds."

Jaekel regards the limb which is preserved with the material as an arm (plate 15, fig. 4), but there is no reason stated for his conclusion. It has all of the characters of the leg and may be regarded as such. Only a part of the lower end of the femur is preserved. The tibia and fibula are preserved as separate rod-like elements with one of the bones longer and larger, probably the tibia. There are 5 toes which have the customary phalangeal formula for a microsaurian foot of 2-3-3-4(?)-3. The tarsals are unossified.

The vertebræ (plate 15) were perforated for the notochord, and are hour-glass-shaped, with the neural arch thickened to support a heavy spine which bore a sculptured plate. These apical plates occur in the dorsal region, but diminish toward the caudal vertebræ. The number of the vertebræ in the dorsal region is very small and in the tail very large. There are possibly 2 vertebræ in the cervical, 11 in the dorsal series; the thirteenth carries the pelvis. There are over 100 vertebræ in the tail.

The ribs have an expanded head and the transverse processes of the vertebræ are long.

The following account is based on the writer's study of the type specimen and he is able to add several points of interest to a knowledge of the anatomy of the type of this interesting microsaurian.

The type specimen consists of 11 consecutive vertebræ with a portion of the skull, the greater portion of the pectoral girdle, parts of both fore limbs, ribs, and ventral scutellæ (plate 14, fig. 4). The species is represented in the collection by yet another specimen, on which Cope based his Tuditanus mordax (plate 22, fig. 5), of which he himself says: "Further examination of the specimen on which the latter (T. mordax) was founded leads to the belief that it is an imperfect cranium of Ceraterpeton (Diceratosaurus) punctolineatum Cope." The plates referred to are rather to be regarded as elements of the pectoral girdle and I believe they represent the clavicle and a portion of the interclavicle.

The skull of the present species is fully described by Jaekel. The type specimen does not offer any evidence in support of Jaekel's "perisquamosal," but rather tends to the idea that he is incorrect in his assumption of the fusion of these elements of the skull. The direction taken by the ridges and grooves on the elements preserved indicate a separation between the supratemporal and the squamosal. I do not find that the grooves have the tendency to arise from a common center of ossification in the squamosal, as suggested in the figures of Jaekel. The horn which projects backward from the squamosal is rather large and heavy for the size of the skull, and after curving slightly inward ends in a blunt point and not sharply, as Jaekel figures in his specimens. The vertebral column is indistinctly preserved and I have nothing to add to Jaekel's account given above.

In the structure of the pectoral girdle my results are greatly at variance with those of Jaekel. I do not find the remarkable elements which Jaekel has figured (347) in his specimens. On the other hand, I find a normal microsaurian pectoral arch (464), such as has been described for numerous other forms. There are present, distinctly preserved in the type specimen, the scapulæ, the clavicles, and the interclavicle, with the possibility of the coracoid. The peculiar element referred to by Cope as resembling a "lacertilian pubis" is without doubt the left scapula of the animal (plate 16, fig. 1). Its form compares very favorably with that of Ceraterpeton as figured by Woodward (630). The coracoid may be represented by the fragment which lies close to the scapula. The sculptured element lying next to the supratemporal horn of the skull is the right clavicle preserved bottom side up. Of the other two sculptured elements, one is the interclavicle, only a portion of which is preserved. The left clavicle lies beside it. The clavicles in this species have a tendency to assume the triangular shape so common in other species of Microsauria, and the interclavicle, so far as can be determined, was shield-shaped. The upper surfaces of the pectoral elements are marked by grooves for the attachment of the pectoral muscles.

Fig. 24.

A. Skull of Diceratosaurus lævis Moodie, from the Linton Coal Measures. × 1. f, frontal; j, jugal; mx, maxilla; n, nasal; or, orbit; par, parietal; pof, postfrontal; po, postorbital; pf, prefrontal; pp, postparietal; sq, squamosal; spt, supratemporal; qj, quadratojugal; pmx, premaxilla; tab, tabulare.

B. Reconstruction of skull outlines of Diceratosaurus robustus Moodie, from the Coal Measures of Ohio. × 0.75. fr, frontal; f, jugal; or, orbit; par, parietal; pof, postfrontal; po, postorbital; pp, postparietal; qj, quadratojugal; spt, supratemporal; tab, tabulare.

The ventral scutellation is present in a small patch (plate 16, fig. 1) near the horn of the skull. The scutæ are oat-shaped and take the usual form. The ribs are not long, are rather stout, and beyond the proximal curve are nearly straight to the obtuse tips. The heads of the ribs are so obscure that it is impossible to determine whether they were two-headed or not. They are expanded proximally and there is a slight tendency to a division of the head.

Portions of both fore limbs are preserved. The right limb possesses the humerus, separate radius and ulna, and 2 metacarpals. The other possesses only the radius, 3 metacarpals, and a portion of a phalanx. The humerus is a very stout bone and at once recalls that of Amblyrhynchus. The ends are expanded and there are roughnesses on the bone for the attachment of muscles. The radius and ulna are subequal in size. They are both expanded more proximally than distally. The carpus was cartilaginous. An additional specimen of this species is figured on plate 16, fig. 5. This adds to our knowledge of the pelvis especially.

The specimen (No. 2566, Am. Mus. Nat. Hist.) on which Cope based his Tuditanus mordax is composed of two plates of the above-described species.

Diceratosaurus lævis Moodie.

Type: Specimen No. 102 (8680 G), American Museum of Natural History, where it forms part of the Newberry collection.

Horizon and locality: Linton, Ohio, Coal Measures.

The species is represented by an almost complete skull, which had been identified previously by Cope as Tuditanus radiatus. The specimen consists of the impressions of the bones of the cranial roof, the bones themselves having disappeared. It is not probable that the dorsum of the skull was smooth. The details in the structure of the skull have been ascertained quite definitely. There can be no doubt that the arrangement of the elements is accurate, as shown in figure 24 A. The supratemporal, as in Erpetosaurus tabulatus Cope, is excluded from the parietal.

The form of the skull at once recalls that of the species D. punctolineatus, as figured by Jaekel (see plate 15). The orbits are located in nearly the same region of the skull and the sutures separating the cranial elements are quite similar in the anterior portions. The species D. lævis is based on the divergent character of the horn-like protuberances which project from the squamosals. The horns of D. punctolineatus are convergent. The present skull is also smaller and the parietals in D. lævis are much larger than in the type species. In the type species, also, the pineal foramen is located well forward in the parietal, while in the present form the foramen is located well posterior.

The skull is almost rectangular. The nostrils are elongate ovals. The orbits are circular and the distance between them is equal to two-thirds of the dimensions of the orbit. They are located well forward in the skull and are bounded laterally by the maxillaries. The nostrils have much the same character as in the type form, being broadly oval.

The premaxillæ are elongate transversely, being about twice as long as wide. They are identical in shape and relations with the same elements in D. punctolineatus Cope. The nasal is nearly square and forms the interior boundary of the nostril. The frontal is elongate in the median length of the skull and it is acuminate posteriorly, where the acumination is inclosed by the parietal and postfrontal. The parietals are by far the largest elements in the cranium. They form together an oval which is elongate in the longitudinal diameter of the skull. They inclose between them, in the median suture, the small pineal foramen. They are acuminate in front, with a broad truncate posterior base, where they are bounded by the postparietals. The postparietal is nearly square, being somewhat wider than long. It joins the tabulare and the parietal. The tabulare is elongate in the long diameter of the skull. It ends anteriorly in a point which is inserted between the postorbital and the parietal, and bears a short protuberance posteriorly, much as does the same element in the type species.

There are four elements which take part in the formation of the posterior border of the skull. These are the postparietal, the tabulare, the squamosal, and the supratemporal. It is very unusual for the supratemporal to reach the posterior edge of the cranium. The pref rental lies anterior to the orbit, of which it forms the anterior border. The lacrimal has not been detected, although Jaekel (347) has indicated it in his drawings of the skull of the type species. The maxilla is elongate and forms the lateral border of the skull. No teeth have been detected, although they were doubtless the same as Jaekel has figured in D. punctolineatus. The jugal is an elongate element joining the maxilla posteriorly. Jaekel included this element in his "perisquamosal," but the sutures are clearly evident in the present specimen and there is no evidence of a structure at all similar to a "perisquamosal." The postorbital is fully as large as the jugal which it joins, forming a part of the posterior border of the orbit and ending posteriorly in a point which is inclosed by the tabulare and the squamosal. The postfrontal with the foregoing element forms the entire posterior border of the orbit and it likewise ends in a point inclosed by the parietal and the postorbital. The quadratojugal has much the same shape and relations as in D. punctolineatus, although it is located further back. The squamosal is also elongate, as are most of the posterior cranial elements, and it also has an acumination which is directed forward and is inclosed by the postorbital and jugal. The anterior suture of this element is rather indistinct, but it is, I believe, as represented (fig. 24). The element is elongate and is prolonged posteriorly to form the horn, which ends in a blunt point and is not sharp, as in the type species.

Jaekel (347) regards the species Diceratosaurus punctolineatus Cope as being unparalleled among known vertebrates in the possession of a "perisquamosal" element. In closely allied species the "perisquamosal" is easily separated into its component elements, and the morphology of the present skull would throw considerable doubt on Jaekel's interpretation of the skull of the type species. Another specimen, described below as another species of this genus, shows no evidence of this fusion. So far as I can learn, there have been no cases of true fusion of cranial elements correctly reported, unless it be that which possibly exists between the two frontals in the skull of Diplocaulus. It was on the basis of such fusions that Maggi (397) proposed to derive the interparietals of the primates from the tabulare of the stegocephalians.

The posterior outline of the skull in the present specimen is not well preserved and the outline as given may be slightly inaccurate. The indentation figured by Jaekel in the posterior border of the skull of the type form is not present in the species under discussion.

Diceratosaurus robustus Moodie.

Type: Specimen No. 8611 G, American Museum of Natural History.

Horizon and locality: Linton, Ohio, Coal Measures.

The present species is indicated by the left portion of a cranium representing a large individual. The characters of the skull are so clearly marked that it seems worthy of description. The presence of horns as given in the restoration of the skull (fig. 24, B) is based on the analogy with the other two species of this genus, in both of which horns are present. The generic determination of the species is based on the large size of the postorbital, which is essentially characteristic of the other species of Diceratosaurus.

The characters which distinguish the species from others of the genus are the large postorbitals and the small parietals, which are excluded from union with the postfrontals on account of the large size of the frontal. In the other two known species the frontal is small and the parietal comes forward to join the postfrontal. The present species exhibits a skull which is nearly twice as large as that of D. lævis and nearly three times the size of the skull of D. punctolineatus.

The portion of the skull preserved shows the cranium to have had a rather acuminate snout, not blunt as in the type species. The orbit is an elongate oval, although it has the same relative position in the skull as in the other species. The nostril is indicated by an oval depression near the anterior edge of the skull. The frontals, as indicated by the sutures present on the portion of the skull which is preserved, are fully as long as the parietals. Whether they were as wide as is represented is uncertain. The postfrontals are very small bones, the sutures of which are somewhat uncertain, although they can not be far from what is represented (fig. 24, B). The postorbital is large and elongate, and is distinctive of this species on account of its unusual size, although it does not attain the same proportions as in other members of the genus. The parietals are elongate and narrow. The pineal foramen is represented by its lateral edge and its position is about midway of the longitudinal diameter of the parietals. The narrow postparietal is represented by its anterior border; as restored (fig. 24, B) it may be too long. The tabulare, also, is represented by an anterior portion and it shows this element to have the position and form which is typical of the form Diceratosaurus lævis. Such other of the cranial elements as are indicated are based on the relations discovered in D. lævis.

The heavy line on the left of the drawing (fig. 24, B) represents the outline of the preserved portion. The skull, as restored, may be a little too long, and the shape of the horns is conjectural. In the orbit there are preserved 2 misplaced teeth showing longitudinal fluting. The longest tooth is about 3 mm.

Genus EOSERPETON Moodie, 1909.

Type: Eoserpeton tenuicorne Cope.

The genus was proposed for the reception of a single species originally referred by Cope to Ceraterpeton (C. tenuicorne). The species can not be placed in the genus Ceraterpeton on account of the form and structure of the skull, which varies widely from that of the type species, Ceraterpeton galvani Huxley, from the Kilkenny Coal Measures of Ireland. The most important character in which the present species differs from C. galvani is the peculiar form taken by the squamosal and by the position of the "horn." These characters will be evident on referring to figure 25. No undoubted remains of Ceraterpeton have been found outside the British Isles. Fritsch referred (251) a species, previously described as Scincosaurus crassus, to this genus, but Andrews (8), Jaekel (347) and Woodward (630) all unite in placing the species in the genus where it was formerly described. Jaekel even says that the Scincosaurus has no horns, so far as he can determine. Cope referred 3 species (123) from the Linton Coal Measures of Ohio to the genus Ceraterpeton, but it has been shown elsewhere (347) that no one of them belongs in the genus, nor in fact do they all belong in one genus.

Eoserpeton tenuicorne Cope.

Type: Specimen in the American Museum of Natural History. There are also specimens Nos. 4472 and 4473 in the U. S. National Museum.

Locality and horizon: Linton, Ohio, Coal Measures.

The species was founded on a complete skull preserved on obverse sides of a block of coal. Cope (123, pl. XLII, fig. 2) figured this skull. The figure is poorly executed and does not do justice to the specimen, which is really well preserved. In general the skull is oval, with the orbits located well towards the acuminate snout. The interorbital space is equal to twice the width of the orbit. The pineal foramen lies near the center of the skull. The quadrate angles are drawn out into slender acuminate, longitudinally striate horns, processes from the squamosal. The "horn" arises from an expanded base, which is a portion of the cranial element at the postero-lateral angle of the skull. This character is taken as the distinctive one of the genus. It is possessed by no other form of Carboniferous air-breathing vertebrate, in association with the other characters of the form.

Measurements of the Type of Eoserpeton tenuicorne. mm.     Median length of skull 26     Width across squamosal enlargement 30     Width across base of horns 25     Length from muzzle to tip of horn 40     Length of horn 10     Width of horn at base   4     Length of orbit   4     Width of orbit   3     Interorbital span   6   Nos. 4472 and 4473, U. S. National Museum, from Linton, Ohio, Coal Measures. mm.     Median length of skull 16     Maximum width of skull 20     Length of horn from base   7     Length of orbit   3     Interorbital width   3.5     Length of vertebral column to sacrum 33     Length of femur   5.5     Length of tibia and fibula   3     Length of second digit (incomplete)   7     Length of metatarsal   1.5     Length of clavicle   6     Width of clavicle   3.5

Fig. 25. Restoration of skeleton of Eoserpeton tenuicorne Cope. × 1 .5. Skull: pmx, premaxilla; n, nasal; fr, frontal; par, parietal; pf, prefrontal; mx, maxilla; pof, postfrontal; po, postorbital; pp, postparietal; j, jugal; qj, quadratojugal; spt, supratemporal; sq, squamosal; tab, tabulare. Skeleton: ic, interclavicle; cl, clavicle; sc, scapula; h, humerus; r, radius; u, ulna; c, carpus; sr, sacral rib (uncertain?); il, ilium; f, femur; fb, fibula; t, tibia; ts, tarsus.

The boundaries of the premaxillæ are indefinite, but what remains of the sutures indicates that the elements were small. No teeth have been detected. The nasal is likewise not clearly defined, but the frontal is an elongate element which occupies the space between the orbits and joins the parietal posteriorly. The parietals form a large oval space, so characteristic of many of the Carboniferous Microsauria, in the anterior third of which occurs the median parietal foramen. The postparietal is almost square, and forms part of the posterior boundary of the skull. The tabulare has the usual position and relations. The prefrontal is ill defined. The postfrontal is small and forms a slender rod on the postero-inner boundary of the orbit. The postorbital is small and its bounding suture with the postfrontal is indefinite. The jugal is only partially represented in the specimen, and that part forms the outer boundary of the orbit. The maxillary sutures are not defined. There are no evidences of teeth, since the skull is compressed dorso-ventrally. The quadratojugal is, apparently, a larger element than usual, with the visual relations. The supratemporal lies in great part in front of the squamosal, but still has the normal relations of that element. The squamosal is the characteristic feature of the skull. It is very tumid at its base and projects into a long, slender, acuminate horn, the tumid portion being ornamented by radiating striæ.

Another specimen of this species presents the greater part of the skeleton. However, very little can be added to our knowledge of the skull structure. It is barely possible that the second specimen may be distinct from the type. The horns are curved inward, but otherwise there is little or no difference. One of the most interesting and important features of the complete specimen is the unusual preservation of a leg, with impressions of 15 or more vertebra;, and traces of curved ribs which are intercentral in position.

The femur is slender and expanded at the ends, with the articular surfaces well formed. The tibia and fibula are mere rods of bone, although the tibia has slightly expanded extremities. There is no osseous tarsus. There are 5 digits in the foot; the second one is entire and contains 4 phalanges; the other digits are incomplete. The foot is remarkably long and slender, and is fully as long as the tarsal space plus the tibia, with the terminal phalanx clawed.

There are impressions of 2 oval and elongate clavicles in the pectoral region. The outer end is not expanded as is usual, and the surface is ornamented with grooves and ridges which radiate from a common center.

The entire remains measure scarcely 3 inches in length and it is to be doubted if the creature attained a length of more than 4 inches. It is probably a young form, but there are no evidences of external gills. The chevron armature is but poorly preserved, but so far as can be determined it is not different from that of other Microsauria, such as Amphibamus.

MOODIE

PLATE 17

Type of Saurerpeton latithorax Cope. × 1.5.
Original in U. S. National Museum.

THE MICROSAURIAN FAMILY AMPHIBAMIDÆ, FROM THE COAL MEASURES OF MAZON CREEK, ILLINOIS.

Family AMPHIBAMIDÆ new family.

Two genera are associated in this family: Amphibamus grandiceps Cope, known from three nearly complete skeletons; A. thoracatus Moodie, known from a single incomplete skeleton; Cephalerpeton ventriarmatum Moodie, anterior portion of body and skull. The species are all from the Mazon Creek shales and the family seems unrepresented elsewhere. It may be necessary to compare the Amphibamidæ with the Hylonomidæ when the latter group is better known, but in the light of our present knowledge the two families are distinct.

The genera may be distinguished as follows:

Genus AMPHIBAMUS Cope, 1865.

Type: Amphibamus grandiceps Cope.

The publication of the type species of this genus began the researches of Professor Cope on the extinct Amphibia of North America, which he continued for so many years with such excellent results (105-177). The description was based on a single specimen (plate 3, fig. 7) belonging to Mr. Joseph Evans, of Morris, Illinois, who loaned it to Dr. Worthen for the Illinois Geological State Survey (107), in order that it might be described. The type has been destroyed by fire; so I am informed by Mr. L. E. Daniels, of Rolling Prairie, Indiana. There are two other known specimens of the species. One is in the collection of Mr. Daniels and the other No. 794, of Yale University Museum.

This genus may be clearly separated from all the other microsaurians by characters which are peculiar to the form. Among these may be mentioned the possession of sclerotic plates in the eyes; the large size of the orbits in comparison with the dimensions of the skull; the short, broad form of the body; the very short tail; the possession of a calcified cartilaginous pubis; clawed phalanges; presacrals 22. The character which places the genus distinctly in the Microsauria is the possession of long, slender, curved ribs, first detected on Mr. Daniels's specimen (plate 14, figs. 1, 2), by Dr. Hay (316). Its stegocephalian characters are evident in every particular of its anatomy the roofed skull, the arrangement of the cranial elements, the presence of a well-developed ventral armature, and the digital formula (4 for the hand and 5 for the foot).

The genus Amphibamus was regarded by Cope as a representative of a new order of vertebrates which he called (105) Xenorachia. He later (123) abandoned this, however. Fritsch (251), Zittel (642), and others regarded Amphibamus as a branchiosaurian. The exact position of the form was uncertain until 1900, when Dr. Hay (316) described the long, curved ribs and suggested its place among the Microsauria. He, however (Cat. Foss. Vert., p. 410), made the mistake of including the branchiosaurian family Protritonidæ, under Microsauria, thus confusing the subject further. The genus (462) has not the slightest relationship with the Branchiosauria.

Amphibamus grandiceps Cope.

Type: Specimen has been destroyed. There is an excellent specimen (plate 4, figs. 5. 6), No. 794 (1234), in Yale University Museum, and another nearly as good in the possession of Mr. L. E. Daniels, of Rolling Prairie, Indiana.

Horizon and locality: Mazon Creek shales, near Morris, Illinois.

The form of the skull of Amphibamus grandiceps Cope is not unlike that of Tuditanus minimus Moodie (462) from the Linton, Ohio, beds, but it is less acuminate than in that form. The large size of the orbits is especially striking. The shape of the skull is triangular, with concavities in the posterior table which correspond to the ear-slits so characteristic of Metoposaurus (242) from the Keuper of Germany. The narrowed posterior table of the skull is truncate, as in several other genera of Microsauria, notably Tuditanus and Saurerpeton. In structure the skull differs but little from many of the other Carboniferous forms, but the arrangement of the elements of the skull is more regular than in other genera.

The premaxillaries are very small elements in the anterior tip of the skull. They border the nares. The skull is rather peculiar among the Microsauria in the possession of a distinct lacrimal. I have detected this element in the cranium of Stegops divaricata Cope. As here defined the lacrimal is triangular, with its posterior border formed exclusively by the prefrontal. Its other relations are the normal ones. The nasal is elongate, with the usual relations of that element. The frontal is slightly longer and broader than the nasal. It apparently forms a portion of the inner border of the orbit. The parietal foramen lies in the anterior fourth of the parietal, a rather unusual position for this structure. The parietals, as in so many of the Microsauria, together form the largest element of the skull and are roughly a triangular area in the postero-median portion of the skull. The postparietal and the tabulare are clearly distinguishable and they have the usual relations for those elements. The maxillary, jugal, and quadratojugal together form the greater part of the maxillary border. The postero-lateral angle of the skull is, as visual, formed by the squamosal. The orbit is bounded posteriorly by the postorbital and the postfrontal, which include in the angle between them the quadrangular squamosal. The orbit is especially remarkable for its size as compared with the dimensions of the skull, being without a parallel among other known Microsauria. Around the border of the orbit in the specimen Cope studied (105) there were found 14 quadrangular plates which he called "superciliary plates." Hay (316) was inclined to regard them as sclerotic plates. In the Yale Museum specimen (plate 4, figs. 5, 6) there are 20 of these plates, and there seems to be no doubt that they are sclerotic elements. In the restoration (fig. 26) 29 sclerotic plates are given, but there is no assurance that this number is the exact one. They may also have been slightly larger, but not as large as in Branchiosaurus.

Fig. 26. Restoration of body outline and skeleton of Amphibamus grandiceps Cope, from Mazon Creek, Illinois, shales. Restoration is based on complete specimens of the species and on Cope's drawing. Form of body is indicated in one specimen, that in possession of Mr. Daniels. × 1.5.

Skull: pmx, premaxilla; n, nasal; fr, frontal; par, parietal; la, lacrimal; pf, prefrontal; pof, postfrontal; po, postorbital; pp, postparietal; spt, supratemporal; mx, maxilla; j, jugal; qj, quadratojugal; sq, squamosal; tab, tabulare.

Skeleton: ic, interclavicle; cl, clavicle; sc, scapula; h, humerus; r-u, radius, ulna; r, carpus; pu, pubis; il, ilium; f, femur; t, tibia; fb, fibula; ts, tarsus; x, ischium.

The vertebral column is preserved nearly entire in the Daniels specimen and quite entire (478) in the Yale specimen. Cope, in his study of the type (105, 107), thought there could be no more than 13 presacrals, but the specimen was poorly preserved and indecisive on this point. Dr. Hay (316) was inclined to the opinion that there were less than 20. The Yale specimen shows 22 centra, which are elongate, hour-glass-shaped bodies, with the neural spine a long, low crest running the entire length of the centrum, with a median elevation, so that in lateral view the spine would be triangular in form. The body of the centrum is expanded laterally into a diapophysis which extends anteriorly. The posterior vertebræ, at least, had the notochord largely persistent. The osseous part of the vertebra seems to have been but a thin shell, and the structure of the zygapophyses can not be determined. That they were dorsal in position is, however, evident from several vertebræ. The points of these structures project laterally. The tail is short and the caudal vertebræ weakly developed.

There are distinct impressions of at least 12 pairs of ribs in the Daniels specimen. They are long, slender, and curved, and there is no definite assurance that there were as many ribs as are indicated (fig. 26) in the restoration (462). The ribs are intercentral (469) and probably occupied the full length of the vertebral column. There may have been as many as indicated in the restoration.

Fig. 27. Skeleton of Amphibamus grandiceps Cope. × 1.4.

c, carpus; cl, clavicle; cr, caudal rib; cv, caudal vertebra; f, femur; h, humerus: il, ilium; s, skin: or, orbit: r, radius; ul, ulna; sc, scapula; sp, sclerotic plates; t, tibia and fibula; ts, tarsus; vs, ventral scutellæ. Specimen No. 794, Yale University Museum.

One of the most interesting features of the Yale specimen is the preservation of a small patch of skin, evidently from the back, lying to one side near the head, measuring 5 mm. in length by 3 mm. in width. The fragment shows the skin to be of tuberculated scales, 4 of which occupy the length of 1 mm. The scales are somewhat hexagonal, almost rounded, and were relatively quite thick. They lie in a close mosaic (fig. 27).

The Yale specimen has, very well preserved, a portion of the ventral scutellæ, of the throat, chest, and belly. The arrangement of the plates on the throat and chest is almost exactly the reverse of what Credner has described (190) for Branchiosaurus amblystomus Cred. On the throat, in the present form, the chevron points anteriorly, and it is the anterior prolongation of the belly scutes with the postero-lateral projection of the gular scutes which form the chest and arm scutellation. The belly chevrons point anteriorly, as in Branchiosaurus, the rods formed by the scutes being straight and not curved as in Branchiosaurus. The entire ventral armature preserved is displaced to the left of the animal and only the anterior portion is preserved.

The pectoral girdle is only partially known. The scapula is crescent-shaped. The other elements are indicated only by fragments and nothing is known of their form.

The arm elements are nearly all known. The humerus is slender and expanded at the ends, with its articular surfaces well developed. The separate radius and ulna are of approximately the same size and length. The carpus is unossified. The complete phalangeal formula for the hand of Amphibamus is unknown. The third digit seems to have 4 elements. The formula 2-2-3-2 has been suggested (462).

The pelvis is very satisfactorily known. The ilium is a long, slender, straight rod, with expanded ends. The ischium is shown on both sides of the vertebral column in the Yale specimen. Its form is almost identical with that of Paleohatteria longicaudata Credner, from the Rothliegenden of Saxony. The ischia are apparently approximate in the median line, though this character is somewhat obscured by the impression of the caudal vertebræ. Their relation with the ilium, other than that they were posterior to it, is uncertain. The pubis is, apparently, calcified cartilage. It is a squarish plate, somewhat corrugated, lying anterior to the ilium in the Daniels specimen. The elements of the pelvis were undoubtedly hung loosely in the flesh, as in modern salamanders, since there is no indication of articular surfaces.

The hind limb is well known, the type having a nearly complete leg with the foot. The Daniels and the Yale specimens supplement and substantiate the type. The femur is longer than the humerus, but more slender, with its articular surfaces about as well developed as in the humerus. The element is a simple rod of bone without muscular crests of any kind. The tibia and fibula are, likewise, slender separate rods of bone. The tarsus is unossified. The phalangeal formula is 2-2-3-4-3, and is fairly definite.

Fig. 28. Restoration of probable appearance of Amphibamus grandiceps Cope on the basis of the material described herewith. × 1.5.

In the type specimen the matrix in the orbit was blackened as if by the pigmentum nigrum of the choroid. The same has been noticed in other specimens. Professor Cope thought this indicated that the animal was nocturnal.

There are many characters in Amphibamus which seem to approximate the reptilian type of structure. Among these may be mentioned the character of the articular surfaces of the limb bones, the intercentral position of the ribs, the incipient double-headedness and the curvature of the ribs, the presence of a cartilaginous calcified pubis, the length of the limbs, and the clawed character of the phalanges.

Amphibamus was a low, flat, short, and undoubtedly a creeping, crawling animal, possibly spending a portion of its time in the water; but it could not have been a swimmer. It was one of nature's first attempts at constructing a land vertebrate.

Measurements of Amphibamus grandiceps Cope.

Collection of Mr. L. E. Daniels, of Rolling Prairie, Indiana: