Transcriber's Note

Many figure captions contain size references (i.e., × 6) these should be taken as an approximate guideline as the displayed images most likely will not match the size of the originally printed versions due to a number of factors (monitor resolution, chosen display size in browser, etc.).

[F]

MEMOIRS OF
THE CONNECTICUT ACADEMY
OF ARTS AND SCIENCES

VOLUME VII DECEMBER, 1920

The Appendages, Anatomy, and Relationships
of Trilobites

BY

PERCY E. RAYMOND, Ph.D.

ASSOCIATE PROFESSOR OF PALAEONTOLOGY, AND CURATOR OF INVERTEBRATE
PALAEONTOLOGY IN THE MUSEUM OF COMPARATIVE ZOOLOGY,
HARVARD UNIVERSITY

NEW HAVEN, CONNECTICUT
PUBLISHED BY THE
CONNECTICUT ACADEMY OF ARTS AND SCIENCES
AND TO BE OBTAINED ALSO FROM THE
YALE UNIVERSITY PRESS

THE TUTTLE, MOREHOUSE & TAYLOR COMPANY

TO THE MEMORY OF

CHARLES EMERSON BEECHER

SKILLFUL WITH HAND, BRAIN, AND PEN; REVEALER OF THE MYSTERIES OF TRILOBITES;

THIS MEMOIR IS DEDICATED

FOREWORD.

By CHARLES SCHUCHERT.

Trilobites are among the most interesting of invertebrate fossils and have long attracted the attention of amateur collectors and men of science. These "three-lobed minerals" have been mentioned or described in books at least since 1698 and now several thousand species are known to palæontologists. To this group of students they are the most characteristic animals of the seas of Palæozoic time, and even though they are usually preserved as dismembered parts, thousands upon thousands of "whole ones" are stored in the museums of the world. By "whole ones" perfect individuals are not meant, for before they became fossils the wear and tear of their time and the process of decomposition had taken away all the softer parts and even most of the harder exterior covering. What is usually preserved and revealed to us when the trilobites weather out of the embrace of their entombing rocks is the test, the hard shell of the upper or dorsal side. From time to time fragments of the under or limb-bearing side had been discovered, first by Elkanah Billings, but before 1876 there was no known place to which one could go to dig out of the ground trilobites retaining the parts of the ventral side.

Students of trilobites have always wanted specimens to be delivered to them weathered out of the rock by nature and revealing the ventral anatomy without further work than the collecting, but the wish has never been fulfilled. In the Utica black shales, near Rome, New York, there was finally discovered in 1892 a layer less than ten millimeters thick, bearing hundreds of Triarthrus becki with most of the ventral anatomy intact. The collector's first inkling that such were present in the Utica formation came to him in a chance find in 1884, and for eight years he sought off and on for the stratum whence this specimen came. His long search was finally rewarded by the discovery of the bed, and lo! here were to be had, in golden color, prostrate specimens with the breathing and crawling legs and the long and beautifully curved feeling organs all replaced by iron pyrites. Fool's gold in this case helped to make a palæontologic paradise. The bed contained not only such specimens of Triarthrus becki, but also, though more rarely, of Cryptolithus tessellatus and exceptionally of Acidaspis trentonensis. This important discovery, which has figured so largely in unraveling the evolution of the Crustacea and even has a bearing on that of most of the Arthropoda, was made by Mr. W. S. Valiant, then curator of the Museum of Rutgers College.

There were, however, great material difficulties to overcome before the specimens revealed themselves with all of their information exposed for study. No surgeon was needed, but a worker knowing the great scientific value of what was hidden, and with endless patience and marked skill in preparation of fossils. Much could be revealed with the hammer, because specimens were fairly abundant. A chance fracture at times showed considerable portions, often both antennæ entire, and more rarely the limbs protruding beyond the test, but the entire detail of any one limb or the variation between the limbs of the head, thorax, and tail was the problem to be solved. No man ever loved a knotty problem more than Charles E. Beecher. Any new puzzle tempted him, and this one of Triarthrus becki interested him most of all and kept him busy for years. From the summer of 1893. when he quarried out two tons of the pay stratum at Rome, until his death in 1904, his time was devoted in the main to its solution by preparing these trilobites and learning their anatomical significance.

The specimens of Triarthrus becki from Rome are pseudomorphs composed of iron pyrites, as has been said, and are buried in a gray-black carbonaceous shale. A little rubbing of the specimens soon makes of them bronze images of the former trilobite and while under preparation they are therefore easily seen. However, as the average individual is under an inch in length and as all the limbs other than the antennæ are double or biramous, one lying over the other, and the outer one fringed with a filamentous beard, the parts to be revealed by the preparator are so small and delicate that the final touch often obliterates them. These inherent difficulties in the material were finally overcome by endless trials on several thousand specimens, each one of which revealed something of the ventral anatomy. Finally some 500 specimens worthy of detailed preparation were left, and on about 50 of these Beecher's descriptions of Triarthrus and Cryptolithus were based.

The black shale in which the specimens are buried is softer than the pseudomorphous trilobites, a condition that is of the greatest value in preparation. With chisel and mallet the trilobites are sought in the slabs of shale and then with sharp chisels of the dental type they are revealed in the rough. At first Beecher sought to clean them further by chemical methods, and together with his friends, the chemist Horace L. Wells, and the petrologist Louis V. Pirsson, several solutions were tried, but in all cases the fossils were so much decomposed as to make them useless in study. Therefore Beecher had to depend wholly oh abrasives applied to the specimens with pieces of rubber. Much of this delicate work was done on a dental lathe, but in the final cleaning most of it was done with patient work by hand. Rubber has the great advantage of being tough and yet much softer than either specimen or shale. As the shale is softer than the iron pyrites, the abrasives (carborundum, emery, or pumice) took away the matrix more quickly than the trilobite itself. When a part was fully developed, the rubbers were cut to smaller and smaller dimensions and the abrading reduced to minute areas. So the work went on and on, helped along from time to time by the dental chisels. Finally Beecher became so expert with these fossils that after one side was developed he would embed the specimen in Canada balsam and fix it on a glass slide, thus enabling him to cut down from the opposite side. This was done especially with Cryptolithus because of the great scarcity of material preserving the limbs, and two of these revealed both sides of the individuals, though they were then hardly thicker than writing paper.

Then came illustrations, which at first were camera-lucida drawings in pencil smoothed out with pen and ink. "In some quarters," however, it has been said, "his methods unknown, their results were not accepted; they were regarded as startling, as iconoclastic, and even unreliable." He therefore decided to rework his material and to illustrate his publications with enlarged photographs. The specimens were black, there was little relief between fossil and matrix, and the ammonium chloride process of coating them white and photographing under artificial light was unsuitable. Nevertheless, after many trials, he finally succeeded in making fine enlarged photographs of the trilobites immersed in liquid Canada balsam, with a contact cover of glass through which the picture was taken, the camera standing vertically over the horizontal specimen. Beecher had completed this work in 1903 and in the winter of 1903-1904 was making the drawings, nearly all of which are here reproduced. On Sunday morning, February 14, 1904, as he was working at home on a large wash drawing of Cryptolithus, death came to him suddenly, leaving the trilobite problem but partially solved.

When the writer, in the autumn of 1904, succeeded Professor Beecher in the chair of Palæontology at Yale, he expected to find considerable manuscript relating to the ventral anatomy of the trilobites, but there was only one page. It was Beecher's method first to prepare and thoroughly study the material in hand, then to make the necessary illustrations, and between times to read what others had written. There was no written output until everything had been investigated and read, certain passages being marked for later reference. Then when all was assimilated, he would write the headings of topics as they came to him, later cutting them apart and arranging them in a logical sequence. When the writer visited him in his home in January 1904, he was primed for his final trilobite memoir, but the writing of it had not been begun.

The writer has never made the trilobites his special subjects for study as he has the brachiopods, and therefore felt that he should not try to bring to light merely the material things that Beecher had so well wrought out. It seemed at first an impossible task to find the specialist and friend to do Beecher justice, but as the years have passed, one of Beecher's students, always especially interested in trilobites, has grown into a full appreciation of their structures and significance, and to him has fallen the continuation of his master's work. If in the following pages he departs here and there from the accepted interpretation and the results of others, it is because his scientific training, in desiring to see with his own eyes the structures as they are, has led him to accept only those interpretations that are based on tangible evidence as he understands such. Furthermore, in seeking the relationship of the trilobites to the rest of the Arthropoda, his wide study of material and literature, checked up by the ontogeny of fossil and recent forms, has led him in places from the beaten path of supposedly ascertained phylogenies. His results, however, have been won through a detailed study of the interrelations of the Arthropoda, starting from the fact that the Trilobita are chronogenetically the oldest and most primitive. The trilobites are held by him to be the most simple, generalized, ancient Crustacea known, and the progenitors, directly and indirectly, of all Arthropoda.

It is now twenty-six years since Professor Beecher began his publications on the class Trilobita, and in commemoration of him and his work, Professor Percy E. Raymond of Harvard University presents this memoir, to bring to fruition the studies and teachings of his honored guide. It has been with Professor Raymond a labor of love, and it is for the writer of this foreword a long-desired memorial to the man to whose position in the Museum and University he had the privilege of succeeding.

Yale University, New Haven, Connecticut.

PREFACE.

The primary object of this memoir is, as has been stated by Professor Schuchert, to. rescue from oblivion the results of the last few years of Professor Beecher's investigations on the ventral anatomy of trilobites. Since he left his data in the form of drawings and photographs, without even rough notes, it became necessary, in order to write a text to accompany the plates, to restudy the entire subject. Under these circumstances, it seemed best to include all that is known about the appendages of trilobites, thus bringing together a summary of present information on the subject.

The growth of the memoir to its present size has been a gradual one. As first completed in 1917, it contained an account of the appendages only. Thoughts upon the probable use of the appendages led to the discussion of possible habits, and that in turn to a consideration of all that is known or could be inferred of the structure and anatomy of the trilobite. Then followed an inquiry into the relationships to other Arthropoda, which ultimately upset firmly established preconceptions of the isolated position of the group, and led to a modification of Bernard's view of its ancestry.

During the progress of the work, I have had the opportunity of examining most of the known specimens retaining appendages. From the Marsh collection in the Yale University Museum were selected the forty-six specimens showing best the appendages of Triarthrus, Cryptolithus, and Acidaspis. Dr. Charles D. Walcott very kindly returned to the Museum of Comparative Zoology the slices of Ceraurus, Calymene, and Isotelus which were the basis of his paper of 1881, and which had been loaned him for further study. He loaned also eight of the more important specimens of Neolenus serratus, and two of Triarthrus becki. At the United States National Museum I saw the specimens of Isotelus described by Mickleborough and the isolated limbs of Calymene from near Cincinnati. The Isotelus at Ottawa I had already studied with some care while an officer of the Geological Survey of Canada.

This memoir consists, as shown in the table of contents, of four parts. The appendages of Neolenus, Isotelus, Ptychoparia, Kootenia, Ceraurus, Calymene, and Acidaspis are discussed, as fully as circumstances warrant, in the first part, and new restorations of the ventral surfaces of Neolenus, Isotelus, Triarthrus, Ceraurus and Cryptolithus are included It is not supposed that these restorations will be of permanent value in all of their detail, but they are put forward as the best approximations to the real structure that the writer is able to present from the materials so far discovered. I am greatly indebted to Doctor Elvira Wood for the care and skill with which she has worked up these restorations from my rather sketchy suggestions. She has put into them not only a great amount of patient work, but also the results of considerable study of the specimens.

[Part II] is a discussion of the internal anatomy of the trilobite and a brief statement of some of the possible habits and methods of life of these animals. [Part III], which begins with a survey of the relationships of the trilobites to other Arthropoda, is largely taken up with an attempt to demonstrate the primitive characteristics of the former, and their probable ancestral position. The form of the ancestor of the trilobite is deduced from a study of the morphology, ontogeny, and phylogeny of the group, and evidence adduced to indicate that it was a depressed, flattened, free-swimming animal of few segments.

In [Part IV] are included somewhat detailed descriptions of a few of the best specimens of Triarthrus and Cryptolithus. Professor Beecher, while an observer of the minutest details, believed in publishing only the broader, more general results of his investigations. This method made his papers brief, readable, and striking, but it also resulted in leaving in some minds a certain amount of doubt about the correctness of the observations. In a matter so important as this, it has seemed that palæontologists are entitled to the fullest possible knowledge of the specimens on which the conclusions are based. The last part is, therefore, a record of the data for the restorations of Triarthrus and Cryptolithus.

The illustrations in the plates were nearly all made by or under the supervision of Professor Beecher, as were also text figures [45] and [46].

In conclusion, I wish to express my thanks to Mrs. Charles E. Beecher for the use of drawings which were the personal property of Professor Beecher; to Doctor Charles D. Walcott for photographs of the limbs of Calymene, and for his kindness in sending me the slices of trilobites from Trenton Falls and specimens of Neolenus and Triarthrus; to Doctor R. V. Chamberlin for suggestions and criticisms in regard to the relationship of trilobites to Insecta, Arachnida, Chilopoda, and Diplopoda; to Mr. Samuel Henshaw, Director of the Museum of Comparative Zoology, for permission to use the time which has been devoted to this work; and to Miss Clara M. Le Vene, for assistance in the preparation of the manuscript. My greatest debt is to Professor Charles Schuchert, to whom the work owed its inception, who has assisted in many ways during its prosecution, and who read the manuscript, and arranged for its publication. To him I can only express my warmest thanks for the favors which I have received and for the efforts which he has put forth to make this a worthy memorial to our friend and my teacher, Professor Charles Emerson Beecher.

Harvard University, Cambridge, Mass.
November, 1919.

TABLE OF CONTENTS.

LIST OF ILLUSTRATIONS.

[Frontispiece.] Charles Emerson Beecher, 1896.

Plates [1]-[5]. Photographs of Triarthrus becki, made by C. E. Beecher.

Plate [6]. Photographs of Triarthrus becki (figs. 1-3), Acidaspis trentonensis (fig. 6), and Cryptolithus tessellatus (fig. 7), made by C. E. Beecher. Photographs of the endopodites of a probable species of Calymene (figs. 4, 5)

Plates [7]-[8]. Photographs of Cryptolithus tessellatus, made by C. E. Beecher.

Plate [9]. Drawings of Cryptolithus tessellatus, made by C. E. Beecher or under his direction.

Plate [10]. Photographs of Isotelus latus and I. maximus, made by C. E. Beecher.

Plate [11]. Drawing of a restoration of Ceraurus pleurexanthemus, made by Elvira Wood.

HISTORICAL REVIEW.

The beginning of the search for the limbs of trilobites was coeval with the beginning of scientific study of the group, knowledge of the appendages being essential to the proper systematic allocation of the animals.

The early search was so barren of results that negative evidence came to be accepted as of positive value, and it was for many years generally believed that such organs as may have been present beneath the dorsal test were so soft as to be incapable of preservation. This view is best expressed by Burmeister (1846, p. 43):

There is good proof that the feet of trilobites must have been soft membranous organs, for the absence of the slightest remains of these organs in the numerous specimens observed is of itself evidence of the fact, and it can indeed scarcely be supposed that hard horny extremities should be affixed to a soft membranous abdominal surface; since they would not have possessed that firm basis, which all solid organs of locomotion require, in order that they may be properly available.

Very well reasoned, and were it not for the discovery of new material in American localities, Burmeister's views would probably never have been proved incorrect. One can not escape the suspicion that some of the accepted hypotheses of today, founded on similar "proof," may yield in time to the weight of bits of positive evidence.

The history of the study of appendages of trilobites may be divided into two periods. The first, in which there was a general belief that the appendages were soft organs, but during which numerous "finds" of limbs were reported, extended from the time of Linné to the year (1876) in which Walcott demonstrated the fact that the animals possessed jointed ambulatory and breathing organs.

The second, much more fruitful period, began with Walcott's publication of 1881, descriptive of the appendages of Ceraurus and Calymene, and for the purposes of this memoir, closes with his great contribution on the anatomy of Neolenus (1918). Beecher's brilliant productions came in the middle of the second period.

In the first period, there were at least two authentic discoveries of appendages, those of Eichwald (1825) and Billings (1870), but since neither of these men convinced his confreres of the value of his finds, the work of neither can be considered as having marked an especial epoch in the history.

As all the authentic finds will be treated in detail on later pages, only a brief résumé of the first period will be given here. This has already been done by Burmeister (1843, 1846) and Barrande (1852, 1872), whose works have been my primary sources of information, but I have looked up the original papers, copies of nearly all of which are to be seen in the libraries in Cambridge and Boston. Brig.-Gen. A. W. Vogdes, U. S. A. (retired), has very kindly placed at my disposal a number of references and notes.

Linné (1759) was the first to report the discovery of appendages of trilobites. Törnquist (1896) has pressed for a recognition of the contribution of the great Swedish naturalist to this problem, but Beecher (1896 B) doubted the validity of the find. Linné figured a specimen of Parabolina spinulosa (Wahlenberg), with what he interpreted as a pair of antennæ attached. He states (translation quoted from Törnquist): "Most remarkable in this specimen are the antennæ in the front, which I never saw in any other sample, and which clearly prove this fossil to belong to the insects." Beecher has shown as conclusively as can be shown without access to the original specimen that the supposed antennæ were really only portions of the thickened anterior border, the appearance being due to imperfect preservation. Brünnich as early as 1781 called attention to the imperfection of this specimen, and it is also referred to by Wahlenberg (1821, p. 39), Brongniart (1822, p. 42), Dalman (1828, p. 73), and Angelin (1854, p. 46).

Audouin (1821) seems to have been the first naturalist with sufficient knowledge of the Arthropoda to be competent to undertake the study of the trilobites. He concluded that the absence of ventral appendages was probably a necessary consequence of the skeletal conformation, and thought if any were discovered, they would prove to be of a branchial nature.

Wahlenberg (1821) in the same year expressed his belief that the trilobites were nearly allied to Limulus and in particular tried to show that the trilobites could have had masticatory appendages attached about the mouth as in that modern "insect" (p. 20). Wahlenberg was also the first to describe an hypostoma of a trilobite (p. 37, pl. 1, fig. 6), but did not understand the nature of his specimen, which he described as a distinct species.

Brongniart (1822, p. 40) devoted five pages of his monograph to a discussion of the affinities of trilobites, concluding that it was very probable that the animals lacked antennæ and feet, unless it might be that they had short soft feet which would allow them to creep about and fix themselves to other bodies.

Schlotheim (1823) thought that the spines on Agnostus pisiformis were segmented and compared them with the antennæ of Acarus.

Stokes (1823) was the first who, with understanding, published an illustration of the ventral side of a trilobite, having figured the hypostoma of an Isotelus. He was followed in the next year (1824) by Dekay, who also figured the hypostoma of an Isotelus, and added some observations on the structure of trilobites. The researches of Barrande, Novak, Broegger, Lindstroem, and others have dealt so fully with the hypostoma that further references to that organ need not be included here.

Dalman (1826, 1828) reviewed the opinions of his predecessors, and thought it not impossible that organs of mastication may have been present under the head shield of the trilobite as in Limulus (1828, p. 18). In this he of course followed Wahlenberg.

Goldfuss (1828) figured sections of Dalmanites hausmanni, Phacops macrophthalma, and Calymene tristani, which remind one of some of Doctor Walcott's translucent slices. So far as one can judge from the illustrations, it is probable that what he took for limbs were really fragments of other trilobites. Such is certainly the case in his figures 9 and 10, where a number of more or less broken thoracic segments are present. The section of Encrinurus punctatus shown in figure 7 may possibly exhibit the position and folds of the ventral membrane beneath the axial lobe, and also, perhaps, the appendages. His figures 4, 5 and 8 show the hypostoma in section.

Pander (1830) described the hypostoma in greater detail than had been done by previous authors, but otherwise added nothing to the subject.

Sternberg (1830) thought he had individuals showing appendages, but judging from his poor figures, he was deceived by fragmentary specimens.

Green (1839 A, B, C) described specimens of Phacops from Berkeley Springs, West Virginia, which had the hypostoma in position, and appear to have had a tubular opening under the axial lobe. While appendages were not actually present, these specimens suggested fairly correct ideas about the swimming and breathing organs of trilobites. They were similar to the ones which Castelnau obtained, and all were perhaps from the same locality.

It is not worth while to do more than enumerate the other authors of this period: Hisinger 1837, Emmrich 1839, Milne-Edwards 1841, for they all shared the same views, and added nothing to what was already known.

Castelnau (1843) described and figured a Phacops said to come from Cacapon Springs, West Virginia, which he thought possessed remains of appendages. There is nothing in the description or figures to indicate exactly what was present, but it is very unlikely that any limbs were preserved. The broad thin "appendage" figured may have been a fragment of a thoracic segment. This specimen was evidently described by Castelnau before 1843, as is inferred from a reference in the Neues Jahrbuch, 1843, P. 504, but I have not seen the earlier publication.

Burmeister (1843-1846), in his "Organization of the Trilobites," reviewed in extenso the history of the search for appendages, and concluded that they must have been so soft as to preclude the possibility of their being preserved as fossils. "Their very absence in fossils most distinctly proves their former real structure" (p. 10). In figures 7 and 8 on plate 6 he gave a restoration of the ventral surface of an Asaphus, the first restoration of the ventral anatomy to be attempted. Since he chose modern branchiopods as his model, he did not go so far wrong as he might have done. Still, there is little in the figure that would now be accepted as correct. The following quotation will serve to give the opinion of this zoologist, who from his knowledge of the Crustacea, was the most competent of the men of his time to undertake a restoration of the appendages of the trilobites:

… in giving a certain form to the feet in the restored figure, I have done so rather intending to indicate what they might have resembled, than with any idea of assuming their actual form. I merely assert that these organs were soft, membranous, and fringed, adapted for locomotion in water, placed on the abdominal portion of the body, and extending sidewise beneath the lateral lobes of the rings, as shown in the ideal transverse section. These feet were also indented, and thus divided into several lobes at the open lower side, and each separate lobe was furnished at the margin with small bristles serving as fins. The last and external lobe was probably longer, smaller, and more movable, and reached to the termination of the projecting shell lobe, bearing a bladder-shaped gill on the inner side (1846, p. 45).

McCoy (1846) observed in several trilobites a pair of pores situated in the dorsal furrows near the anterior end of the glabella. He showed that the pits occupy precisely the position of the antennæ of insects and suggested that they indicated the former presence of antennæ in these trilobites (chiefly Anipyx and "Trinucleus"). The evidence from Cryptolithus, set forth on a later page, indicates the correctness of McCoy's view.

Richter (1848, p. 20, pl. 2, fig. 32) described and figured what he took to be a phyllopod-like appendage found in a section through a Phacops. Without the specimen it is impossible to say just what the structure really was. The outline figure is so obviously modeled on an appendage of Apus that one is inclined to think it somewhat diagrammatic. In calling attention to this neglected "find," Clarke (1888, p. 254, fig.) interprets the appendage as similar to the spiral branchiæ of Calymene senaria, and adds that he himself has seen evidence of spiral branchiæ in the American Phacops rana.

Beyrich (1846) described a cast of the intestine of "Trinucleus," and Barrande (1852) further elaborated on this discovery.

Corda (1847) made a number of claims for appendages, but all were shown by Barrande (1852) to be erroneous.

Barrande (1852, 1872) gave a somewhat incomplete summary of the various attempts to describe the appendages of trilobites, concluding that none showed any evidence of other than soft appendages, until Billings' discovery of 1870.

Volborth (1863) described a long chambered tubular organ in Illænus which he believed to represent a cast of the heart of a trilobite, but which has since been likened by writers to the intestinal tract in "Trinucleus."

PART I.

THE APPENDAGES OF TRILOBITES.

Terminology.

The terminology employed in the succeeding pages is essentially the same as that used by Beecher, with two new terms added. Beecher assigned to the various segments of the limbs the names suggested by Huxley, but sometimes used the name protopodite instead of coxopodite for the proximal one. It is obvious that he did not use protopodite in the correct sense, as indicating a segment formed by the fusion of the coxopodite and basipodite. The usage employed here is shown in [figure 1].

Fig. 1.—Triarthrus becki Green. Diagram of one of the limbs of the thorax, viewed from above, with the endopodite in advance of the exopodite. 1, coxopodite, the inner extension being the endobase (gnathobase on cephalon); 2, basipodite, springing from the coxopodite, and supporting the exopodite, which also rests upon the coxopodite; 3, ischiopodite; 4, meropodite; 5, carpopodite; 6, propodite; 7, dactylopodite, with terminal spines.

The investigation of Ceraurus showed that the appendages were supported by processes extending downward from the dorsal test, and on comparison with other trilobites it appeared that the same was true in Calymene, Cryptolithus, Neolenus, and other genera. Thin sections showed that these processes were formed by invagination of the test beneath the dorsal and glabellar furrows. While these processes are entirely homologous with the entopophyses of Limulus, I have chosen to apply the name appendifer to them in the trilobites.

The only other new term employed is the substitution of endobase for gnathobase in speaking of the inner prolongation of a coxopodite of the trunk region. The term gnathobase implies a function which can not in all cases be proved.

The individual portions of which the limbs are made up are called segments, and the articulations between them, joints. Such a procedure is unusual, but promotes clearness.

The Appendages of Neolenus.

HISTORICAL.

The first mention of Neolenus with appendages preserved was in Doctor Walcott's paper of 1911, in which two figures were given to show the form of the exopodites in comparison with the branchiæ of the eurypterid-like Sidneyia. In 1912, two more figures were presented, showing the antennules, exopodites, and cerci. The specimens were found in the Burgess shale (Middle Cambrian) near Field, in British Columbia. This shale is exceedingly fine-grained, and has yielded a very large fauna of beautifully preserved fossils, either unknown or extraordinarily rare elsewhere. It was stated in this paper (1912 A) that trilobites, with the exception of Agnostus and Microdiscus, were not abundant in the shale.

In discussing the origin of the tracks known as Protichnites, Walcott presented four figures of Neolenus with appendages, and described the three claw-like spines at the tip of each endopodite.

Three new figures of the appendages were also contributed to the second edition of the Eastman-Zittel "Text-book of Paleontology" (1913, p. 701). Later (1916, pl. 9) there was published a photograph of a wonderful slab, bearing on its surface numerous Middle Cambrian Crustacea. Several of the specimens of Neolenus showed appendages.

Finally, in 1918, appeared the "Appendages of Trilobites," in which the limbs of Neolenus were fully described and figured (p. 126), and a restoration presented. Organs previously unknown in trilobites, epipodites and exites, attached to the coxopodites, were found.

Neolenus serratus (Rominger).

(Text figs. [2]-[8].)

Illustrated: Walcott, Smithson. Misc. Coll., vol. 57, 1911, p. 20, pl. 6, figs. 1, 2 (exopodites of thorax and cephalon);—Ibid., vol. 57, 1912, p. 191, pl. 24, figs., 1, la (antennules, caudal rami, and endopodites of thorax);—Ibid., vol. 57, 1912, p. 277, pl. 45, figs. 1-4 (antennules, endopodites of cephalon and thorax, caudal rami);—Text-book of Paleontology, edited by C. R. Eastman, 2d ed., vol. 1. 1913, p. 701, fig. 1343 (exopodites), p. 716, fig. 1376 (abdominal appendages), fig. 1377 (appendages of thorax and pygidium);—Ann. Rept. Smithson. Inst. for 1915, 1916, pl. 9;—Smithson. Misc. Coll., vol. 67, 1918, pp. 126-131 et al., pl. 14, fig. 1; pls. 15-20; pl. 21, fig. 6; pls. 22, 23; pl. 31 (restoration); pl. 34, fig. 3 (restored section); pl. 35, fig. 4; pl. 36, fig. 3 (hypostoma).

The following description of the appendages of Neolenus is summarized from Walcott's paper of 1918, and from a study of the eight specimens mentioned below.

Cephalon.

The antennules are long, slender, and flexible, and lack the formal double curvature so characteristic of those of Triarthrus. There are short fine spines on the distal rims of the segments of the proximal half of each, thus giving great sensitiveness to these organs. In the proximal portion of each, the individual segments are short and wider than long, and in the distal region they are narrow and longer than wide.

There are four pairs of biramous cephalic appendages, which differ only very slightly from the appendages of the thorax. All are of course excessively flattened, and they are here described as they appear.

The coxopodites, shown for the first time in Walcott's paper of 1918, are broad, longer than wide, and truncated on the inner ends, where they bear short, stout, unequal spines similar to those along the anterior margin. The gnathobases are but slightly modified to serve as mouth parts, much less so than in Triarthrus, but the coxopodites of the cephalon are shorter and wider than those of the thorax.

At the distal end of the coxopodite arise the endopodite and exopodite. The endopodite consists of six segments, the distal ones, propodite and dactylopodite, more slender than the others, the last bearing three terminal spines. The first endopodite is shorter than the others and slightly more slender (pl. 16, fig. 1)[1] and the anterior appendages turn forward more or less parallel to the sides of the hypostoma (pl. 22). The basipodite, ischiopodite, meropodite, and carpopodite are, in their flattened condition, roughly rectangular, only a little longer than wide, taper gradually distally, each bears small spines on the outer rim, and some of the proximal ones usually have a row along the margin.

[1] Nota bene! All references in this section are to the plates of Doctor Walcott's paper in 1918.

The exopodites of the cephalon, as of the body of Neolenus, are very different from those of any other trilobite whose appendages were previously known. As shown in the photographs (pl. 20, fig. 2; pl. 22), each exopodite consists of a single long, broad, leaf-like blade, not with many segments as in Triarthrus, but consisting of a large basal and small terminal lobe. It bears on its outer margin numerous relatively short, slender, flat setæ. The long axes of the exopodites point forward, and the setæ are directed forward and outward. They stand more nearly at right angles to the shaft on the cephalic exopodites than on those of the thorax. This same type of broad-bladed exopodite is also found on the thorax and pygidium.

The number of functional gnathobases on the cephalon is unknown. That four endopodites were present on one side is shown pretty clearly by specimen 58591 (pl. 16, fig. 3) and while no more than two well preserved exopodites have been seen on a side, there probably were four. Specimen 65513 (pl. 16, fig. 1) shows gnathobases on the second and third appendages of that individual as preserved, but there is no positive evidence that these are really the second and third appendages, for they are obviously displaced. The hypostoma of Neolenus is narrow but long, several specimens showing that it extended back to the horizon of the outer ends of the last pair of glabellar furrows. It is not as wide as the axial lobe, so that, while gnathobases attached beneath the first pair of furrows would probably not reach back to the posterior end of the hypostoma, they might lie parallel to it and not extend beneath. It seems possible, then, that there were four pairs of endobases but that the second rather than the first pair served as mandibles, as seems to be the case in Ceraurus.

Thorax.

The thorax of Neolenus consists of seven segments, and the appendages are well shown (pl. 17, fig. 1; pl. 18, figs. 1, 2; pl. 20, fig. 1.), The endopodites of successive segments vary but little, all are slender but compact, and consist of a long coxopodite with six short, rather broad segments beyond it. In the figures, the endopodites extend some distance in a horizontal direction beyond the edges of the dorsal test, as many as four segments being in some cases visible, but measurements show that the appendages tended to fall outward on decay of the animal. The dactylopodites are provided with terminal spines as in Triarthrus. The coxopodites are long, straight, and slender. They are well shown on only one specimen (pl. 18), where they are seen to be as wide as the basipodite, and the endobases are set with spines on the posterior and inner margins. They are so long that those on opposite sides must have almost met on the median line. The segments of the endopodites are mostly but little, if any, longer than broad, and at the distal end each shows two or more spines. The propodite and dactylopodite are notably more slender than the others. The exopodites of the thorax are broad and flat, and each shaft has two distinct parts with different kinds of setæ. The posterior edge of the proximal lobe is fringed with a slender, flat, overlapping hairs which are a little longer than the width of the lobe, and stand at an angle of about 60 degrees with the direction of the axis of the appendage. The outer lobe is at an angle with the main one, and has short, very fine setæ oh the margin. One or two specimens show some evidence of a joint between the inner and outer lobes, but in the great majority of cases they seem to be continuous; if originally in two segments, they have become firmly united. The exopodites of the thorax, like those of the cephalon, are directed diagonally forward and outward. (pl. 21, fig. 6; pl. 22.)

Pygidium.

The pygidium of Neolenus serratus is large, and usually shows five rings on the axial lobe and four pairs of ribs on the sides. There are five pairs of biramous appendages belonging to this shield, and behind these a pair of jointed cerci. That the number of abdominal appendages should correspond to the number of divisions of the axial lobe rather than to the number of ribs on the pleural lobes is of interest, and in accord with other trilobites, as first shown by Beecher.

The endopodites of the pygidium have the same form as those of the thorax, are long, and very much less modified than those of any other trilobite whose appendages are known. On some specimens, they extend out far beyond the dorsal test, so that nearly all the segments are visible (pl. 17, fig. 3; pl. 18; pl. 19; pl. 20, fig. 1), but in these cases are probably displaced. The segments are short and wide, the whole endopodite tapering gradually outward. The dactylopodite bears terminal spines, and the individual segments also have outward-directed spines.

The cerci appear to have been long, slender, very spinose organs much like the antennules, but stiff rather than flexible. They are a little longer than the pygidium (pl. 17, figs. 1, 2), and seem to be attached to a plate on the under surface of the posterior end and in front of the very narrow doublure. The precise form of this attachment can not be determined from the published figures. They bear numerous fine spines (pl. 17, fig. 3).

Epipodites and Exiles.

Doctor Walcott has found on several specimens of Neolenus remains of organs which he interprets as epipodites and exites attached to the coxopodites. A study of the specimens has, however, convinced me that both the large and small epipodites are really exopodites, and that the exites are badly preserved and displaced coxopodites. Detailed explanation of this interpretation is given below in the description of the several specimens involved.

Description of Individual Specimens.

Doctor Walcott was kind enough to send me eight of the more important specimens of Neolenus figured by him, and since my interpretation of them does not agree in all respects with his, I have thought it fairer to the reader to present here rather full notes explaining the position I have taken. I understand that since I communicated my interpretation of the epipodites and exites to him, Doctor Walcott has submitted the specimens to several palæontologists, who consider that epipodites are really present. Since I am not able to convince myself that their conclusion is based upon sound evidence, I give here my own interpretation. There is of course, no a priori reason why trilobites should not have had epipodites.

Specimen No. 58589.

Illustrated: Walcott, Smithson. Misc. Coll., vol. 57, 1912, pl. 45, fig. 2;—Zittel-Eastman Text-book of Paleontology, vol. 1, 1913, fig. 1377;—Smithson. Misc. Coll., vol. 67, 1918, pl. 18, fig. 1; pl. 20, fig. 1.

This is one of the most important of the specimens, as it shows the coxopodites of three thoracic limbs and the well preserved endopodites of six thoracic and five pairs of pygidial appendages.

The appendages are all shifted to the left till the articular socket of the coxopodite is. about 8 mm. outside of its proper position. The endopodites extend a corresponding amount beyond the edge of the dorsal test and are there so flattened that they are revealed as a mere impression. The coxopodites, which are beneath the test, seem to have been somewhat protected by it, and while hopelessly crushed, are not flattened, but rather conformed to the ridges and grooves of the thorax.

Fig. 2. Neolenus serratus (Rominger). A sketch of the coxopodites and endopodites of two thoracic segments. Note notch for the reception of the lower end of the appendifer. × 3.

The coxopodite of the appendage of the last thoracic segment is best preserved. It is rectangular, about one third as wide as long, with a slight notch in the posterior margin near the outer end. The inner end is obliquely truncated and shows about ten sharp spines which do not appear to be articulated to the segment, but rather to be direct outgrowths from it. There are similar spines along the posterior margin, but only two or three of what was probably once a continuous series are now preserved. On the opposite margin of the coxopodite from the slight depression mentioned above, there is a slight convexity in the outline, which is better shown and explained by the coxopodite just in front of this. That basal segment has the same form as the one just described, but as its posterior margin is for the greater part of its length pushed under the one behind it, the spines are not shown. On the posterior margin, two-thirds of the length from the proximal end, there is a shallow notch, and corresponding to it, a bulge on the anterior side. From analogy with Ceraurus and Calymene it becomes plain that the notch and bulge represent the position of the socket where the coxopodite articulated with the appendifer. Since these structures have not been shown in previous illustrations, a drawing giving my interpretation of them is here inserted ([fig. 2]). It is evident from the position of the notch that the row of spines was on the dorsal (inner) side of the coxopodite and that the truncation was obliquely downward and outward.

The endopodite of the last thoracic appendage is well preserved and may be described as typical of such a leg in this part. The basipodite is as wide as the coxopodite, and it and the three succeeding segments, ischiopodite, meropodite, and carpopodite, are all parallel-sided, not expanded at the joints, and decrease regularly in width. The propodite and dactylopodite are also parallel-sided, but more slender than the inner segments, and on the end of the dactylopodite there are four little spines, three of them—one large and two small—articulated at the distal end, and the fourth projecting from the posterior outer angle. Each segment has one or more spines on the outer articular end, and the ischiopodite has several directed obliquely outward on the posterior margin. All of the four proximal segments show a low ridge parallel to and near the anterior margin, and several endopodites of the pygidium have a similar ridge and a row of spines along the posterior margin of some of the segments. These features indicate that the segments in question were not cylindrical in life, but compressed. From the almost universal location of the spines on the posterior side of the limbs as preserved, it seems probable that in the natural position the segments were held in a plane at a high angle with the horizontal, the ridge was dorsal and anterior and the row of spines ventral and posterior. Because the spines on the endobases are dorsal it does not follow that those on the endopodites were, for the position of the coxopodite in a crushed specimen does not indicate the position of the endopodite of even the same appendage.

The endopodites of the pygidium are similar to the one just described, except that some of them have spines on the posterior margin of the segments, and a few on the right side have extremely fine, faintly visible spines on the anterior side. The specimen shows fragments of a few exopodites, but nothing worth describing. In the middle of the right pleural lobe there is a small organ which Walcott has interpreted as a small epipodite. It is oval in form, broken at the end toward the axial lobe, and has exceedingly minute short setæ on the posterior margin. From analogy with other specimens, it appears to me to be the outer end of an exopodite.

Measurements: The entire specimen is about 64 mm. long and 52 mm. wide at the genal angles. The thorax is about 41 mm. wide (disregarding the spines) at the seventh segment, and the axial lobe about 13 mm. wide at the same horizon. The measurements of the individual segments of the seventh left thoracic limb are:

The five distal segments of the last pygidial endopodite are together 10.5 mm. long. The whole six segments of the endopodite of the third thoracic segments are together 21 mm. long. The distance from the appendifer of the third segment to the outer end of the spine is 17 mm. From the center of the notch in the coxopodite to the outer end is 1.5 mm., which, added to the length of the endopodite, 21 mm., makes a distance of 22.5 mm. from the appendifer to the tip of the dactylopodite, showing that if projected straight outward, the endopodites of the thorax would project 5.5 mm. beyond the test, including spines.

The distance across the axial lobe from appendifer to appendifer on the seventh thoracic segment is 12.5 mm. Measured along the top of the coxopodite, it is 6 mm. from the middle of the notch to the inner end, and measured along the bottom it is 8 mm. From the truncated form of the ends it is evident that the coxopodites extended inward and downward from the appendifers, and with the dimensions given above, the inner toothed ends would practically meet on the median line.

Measurements on the appendages of the pygidia show that on this specimen they extend back about twice as far beyond the edge of the pygidium as they should, all being displaced.

Specimen No. 65514.

Illustrated: Walcott, Smithson. Misc. Coll., vol. 67, 1918, pl. 19, figs. 1-3.

This specimen is so twisted apart that it is not possible to determine to what segments the appendages belong, but it exhibits the best preserved exopodites I have seen. The best one is just in front of the pygidium on the matrix, and shows a form more easily seen than described (our fig. 3). There is a broad, flat, leaf-like shaft, the anterior side of which follows a smooth curve, while in the curve on the posterior side, which is convex backward, there is a re-entrant, setting off a small outer lobe whose length is about one third the length of the whole. This lobe seems to be a continuation of the shaft, and the test of the whole is wrinkled and evidently very thin. The main and distal lobes of the shaft both bear numerous delicate setæ, but those of the outer lobe are much shorter and finer than those on the main portion. The latter are flattened and blade-like.

The anterior edge of the shaft shows a narrow stiffening ridge and the setæ are but little longer than its greatest width. The second segment of the pygidium has another exopodite like this one, but shows faintly the line between the two lobes, as though there were two segments.

This specimen also shows some very well preserved endopodites, but they differ in no way from those described from specimen No. 58589. Walcott mentions two large epipodites projecting from beneath the exopodites. I judge that he has reference to the distal lobes of the exopodites, but as these are continuous with the main shaft, there can be no other interpretation of them than that which I have given above.

Measurements: The pygidium is 19 mm. long (without the spines) and about 34 mm. wide at the front. The exopodites show faintly beneath the pygidial shield, but their proximal ends are too indistinct to allow accurate measurement. Apparently they were just about long enough to reach to the margin of the shield. The best preserved one, that of the second segment in the pygidium, is about 11 mm. long, 2.5 mm. wide at the widest; the distal lobe is 2.5 mm. long, and the longest setæ of the main lobe 3.5 mm. long. The pleural lobe of the pygidium is just 11 mm. wide at this point.

The endopodites project from 8 to 12 mm. beyond the pygidium, showing about four segments.

The thoracic exopodite described above is 11 mm. long and 2.75 mm. wide at the widest part. The distal lobe is 3.5 mm. long and 2.25 mm. wide, and the longest setæ on the main lobe 3 mm. long.

Specimen No. 65519.

Illustrated: Walcott, Zittel-Eastman Text-book of Paleontology, vol. 1, 1913, fig. 1343;—Smithson. Misc. Coll., vol. 67, 1918, pl. 21, fig. 6.

This specimen is somewhat difficult to study but is very valuable as showing the natural position of the exopodites of the anterior part of the thorax. Walcott's figures are excellent and show the broad leaf-like shafts, the distal lobes with the re-entrant angles in the posterior margin, and the long fine setæ of the main lobes. None of the distal lobes retains its setæ. All extend back to the dorsal furrows, but the proximal ends are not actually shown.

The specimen is especially important because it shows the same distal lobes as specimen No. 65514, and demonstrates that they are a part of the exopodite and not of any other structure.

Measurements: The exopodite belonging to the fourth thoracic segment is 23 mm. long and 4 mm. wide at the widest part. The longest setæ are 7 mm. in length.

Specimen No. 65520.

Illustrated: Walcott, Smithson. Misc. Coll., vol. 67, 1918, pl. 20, fig. 2; pl. 22, fig. 1.

This is a practically entire specimen, on two blocks, one showing the interior of the shell, and the other the one figured by Walcott, a cast of the interior. The first shows the low rounded appendifers at the anterior angle of each axial tergite. They are almost entirely beneath the dorsal furrows and do not project so far into the axial lobe as those of Ceraurus and Calymene. In fact, only those at the anterior end of the thorax project inward at all. As expected, there are five pairs on the pygidium. The cephalon is unfortunately so exfoliated that the appendifers there are not preserved. The doublure of the pygidium is extremely narrow.

The cast of the interior shows, rather faintly, the exopodites of the right side of the thorax and of the left side of the cephalon, and, still more faintly, the caudal rami and a few pygidial endopodites. The exopodites on the right side are in what seems to be the customary position, directed obliquely forward and outward, and the tips of their distal lobes project slightly beyond the edge of the test. These lobes were interpreted by Walcott as epipodites, but after comparing them with the terminal lobes of the exopodites of specimens No. 65519 and 65514 I think there can be no doubt that they represent the same structure. The pleura of the individual thoracic segments on this side of the specimen have an unusual appearance, for they are bluntly rounded or obtusely pointed, instead of being spinose.

The interpretation of the appendages of the cephalon is somewhat difficult. At the left of the glabella there are two large exopodites, the anterior of which lies over and partially conceals the other. These show by their position that they belong to the fourth and fifth cephalic appendages. In front of these lie two appendages which may be either endopodites or exopodites, but which I am inclined to refer to the latter. Both are narrow and shaped like endopodites, but bear on their outer edges close-set fine setæ. They also show what might be considered as faint traces of segmentation. If the first of these ran under the end of the exopodite behind it, as shown in Walcott's figure (pl. 22), then it would be necessary to interpret it as an endopodite, but it really continues down between the exopodite and the glabella, and seems to be attached opposite the middle of the eye. The specimen does not indicate clearly whether this appendage is above or below the exopodite behind it, but one's impression is that it is above, in which case it also must be an exopodite. The appendage in front, being similar, is similarly interpreted. If this be correct, then the exopodites of the second and third cephalic appendages are much shorter and narrower than those of the fourth and fifth. All of these appendages are obviously out of position, for the cheek has been pushed forward away from the thorax, though still pivoting on its inner angle at the neck-ring, till the eye has been brought up to the dorsal furrow. In this way the anterior exopodites have been thrust under the glabella and all the appendages have been moved to the right of their original position. The anterior exopodite is very poorly shown, but seems to be articulated in front of the eye. The posterior exopodites are very similar to those on the thorax. The distal lobe is shown only by the second from the last. It has the same form as the distal lobes on the thoracic exopodites, and like them has much finer setæ than the main lobe, but it does not stand at so great an angle with the axis of the main lobe, nor yet is it so straight as shown in Walcott's figure.

Measurements: The specimen is about 72 mm. long and 54 mm. wide at the genal angles. The pygidium is 22 mm. long and 37 mm. wide. The doublure is 1.5 mm. wide. The exopodite of the third thoracic segment is 19.5 mm. long. The pleural lobe at this point is 13 mm. wide without the spines and 18.5 mm. wide with them. The third exopodite of the cephalon was apparently about 15 mm. long when complete.

Specimen No. 65515.

Illustrated: Walcott, Smithson. Misc. Coll., vol. 67, 1918, pl. 20, figs. 3, 4.

This is a small piece of the axial portion of a badly crushed Neolenus, showing appendages on the left side as viewed from above. On the posterior half there are three large appendages which have the exact form of the exopodites of other specimens. There is a broad, oval, proximal lobe and a distal one at an angle with it. The proximal part of the shaft has fine setæ or the bases of them, and the distal lobe faint traces of much finer ones. The form, and the setæ so far as they are preserved, are exactly like those of the exopodites on the specimens previously described. (See [fig. 4, page 26].) Beneath them there are slender, poorly preserved endopodites.

In front of the exopodites and endopodites lie a series of structures which Walcott has called exites, but for which I can see another explanation. Walcott has shown them as four broad rounded lobes, but his figure must be looked upon as a drawing and not as a photograph, for it has been very much retouched.

For convenience of discussion, these lobes may be called Nos. 1, 2, 3, and 4, the last being the posterior one ([fig. 5]). This lobe is best shown on the matrix, where the anterior end is seen to be margined by stout spines, while the posterior end lies over the endopodite and under the exopodite behind it. No. 3 is sunk below the level of the others, and only a part of it has been uncovered. Its margin bears strong spines of different sizes. Its full shape can not be made out, but it has neither the shape nor the form of spines shown in figure 3, plate 20 (1918). Lobes 2 and 1 and another lobe in front of 1 seem to form a continuous series and to be part of a single appendage. They are all in one plane, arc so continuous that the joints between them can be made out with difficulty and if they do belong together, can easily be explained.

Before calling these structures new organs not previously seen on trilobites, it is of course necessary to inquire if they can be interpreted as representing any known structures. That they can not be exopodites is obvious, since they are bordered by short stout spines instead of setæ. The same stout spines that negate the above possible explanation at once suggest that they are coxopodites (compare [fig 6]). At first sight, the so-called exites seem too wide and too rounded to be so interpreted, but if reference be had to the specimens rather than the figures, it will be noted that the only well preserved structure (No. 2) is longer than wide, has spines only on one side and one end, and does not differ greatly from the coxopodite of specimen No. 58589 (pl. 18, 1918). If structures 2, 1, and the segment ahead of 1 are really parts of one appendage, it can only be an endopodite, of which No. 2 is the coxopodite, No. 1 the basipodite, and the next segment the ischiopodite. If one looks carefully, there are no traces of spines on either end of No. 1, but only on the margin. The extreme width of No. 2 is against this interpretation as a coxopodite (see, however, [fig. 6]), but it may be rolled out very flat, as this is an unusually crushed specimen. No. 2 is 10 mm. long and 6 mm. wide at the widest point. No. 1 is 5 mm. long and 3.5 mm. wide.

The crucial point in this determination is whether 2 and 1 are parts of the same appendage. I believe they are, but others may differ.

Specimen No. 65513.

Illustrated: Walcott, Smithson. Misc. Coll., vol. 57, 1912, pl. 45, fig. 3;—Ibid., vol. 67, 1918, pl. 16, figs. 1, 2.

This is nearly all of the right half of an entire specimen, but the only appendages of any interest are those of the cephalon. Five endopodites emerge from beneath that shield, but as all are displaced it is not possible to say how many belong to the head. When held at the proper angle to the light, the second and third from the front show faintly the partial outlines of the coxopodites. The anterior side and end of the best preserved one shows irregular stout spines of unequal sizes, and the inner end is truncated obliquely ([fig. 6]). These coxopodites are like those on the thorax of specimen No. 58589, but shorter and wider. This of course suggests that the "exite" No. 2 of specimen No. 65515 may be a cephalic coxopodite. The endopodite of this appendage, like the others on this cephalon, is shorter and stouter than the thoracic or pygidial endopodites of the others described.

Fig. 7.—A restored section across the thorax of Neolenus serratus, showing the probable form of attachment of the appendages, their relation to the ventral membrane, and the jaw-like endobases of the coxopodites.

Measurements: The cephalon is 24 mm. long and about 60 mm. wide. The coxopodite of the third appendage is about 10 mm. long and 5.5 mm. wide at the widest point. The corresponding endopodite is 19 mm. long and projects 11 mm. beyond the margin, which is about 5 mm. further than it would project were the appendage restored to its proper position.

RESTORATION OF NEOLENUS.

(Text figs. [7], [8].)

This restoration is based upon the information obtained from the studies which have been detailed in the preceding pages, and differs materially from that presented by Doctor Walcott. The appendages are not shown in their natural positions, but as if flattened nearly into a horizontal plane. The metastoma is added without any evidence for its former presence.

The striking features of the appendages are the broad unsegmented exopodites which point forward all along the body, and the strong endopodites, which show practically no regional modification. Although the exopodites have a form which is especially adapted for use in swimming, their position is such as to indicate that they were not so used. The stout endopodites, on the other hand, probably performed the double function of natatory and ambulatory legs.

Fig. 8.—Neolenus serratus (Rominger). A restoration of the ventral surface, with the endopodites omitted from one side, to permit a better exposition of the exopodites. The position and number of the appendages about the mouth are in considerable doubt. Restored by Doctor Elvira Wood under the supervision of the writer. About one-half larger than the average specimen.

Nathorstia transitans Walcott.

Illustrated: Walcott, Smithson. Misc. Coll., vol. 57, 1912, pl. 28, fig. 2.

The badly preserved specimen on which this genus and species was based is undoubtedly a trilobite, but for some reason it does not find a place in Walcott's recent article on "Appendages" (1918). The preservation is different from that of the associated trilobites, being merely a shadowy impression, indicating a very soft test. The general outline of the body, the position of the eye, and even a trace of spines about the pygidium (in the figure) are similar to those of Neolenus, and I would venture the suggestion that Nathorstia transitans is a recently moulted Neolenus serratus, still in the "soft-shelled" condition. Even if not a Neolenus, it is probable, from the state of preservation, that it is an animal which had recently cast its shell.

Walcott describes such fragments of appendages as remain, as follows:

Head. A portion of what may be an antenna projects from beneath the right anterior margin; from near the left posterolateral angle a large four-jointed appendage extends backward. I assume that this may be the outer portion of the large posterior appendage (maxilla) of the head.

Thorax. Traces of several slender-jointed thoracic legs project from beneath the anterior segments and back of these on the right side more or less of six legs have been pushed out from beneath the dorsal shield; these are composed of three or four long slender joints; fragments of the three proximal joints indicate that they are shorter and larger and that they have a fringe of fine setæ. Indications of a branchial lobe (gill) are seen in two specimens where the legs are not preserved. This is often the case both among the Merostomata (pl. 29, fig. 3, Molaria) and Trilobita (pl. 24, fig. 2, Ptychoparia).

Two caudal rami project a little distance beneath the posterior margin of the dorsal shield.

This latter feature of course suggests Neolenus. The other appendages are too poorly preserved to allow comparison without seeing the specimen.

The specific name was given "on account of its suggesting a transition between a Merostome-like form, such as Molaria spinifera, and the trilobites." In what respect it is transitional does not appear.

Formation and locality: Same as that of Neolenus serratus. One nearly complete specimen and a few fragments were found.

The Appendages of Isotelus.

HISTORICAL.

The first specimen of Isotelus with appendages was described orally by Billings before the Natural History Society of Montreal in 1864, and in print six years later (1870, p. 479, pls. 31, 32). The specimen is described in detail on a later page. Billings recognized the remains of eight pairs of legs on the thorax, a pair for each segment, and he inferred from the fact that the appendages projected forward that they were ambulatory rather than natatory organs. He was unable to make out the exact number of the segments in the appendages, but thought each showed at least four or five.

Having examined the individual sent to London by Billings, Woodward (1870, p. 486, fig, 1) reviewed the collection from the American Trenton in the British Museum and found a specimen in the "Black Trenton limestone," from Ottawa, Ontario, in which, alongside the hypostoma, was a jointed appendage, which he described as the "jointed palpus of one of the maxillæ." This has always been considered an authentic "find," but I am informed by Doctor Bather that the specimen does not show any real appendage. For further discussion, see under Isotelus gigas.

In 1871, Billings' specimen was examined by Professors James D. Dana (1871, p. 320), A. E. Verrill, and Sydney I. Smith, who agreed that the structures identified by Billings as legs were merely semicalcified arches of the membrane of the ventral surface, which opinion seems to have been adopted by zoologists generally in spite of the fact that the most elementary consideration of the structure of the thorax of a trilobite should have shown its falsity. While the curvature of the thoracic segments was convex forward, that of the supposed ventral arches was convex backward, and the supposed arches extended across so many segments as to have absolutely prevented any great amount of motion of the segments of the thorax on each other. Enrollment, a common occurrence in Isotelus, would have been absolutely impossible had any such calcified arches been present.

Walcott, in his study of trilobites in thin section (1881, pp. 192, 206, pl. 2, fig. 9), obtained eleven slices of Isotelus gigas which showed remains of appendages. He figured one of the sections, stating that it "shows the basal joint of a leg and another specimen not illustrated gives evidence that the legs extended out beneath the pygidium, as indicated by their basal joints."

The second important specimen of an Isotelus with appendages was found by Mr. James Pugh in strata of Richmond age 2 miles north of Oxford, Ohio, and is now in the U. S. National Museum. It was first described by Mickleborough (1883, p. 200, fig. 1-3). In two successive finds, a year apart, the specimen itself and its impression were recovered. Since I am redescribing the specimen in this memoir (see p. 35), it only remains to state here that Mickleborough interpreted the structures essentially correctly, though not using the same terminology as that at present adopted. His view that the anterior appendages were chelate can not, however, be supported, nor can his idea that the sole appendages of the pygidium were foliaceous branchial organs.

Walcott (1884, p. 279, fig. 1) studied the original specimens and presented a figure which is much more detailed and clear than those of Mickleborough. By further cleaning the specimen he made out altogether twenty-six pairs of appendages. He stated that one of these belonged to the cephalon, nine to the thorax,[1] and the remaining sixteen to the pygidium. He showed that the endopodites of the pygidium were of practically the same form as those on the thorax, and stated that the "leg beneath the thorax of the Ohio trilobite shows seven joints in two instances; the character of the terminal joint is unknown." His figure shows, and he mentions, markings which are interpreted as traces of the fringes of the exopodites.

[1] The posterior one of these he believed to have been crowded forward from beneath the pygidium.

In the same year Woodward (1884, p. 162, fig. 1-3) reproduced all of Mickleborough's figures, and suggested that the last seven pairs of appendages on the pygidium of Calymene and Isotelus were probably "lamelliform branchiferous appendages, as in Limulus and in living Isopoda."

Professor Beecher published, in 1902, an outline taken from Mickleborough's figure of this specimen, to call attention to certain discontinuous ridges along the axial cavity of the anterior part of the pygidium and posterior end of the thorax. These ridges are well shown in Mickleborough's figure, though not in that of Walcott, and their presence on the specimen was confirmed by a study by Schuchert, who contributed a diagrammatic cross-section to Beecher's paper (1902, p. 169, pl. 5, figs. 5, 6). Beecher summarized in a paragraph his interpretation of this specimen:

The club-shaped bodies lying within the axis are the gnathobases attached at the sides of the axis; the curved members extending outward from the gnathobases are the endopodites; the longitudinal ridges in the ventral membrane between the inner ends of the gnathobases are the buttresses and apodemes of the mesosternites; the slender oblique rod-like bodies shown in the right pleural region in Walcott's figure are portions of the fringes of the exopodites.

In 1910, Mr. W. C. King of Ottawa, Ontario, found at Britannia, a few miles west of Ottawa, the impression in sandstone of the under surface of a large specimen of Isotelus arenicola, described on a later page (p. 39).

Finally (1918, p. 133, pl. 24, figs. 3, 3a; pl. 25), Walcott has redescribed the specimen from Ohio, presenting a new and partially restored figure. He refers also to the specimen from Ottawa under the name Isotelus covingtonensis? Foerste (not Ulrich). He advances the view, which I am unable to share, that the cylindrical appearance of the segments of the appendages of Isotelus is due to post-mortem changes.

Isotelus latus Raymond.

([pl. 10, fig. 1.])

Illustrated: Asaphus platycephalus Billings, Quart. Jour. Geol. Soc., London, vol. 26, 1870, pl. 31, figs. 1-3; pl. 32, figs. 1, 2.—Woodward, Geol. Mag., vol. 8, 1871, pl. 8, figs. 1, 1a.—Gerstäcker, in Bronn's "Klassen u. Ordnungen d. Thier-Reichs," 1879, pl. 49, fig. 1.—von Koenen, N. Jahrb. f. Min., etc., vol. 1, 1880, pl. 8, fig. 8.—Milne-Edwards, Ann. Sci. Nat., Zoologie, ser. 6, vol. 12, 1881, pl. 12, fig. 45.

Isotelus latus Raymond, Bull. Victoria Mem. Mus., Geol. Survey Canada, No. 1, 1913, p. 45 (species named).

Isotelus covingtonensis? Walcott (not Foerste), Smithson. Misc. Coll., vol. 67, 1918, p. 134.

Knowledge of the appendages of this species is derived from the specimen which Billings described in 1870. It was found in the Trenton, probably the Middle Trenton, near Ottawa, Ontario, and is preserved in the Victoria Memorial Museum at Ottawa.

Viewed from the upper surface, it shows a large part of the test, but is broken along the sides, so that parts of the free cheeks, considerable of the pleural lobes of the thorax, and one side of the pygidium are missing. Viewed from the lower surface, the appendages are practically confined to the cephalon and thorax.

A short time before his death, Professor Beecher had this specimen and succeeded in cleaning away a part of the matrix so that the appendages show somewhat more clearly than in Billings' time, but they are not so well preserved as on the Mickleborough specimen, found in Ohio somewhat later.

The hypostoma is in place and well preserved; the posterior points are but 3 mm. in advance of the posterior margin of the cephalon. Behind the hypostoma there are only two pairs of cephalic appendages, the first of which is represented by the coxopodite and a trace of the endopodite. The outer end of the coxopodite is close to the outer margin of one of the prongs of the hypostoma and about 3 mm. in front of its posterior end. The gnathobase curves backward and inward, and appears to pass under the tip of the hypostoma. There were probably two appendages in front of this, whose gnathobases projected under the hypostoma, but the specimen shows nothing of them unless it be that one small fragment about 2 mm. back of the center is really a part of a gnathobase.

The specimen retains only the coxopodite and basipodite of the posterior cephalic appendage on the left side. The coxopodite is long and apparently cylindrical, the cross-section being of uniform diameter throughout the length. The inner portion is nearly straight, while the outer part is curved gently forward.

It is possible to make out remains of eight pairs of appendages on the thorax, some of them represented by coxopodites only, but most with more or less poorly preserved endopodites as well. No exopodites are visible. The coxopodites of the thorax seem to be of the same form as the last one on the cephalon, but slightly less curved. All are long and heavy, and there seems to be no decrease in size toward the pygidium. The endopodites are very imperfectly shown. They seem to be longer than those of Isotelus maximus, and the segments, while of less diameter than the coxopodites, do not show so great a contrast to them as do those of that species. The direction of the endopodites is diagonally forward, and the outer portions do not appear to be curved backward as in Isotelus maximus. It would appear also that the endopodites were nearly or quite long enough to reach the outer margin of the dorsal test. On no endopodite can more than three segments be definitely distinguished, but the longest ones are the most obscurely segmented.

No appendages are preserved on the pygidium, but at one side of the median groove there are two projections which may be processes to which the appendages were attached.

Measurements: Total length of specimen, 109 mm. Probable length when complete, 116 mm. Length of cephalon, 40 mm.; width at genal angles, restored, about 62 mm. (Billings' restoration). Width of doublure of front of cephalon on median line, 17 mm.; length of hypostoma, 20 mm. Length of coxopodite of last appendage on left side of cephalon, 10.5 mm.; length of basipodite of the same appendage, 5 mm. Diameter of coxopodite, 2 mm.; diameter of basipodite, 1.5 mm. Length of coxopodite on left side of the second segment of the thorax, 11 mm.; diameter, about 2.5 mm. Length of basipodite of the same, 5 mm.; diameter, about 1.5 mm. Length of ischiopodite, 3.5 mm.; diameter, about 1.5 mm. Length of meropodite, 2.5 mm. (this may be less than the total length as the segment is not completely exposed.) Distance between proximal ends of gnathobases of the fifth thoracic segment, about 7 mm. Distance between outer ends of the coxopodites of the first thoracic segment (estimated from measurements on the left side), 27 mm Distance apart of the dorsal furrows at the first thoracic segment, 27 mm. Length of the longest exopodite which can be traced, about 20 mm.

Isotelus maximus Locke.

([pl. 10, fig. 2.])

Illustrated: Mickleborough, Jour. Cincinnati Soc. Nat. Hist., vol. 6, 1883, p. 200, figs. 1-3 (endopodites and coxopodites). Walcott, Science, vol. 3, 1884, p. 279, fig. 1 (endopodites, coxopodites, and traces of exopodites). Woodward, Geol. Mag., dec. 3, vol. 1, 1884, p. 162, figs. 1-3 (copies of Mickleborough's figures). Bernard, The Apodidæ, 1892, text fig. 49. Beecher, Amer. Jour. Sci., vol. 13, 1902, p. 169, pl. 5. figs. 5, 6 (outline from one of Mickleborough's figures and an original figure). Walcott, Smithson. Misc. Coll., vol. 67, 1918, p. 133, pl. 24, figs. 3, 3a; pl. 25, fig. 1.

This specimen, which conies from the Richmond strata 2 miles north of Oxford, Ohio, is the best preserved of the specimens of Isotelus with appendages which has so far been found. The individual consists of two parts, the actual specimen, and the impression of the ventral side.

To describe it I am using very skillfully made plaster reproductions of both parts, presented to the Museum of Comparative Zoology by Doctor Charles D. Walcott, and presumably made after he cleaned the specimen as described in Science (1884). I have also an enlarged photograph (pl. 10, fig. 2) which seems to have been made after some later period of cleaning, probably by Professor Beecher, and I have examined the original specimens in Washington.

Viewed from the dorsal side, it is seen that the individual is very imperfect, the greater part of the cephalon being removed by a diagonal break which cuts off the anterior third of the left eye and extends to the front of the second thoracic segment on the right side. The ends of the pleura of both sides of the thorax are broken away, as are also the greater parts of the pleural lobes and the posterior end of the pygidium. On the ventral side, merely the posterior tips of the hypostoma remain, but the distal ends of the appendages were so far within the outer margin that the appendagiferous area is quite fully retained.

The most conspicuous feature of this specimen is the presence of nine pairs of large coxopodites behind the hypostoma, and of the remains of ten pairs of endopodites, making in all ten pairs of appendages which are easily seen. The apportionment of these segments to cephalon, thorax, and pygidium is not agreed upon by the people who have examined the specimens, but if one remembers that it is the outer and not the inner end of the coxopodite which articulates with the appendifer, it at once becomes evident that the first two pairs of appendages on the specimen are the last two pairs belonging to the cephalon, and that the next eight pairs are those of the thorax.

The impressions of fourteen pairs of coxopodites are readily counted on the pygidium, and as Doctor Walcott noted sixteen pairs on the actual specimens, his number was probably correct.

Cephalon.

Projecting the line of the back of the cephalon through from the dorsal side, it is found that the posterior tips of the hypostoma are 7 mm. in front of the posterior margin of the cephalon, and that the points of attachment of the posterior pair of cephalic appendages (the second pair shown on the specimen) are just within the posterior margin. The gnathobases of this pair of appendages extend back some distance beneath the thorax, and so give the impression that they belong to that part of the body. So far as can be determined, the cephalic appendages do not differ in any way from those of the thorax. On the mould of the ventral surface, just outside of the lateral edge of the right lobe of the hypostoma, is the somewhat imperfectly shown impression of the endopodite of the third cephalic appendage. The point of junction of the endopodite and coxopodite is about 2 mm. in front of the tip of the adjacent branch of the hypostoma, and the gnathobase is curved around just behind it. This accounts for three of the pairs of cephalic appendages. The second cephalic appendages must have thrust their gnathobases under the prongs of the hypostoma, and the endopodites were probably close to its edge. No trace of this pair appears on the specimen.

Thorax.

The thoracic appendages are the best preserved of any, and show the large coxopodites and the more slender endopodites which do not extend to the outer margin of the test. The latter extend forward and outward for about one half their length, then turn backward in a graceful curve.

Walcott's figure in Science shows hair-like markings on the under side of the right half of the thorax. These were interpreted by both Walcott and Beecher as fringes of the exopodites, but since the setæ of those organs on all other trilobites are always above the endopodites, while these are represented as below them, it would seem doubtful if this interpretation can be sustained. Furthermore, I find no trace of them on either cast or mould, and the actual specimen does not now show them.

Pygidium.

The coxopodites and endopodites of the pygidium seem to be similar to those on the thorax, but both are shorter and more slender, and the former decrease in length rapidly toward the posterior end. As mentioned above, it is not perfectly plain how many appendages are present, but I have accepted Doctor Walcott's count of sixteen pairs. Of the endopodites only the barest traces are seen, and of exopodites nothing.

One point of considerable interest in this specimen is the thickness, as it probably gives some measure of the space occupied by the animal. In Triarthrus and other trilobites from Rome, New York, the appendages are pressed directly against the dorsal test, but in this specimen a considerable space intervenes between the plane of the appendages and the shell. Between the central furrow and the inner surface of the dorsal test at the anterior end of the thorax is a distance of 13 mm. and under the dorsal furrows the thickness is about 7 or 8 mm., no accurate measurement being possible in the present state of the specimen.

Measurements: Length of specimen on median line, 121 mm.; probable original length, about 195 mm. (Walcott's restoration). Length of thorax, 58 mm.[1] Width of axial lobe at the first thoracic segment, 45 mm.; total width as preserved, 92 mm.; width as estimated from the mould of the ventral surface, no mm.; Walcott's restoration, 105 mm.

[1] If this specimen had the same proportions as specimens of Isotelus maximus from Toronto, the total length would be only 174 mm. The cephalon would be about 52 mm. long, the thorax 58 mm., and the pygidium about 64 mm. long.

Length of coxopodite of fourth left cephalic appendage, about 18 mm.; diameter, about 2.5 mm. Length of coxopodite of last left cephalic appendage, about 18.5 mm. Distance apart of inner ends of gnathobases of fourth cephalic appendages, about 4 mm. Distance apart of inner ends of endobases of first thoracic segment, about 6 mm. Distance apart of outer ends of coxopodites of first thoracic segment, about 43 mm.

Length of coxopodite of seventh left thoracic appendage 16 mm., diameter about 3.5 mm.; length of basipodite of the endopodite of the same appendage 6 mm.; diameter about 2 mm.; length of ischiopodite 5 mm.; length of meropodite 4.5 mm.; length of carpopodite 4.5 mm.; length of propodite 3 mm.; length of dactylopodite 2.75 mm.; total length of endopodite 25.75 mm.

Length of coxopodite of fourth left thoracic appendage 20 mm., diameter 4 mm.; length of five proximal joints of the endopodite 25 mm.; diameter of basipodite about 2 mm.

RESTORATION OF ISOTELUS.

(Text [fig. 9].)

The exopodites have been omitted from this restoration since nothing is known of their actual form. The chief reason for the figure is to contrast the greatly developed coxopodites of the posterior part of the cephalon and thorax with those of other trilobites. The antennules and first two pairs of biramous appendages of the cephalon are more or less hypothetical, and less is known of the appendages of the pygidium than is shown here. The restoration is based somewhat upon Walcott's figure in Science. The outline is that of a specimen of Isotelus maximus from Toronto, Ontario.

Isotelus gigas Dekay.

Illustrated: Woodward, Quart. Jour. Geol. Soc., London, vol. 26, 1870, text fig. 1; Geol. Mag., dec. 3, vol. 1. 1884, p. 78, text fig. Milne-Edwards, Ann. Sci. Nat, Zoologie, ser. 6, vol. 12, 1881, pl. 12, fig. 46. Walcott, Bull. Mus. Comp. Zool., Harvard Coll., vol. 8, 1881, pl. 2, fig. 9; Geol. Mag., dec. 4, vol. 1, 1894, pl. 8, fig. 9; Proc. Biol. Soc. Washington, vol. 9, 1894, pl. 1, fig. 9.

The specimen in the British Museum which Woodward called Asaphus platycephalus, is, in all probability, an Isotelus gigas. Woodward says of it:

I was at once attracted by a specimen of Asaphus, from the Black Trenton Limestone (Lower Silurian), which has been much eroded on its upper surface, leaving the hypostoma and what appear to be the appendages belonging to the first, second, and third somites, exposed to view, united along the median line by a longitudinal ridge. The pseudo-appendages, however, have no evidence of any articulations. But what appears to me to be of the highest importance, as a piece of additional information afforded by the Museum specimen, is the discovery of what I believe to be the jointed palpus of one of the maxillæ, which has left its impression upon the side of the hypostoma—just, in fact, in that position which it must have occupied in life, judging by other Crustaceans which are furnished with an hypostoma, as Apus, Serolis, etc.

The palpus is 9 lines in length, the basal joint measures 3 lines, and is 2 lines broad, and somewhat triangular in form.

There appear to be about 7 articulations in the palpus itself, above the basal joint, marked by swellings upon its tubular stem, which is 1 line in diameter.

Fig. 9.—A restored composite of Isotelus maximus and I. latus. The exopodites are left out because entirely unknown. Drawn by Doctor Elvira Wood. Natural size.

Desiring to know more of this individual, I wrote to Doctor Bather and was surprised to learn that the specimen which was the basis of Woodward's observations is so badly preserved as to be of no real value. With his permission, I append a note made by Doctor Bather some years ago when selecting fossils to be placed on exhibition:

Asaphus gigas Dekay. Ordovician, Trenton Limestone. N. America, Canada. Descr. H. Woodward, 1870, Q. J. G. S., XXVI, pp. 486-488, text fig. 1, as Asaphus platycephalus. Coll. and presd. J. J. Bigsby, 1851. Regd. I 14431.

This specimen is in the Brit. Mus. Geol. Dept. I 14431. The supposed hypostome is exceedingly doubtful; it lies dorsad of the crushed glabellar skeleton. The "appendage" is merely the edge of a part in the head-shield; the maxilla is some calcite filling, between two such laminæ.

13 Sept. 1911. (Signed) F. A. BATHER.

Walcott figured a slice of Isotelus gigas from Trenton Falls, New York, which shows a few fragments of appendages, but is of particular importance because it shows the presence of well developed appendifers beneath the axial lobe.

Isotelus arenicola Raymond.

Illustrated: Ottawa Nat, vol. 24, 1910, p. 129, pl. 2, fig. 5.

The following quotations from my paper are inserted here to complete the record of appendage-bearing specimens:

A rather remarkable specimen of this species was found by W. C. King, Esq., on the shore of Lake Deschenes at Britannia [near Ottawa, Ontario]. This specimen is an impression of the lower surface of the trilobite, and shows a longitudinal ridge corresponding to the central furrow along the axis of the ventral side of the animal, ten pairs of transverse furrows, and the impression of the hypostoma. The doublure of the pygidium has also left a wide smooth impression, but in the cephalic region the hypostoma is the only portion of which there are any traces remaining. The specimen was found on a waterworn surface of the beach, partially covered by shingle….

The transverse furrows are the impressions left by the gnathobases of the basal joints of the legs. They were evidently long and very heavy, but the specimen has been so abraded that all details are obscured. The first six pairs of impressions are longer and deeper than the four behind. The first eight pairs seem to pertain to the thoracic appendages, while the last two belong to the pygidium. From the posterior tips of the hypostoma to the first gnathobases of which traces are present there is a distance of about 22 mm. without impressions. In Isotelus gigas the hypostoma normally extends back to the posterior margin of the cephalon, so that it seems that in this specimen the impressions of the first two pairs of gnathobases under the thorax may not have been preserved. In that case, the six pairs of strong impressions may represent the last six pairs of thoracic segments, and the pygidium might begin with the first of the fainter ones.

Horizon and locality: From the sandstone near the base of the Aylmer (Upper Chazy) formation at Britannia, west of Ottawa, Ontario. Specimen in the Victoria Memorial Museum, Geological Survey of Canada, Ottawa.

The Appendages of Triarthrus.

Triarthrus becki Green.

(Pls. [1]-[5]; pl. [6, figs. 1-3]; text figs. [1], [10], [11], [33], [42].)

(Also see [Part IV].)

Illustrated: Matthew, Amer. Jour. Sci., vol. 46, 1893, pl. 1, figs. 1-7;—Trans. N. Y. Acad. Sci., vol. 12, pl. 8, figs. 1-7.—Beecher, Amer. Jour. Sci., vol. 46, 1893, text figs. 1-3;—Amer. Geol., vol. 13, 1894, pl. 3;—Amer. Jour. Sci., vol. 47, pl. 7, text fig. 1;—Amer. Geol., vol. 15, 1895, pls. 4, 5;—Ibid., vol. 16, 1895, pl. 8, figs. 12-14; pl. 10. fig. 1;—Amer. Jour. Sci., vol. 1, 1896, pl. 8; Geol. Mag., dec. 4, vol. 3, 1896, pl. 9;—Eastman-Zittel Text-book of Paleontology, vol. 1, 1900, text figs. 1267-1269;—2d ed., 1913, fig. 1375; Studies in Evolution, 1901, reprint of all previous figs.;—Amer. Jour. Sci., vol. 13, 1902, pl. 2, figs. 1-5; pl. 3, fig. 1; pl. 4, fig. 1; pl. 5, figs. 2-4;—Geol. Mag., dec. 10, vol. 9, 1902, pls. 9-11, text figs. 1-3.—Walcott, Proc. Biol. Soc. Washington, vol. 9, 1894, pl. 1 figs. 1-6;—Geol. Mag., dec. 4, vol. 1, 1894, pl. 8;—Smithson. Misc. Coll., vol. 67, 1918, pl. 29, figs. 1-11; pl. 30, figs. 17-20; pl. 32; pl. 34, figs. 4-7; pl. 35, fig. 5.—Bernard, Quart. Jour. Geol. Soc., London, vol. 50, 1894, text figs. 11, 12.—Œhlert, Bull. Soc. Géol. France, ser. 3, vol. 24, 1896, text figs. 1-17, 34.—Jaekel, Zeits. d. d. geol. Gesell., vol. 53, 1901, text fig. 24. Moberg, Geol. Fören. Förhandl., vol. 29, pl. 5, 1907, pl. 4, fig. 2; pl. 5, fig. 1.—Handlirsch, Foss. Insekten, 1908, text fig. 6.—Tothill, Amer. Jour. Sci., vol. 42, 1916, p. 380, text fig. 5.—Crampton, Jour. N. Y. Entomol. Soc., vol. 24, 1917, pl. 2, fig. 20.

Historical.

Specimens of Triarthrus retaining appendages were first obtained by Mr. W. S. Valiant from the dark carbonaceous Utica shale near Rome, New York, in 1884, but no considerable amount of material was found until 1892. The first specimens were sent to Columbia University, and were described by Doctor W. D. Matthew (1893). This article was accompanied by a plate of sketches, showing for the first time the presence of antennules in trilobites and indicating something of the endopodites and exopodites of the appendages of the cephalon, thorax, and pygidium. Specimens had not yet been cleaned from the lower side, so that no great amount could then be learned of the detailed structure. Matthew concluded that "The homology with Limulus seems not to be as close in Triarthrus as in the forms studied by Mr. Walcott; but the characters seem to be of a more comprehensive type, approaching the general structure of the other Crustacea rather than any special form."

Professor Beecher's first paper, dated October 9, 1893, merely mentioned the fact that the Yale University Museum had obtained material from Valiant's locality, but was quickly followed by a paper read before the National Academy of Sciences on November 8, and published in December, 1893. This paper described particularly the thoracic appendages.

This was followed in January (1894 A) by an article in which some information about the mode of occurrence of the specimens was added, and in April (1894 B), the limbs of the pygidium were described and figured. The determination of the structure of the appendages of the head evidently presented some difficulty, for the article describing this portion of the animal did not appear until the next February (1895 A). This cleared up the ventral anatomy of Triarthrus, and was followed by a short article (1896 A) accompanied by a restoration of the trilobite showing all the appendages.

This ended Professor Beecher's publications on Triarthrus until his final paper in 1902, although he contributed some of his results and figures to his chapter on the trilobites in the Eastman-Zittel Text-book of Paleontology in 1900.

The discovery of these excellent specimens had of course excited very great interest. Doctor Walcott also studied a number of specimens from Valiant's locality, and published in 1894, with some original figures, the results of his comparison of the appendages of Triarthrus with those of Calymene and Ceraurus.

In his article on the "Systematic Position of the Trilobites," Bernard (1894) used the results of Professor Beecher's studies of 1893, and also quoted the papers by Matthew (1893) and Walcott (1894), though the article by the latter appeared too late to be used except for a note added while Bernard's paper was in press. A final footnote quoted from Professor Beecher's paper of April, 1894 (1894 B).

Œhlert (1896) gave an excellent summary in French of the work of Beecher and Walcott on Triarthrus, with reproductions of many of their figures.