THE UNIVERSITY OF ILLINOIS LIBRARY — 0-1096 FIELD MUSEUM OF NATURAL HISTORY PUBLICATION 224 ZOOLOGICAL SERIES VOL. XIV, No. 3 THE BRAINS OF THE SOUTH AMERICAN MARSUPIALS CAENOLESTES AND OROLESTES BY JEANNETTE BROWN OBENCHAIX Hull Laboratory of Anatomy, University of Chicago WILFRED H. Oscoon Curator, Department of Zoology CHICAGO, U. S. A January 26, 1925 TBt U8HAHT Of THf APR 9 1925 UNIVERSITY «r ILUNM8 LiKKAKT !J' '! OCT 2 8 1924 UNIVERSITY OF ILLINOIS FIELD MUSEUM OF NATURAL HISTORY PUBLICATION 224 ZOOLOGICAL SERIES VOL. XIV, No. 3 THE BRAINS OF THE SOUTH AMERICAN MARSUPIALS CAENOLESTES AND OROLESTES BY JEANNETTE BROWN OBENCHAIN Hull Laboratory of Anatomy, University of Chicago WILFRED H. OSGOOD Curator, Department of Zoology CHICAGO, U. S. A. January 26, 1925 THE UflfiAKT OF TBf APR 9 1925 UNIVERSITY «F ILLINOIS THE BRAINS OF THE SOUTH AMERICAN MARSUPIALS CAENOLESTES AND OROLESTES BY JEANNETTE BROWN OBENCHAIN CONTENTS PAGE Introduction 175 External anatomy of the brains of Caenolesies and Orolestes 178 Internal anatomy of the cerebral hemisphere of Caenolestes 188 Primary olfactory area 189 Secondary olfactory areas 189 Anterior olfactory nucleus 191 Tuberculum olfactorium 195 Septal region 199 Pyriform lobe 200 Hippocampal formation 206 Corpus striatum 220 Neopallium 221 General considerations 222 A Summary 225 ^ Acknowledgments 227 Bibliography 228 Abbreviations 231 ^ INTRODUCTION Caenolestes and Orolestes are tiny shrewlike marsupials five inches *'m length of head and body exclusive of the slender tail. They are Natives of high Andean forests from Venezuela southward into Peru, .^and although discovered in 1863, have not been known to science, CA except by imperfect material, until recent years. Their peculiar denti- tion has made their assignment to one or the other of the marsupial suborders (Diprotodontia and Polyprotodontia) a matter of much un- ^ certainty. This question remains even yet, perhaps, in suspense, since 1* the one brain character of evidential value seems to reverse the latest ^conclusion, based on careful sifting of all non-nervous characters ^ (Osgood, 1921; Obenchain, 1923). 175 176 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. Aside from its bearing upon classification, the study of the brains of these species is warranted by many other considerations. These small marsupials, surviving members of an ancient American group, are characterized by the retention of many primitive features, by the absence of any great degree of specialization, by few non-marsupial characters, by numerous resemblances to the modern peramelids (Aus- tralian polyprotodonts), and by lack of affinity to the other living American marsupials, the polyprotodont Didelphiidae (Osgood, 1921, pp. 150-151). Material. — The material which forms the basis of the present study consists of the brains of three female specimens, one of Caenolestes obscurus and two of Orolestes inca. The first brain came from an adult female of Caenolestes collected in 1911 in the Colombia- Venezuela boundary region by Dr. Wilfred H. Osgood of the Field Museum of Natural History; it was examined by Dr. C. Judson Herrick, who described and figured its external surface in a short paper (1921, antea, pp. 157-162, 3 figures) appended to Dr. Osgood's long monograph on the anatomy and zoological position of Caenolestes. Dr. Herrick later offered me the opportunity of studying this brain, if I could con- vert it into serial sections. The other two brains, from adult females of Orolestes inca, were collected by Mr. E. Heller of the Yale National Geographic Society Peruvian Expedition of 1914-15, and were loaned by the U. S. National Museum to the Field Museum to be transferred to me for the purpose of this study. The first specimen alone (Caenolestes) has been sectioned and stained. The description of the internal anatomy is therefore based entirely on this one trans- verse series, which, although somewhat imperfect because of incom- plete fixation within the unopened skull, has stained with unexpected brilliancy by the iron-haematoxylin method. This series, owing to the fact that both cells and fibers have been stained, serves remarkably well, all things considered, for the study here undertaken. Since the brain was sectioned whole, its medial surface is known only by means of wax and linear reconstructions made from the sections. These are believed to be reasonably trustworthy, but since they are made from an imperfect series, additional corroboration was greatly to be desired. Therefore the two brains of Orolestes were especially welcome. I take this opportunity to thank Mr. W. J. Owen of the Anatomical Institute of Melbourne, Australia, for his kind assistance in the at- tempt to remove them as nearly intact as possible from the skulls. The smaller of the two brains (No. 194948) was divided by a sagittal cut, the right hemisphere was detached from the brain stem, and drawings 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 177 of the medial surface of both the whole brain and the detached hemi- sphere were made by Mr. Kenji Toda, of the University of Chicago, to whom I am indebted for the five beautiful and accurate air brush pictures of the brain of Orolestes (Figs. 1-5). The larger brain (No. 194921) was used for the external views, with correction and supple- ment from the other where necessary, as in the case of the parafloccular lobes, which the larger brain had lost. The business of clearing these small and fragile brains from membranes and coagulated cerebro-spinal fluid was both tedious and difficult, but fairly successful in the end, and the figures are offered as faithful representations of the actual specimens. Three air brush drawings of the brain of Caenolestes by Mr. A. B. Streedain, formerly of the Department of Anatomy, also may be consulted (Herrick, antea, pis. XXI-XXII). This brain, owing to the difficulty of removing membranes and debris, was in somewhat less favorable condition for drawing than were the other two. This applies particularly to the ventral surface, which was further marred by a considerable hole due to faulty fixation. All three specimens were originally preserved whole, with only a ventral longitudinal opening of the body, in 10% formalin, and sub- sequently transferred to alcohol. In the case of Caenolestes, the brain, when freed from the skull, ballooned out with a glistening white sur- face, but later shrunk and darkened. The other two brains (Oroles- tes) were quite rigid when uncovered, and presented a dull, light brown surface. All three completely filled the skull. It is hoped that the two brains of Orolestes may be successfully converted into sagittal and horizontal series. Method. — The brain of Caenolestes, infiltrated in 42 degres paraf- fine and blocked in 56 degree paraffine (to minimize brittleness), was cut 10 micra anteriorly and 15 micra farther back, and made about 1 200 sections. About 850 sections were drawn at a magnification of 25 diameters with the aid of the Edinger projection apparatus. These have proved a very useful close series for annotation. A wax model made from these drawings, while somewhat disappointing, especially in the. midregion, owing to the emptiness of the greatly expanded ven- tricles (the fragility of the specimen precluded aspiration) and conse- quent wrinkling and uneven spreading of the thinner portions of the walls, has been an invaluable aid to study. Linear reconstructions drawn to scale have also proved indispensable. An additional series of 31 Edinger drawings with a magnification 85 was made from the precommissural hippocampal region. 178 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. EXTERNAL ANATOMY Dr. Herrick, in the first account ever given of the brain of Caeno- lestes (or of any caenolestid brain), emphasized its extreme simplicity, the enormous development of its rhinencephalon, the smoothness and limited extent of its neopallium, and called attention to the fact that it most closely resembles the brains of the lowly Australian polyproto- donts, Notoryctes and Perameles. To this it will be necessary to add the description of the median section of the whole brain and median surface of the hemisphere, as well as a discussion of the other sur- faces in the light of knowledge of the internal structure of Caenolestes, with reference to changes of nomenclature, or in comparison with the brain of Orolestes. This will of course involve some unavoidable repe- tition. The two types, Caenolestes and Orolestes, exhibit only minor dif- ferences in external anatomy. Although all three brains are supposed to have come from adult female specimens, they differ somewhat in size and proportions, as may be observed in the table of measurements. Caenolestes is the smallest of all, but it has the longest hemisphere. Its cerebellum is smaller and less plump than the other two, and ex- hibits an extra furrow and convolution — not a matter of great impor- tance, as will be seen below under the description of the cerebellum. It is, however, to be remembered that the three brains represent two genera of caenolestids. The olfactory bulbs are truly enormous, almost half the total length of the hemisphere at the medio-ventral angle. They are of the "sessile" type, with no visible olfactory peduncle in the intact brain. Following Livini (1908), we have called the deep constricting sulcus which marks the posterior border of the olfactory formation, the fissura circularis (fs.circ., Figs, 6, 12, 15, 25-28; unlabelled, Fig. 1-5). The circu- lar fissure is a compound structure whose component parts are homol- ogous to portions of fissures described in the brains of other animals. Medially it represents the anterior portion of the fissura prima of His, above the antero-medial tip of the rhinal arcuate fissure (fs.rh.acr., Figs. 12, 28) ; dorsally, after meeting the medial prolongation of the rhinal fissure (fs.rh.a., Figs. 15, 25) at its medial end, it diverges from the latter to leave the interval for an exposed triangle of dorsal olfac- tory peduncle, and then meets the rhinal fissure again as the latter turns caudad on the lateral stirface of the brain (Fig. 14) ; meets also the anterior end of the endorhinal fissure (fs.erh., Fig. 13) just below, and then drops downward across the wide lateral olfactory tract (tr. 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 179 ol,L, Fig. 13) overlying the pyriform lobe, to continue as a deep con- striction across the base of the brain between the olfactory bulb (b.ol. and the olfactory tubercle (t.ol., Fig. 15), finally joining the anterior end of the medial portion of the rhinal arcuate fissure (Fig. 15), and thus completely defining the caudal border of the olfactory formation. Dorsally the olfactory bulb is overhung by the projecting frontal pole of the hemisphere, which conceals the exposed olfactory peduncle from view. The rest of the olfactory peduncle (see anterior olfactory nucleus, page 191) is thrust deep into the heart of the olfactory bulb, as recognized long ago in this type of brain by Elliott Smith and others. This condition is partly the result of the great caudal extension of the olfactory bulb itself, partly one of the numerous expressions of the very considerable fore-and-aft "telescoping" characteristic of this brain, which puts it in rather marked contrast to more elongated brains like that of the Virginia opossum (Didelphis virginiana), for example. Didelphis marsupialis, however, resembles the caenolestid brain in this respect (Beccari, 1910, Fig. 18). The great hemispherical mass of the tuberculum olfactorium (t.ol.} is also delimited by an encircling sulcus, the fissura rlrinalis arcuata (fs.rh.arc., an adaptation of one of Retzius' names; see Herrick, 19243). Medially the fissure is homologous to the caudal portion of the fissura prima of His. Its posterior more transverse portion has been called the fissura diagonalis by Beccari (1910), but his median continu- ation of it upward in the septum is not followed here (see tuberculum olfactorium, page 196). Laterally the rhinal arcuate fissure is usually taken to be homologous to the endorhinal fissure, but this is not strictly accurate, since the endorhinal fissure lies wholly within the pyriform lobe and therefore above the tuberculum, as has been noted in other forms by Smith, Johnson and others. The endorhinal fissure really marks the dorsal border of the massive part of the lateral olfactory tract, which is very wide rostrally, but rapidly diminishes caudally through the loss of fibers to the pyriform lobe and tuberculum olfactorium, so that in its caudo-ventral course it gradually approaches the rhinal arcuate fissure and finally meets it near the posterior limit of the tuberculum (Figs. 13, 15, 33-34). The greatly reduced mas- sive part of the lateral olfactory tract lies at that level within the endorhinal fissure, where the latter fuses with the rhinal arcuate fissure and so comes to an end. In its caudo-ventral course the lateral ol- factory tract leaves a thin film of fibers (tr.ol.l.d., Figs. 25-33) cover- ing the pyriform cortex almost to the height of the rhinal fissure and far toward the caudal pole of the pyriform lobe. 180 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. Another important fissure visible on the lateral surface of the hemi- sphere of these brains is the rhinal fissure (fs.rh., to be seen in the majority of the figures), which separates the dorsal or neopallial con- vexity from the latero-ventral pyriform lobe. It traces an almost exactly horizontal course backward over the lateral surface of the brain at the very high level characteristic of the lowly marsupials Notoryctes (where it is internal only) and Perameles and the insectivore Erina- ceus, forms whose neopallial development is the least extensive among mammals. Anteriorly the rhinal fissure is a continuation of the deep groove between the projecting frontal pole of the neopallium and the exposed dorsal part of the olfactory peduncle (fissura rhinalis medialis of some writers, but here called the fissura rhinalis anterior, fs.rh.a., Figs. 12-15), which bends sharply on to the lateral surface of the hemisphere just above the rostral end of the endorhinal fissure and runs horizontally backward at this high level to round the caudal pole of the hemisphere, where it appears upon the median surface as the median rhinal fissure running obscurely forward into the subicular border of the hippocampus (fs.rh.m., Figs. 6, 12, 42-43; rh., Fig. 5, where it probably coincides only caudally with the diagonal groove shown — there has been no control with sections in Orolestes so far). Even where imperfectly developed externally the position of the fis- sura rhinalis is unmistakably discernible throughout, and internally it is in Caenolestes always clearly defined. The dorsal convexity or neopallium contains only one other ex- ternally visible fissure. In Caenolestes, "about one-fifth of the dis- tance from the frontal to the posterior pole of the hemisphere there is a distinct, though shallow, transverse sulcus which probably repre- sents the sulcus orbitalis of Elliot Smith's descriptions (Herrick, 1921, antea, p. 158). Ventrally the orbital fissure is "obscurely con- fluent" with the rhinal fissure. In the smaller of the two brains of Orolestes this fissure seems to be absent as an external groove, but its position is apparently marked by blood vessels. In the reconstruc- tions of Caenolestes also it does not show up, and a cursory examina- tion of the sections has failed to reveal it. A more careful study might disclose its position. The pyriform lobe comprises the larger part of the lateral and ven- tral aspects of the hemisphere. Its dorsal boundary is the rhinal fis- sure, which is at the same time the latero-ventral boundary of the neo- pallium. Antero-laterally the pyriform lobe contains the very sharp endorhinal fissure (as explained above, page 179) and below this fis- stire its antero-lateral boundary is the lateral portion of the rhinal 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 181 arcuate fissure (sharply incised in Orolestes but not depressed at all in Caenolestes) , a fissure marking the ventral border of the massive portion of the lateral olfactory tract, as the boundary of the tuber- culum, since the latter appears just ventral to this line. The greatly widened caudal portion of the pyriform lobe is divided into two distinct parts by a very important fissure so faintly indicated externally that its existence was unsuspected before the brain was sectioned. This is the fissura amygdaloidea (1913, 1923) of Johnston (fs.amg., Figs 2, 3, 6, 12, 13, 15, 36-43). In the sections it branches quite clearly from the rhinal arcuate fissure a short distance in front of the caudal pole of the tuberculum (between Figs. 35 and 36), and very soon disappears as a recognizable fissure. Internally it is quite distinct throughout its course, marking the quasi -horizontal line of junction between the sharply defined ventral edge of the pyriform cor- tex and the lateral border of the amygdaloid complex which occupies practically the entire base of the hemisphere behind the tuberculum. The rather pronounced sculpturing of this region is due to the state of development of the different amygdaloid nuclei, which have been so thoroughly worked out in the Virginia opossum by Johnston (1923). These will be more fully described further on, in connection with the internal anatomy of this brain (page 201). In Caenolestes a very salient caudo-ventro-lateral prominence cor- responds to the "natiform eminence" of Elliot Smith (i895b) in Notoryctes (em.nat., Figs. 37-41 ; unlabelled, Fig. 2, pi. XXI, antea, Herrick; cf. Orolestes, figure 3 here, which shows it much less dis- tinctly). It is due both to an actual thickening of the lateral wall of the hemisphere in this region and, in greater degree perhaps, to the lateral cupping of the hemisphere to accommodate the wide midbrain. Elliot Smith thinks that the temporal bending of the hemisphere has been a main factor in the formation of this eminence. This is much more strongly suggested in Notoryctes than in Caenolestes (see page 213)- The median section of the brain of Caenolestes is known only in wax and linear reconstructions, but the smaller of the two brains of Orolestes was divided sagittally, and figures 4 and 5 show respectively the left median section of the whole brain and the median surface of the right hemisphere. Figure 6 is a linear reconstruction of the- left hemisphere (reversed) of Caenolestes. The third ventricle and external medial surface of the brain of Orolestes were so filled or coated with a hard coagulum of the same color as the tissue and so tightly adherent to it that it was not found iS>2 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. possible to remove it without some laceration of the tissue. Thus the edges of the massa intermedia (vn.i., Fig. 4) were somewhat frayed, the aqueduct was probably widened, and the epiphyseal attachments and choroid roof of the third ventricle were torn. But these defects, regrettable as they are, do not greatly impair the value of the section for present purposes. At least the sacrifice of the other brain with the prospect of no better success did not in the circumstances seem to be justified. The large anterior or ventral commissure of Orolestes (y., Figs. 4 and 5) is slightly oval and considerably darker than the rest of the surface, standing out sharply. That of Caenolestes (v., Fig. 6) is similar in size, shape and position. The dorsal commissure (rf.) of Orolestes appears as a moderately darker and distinctly bilaminar mass occupying the dorsal and posterior borders of this portion of the lamina terminalis (lamina supraneuroporica of Johnston), en- closing a small whitish triangular space which gives the impression of being without transverse fibers. The dorsal commissure of Orolestes (counting it only as the darkened bilaminar mass) thus exhibits mac- roscopically the typical marsupial form, while that of Caenolestes (d., Fig. 6) as reconstructed from sections, shows only the merest hint of bilaminarity in the slightly reentering anterior angle and the dorso-caudal prolongation ("splenium"), suggesting an intermediate type between the solid rounded type of the monotremes and the bilam- inar type of the marsupials. The reconstruction of Caenolestes was made from the exact midline of the sections, the outline enclosing only the area through which commiss'ural fibers were coursing. These fibers are much less dense in the ventral than in the dorsal region of the commissure, owing to the intermingling of many cells (bed or nucleus of the dorsal commissure) with the fibers in its ventral region. In Orolestes the third ventricle (Fig. 4), as seen from the median surface, has the form of a tall parallelogram tipped slightly backwards and downwards. Its walls are formed anteriorly by the lamina ter- minalis, whose upper half is greatly thickened by the two commissures, while the lower half remains very thin ; posteriorly, by the almost ver- tical mammillary body and tuberculum posterius in line with the ros- tral opening of the aqueduct (aq.) and the thick anterior end of the tectum mesencephali (tect.} ; dorsally by the choroid roof and the epiphysis ; ventrally by the thin floor plate containing the small chiasma (ch.) rostrally, and the infundibular lumen more caudally; laterally, by the median surfaces of the thalami. The upper half of the third ventricle is almost filled by the enormous oval massa intermedia, which 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 183 so nearly approaches the neighboring walls as to leave on three sides of it only two very narrow passages leading into the aqueduct. The lower half of the ventricle is quite open, and the thalamic lateral wall displays a diagonal sulcus running backward and downward from the middle of the lower border of the massa intermedia to the posterior wall of the ventricle just above the floor. The ventricle is also deep- ened laterally just above the floor in two places — the preoptic recess and the mammillary recess. The interventricular foramen opens into the narrow canal between the dorsal commissure and the massa inter- media. A very sharp midbrain flexure results in the formation of a deep and narrow vertical cleft across the base of the brain which folds a section of it between the pons and the mammillary region together like the pages of a book. The tectum mesencephali (tect.) is entirely con- cealed from above by overlying structures, the rostral three-quarters by the cerebral hemispheres and the caudal quarter by the anterior tip of the median lobe of the cerebellum. (Figure 4 shows a gap between the hemisphere and the cerebellum, due to a mass of coagulum after- wards cleared away ; the relations in Caenolestes and Orolestes are the same.) "This is in contrast to the usual marsupial arrangement, for the corpora quadrigemina are in most cases well exposed dorsally. (Petaurus is another exception; see Elliott Smith, 1895, p. 188)" (Her- rick, 1921, antea, page 158). The condition in Orolestes and Caeno- lestes is apparently one of the many expressions of the fore-and-aft compression already mentioned. Partly in response to the crumpling of the brain base as described above, the tectum displays a marked caudal prolongation, emphasized by its recurved keel, as indicated by the forward point of attachment of the anterior medullary velum (v.m. a., Fig. 10). Laterally the posterior colliculus is even more caudally extended (col. p., Figs. 7, 8, 10). The cerebellum, which will be de- scribed below, fits snugly into and around these structures, behind an almost perpendicular anterior medullary velum, and below, behind and above the tectum. The median surface of the cerebral hemisphere of Orolestes (Fig. 5) and of Caenolestes (Fig. 6) exhibits a flattened surface anteriorly, in close apposition with its fellow, and a deeply concave postero-median surface, hollowed out to accommodate the bulky midbrain. It is tra- versed by four prominent arcuate fissures, concentrically arranged and different in extent and curvature. These are, in order from without in, the hippocampal, fimbrio-dentate, fimbrio-alvear (to coin a term), and choroid fissures. 184 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. The outermost or hippocampal fissure (fs.hip., Figs. 5 and 6) ap- parently describes almost three-quarters of a circle in Orolestes (Fig. 5), stretching from the fissura circularis (fs. circ.) upward, backward, downward and forward nearly to the posterior end of the tuberculum olfactorium. The same fissure in the reconstruction of Caenolestes (fs.hip., Fig. 6) falls short of this at both ends. It does not reach the fissura circularis rostrally and it ends more briefly caudally near the medial prolongation of the fissura amygdaloidea, as the latter goes forward towards the posterior end of the choroid fissure. Perhaps also in Orolestes what appears to be the strongly recurved temporal end of the hippocampal fissure is really the fissura amygdaloidea me- dialis. Indeed this interpretation is suggested by the sudden pro- nounced shallowing and narrowing of the fissure at the exact point where a slight diagonal fissure runs into it from behind. The pos- terior end of the diagonal fissure coincides with the apparent level of the caudal end of the amygdaloid fissure on the lateral surface, which on the figured lateral view of the left hemisphere (Fig. 3) is much less clearly indicated than on the right (Fig. 5). The sections of this particular right hemisphere will, when made, clear up this point, which now rests partly on the external configuration of the hemisphere of Orolestes and partly on the internal configuration of that of Caenolestes. The second fissure, the fimbrio-dentate fissure (fs.fim.d.) extends also in Orolestes (Fig. 5) to the fissura circularis, just below the fis- sura hippocampi. In Caenolestes, however, it drops sharply downward away from the hippocampus entirely, and ends briefly just in front of the dorsal commissure (Fig. 6). This portion of the fissure seems to be purely a response to the pressure of a blood vessel which is lodged in the canal formed by the corresponding fissures of the two hemispheres (Figs. 6, I7b). The postcommissural and main portion of the fimbrio- dentate fissure separates the gyrus dentatus (gy.dent.) not from the massive fimbria, but from the extraventricular ammon's horn or in- verted hippocampus, with its thin coating of subpial alveus fibers, and it is therefore really an o/wo-dentate fissure in this region. Posteriorly it is both less extensive and less recurved than the hippocampal fissure. The third fissure seems to have no name, but it is nevertheless quite distinct as the line of demarcation between the massive fimbria and the alveus-coated extraventricular hippocampus (ammon's horn), defining its rolled ventral limit. It might be designated as the fimbrio- alvear fissure (fs.fim.al., Figs. 6, 173; unlabelled, Figs. 5, 35-38). It begins rostrally just behind the dorsal commissure as an offshoot of 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 185 the fimbrio-dentate fissure, dropping rapidly away from the fimbrio- dentate fissure as the extraventricular hippocampus widens, and it runs out caudally as the massive fimbria tapers down to nothing at the posterior end of the choroid fissure (just in front of figure 39). The innermost or choroid fissure (unlabelled, Fig. 5; fs.ch., Figs. 6, 173, 35-39; Fig. 39 is really just behind the caudal end of the choroid fissure) extends from the ventral tip of the dorsal commissure backwards and downwards in a smooth curve outlining the ventral border of the fimbria and the thalamo-hemi spheric junction. In Caeno- lestes, as noted above, the internal medial prolongation of the fissura amygdaloidea can be quite clearly traced forward to the neighborhood of the caudal end of the choroid fissure. Other features of the medial surface of the hemisphere are the tuberculum olfactorium in profile behind the olfactory bulb, followed immediately in Orolestes by the small ventro-median eminence (Fig. 5, unlabelled) which is probably due to the nucleus of the lateral ol- factory tract (nuc.tr.ol.l., Figs. 36-38). This small tubercle appears in Notoryctes (Elliot Smith, i895b) and in a number of other mar- supials and lowly eutherian mammals, but is apparently absent or not very prominent in Caenolestes, although the nucleus, as will appear below, is well developed. It is, however, partly covered by the tuber- culum olfactorium, and this masks it to some extent. The amygdaloid complex (amg., Fig. 6; unlabelled, Fig. 5) extends upward as high as the position of the medial portion of the amygdaloid fissure (fs.amg. w.) already described, and the posterior pyriform cortex (cx.pir.p.) curves around the caudal pole of the hemisphere to form the more temporal subicular border between the medial extensions of the rhinal and amygdaloid fissures, in the familiar mammalian pattern. From the postero-medial extension of the rhinal fissure (fs.rh.m.) forward to the antero-medial extension of the same fissure (fs.rh.a.) the space above the hippocampal fissure is occupied by neopallium. In front of the anterior rhinal fissure the hippocampal cortex is in cellular con- tinuity with the dorso-medial portion of the anterior olfactory nucleus (nuc. ol. ant.d., Fig. 26), the peduncular or postbulbar gray matter which laterally merges without interruption into the pyriform cortex (see page 192). Save for these rostral and caudal junctions the lateral and medial olfactory cortices are split apart dorsally by the wedge- like neopallial cap of the hemisphere. The entire length of the fissura prima of His appears upon the median surface of the hemisphere as the visible portions of the fis- sura circularis rostrally and fissura rhinalis arcuata caudally (see page i86 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. 178). Between it and the lamina terminalis lies the precommissural area (paraterminal body) of Elliot Smith, the parolfactory region of Johnston, the septum of most neurologists (a. prcom., Fig. 6; Unla- belled, Fig. 5). The nucleus parolfactorius medialis, the nucleus of the diagonal band of Broca, the precommissural fornix, olfacto-hippo- campal fibers, and the medial forebrain bundle (fasciculus medialis telencephali), are the more superficial structures lying above the pos- terior part of the fissura prima (medial part of the fissura rhinalis arcuata). The sharply grooved dorso-medial margin of the hemisphere is the imprint of the longitudinal sinus, which is seen in position in fig- ure 4. The enormously expanded lateral ventricle of Caenolestes (v.l., Figs. 28-42) is probably not pathological, but the result of defective fixation. As a cursory inspection of the more caudal sections will show, a recurved temporal horn has not even begun to form. The posterior horn, present in the Virginia opossum, is also absent in Caenolestes. In the lower rostral wall of the ventricle a flaring opening leads into a narrow crooked canal, which in turn expands into a widened terminal sac. These constitute the olfactory ventricle (v.ol., Figs. 12- 15, 17,23-27). Cerebellum. — (Figs. I, 3, 4, 7-10, ica and lob, 41-44; also Herrick, 1921, antea, pi. XXI.) The cerebellum of Caenolestes and Orolestes (the two forms are practically identical in general structure) very neatly fills the gap between those of Notoryctes and Perameles, the two simplest mammalian cerebella hitherto described. (Elliot Smith, i9O2b, I9O3C, d, e). The cerebella of the insectivores Macroscelides and Erinaceus (also described by Elliott Smith, I9O2C, i9O3c) probably oc- cupy the fourth and fifth places in the series. The cerebellum of the Virginia opossum is considerally more complex than any of those mentioned. Of the three fundamental cerebellar lobes separated by the fissura prima (fs.pr.) and fissura secunda (fs.s.), the median and posterior lobes differ little in the first four forms named above. But the anterior lobe grades very clearly from the simple unfissured lobe of Notorcytes (loa), through Caenolestes (Figs. 4, 7-10) or Orolestes, with two lobules separated by the deep fissura preculminata (fs.prcul.), and a third well developed lobulus, the lingula (/#.), which is not separated from the rest by a fissure, to Perameles (Fig. lob), with an anterior lobe crossed in the midline by four fissures of varying depth. The anterior lobe and the fissura prima are entirely concealed in the intact 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 187 brain. The visible portion of the cerebellum is formed by the exposed parts of the median and posterior lobes. It is crossed in the midline by a ventrally concave fissure, the fissura secunda (/J.J.), separating a small postero-medial convolution, the uvula (uv.) belonging to the posterior lobe, from the larger median lobe above. In Caenolestes but not in Orolestes, the secondary fissura suprapyramidalis (fs.spyr.), shallow but sharp, divides the median lobe into a large suprapyramidal region (p.spyr.) and a slender curved ventral convolution, the pyramis (pyr.), concentric with the uvula, which fills the ventral concavity of the fissura secunda. The other lobule of the posterior lobe, the nodulus (nod.), lies entirely hidden on the ventral surface of the cerebellum, being separated from the uvula by the concealed fissura postnodularis (fs. pnod.}. The cerebellar ventricle (v.cb.) is seen between the nodulus and the lingula. The large mushroom-shaped pedunculate lateral masses projecting beyond the lateral lobes (these indicated only by a very slight partial constriction) are the paraflocculi (pfloc.), the dorsal components of the floccular lobes. The ventral components are the tiny flaplike flocculi, concealed from view, being plastered against the medulla beneath the paraflocculi, from which they are separated by the fissura floccularis (fs.floc.). The fissura parafloccularis (fs. pfloc.), visible laterally, separates the paraflocculus from the median lobe. It cuts down to the floccular peduncle between these structures. All the cerebellum save the floccular lobe and the lateral lobes (area pteroidea, a.pt., in part) corresponds probably to the vermis of higher cerebella. The lateral lobes similarly are probably homologous with the cerebellar hemispheres. The deep nuclei (deep nuc.) form a pair of large oval masses almost meeting in the midline, with a large area of their ventral surface ex- posed in the roof of the fourth ventricle. They exhibit the mammalian characteristic of complete separation from Deiters' nucleus, as van Hoevell (1916) found in some other marsupials. There are only slight indications of differentiation into separate nuclei (dentate and roof, nuclei). Since the cerebellum of Caenolestes and Orolestes will form part of the next report on these brains, it need not be further discussed at this time. The intra- and extracerebellar relations will be partially clarified by the diagrammatic reconstructions, figures 7-10, together with the sections, figures 41-44. The nomenclature follows the usage of Elliot Smith. i88 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. Measurements. — The dimensions of these three brains as meas- ured on the alcoholic specimens are as follows: C.obs.* O.inca 18507 194948 194921 F.M. U.S.N.M. U.S.N.M. 1. Total length, tip of olfactory bulb to first spinal nerve 14.1 mm. 16.5 mm. 19.1 mm. 2. Length, tip of olfactory bulb to end of cerebral hemisphere 2.6 3.0 3.6 3. Length of cerebral hemisphere 10.0 9.3 8.8 4. Length of cerebellum on longitudinal axis of brain in median plane 3.0 5.0 6.2 5. Greatest width of olfactory bulbs 7.6 7.8 7.9 6. Greatest width of both cerebral hemispheres n.8 n.8 12.6 7. Total width of cerebellum and floccular lobes n.o 11.7 — 8. Width of cerebellum exclusive of floccular lobes 8.8 i i.o i i.o 9. Maximum vertical height of cerebral hemi- spheres 7-9 8.9 10. Distance of orbital fissure behind rostral tip of hemisphere (2.0) * 2.3* 2.4 11. Distance of rhinal fissure below vertex at orbital fissure 2.7* 4.2 12. Length, head and body 107.0 (89.0) ( 102.0) 13. Length, tail 118.0 115.0 116.0 14. Length, total 225.0 204.0 218.0 15. Weight of brain, immediately out of 80% alcohol 760.9 mg. 961.0 mg. * C. obscurus, measurements i-n, from Herrick (1921) ; 12-14, from Dr. Osgood. * Position of orbital fissure in C. obscurus is "about one-fifth of the distance backward from the frontal to the posterior pole of the hemisphere". * Orbital fissure not very clear in O. inca, No. 194948. O. inca No. 194921 had probably all of cervical cord attached, but lost parafloc- cular lobes on removal from the skull. O. inca No. 194948 was broken off just behind cerebellum, in removing it from the skull. INTERNAL ANATOMY The twenty-two simplified cross sections from the hemisphere of Caenolestes obscurus (Figs. 23-44) and the linear reconstructions of the hemisphere or of parts of it (Figs. 12-15, J7a and b) may perhaps, with the aid of the figures and descriptions of the external surfaces, render a brief and incomplete account of the internal anatomy com- prehensible. The following table of critical levels and their section numbers may also assist in the orientation of internal structures. Sec. i — Anterior tip of the olfactory bulb. Sec. 225 — Anterior tip of the frontal lobe of the neopallium (Fig. 23). Sec. 400 — Caudal (medio-ventral) limit of the olfactory bulb, adjoin- ing the tuberculum olfactorium (Fig. 25). Sec. 710— Caudal limit of the tuberculum, adjoining the amygdaloid complex (Fig. 36). Sec. 935 — Posterior tip of the cerebral hemisphere (Fig. 44). 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 189 PRIMARY OLFACTORY AREA The description of the histological pattern of the bulbar formation, both ordinary and accessory, may be dispensed with here, in view of its similarity to that of the Virginia opossum, already described in detail by McCotter (1912). Contrary to the condition in the opossum, however, the horizontal diameter of the olfactory bulb in Caenolestes is greater than the vertical diameter. The usual medial displacement of the olfactory ventricle is emphasized here by the considerable dorso- ventral flattening of the olfactory bulb. This displacement may per- haps be a condition of interest in connection with the disposition of the olfactory tract fibers. The well developed accessory olfactory bulb (b.ol.ac., Figs. 12, 14, 23-24) is embedded in the postero-dorsal olfactory formation beneath the overhanging frontal pole of the neopallium. It is outlined by a slight fissure. Its peripheral nerve, the vomeronasal nerve (n.vn., Figs. 12, 14, 23-24) is quite clear as it curves up over the medio-dorsal angle of the olfactory bulb and spreads out to cover the surface of the accessory bulb. The secondary tract of the accessory bulb passes lat- erally and superficially and helps to make up the pars dorsalis of the lateral olfactory tract (tr.ol.l.d., Figs. 23 ff.) The olfactory tracts, which, as Cajal pointed out (1911), arise indiscriminately from all parts of the olfactory formation and are without specificity (save for the fibers from the accessory bulb), first condense in the center of the bulb some distance in front of the olfactory ventricle (tr.ol., Figs. 12-15). Since they so quickly begin to distribute to the secondary olfactory areas and even in part (intermediate olfactory tract) to re- ceive fibers from them, their description will be continued below Under that head. SECONDARY OLFACTORY AREAS The olfactory fibers spin a whorl around the olfactory ventricle, much thicker on the lateral side than elsewhere. A second independent half-whorl forms within the first, between its thick lateral portion and the ventricle. This mass of fibers, composed of secondary (direct) and tertiary olfactory fibers, soon rounds up in the lateral wall of the ventricle as the intermediate olfactory tract (tr.ol.i., Figs. 23-28), largely forming the rostral limb of the anterior commissure (c.a., Figs. 24-28). Meanwhile, between it and the lateral part of the first whorl (the massive portion of the lateral olfactory tract, tr.ol.i.) appear the most anterior cells of the peduncular grey or anterior olfactory nucleus, pars lateralis (nuc.ol.ant.L, Figs. 12-15, 17, 23-25). This growing 190 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. cell mass rapidly increases the widening interval between the lateral part of the big whorl externally and the half -whorl internally, and immediately begins to contribute tertiary olfactory fibers to the inner half-whorl, the intermediate or commissural olfactory tract, which also continues to receive secondary olfactory fibers from all parts of the olfactory formation. Still farther back the tuberculum olfactorium (t.ol., Figs. 13, 28) cuts the outer whorl into two parts, the lateral olfactory tract, much the largest of all the olfactory tracts, and the slender medial olfactory tract (tr.ol.m., Fig. 28). The intermediate tract, which retains its subependymal position back to the point where the head of the caudate nucleus (nuc. caud., Fig. 28) overlies it, soon con- tains perhaps more tertiary than secondary fibers, and is therefore labelled only anterior commissure (c.a.) from figure 24 back. The lateral olfactory tract, the most widely distributed of the three, becomes superficial (tr.ol.l., Figs. 3, 12, 25) behind the fissura circu- laris, forming the medullated external fiber layer of the lateral sur- face of the hemisphere over most of its extent below the rhinal fissure. It may be divided into several parts : anteriorly, two peduncles, a sub- ventricular ventral peduncle (tr.ol.l.p.v., Figs. 24-27) a part of which caudally becomes independent and is then called the medial olfactory tract {tr.ol.m., Fig. 28) as noted above, and a supraventricular dorsal peduncle (tr.oLl.p.d., Figs. 23-24) with two roots, a ventral root, below the accessory olfactory bulb, and a dorsal superficial root which is con- tinued caudally as the pars dorsalis of the lateral olfactory tract (tr.ol.- l.d., Figs 25 ff.) ; a clubshaped massive portion (tr.ol.l., Figs. 23-35) filling the interval between the endorhinal and the rhinal arcuate fis- sures, wide anteriorly but decreasing to the vanishing point as these fissures meet (fs.erh., fs.rh.arc., Figs. 13, 33) ; posteriorly a pars ventralis (tr.ol.l.v., Figs. 29 ff.), distributing to the tuberculum, nucleus of the lateral olfactory tract (nuc. tr.ol.l., Figs. 12, 36-38) and other amygdaloid nuclei, and perhaps also to the diagonal band nucleus (nuc.d.b., Figs. 32-33). The small medial olfactory tract (tr.ol.m., Fig. 28) can be seen turning sharply downward into the plexiform layer of the tuberculum at its medial border, where some observers (e.g., Livini, 1908) recognize a terminal nucleus of the median olfactory tract in a large median rolled portion of the tubercular formation. Such a structure is present here (nuc.ol.m., Fig. 31). Beccari (1910) inclines to doubt the distribution of secondary olfactory fibers to any extent to the tuberculum, but such evidence as these sections give, while not at all conclusive, seems to point to it. 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 191 Anterior olfactory nucleus. — The less differentiated postbulbar gray matter of the rhinencephalon has been termed by Herrick (1910, pp. 191-2, Figs. 9 and 10) the anterior olfactory nucleus. In Caeno- lestes it is almost entirely intrabulbar and in direct continuity with the rest of the more ventral gray subjacent to the lateral olfactory tract. Rostrally therefore it completely surrounds the olfactory ventricle. It corresponds essentially to the peduncular gray of Cajal, and is not included in his pyriform lobe. It is here excluded from the pyriform lobe, though earlier, following Johnston's definition of the pyriform lobe as all the gray underlying the lateral olfactory tract, I counted it a part of the pyriform lobe (Obenchain, 1923). Since its lateral border is farthest advanced rostrally, this peri ventricular rhinencephalic ring appears incomplete in the more rostral sections (Figs. 23-26), and since the pyriform cortex is also advanced rostrally beyond some por- tions of the anterior olfactory nucleus, the whole of the latter is never seen in any transverse section. The anterior olfactory nucleus, on both topographical and histological grounds, may be subdivided into several parts, all in cellular continuity rostrally. I have, as far as seemed advisable, conformed to the terminology applied to the different parts of the anterior olfactory nucleus in the Virginia opossum (Herrick, 19243). Divergences will be noted as they occur. The lateral part (nuc.ol.ant.L, Figs. 12-15, I7> 23~25)> rostrally most advanced, passes caudally without interruption into true pyriform cortex, and therefore has no really definite posterior limit. The pars lateralis is extended medially above the olfactory ventricle as the pars dorsalis (nuc.ol.ant.d., Figs. 12-15, I7> 25~26) ; it corresponds to Cajal's superior peduncular nucleus. Just in front of the antero- median extension of the rhinal fissure (fs.rh.a.) it comes to the sur- face, as noted above, and just behind it fuses with the cell mass of the overlying neopallial frontal pole. More caudally and below the ventricle the anterior olfactory nucleus is also prolonged medially by another hooklike extension, the pars later o-ventralis (nuc.ol.ant.l.v., Figs. 12, 15, 17, 26), and beyond that the pars posterior (nuc.ol.ant.p., Figs. 12, 15, 17, 27-29), which fills the wedgelike interval in front of the fusion of the head of the caudate nucleus and the tuberculum olf actor ium. The pars posterior extends further back than other portion of the anterior olfactory nucleus, 'but it is neither in continuity with the caudate nucleus — forming the rostral portion of its head, as in the turtle, Cistudo Carolina (Johnston, 1915) — nor with the deeper layer of the tuber- culum, as in the alligator (Crosby, 1917). It stops bluntly just ros- tral to the fusion of the caudate head and the tuberculum. This part 192 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. of the anterior olfactory nucleus never shows any appreciable degree of differentiation, probably owing to the too purely olfactory character of its connections. The latero-ventral and posterior parts of the an- terior olfactory nucleus as here described do not exactly coincide with the divisions in the Virginia opossum. More caudally the dorsal and posterior parts meet medially, thus completing the anterior periventric- ular rhinencephalic ring by means of a pars medialis (nuc.ol.ant.m., Figs. 12, 15, 17, 26). At the dorso-medio-rostral border of this ring there is a small superficial condensation of cells. This is the anterior tip of the hippocampal formation (hip. a., Figs. 12, 17; ex. hip. a., Fig. 26). It corresponds to a similar dorso-lateral pyriform condensation (cx.pir.a., Fig. 25), which, although it appears in more rostral sections, does not quite reach the rostral border of the ring. The lateral and medial olfactory cortices (pyriform and hippocampal) are thus only indirectly continuous rostrally by a double bond (supra- and infra ventricular) through the agency of the anterior olfactory nucleus, while, as we have seen, they are directly continuous caudally. There are thus two peri- ventricular rhinencephalic rings, a smaller transverse and a larger hori- zontal one, which are partially fused anteriorly. The medial part of the anterior olfactory nucleus soon passes ob- scurely backward into the lateral parolfactory nucleus (nuc.pol.L, at a level between figures 26 and 27). The latero-ventral part, which is transitional between the lateral and posterior parts, continues back- ward into the ventral pyriform cortex underlying the massive portion of the lateral olfactory tract along the lateral border of the tuberculum (pars ventralis, lobus piriformis, Gray, 1924; ex.pir.v., Figs. 27-33), and still more caudally perhaps into the diffuse region of the anterior perforated space (Johnston, 1923) (l.perf.a., Figs. 35-38), along the anterior portion of the amygdaloid fissure (fs.atmg., Figs. 12, 15, 36-38). One other — and the most interesting — part of the anterior olfactory nucleus remains to be described, the pars externa (nuc.ol.ant. ex., Figs. 13, 15, 24-26). It is apparently the result of the doubling of the in- trabulbar portion of the lateral part of the anterior olfactory nucleus, which is the only portion to exhibit this phenomenon. The external nucleus appears rostrally as a vertical plate of smaller cells between the pars lateralis and the massive lateral olfactory tract. It is separated from the subjacent pars lateralis by a wide plexiform layer narrowing ventrally to suggest cellular continuity between the two nuclei. Ros- trally a very narrow but definite external plexiform layer separates it 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 193 from the fibers of the lateral olfactory tract. Caudally the external nucleus, or external part of the lateral part of the anterior olfactory nucleus (to give it its full designation) slides downward along the internal surface of the lateral olfactory tract, gradually diminishing and rounding up into a small mass of cells which, as it slips medial ward around the inside of the "elbow" of the lateral tract, becomes counter- sunk in a space due to the clearing away of some of the blue tract fibers. It is here opposite to the pars latero-ventralis, but its main bulk lies opposite the pars lateralis — there is no definite boundary be- tween these two parts. The general form of the external nucleus is that of a long and slender pennant (Fig. 13) whose caudo-ventrally extended tail curves far medialward to end almost beneath the olfac- tory ventricle (Fig. 15). The external nucleus of Caenolestes is thus entirely intrabulbar, so that it faces the granular olfactory formation across the olfactory tract. Dr. Herrick finds that the Virginia opos- sum also possesses an essentially similar nucleus, save that it lacks an external plexiform layer, and is caudally unevenly swallow-tailed. In Caenolestes a small dorsal group of cells tends to be separated from the rest by the fibers of the ventral root of the dorsal peduncle of the lateral olfactory tract. Rothig (1910, Fig. i) figures but does not name a small vertical cell plate in Didelphis marsupialis, which is plastered against the inner aspect of the dorsal thin part of the lateral olfactory tract behind the bulbar formation, a' dorsal shift which would bring it nearer to the remaining bulbar formation at this level. The entire width of the plexiform layer separates it from the underlying pyriform cortex. As this is the most anterior section figured by Rothig, the more rostral extent and relations of this cell plate are unknown to me. Livini (1908, Fig. 2) shows in Hypsiprymnus rufescens a doub- ling of the intrabulbar portion of his anterior pyriform lobe (which cor- responds to the lateral part of the anterior olfactory nucleus as de- scribed here) into two nearly equal vertical plates of cells only narrowly separated, the outer of which seems to be composed of somewhat smaller cells. In a Nissl series of a white rat brain I find the anterior olfac- tory nucleus also doubling laterally in the olfactory peduncle, in the gap between the dorsal and ventral edges of the bulbar formation, which in cross section has just broken in two — in other words, just behind the fissura circularis. The external portion is a more condensed ver- tical plate of cells, immediately beneath the massive lateral olfactory tract, bridging the gap in the olfactory formation and therefore in contact with it at its dorsal and ventral margins. More caudally, with the widening of the gap between the receding edges of the olfactory 194 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. formation, the external nucleus separates into two parts, dorsal and ventral, which diminish as they separate more and more, and finally end as tiny cell masses at the upper and lower borders of the bulbar formation near its caudal limits. Its general form is that of, an evenly swallow-tailed pennant. The Winkler-Potter rabbit and cat atlases (1911, 1914) give no figures showing such a nucleus. Cajal (1911) neither mentions nor figures it, a circumstance which is perhaps to be ascribed to its failure to impregnate in his rich collection of Golgi series of the brain of the mouse. His only figure showing the lateral peduncular gray (1911, II, Fig. 431, p. 696, mouse, 5 days old, after Calleja) is apparently a horizontal section, which may miss the level of the external nucleus, if there be such a nucleus in the mouse. I have so far found no reference in the literature to this curious cell mass beyond the Rothig and Livini figures cited above, in which it was unlabelled. The interpretation of this nucleus offers apparently no great diffi- culty. It is, probably, some sort of reenforcing or stepping-up device for olfactory stimuli, an accessory olfactory nucleus, discharging its afferent fibers into the subjacent lateral part of the anterior olfactory nucleus. It is a regulatory response probably provoked in part by the antero-posterior compression of this highly macrosmatic type of brain at the point where the neurobiotactic attraction of accumulated secondary fibers is strongest. Does this nucleus represent the morpho- logical anterior end of the anterior olfactory nucleus, detached and car- ried back by the more anterior collaterals of the lateral olfactory tract, either actually or by being held fast by them during the progress of the compression which telescopes bulb and anterior end of the extrabulbar rhinencephalon ? Or is it delaminated in situ from the subjacent cell mass, being composed of the cells which possess no basilar dendrites and are therefore unable to resist the unopposed neurobiotactic influence of the overlying fibers? It is not inconceivable that both delamination and dislocation may have operated in its formation. Although Cajal neither pictures nor describes the external nucleus, and gives but one copied figure of the lateral peduncular formation in connection with which it arises, his figure and description of the region immediately behind its due position furnish apparently unmistakable clues, in the light of Kappers' concept of neurobiotaxis, of the mode and causes of its formation further forward (1911, II., Fig. 433, p. 680, rabbit, aged 25 days). In the cortex of his "frontal lobe" (anterior area of pyri- form cortex, Gray, 1924) the third layer (outer cell layer, superficial polymorphs or medium pyramids) contains cells without basilar den- 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 195 drites, all their dendrites being directed outward toward the fibrillar layer (lateral olfactory tract). These cells form, according to Cajal, a rather precise wavy band, not however separated off from the cells below, and they are extremely variable in form. The deeper pyramidal cells and the still deeper polymorphic cells are provided with both ascending and descending dendrites, and are therefore doubly anchored and not subject to extreme outward displacement by unbalanced olfac- tory stimuli of the lateral olfactory tract, as are the more superficial cells without descending dendrites. The balancing attraction resides, probably, in the heterolateral secondary and perhaps in the tertiary fibers in the rostral limb of the anterior commissure, which lies beneath the lateral portion of the anterior olfactory nucleus. Whether any other fibers reach it from this direction, as, for example, homolateral second- ary fibers of the intermediate olfactory tract, I do not know. At any rate this external part of the anterior olfactory nucleus is slung like a hammock between two opposing neurobiotactic forces, and the doubling or splitting of the nucleus in this region expresses the resolution of the situation. The deep and main portion of the nucleus in this region lies in Caenolestes closer to the anterior commissure than to the lateral olfactory tract. Cajal (1911, II., p. 678) says that the cortex of the olfactory peduncle and of the "frontal lobe" (which lies next behind it) are essentially the same in structure. We should indeed expect to find the lateral peduncular gray less differentiated, with perhaps shorter axons and fewer descending dendrites, and with only a slight tendency toward cortical lamination. If the situation is as sketched we have a clear and exquisite illustration of the two-fold activity of neurobiotaxis at work. Since the lateral olfactory tract is always in macrosmatic ani- mals an exceedingly heavy mass of fibers it is not improbable that the external olfactory nucleus is well developed in at least all those forms whose olfactory peduncle (anterior olfactory nucleus) is jammed for- ward and largely enclosed within the bulbar formation, as in the forms mentioned here. We should expect to find vestigial traces of it wide- spread among mammals in general, and it has most likely already found its way into the literature in some form or other. Tuberculum olfactorium. — The enormous tuberculum olfactorium begins in these sections latero-ventrally (t.ol. (i.C.), Figs. 2-5, 12, 13, 15, 28-36), instead of medio-ventrally as in the Virginia opossum. It rapidly expands medialward to occupy the entire width of the base of the brain, and caudally to a point beyond the middle of the hemisphere (section 710 in a hemisphere numbering 935 sections). Antero-medially it turns up on the medial surface for a short distance, where it is de- 196 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. limited by the medial portion of the fissura rhinalis arcuta (fs.rh.arc.}, which does not as here described quite agree with Beccari's account ( 1910, Fig. 18) ; he extends the anterior and posterior parts of the encircling fissure of the tuberculum upward in the medial wall to meet at the ventral border of the hippocampal formation ; but I have thought it simpler (and more in keeping with corresponding conditions of the lateral wall) to carry it across the base of Beccari's median triangle, in a fissure which is present and which delimits the cellular formation characteristic of this region, just as the arcuate fissure does laterally. In the more rostral sections it will be seen that the median part of the rhinal arcuate fissure is much shallower than the very sharp one lying immediately below it within the tuberculum. The two fissures define a rather prominent rolled portion of the tuberculum which is probably to be identified with the nucleus of the medial olfactory tract of Livini (1908) and others. Above the rhinal arcuate fissure the characteristic formation of the tuberculum falls away from the surface and runs up along the median border of the nucleus accumbens (nuc. ac., Fig. 30), almost if not quite to the ventricular ependyma beneath the more ros- tral portion of the body of the anterior commissure, where the latter breaks across the ventricle to reach the, septum. It thus intervenes be- tween the sharply defined nucleus accumbens and the ventral portion of the precommissural body or septal formation. Its own medial boundary is also sharply defined from the septal formation. The tuberculum ap- parently receives secondary olfactory fibers from both the lateral and medial olfactory tracts (Fig. 28), whose fibers may be seen bending down into the plexiform layer of the tuberculum on its lateral and medial borders. Beccari (1910) questions this (see page 190 above), finding evidence of other sources of origin (the pyriform lobe in par- ticular) for the fibers of the external plexiform layer of the tuberculum. He thinks that the olfactory tract fibers, if present, exist only rostrally in this layer. It looks otherwise here, but this is not really decisive material. In front of the strio-tubercular fusion (see Figs. 29-30), which takes place behind the caudal end of the posterior part of the anterior olfactory nucleus (see page 191) essentially in the fashion so clearly described by Livini ( 1908) in Hypsiprymnus, a rather wide deep plexi- form layer, continuous with the superficial encircling plexiform of the section, intervenes between the tuberculum and the rest of the cellular formation of the section. This layer is crowded with a wealth of fibers of diverse origin and destination, whose adequate analysis is impossible in this, series. They include fibers from the lateral olfactory tract, an- 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 197 terior olfactory nucleus, intermediate olfactory tract (probably both crossed and uncrossed), from the tuberculum, which are destined for the septal nuclei, hippocampus, frontal neopallial pole. Probably all of these systems are to some extent doubly oriented. And some of them also apparently contribute to the median forbrain bundle, fasciculus medialis telencephali (f.med.t., Figs. 29-35), which begins collecting in the base of the septum quite far forward ; these fibers are projection fibers to stem centers and probably from them also. It is the mul- titudinous ascending precommissural fibers which are mainly responsi- ble for the differentiation of the rostral portion of the hippocampus in lower mammals. The histological development of the tuberculum is in Caenolestes spectacular in the highest degree. The external cell layer of medium darkstaining pyramidal or polymorphic cells is thrown into battlement- like folds, interrupted irregularly by islands of Calleja (i. C., Figs. 29- 32), composed of extremely small, pale, round cells densely crowded, glomerulus-like roundish or vermiculate areas, and including sometimes a few pale giant cells. These masses vary greatly in size and shape, and they are so sharply delimited and so different from their surroundings as to suggest pathological growths. They probably correspond to Beccari's (1910) Type 2 islands, while deeper ones of the same general character belong to his Type 3 islands. The Type I islands consist of thickenings of the crenulations of the external cell layer ; these are some- times fringed with pale granules like those of the other islands. While these types are sharply differentiated, there are, as Becarri found, intermediate types. Small isolated cell masses in the external plexiform layer of the more caudal sections especially are all traceable into the main mass of the tuberculum. None of Beccari's figures show so great a histological complexity as Caenolestes. The conditions in the Virginia opossum are much less complex (Gray, 1924). What is the function of these highly elaborated and integrated organs icithin an organ? They suggest some sort of elaborate rehandling and sorting of incoming stimuli — a physiological analysis by means of differ- ential thresholds, effecting a secondary specificity from mass stimuli dis- sociated, reenforced and more or less independently projected? A searching study of the tuberculum at its most bizarre stage of develop- ment in the lower mammals ought surely to discover valuable clues to modes of nervous organization. The deep layers of the tubercular "cortex" are best described in con- nection with the immense fiber and giant cell stream which, like a great diagonally slung hammock, extends from the ventral pyriform cortex 198 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. and lateral anterior commissure region across the base of the hemi- sphere to the ventro-medial "regio innominata", the vestibule to the thalamus and lower stem centers. This great and complex system, composed of an immense number and variety of fibers, both pro- jection and associational, is certainly one of the striking features of the middle subventricular region of the hemisphere. It is generally called the basal olfactory bundle (of Wallenberg) in these lower mammalian brains. It forms a part of the practically unanalyzable mass of fibers (in this series) traversing the basal region of the hemisphere, and in- cluding such systems as the olfactory projection of Cajal, striatal systems (ansa lenticularis), etc., the whole mass, save the association fibers, drifting ventro-medially to join the medial forebrain bundle and continue spinalward with it. Therefore, following some writers, I have called it the lateral limb of the medial forebrain bundle (fasci- culus medialis telencephali, pars lateralis, f.med.t.1., Figs. 29-33). It is largely, in all probability, a doubly oriented system. Its more ros- tral portion is characterised by the presence of a great multitude of pale giant cells strewn among the fibers (nucleus of the basal olfactory bundle, Figs. 29-33). A similar condition occurs in the more median fiber tangle, whose giant cells correspond probably to the "border nucleus" of Volsch (1906). The ansa lenticularis component contains few cells. But for many of its fibers the cells of origin are the giant cells characteristic of the globus pallidus (gl.p., Figs. 12, 15; glob, p., Fig. 34), and these are just like those strewn so thickly among the fibers of the more rostral stream of the medial forebrain bundle, especially in its lateral limb and in the more medial portion, where they seem to correspond to the "border nucleus" of Volsch (1906). The globus pallidus cells are certainly motor projection cells, and very likely the other giant cells mentioned also send long axons to stem centers. The fact that the red nucleus is also mainly composed of pale giant cells of the same type further tends to support this supposition. With the reduction of the lateral and basal olfactory centers in higher mammalian brains, the more posterior ansa lenticularis com- plex becomes so preponderant as to throw the more anterior olfactory complex into the shade, and so the name "ansa lenticularis" comes to be applied to the whole stream. But in lower mammalian brains it is the more anterior component, mediating correlated olfactory stimuli, which seems more conspicuous. Many longitudinal fibers and fiber bundles (association tracts between rostral and caudal regions of the base of the hemisphere) further complicate the situation. 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 199 Septal region. — The precommissural area of the median surface of the hemisphere is the external surface of the thick paraterminal body of Elliot Smith or septum of ordinary terminology. This region con- tains the ascending olfactory systems already enumerated (olfacto- septal, olfacto-cortical) and the descending precommissural fornix, septo-amygadaloid (Johnston's stria terminalis bundle 4), septo-habe- nular (of stria medullaris system) and median forebrain bundles. The bundle labelled olfacto-f rental (tr.ol.fr., Fig. 27), extending from the frontal pole of the neopallium into the septum, where it mingles with the septal fibers which sweep laterally beneath the ventricle or below the lateral mass of the anterior commissure, cannot itself be disen- tangled and followed with precision in these sections. Arising in the frontal pole of the neopallium just in front of the laterally directed corona radiata fibers destined for the internal and external capsules, these fibers inevitably suggest the possibility of part of the coronal fibers being diverted, or rather persisting, medially into the septum as projection fibers to stem centers by way of the median forebrain bundle, as a vestige of what Edinger has called the septomesencephalic tract of submammalia. In this connection Pedro Cajal's findings (1917, 1919) in Varanus and Lacerta, of a septal passage for descending fibers from the entire cortex, including the depressed portion between the medio-dorsal hippocampal cortex and the lateral pyriform cortex, which he considers the "general cortex", seem highly significant. One is led to recall also that the only cortical projection tract in birds is a septal one, and that it does not proceed from what seems to be olfactory cortex. It is not inconceivable that in the neopallial frontal pole of the lowly mammalian type of brain, at a level where the hippocampal formation is still rather insignificant, the more medial neopallial fibers might have preserved the shorter septal path. Two well developed nuclei lie in the septum — the lateral and medial parolfactory nuclei (Herrick). The lateral parolfactory nucleus (nuc.pol.L, Figs. 28-30) occupies the septal wall lateral to the pre- commissural fornix fibers. It corresponds to Johnston's (1913) pri- mordium hippocampi and not to his lateral parolfactory nucleus, which is here the nucleus accumbens (nuc.ac.}, the medial portion of the head of the caudate nucleus. The medial parolfactory nucleus (nuc. pol.m., Figs. 30-31), located medially and ventrally, corresponds to Johnston's riucleus of the same name, and, as he finds in the Virginia opossum and in other forms (Johnston, 1923), it passes back insen- sibly into the nucleus of the diagonal band of Broca (nuc.d.b., Figs. 2OO FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. 32-33). Apparently it grades also into the preoptic nucleus of the telencephalon medium (nuc, prop., Fig. 34). The dorsal (d.) and anterior (c.a.) commissures have already been briefly described in connection with the median section of the brain. The dorsal commissure will be considered also below in connection with the hippocampus. The composition of the anterior commissure offers in Caenolestes nothing of unusual interest. But in view of the disagreement among zoologists with regard to the assignment of Caen- olestes to one or the other of the two marsupial subgroups — Diproto- dontia and Polyprotodontia — with its bearing on marsupial distribu- tional problems, the arrangement of the fibers of the anterior commis- sure was, since it seemed to be the only remaining anatomical evidence to be expected on the question, a matter of great interest (Obenchain, 19233, I923b). After an examination of the brains of every mar- supial genus except Caenolestes (which was not available) Elliot Smith (i9O2a and b) found that all diprotodont brains exhibit one feature never found in any polyprotodont brain, or indeed in any other ver- tebrate brain. This exclusive diprotodont character he named the aberrant bundle (fasciculus aberrant) of the anterior commissure, considering it a true diagnostic character of diprotodont brains. It is merely the dorsal portion of the anterior commissure which in dipro- todonts splits off from the rest of the commissure to pass upward by way of the internal instead of the external capsule, the common route in all other brains. The aberrant bundle is absent in Caenolestes, a fact which, if this feature be decisive, would ally it with the polypro- todonts. The brains of the fossil caenolestids, however, can never be known, but the presumption of the absence of the aberrant bundle in them also would perhaps be justifiable. The effect of this would occa- sion no further disturbance of Dr. Osgood's marsupial "family tree" than the lengthening of the polyprotodont bracket to include Caeno- lestes, leaving it still in place between the generalised polyprotodont Perameles and the diprotodonts of Australia; or at most, in view of the intermediate position of the exposed gyrus dentatus of Caenolestes, the latter might be shifted to a position between Perameles and Notoryctes. Pyriform Lobe*. — The pryiform lobe consists of two distinct parts : (i) the pyriform cortex, mostly confined to lateral wall of the hemi- *The limits here assigned to the pyriform lobe do not exactly coincide with those assigned either by Cajal or Johnston, although the actual descriptions of the areas involved vary little or not at all. Cajal (1911) restricts the pyriform lobe to include, besides the amygdala, only the median pyriform cortex 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 201 sphere, between the rhinal fissure above and the rhinal arcuate and amygdaloid fissures below; and (2) the amygdaloid complex, mostly in the ventral wall between the amygdaloid and choroid fissures. The amygdaloid complex, thanks to Johnston's illuminating com- parative analysis (1923), which builds upon and completes Volsch's intensive study (1906), is no longer the mysterious territory it has been. Since it is, strictly speaking, a subcortical center, it will be described first here, leaving the pyriform cortex to precede the hippo- campus. Owing to its enormous extent, the amygdaloid complex is perhaps the most spectacular part of the brain — unless it share this distinction with the tuberculum olfactorium. The superficial extent of the amygdaloid complex has already been described. Internally it exceeds this both in length and width, overlapping the posterior por- tion of the tuberculum anteriorly and the pyriform cortex laterally, and, as Johnston (1923) points out in the Virginia opossum, stretch- ing rostrally far towards the anterior commissure. Behind the caudate nucleus and the putamen (put.) it floors the ventricle, while the cellu- lar bed of its great fiber system, the stria terminalis (st.t., Figs. 32 ff.) forms the median strip of the floor behind the anterior commissure. I am not, however, able to follow the stria bed in these sections into the anterior olfactory nucleus, as Johnston does in the opossum. It is clear anteriorly here only as it lies upon the anterior commissure and more caudally upon the internal capsule or cerebral peduncle. The nuclei of the amygdaloid complex in Caenolestes comprise the six described by Johnston (1923), and include also the extra seventh one he found in the Virginia opossum : the nucleus of the lateral olfac- tory tract, the central, medial, lateral, basal, accessory basal, and corti- cal amygdaloid nuclei. (his anterior pyriform cortex), receiving terminals of the lateral olfactory tract, and the posterior pyriform cortex, receiving no direct olfactory fibers (his superior temporal cortex). This makes it coincide approximately with the gyrus hippocampi of primates. Johnston (1915) defines the pyriform lobe as the gray matter underlying the lateral olfactory tract, which would include the entire anterior olfactory nucleus, but not the posterior pyriform cortex. By this definition the pyriform lobe and hippocampus are directly continuous anteriorly, but not in contact posteriorly, since the posterior pyriform cortex would be included in the neopallium. Since this area is the field of origin of the temporo-ammonic tract (Cajal), the main afferent tract to the hippocampus in mammals, it seems more logical to include it within the pyriform lobe. Its exclusion by Johnston would explain the complete caudal separation of the pyriform lobe and hippocampus in the turtle, Cistudo Carolina (1915). The anterior pyriform cortex here coincides with Cajal's "frontal lobe", which he excludes from the pyriform lobe, on the basis that it, like the peduncular gray (anterior olfactory nucleus) receives mainly collaterals of the lateral ol- factory tract. 202 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. The well delimited nucleus of the lateral olfactory tract (nuc.tr. ol.L, Figs. 12, 15, 36-38) is a duplex structure, consisting of a large-celled dorsal part and a small-celled ventral part, separated by a narrow plexiform layer. It is tilted up rostrally, and shifted forward for about half its length above the caudal portion of the tuberculum, and there- fore only its posterior half lies beneath the medio-ventral tubercle which in Caenolestes is apparently not so pronounced as in Orolestes, of which no sections are now available for comparison. The nucleus of the diagonal band (nuc.d.b., Figs. 32, 33) curves around from the median wall here, and the diagonal band fibers pass laterally in the diffuse region above the posterior end of the nucleus of the lateral tract towards the pyriform cortex, probably also effecting connections with the intermediate region. The ventral small-celled portion of the nucleus is so sharply delimited as to appear almost encapsulated, an appearance heightened at its caudal pole, which extends beyond the dorsal part as a sort of island of cells in the plexiform layer of the hemisphere. The corresponding n'ucleus of the Virginia opossum ex- hibits neifher the forward shifting above the tuberculum nor the histological differentiation seen in Caenolestes (Gray, 1924). The nucleus of the lateral olfactory tract receives secondary olfactory fibers from the pars ventralis of the lateral olfactory tract (unlabelled, Figs. 36-38), and dorsally a great fan of fibers from the stria terminalis (st.t.i, Figs. 33-36), its most anterior contingent. Johnston (1923) identifies this with the commissural bundle of the stria, but these sec- tions do not actually permit this. The central nucleus of the amygdala (c., Figs. 12, 15; nuc.amg.c., Figs. 35-37) lies above the nucleus of the lateral olfactory tract. It is confluent with the strongly developed "intercalary plate" (int. plate, Fig. 36) of Johnston (1923), which is the most hypertrophied part of the stria bed. The central n'ucleus advances farther forward than any other amygdaloid nucleus, but apparently not so far as Johnston found in the Virginia opossum, up to the region of the anterior com- missure. Its limits are not well defined nor its cellular structure strik- ing. The medial amygdaloid nucleus («*., Figs. 12, 15; nuc.amg.m., Figs. 36, 37) occupies the ventro-medial angle of the hemisphere, and also lacks well defined limits or strikingly marked cell structure. It is impossible to fix its ca'udal limit here. It also is continuous with the intercalary plate and with the central nucleus, as Johnston found. The lateral amygdaloid nucleus (I., Figs. 12, 15; nuc.amg.1., Figs. 35-39) is the "poststriatum" of earlier writers. It is a large and ex- 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 203 tremely well defined antero-posteriorly elongated nucleus, oval in sec- tion and characterized by good-sized pyramidal or polymorphic cells. It lies in the concavity of the external capsule (cap.e), between it and the ventral part of the putamen (put.), and behind the latter it rises to form the more lateral part of the ventricular floor. It clearly receives external capsule fibers, more probably originating in the pyriform cor- tex, and it seems also to receive a large number of stria terminalis fibers, coursing horizontally just above the basal amygdaloid nucleus (st.t. 3, Fig. 38). This last is contrary to Johnston's observations. The large-celled well defined basal nucleus of the amygdala (&., Figs. 12, 15; nuc.amg.b., Figs. 37-42) arises medial to the more caudal portion of the lateral nucleus, near its ventral border, and increases to large proportions as the lateral nucleus diminishes. Behind the flat- tened tail of the caudate nucleus it occupies a large part of the floor of the ventricle, extending far medialward and lying caudally directly beneath the ependyma. Both it and the lateral nucleus are in cellu- lar continuity with the deep cells of the ventral border of the pyri- form cortex at the level of the amygdaloid fissure. Johnston (1923) considers them to have been derived from the pyriform cortex by a process of infolding along this line, and the situation in Caenolestes seems to support this view. The accessory basal nucleus of the amygdala (b.ac., Fig. 15; nuc. amg.b.ac., Figs. 37-38) is a less clearly defined nucleus of medium dark cells lying among the ventral fibers of the external capsule as they fan out in the postero-ventral amygdaloid region, below the caudal ends of the lateral and basal nuclei and obscurely confluent with them. It is not only present in Caenolestes, but it would be only too easy to subdivide the heterogeneous caudal amygdaloid region into further nuclei. The cortical amygdaloid nucleus (amg., Fig. 12; nuc.amg.cort., Figs. 38-43) forms almost the entire superficial portion of the amygda- loid complex. Laterally it approaches or adjoins the ventral edge of the pyriform cortex and medially the ventral edge of the hippocampus behind the choroid fissure (fs.ch.) along an internally well marked and an externally partially obvious medial continuation of the amygda- loid fissure (fs.amg.m). Its histological structure assumes in places a structure similar to that of the pyriform cortex. Its deeper portion is very heterogeneous. Some confusion, however, results from the tangential nature of the posterior sections of the hemisphere. The cortical amygdaloid nucleus receives secondary fibers by way of the pars ventralis of the lateral olfactory tract. The hippocampal-amygda- 2O4 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. loid junction (subic. (amg.), Figs. 40-42) along the median extension of the amygdaloid fissure is also evidence of fiber connections, and Johnston (1923) has demonstrated them in the Virginia opossum. The stria terminalis (st.t., Figs. 32-38), the great fiber system re- lated to the amygdaloid complex, has also been carefully analyzed by Johnston (1923). Practically its entire course is quite clear in these sections, and rostrally and caudally its five component bundles as iden- tified by Johnston, can easily be recognized, but in the compact medial portion of the tract, where it lies in the stria bed between the cerebral peduncle and the ependyma of the lateral ventricle they cannot be in- dividually identified. Johnston's numbers have been affixed to these bundles as traced by him in the Virginia opossum. Rostrally they are certainly correctly applied, but caudally they are applied without direct evidence from the brain of Caenolestes of continuity with the respective rostral bundles. The lateral part of the anterior olfactory nucleus passes without break directly over into the least differentiated anterior pyriform cor- tex behind it. The cortex of the pyriform lobe covers most of the lateral surface of the pyriform lobe. Antero-laterally it may be divided into three regions: the anterior pyriform cortex (cx.pir.a., Figs. 13, 15, 26-31; area piriformis anterior, Gray, 1924; frontal lobe of Cajal, 1911), which passes more caudally by very gradual transition into still more differentiated medial pyriform cortex (cx.pir.m., Figs. 13, 15, 32-34; area piriformis medialis, Gray; anterior pyriform cortex of Cajal; this in turn merges more abruptly into the most specialized posterior pyriform cortex (cx.pir.p., Figs. 12, 13, 15, 42-44; area piriformis posterior, Gray; superior temporal or angular nucleus or center of Cajal. Since the lateral part of the anterior olfactory nucleus merges insensibly with the anterior pyriform cortex, and the latter merges insensibly into the medial pyriform cortex, no definite boundaries can be made out between them. The second cortical layer (not counting the external fibrillar layer) becomes progressively more and more condensed, and any lines of demarcation are, in these sections, merely arbitrary. But the posterior pyriform cortex (cx.pir.p., Figs. 42-44), which is on a different plane functionally from the remainder of the pyriform cortex because (by Cajal's definition) it receives no secondary olfactory fibers, also differs histologically. Its histological development is connected not only with the absence of secondary olfac- tory fibers, but perhaps even more with the increase of non-olfactory fibers, and its consequent elevation into an associational area second 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 205 in rank only to the neopallium. It displays a very striking anatomical character. This is mainly due to the development within it of a wide and very dense plexus in the third layer of the cortex. In Caenolestes this plexus, which according to Cajal is in the mouse of extraordi- nary density, almost fills the pyriform wall dorso-caudally. Its broad, rounded head rises slightly above the level of the rhinal fissure. It contains many pale giant cells, found nowhere else in the pyriform cortex of Caenolestes. The posterior pyriform region is the field of origin of the great temporo-ammonic or angular bundle of Cajal (1911, spheno-ammonic, 1906), which delivers a huge stream of highly correlated olfacto-somatic impulses to both the ammon's horn and the gyrus dentatus, for almost their entire length — certainly reaching as far forward in Caenolestes as the anterior level of the commissural region, where its structural influence is suddenly and strikingly felt. Since it mingles with other pyriform and with neopalial association fibers to form the cingulum limitans (ci.lim., Figs. 3042) at the inner dorsal angle of the lateral ventricle, and the cingulum ammonis (ci.am., Figs. 30-42) in the outer plexiform layers of the ammon's horn and gyrus dentatus, it is not separately named in the sections given here. In the posterior region of the hemisphere this avalanche of fibers may be seen pouring above and behind the ventricle into the presubicular and subicular regions and through them, either directly or indirectly, as perforating fibers to the ammon's horn and gyrus dentatus. Anteriorly the lateral olfactory nucleus certainly sends tertiary ol- factory fibers above the olfactory ventricle into the anterior hippo- campal formation (Figs. 26-27). The sections suggest that the dorsal path between the pyriform and hippocampal cortices might perhaps be patent for practically the entire length of the neopallium, in the deep layer of the corona radiata. Owing to the height of the rhinal fissure the distance is nowhere very great, and this would certainly be the shortest path for the more dorsal pyriform cortex. The anterior pyriform cortex must also contribute very largely to the subventricular systems, both to the olfacto-septal, olfacto-hippocampal and olfacto- f rontal systems anteriorly, and to the more posterior and extensive com- plex included under the head of the lateral limb of the medial forebrain bundle, as explained above (see page 198). The great efferent pyri- form path to stem centers is of course the olfactory projection path of Cajal, probably, like so many of these fiber systems, a doubly oriented one. The stria terminalis component of this system (John- ston's bundle 2, 1923) arises from the amygdaloid complex (so far as it is a descending bundle). The fiber connections between the 206 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. pyriform cortex and the amygdaloid complex have already been men- tioned (page 203). In addition to the antero-posterior divisions of the pyriform cortex, Gray (1924) has described several narrow longitudinal bands along the boundary fissures in the Virginia oppossum. Three such areas accom- pany the rhinal fissure: the area perirhinalis, the area piriformis dor- salis, and the area piriformis fissuralis, in order from above down- wards. The extreme shallowness of the rhinal fissure -in Caenolestes is not conducive to great development of these areas. The area peri- rhinalis, the transition area between neopallial and pyriform cortex, is probably individualized, but the other two are feebly developed and not distinct from one another. Ventrally also the area piriformis ventralis and the area subpiriformis of Gray tend to fuse into one band quite well marked, the ventral pyriform cortex (cx.pir.v., Figs. 27-34), especially anteriorly where it describes a deep reentering angle subjacent to the massive lateral olfactory tract. This band is here regarded as following caudally upon the latero-ventral part of the anterior olfactory nucleus (nuc.ol.ant.l.v., Fig. 26). Behind the tuber- culum, and even more rostrally, where the diffuse anterior perforated space (l.perf.a., Figs. 35-38) of Johnston (1923) intervenes between it and the nucleus of the lateral olfactory tract, the pyriform cortex exhibits a very sharp ventral margin back to the point where it gives evidence of infolding at the level of the lateral and basal amygdaloid nuclei. Still further back it becomes more or less continuous with the outer layer of the cortical amygdaloid nucleus, and histological differen- tiation tends to fade out. The tangential nature of the sections in this caudal region is somewhat confusing, and control by sections in other planes is unfortunately lacking. HIPPOCAMPAL FORMATION The hippocampal formation of Caenolestes (Figs. 6, 12, 17, 26-43) begins with a small patch of cells condensed upon the outer aspect of the rostral border of the anterior olfactory nucleus, at its dorso-medial angle, that is to say, immediately caudal to the b'ulbar formation (hip. a., Figs. 12, 17; cx.hip.a., Figs. 26-29). It thus takes part in the forma- tion of the rhinencephalic ring surrounding the olfactory ventricle, by means of which the two olfactory cortices are indirectly continuous with one another through the anterior olfactory nucleus. It is to be noted that while the hippocampal formation begins at the very rostral margin of this ring, in its intrabulbar portion, just caudal to the olfac- tory formation itself, a similar lateral (pyriform) condensation falls 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 207 short of this, being separated from the bulbar formation by uncorti- cated anterior olfactory nucleus. The explanation of this precocity of hippocampal differentiation might conceivably be tied up with the prevailing latero-ventral drift of the secondary olfactory fibers toward the lower olfactory correlation centers of the lateral wall and base of the hemisphere — pyriform lobe and tuberculum olfactorium. This trend is clearly brought out in the bulb, where there is never any considerable accumulation of secondary fibers in the dorso-medial re- gion, those arising there apparently for the most part making as rap- idly as possible for the laterally and ventrally situated lateral, inter- mediate and medial olfactory tracts. The only exception to this is the presence of the small more median portion of the pars dorsalis of the lateral olfactory tract (tr.oLd.in., Fig. 26), which may, and prob- ably does (as silver preparations indicate in the Virginia opossum, Herrick, 19243), give off some collaterals to the hippocampal forma- tion, which is, however, already strongly under the influence of ter- tiary olfactory fibers. It seems but natural that such a region, receiv- ing only a minimum of direct olfactory fibers and beginning almost, if not quite, at once to receive tertiary olfactory fibers from the more lateral portions of the anterior olfactory nucleus (these can be seen crossing above the olfactory ventricle) and presently still more highly correlated olfactory stimuli from the tuberculum by way of the sep- tum, should undergo accelerated differentiation in comparison with a region preponderantly invaded by secondary olfactory fibers (see page 219). Dorsally the continuity between the hippocampal formation and the pars dorsalis of the anterior olfactory nucleus becomes transformed at the junction between the neopallium and the rhinencephalon into a continuity between the hippocampus and the neopallium on the one hand, and between the pyriform cortex and the neopallium on the other (Figs. 26-27). Ventrally the hippocampal formation, at first continuous with the pars posterior of the anterior olfactory nucleus (nuc.ol.ant.p., Figs. 27-28), breaks away from the latter and becomes separated from it by an increasing interval of diffusely scattered small cells through which some fine pale fibers pass from the deeper septal region to the plexiform layer (Fig. 29). This area is replaced more caudally by the medial parolfactory nucleus (nuc.pol.m., Figs. 30-31). The ventral end of the precommissural hippocampus (Fig. 29) thins to a slender needle point which, as it recedes upward in the septal wall, approaches the pial surface. This slender column of cells widens 208 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. above very quickly as it merges with the neopallial gray. It soon begins to show a slight concavity toward the pial surface (fs.hip., Fig. 29), which more caudally comes to expression as the external hippo- campal fissure (fs.hip., Fig. 30). The cells diminish in size and in- crease in density from above downward, to the sharply pointed ventral end of the hippocampal formation, which curves upwards and back- wards in line with the ventral (medial) edge of the definitive gyrus dentatus at the fimbrio-dentate fissure near its rostral end at the antero- dorsal angle of the hippocampal commissure (fs.fim.d., c.d., cx.dent., Figs. 30-31). As the hemisphere increases in height the hippocampal formation lengthens, the upper thicker and larger-celled portion more rapidly than the lower, so that it comes to exceed the lower smaller- celled portion in length. It becomes at the same time thinner and denser. It is interesting to note that rostrally the hippocampal for- mation differentiates more rapidly at its peripheral (dentate) border, and that differentiation travels upward, the orderly condensed files of definitive ammon's pyramids appearing tardily. Caudally also we find the cortex dentatus apparently leading the way in the formation of the recurved temporal pole of the hippocampus (see page 215). When I was making the first Edinger drawings I noted a curious breaking up or disorganization in the precomrnissural hippocampal cell plate, due to the loosening up and paler staining of some of the cells of the upper region (*, Figs. 12, i/a). This rift or line of fracture I interpreted as the locus of the interpositio medialis, the break between the definitive gyrus dentatus and the ammon's horn. But to my great surprise the interpositio medialis formed rather sud- denly at a distinctly lower (more ventral) level at the anterior border of the dorsal commissure (ip.m., Fig. 31). A provisional explanation of this puzzling rift in the cell plate suggested itself later, however, in connection with Cajal's statement (1911, II., p. 754) that the gyrus dentatus and the more ventral half of the ammon's horn (his extra- ventricular ammon's horn or region of grand pyramids) seem to form an indissoluble anatomical and functional unit, owing to the exclusive distribution of efferent gyrus dentatus axons (the "mossy fibers") to the extraventricular ammon's horn. The line of fracture in question might presumably be interpreted, then, as the upper limit of the dis- tribution of the "mossy fibers", the boundary between intra- and ex- traventricular ammon's horn. (For a provisional explanation of its absence in the supra- and postcommissural ammon's horn, see page 209 below). The material available here does not furnish proof of this hypothesis, which is offered merely as a suggestion. 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 209 At the level of the anterior face of the dorsal commissure the pre- commissural hippocampus, which, as Elliot Smith (i8o,6c) long ago pointed out, exhibits in marsupials all stages of development to be seen in the reptilian hippocampus, suddenly assumes the definitive mam- malian appearance of double interlocked arcs by the abrupt formation of the interpositio medialis separating the ammon's horn and the gyrus dentatus, and the synchronous humping up of the latter into a horse- shoelike form into whose ventral concavity the terminal lamina (Levi, 1904; nucleus fasciae dentatae, Elliot Smith, i896c) of the ammon's horn is displaced. The explanation of these sudden changes is proba- bly three-fold: external pressure, internal pressure and neurobiotactic influence. External pressure is exerted on three sides : from above by the downward pressure of the neopallium (as pointed out by Elliot Smith, Levi and many other neurologists), medially by the opposite hemisphere, and ventrally by the rigid barrier of the dorsal commis- sure and subpallial structures, beneath the choroid fissure (the atro- phic choroid barrier so strongly emphasized by Levi, 1904). The in- ternal pressure arises from the intrinsic growth of the hippocampus, chiefly the ammon's horn, which must crumple or roll to adjust itself to the space allotted. The neurobiotactic influence is in the first place due to the "perforating fibers" of the great temporo-ammonic tract and fibers associated with it in the cingulum limitans and cingulum ammonis (see pages 201 and 205). These fibers distribute equally to the gyrus dentatus and the ammon's horn. Those fibers which dis- charge into the granule cells of the gyrus dentatus break across the hippocampal fissure as it deepens and cause a more or less partial obliteration of the fissure by a secondary fusion of its two lips. They pull the sheet of granules upward through its entire length along a longi- tudinal axis (nearly median in Caenolestes}, so that in cross section the gyrus dentatus describes a horseshoe curve. The efferent fibers of the gyrus dentatus (axons of the granules, "mossy fibers"), con- verging upon the terminal lamina of the ammon's horn, cause a break at the position of the interpositio medialis (ip.tn., Fig. 31), and the dis- placement of the lamina terminalis into the ventral concavity of the gyrus dentatus. Figure na, b, c illustrates diagrammatically the mode of formation of the definitive mammalian "hippocampal figure" in Caenolestes under the three influences named. The arrows indicate the directions in which the force is applied and its relative strength. It might be possible to invoke also Kappers' ever useful concept of neurobiotaxis to explain the obliteration of the precommissural rift or fracture in the hippocampal formation (*, Figs. 12, i?a) noted above 2io FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. page 208) as the possible boundary line between the intra- and extra- ventricular ammon's horn; the extraventricular ammon's horn alone receives "mossy fibers" from the gyrus dentatus, a circumstance that might be held to account for the rift or fracture line in question; but the grand pyramids of this region, according to Cajal, send recurrent collaterals into the external plexiform layer of the ammon's horn (stratum lacunosum) which rake the whole extent of the ammon's pyramids, both intra- and extraventricular, and so tend to close any gap in them. Furthermore the afferent commissural fibers in the alveus might contribute to the same result, as well as to the forma- tion of the interpositio lateralis, by a general neurobiotactic compacting of the whole line of pyramids. More than this, the temporo-ammonic fibers which distribute to the ammon's pyramids either by way of the superficial plexiform layer external to the stratum lacunosum, or by way of the alveus (CajaPs temporo-alvear tract), engage, according to Cajal, not only the intraventricular but also some of the nearer extraventricular pyramids, perhaps most of them. At any rate it would seem as though the initial agent in the production of such dislocations as those considered is always the functional activity of the intrinsic structures concerned, and not the mechanical action of the fibers which may later not only occupy the break, but enlarge it, even to the extent, conceivably, of interfering with the function which originally pro- duced it. The dorsal or hippocampal commissure (c.d., Fig. 31) exhibits a narrower and denser dorsal portion, the psalterium dorsale (ps.d.), which Cajal (1911) considers the commissural path of the crossed portion of the temporo-ammonic system, and the wider and much more diffuse psalterium ventrale (ps.v.) the commissure of the ammon's axons. Many cells are mingled with these latter fibers (the nucleus of the commissure), so that on the whole the dorsal commissure in comparison with the much denser anterior commissure seems larger than it really is. The precommissural fornix fibers (f.prcom., Fig. 30) may be seen passing vertically downward in front of the commissures, and above the psalterium dorsale the dense short mass of the fornix longus, bending downward and partly interweaving with the psalteri'um dorsale, forms the "knieformiges Biindel" (Koelliker; Livini, 1908); they belong to the median striae Lancisii (Johnston, 1913). The descending columns of the fornix (c.for., Fig. 31) collect in the usual way close to the midline as deeply stained oval bundles above the an- terior commissure and pass down behind it into the hypothalamic re- gion. Two similar oval bundles of pale fibers located between the 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 211 fornix bundles and the anterior commissure on either side of the inferior recess (V.».) may be the supracommissural bundle of the stria terminals (Johnston's bundle 4, 1923). They seem to be receiv- ing more lateral fibers disposed in the same diagonal direction as some of the darker fibers entering the fornix bundles. I cannot tell whether they are of septal or hippocampal origin, or both. These bundles can- not in my sections be definitely followed laterally into the stria bed above the anterior commissure. There is also the possibility that they belong to the stria medullaris system, but I cannot follow them to the habenula. As Johnston (1913) was the first to point out, marsupials pos- sess well developed medial striae Lancisii above the dorsal commissure, a level at which full-bodied hippocampus also exists, proving that the lateral and not the medial stria of the indusium are the vestigial rem- nants of degenerate hippocampus in "callosal" brains. These median striae are perfectly clear in Caenolestes, but they are omitted from the reduced sections figured here. They are seen as cross cut fibers upon the dorsal surface of the hippocampal commissure and as vertical fibers cutting across the anterior dorsal commissure fibers and partly interwoven with them (see knieformiges Biindel, page 210). In connection with the structures under discussion it may be well to describe the form taken by an interesting ventral diverticulum of the superior recess (of Elliot Smith), which has been described and figured in sagittal section of Johnston (1913) for the Virginia opos- sum (Fig. i6b). In Caenolestes this subcommissural pouch (Fig. i6a) runs its course between section 510 (Fig. 30 shows the beginning of the glial mass in which it is embedded) and section 528. The dor- sal recess or sac bends down around the rostral surface of the dorsal commissure, carrying with it membranous roof tissue which hypertro- phies in the septum between the commissures and in the pial side of the precommissural area to form a rather thick mass of glial tissue between the medial fornix fibers. This mass, composed of densely crowded fine pale granules, is sharply separated from the nervous tissue adjacent to it. Rostrally it is bifurcated and caudally it comes to a median point in the septum beneath the dorsal commissure. It contains a lumen of similar shape, lined with ependyma, which ends caudally as a median recess in front of and above the inferior recess and bifurcates rostrally to end in two smaller diverticula in the pre- commissural septal walls. In Caenolestes the communicating canal to the dorsal recess or pouch is collapsed and obscured in dark-stained membranes and blood vessels. Figure i6a shows a reconstruction of 212 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. the ventral diverticulum of the dorsal recess in Caenolestes in the hori- zontal plane below the entrance of the communicating canal ; figure l6b the sagittal section of this diverticulum in the Virginia opossum from Johnston (1913, part of figure 35, redrawn). The middle portion of the hippocampus (Figs. 32-37) is naturally less interesting than the two extremities, which offer more develop- mental and adaptational clues, and it therefore calls for but few re- marks here. With respect to the amount of gyrus dentatus exposed upon the median surface of the hemisphere, Caenolestes is intermediate between Perameles and Notoryctes, a fact which indicates that the neopallial pressure, apparently the chief varying factor here, exceeds that in the latter and falls below that in the former. According to Elliot Smith ( i895b, Fig. 6) Notoryctes has the least extensive neo- pallium found among mammals, and "in no other animal does one find the simplicity of arrangement which the hippocampus of Notoryc- tes presents, an appearance which recalls the foetal hippocampus of Perameles or Macro pus." (This statement refers only to what I have called the "hippocampal figure" as seen in cross section and not to the development of the temporal pole of the hippocampus, which will be discussed below.) The intermediate condition of the exposed gyrus dentatus in Caenolestes is especially interesting in view of the, simi- larly intermediate condition of the cerebellum in Caenolestes. It should be held in mind that the more caudal sections of the hippocampus become progressively more tangential to the structure itself and present an increasing distortion of the "hippocampal figure" of double interlocked arcs, as well as an exaggeration of its size. We should have a true picture of the hippocampus only if we could make all sections vertical to the hippocampal axis, which, since that is curved, would make them radial. The sections show that the hippocampus as a whole has in Caenolestes not yet "turned the corner"; that is to say, the typical "hippocampal figure" does not appear twice, like an object and its inverted image, in any one cross section, as it does, for example, in the rabbit's brain (Fig. 2oa, b, c, Winkler-Potter, re- drawn) ; or once, the "mirror image" only, as, for example, in the lion (Fig. 21, Elliot Smith, redrawn) and in man. In Caenolestes the gyrus dentatus alone has recurved and therefore it occurs in two places, one dorsal and one ventral, in some sections (Fig. 39). Neither the fimbria nor the ammon's horn has, however, recurved, and hence the ventral hippocampal figure of double interlocked arcs is incom- plete and the temporal hippocampus is practically unrecognizable at first glance. The Virginia opossum has advanced one step farther in 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 213 the recurving process by the formation of a small forwardly directed pouch on the rostral face of the ammon's horn at its ventral border below the caudal end of the fimbria (Fig i8a-g, cross sections, from unpublished drawings, Streeter, redrawn; Fig. i8h, reconstruction macle from same). This pouch is evidently formed in response to the neurobiotactic attraction of the gyrus dentatus, which has here grown forward beyond the field of the unrecurved ammon's horn. This change still does not result in a complete "hippocampal figure". But in higher marsupials, like Hypsiprymnus (Fig. 193, b, c, Livini, 1908, re- drawn), the temporal development has advanced to a point which allows the typical hippocampal figure to appear twice in the same section. In Notary ctes (Fig. 22, Dart, 1920, Fig. 13, redrawn), to my great surprise, the temporal pole of the hippocampus is apparently quite as well developed as in Hypsiprymnus, and considerably better developed than in Caenolestes and the Virginia opossum. Notoryctes is a sightless form without external ears, and the elongation of the hippocampus and great amount of recurving of its temporal pole is perhaps to be regarded as compensatory, in view of the absence or extreme reduction of the vis'ual and auditory systems. The median surface of the hemisphere (Elliot Smith, i895b, Fig. i) has a peculiar peaked appearance, which may perhaps be due to the elongation of the hippocampus and its consequent caudal bowing. The pressure of the slightly developed neopallium is not sufficient to cause a great degree of inrolling, not so much as in Caenolestes. Ornithorhynchus also has an elongated slender but well formed temporal pole of the hippocampus (Elliot Smith, i896b, Figs. 4, 5, 5", pp. 472-3 — the section given through the "tail" of the hippocampus is not very near its temporal end, however). In this case the elongation of the hippocampus and its temporal recurving, which is pronounced, is probably to be corre- lated with the very great size of the neopallium and consequent com- pression of the hippocampus. I do not know the condition in Echidna, but should expect a similar situation from similar conditions. The temporal pole of Perameles is also unknown to me. In the rabbit's brain, a "callosal" one, in which the hippocampal commissure is caudally displaced by the corpus callosum, there are also two typical hippocampal figures, posed in opposite directions, joined by the tangentially cut fimbria (Fig. 2Oa, b, c, Winkler-Pot- ter, 1911, pis. XII, XIII, XIV, redrawn). Here we have two exactly identical and completely detached images in front of the mid- dle of the fimbria (Fig. 2oa and b). I do not know whether this is 214 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. the case also in Hypsiprymnus (Fig. ipb). Livini does not give a section showing the fimbria in two parts. The more rostral section (Fig. 193) shows, however, a little ventral "island" which may be the projecting tip of the uncus, and between this and the next caudal sec- tion given the fimbria may appear in two parts; but even so it would probably be less recurved than in the case of the rabbit. In the case of the higher and more microsmatic mammals, as in the lion (Fig. 21), and in man, the dorsal "figure" disappears as the corpus callosum elongates and the supracallosal hippocampus (lateral striae Lancisii and the accompanying gray) stretches, and only the ventral figure persists completely. We can thus easily assemble a short series which will adequately illustrate the remarkable complete reversal of the hippocampus in the course of phylogeny by the agency of the combined action of callosal growth, tremendous neopallial hypertrophy, and the anchoring of the hippocampus in a rostral position by the pyriform cortex, which is itself strongly anchored to the bulbar forma- tion. The tuberculum and the amygdala, as they diminish, retain their old places one behind the other between the bulb and the temporal end of the hippocampus, and medial to the lateral peduncular gray (lateral olfactory gyrus) and pyriform cortex. The neopallium expands enor- mously in the caudal direction, further accentuating the hippocampal reversion. The pyriform cortex loses its distinctive histological char- acter and assumes a progressively more neopallial appearance, until the primate condition is attained, where, as "gyrus hippocampi" it becomes histologically practically identical with the neopallium (save in the pres- ence of the external fibrillar layer in its anterior portion). The pos- terior pyriform region, that devoid of secondary olfactory fibers and richly supplied with non-olfactory fibers, begins, even in the Virginia opossum, (Gray, 1924) to resemble neopallial cortex rather closely along the border for the caudal prolongation of the rhinal fissure. In the mouse, as Cajal's (1911) intensive studies show, this area has developed an exceedingly characteristic histological structure of its own, and assimilation to the neopallial cortical pattern has not apparently made so much progress. The same is true in Caenolestes, in which the boundary between the neopallium and the posterior pyriform area is pretty definite (Figs. 41-42), so far as can be construed from frontal sections. The conditions existing at the temporal extremity of the hippocampus in Caenolestes, the Virginia opossum, Hypsiprymnus and the rabbit certainly stress the dynamic character of the temporal dis- placement of the hippocampus — "the brain is not 4a rigid mosaic of morphological units which were laid down in the primordial vertebrate 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 215 and thereafter preserved inviolate" (Herrick, 1922, p. 199). The little, rostrally directed ammon's horn pouch in the Virginia opossum is clear evidence of the regulative nature of the changes going on. In Caenolestes the forward advance of the temporal gyrus dentatus has not removed it from the unrecurved ammon's field, but in the Virginia opossum this would have happened if the ammon's horn had not responded by the formation of the pouch. It is to me rather surprising to find the gyrus dentatus leading in the reversal of the hippocampus when rostrally it appears in definitive form tardily. But this tardiness is perhaps more apparent than real, since the lower portion of the precommissural hippocampus is in mam- mals, as Elliot Smith has stated for the same region of the posterior part of the reptilian hippocampus, clearly "on the way" towards differ- entiation into definitive gyrus dentatus. In the mammalian precom- missural hippocampus it is the ventral portion, that directly contin- uous with the definitive gyrus dentatus, which takes the lead in differ- entiation. Differentiation apparently travels from below upwards, and the definitive ammon's horn or rather, perhaps, its "intraventricular" portion, is perhaps the last to develop its distinctive structure and extent. It is at least the last to begin differentiating. Pedro Cajal's illuminating studies on the reptilian brain (Varanus and Lacerta, 1917, 1919) by the silver methods show that in the ventral or medial small-celled portion of the hippocampal formation (quasi gyrus dentatus) there is a clear transition in cell type from the deeper to the outer cell ranks. The cells of the outermost cell ranks (nearer the pia) are practically true granules, the innermost true pyramids of the ammon's type. This is very significant, and it would not, I think, be surprising to find in full-bodied precommissural hippocampus of lower mammals an analogous state of affairs. The situation in this Caenolestes series strongly suggests, far in advance of the interpositio medialis and the inrolling of the gyrus dentatus, that the ventral cells of the more ventral region of the hippocampus are rapidly verging towards the true granule type, so that the transition to definitive gyrus dentatus involves apparently no sudden cytological changes. The hip- pocampus is apparently affected by two distinct waves of differentiation, both starting at the ventral border and traveling upward. The first de- velops cells of pyramidal type : the second, following in its wake, trans- forms pyramids into granules (phylogenetically speaking). The first wave affects the entire width of the hippocampal formation, reaching its dorsal border (locus of the interpositio lateralis), the second stops at the level of the interpositio medials, affecting the "dentate" region but 216 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. never the ammon's region above it. The two main types of cells are perhaps to be interpreted as an expression of the enhancement of function in the hippocampus by its organization into a highly inte- grated duplex structure, the ammon's horn from which spring the long projection axons (which alone form the efferent path of the hippocam- pus), and the gyrus dentatus, whose shorter axons of a specialized type ("mossy fibers") deliver stimuli to part of the ammon's pyramids directly and to many others indirectly (recurrent collaterals of grand pyramids). We can never perhaps get the complete phylogenetic story in any one form, certainly not in the lowest and highest, where either later or earlier steps fail to appear. Although the general form of the ammon's horn and the gyrus dentatus cell sheets in Caenolestes is really simple, it is difficult to show them both in the same diagram. Therefore dissected recon- structions (Figs. i?a and i?b) attempt to show them separately, one or the other being cut away at the commissural level of the hippocam- pus. The two parts are shown in place together in figure 12. It is seen that the gyrus dentatus takes the form of a sort of helmet or hood with the top of the crown or apex pointing caudally; the front or visor of the helmet corresponds to the elongated dorsal portion of the gyrus dentatus extending above the commissure where it ends in a little pouched thickening (Fig. 30) ; the back or neck portion cor- responds to the ventral or temporal recurved part of the gyrus denta- tus. The ammon's horn sheet takes the form of a loose scroll, greatly widened posteriorly in the sections (Figs. 38-40), so that it is really more like a cornucopia, with the wide end caudo-ventral. The up- turned median flap of this scroll, covered with a thin coating of alveus fibers, is exposed upon the median surface of the hemisphere, save where its caudally directed corner is inserted into the hooded portion of the gyrus dentatus. This upturned ammon's horn forms Elliot Smith's extraventricular or inverted hippocampus ("dorsales Blatt" of Koelliker). It is seen in the sections to diminish in length as its lower edge recedes upward between the approaching dorsal and ventral portions of the gyrus dentatus (Figs. 34-40), which still more caudally unite to form an oval ring (the "crown" of the hood) containing the caudal tip of the inverted hippocampus (a few scattered cells, Fig. 41). At the caudal end of the choroid fissure the continuity of the two parts of the ammon's horn (in sections) is dissolved (Figs. 40-42), and the lower end of the extraventricular ammon's horn begins to recede upward, while that of the intraventricular ammon's horn seems to turn laterally under the median angle of the ventricle (Volsch's 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 217 subventricular hippocampus, 1906), where it comes in contact with the amygdaloid complex (stria bed and cortical amygdaloid nucleus). This contact extends backward along the line of the medial continua- tion of the amygdaloid fissure (fs.amg.m., Figs. 40-43), always very strongly marked internally and sometimes externally apparent. As the sections show, the more temporal extremity of the gyrus den- tatus (Fig. 39) is contiguous with the ammon's horn and perhaps with the amygdala where the two structures adjoin. The fiber connection between the hippocampus and the accessory basal nucleus of the amyg- dala (Johnston, 1923, Fig. 55) is not apparent in this series. This by no means proves its absence. The emphasis here placed upon the amount of recurving of the hip- pocampus is intended to apply specially to mammals, where it is to be considered in connection with the inrolling of the hippocampus under the double necessity of enhancement of function and economy of space. Even mammals, as we have seen above (Notoryctes), do not present a quite orderly series in this respect. The degree of recurving alone is not a criterion of advancement in hippocampal development. In some reptiles (e.g., Cistudo Carolina (Johnston, 1915, Figs. 6, 12, 13, 45) the recurving of the temporal pole is more pronounced than in Caenolestes — in single cross sections two separate "hippocampal figures", complete for this brain, appear. The factors provoking the recurving in such cases are easily recognisable and need not be detailed here. The factors operating in mammals to bring about inrolling have apparently provoked a method of increase for the hippocampus, which, in concert with other existing conditions, may result in less recurving of the temporal pole than in lower forms. In higher forms the caudal displacement of the hippocampus due to callosal elongation, reverses or recurves the inrolled hippocampus exactly as in the case of simpler brains alluded to. What we seem to have always before us is the structural record of the solution of various problems of regulative be- havior, from which we may attempt to reconstruct a phylogeny of function. We should remember that there are always a great number of factors working in concert or as more or less independent variables, and that the outstanding potency of any particular factor does not mean its exclusive activity. Above the medial amygdaloid fissure (fs.amg.m., Figs. 6, 12, 17, 40-43) the dorsal or subicular border of the ammon's horn adjoins, as it curves upward, the posterior pyriform cortex along a line reaching upward to the medial extension of the rhinal fissure (fs.rh.m.} thus completing the great horizontal rhinencaphalic cortical ring, as de- 218 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. scribed in the earlier part of this paper dealing with the external form of the brain (page 185). In Caenolestes this line is longer than in the Virginia opossum, because the position of the rhinal fissure is consider- ably higher. The exact point at which the rhinal fissure hits the subic- ular edge of the hippocampus is not determinable, owing to the fading out of histological differentiation, and therefore the extended portion of the fissure is shown as a broken line (fs.rh.m., Figs. 6, 12, 17, 42). Elliot Smith (i895b, p. 183), in discussing the hippocampus of Notoryctes, makes the following remarks : "The hippocampus is equally convoluted in all mammals, because it reaches its maximum develop- ment quite early in the phylogenetic history of the individual. Thus in Platypus it possesses a histological differentiation quite as complex and fine as is found in the highest mammals. Like the pyriform, it is developed early both in phylogeny and ontogeny in accordance with the development of the olfactory apparatus. Because part of the smell center should reach a high state of development, when the pallium is not proportionately intricate, is no argument that the cortex of the smell center does not behave like the rest of the cortex in similar circum- stances. It should be noted, however, that, intimate as is the connec- tion between the hippocampus and the olfactory lobe, the relative sizes of the two parts are by no means constant. Thus, in spite of the marked difference in the sizes of the olfactory bulb in Ornithorhynchus and Perameles, there is no appreciable difference in the sizes of their hippocampi. In Notoryctes the size of the hippocampus is relatively small, considering its huge olfactory. What determines the size of the hippocampus is hard to say." (Italics mine.) These statements are very suggestive, and in my study of the cere- bral hemisphere of Caenolestes they have often recurred to me. Are we, however, warranted in saying that the hippocamp'us reaches its "maximum development" in the lower mammals, and that in Ornitho- rhynchus the histological differentiation is "quite as complex and fine as is found in the highest mammals"? It is true, as explained above, that the definitive mammalian "hippocampal figure" appears suddenly and at a constant level in the lowest mammals. But the amount of inrolling and the development of the temporal pole of the hippocampus both vary considerably, as we have seen, in the marsupials alone. Now in the marsupials we find also a great variation in the amount of neo- pallium present, without, however, a greatly elongated eutherian type of corp'us callosum to account for the temporal pole by mere displace- ment of the entire supra- and postcommissural hippocampus. Fur- 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 219 thermore, in considering the histological differentiation of the hippo- campus, we now know that if we take into account not merely the simple ranks of ammon's pyramids and the granules of the gyrus den- tatus as they are stained by ordinary cell methods, such as Nissl, but also the many types of cells in the other layers of the hippocampal cortex, as they were demonstrated by Cajal (1911) in the mouse by the silver methods, we get a tremendously suggestive picture of the possibilities for progressive complication within the hippocampus. I attempted to count the different types of intrahippocampal reenforcing and stepping-up devices as described and figured by Cajal in the mouse, and found some twenty-five or thirty of them. Only a com- parative study by the silver methods (such as that being carried on by del Rio-Hortega, 1919) of the whole range of mammalian brains could reveal the stage at which the maximum development (maxi- mum histological differentiation) of the hippocampus was attained. I should rather expect to find it nearer the upper than the lower end of the mammalian phylum, among those brains in which the discrep- ancy between the olfactory bulbs and the neopallium is marked, but in which smell is still an active function. As Elliot Smith remarks, the hippocampus does not, like the pyriform lobe and the tuberculum olfactorium, decrease pari passu with the olfactory bulbs. On the contrary, it holds its own (disregarding the amosmatic brains, and even they manage to retain some recognizable hippocampus), perhaps because it can so efficiently combine enhancement of function with economy of space (see page 217) as the neopallium increases, and this, apparently because of the increasing activity back and forth be- tween the two. In this connection it is to be noted (see Herrick, 1922, page 196) that the primordium hippocampi has as early as the preganoidean stage no direct somatic connections, and in the ganoidean stage begins to lose direct olfactory connections, and very soon begins to develop a characteristic structure in the presence of indirect cor- relation and associational connections from various sources, "result- ing in a topographical rearrangement which prepared the way for the differentiation within this area of true hippocampal cortex in higher forms" (Herrick, 1922, page 196). The ability of the hippocampus to preserve a sort of structural constancy or individuality with almost total loss of direct olfactory connections, and in the face of great reduction of primary and secondary olfactory centers, has apparently undergone no abatement in the presence of the developing mammalian neopallium. Its peculiar structural pattern has merely unfolded in its own way; it has not approximated the neopallial structural pattern. 220 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. Functionally the increasing neopallial-hippocampal bond is of tre- mendous import. On reflex and lower psychic levels ol faction is mainly if not exclusively linked with food and sex; on higher psychic levels it becomes increasingly significant esthetically and intellectually, in correspondence with the increase of the neural connections between the hippocampus and the somatic pallium which apparently accompanies the decrease of neural connections between the hippocampus and the diminishing olfactory bulbs and their immediately dependent secondary olfactory centers. This does not mean that the hippocampus increases phylogenetically in the same ratio with the somatic pallium, or in any constant ratio with it. Apparently neither is true. But it seems hard to believe that a direct relation does not exist between the neopallium and the hippocampus, a relation which is responsible not only for the remarkable preservation of pattern (not its initiation) and maintenance of size of the latter structure in higher and relatively microsmatic brains, but finally for the sublimation of smell into a "nobler" sense, practically as truly subservient to higher psychic life as are vision and audition. Language is a form of behavior, and the rich poetic and spiritual imagery clustering around such words as "fragrance" and "incense" amply testifies at once to the importance of the existing sense of smell in man and to its sublimation. So too does the tre- mendous evocatory power of smell, when it recreates vanished ex- periences which arouse high and beautiful thoughts or emotions. CORPUS STRIATUM The corpus striatum consists of a caudate nucleus (nuc.caud., Figs. 12, 15, 28-38) projecting into the ventricle and extending farther ros- trally and caudally than any other part of the striatum, and a lenti- form nucleus lateral and ventral to it, in two parts, the putamen (put., Figs. 29-35) and globus pallidus (gl.p., Figs. 12, 15; glob.p., Fig. 34) The medial or septal part of the head of the caudate nucleus, the nucleus accumbens (nuc.ac., Figs. 12, 20-30; the lateral parolfactory nucleus of Johnston, 1913 — see page 199), has already been mentioned in connection with the fusion of the head of the caudate with the tuber- culum (page 196). Rostral to this fusion the large head of the caudate is laced with small diagonal fiber bundles, like darning stitches, and heavily fringed on its ventral border with diagonal fibers, all the way from the septum to the lateral arm of the anterior commissure (Fig. 29). It is split dorsally by a small caudally directed ventral di- verticulum of the lateral ventricle, roofed by the lower surface of the anterior commissure which here breaks across the ventricle to reach 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 221 the median wall (Figs. 30-31). The large and compact internal cap- sule (cap. i.,, Figs. 29-34) separates the caudate nucleus from the large putamen, whose lateral boundary is the curve of the external capsule (cap.e., Figs. 28-35). The caudate nucleus, including the nucleus accumbens, and the putamen are in this series characterized by a peculiar dark, gun-metal gray background, seen also in the lateral and basal amygdaloid nuclei, setting all these structures sharply off from neighboring ones. I do not know what it may be worth as a criterion, but it is interesting to note that the nucleus accum- bens, which Johnston (1913) considers to be the lateral parolfactory nucleus", is uniform with the rest of the caudate in this respect, being sharply delimited from the septum by this ground color.* The globus pallidus (gl.p., Figs. 12, 15; glob. p., Fig. 34) is a prom- inent mass of pale giant cells in a rich tangle of fibers, situated near the center of the sections in which it appears, just below the internal capsule as it passes into cerebral peduncle, and not very far behind the anterior commissure level. Small giant cells trailing irregularly towards the basal forebrain bundle in the innominate region (prethala- mus, border nucleus of Volsch) may belong to the basal nucleus of the palacostriatum (De Vries, 1910; Kappers, i92ib). NEOPALLIUM No attempt has been made to analyze the neopallium of Caenoles- tes into anatomical regions, although even this inadequate series gives some evidence of its possibility. It is hoped that the two series to be made from the brains of Orolestes may, with the aid of the intensive cortical analysis of the Virginia opossum in this laboratory (Gray, 1924), render a fairly adequate analysis of the somatic cortex possible in Caenolestes and Orolestes. The claustrum, which Brodmann and others consider to be a neo- pallial derivative, is certainly present in the more dorsal part of the *I was much interested to read (Kappers, 1923, page 365) of the "Spatz reaction" of sulphur ammonium, based on the presence of iron, which gives a much stronger blue color in the palaeostriatum than in the neostriatum. While not inclined to attribute very great test force to his reaction, Kappers thought it interesting that results on the chick and man corresponded. In this iron- haematoxylin series of Caenolestes the dark coloration has a very different dis- tribution. The globus pallidus is conspicuously without it. The caudate nucleus (nearly all but the tail) and parts of the tuberculum fusing with it, the putamen, and the two newest amygdaloid nuclei (according to Johnston, 1923), the lateral and the basal, are the regions strikingly affected by the dark coloration. Very pronounced in the rostral portion of all the structures named, it tends to fade out caudally in each of them. I do not know enough about either the Spatz reaction or the dark coloration here to form an opinion as to any relation between them. 222 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. anterior pyriform region, but is not clearly delimited. It apparently forms an irregular cell plate between the external capsule and a cell- poor strip in which some fibers appear which may be an incipient cap- sula extrema (clau., cap. ex., Figs. 27-34). Elliot Smith (19193, b) considers the claustrum a derivative of the upturned lower edge of the pyriform cortex. This material affords no real evidence either way, although the claustrum here seems to be continuous with the deeper layers of the neopallial cortex. The cells of this brain, which are remarkably well stained, con- sidering the method used, may be divided first of all into two groups, which Volsch (1906) has called round and pyramid cells, referring to the hemisphere only. The pyramidal cells (Using the term very loosely for any angular cell) are black and show dendritic stumps, sometimes very long. The round cells are pale-stained with the nucleus clearly visible. The variation in size is very much greater for the round cells than for the pyramidal cells. All the giant cells of the hemisphere belong to the former type, those of the basal olfactory region, in the anterior portion of the lateral limb of the median fore- brain bundle (see page 198), those of the regio innominata, prethala- etc. The smallest cells also belong to this type, the olfactory granules, those of the gyrus dentatus, the exceedingly fine granules of some of the islands of Calleja in the tuberculum, and the scattered masses of very tiny granules tucked between adjacent structures like packing. Volsch considers these last to be non-nervous, I do not know on what evidence other than their size and resemblance to stained glia nuclei. The largest cells of the pyramidal or dark angular type are probably those found in the basal nucleus of the amygdala, for the hemisphere at least. GENERAL CONSIDERATIONS In the beginning of this study of a type of brain totally unknown to me I saw that no rational progress was possible save upon a broad and strictly comparative basis. In the course of parallel studies of the literature and of the sections of this particular brain, some more general considerations arose, first as persistent questions, usually en- tirely unanswerable by me at least, but sometimes coming halfway to rest in my mind. Two of these are briefly outlined below. One of the most puzzling aspects of such neurological studies as this comes out in connection with the very great diversity of anatomi- cal pattern of structures whose afferent and efferent connections seem almost or practically identical. How shall we explain the enormously 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 223 discrepant anatomical structure? The real functional adequacy of every existing structure must of course be taken for granted as the basis of its existence. But more than this must apparently be called in to explain its particular anatomical type of structure. Phyletic tra- dition clarifies some things — inherited type of anatomical pattern (or the potency to develop it) retained, embellished, even accentuated, so long as it is functionally utilisable or perhaps not actually disadvantag- eous. The hippocampus is perhaps the classic example under this head. Even when no longer adequate, as evidenced by reduction and degeneration, it seems very hard to be got rid of entirely. Fortunately for the comparative neurologist, it hangs on, sometimes only a pale and shrunken relic, long after it has been more or less supplanted by other structures, with similar or unlike functions, which have become more important in the action-system of the animal either because of fundamentally more serviceable type of anatomical structural pattern, or because of more useful afferent or efferent connections, or for both reasons. Thus a sort of natural selection (Roux's "struggle of parts"?) is constantly operating among structural patterns. A struc- tural pattern is safe so long as the demands upon it are not too heavy, or so long as a competing pattern with greater possibilities does not outrun it. The superiority and final supremacy of the competing pat- tern may depend not only upon its intrinsic possibilities, but also upon its topographical position and the character and activity of neighbor- ing structures. Thus the early "physiological isolation" (to use Child's now familiar term, 1915, 1921) of the primordium hippocampi from direct sensory stimuli (see page 219) led to "topographical and physio- logical relationships [which have] prepared the way for the differentia- tion within this area of true hippocampal cortex in higher forms" (Herrick, 1922, page 196). The trail of the Law of Neurobiotaxis is to be seen everywhere, and this is apparently the most potent single factor which can be invoked immediately to "explain" structural pat- tern (i.e., to translate it into functional pattern) within the central nervous system. Acting in concert with phyletic tradition, it might well result in exaggerations of structural pattern which do not seem to parallel the development of functional pattern. We must, of course, in comparing similar lists of afferent and efferent connections of very dissimilar structures, take into account the quantitative and positional differences of all the factors, as well as the selective operation of synap- ses. On the surface, perhaps the discrepancies seem much more strik- ing on the 'structural than on the functional side, although sometimes the reverse seems to be true. We must of course believe that no two 224 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. structures that are visibly different do have absolutely identical func- tions, not even right and left members of paired structures in the same brain. On the whole, however, the physiologist's frequently voiced objection to the anatomist's too great emphasis upon structural pattern may not perhaps be an unjustifiable one. The modern anatomist, of course, regards structure as merely a useful and convenient clue to function. Another question which naturally arises in such a study as this concerns what, if any, light is thrown upon the method of progressive evolution in the central nervous system. Davidson Black (1913, page 366) quotes Roux's definition of the two phases of the development of an active tissue: "self-differentiation", which goes on without re- gard to functional differentiation, and "dependent differentiation", which cannot proceed normally in the absence of functional connec- tions— in the nervous system, functional continuity of neuron systems, as Bechterew pointed out. The transition point between these two phases of growth Black calls the "critical point". Apparently the essential character or activity of progressive evolution in the central nervous system is the pushing forward of this "critical point", relegat- ing more and more of the period of "dependent differentiation" back into the period of "self-differentiation", where it undergoes compres- sion, abridgement, and, so to speak, distortion, thus building the founda- tion for newer and higher development in the period of dependent differentiation — like using the capstones of a newer structure to strengthen and enlarge the old foundation and so to fit it for still newer and more ambitious structures. What this "antedating" of structure implies in physiological terms, what metabolic changes or reorganization by which, or accompanying which, dissociations take place, some phases dropping out, others being temporally dislocated with reference to the stimuli formerly necessary to elicit them, or whether the dissociations are more apparent than real, I am not com- petent to discuss. In attempting to reconstruct the phyletic history of any form, it is to be remembered that not all the foundation blocks, original or "second-hand", have been retained in the enlarged foundation — some have been rejected entirely, all have been recut, and many have altered their relations with reference to others. The higher the form the greater the compression, abridgment and rearrangement, and there- fore the greater the necessity of prudence and hesitation in the recon- struction of the detail of phyletic pattern, no matter how simple and obvious, perhaps deceivingly so, the ontogenetic pattern may appear. 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 225 SUMMARY 1. The brain of Caenolestes (and of Orolestes) is of the extreme macrosmatic type, characterized by the great size of the olfactory bulbs and the great development of the higher rhinencephalic centers (Herrick, 1921). 2. The neopallium, in contrast, forms only the shallow cap of the cerebral hemisphere above the high rhinal and hippocampal fissures, and may perhaps be relatively the least extensive or the second least extensive among mammals (Herrick, 1921). 3. The olfactory cortices — pyriform and hippocampal — form, with the aid of the anterior olfactory nucleus, two periventricular rhinence- phalic rings partially united anteriorly : a smaller quasi- vertical one ros- trally, in which the pyriform and hippocampal cortices are doubly (supra- and infra ventricularly) but indirectly united through the an- terior olfactory, nucleus ; and a much larger horizontal ring formed by the additional direct union of the posterior pyriform cortex and the subicular margin of the ammon's horn on the median surface of the hemisphere between the medial prolongations of the rhinal and amygda- loid fissures. The two olfactory cortices are split apart dorsally by, the vvedgelike neopallium and ventrally by a similar wedgelike formation composed of the tuberculum olfactorium rostrally and the amygdaloid complex caudally. 4. The marked antero-posterior foreshortening of this brain is one of its characteristic features, frequently in evidence. It is probably to be explained, in part at least, by the exaggeration of the subcortical centers. 5. The hippocampus begins immediately behind the olfactory (bul- bar) formation, passes through practically all reptilian stages precom- missurally, and at the dorsal commissure rather suddenly assumes defin- itive mammalian form, under the three-fold influence of external pressure, internal pressure and neurobiotactic attraction. Its temporal end has just begun to recurve, the gyrus dentatus alone being involved. In the Virginia opossum the ammon's horn has begun to follow suit. More advanced critical stages can be added to these to form a complete and illuminating series of the phyletic development of the mammalian hippocampus, omitting the monotremes and the lowly marsupial Notor- yctes, aberrantly advanced in this respect. 6. In the amount of gyrus dentatus exposed upon the median sur- face of the hemisphere (an index of the degree of inrolling of the hippocampus) Caenolestes is intermediate between Notorcyctes and Perameles. 226 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. 7. The enormous size of the amygdaloid complex is a striking char- acteristic of this brain, as is also its prolonged contact with the temporal hippocampus, both ammon's horn and gyrus dentatus. 8. The absence of the aberrant bundle of the anterior commissure would on the basis of Elliot Smith's definition of it as a diagnostic character of diprotodont brains, seem to put Caenolestes among the polyprotodonts, and even to raise the question of the exclusion of the diprotodonts from America. 9. The cerebellum of Caenolestes and Orolestes is almost exactly intermediate between those of Notoryctes and Perameles, the two simplest mammalian cerebella hitherto described. 10. In consideration of the marked specialization and aberrancy of the two monotreme brains, the brain of Caenolestes (and of Orolestes so far as can be judged from the external anatomy) ranks as one of the three simplest and most generalised mammalian brains known at the present time. In view of the exaggerated development of the temp- oral pole of the hippocampus in Notoryctes, it really takes first rank. But on the whole, it is fairly intermediate between Notoryctes and Perameles, the two simplest and most generalized mammalian brains hitherto described (Elliot Smith), which it most closely resembles. It should offer — and actually does offer — especially promising clues for the reconstruction of the presumptive phyletic stages involved in the transition from the reptilian to the mammalian type. And since it is at once a mammalian brain and so simple and generalized — almost, indeed, a mammalian brain reduced to lowest terms — it also offers hints of Unusual legibility for the verification of the structural activities of the nervous system, viewed as regulatory behavior of a more or less plastic material. The cerebral hemisphere of Caenolestes, the only part of the brain yet studied in any detail, fairly swarms with exquis- itely clear examples of structural evidences of the operation of neuro- biotaxis, the principle whose conception and definition by Dr. Kappers of Amsterdam has transferred the study of brain morphology from a static to a dynamic basis. ACKNOWLEDGMENTS I am glad to acknowledge here my indebtedness to Dr. C. Judson Herrick, who set me upon this research and who has aided me with practical advice ; to Dr. Wilfred H. Osgood, who procured for me the use of further material and in other ways furthered this work; to Dr. Percival Allen Gray, Jr., who did me the very great service of reading the first draft of this paper without reference to the figures, and offered 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 227 many suggestions which have greatly contributed to the lucidity of the final draft; to Miss Helen Kates for welcome assistance in the tedious work of labelling the figures. I wish it were possible also to acknowledge fully the pleasure and benefit I have received from all those who were interested enough in this material to examine or dis- cuss it with me. BIBLIOGRAPHY BECCARI, N. 1910. II lobo parolfattorio nei mammiferi. Arch. anat. e embriol., g, pp. 173-220. BLACK, DAVIDSON. 1913. The study of an atypical cerebral cortex. Jour. Comp. Neurol., 23, PP. 351-369. CHILD, C. M. 1915. Individuality in organisms. The University of Chicago Press. 1921. The origin and development of the nervous system. The University of Chicago Press. 1924. Physiological foundations of behavior. New York. CROSBY, ELIZABETH C. 1917. The forebrain of Alligator mississippiensis. Jour. Comp. Neurol., 27, pp. 325-402. DART, R. A. 1920. A contribution to the morphology of the corpus striatum. Jour. Anat., 55, PP. 1-26. GRAY, P. A., JR. 1924. The cortical lamination pattern of the opossum, Didelphys virginiana. Jour.Comp. Neurol, 37, pp. 221 — 263. HERRICK, C. JUDSON. 1910. The morphology of the forebrain in Amphibia and Reptilia. Jour. Comp. Neurol., 20, pp. 413-549. 19213. The brain of Caenolestes obscurus. Field Museum of Nat. Hist., Pub. 207, Zool. Ser., 14, pp. 157-162. 3 figures. I92ib. The connections of the vomeronasal nerve, accessory olfactory bulb and amygdala in amphibia. Jour. Comp. Neurol., 33, pp. 213-280. 1922. Functional factors in the morphology of the forebrain of fishes. Libro en honor de S. Ramon y Cajal, Madrid, pp. 143-244. 19243. The nucleus olfactorius anterior of the oppossum. Jour. Comp. Neurol., 37, PP- 3I7-359- I924b. Neurological foundations of animal behavior. New York. VAN H6EVELL, J. J. L. D. 1916. The phylogenetic development of the cerebellar nuclei. K. Akad. van Wetenschappen, Amsterdam, 18, pp. 1421-1434. JOHNSTON, J. B. 1913. The morphology of the septum, hippocampus, and pallial commissures in reptiles and mammals. Jour. Comp. Neurol., 23, pp. 371-478. 1915. Cell masses in the forebrain of the turtle, Cistudo Carolina. Jour. Comp. Neurol., 25, pp. 393-468. 1923. Further contributions to the evolution of the forebrain. Jour. Comp. Neurol., 35, pp. 337-48i. KAPPERS, C. U. ARIENS 1917. Further contributions to neurobiotaxis. IX. An attempt to compare the phenomena of neurobiotaxis with the phenomena of taxis and tropism. The dynamic polarization of the neurone. Jour. Comp. Neurol., 27, pp. 261-298. 228 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. 19213. On structural laws in the nervous system. The principles of neurobio- taxis. Brain, 44, pp. 125-129. I92ib. Die vergleichende Anatomic des Nervensystems der Wirbeltiere und des Menschen. II. Abschnitt: Vergleichende Anatomic des Kleinhirns, des Mittel — und des Zwischenhirns und des Vorderhirns. Haarlem. 1923. Le developpement ontogenefigne du corps strie des oiseaux en com- paraison avec celui des mammiferes et de l'homme. Schweizer Arch. F. Neurol. u. Psychiat., 13, pp. 348 — 370. LEVI, G. 19043. Ueber die Entwicklung und Histogenese der Ammonshornformation. Schultze's Arch., 64, pp. 389-404. i9O4b. Sull'origine filogenetica della formazione ammonica. Arch. ital. di anat. e embriol., 3, pp. 234-247. LIVINI, G. 1908. II proencefalo di un marsupiale. Arch. ital. di anat. e embriol., 6, pp. 549-584. McCOTTER, R. E. 1912. The connections of the votneronasal nerve and the accessory olfactory bulb in the opossum and other mammals. Anat. Rec., 6, pp. 299-318. OBENCHAIN, JEANNETTE B. 19233. The brain of Caenolestes obscurus. Anat. Rec., 25, p. 145. 192313. The brain of Caenolestes obscurus. Trans. Illinois State Academy of Science, 16, pp. 100-106. OSGOOD, W. H. 1921. A monographic study of the American marsupial, Caenolestes. Field Mus. of Nat. Hist., Pub. 207, Zool. Ser., 14, pp. 1-156. RAMON y CAJAL, P. 1917-19. Nuevo estudio del encefalo de los reptiles. Trabajos del Lab. de Investigaciones biol. de la Univ. de Madrid, 15, pp. 83-99; J6, pp. 309-333. 1922 El cerebro de los batracios. Libro en honor de S. Ramon y Cajal, Madrid, pp. 13-58. RAMON Y CAJAL, S. 1911. Histologie du systeme nerveux de l'homme et des vertebres, tome II., Paris. RlO-HoRTEGA, P. DEL 1919. Particularides histologicas de la fascia dentata de algunos mamiferos. Trabajos del Lab. de Investigaciones de la Univ. de Madrid, 16, pp. 291-308. ROTHIG, P. 1910. Riechbahnen, Septum, und Thalamus bei Didelphys marsupialis. Abh. Senckenb. Nat. Ges. Frankfurt a. M., 31, pp. 1-19. SMITH, G. ELLIOT 1894. A preliminary communication upon the cerebral commissures of mam- malia, with especial reference to the Monotremata and Marsupialia. Proc. Linn. Soc. New. So. Wales, g, pp. 635-656. 18953. The cerebrum of the marsupial mole, Notoryctes typhlops. Zool Anz., 1 8, pp. 480-482. 1895!). The comparative anatomy of the cerebrum of Notoryctes typhlops. Trans. Roy. Soc., So. Australia, 19, pp. 167-193. 18963. The brain of foetal Ornithorhynchus. Quart. Jour. Micr. Soc., 39, pp. 191-206. i896b. Structure of the cerebral hemisphere of Ornithorhynchus. Jour. An3t. and Physiol., 30, pp. 463-487. i896c. The fascia dentata. Anat. Anz., 12, pp. 119-126. i896d. The morphology of the limbic lobe, corpus striatum, the reptilian septum pellucidum, and the fornix. Jour. Anat. and Physiol., 30, pp. 157-167, and pp. 185-203. 18973. The morphology of the indusium and the striae Lancisii. Anat. Anz., 13, PP. 23-27. 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 229 18975. The origin of the corpus callosum; a comparative study of the hippo- campal region of the cerebrum of Marsupialia and certain Cheiroptera. Trans. Linn. Soc. Lond., 7, pp. 47-69. i897c. Further observations upon the fornix, with especial reference to the brain of Nyctophilus. Jour. Anat. and Physiol., 32, pp. 231-246. :897d. The fornix superior. Jour. Anat. and Physiol., 31, pp. 80-92. 18976. The relation of the fornix to the margin of the cerebral cortex. Jour. Anat. and Physiol., 32, pp. 23-58. 1899. The brain in the Edentata. Trans. Linn. Soc. Lond., 2nd ser., Zool., 7, pp. 277-394. i9o2a. Descriptive and illustrated catalogue of the physiological series of com- parative anatomy contained in the Museum of the Royal College of Surgeons of England. Vol. 2. London. 2nd ed. i9O2b. The primary subdivision of the mammalian cerebellum. Jour. Anat and Physiol., 36, pp. 381-385. I9O2C. Notes on the brain of Macrosc elides and other Insectivora. Jour. Linn. Soc., Zool. Ser., 28, pp. 443-448. iox)2d. On a peculiarity of the cerebral commissures in certain Marsupialia not hitherto recognised as a distinctive feature of the Diprotodontia. Proc. Roy. Soc., 70, pp. 226-231. 19033. On the so-called "gyrus hippocampi". Jour. Anat. and Physiol., 37, pp. 324-328. i903b. On the morphology of the cerebral hemisphere in the vertebrata, with especial reference to an aberrant commissure found in the brains of certain reptiles. Trans. Linn. Soc. Lond., 2nd Zool. Ser., 8, pp. 455-500. I9O3C. On the morphology of the brain of the mammalia, with especial refer- ence to that of the Lemurs, recent and extinct. Trans. Linn. Soc. Lond., 8, part 10, pp. 319-432. i9X>3d. Notes on the morphology of the cerebellum. Jour. Anat. and Physiol., 37, PP- 329-332. 19036. Further observations on the natural mode of subdivision of the mam- malian cerebellum. Anat. Anz., 23, pp. 368-384. 19193. A preliminary note on the morphology of the corpus striatum and the origin of the neopallium. Jour. Anat., 53, pp. 272-291. I9i9b. A note on Professor Landau's memoir on "The comparative anatomy of the nucleus amygdalae, the claustrum and the insular cortex." Jour. Anat., 53, PP- 361-362. VOLSCH, M. 1906. Zur vergleichenden Anatomic des Mandelkern und seiner Nachbar- gebilde. Anat. Anz., 68, pp. 573-683; ?6, PP- 373-523- DE VRIES, E. 1910. Das Corpus Striatum der Saugetiere. Anat. Anz., 37, pp. 385-405. WINKLER, C. and POTTER, ADA 1911. An anatomical guide to experimental researches on the rabbit's brain. Amsterdam. 1914. An anatomical guide to experimental researches on the cat's brain, Amsterdam. 230 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. ABBREVIATIONS al. a. In. amg. a. prcom. a. pt. aq. b. b. ac. b.ol. b. ol. ac. c. (Fig. 8) c. c. a. c. am. cap. e. cap. em. cap. i. ch., ch. op. ci. am. ci. lim. clau. c. c. c.d. c. g. 1. c. g. m. col. a. col. p. cort. cul. dec. pyr. deep nuc. em. nat. ex. am. ex. am. iven. ex. am. xven. ex. dent, ex. hip. a. ex. pir. ex. pir. a. alveus area lunata cerebelli nucleus amygdalae ; amygdaloid complex area precommissuralis area pteroidea cere- belli aqueduct of Sylvius nucleus basalis amyg- dalae nucleus basalis acces- sorius amygdalae bulbus olfactorius bulbus olfactorius ac- cessorius brachium conjucti- vum nucleus centralis amygdadae commissura anterior sive ventralis ; see v. cornu ammonis capsula externa capsula extrema capsula interna chiasma opticum cingulum ammonis cingulum limitans claustrum corpus callosum commissura dorsalis ; see d., ps. d., ps.v. corpus geniculatum laterale corpus geniculatum mediale colliculus anterior colliculus posterior nucleus c o r t i c a 1 i s amygdalae culmen, pars culmina- ta cerebelli decussatio pyramida- lis deep nuclei of cere- bellum (undivided) eminentia natiformis cortex ammonis cortex ammonis intra- ventricularis cortex ammonis extra- ventricularis cortex dentatus cortex hippocampalis anterior (precom- missuralis) cortex piriformis cortex piriformis an- terior ex. pir. m. cortex piriformis me- dialis ex. pir. p. cortex piriformis pos- terior d.b. diagonal band of Bro- ca d.v. diverticulum ventralis (of superior re- cess) f ila ol. fila olf actor ia fim. fimbria floe. flocculus f . med. t. fasciculus medialis te- lencephali (medial forebrain bundle) f . med. 1. 1. fasciculus medialis te- lencephali lateralis (lateral limb of m. f. b.) form. bul. formatio bulbaris (ol- factorius) f. prcom. fornix praecommis- suralis f s. amg. fissura amygdaloidea f s. amg. m. fissura amygdaloidea medialis fs. circ. fissura circularis fs. ch. fissura choroidea f s. di-tel. fissura di-telencepha- lica f s. erh. fissura endorhinalis fs. f . d. fissura fimbrio-dentata f s. fim. al. fissura fimbrio-alvea- ris f s. fim. d. fissura fimbrio-dentata f s. h., f s. hip. fissura hippocampi fs. pin. fissura postlunata f s. pnod. fissura postnodularis f s. orb. fissura orbitalis f s. pr. fissura prima cerebelli f s. prcul. fissura preculminata f s. rh. fissura rhinalis f s. rh. a. fissura rhinalis ante- f s. rh. arc. f s. rh. m. f s. s. f s. spyr. fs. tr. cb. g. bas. op. g-S gl. p., glob. p. gy. dent. nor fissura rhinalis arcua- ta fissura rhinalis me- dialis fissura secunda cere- belli fissura suprapyrami- dalis fissura transversa ce- rebelli ganglion basale op- ticum ganglion semilunaris globus pallidus gyrus dentatus 1925. BRAINS OF CAENOLESTES AND OROLESTES — OBENCHAIN. 231 hab. hip. a. i.C. inf. interpos. 1. interpos. m. int. plate ip. m. 1. lob. a. cb. lob. m. cb. lob. p. cb. 1. perf . a. m. med. m. i. neop. n. 2 n. 5 nod. n. t. o. 1. nuc. ac. nuc. amg. b. nuc. amg. b. ac. nuc. amg. c. nuc. amg. cort. nuc. amg. 1. nuc. amg. m. nuc. caud. nuc. d. b. nuc. ip. nuc. ol. ant. nuc. ol. ant. d. nuc. ol. ant. ex. nuc. ol ant. 1. habenula hippocampus anterior ( praecommissuralis ) island of Calleja infundibulum interpositio lateralis interpositio medialis intercalary plate of Johnston (1923) interpositio medialis nucleus lateralis amygdalae lingula cerebelli lobus anterior cere- belli lobus medialis cere- belli lobus posterior cere- belli locus perforatus ante- rior nucleus medialis amygdalae medulla oblongata massa intermedia neopallium nervus opticus nervus trigeminus nodulus cerebelli nucleus tracti olfacto- rii lateralis nucleus accumbens nucleus basalis amyg- dalae nucleus basalis acces- sorius amygdalae nucleus centralis amygdalae nucleus corticalis amygdalae nucleus lateralis amygdalae nucleus medialis amygdalae nucleus caudatus nucleus of diagonal band of Broca nucleus interpeduncu- laris nucleus olfactorius anterior (Her rick) nucleus olfactorius anterior, pars dorsa- lis nucleus olfactorius anterior, pars exter- nus nucleus olfactorius anterior, pars late- ralis nuc. ol. ant. 1. v nuc. ol. ant. m. nuc. ol. ant. p. nuc. pol. 1. nuc. pol. m. nuc. tr. ol. 1. ped. floe. pfloc. pi. ch. 3 (4) p. prcul. p. spyr. put. pyr. r. rec. i. rec. s rad. thai. st. glom. st. gr. st. m. c. st. med.. st. t. st. 1. 1 st. t. 2 st. t. 3 st. t. 4 st. t. 5 st. t. bed tect. nucleus olfactorius anterior, pars late- ro-ventralis nucleus olfactorius anterior, pars me- dialis nucleus olfactorius anterior, pars pos- terior nucleus parolfactorius lateralis nucleus parolfactorius medialis nucleus tracti olfacto- rii lateralis; see n. t. o. 1. pedunculus flocculi paraflocculus plexus choroidea ven- triculi tertii (quar- ti) pars preculminata ce- rebelli pars suprapyramidalis cerebelli putamen pyramis cerebelli corpus restiforme recessus inferior recessus superior radiatio thalamica stratum glomeruli stratum granulare stratum cellularum mi- tralium stria medullaris, see st. *: 5 stria terminalis stria terminalis, bundle i (Johnston, iQ23)=com. bundle stria terminalis, bundle 2 (Johnston, iQ23)=ol. proj. tr. (amg.) stria terminalis, bundle 3 (Johnston, 1923) =subcom. bundle stria terminalis, bundle 4 (Johnston, 1923) = supracom. bundle stria terminalis, bundle 5 (Johnston, 1923) =st. med. bundle stria terminalis bed tectum mesencephali 232 FIELD MUSEUM OF NATURAL HISTORY — ZOOLOGY, VOL. XIV. t.ol. tr. ol. tr. ol. d. m. tr. ol. f r. tr. ol. i. tr. ol. 1. tr. ol. 1. d. tr. ol. 1. p. d. tr. ol. 1. p. v. tr. ol. 1. v. tr. ol. m. tr. ol. s. + c. tr. op. tr. st. thai. tuberculum rium olfacto- tr. t. a. tractus olfactorius tractus olfactorius dorso-medialis tractus olfacto-fronta- lis tractus olfactorius in- termedius (commis- suralis) tractus olfactorius la- teralis (massive part) tractus olfactorius la- teralis, pars dorsa- lis tractus olfactorius la- teralis, pedunculus dor sal is tractus olfactorius la- teralis, pedunculus ventralis tractus olfactorius la- teralis, pars ventra- lis tractus olfactorius me- dialis tractus olfacto-septa- lis et corticalis tractus opticus tractus strio-thalami- cus t. v. 4 uv. v. cb. v. 1. v. m. a. v. ol. v. 3 (4) xven. alv. xven. ex. am. tractus temporo-am- monis of C a j a 1 (1911; sphenotemp., 1906) ; see ci. lim. (cingulum limitans) and ci. am. (cingu- lum ammonis) taenia ventriculi quarti uvula cerebelli ventriculus cerebelli ventriculus lateralis velum medullare ante- rius ventriculus olfactorius ventriculus t e r t i u s (quartus) extraventricular alveus extraventricular cor- tex ammonis line of fracture sup- posed to mark boundary between extra- and intraven- tricular precommis- sural ammon's cor- tex. arrows to show direc- tion and approxi- mate degree of force operating in the formation of the definitive mamma- lian "hippocampal figure", figure u line of amputation of cortex dentatus, fig- ure I7a line of amputation of cortex ammonis, figure I7b FIELD MUSEUM OF NATURAL HISTORY. ZOOLOGY, VOL. XIV, PL. XXIV. fs'.rh. fs. s.' OROLESTES INCA. DORSAL VIEW OF BRAIN. Five times natural size. FIELD MUSEUM OF NATURAL HISTORY. ZOOLOGY, VOL. XIV, PL. XXV. •tr. ol.lat. ~.fs.rh.arc. erh. fs. amg. OROLESTES INCA. VENTRAL VIEW OF BRAIN. Five times natural size. FIELD MUSEUM OF NATURAL HISTORY. ZOOLOGY, VOL. XIV, PL. XXVI. f,5. orb. -. Vs. erh. \\[''tr. ol.lat. . \ fs. rfi. arc. tub. ol. fs prcul. . fS.pnod v. cb. OROLESTES INCA. FIG. 3. LATERAL VIEW OF BRAIN. FIG. 4. MEDIAN SECTION OF BRAIN. Five times natural size. FIELD MUSEUM OF NATURAL HISTORY. ZOOLOGY, VOL. XIV, PL. XXVII. hip. a. MEDIAN SURFACE OF HEMISPHERE. FIG. 5. OROLESTES INCA. FIG. 6. CAENOLESTES OBSCURUS. RECONSTRUCTION. Five times natural size. FIELD MUSEUM OF NATURAL HISTORY. ZOOLOGY, VOL. XIV, PL. XXVIII. prcul Tned. r > V^_r/.v.c NOTORYCTES (Elliot Smith) 10 a 10. MEDIAL FIGS. 7 TO 10. CAENOLESTES OBSCURUS. CEREBELLUM. RECONSTRUCTIONS. X 8. FIGS. 10a-10b. CEREBELLA OF NOTORYCTES AND PERAMELES. FIG. 11. DEVELOPMENT OF MAMMALIAN "HIPPOCAMPAL FIGURE. " o o " a FIELD MUSEUM OF NATURAL HISTORY. ZOOLOGY, VOL. XIV, PL. XXX. interpos. lat. (sxibic.) fs.fm.d inierpos. m 0 (1560) '" :' :!:V-:: &% gy. d«nt. \ " !•••• .xven. c. OTTI C. am _= ::'~ ^ \ ! ive-n. c. am. ; *.• .~r I'll '' * £ gy. dent \ $*<:•;:. ""%£ \ % ^ ^.••'"•eCI380> a-g,5treeter, Op. C. ioi a b c de fg ...X. OTA. FIG. 17. CAENOLESTES OBSCURUS. DISSECTED DIAGRAM OF HIPPOCAMPUS. X 12. FIG. -8. DIDELPHIS VIRGINIANA. TEMPORAL HIPPOCAMPUS. X 5. FIELD MUSEUM OF NATURAL HISTORY. ZOOLOGY, VOL. XIV, PL. XXXI. cx am ex. dent. s.rk. M4 & CX.deni... fs.arng. m.^T 22 TEMPORAL RECURVING OF HIPPOCAMPUS. FIG. 19. HYPSIPRYMNUS (LiviNi) X 3. FIG. 20. RABBIT (WlNKLER-PoiTER). FIG. 21. FELIS LEO (ELLIOT-SMITH) X Va. FIG. 22. NOTORYCTES (DART). DIVERSITY OF Illinois IIBBAH> FIELD MUSEUM OF NATURAL HISTORY. ZOOLOGY, VOL. XIV, PL. XXXII. b. ol. ac trollpd... nucolantl troll.. troll st.gr st me., st. pi ex. •^ : :ir trollp.w laol. trollp.d nuc. ot ant. d troll. nuc.olant.6K. •nuc. ol ant. I troliU.a).. tr.ollpv.. neop. 23 (SEC. 243) 24 (SEC. ZT1) t\uc ol.ant. d ex. pit. a fs.rh.. trolld':. fs. circ.. tr ol. I... •nuc. ol.ant. 1. Txuc.olant ex.. c a . tr ol t pi v nuc. olant.d fs. erh. troll nuc.ol.ont.lv fs.circ. C.«L-- , nucolant.ex \Jonn.buL 26(StC 445) ex. pir. a trolld- fs erh.. trol'l... ex. pirLv, nuc.ol ant. p. nuc ol anf ex fs.circ or. hip. a form bul 27 (5tC 3B2) a. hi p. a. trols+c. fs.circ. .js. circ. t.ol U.C.) nucolontp. Vtrolm. /orm. buL 28 (SEC 4-10) CAENOLESTES OBSCURUS. x s. FIELD MUSEUM OF NATURAL HISTORY. ZOOLOGY, VOL. XIV, PL. XXXIII. , cor rod nuc. caud. ex. pir a. trol. Id. fs erK troll..' ex. pir v. js rK.arc fmedll,- trolly, rh'arcV ex. pir v..\..' .•nue.olar.tp. fs-JJJ fs.rkarc t.ol. (i.C; (+ giant cell n ilol. li.C.) 29 (SEC4J2) 30 (Sec. sio) ci. lim. ex. pir a. trol.l.d. ex. pir v... fs. erh trol.C. fs..Th.arc f.7ned.tl..- tn ft thai 31 (SEC. 542) (Sec. CAENOLESTES OBSCURUS. x s. DIVERSITY OF RLWOISUBBW FIELD MUSEUM OF NATURAL HISTORY. ZOOLOGY, VOL. XIV, PL. XXXIV. C'L llm. nuc.caadL cor. rod cap. cap.i put. cap.e. cx.piTTn.lL..I - \ glob. p..V "- ttoUi. «.plTV.> frnelt . ' jfs. di-tel. )s. Th.arc. 33. (Sec. S9S) 34: (SEC. 642) COT. rad:. nuc.ca- clau l.ci.lim, .exam ci.am. ;x.dent. .fs.hip. i , .-ittb. WWrwL cor rad fs.rK cx.pir.m./..4 nuc.caud int plate cap. e. . nuc.ama.l nuc.amg.c. nucamg.m I perf a. is. amg.- nuc.tr ol. I. fs.Tti.a ,. bas.op. -tel 35. CSEC.668) 36 (5EC.ro4) CAENOLESTES OBSCURUS. X 8. FIELD MUSEUM OF NATURAL HISTORY. ZOOLOGY, VOL. XIV, PL. XXXV. cor rod frol.l.d.. •nuc. caul cx.plr m Tiuc.amg. I. .. em.nat: nuc. amg.b. cap. e . I. perf. a, nuc.tr.ol. I 37 (SEC.T30) 38 (Sec. 760) i.lim. ci.am. cx.pirm etn.natl '-., nuc.aTng l./X-.'.': •nuc. ati\fl. b. true, amg.b. ac. (s. aing 4-0 (SEC.83Z.) CAENOLESTES OBSCURUS. x 8. FIELD MUSEUM OF NATURAL HISTORY. ZOOLOGY, VOL. XIV, PL. XXXVI. Job. m.cb. (p.spyr) P- Krh. nuc.o-mg.b.ai f s. amg nuc.amg. cort. 41 (SEC. 851) 42 (SEC 879) fs. pr fs.tr cb. a.pt.cb. .p.spirllob.m.cb.) pspyr.cb..^< {s. prcul. fs.tr cb... a. PIT p. 43 (Stc. 900) 44 (SEC.920) CAENOLESTES OBSCURUS. x s. UNIVERSITY OF ILLINOIS-URBANA