Book - The brain of the tiger salamander 11

From Embryology
Revision as of 15:29, 28 June 2018 by Z8600021 (talk | contribs)
Embryology - 28 Mar 2024    Facebook link Pinterest link Twitter link  Expand to Translate  
Google Translate - select your language from the list shown below (this will open a new external page)

العربية | català | 中文 | 中國傳統的 | français | Deutsche | עִברִית | हिंदी | bahasa Indonesia | italiano | 日本語 | 한국어 | မြန်မာ | Pilipino | Polskie | português | ਪੰਜਾਬੀ ਦੇ | Română | русский | Español | Swahili | Svensk | ไทย | Türkçe | اردو | ייִדיש | Tiếng Việt    These external translations are automated and may not be accurate. (More? About Translations)

Herrick CJ. The Brain of the Tiger Salamander (1948) The University Of Chicago Press, Chicago, Illinois.

Historic Disclaimer - information about historic embryology pages 
Mark Hill.jpg
Pages where the terms "Historic" (textbooks, papers, people, recommendations) appear on this site, and sections within pages where this disclaimer appears, indicate that the content and scientific understanding are specific to the time of publication. This means that while some scientific descriptions are still accurate, the terminology and interpretation of the developmental mechanisms reflect the understanding at the time of original publication and those of the preceding periods, these terms, interpretations and recommendations may not reflect our current scientific understanding.     (More? Embryology History | Historic Embryology Papers)

Part I. General Description and Interpretation 1. Salamander Brains | 2. Form and Brain Subdivisions | 3. Histological Structure | 4. Regional Analysis | 5. Functional Analysis, Central and Peripheral | 6. Physiological Interpretations | VII. The Origin and Significance of Cerebral Cortex | VIII. General Principles of Morphogenesis Part 2. Survey of Internal Structure 9. Spinal Cord and Bulbo-spinal Junction | 10. Cranial Nerves | 11. Medulla Oblongata | 12. Cerebellum | 13. Isthmus | 14. Interpeduncular Nucleus | 15. Midbrain | 16. Optic and Visual-motor Systems | 17. Diencephalon | 18. Habenula and Connections | 19. Cerebral Hemispheres | 20. Systems of Fibers | 21. Commissures | Bibliography | Illustrations | salamander

Chapter XI Medulla Oblongata

THERE is no systematic description of the medulla oblongata of adult Amblystoma comparable with my paper on Necturus ('30) . Coghill's papers give many details of the early stages of development, but no comprehensive description has been written. The 38mm. larva was described in 1914, and subsequent study convinces me that the conditions found there and in Necturus are not fundamentally different in adult Amblystoma, though the latter is considerably more specialized. This is evident upon comparison of the figures of adult Amblystoma recently published ('446) with those of the two papers just cited, to which the reader is referred for general discussion. At metamorphosis the changes are far less in Amblystoma than in Salamandra and some other urodeles. Descriptions of the medulla oblongata and cerebellum of Cryptobranchus, Siren, Proteus, Salamandra, and some other urodeles are found in the papers cited on page 11. Some details of the schematic outline given in chapters iv and v will now be filled in.


SENSORY ZONE

The sensory zone forms the massive dorsolateral wall of the fourth ventricle from the calamus scriptorius forward, to and including the thick auricle under the cerebellum (p. 44). This zone is reduced in size but not suppressed in the floor of the lateral recess of the ventricle spinal ward of the auricle. In the remainder of the medulla oblongata the gray substance forms a low ridge bordering the taenia of the fourth ventricle and projecting into the ventricle, the acousticolateral area, which is smaller in the adult than in the larva (figs. 9, 89, 90). Ventrally of this area is a narrower ridge, which expands posteriorly — the visceral lobe (figs. 9, 88). This contains the nucleus of the fasciculus solitarius.


The white substance of the sensory zone is composed chiefly of the fascicles of sensory root fibers and the associated neuropil. There are, in addition, secondary arcuate fibers and two longitudinal bundles of correlating fibers — tracts a and h of Kingsbury (figs. 7, 9, 87-90).


The sensory root fibers are related peripherally with end-organs which seem to be physiologically as specific as those of man, though the specificities are of different types; but at the first synapse this specificity is no longer preserved in terms of localized gray centers or pathways of conduction. A neuron of the second order may have functional connection with afferent fibers of tactile, vestibular, lateral-line, or gustatory roots, singly or in any combination.


Since most of the sensory root fibers span the entire length of the medulla oblongata and have numberless collateral branches throughout this length, evidently there is little provision for localization of function in terms of fore-and-aft relations in space. Mass responses of the entire musculature are readily evoked by excitation of any or all components of this sensory complex, but how are selective local responses effected? When the arrangement of the secondary connections of these sensory root fibers is examined, it is evident that this question cannot be answered in terms of any mechanism of switchboard type. The axons of these secondary neurons do not take random courses. They tend to be assembled in bundles of fibers, which have a common direction and destination. Such bundles as ascend to higher levels (the lemniscus systems) evidently have some kind of specificity, for they retain their identity up to termination in definite and different places. This specificity may be determined by the source of their excitation — the modality of sense and the location in the body of the receptive end-organs — or it may be determined by the nature of the response to be evoked — muscular contraction, secretion, and the location of the effector organs activated. Probably both these factors are operative in establishing the pattern of arrangement of the higher centers of adjustment and their connecting tracts of fibers.


The problem of the apparatus employed in making discriminative responses to specific types of sensory excitation is not simplified by these observations. In higher animals the specificity of the various modalities of sense is preserved in central pathways of conduction leading up to higher centers of adjustment in the thalamus and the cerebral cortex. But somewhere in the course of this transmission these diverse, localized, functional systems are brought into relation with one another and are integrated in such a way as to result in appropriate responses to the total situation. In these amphibians the process of integration begins at the first synapse. The organization of the neurons of the second order is such as to facilitate activation of


MEDULLA OBLONGATA 155

large masses of musculature by sensory excitation of any kind, that is, reactions of total-pattern type as defined by Coghill. This is the only kind of response which can be made by Amblystoma in the earliest stages of development of motility ; and in the adult animal it is clear, from both the structural organization of the nervous system and the overt behavior, that the action system as a whole is predominantly of total-pattern type. One of the coarser features of the structural mechanism of this integrated behavior is seen in the waj^ the neurons of the second order act as collectors of excitations of diverse modalities and widely distributed peripheral origin.


Within this primordial apparatus of integrated mass action, as development advances, local areas are differentiated with progressively more restricted functional connections. This is manifested anatomically by contraction of the spread of the dendrites of the secondary elements so as to have synaptic contact with a smaller number of fascicles of root fibers. In midlarval stages some of these elements spread their dendrites through the whole extent of the white substance of the sensory zone, thus having contact with terminals of all fascicles of root fibers (fig. 9, no. 1). Some of these dendrites also reach downward into the zone of correlation, the reticular formation (fig. 9, no. 4). Other elements are in functional relation with a few or only one of the root fascicles (fig. 9, nos. 2, 3). In the adult the dendritic spread is not so wide, and a larger proportion of the neurons are in synaptic relation with one or a few contiguous fascicles of root fibers; yet even here there are few neurons of the second order which may be activated by root fibers serving a single modality of sense.


There is, accordingly, as development advances a progressive restriction and local specialization of the central apparatus of analysis of the manifold of sensory experience, and this is an important factor in the acquisition of the local reflexes which are individuated within the larger total patterns of behavior as this process has been described by Coghill. Furthermore, this process of local specialization of structure within the sensory field, as exhibited in ontogeny, has a parallel in phylogenetic history, as illustrated, for instance, by comparison of the organization of this field of Amblystoma with that of the frog, as described by Larsell ('34, p. 504), and with that of the specialized fishes ('44&).


In the tracts of fibers which ascend to higher levels of the brain stem from this sensory field — the lemniscus systems as described below — there is a similar lack of segregation of separate pathways for the several modalities of sense. The general bulbar lemniscus {Im. of the figures) carries fibers which may be activated from any or all of the sensory nerve roots. The segregation of some specific sensory pathways is incipient in the lemniscus systems, and this is seen with especial clarity in the secondary ascending visceral-gustatory path (tr.v.a.).


INTERMEDIATE ZONE

The reticular formation of the spinal cord is continued forward into the medulla oblongata (fig. 9; '14a, p. 378; '30, p. 59). In Weigert sections (figs. 89, 90) it may be recognized as a lighter field between the bulbar lemniscus (Im.) and the spinal lemniscus (Im.sp.). Here the spino-bulbar, spino-cerebellar, and spinal lemniscus tracts are assembled in the alba.


The neuropil of this field receives dendrites from the underlying gray and from the adjoining sensory and motor zones (fig. 9; '14a, figs. 22-43; '446, figs. 7, 8). Some of these cells are small, with local connections; others are very large, with wide dendritic spread and long axons (most of them myelinated), which descend as internal arcuate fibers to the ventral funiculi. Many of them decussate in the ventral commissure; and both crossed and uncrossed fibers, before or after crossing, bifurcate into ascending and descending branches. The descending bulbo-spinal fibers and the ascending bulbo-tegmental fibers are part of the neuromotor apparatus of bulbar and spinal reflexes. The lemniscus systems, on the other hand, arise chiefly in the sensory zone, though some of their fibers may be axons of neurons of the intermediate zone, a distinction of no great significance, in any event, because many neurons have extensive dendritic spread in both zones. The lemniscus systems also differ from the neuromotor apparatus just described, in their terminal connections, passing not to the motor zone but to higher levels of the sensory zone.


Among the large cells of the intermediate zone the two giant cells of Mauthner are of special interest. They lie at the level of the vestibular roots and have large dendrites directed outward among terminals of these root fibers ('14a, fig. 12). In the 12-mm. larva they are of enormous size ('14a, fig. 53), with thick dendrites, which ramify throughout the entire area of the white substance. In the adult the dendritic spread is less extensive, yet it is sufficiently wide to embrace the greater part of the alba of the intermediate and motor zones and especially the terminal area of vestibular fibers. The thick


MEDULLA OBLONGATA 157

myelinated axons immediately decussate and descend in the f . longitudinalis medialis tlirough the entire length of the spinal cord. These two cells are collectors of nervous impulses from many sources and are part of the apparatus of control of mass movements of the musculature of the trunk. Their specific functions have not been fully explained, though they have been much studied. They are magnified and highly specialized examples of the elements of the nucleus motorius tegmenti of the medulla oblongata.


We lack sufficient knowledge of the structural and physiological properties of this tegmentum to frame satisfactory hypotheses of the actual mechanism of the integrative and co-ordinating functions which seem to be operating here. The intermediate tegmentum is structurally so intimately interwoven with the tegmentum of the motor zone that both probably act in these operations as a functional unit.


At the anterior end of the medulla oblongata the trigeminal tegmentum merges with the isthmic tegmentum, and dorsally of this junction the nucleus cerebelli occupies the position of the intermediate zone. In higher animals the gray of this nucleus is incorporated within the body of the cerebellum as the deep nuclei. The definitive cerebellum, accordingly, is a derivative of both sensory and intermediate zones.


MOTOR ZONE

Little can be added here to the general description in chapter v. The larger cells of the motor field may spread their dendrites through the whole of the motor and intermediate zones and also upward into the sensory zone and across the ventral commissure to the opposite side ('14a, figs. 29, 30, 31, 37, 41, 42; '446, figs. 7, 8). This type of structure is extended forward into the isthmus, where it is more specialized and the small and large cells are segregated, though not completely so. Most axons of the large cells are myelinated in the adult. In midlarval and late larval stages, thick unmyelinated axons of the large cells take curious courses. Some descend uncrossed in the ventral or ventrolateral funiculi. Others decussate in the ventral commissure. The latter may cross transversely as internal or external arcuate fibers, or they may take long ascending or descending courses before crossing obliquely and then turning spinal ward. Before or after crossing, these fibers may divide into long ascending and descending branches, and they may give long branched collaterals along the entire course. The terminals of an individual fiber may reach a large part of the motor field of the medulla oblongata of both sides. How far forward the ascending branches may extend has not been determined; some of them certainly pass beyond the isthmus. Some illustrations of these fibers have been drawn ('396, figs. 42, 46, 51, 57, 58, 61, 66, 68). In the medulla oblongata they are mingled with similar fibers of the thalamo-bulbar, pedunculo-bulbar, and tegmento-bulbar systems ('396, fig. 23). Relatively few of them enter the f. longitudinalis medialis. The extensive and diffuse spread of the terminals of these fibers makes it evident that, if they are employed in any definitely local reactions, the localization is effected by some device other than the arrangement in space of their terminals.


In mammalian neurology the transverse segment of the brain stem, which lies under the cerebellum, is named the "pons" after its most conspicuous external feature. This name is obviously inappropriate here, where there is no pons or any recognizable primordium of it.


FIBER TRACTS OF THE MEDULLA OBLONGATA

The white substance of the medulla oblongata contains many compactly arranged, well-myelinated fibers. Some are very coarse and some thinner. Mingled with these, especially in the sensory and intermediate zones, there are many unmyelinated fibers. In most preparations the several systems are not clearly fasciculated, so that analysis is difficult in the adult. In larvae elective Golgi preparations have revealed many details, though much remains obscure. Most of the ventromedial fibers are descending. Laterally within and surrounding the reticular formation are fibers of local correlation and the ascending lemniscus systems; and dorsolaterally there are the fascicles of sensory root fibers, together with some correlating systems. Most of the latter appear as arcuate fibers, which are very numerous at the outer border of the gray, with some at intermediate and superficial depths of the alba (figs. 87-90).


Sensory fibers of the second order are of five sorts: (1) reflex connections by arcuate fibers with the bulbar motor zone of the same and the opposite side, either directly or with synapse in the intermediate zone; (2) bulbo-spinal connections, crossed and uncrossed, descending in the ventral and ventrolateral funiculi; (3) bulbo-cerebellar connections (see chap, xii); (4) correlation tracts a and h (of Kingsbury), intrinsic to the sensory zone; (5) the lemniscus systems, passing between the sensory zone and higher levels of the same zone, most of


MEDULLA OBLONGATA 1.59

them decussating in the ventral commissure. The fibers of all these kinds are widely dispersed, and their analysis is still incomplete. Arcuate fibers are everywhere present in the medulla oblongata, and each of the five groups of secondary connections mentioned above is represented in these arcuates. There are also deep arcuates from the intermediate and motor zones to the ventral commissure. These are doubtless part of the apparatus by which the patterns of performance of synergic groups of muscles are organized, but the details of actual operation of this apparatus are unknown. It seems to be well established that the primary organization of these motor patterns is intrinsic to the motor and intermediate zones. These intrinsic mechanisms may be activated (1) in a nonspecific way from the sensory zone, with resulting mass movement; (2) more specifically from parts of the sensory field dominated by one or another system of peripheral sense organs, with local motor response; (3) from higher centers of control which co-ordinate all motor activity in the interest of integrated behavior of the body as a whole. The second and third of these types of behavior are recognizably present in Ambly stoma, though at a very primitive level of specialization.


SENSORY CORRELATION TRACTS a AND b

All peripheral somatic sensory fibers which enter the medulla oblongata terminate in a common pool of neuropil, which pervades the sensory zone within which the various qualities of sense seem to be more or less completely merged. The secondary fibers which leave these fields are, however, not strictly equipotential. There is evidently an incipient localization of function among these fibers, but the apparatus by which this specialization is achieved is not clear. It seems probable that part of this apparatus is to be sought in two longitudinal tracts of correlation, which were first described by Kingsbury ('95) in Necturus. To these he gave noncommittal names, "tracts a and 6," one lying above, the other below, the fascicles of lateral-line nerve roots (figs. 9, 87-90). I have written a general description of these tracts of Amblystoma ('446) and, in particular, of their relations posteriorly with the dorsal funiculus of the spinal cord and the nucleus funiculi and anteriorly with the cerebellum. Some additional features may now be added.


Tract a. — This tract follows the taenia of the fourth ventricle for most of its length, anteriorly passing below and through the dorsal island of neuropil, under the floor of the lateral recess of the fourth


160 THE BRAIN OF THEiTIGER SALAMANDER

ventricle, and then arborizing in the lateral vestibular area of the auricle (figs. 7, 32, 33; '446, fig. 14). Posteriorly, at the calamus scriptorius, it converges with its opposite fellow, and the two tracts enter, respectively, the medial fascicles of the dorsal funiculi of the spinal cord. Tract a is regarded as a mixed collection of fibers of correlation, related primarily with the lateral-line nerve roots, a supposition which is supported by Kreht's observation ('30, p. 316; '31, p. 422) that in adult Salamandra, in which the lateral-line roots are atrophied, tract a also disappears. It is connected by arcuate fibers with the intermediate and motor zones of both sides, and probably some of these fibers are commissural, connecting the acousticolateral areas of the two sides.


The nucleus of the dorsal funiculi extends for a considerable distance anteriorly and posteriorly of the calamus scriptorius. Fibers of the slender median fascicle of the dorsal funiculus of the cord terminate in this nucleus, and some of them continue forward for an undetermined distance in tract a. The nucleus funiculi is connected through tract a with the lateral-line area farther forward by fibers, some of which are probably descending and some ascending. The latter may go as far as the cerebellar primordium in the auricle, though there is no demonstration of this. In Triturus, as described by Larsell ('31, p. 48), unmyelinated fibers descend into tract a from the auricle, thus providing for cerebellar discharge into the lateral-line area.


Trad h. — This tract is a mixed bundle of myelinated and unmyelinated fibers, believed to be related primarily and perhaps specifically with the vestibular roots (figs. 7, 31). The literature contains several descriptions of tract b of Amblystoma ('14a, p. 373; '396, p. 604; '446, p. 317; Larsell, '32) and other urodeles. In the aggregate, these give an incomplete account of it. From the data at hand it is concluded that it contains a variety of crossed and uncrossed, ascending and descending, fibers connecting the vestibular areas of both sides with each other, with the underlying tegmentum, with the cerebellum, and with the spinal cord. Kreht ('31, p. 423) describes in Proteus a probable connection with the f . longitudinalis medialis.


Anteriorly, tract h converges with tract a into the vestibular neuropil at the lateral border of the auricle (fig. 91). This is cerebellar territory, the primordium of the flocculus of mammals. Silver impregnations show that many of these fibers end in free arborizations in the auricle ('396, figs. 43, 67, 77, 98). This connection is clear in


MEDULLA OBLONGATA IGl

Triturus also (Larsell, '31, p. 50) and is comparable with the secondary vestibulo-cerebellar tract of mammals. Fibers descending from the auricle into tracts a and b are probably ancestral to the mammalian f. uncinatus of Russell.


At the bulbo-spinal junction, tract b merges with the spinal root of the vestibular nerve, and its fibers descend for an undetermined distance in the lateral fascicles of the dorsal funiculus of the cord. There is, accordingly, a dorsal vestibulo-spinal connection by both peripheral and secondary fibers, a connection which puts the vestibular apparatus into especially intimate relation with the neuropil of the nucleus funiculi. It is probable that fibers ascend from the nucleus funiculi in tract b, and some of these may extend as far as the auricle. As I have pointed out ('44&), this cerebellar connection, if confirmed, would be the obvious precursor of the system of arcuate cerebellar fibers described in primates by Ferraro and Barrera ('35) as passing from the external cuneate nucleus to the cerebellum by way of the corpus restiforme. Tract b connects only with the vestibular field of the cerebellum, and this cerebellar connection of the nucleus funiculi (if present) would have a physiological significance quite different from that of the larger connection by way of the tractus spinocerebellaris (fig. 3), for the latter connects only with the non vestibular body of the cerebellum.


In addition to the dorsal uncrossed vestibulo-spinal connection just described, there is an extensive crossed and uncrossed connection between tract b and the motor field by arcuate fibers. Many of these fibers bifurcate, with branches which ascend and descend in the ventral funiculus. These may correspond with the secondary vestibular fibers of the mammalian f. longitudinalis medialis, but here they are dispersed, and the details of their courses have not been described. In the absence of elective impregnations, our material does not reveal their courses. Their presence may be expected, in view of the fact that they comprise an important constituent of this fasciculus in mammals. An incomplete impregnation in an advanced larva gives some evidence of them. In figure 38 the contorted fiber directed medially from the region of entrance of the VIII root is probably a secondary vestibular fiber, which enters the f . longitudinalis medialis, and the two impregnated fibers seen in this fasciculus may belong to this system.


The inferior olive has not been identified in urodeles, but among the dispersed arcuate fibers are some with connections which suggest the presence of a primordium of this structure ('396, p. 604). Some of the smaller tegmental neurons in the vicinity of the VIII roots send axons laterally into tract b, where they divide with ascending and descending branches. Similar fibers are seen to enter this tract at all levels between the V and X roots. Some of these come from the opposite side. The ascending branches of these fibers pass under the auricular gray and recurve dorsalward to arborize within the gray of the rostral face of the auricle. This may represent a vestigial remnant of the large olive of some fishes, here reduced to insignificant proportions because of the small size of the cerebellum (Ariens Kappers, Huber, Crosby, '36, pp. 668-89).


THE LEMNISCUS SYSTEMS

The less myelinated field of alba of the medulla oblongata between the sensory and motor zones, which I term the "reticular formation," contains the ascending fibers of the lemniscus systems and a neuropil, which receives dendrites of neurons of the underlying gray and also of many neurons of the adjoining sensory and motor zones. This seems to be the field in which patterns of local bulbar reflexes are organized, but its texture and connections have not been satisfactorily analyzed. From the reticular formation and the overlying sensory gray, thick axons, some of which are myelinated, descend as internal and external arcuate fibers to the ventral funiculus. Many of these pass to the motor zone as part of the neuromotor apparatus of bulbar and spinal reflexes. These enter a mixed spino-bulbar and bulbo-spinal tract associated with the spinal lemniscus, as shown in figures 38 and 39. Others, usually thinner fibers, immediately decussate in the ventral commissure and then ascend in the lemniscus tracts to higher levels of the sensory zone.


In the lemnisci of mammals, fibers of the various sensory systems are segregated in separate tracts. In Ambly stoma the arrangement is radically different. Segregation of functional systems is incipient but so little advanced that mammalian names are inapplicable. Under this heading there are included here all ascending fibers from the spinal cord and medulla oblongata, which terminate in the sensory and intermediate zones at higher levels. These are arranged in four groups which take separate courses: (1) the spinal lemniscus complex (Im.sp.), (2) the general bulbar lemniscus (hi.), (3) tr. bulbo-tectalis lateralis (tr.t.b.L), and (4) the ascending secondary visceral-gustatory tract {tr.v.a.). The spino-cerebellar tract is closely associated with the


MEDULLA OBLONGATA 163

first group. There is some evidence of a separate trigeminal lemniscus, as described below.


1. THE SPINAL LEMNISCUS (fIGS. 3, 9, 10, 11, Im.Sp.)

The spinal lemniscus is an extensive system of fibers ascending from the spinal cord and here lying immediately ventrally of the spinal V root. In the cord these fibers arise as axons of cells of the dorsal gray column, which decussate in the ventral commissure. In the region of the calamus this tract receives extensive additions from the nucleus funiculi, the spinal V nucleus, and the commissural nucleus of the opposite side (figs. 3, 87). Some of these crossed fibers from the calamus region have thinner collaterals, which descend for a short distance in the ventral funiculus of the spinal cord. The ascending fibers, at first, lie in the ventral funiculus (figs. 41, 42) but soon turn dorsally to join those from lower levels of the cord ('446, figs. 9-11). The most lateral and ventral fibers of this mixed bundle terminate in the reticular formation of the medulla oblongata — tr. spino-bulbaris (figs. 3, 88, 89, 90). The others at the level of the V nerve roots separate into spino-cerebellar and spino-tectal tracts. Many of the spino-cerebellar fibers are collaterals of the spino-tectal fibers (fig. 10; '14a, p. 376). Some fibers of the spino-tectal tract continue past the midbrain to end in the dorsal thalamus (fig. 34). The entire course of this complex is shown in a diagram ('396, fig. 26) drawn from three adjoining sagittal sections of a larva, and this is confirmed by elective Golgi impregnations of the adult. Terminals of this tract in the dorsal thalamus are seen in figure 44.


2, THE GENERAL BULBAR LEMNISCUS (hn.)

This large fascicle, the general bulbar lemniscus, receives fibers from all parts of the sensory zone of the medulla oblongata (figs. 9, 11). These decussate in the ventral commissure and ascend medially of the reticular formation (figs. 89, 90). In the isthmus they turn dorsally and traverse the midbrain ventrally of the tectal formation, some fibers continuing forward into the dorsal thalamus (figs. 27-34, 91-94). These fibers may apparently carry nervous impulses activated by all kinds of sensory fibers which enter the medulla oblongata. Their terminals are spread quite uniformly throughout the entire extent of the tectum, and here they mingle with those of the spinal lemniscus (fig. 11; '396, fig. 96). There is little evidence of physiological specificity in either of these tracts, save that one comes from the spinal cord and the other from the medulla oblongata and that the latter may be activated by a wider variety of sensory excitations, including the vestibular, lateral-Hne, and visceral-gustatory systems. This lemniscus is one of the largest tracts of the urodele brain and is very large in fishes also. In anurans and higher forms it can be recognized with difficulty because its fibers are dispersed or segregated in other specialized tracts like the trigeminal lemniscus. Amblystoma has nothing comparable with the medial lemniscus of mammals.


There is some evidence of an incipient trigeminal lemniscus, though most of the secondary general cutaneous fibers ascend in the general bulbar lemniscus ('39&, p. 606). From the region of the superior sensory V nucleus under the auricle a large number of arcuate fibers descend to the ventral commissure. Some of these enter the general bulbar lemniscus. Others are added from the entire length of the spinal V nucleus, and in the calamus region similar fibers join the spinal lemniscus. These connections were observed by Woodburne ('36, p. 455). From the region of the superior V nucleus, fibers ascend uncrossed as far as the isthmic tegmentum, where they end. This may be the precursor of an uncrossed trigeminal lemniscus. In the preparation from which figure 39 was drawn, no fibers of the lemniscus systems are impregnated. The visible external and internal arcuate fibers make local bulbar connections, descend to the spinal cord, and ascend only as far as the isthmus. Some of these fibers divide into ascending and descending branches before or after decussation. Anteriorly on the right, a compact fascicle of external arcuates descends from the superior trigeminal neuropil, turns forward, and decussates close to the ventral surface within the interpeduncular neuropil. Most of them descend in the bulbo-spinal tract, but some turn forward as tr. bulbo-isthmialis. This may be a precursor of a separate crossed trigeminal lemniscus. Their course is parallel with tr. bulbo-tectalis lateralis (primordial lateral lemniscus), but deeper.


3. TRACTUS BULBO-TECTALIS LATERALIS (fIGS. U, 91, 9'2, tr.b.t.l.)

The tr. bulbo-tectalis lateralis is a mixed system of fibers, closely associated with the general bulbar lemniscus and of similar origin from the sensory zone, chiefly from its middle part in the vicinity of the vestibular nerve roots. In the isthmus it lies externally of the general bulbar lemniscus, and it terminates exclusively in the primordial inferior colliculus (nucleus posterior tecti) and the underlying isthmic neuropil. These fibers are mingled with those of a system


MEDULLA OBLONGATA 1G5

which descends from the inferior colhculus — tr. tecto-bulbaris posterior (figs. 1)2, 13, 29-34, 89, 90, tr.t.h.p.). Their segregation from those of the general bulbar lemniscus in the urodeles seems to be determined primarily by their terminal distribution; they end in the inferior colliculus, while the larger bulbar lemniscus ends chiefly in the superior colliculus.


This tract is probably the precursor of the lateral lemniscus, which first makes its appearance in definitive form in anurans. A "lateral lemniscus" has been described in the brains of fishes by many authors, but this is a misnomer. That large tract of fishes is comparable with the general bulbar lemniscus of urodeles and has none of the distinguishing features of the mammalian lateral lemniscus, though it doubtless includes the primordium of that tract. Its current name, "f. longitudinalis lateralis," is more appropriate.


Both these bulbar lemniscus tracts are well developed in Necturus and the reader is referred to their description in that animal for discussion of the phylogenetic relationships ('30, pp. 51 ff.; Ariens Kappers, Huber, Crosby, '36). Necturus lacks a cochlear rudiment, and hence the lateral bulbar lemniscus here probably has a very imperfect auditory function, if any. In Ambly stoma and most other urodeles there is such a rudiment, and here audition of a primitive type may be represented, along with other functions, in this tract. In adult Anura there is a well-developed primordium of the cochlea, from which a cochlear division of the VIII nerve arises. Correlated with this, there are cochlear nuclei, from which a true lateral lemniscus passes to the inferior colliculus. Parallel with this differentiation, both general and lateral lemniscus tracts, as seen in urodeles, are radically reorganized. Larsell ('34, p. 521), after thorough examination of this question, accepts my interpretation of the lateral tract of urodeles as an incipient lateral lemniscus, citing my remark ('30, p. 58) : "It may be that this incipience of an auditory lemniscus in Necturus is a central anticipation of a later peripheral specialization of the cochlear rudiment— ^that is, an apparatus for sorting out centrally the meager auditory component of the mixed functional complex of the undifferentiated VIII nerve. But it is more likely another evidence that Necturus is a degenerated or an arrested derivative of some more highly differentiated ancestor." A good brief summary of the evolution of the auditory apparatus has been given by Papez ('36). For further comments on this tract see pages 188, 214.


4. THE SECONDARY VISCERAL-GUSTATORY TRACT {tr.V.a.)

The fibers of the secondary visceral-gustatory tract are segregated in terms of physiological specificity more definitely than are those of the other lemniscus systems. They are axons of neurons of the nucleus of the f . solitarius and commissural nucleus, which ascend uncrossed ventrally of the spinal V root. Most of the visceral sensory fibers of the VII root are gustatory, and most of those of the vagus roots are general visceral. It is probable that the secondary fibers arising more anteriorly are activated mainly from taste buds, but these cannot be distinguished from those of general visceral sensibility. The central courses of these fibers have been fully described and illustrated in the larva ('14a), in the adult ('446), and in Necturus ('30). A useful summary of the comparative anatomy of this system has been published by Barnard ('36). The adult arrangement is shown here in figures 7, 8, 9, 23, 30, 37, 38, 87-90.


One is impressed by the stability of the general plan of these arrangements throughout the vertebrate series, despite the most extreme modifications of the details in adaptation to different modes of life. This applies particularly to the peripheral and bulbar connections. The ascending connections are less well known, and they probably show more radical changes in the series from lower to higher vertebrates. Fishes and amphibians exhibit a common plan, with infinite variety of detail; the arrangement in Amblystoma appears to present this plan reduced to its simplest form, and this generalized plan may be taken as a point of departure from which the cerebral apparatus of visceral-gustatory adjustments of higher animals has been derived.


All the peripheral fibers enter the f. solitarius, which spans the entire length of the medulla oblongata. They are thin, and many of them are myelinated. Most of those from the geniculate ganglion, which form the visceral sensory VII root, have a T-form division, with ascending and descending branches as they eiiter the f . solitarius. Most (perhaps all) of the prefacial f. solitarius is composed of these ascending facialis fibers, which are less heavily myelinated than are their descending branches and, presumably, are gustatory in function. These fibers ascend to the auricle, where they end in the same neuropil as the ascending V root (figs. 7, 30, 38; '14a, p. 365). By this arrangement, correlation may be effected between visceralgustatory and general somatic sensibility of the mouth cavity. The visceral sensory fibers of the IX and X roots are less myelinated than


MEDULLA OBLONGATA 167

are those of the VII root. Most of them descend without division in the f. soHtarius, but some divide with branches, which ascend for short distances (fig. 37; '446, figs. 14, 17).


In Amblystoma, as in all other vertebrates, the two f . sohtarii converge at the calamus scriptorius into the commissural nucleus, and part of their fibers decussate here in the dorsal commissura infima of Haller. Here at the bulbo-spinal junction there is an important field of correlation between the visceral-gustatory systems and the general somatic sensory systems of the entire body. The details of this structure in Amblystoma, the comparative anatomy of this region, and its strategic importance for fundamental physiological problems have recently been discussed ('446) and are summarized here in chapter ix.


The nucleus of the f. solitarius contains some neurons with dendritic connections exclusively with this fasciculus, but most of these elements have other connections also. Some of their axons cross in the ventral commissure and are lost to view in the vicinity of the general bulbar lemniscus; others ascend uncrossed in the secondary visceral-gustatory tract to the isthmus, midbrain, and hypothalamus. Some of these secondary fibers bifurcate, with branches taking both these courses (fig. 9) . The first of these pathways is probably more primitive, for Barnard ('36, p. 513) writes : "There is no differentiated, uncrossed secondary gustatory tract in the lamprey. All connections with higher centers are through the bulbar lemniscus system." It is believed that in Amblystoma the decussating fibers are concerned chiefly with bulbar reflexes; whether any of them ascend to higher levels in the bulbar lemniscus is not clear. In any event the uncrossed tract is evidently the chief pathway to higher centers. Rostrally of the V roots this tract divides, some of its fibers continuing forward to area ventrolateralis pedunculi and some turning dorsad to the superior visceral nucleus in the isthmus. We have many elective preparations of these fibers which have been cut in various planes, and the entii-e course is clear.


In the ganoid fishes (Johnston, '01; Barnard, '36, p. 517) the secondary visceral tract takes essentially the same course as in Amblystoma; but the superior visceral nucleus lies more ventrally and posteriorly, under the cerebellum. In those teleosts in which the gustatory system is enormously enlarged (carp and catfish), the secondary visceral-gustatory nucleus shows corresponding enlargement and is displaced forward and dorsalward in the isthmus to a position similar to that seen in Amblystoma. This makes it probable that in Amblystoma the fibers of the secondary tract that pass through the isthmus and go directly forward to the peduncle and hypothalamus are general visceral in function and that those that take the more dorsal course to the superior nucleus are chiefly gustatory in function, though this distinction may not be sharply drawn. In support of this conclusion, Barnard ('36, p. 595) mentions the fact that this superior nucleus disappears in birds, in which the gustatory apparatus is greatly reduced. If we may accept the suggestion of Fox ('41, p. 418) and others that the ventral tegmental nucleus of mammals is comparable with the superior visceral-gustatory nucleus of lower forms, it is evident that the mammalian nucleus has a more posteroventral position, similar to that of the generalized ganoid fishes.


The isthmic secondary visceral nucleus was first identified in larval Amblystoma ('14a, pp. 364, 373) and subsequently in Necturus ('17, p. 248; '30, p. 62). Its cells form a low ventricular eminence at the posterior dorsal tip of the isthmic tegmentum (fig. 2B). They are of medium size, scattered and clumped in an open neuropil (fig. 34), with outlying cells dispersed in the alba. As seen in transverse sections, they lie immediately ventrally of the posterior end of the isthmic sulcus ('25, fig. 19; '42. fig. 43). Their long dendrites are directed laterally and ventrally into the posterior isthmic neuropil, where they engage terminals of the secondary visceral tract, ascending root fibers of the trigeminus, tr. bulbo-tectalis lateralis, tr. bulboisthmialis (fig. 38), collaterals of tr. tecto-bulbaris rectus (fig. 37; '42, fig. 67), and terminals of the dorsal tegmental fascicles, notably those of group (9), as described in chapter xx, from the cerebral hemispheres. There is also a diffuse connection with the cerebellar formation, and this is much more intimate in some fishes ('05) .


An extension of the secondary visceral-gustatory tract to the tectum has been described by Brickner ('30) in some teleostean fishes. In these fishes there is a large commissure connecting the two visceral-gustatory nuclei, which includes many decussating fibers of the secondary visceral-gustatory tract. A separate fascicle of the latter fibers turns forward, as tr. gustato-tectalis, to reach the anterior part of the tectum opticum. This connection has not been observed in Amphibia, but it may exist. It makes provision for correlation of visceral-gustatory with visual experience, and it is accompanied by


MEDULLA OBLONGATA 169

a return path from the tectum to the visceral-gustatory nucleus byway of tr. tecto-bulbaris rectus — a typical reflex circle (p. 76).


At the outer border of the gray of this nucleus, in a few preparations there are large spindle-shaped neurons, with dendrites extending for long distances tangentially to the gray layer. One dendrite may ramify among terminals of the secondary visceral tract and dorsal tegmental fascicles of groups (7), (8), and (9) and the other in the deep neuropil and the f. tegmentalis profundus. One such element is illustrated in figure 101, showing an axon directed dorsally toward the decussatio veli, possibly a commissural connection between the two visceral nuclei similar to that described in the frog and in fishes.


These afferent connections suggest that this isthmic nucleus is a correlation center, where gustatory, general visceral, and a considerable variety of other types of experience are brought into relation, with the gustatory component dominant and all in the interest of reflexes concerned with feeding. The afferent fibers of the secondary gustatory tract terminate not only in the neuropil of the visceralgustatory nucleus but also in that of the adjacent isthmic and dorsal tegmentum ('25, fig. 19; '42, fig. 43), where reflexes of the jaw and hyoid musculature are believed to be organized (p. 190).


This hypothesis is supported also by the courses taken by the efferent fibers. These go out in several directions. First, there is a dispersed group of fibers which spread in the underlying tegmentum (figs. 8, 13). This is a direct connection with the neuromotor apparatus of the mouth and pharynx. The remainder enter the tertiary visceral tract (fig. 8, tr.v.t.), which has rather wide distribution. This tract of Amblystoma is relatively small, and its fibers are dispersed and so intimately mingled with others that analysis is difficult. The diagrams show what has been clearly seen. In figure 8 the dotted component of tr.v.t. is highly probable, though not confirmed by elective impregnations.


In the description of the brachium conjunctivum (p. 176) this fascicle is shown to descend close to the gray in the posterior lip of the isthmic sulcus to reach the decussation. In this part of their course the cerebellar fibers comprise a large component of the f . tegmentalis profundus. As they emerge from the cerebellar formation, they pass through the superior visceral nucleus, and here they are joined by a smaller number of similar thin unmyelinated axons of cells of this nucleus. These two systems of fibers are mingled, and analysis is impossible except in electively impregnated material. We have few such specimens of either Amblystoma or Necturus, but those we have show that in the ventral commissure immediately spinalward of the decussation of the f. retroflexus some of the visceralgustatory fibers turn forward in company with those of the ventral secondary visceral tract, and both groups of fibers arborize in the ventrolateral neuropil of the peduncle. Some of these fibers probably pass through this neuropil without synapse to reach the hypothalamus, as in fishes, in company with pedunculo-mamillary fibers and mingled with those of the mamillo-peduncular, mamillo-tegmental, and mamillo-interpeduncular tracts. At their decussation the tertiary visceral fibers spread in the alba of the isthmic tegmentum before and after crossing, and some of them probably reach the interpeduncular neuropil, though this needs confirmation. All these systems and others (including, perhaps, mamillo-cerebellar fibers) are mingled, and none of the tracts are closely fasciculated.


In some fishes these tracts are separately fasciculated, and the amphibian arrangement, so far as revealed in our material, conforms with the teleostean pattern. The tertiary visceral-gustatory tract is very large in some of these fishes ('05, pp. 420, 436, figs. 20, 23, 37, 38), passing from the secondary nucleus directly to the hypothalamus. This direct connection has been described by Larsell ('23, p. 109) in the frog; and in both Amblystoma and Triturus he saw some indications of it. In the frog these direct fibers do not take the deep course, as described above for Amblystoma, but are superficial, accompanying mamillo-cerebellar fibers. I find evidence of this superficial connection in Amblystoma also, but in the absence of elective impregnations no accurate description can be given. These fibers in some of our preparations are seen as a well-fasciculated tract. In the Cajal sections, from which figures 25-36 were drawn, it is not so clearly shown as in some other specimens, but its location can be identified. These fibers assemble at the lateral surface rostrally of tr.v.a. (fig. 34), in company with those of tr.t.b.p., as drawn in that figure. Ten sections farther ventrally, they are drawn, but not labeled, in figure 33 superficially of tr.t.h.r. In some Golgi preparations, fibers in this position connect with the hypothalamus, and this is the course taken by the tertiary visceral tract in the frog (Larsell, '23, figs. 2, 4). This is the position also of the tr. mamillo-cerebellaris,


MEDULLA OBLONGATA 171

which I provisionally identified in Necturus ('14, p. 8) and subsequently ('17, p. 250) questioned and the presence of which has been confirmed by Larsell.


In the light of such fragmentary evidence as we now have, it seems to me probable that in Amblystoma the ascending tertiary visceralgustatory fibers take a deep course, accompanying the brachium conjunctivum, and that the superficial fascicle described above is a mamillo-cerebellar tract, which also connects with the isthmic tegmentum and the isthmic visceral-gustatory nucleus.


The bulbar apparatus of visceral-gustatory adjustment is simply organized in Amblystoma, as, indeed, it is in man, with minimal provision for local specialization. The division of the ascending path into dorsal and ventral moieties in the isthmus is similar to the divergence of the olfactory paths to epithalamus and hypothalamus. In both cases somatic correlations are effected dorsally and olfactovisceral correlations ventrally. The dorsal gustatory nucleus may directly activate the skeletal musculature concerned with feeding, through its connections with the underlying tegmentum, or it may discharge into the hypothalamus, where all visceral adjustments are organized.


The swallowing of a morsel of food in response to excitation of taste buds is one of the simplest acts of which the body is capable. This may be done reflexly through the local activation of the bulbar connections of the f. solitarius; but, simultaneously with this local action, nervous impulses may be transmitted to higher centers which are so interconnected as to bring this simple act into relation with any other activities that may be in process in response to internal or external stimuli. The final result of any particular sensory excitation is dependent upon the central excitatory state of the central nervous system as a whole and of every local part of it. And the quality of this excitatory state is determined not only by the stimuli presently acting upon it but also by the^past experience of the race and of the individual. This theme is amplified in a recent discussion of the apparatus of optic and visceral correlation ('44a).



CHAPTER XII CEREBELLUM

IN URODELES the cerebellum is incompletely separable from the brain stem, and the entire rhombencephalon evidently acts as a closely integrated unit. Larsell's papers ('20, '31, '32, '45) on the cerebellum of Amblystoma and Triturus give good accounts of its development and structure.


Three of the primordia from which the mammalian cerebellar complex has been assembled are here clearly separate in relations easily recognized. The fourth mammalian component — the pontile system — is not represented in Amblystoma. The three components present are: (1) the vestibulo-lateralis system in the auricles, primordia of the floccular part of the fiocculonodular lobes ; (2) the median body of the cerebellum, which is ancestral to the larger part of the vermis and adjoining regions; and (3) the nucleus cerebelli, internal to the other two and in intimate relations with both of them. The topographic arrangement of these parts as seen in transverse sections is shown in figure 91. In higher brains these three components are variously merged, and the nucleus cerebelli is subdivided and incorporated within the cerebellar mass as the deep nuclei. The connections of the vestibular and lateral-line systems with the auricle have been described above. The afferent connections of the body of the cerebellum are shown in figure 10. These include primary and secondary sensory trigeminal fibers, spino-cerebellar fibers, tectocerebellar fibers, and probably a hypothalamo-cerebellar tract. Evidence from both comparative anatomy and embryology indicates that cerebellar differentiation began within the sensory zone, and in the adults of all vertebrates it retains some features characteristic of this zone, including termination within it of vestibular root fibers and (in many species) of trigeminal root fibers also. It is morphologically supra-segmental and physiologically supra-sensory, an adjustor of higher order, not primarily concerned with determining the pattern of performance but rather with facilitation and regulation of its execution. The motor zone plays no part in it genetically or functionally except as a sort of accessory after the fact. In later phylogenetic

172


CEREBELLUM 173

stages, with the appearance of the pons and cortico-pontile connections, this situation is changed; but even in mammals these connections, though topographically in the motor zone, are foreign to its intrinsic organization, for the pons and all its connections are dependencies of a supra-segmental apparatus and the cells of the pontile nuclei are embryologically derived from the sensory zone. In the most primitive vertebrates in which cerebellar differentiation is minimal (almost absent in myxinoids, Jansen, '30; Larsell, '47a) the sensory zone of the medulla oblongata abuts against that of the midbrain, which in these animals is the dominant adjustor of all somatic (nonvisceral) activities. Necturus presents a very early stage in the fabrication of the cerebellar complex ('14), scarcely more advanced than in cyclostomes. In Ambly stoma the process is further advanced. It is evident that cerebellar differentiation began under the influence of several kinds of sensory excitations, viz., the vestibular, lateralline, general cutaneous, and deep sensibility of the muscles, joints, etc. ('24 ; Larsell, '47, '47a) . Connection with the visual system of the tectum was effected in early phylogenetic stages. With the elaboration of the cerebral cortex in mammals, another major component was added — the cortico-pontile system. This history has been thoroughly explored and documented by Larsell ('20-'47a). Brodal ('40) and Brodal and Jansen ('46) have greatly extended our knowledge of the connections between the inferior olive and the cerebellum in mammals, but we have little exact information about them in lower forms.


The isthmus is a critical junctional field. Immediately spinalward of it the sensory and intermediate zones of the medulla oblongata are enlarged to form the massive auricle, within which are terminals of the ascending vestibular and associated lateral-line roots (fig. 7). This is probably the first primary component of cerebellar architectonic. More medially the ascending sensory root of the trigeminus reaches its terminal (superior) nucleus. Some of these V root fibers, accompanied by secondary V fibers from the superior nucleus, continue into the median body of the cerebellum (figs. 10, Sl-33, 91), where they are joined by fibers of the spino-cerebellar tract. Some fibers of the mixed bundle decussate in the commissura cerebelli (com.cb.). This is the second primary component of the cerebellum from which the median mass (corpus cerebelli) of larger cerebella has been derived and into which converge sensory influences of all sorts except those from the internal ear and the lateral-line organs. This median mass receives fibers from the mesencephalic tectum and (in lower-vertebrates) also from the hypothalamus. The tecto-cerebellar tract may be activated from the mesencephalic V nucleus and from terminals of the optic and lemniscus systems.


The very rudimentary cerebellum of Myxine has been the subject of much controversy. Holmgren ('46, p. 54) described in the embryo two possible primordia, one in the auricle of the rhombencephalon and one at the posterior end of the mesencephalic tectum. Larsell's comprehensive study ('47, '47a) of this region of cyclostomes has clarified the problem. Myxine has the most primitive cerebellum known, and, so far as its differentiation has gone, it conforms with the typical vertebrate pattern.


The spino-tectal terminals in the inferior colliculus are primitively contiguous with the vestibular and lateralis terminals in the auricle, and from the wedlock of these two systems the cerebellum was born. The primordial inferior colliculus, accordingly, was primitively concerned with proprioception. Upon this foundation the cochlear system of higher animals was built in much the same way that the bulbar cochlear nuclei emerged from the lateral-line nuclei of the medulla oblongata (p. 138). It is significant that in Amblystoma the primordial inferior colliculus contains dense collections of cells of the mesencephalic V nucleus, that fibers of the mesencephalic root of the V nerve penetrate cerebellar tissue, and that cells of the mesencephalic V nucleus are found sparsely scattered in the cerebellar gray (p. 140). It seems probable that the mesencephalic V nucleus was first differentiated at the posterior end of the tectum and that from this focus its cells spread forward into the optic tectum and, in smaller number, backward into the cerebellum.


In the gray of both the median body of the cerebellum and the lateral auricle some of the neurons are specialized as Purkinje cells, in contrast with smaller cells, which are precursors of the cerebellar granules. The largest and best developed Purkinje cells are loosely arranged at the outer border of the central gray, but they are not delaminated from the granular layers as in mammals. If cortex is defined as laminated superficial gray, Amblystoma has no cerebellar cortex, though its primordium is clearly evident. The Purkinje cells, though of simple form, are quite distinctive. From them and from the nucleus cerebelli, efferent fibers stream downward, forward, and backward as cerebello-tegmental fibers, and one large fascicle of


CEREBELLUM 175

these extends farther forward parallel with the sulcus isthmi as brachium conjunctivum.


In, my earlier papers a ventricular swelling under the cerebellum was termed "eminentia subcerebellaris tegmenti." This has subsequently been analyzed into several distinctive areas, of which the posterodorsal member was named "eminentia cerebellaris ventralis" ('35a, fig. 1). It is now clear that this is the primordium of the deep nuclei of the cerebellum, and it is here named "nucleus cerebelli" (figs. 2, 10, 32, 33, 91, nvc.cb.). This is an ill-defined region of the intermediate zone, not clearly separable from its surroundings. Some fibers of the brachium conjunctivum arise from these cells, though only a small part of them.


The spino-cerebellar tract ascends in company with the spinal lemniscus, some cerebellar fibers being collaterals of lemniscus fibers, as described above. Cerebellar fibers join this tract from the nucleus of the dorsal funiculus and the sensory zone of the medulla oblongata. One confused issue regarding this tract of Amblystoma can now be clarified. A dorsal slip of the spino-cerebellar tract passing through or above the entering fibers of the sensory V root was described by me ('14, p. 7) and by Larsell ('20, p. 277). These observations could not be confirmed by either of us (Herrick, '14a, p. 375; Larsell, '32, p. 414), the supposed tract being then interpreted as the ascending trigeminal root. It now appears that both the earlier and the later accounts are correct. In one of our adult brains, cut in a favorable oblique plane (no. 2245), there is partial Golgi impregnation of the sensory V root fibers and of the spino-cerebellar tract, with no other fibers stained in this vicinity. A separate fascicle of the ventral spinocerebellar tract turns dorsally at the level of the V nerve, its fibers interdigitate with those of the entering sensory V root, and rostrally of this level the spino-cerebellar fibers and the ascending sensory V fibers are mingled as far as their entrance into the cerebellar commissure. Here a fortunate elective impregnation clarifies the confusion as neatly as could be done by an experimental degeneration (compare Salamandra as described by Kreht, '30, p. 294).


In the literature our tractus spino-cerebellaris is sometimes designated "tractus cerebello-spinalis," implying conduction spinalward. Some descending fibers may be present in this tract of Amblystoma, though our preparations have not revealed them. That the larger part of the tract is ascending is clearly evident.


The tr. tecto-cerebellaris was originally described ('*25, p. 483 and figs. 39-41) from two series of sagittal Golgi sections of the adult, and this description has been confirmed by Larsell ('3'2, p. 425). Additional details can now be added. In several series of Golgi sections of advanced larvae and adults these thin unmyelinated fibers arise from neurons of the nucleus posterior tecti with the dendrites directed forward into the posterior border of the optic tectum. Their axons form a rather compact fascicle close to the dorsal surface of the nucleus posterior, which passes backward though the anterior medullary velum and laterally of it, accompanying thick myelinated fibers of the mesencephalic root of the V nerve which take a similar course. At the anterior border of the cerebellum they turn ventrad and spread out along the anterior side of the body of the cerebellum, some reaching the underlying nucleus cerebelli.


A tr. mamillo-cerebellaris was described by Larsell ('20, p. 279; '32, p. 424), and some evidence of it has been seen in our preparations; but in the absence of elective impregnations no satisfactory description can be given. In view of the large size of this connection in fishes, its presence in urodeles may be expected.


BRACHIUM CONJUNCTIVUM

Large numbers of unmyelinated and lightly myelinated fibers descend from the body of the cerebellum, the auricle, and the nucleus cerebelli. These cerebello-tegmental fibers are in addition to the connections of the auricle with the vestibulo-lateralis sensory field in correlation tracts a and b; they are directed in dispersed arrangement ventrally, posteriorly, and anteriorly. A large and rather compact fascicle of these fibers, a few of them myelinated, is directed antero ventrally and is clearly a primordial brachium conjunctivum (figs. 10, 21, 71). These axons arise from cells of the body of the cerebellum (fig. 47) and the nucleus cerebelli (14a, fig. 52); they assemble anteriorly of the body of the cerebellum and pass forward through the visceral-gustatory nucleus in the isthmus (figs. 33, 34). Here they are joined by fibers of the tertiary visceral tract, which arises in this nucleus (p. 169).


This mixed tract with accessions of fibers from other sources has been termed "fasciculus tegmentalis profundus" ('36, p. 304, figs. 14, 23, f.teg.p.) ; its analysis is possible only with the aid of elective impregnations, and these, fortunately, are available (see the further de


CEREBELLUM 177

scription in chap. xx). This fascicukis extends forward and ventralward from the cerebellum at the outer border of the gray in the lips and floor of the isthmic sulcus, finally to reach the ventral commissure, where some of its fibers decussate a short distance spinalward of the fovea isthmi, mingled with those of the decussation of the f. retroflexus and spinalward of it. Its course as seen in horizontal sections is shown in figures 29-32 (f.teg.p.), and in sagittal sections in figure 104 and previously published figures of the same specimen ('36, figs. 19-21, trxb.teg.).


Elective impregnations show that the brachium conjunctivum is the largest component of this mixed fasciculus (figs. 71, 72). Many of its fibers spread outward into the alba of the isthmic tegmentum on the side of origin, and these include all the myelinated fibers of the brachium. A residue of the unmyelinated fibers decussates in the ventral commissure. Elective Golgi impregnations show that, at the decussation, fibers of the brachium conjunctivum and tertiary visceral tract are mingled and that both enter a dense superficial neuropil laterally of the crossing. Here the tertiary visceral fibers turn forward in company with those of the secondary visceral tract to reach the area ventrolateralis pedunculi, and the cerebellar fibers turn posterodorsally and spread out in the alba of the isthmic tegmentum. This primordial brachium conjunctivum may activate almost the entire extent of the isthmic tegmentum diffusely on both sides. No evidence has been found of a concentration of cells related with it which could be regarded as a nucleus ruber.


THE CEREBELLAR COMMISSURES

In the body of the cerebellum there are two commissures, which differ in position, connections, and functions. They are specifically related with the two chief subdivisions of the cerebellar complex.


The com. cerebelli (com.cb.) is a compact fascicle of well-myelinated fibers crossing more ventrally than the less myelinated com. vestibulo-lateralis cerebelli (com.cb.i.), as shown in figures 10 and 91. It crosses between the levels of figures 34 and 35. The com. cerebelli is composed of fibers from two sources, which terminate chiefly in the median body of the cerebellum. These are (1) fibers of the tr. spinocerebellaris (figs. 32, 33, 91) and (2) fibers of the sensory trigeminus system. The latter are, in part, fibers of the ascending sensory V root terminating in the cerebellum and, in part, axons of cells of the superior sensory V nucleus terminating in the cerebellum and the neuropil of the V nucleus of the opposite side (figs. 31, 32, 91). The latter is a true trigeminal commissure.


The com. vestibulo-lateralis cerebelli is a more dispersed collection of unmyelinated and lightly myelinated fibers assembled from the auricles (figs. 31 36, 91). Included among them are root fibers of the vestibular nerve {comxb.VIII) and secondary fibers of the lateralline system of nerves {com.cb.l.L). These are mingled at their crossing in the cerebellar alba. Whether any primary root fibers of the lateralline nerves decussate in this commissure is not clear in our material ; apparently they do not.


PROPRIOCEPTIVE FUNCTIONS OF THE CEREBELLUM

In the initial differentiation of the cerebellum and in its normal functions in all animals, proprioception has played a dominant part. Nevertheless, it must be recognized that all sensory systems may participate in cerebellar control of muscular movements (Larsell, '45), a control that is applied to the motor adjustors as going concerns at every phase of their operations. The significance of the cerebellum as "the head ganglion of the proprioceptive system" (Sherrington) is discussed in chapter x in connection with a critique of the proprioceptive system as a whole.



CHAPTER XIII ISTHMUS

THE strategic position of the isthmus as a transitional sector of the brain stem between rhombic brain and cerebrum has been commented upon (pp. 45, 118). In early neural-tube stages of the developing human brain it is a rather large sector, but in the adult it is hardly recognizable as a distinct entity because its tissues are dispersed among other structures of relatively recent phylogenetic origin. In contrast, the isthmus of adult Amblystoma retains its embryonic separateness and, indeed, accentuates it. Here its distinctive features are most clearly seen, and these will be described as fully as available material permits. Its dorsal part is contracted and includes only the anterior medullary velum and some adjoining structures. To the intermediate zone is assigned the superior visceral-gustatory nucleus, and for convenience of description the remainder is included in the motor zone, though its functions are, in large part, of intermediate type. The interpeduncular nucleus is an isthmic structure, to which a separate chapter is devoted.


DEVELOPMENT

The configuration of this sector of the brain exhibits an interesting series of changes in the course of development, due to the rapid and radical dislocations caused by the cerebral flexures and to other inequalities of intrinsic growth. Successive modifications of the isthmic sulcus provide a useful indicator of the course of these changes.


From the coil stage (Harrison's stage 35) to the adult, the isthmic tegmentum is intimately joined with the trigeminal tegmentum, with no constant external or ventricular boundary visible. "Isthmus rhombencephali," therefore, is an appropriate name. But its anterior boundary is sharply marked by the fovea isthmi, the deep ventricular sulcus isthmi, and the external fissura isthmi. Even in early stages these grooves are not exactly parallel, and as development advances they deviate from this relationship more and more widely because of unequal growth of the deep and superficial structures.


Reference was made on page 177 to von Kupffer's sulcus intraencephalicus posterior as the precursor of the sulcus isthmi. In coil ('37, fig. 1), S-reaction ('37, figs. 7, 8), and early swimming stages ('37, fig. 2) the external fissura isthmi lies a little farther spinalward than the internal sulcus. In slightly older early swimmers ('38, figs. 15-18) both the isthmic fissure and the isthmic sulcus are deep, and the sulcus is displaced farther rostrally of the fissure. This means that the gray of the isthmus is proliferating rapidly and pushing the isthmic sulcus farther forward. The sulcus is now a very deep cleft and lies approximately transversely to the long axis of the rhombencephalon. From this stage the enlargement of the isthmic tegmentum continues on the ventricular side, and in the early feeding stage (Harrison's stage 46) the isthmic sulcus extends from the cerebellum almost horizontally forward before dipping downward to the fovea isthmi ('386, figs. 1, 2). The sulcus and the fissure are still deep, and the sulcus has come to lie far rostrad of the fissure. This is not evident in transverse sections but is shown clearly in horizontal sections of this and subsequent stages ('396, fig. 16). In later larval stages, internal differentiation, particularly the elaboration of the neuropil and thickening of the auricle, result in smoother contours both internally and externally, and the isthmic sulcus and fissure are shallower in the adult.


At the close of the early swimming stage (about Harrison's stage 38) the cerebellum and auricle are so little developed that the plica rhombo-mesencephalica is marked externally by a wide depression ('38, fig. 17). Between this and the early feeding stage (Harrison's stage 46) the sharp cerebral flexures are straightened, resulting in a backward thrust of the plica rhombo-mesencephalica in the dorsal wall, while the ventral wall of the isthmus remains fixed (compare '38, fig. 18, with '386, fig. 1). Though there has not yet been any very great enlargement of the isthmic tegmentum, the isthmic sulcus in early feeders lies more nearly horizontally than vertically. In midlarval stages the deep sulcus isthmi (shown but riot named in fig. 2 of '14a) is inclined more than 45° from the vertical, and between midlarval and adult stages there is enormous enlargement of the gray of the isthmic tegmentum, which thrusts the sulcus isthmi forward and upward toward the horizontal position.


The external fissura isthmi does not follow the sulcus isthmi in these shifts of position, but, as mentioned above, it lies in the adult a considerable distance spinalward of the sulcus. That this involves a mechanical stretching of the wall between the fissure and the sulcus


ISTHMUS 181

in later stages is evident from the arrangement of the ependymal elements in the lips of the isthmic sulcus. This traction begins early. It was noticed in early swimmers ('38, p. 220) that the ependymal elements of the sulcus isthmi are thickened and sharply bent outward and backward. This was illustrated also in early feeding stages ('386, figs. 11, 12) and in Golgi impregnations of midlarval stages ('396, p. 518, and figs. 31, 93). At all stages, including the adult, these thick elements form a dense band of parallel fibers bordering the sulcus on both sides and extending outward and backward to an attachment at the pial surface in the floor of the isthmic fissure. In the adult the sulcus isthmi usually does not extend ventrally so far as the fovea isthmi, but its locus here is indicated by a band of modified ependyma. Two mechanical factors, accordingly, can be recognized in the shifting relations of the isthmic fissure and sulcus : first, the backward thrust of the roof during the straightening of the cerebral flexures between early swimming and early feeding stages and, second, unequal growth of superficial and deep structures, especially in later stages. The rapid enlargement of the gray of the isthmic tegmentum in late larval stages produces an upward and forward thrust ventrally in the direction opposite to the dorsal backward thrust of early stages.


The developmental history and adult position of the sulcus isthmi of Amblystoma are radically different from those of Hynobius as described by Sumi ('26), who illustrates models showing the ventricular sculpturing from closure of the neural tube to the adult. At none of these stages is von Kupffer's sulcus intraencephalicus posterior visible at the plica rhombo-mesencephalica. The sulcus isthmi appears rather late at the fovea isthmi, and it does not at any stage reach the dorsal surface.


SENSORY ZONE

The sensory field of the isthmus is contracted. Its scanty gray substance includes a compact cluster of cells of the mesencephalic V nucleus within and adjacent to the anterior medullary velum. These cells are smaller and more densely crowded than are those of this nucleus in the tectum, and they may have dift'erent functional significance (p. 140). Their connections are unknown. Associated with them are other cells at the posterior border of the inferior colliculus which appear to have peripheral connections. From this region a few myelinated, and many unmyelinated, fibers go out to the meninges and chorioid plexus of the fourth ventricle and some with the IV nerve.


These fibers appear to be related with cells bordering the velum, but no clear evidence of their connections has been found ('36, p. 343; '42, pp. 255, 291). Within and adjoining the anterior medullary velum are fibers of passage of several kinds, including the decussation of the IV nerves, mesencephalic root of the V nerve, and tractus tecto-cerebellaris.


INTERMEDIATE ZONE

The intermediate zone as defined here is represented by the small gray area of the superior secondary visceral-gustatory nucleus and associated neuropil. This gray lies dorsally of the isthmic tegmentum and incompletely separable from it. In figures 2 and 13 it is shown wedged between the isthmic tegmentum below and the nucleus cerebelli and dorsal tegmentum above. The cells of the superior visceral nucleus are of medium size, scattered and clumped in an open neuropil, with a few outlying cells in the alba. Their dendrites are directed ventrolaterally through the alba into the posterior isthmic neuropil (described below), where they engage terminals of the ascending visceral tract ('42, pp. 244, 253, and fig. 43) and of fibers from many other sources.


There is an obscure vestige here of the nucleus isthmi, which is well differentiated in the frog ('42, pp. 244, 253). In anurans and reptiles this nucleus is in intimate relations with the lateral lemniscus, and it is probably part of the auditory reflex apparatus, with, perhaps, vestibular connections also. In Ambly stoma the auditory system is poorly developed, and the crawling habit maintains the balance of the body without nervous control; the nucleus isthmi, accordingly, is undeveloped.


MOTOR ZONE

This field is so intimately related anatomically and physiologically with the tegmentum of the trigemino-facialis region that these will be described together. The tegmentum isthmi is bounded anteriorly by the isthmic sulcus and fovea isthmi, separating it from the cerebral peduncle and dorsal tegmentum. Posteriorly it is bounded by the nucleus cerebelli and the trigeminal tegmentum. Its gray substance includes the nucleus of the IV cranial nerve (as described in chap, x) ; a compact central nucleus of the isthmus (fig. 29, nuc.is.c.) ; and, surrounding this, a pars magnocellularis which is continuous spinal ward with the large-celled component of the trigeminal tegmentum (fig. 30).



ISTHMUS 183

TRIGEMINAL TEGMENTUM

The tegmentum of the trigeminal field forms a low eminence in the floor of the fourth ventricle internally of the V roots, the erainentia trigemini (figs. 2C, 90), which is evident in Necturus also ('30, figs. 11, 12, em.V). The gray under this eminence comprises a small-celled dorsal and medial part and a large-celled ventral part (fig. 27) . The small-celled part is continuous medially with a band of similar cells, which extends spinalward under the ependyma of the paramedian sulcus (fig. 28), and anterodorsally it is continuous with the ventral border of the central nucleus of the isthmic tegmentum and more intimately with the deep periventricular gray of the isthmus (fig. 29). The large-celled part is continuous anterodorsally with the largecelled component of the isthmic tegmentum (figs. 28, 29, 30). Imbedded among these large tegmental cells are the motor V nuclei. There are two motor V roots (fig. 27), but the two motor V nuclei seen in Necturus ('30, p. 13) are here merged. Figure 40 shows three of these cells in Golgi impregnation. The heavily myelinated fibers of the mesencephalic V root are more dorsal, passing upward toward the tectum at the outer border of the gray (figs. 13, 29-32).


The configuration at the boundary between trigeminal and isthmic tegmentum is variable, and the position of this boundary is debatable. A posterodorsal extension of the isthmic tegmentum is continuous with a low ventricular eminence, the superior visceral nucleus. The latter is separated by a shallow and variable sulcus from the nucleus cerebelli posteriorly and a tegmental area ventrally. Comparison of the two key drawings, figures 2B and 2C, reveals differences due partly to individual variations in relative sizes of parts and partly to lack of any well-defined boundary between the large-celled components of the isthmic and trigeminal tegmentum. In figure 2B the dorsal part of the area marked teg.is.m. contains only large cells (fig. 31), and these may be assigned to either the isthmic or the trigeminal tegmentum. More ventrally (figs. 29, 30), both large and small cells occupy a similar ambiguous position.


The large cells of the trigeminal tegmentum take various forms, one of which is shown in figure 62 (others are illustrated in '14a, figs. 21-26). Like those of the nucleus motorius tegmenti elsewhere, they are concerned with motor co-ordination in patterns as yet imperfectly known. Some of these cells in the auricle have dendrites directed forward into the posterior isthmic neuropil, where they engage terminals of the tegmental fascicles (fig. 46). Other similar neurons direct their dendrites laterally among terminals of the ascending sensory V root and related tracts ('396, figs. 47-49, 53, m, 68). The axons of some of these elements are directed forward, as shown in figure 43. These short fibers pass from the trigeminal tegmentum to arborize within the gray of the isthmic tegmentum and are doubtless part of the apparatus of intrinsic co-ordination of patterns of action within the motor zone.


Some of the small cells at the anterior end of the trigeminal tegmentum are shown in figures 66 and 91. These resemble those of the adjacent interpeduncular nucleus. Their dendrites extend laterally through the entire thickness of the alba of the ventral part of the auricle, where they are in contact with all kinds of fibers passing this region. The axons of some of them descend to the interpeduncular neuropil (figs. 63, 80, 84; compare similar cells of the isthmic tegmentum, figs. 60, 84).


ISTHMIC TEGMENTUM

Horizontal sections are most instructive in the analysis of this region (figs. 29-33). The nervous elements are arranged in five groups: (1) A narrow and inconstant layer of subependymal cells. (2) These are separated by a thick sheet of neuropil from the lens-shaped area of small and medium cells, here termed the "central nucleus of the isthmus." (3) Externally and spinalward of this area is the pars magnocellularis, containing scattered nerve cells of medium or large size. The extensive neuropil within which these cells are imbedded is broadly continuous with that of the surrounding alba. This largecelled area envelops the lentiform area of small cells laterally, ventrally, and caudally and is continuous spinalward with a similar field of the trigeminal tegmentum. In figure 2B the dotted line separating the pars magnocellularis (teg.is.m.) is arbitrarily drawn to indicate a posterior sector which may be assigned to either isthmic or trigeminal tegmentum. (4) Dorsally of all the preceding, the gray of the isthmic sector extends upward into contact with that of the dorsal tegmentum and ventral cerebellar nucleus. This area is the superior secondary visceral-gustatory nucleus (fig. 2B, iiuc.vis.s.). (5) A probable vestige of the nucleus isthmi.


The first component of the preceding list is the ventral part of the zone of deep subependymal gray, which everywhere surrounds the aqueduct. In the isthmus it is thin and contains few nerve cells, but more posteriorly in the trigeminal tegmentum it expands to a wide zone of subependymal small cells (figs. 29, 30). This and the associ


ISTHMUS 185

ated neuropil comprise one of the most primitive and conservative features of this part of the brain and also of the diencephalon and the mesencephalon. In mammals it contributes fibers to the dorsal longitudinal fasciculus of Schiitz, as described in the opossum by Thompson ('4'-2).


The second comi)onent, the central nucleus, is clearly delimited in its middle part. Anteriorly, it merges with the small cells of the posterior gray of the peduncle (figs. 29, 30, 31) and more dorsally with the ventral border of the dorsal tegmentum (fig. 32). Posteriorly, it is continuous with the small-celled component of the trigeminal tegmentum (fig. 91), though the boundary between these is clearly marked in many preparations. At its ventral border the gray is continuous with that of the interpeduncular nucleus, with difference in form and arrangement of the cells, though many transitional forms are seen ('42, fig. 43). The boundaries here mentioned are in some preparations quite indeterminate; in some reduced silver preparations they are evident; and in Golgi preparations each of the regions mentioned has characteristic structure and connections. All these areas are intimately connected by a web of interstitial neuropil, the structure of which is characteristically different in each of them. All this gray is connected with the interpeduncular nucleus by fibers passing in both directions, and in chapter xiv, devoted to that nucleus, it is suggested that the central nucleus of the isthmus contains, among other constituents, the primordium of the dorsal tegmental nucleus of mammals.


The large-celled component 3 surrounds the central nucleus on all sides, and many of these larger neurons lie deeper, mingled with those of the small-celled component. These cells have massive dendrites, which spread widely in the overlying alba. Among these are cells of the nucleus of the nervus trochlearis (fig. 70). Figure 61 shows three neurons from this region which probably belong to the IV nucleus. They are surrounded by other similar elements of the tegmentum. Most of the axons fi-om the large tegmental neurons descend, crossed or uncrossed, in the tegmento-bulbar tracts, as described in the larva ('396, p. 590). Some axons from the isthmic tegmentum, as also from the trigeminal tegmentum, are directed forward. So far as observed, these take short courses, though some may extend farther (figs. 43, 46). This may be a precursor of the connection from the substantia nigra to the corpus striatum (Kodama, '29) which has recently been confirmed by several students of nianinialian neurology. Component 4, I lie superior visceral nucleus, has been described above (p. 18'-2).


WHITE SUBSTANCE

In the white substance of the isthmus the most conspicuous features are the numbered tegmental fascicles described in chapter xx. Externally of these fascicles are the massive teclo-bulbar and tectospinal tracts and, suj)erficially close to the pial surface, a layer of line fibers, most of which are unmyelinated. The latter includes tr. thalamo-tegmentalis rectus from dorsal and ventral thalamus (fig. 94; '36, }). .'UO) and, more dorsally, similar fibers from the dorsal thalamus, which decussate in the postoptic commissure — tr. thalamo-tegmentalis dorsalis cruciatus (p. '-29!); '39, pp. 95, 116).


Mingled with these fibers and more abundantly at deeper levels are fibers of tr. tecto-peduncularis and tecto-tegmentalis (figs. 18, 22, 24; '42, p. 267 and figs. 14, 30; Necturus, '17, figs. 9-14, 32, 33, tr.t.ped.p.). Short fibers from the entire tectum and dorsal tegmentum to the isthmic tegmentum are dispersed in both white and gray substance, and a large number of the longer fibers are assembled in dorsal tegmental fascicles of group (7) accompanying tr. tecto-bulbaris rectus (p. 283).


Figure 21 shows only the more prominent att'erent connections; there are many others. In the aggregate these include fibers from the motor zone above and below the isthmus, collaterals of the lemniscus sj^stems in the isthmic neuropil, terminals of the ascending visceralgustatory tract, cerebello-tegmental fibers, short fibers from the overlying dorsal tegmentum and tectum, longer fibers from the tectum which decussate in the postoptic and ventral commissures, direct fibers from the dorsal and ventral thalamus, strong tracts from the dorsal and ventral thalamus which decussate in the postoptic commissure, fibers from both cerebral hemispheres by way of the lateral forebrain bundle and tr. olfacto-peduncularis, and strong tracts from the hypothalamus, crossed and uncrossed. To this list one might add the fasciculus retroflexus, passing from the habenula to the interpeduncular nucleus and, through the latter, acting upon the isthmic tegmentum.


Most of the efferent fibers descend in the f. tegmentalis profundus (p. 286) and the ventral tegmental fascicles. Some are dispersed in the neuropil accompanying tr. interpedunculo-bulbaris dorsalis, and these are regarded as jn-ecursors of the bulbar part of the f. longitudinalis dorsalis of Schlitz (p. 208). The relativelj^ short fibers which


ISTHMUS 187

descend from the hir^e cells of theisthmic<ind trigeminal tegmentum to the motor zone of the medulla oblongata are especially clearly seen in very young larvae from early swimming to early feeding stages and are interpreted as provision for activation of the mandibular and hyoid musculature involved in feeding. Longer fibers descend into the spinal cord; but these, so far as is sliown in our material, are not numerous.


ISTIIMIC NEUROPIL

Both the gray and the white substance of the isthmus are permeated with dense neuropil, the axonic component of which is composed mainly of terminals of afferent fibers and collaterals of fibers of passage. The superficial neuropil is composed chiefiy of branched terminals of the unmyelinated thalamo-tegmental and tecto-tegmental tracts and dendritic terminals from the underlying gray. The intermediate neuropil is in the form of sheets or plaques insinuated between the tegmental fascicles and composed largely of terminals and collaterals of fascicular fibers. The deep neuropil of the alba and the grisea is a c(mfused entanglement of fibers and dendrites, the analysis of which is quite impossible except where clarified l^y elective Golgi impregnations. These preparations reveal a num})er of more or less well-defined tracts imbedded within this interstitial matrix. In the gray substance the cells are arranged in lamellae separated by sheets of neuropil, which is continuous with that of all neighboring parts. A large proportion of these axons trend dorsoventrally from tectum and dorsal tegmentum to isthmic tegmentum and interpeduncular neuropil and posteroventrally, with or without decussation in the ventral commissure. The latter axons accompany thicker fibers of the tr. tegmento-bulbaris arising from large cells of the motor tegmentum.


POSTERIOU ISTIIMIC xVEUKOPIL

The name "posterior isthmic neuropil" is given to a wide zone of neuropil in the alba at the boundary between the isthmus and the bulbar and cerebellar structures. It invades the auricle below and the nucleus posterior tecti above. It is permeated by dendrites from all surrounding parts, including the trigeminal tegmentum (fig. 46; 'S9b, figs. 47-49, 62, 6P>), auricle ('396, fig. 53), cerebellum (fig. 47), superior visceral nucleus ('42, fig. 43), and dorsal and isthmic tegmentum. It receives terminals of peripheral sensory fibers of the V, VH, VIII, and lateral-line nerves, of a bulbo-isthmic tract (figs. 38, 39), of the secondary visceral-gustatory tract; terminals and collaterals of the lemniscus systems; vast numbers of terminals of descending fibers of the tegmental fascicles; and collaterals of tr. tecto-bulbaris rectus. This evidently is a strategic region in the organization of both ascending and descending systems of conduction. Within this undifferentiated field the primordia of a considerable number of specific nuclei of other animals can be recognized, including specialized structures peculiar to fishes and a number of mammalian nuclei. Among these components (with their probable derivatives in more specialized species) the following can be identified:

1. The chief sensory nucleus of the trigeminus. Some ascending sensory V root fibers terminate here, though most of them more posteriorly, and some pass through to reach the commissura cerebelli and the corpus cerebelli of the same and of the opposite side. Secondary fibers go from this area upward into the cerebellum and forward with or without decussation as the primordial trigeminal lemniscus.


2. Similarly, some of the ascending root fibers of the VIII and lateral-line nerves terminate more laterally in the posterior isthmic neuropil (superior vestibular nucleus), and some of the VIII fibers reach the lateral (auricular) part of the cerebellum of the same and of the opposite side, the latter decussating in the com. vestibulo-lateralis cerebelli.


3. The longitudinal bulbar correlation tracts a and b reach the ventral part of the same neuropil. This field is incorporated within the cerebellum of mammals.


4. Spinal lemniscus fibers, as they recurve dorsal ward along the anterior face of the auricle to reach the tectum, are spread through the middle part of this neuropil, with collaterals and branched terminals within it (figs. 42, 43, 44).


5. Associated with the spinal and bulbar lemniscus and laterally of them in the isthmus are fibers of the lateral bulbo-tectal tract (fig. 11, tr.b.t.L), which is the probable primordium of the lateral lemniscus. These ascending fibers are mingled with descending fibers of tr. tecto-bulbaris posterior (fig. 12, tr.t.b.p.; '396, p. 606 and fig. 96). The position of the mixed fascicle in horizontal sections is shown in figures 31-34, here marked tr.t.b.p. Terminals and collaterals of the ascending lateral bulbo-tectal fibers spread throughout the posterior isthmic neuropil (figs. 33, 34), and these endings are comparable with those of the lateral lemniscus in the nucleus isthmi of the frog (Larsell, '24). This field also contains precursors of some of the nuclei of the lateral lemniscus of mammals, specifically the dorsal nucleus of the lateral lemniscus (Ariens Kappers, '29, p. 304; Clark, '33).



ISTHMUS 189

6. The nucleus isthmi, which is large in the frog (Larsell, '24), is probably represented in Amblystoma by a few outlying cells in the dorsal alba of the isthmus, which are more numerous in the larva ('42, pp. 244, 254). If present here, it is an insignificant vestigial structure.


7. The neuropil associated with the superior visceral-gustatory nucleus (fig. 33, nucris.n.) is an important field of sensory correlation. It is penetrated by dendrites of its nucleus ('42, fig. 43) and by terminals of the ascending secondary visceral tract ('25, fig. 19) and also by fibers from several other sources, including the lemnisci and collaterals of tr. tecto-bulbaris rectus (fig. 37; '42, fig. 67). This complex is very large in fishes (the Uebergangsganglion of Mayser) and is present, though small, in the frog (Larsell, '24).


8. The neighboring large cells of the motor tegmentum are undoubted precursors of several mammalian nuclei of this region, including the substantia nigra and the locus coeruleus.


PHYSIOLOGICAL INTERPRETATION

The afferent fibers which terminate in the isthmus come, directly or indirectly, from practically all parts of the brain, and they carry nervous impulses activated from all the major functional systems. The efferent discharge is also widely spread, but the chief distribution seems to be to upper levels of the medulla oblongata which are concerned with movements of the musculature of the head.


It is obvious that the isthmus contains a motor adjustor of prime importance. Its distinctive features can perhaps best be appreciated by comparing it with some of the other important centers of adjustment. The tectum of the midbrain and the dorsal thalamus, which in this animal is ancillary to it, comprise the dominant apparatus of correlation, on the sensory side, of all exteroceptive and proprioceptive systems. In the habenula the olfactory organ as an exteroceptor is tied in with the other exteroceptive systems. In the hypothalamus the olfactory system is similarly related with the visceral and gustatory systems. All fields of sensory correlation discharge into the corpus striatum, ventral thalamus, and peduncle, where they are integrated and co-ordinated primarily in the interest of mass movements of the skeletal musculature, such as those involved in locomotion. Efferent fibers from all the centers just enumerated and from others converge into a common pool in the isthmic tegmentum. Since apparatus adequate for executing the primary mass movements of the trunk and limbs is found elsewhere — primarily in the midbrain.


acting through the ventral tegmental fascicles upon the lower bulbar and spinal apparatus — and, since the isthmic tegmentum matures later in development, with notable enlargement preceding metamorphosis, it seems safe to infer that the isthmic apparatus is the chief regulator of the musculature concerned with feeding. The courses of the efferent fibers from the isthmus support this supposition. In Ambly stoma the feeding activities are thoroughly integrated movements, where posture of the body, conjugate movements of the eyes, and action of jaws, pharynx, and esophagus are organized as a "total pattern," as Coghill ('36) has demonstrated. For further consideration of the olfactory and hypothalamic components of feeding reactions see pages 210 and 252 and 19346, page 384.


It is evident that control of feeding reactions is not the only function of the isthmic sector. For instance, it has been shown by Aronson and Noble ('45) that the removal of the entire brain of the male frog in front of this region (including the tectum, cerebellum, and anterior part of the tegmentum) did not interfere with spawning movements but that "lesions in the tegmentum at the level of the motor nucleus of the trochlear nerve markedly disturbed or completely abolished these spawning responses."

In addition to this apparatus of activation of behavior, there are inhibitory functions here also, implying a participation in motor control of wide import. The intricate connections of the isthmic tegmentum described in the next chapter are part of this inhibitory apparatus.


The unitary character of most of the activities of Amblystoma, involving the synergic action of large masses of musculature in invariable orderly sequence, with relatively less capacity for the local autonomous action of the individual members than in higher animals, explains the simplicity and relative homogeneity of the nervous apparatus involved — also why the isthmic tegmentum is so large in Amblystoma. The integrated functional complexes which here are controlled from this major pool are individually differentiated in man, with corresponding specialization and segregation of the apparatus of the several component parts; and, parallel with this differentiation and increase of local autonomy, the central control apparatus has been transferred from the stem to the cerebral cortex. Accordingly, the isthmic tegmentum, which bulks so large in the brain of Amblystoma (and in the early human embryo), is dismembered in the adult man, and its parts are submerged within surrounding, structures.



CHAPTER XIV INTERPEDUNCULAR NUCLEUS

THE interpeduncular nucleus, like the habenula with which it is connected, is one of the most conservative structures in the brains of vertebrates, but in urodeles it has some characteristics which have not been reported elsewhere. The general features are outlined on page 46, and the details are described here, together with some speculations about the probable functions. The chief connections are shown very diagrammatically in figures 19 and 20, the spiral endings of the tractus habenulo-interpeduncularis in figure 50, the composition of the glomeruli in figure 84, and some typical connections of its neurons in figure 83.


COMPARATIVE ANATOMY

This nucleus was discovered by Forel in 1872 and described in the rabbit in 1877, with confirmation by von Gudden in 1881 and Ganser in 1882. It is a constant feature in all vertebrate brains, much larger in some lower groups than in the higher. In urodeles it is a well-defined column of cells, embracing the ventral angle of the ventricle and extending from the fovea isthmi backward to the level of the V nerve roots. It is a ventromedian structure, differentiated in situ in the floor plate and the adjoining borders of the basal plates'. The fovea isthmi marks the anterior end of the embryonic floor plate as defined by Wilhelm His, and this plate is generally regarded as a nonnervous structure characterized by specially differentiated ependyma (Kingsbury, '30, p. 182) ; but Coghill ('24, Paper III) has shown that neuroblasts are differentiated here intrinsically. Some of these persist in the adult medulla oblongata as nucleus raphis, and between the fovea isthmi and the V nerve roofes this differentiation goes much further, producing the interpeduncular nucleus, as pointed out in my description of this nucleus of Necturus ('34c).


The differentiation is primarily isthmic, as was recognized by van Gehuchten ('00, 2, 199). In Amblystoma it extends spinalward into trigeminal territory, and below the level of the V nerve roots it is continuous with the nucleus raphis of the medulla oblongata. In cyclostomes the fasciculus retroflexus extends spinalward as far as the calamus scriptorius (Johnston, '02, p. 31 and figs. 8-16, 30). A mesencephalic sector of the nucleus, rostrally of the fovea isthmi, has been described in some vertebrates, and there is some evidence of this in Necturus ('34c) and Amblystoma. It has been found in some fishes, reptiles, and mammals; but in urodeles there is no group of differentiated cells in the midbrain comparable with this nucleus in the isthmus, and the secondary connections of mesencephalic terminals of fibers from the habenula are radically different from those of the habenulo-interpeduncular tract. In this account, accordingly, the description relates only to those structures spinal ward of the fovea which have habenular connection. This involves a terminological inconsistency, for in my analysis of the brain of Amblystoma the "peduncle" is defined as a strictly mesencephalic structure, and here none of the interpeduncular nucleus lies in the mesencephalon. Since this nucleus, however, is so obviously homologous with the structure so named in mammals, the mammalian name is retained.


The published description of the interpeduncular nucleus of Necturus ('34c) was based on insufficient material and evidently is incomplete. The structure there described is similar to that of Amblystoma but much simpler. Whether this simplicity indicates that Necturus is really more generalized (or more retrograde) or merely that the material available does not adequately show the actual structure cannot be determined. Probably both these suppositions are true. The more abundant Golgi material of Amblystoma reveals a wealth and intricacy of detail that baffle analysis, and much remains obscure. The exceptionally large and elongated nucleus in these animals is favorable for analysis, but the fact that all the connections are unmyelinated and dispersed in very dense neuropil makes accurate description impossible, except where elective silver impregnation isolates particular components. The erratic incidence of these impregnations and the well-recognized hazards of error in their interpretation necessitate great caution and restraint in drawing conclusions. Most of the observations here recorded have been seen in many preparations and can be adequately documented. Some of them, however, rest on slender and ambiguous evidence and require confirmation. The attempt to assemble these fragmentary observations into a coherent unity, such as is shown in figures 19, 83, and 84, is fraught with danger; and it is emphasized that these pictures and the physiological theories expressed are not regarded as final. They present my opinion of the things seen, subject to revision when information is more complete. A few years ago a brief summary of the observations made up to that time was pubhshed ('396, p. 584), and details were later added ('42, figs. 40-43). Here all these data are assembled, some errors in a brief reference in the paper of 1927 (p. 280) are corrected, and further observations and interpretation are recorded.


HISTOLOGICAL STRUCTURE

In Weigert and cytological preparations the cells of this nucleus are densely crowded under and adjacent to the floor of the ventricle. Between the evels of the III and IV nuclei this gray column is thin. Ventrally of it there are, successively, the ventral median tegmental fascicles of group (1), the decussating fibers of the ventral commissure, and the superficial interpeduncular neuropil (figs. 60, 80-84, 92; '25, figs. 8, 9; '27, fig. 39; '42, figs. 30, 41-43). At the level of the IV nucleus the median fascicles, /.m.^.(jf), turn away from the midplane, and the texture of the ventral medial tissue posteriorly of this level is more open. Some of the cells of the interpeduncular nucleus lie more ventrally among strands of the ventral commissure (fig. 91). This gray is continuous dorsally on each side with that of the isthmic and trigeminal tegmentum, but in cell preparations the boundary is usually evident. Golgi sections show that at this boundary the cells are of transitional form, with some dendrites ramifying in the interpeduncular neuropil and others directed laterally into the alba of the tegmentum (figs. 61, 65, 66).


EPENDYMA

The description of the ependyma of Necturus ('34c, p. 118) applies with little change to Ambly stoma. The ependyma at the fovea isthmi is distinctive (fig. 79) and similar to that bordering the sulcus isthmi (p. 181). Posteriorly of this, the ependymal elements are of two forms. In and near the median raphe they are compact fascicles of thick, thorny fibers (fig. 70). More laterally (figs. 63, 64, 81), each element has a much more widely branched arborization of slender, thorny fibers, many of which end at the pial surface in bulbous enlargements. The external limiting membrane is apparently composed of these expanded ependymal terminal bulbs. Both these types of ependyma are seen in midlarval stages ('396, figs. 89, 94, 95).


The interpeduncular ependyma is an exceptionally dense mat of widely branched and closely interwoven fibers. Many of these branches do not reach the pial surface but arborize throughout the interpeduncular neuropil and especially within the interpeduncular glomeruli (described below), of which they form an integral part. The participation of specially modified ependyma in the formation of these glomeruli raises interesting questions. There is no obvious demand for unusual mechanical support here to account for especially strong ependymal framework, except perhaps in the median raphe, where the ependymal elements are more sturdy. Elsewhere in the interpeduncular field the slender ependymal fibers branch freely, and in the glomeruli they have tufted endings similar to those of dendrites, to be described shortly. This elaboration of the ependymal fabric within specialized synaptic fields suggests that the ependyma has some part to play in metabolism at the synaptic junctions.


INTERPEDUNCULAR NEUROPIL

Ventrally of the interpeduncular gray and the commissure there is a sharply circumscribed area of very dense superficial neuropil, oval in cross-section, which in some of my published figures is marked nuc.inp. It is preferable to apply the word "nucleus" to the gray area only. This band of differentiated neuropil is, accordingly, currently called the "specific interpeduncular neuropil" (inp.n. or nucinp.n.), though it should be kept in mind that the entire interpeduncular field is permeated by neuropil, of which this is a specialized part. Its axonic component is a web of interlaced fibers, which is continuous with that of surrounding parts. Imbedded within this fabric are two specialized structures — the spiral terminals of the habenulo-interpeduncular tract and the glomeruli. The whole interpeduncular area is richly vascularized with capillary net and is a synaptic field of great importance. This activity evidently is especially concentrated in the glomeruli and the area of specific superficial neuropil.


The neuropil of the interpeduncular field is continuous dorsally with the deep neuropil of the gray of the isthmic and trigeminal tegmentum and laterally with the intermediate neuropil of the alba of these tegmental areas, as shown very inadequately in figures 60-66. Posteriorly it is continuous with the diffuse neuropil which pervades both the gray and the white substance of the medulla oblongata (figs. 79, 80, 81). In these drawings the structure of the neuropil is shown greatly simplified and schematized, for it is impossible on this scale of magnification to portray the intricacy of its texture. Since all nerve fibers related with the interpeduncular system are unmyelinated and most of the tracts are dispersed in the neuropil, analysis is


INTERPEDUNCULAR NUCLEUS 195

very difficult and, indeed, impossible except when aided by elective impregnations. The interpretations given in the text and in the diagrams (figs. 19, 83, 84) are based on such evidence. The evidence is incomplete and in many preparations ambiguous, and attention is again called to the fact that the conclusions reached are tentative and subject to revision in detail. It is believed, however, that the main features of the histological analysis are based on adequate evidence and are reliable.


GLOMERULI

Glomerulus-like structures have been known in the mammalian interpeduncular nucleus since 1877 (Forel). In Amblystoma these are small and very numerous elongated areas of very dense neuropil. They are distributed throughout the interpeduncular neuropil, in many places densely crowded; most of them are oriented vertically and extend downward into the ventral area of specific neuropil. Histological analysis of their structure is even more difficult than is that of the rest of the interpeduncular neuropil, for each of the constituent elements may present quite different appearances in Golgi sections, depending on the quality of the impregnation.


These glomeruli resemble in some respects those of the olfactory bulb ('246), though of much more complicated structure. In addition to the capillary net, there are three constituent elements: (1) a condensation of the ependymal framework as already described; (2) tufted terminals of dendrites ; and (3) tufted axonic terminals composed of thin contorted fibers interwoven with the dendritic terminals. Any one of these three constituents may be electively impregnated, or two or three of them may be seen in the same section. In some Golgi preparations the three components are so similar that they cannot be distinguished with certainty except where terminals can be followed to their connections outside the glomerulus, though when two of them are impregnated in the same section the difference betw^een them is usually obvious. In the preparation from which figure 66 was drawn, axons and dendrites are intermingled in each glomerulus. Only the dendritic component is drawn on the left side, and only the axonic component on the right. The visible tufted axons are from interpeduncular neurons.


Most of the glomerular dendrites come from cells of the interpeduncular nucleus, though some of them are branches from the transitional neurons at the ventral border of the tegmental gray. The axonic tufts, as far as observed, come from neurons of the interpeduncular nucleus, cells of the overlying tegmentum, and collaterals of fibers of tr. tegmento-bulbaris. No fibers from other sources seem to participate significantly in the formation of glomeruli. Many axons from the three sources mentioned do not enter glomeruli but arborize in the surrounding neuropil. Glomeruli with dendrites and axons derived chiefly from the interpeduncular nucleus tend to be disposed horizontally, the others vertically. The horizontal glomeruli have been illustrated from the larva ('396). Figure 58 of the paper cited shows two of the dendritic tufts, and figure 61 shows the axonic component of several of these horizontal glomeruli. Most of the horizontal glomeruli are imbedded in the general neuropil dorsally of the ventral specific neuropil. The vertical glomeruli generally extend downward into the specific neuropil, where they are penetrated by fibers of the interpeduncular spiral.


NEURONS

In transverse Golgi sections the dendrites of the interpeduncular neurons are directed ventrally, passing through the ventral commissure. Within the interpeduncular neuropil they branch widely, and many of the branches have tufted endings in the glomeruli (figs. 19, 65, 66, 83, 84). Longitudinal sections show that the spread of these dendrites is much greater fore and aft than it is transversely (figs. 80, 81, 82; '396, figs. 57, 58), and the tufted terminals may be oriented either vertically or horizontally. Some of the cells of the isthmic tegmentum at the border of the interpeduncular nucleus may send dendrites into the interpeduncular neuropil, and these, too, may have tufted endings ('396, figs. 78, 95). Sagittal sections in which both interpeduncular dendrites and ependyma are fully impregnated give spectacular demonstration that the interpeduncular formation ends anteriorly at the fovea isthmi and isthmic sulcus. These interwoven elements directed anteroventrally form a dense mat, which ends abruptly at the locus of the sulcus isthmi. Posteriorly, the interpeduncular formation merges insensibly with the trigeminal tegmentum.


The slender axons of the interpeduncular neurons arise from the cell body or the base of the dendrite (figs. 65, 66, 70, 84; '396, figs. 61, 66) and immediately branch profusely. These fibers take tortuous courses within the interpeduncular neuropil and can be followed only when electively impregnated. In such preparations the longer branches are seen to be directed spinalward and to form the dorsal and ventral interpedunculo-bulbar tracts, as in Necturus ('30, p. 80; '34c, p 120).



INTERPEDUNCULAR NUCLEUS 197

Other branches spread widely in the general interpeduncular neuropil, of which they form an important component. Most of these fibers are dispersed, but some of them form compact longitudinal tufts within the glomeruli (fig. 84: '396, figs. 41, 61).


AFFERENT CONNECTIONS

Fibers from remarkably diversified sources terminate in the interpeduncular neuropil. The widely branched terminals of some of these fibers spread diffusely ; the endings of others are specialized in diverse ways. The dominant member of this complex evidently is tr. habenulo-interpeduncularis, which is the chief component of the f. retroflexus as described in chapter xviii. The interpeduncular connections of this tract are similar in Amblystoma and Necturus ('34c).


Tractu.s habenulo-interpediaicularis. — The two bundles of unmyelinated fibers converge immediately below the fovea isthmi, and here most of their fibers decussate close to the ventral surface. Their f urthei- course is shown in figure 50. Immediately after crossing, each fiber reverses its course, and this is repeated so that a compact spiral is formed, which extends the entire length of the interpeduncular neuropil, diminishing spinalward. This spiral is coextensive with the specific interpeduncular neuropil and is its most characteristic feature. There are many series of sections cut in various planes in which this is clearly shown. In some of these, only one of the habenulo-interpeduncular tracts is impregnated (figs. 51, 52, 56, 57,. 58).


Figures 53 and 54 are drawn from a specimen in which tr. habenulo-interpeduncularis is impregnated on both sides at the origin in the habenula. The impregnation fails on the left side from the level of the dorsal thalamus downward. On the right side it is well stained as far as the decussation, below which only a few fibers are blackened, thus revealing clearly the courses of individual fibers, one of which is drawn in figure 54 through several turns of the spiral (cf. fig. 55). These fibers are slightly varicose and thorny. No specialized endings of any sort have been seen, and they have not been observed to branch except for occasional short forked terminals. The spiral penetrates the glomeruli, and its fibers are in synaptic connection with the glomerular dendrites.


Transverse sections show that the spiral is flattened dorsoventrally ('42, figs. 40-43). In the specimen from which these figures were drawn the spiral is fully impregnated, and few other fibers of the interpeduncular neuropil are stained so that the structure of the spiral is clear. No fibers leave the spiral to pass beyond the specialized interpeduncular neuropil. There is a small fascicle of uncrossed fibers at each lateral border of the sjjiral, and these a})parently enter the spiral at successive levels. Another series of transverse sections (no. 'i'iB'i) from the same lot of young adult s])ecimens shows almost identically the same structure.


Not all the habenulo-intcrpeduncular fibers enter the initial decussation. As mentioned above, a considerable fraction of them, including some of thicker caliber, descend uncrossed along the lateral borders of the interpeduncular neuroi)ih which they enter farther spinalward. In some preparations both crossed antl uncrossed fibers are imi)regnated, in some only the spiral fibers (fig. 50), and in one sagittal series only the uncrossed fibers are blackened (fig. S'i) . In the latter case the f. retroflexus is partially impregnated on each side for its entire length from the habenula to the interpeduncular nucleus. At the decussation none of the crossing fibers or the spiral fibers below it are darkened, but there is a large fascicle of uncrossed fibers, which descends laterally of the decussation and the spiral for almost the entire length of the interpeduncular nucleus. These fibers terminate within the interpeduncular neuropil among those of the unimpregnated spiral.


In 1894 van (iehuchten described thef. retroflexus of the trout. The origin in the habenula is similar to that of Amblystoma, and the course is the same. These fibers decussate in the interpeduncular nucleus and recross with much branching, but they do not form a compact spiral. They are varicose and have no specialized terminals. Simpler spiral terminals of mammals have been many times described (rabbit, von Gudden, '81; mole, Ganser, '82, p. 682; mouse, Ramon y Cajal, '11,2, 27-1; Calderon, '28; and many others).


Tractuff mamillo-i)ifi'rpeduncuIaris. — These fibers comprise group (2) of the tegmental fascicles ('3G, pp. 303, 338, figs, 3, 8). They arise in the dorsal part of the hypothalamus in company with similar fibers for the peduncle and tegmentum (figs. 19, 71, 79). The mixed bundle decussates partially in the retroinfundibular commissure, which is component 1 of the commissure of the tuberculum posterior (p. 302). The interpeduncular fibers, crossed and uncrossed, pass the fovea isthmi at the ventral surface close to the mid-plane and then descend along the lateral border of the specific interpeduncular neuropil (figs. 27-30, 60; '396, figs. 22, 41, 42, 57-61; '42, fig. 3), within which they end in open arborizations. They do not join the spiral or participate


INTERPEDUNCULAR NUCLEUS 199

notably in the formation of glomeruli. Most of them end in the rostral half of the interpeduncular neuropil.


Tradus olfacto-pedujicularis. — This important component of the basal forebrain bundles (fig. 101; 'SOb, p. 534, figs. 1, 57), after contributing fibers to the hypothalamus and peduncle, continues spinalward along the lateral border of the specific interpeduncular neuropil. The endings of these fibers are similar to those of tr. mamillointerpeduncularis (figs. 18, 19, 21, 25-30, 53, 54, 59, 72, 82). This tract in the interpeduncular region lies laterally and dorsally of tr. mamillo-interpeduncularis, extending spinalward beyond the posterior end of the interpeduncular nucleus. There are some terminals of both this and the preceding tract in the dispersed interpeduncular neuropil.


Nervus terminalis. — In Necturus the longest fascicles of this nerve root probably reach the interpeduncular neuropil ('34c, p. 124). In Ambly stoma a small fascicle was observed by McKibben ('11, p. 270) to reach the "dorsolateral part of the hypothalamus and the interpeduncular region." Some of these fibers may connect with the interpeduncular nucleus, but this has not been demonstrated.


Tradus tegmento-ititerpedimcularis. — This is a very extensive connection, which takes two forms. There is, first, a system of short axons dispersed in the neuropil of the gray and the deep neuropil of the alba which arise from small cells of the dorsal, isthmic, and trigeminal tegmentum and descend directly to the interpeduncular neuropil. Some may come from the tectum. They are especially concentrated in the f. tegmentalis profundus (p. 286). A second type of connection is made by collaterals of the thicker axons of tr. tegmentobulbaris at their decussation in the ventral commissure.


1. The short tegmento-interpeduncular fibers of the first grouj) are unmyelinated. Their general arrangement is well shown in the sagittal sections, figures 79 and 80. They arborize in the neuropil of the gray of the interpeduncular nucleus and at all depths of the neuroj^il of the alba, forming an important component of the diffuse neuropil. Transverse sections show that, in addition to the diffuse spread of these fibers in the neuropil, there are also tufted endings which form part of the axonic component of the glomeruli (figs. 61, 62, 63, 66, 84).


2. The tr. tegmento-bulbaris cruciatus as described in the larva ('396, p. 590) is a series of thick axons of large cells of the isthmic and trigeminal tegmentum, which decussate dispersed in the ventral commissure and then descend in the ventral and ventrolateral fascicles of the medulla oblongata. Horizontal sections of one of these larvae show slender collaterals of these decussating fibers, which are directed posteroventrally into the interpeduncular neuropil and here have tufted endings in the interpeduncular glomeruli ('396, p. 600 and figs. 46 and 59). Figures 67 and 68 of the present work show similar collaterals, as seen in transverse sections from the same lot of larvae. Each compact glomerulus-like tuft contains terminals of several of these collaterals. Figure 69 shows similar collaterals in an adult brain. In this preparation only a few of these decussating fibers are impregnated, and these can be followed from section to section. Nothing else is stained here, so there is no opportunity for confusion. Figure 60 is from another adult specimen which shows some of these features more clearly, though the impregnation is less selective. This section is from the series illustrated in 1927 and is a few sections rostrally of figures 39 and 40 of that article. The tufted axonic terminals shown there are similar to the one seen in the present figure 60, and they are now known not to be derived from the mamillo-peduncular tract as there suggested ('27, p. 280).


Tertiary visceral-gustatory tract. — These fibers, as elsewhere described (p. 169), pass from the secondary visceral nucleus in the isthmus to the area ventrolateralis pedunculi. They lie close to the gray in the posterior lip of the sulcus isthmi and form one component of the complex f. tegmentalis profundus. Some fibers of this fasciculus decussate in the ventral commissure dorsally and spinalward of the decussation of the f. retroflexus, and here in some preparations there is evidence that fibers of the tertiary visceral tract turn spinalward to enter the interpeduncular neuropil. Several other systems of similar fibers are mingled here, and, in the absence of elective impregnation of this connection, its presence remains doubtful. Whether or not this direct connection exists, the visceral-gustatory system is related less directly with the interpeduncular nucleus by way of the hypothalamus. The secondary visceral nucleus is connected with the dorsal (mamillary) part of the hypothalamus and from here the large tr. mamillo-interpeduncularis may transmit visceral sensory and gustatory influences to this nucleus.


Other afferent fibers to the interpeduncular nucleus have been described in various animals. Gillilan ('41) reports that in bats and rodents the basal optic tract sends some fibers directly to this nucleus.


This connection may be present in Amblystoma also, though we have no demonstration of it.


EFFERENT CONNECTIONS

The widely branched axons of the interpeduncular neurons apparently ramify throughout the entire interpeduncular field, and some of them pass beyond this field. The latter go out in three directions and are designated as three tracts named after their terminal connections — tr. interpedunculo-bulbaris dorsalis and ventralis and tr. interpedunculo-tegmentalis (figs. 19, 83, 84). All these efferent fibers go out in a web of mixed neuropil, within which the two bulbar groups are more or less clearly fasciculated in elective impregnations. The tegmental group of efferents are nowhere fasciculated, and we have no Golgi preparations in which they are electively impregnated.


Sagittal and transverse sections show that the interpeduncular neuropil extends dorsally into the isthmic and trigeminal tegmentum. Most of these fibers trend dorsoventrally, these comprising part of the mixed system, termed "fasciculus tegmentalis profundus." These fibers are marked f .teg. p. in the following figures: 29, 30, 31, 79; '42, figs. 30, 41, 42, 43. Elective impregnations have revealed some of the components of this fasciculus, including the brachium conjunctivum, tertiary visceral tract, and tr. tegmento-interpeduncularis. The latter is shown in figures 61, 62, 63, 65, and 66. These sections show also (though not clearly drawn in the figures) that from the interpeduncular neuropil a large number of fibers swing dorsolaterally into the alba of the tegmentum, where they engage dendrites of tegmental neurons. These fibers are interpreted as an interpedunculo-tegmental connection. In the light of the mammalian connections of the chief efi^erent tract from the interpeduncular nucleus to the dorsal tegmental nucleus discussed below, it seems probable that in Amblystoma the small-celled central nucleus of the isthmic tegmentum contains the precursor of the dorsal tegmental nucleus, though this evidently is only one of its relationships.


The interpedunculo-bulbar connections are clearly seen in our material. These fibers emerge from the diffuse interpeduncular neuropil in dorsal and ventral strands, which are connected with each other and with the surrounding neuropil, as shown in figures 79 and 81. The tr. interpedunculo-bulbaris ventralis descends for an undetermined distance at the lateral border of the interpeduncular neuropil and farther spinalward in the lip of the ventral fissure. The dorsal tract descends dorsally of the f. longitudinalis medialis (fig. 92). As pointed out below, this position of the dorsal tract agrees with that of the dorsal longitudinal fasciculus of Schiitz, and its fibers take a similar course, turning laterally to spread in the trigemino-facial tegmentum and as far back as the level of the motor nucleus of the IX nerve. In fact, the dorsal bulbar tract is really an extension posteriorly of the tegmental tract, for these comprise a continuous series of efferents to the tegmentum. The dorsal longitudinal fasciculus of Amblystoma is not a well-formed tract as in mammals, but comparable fibers can be recognized in the deep neuropil of the brain stem. In both lower and higher animals this complex seems to be concerned with general central excitatory state and disposition of the individual rather than with specific sensori-motor responses (pp. 208, 217). The inteipedunculo-tegmental and dorsal interpedunculo-bulbar system of fibers is evidently the primordium of an important component of the "caudal division of the dorsal longitudinal fasciculus" as described in the opossum by Thompson ('42), This morphological conclusion rests, I think, on sufficient evidence, whether or not the following physiological speculations can be confirmed experimentally.


INTERPRETATION

Though much remains obscure, the main features of the histological structure and connections of the urodele interpeduncular nucleus are now clear, and these are assembled schematically in figures 19, 83, and 84. The most striking feature is the ventral plaque of specific neuropil containing spiral terminals of the habenulo-interpeduncular tract and glomeruli. Other afferent fibers enter both the specific and the diffuse interpeduncular neuropil and are so arranged as to permit their activation from almost all parts of the cerebrum. All these fibers come from areas of intermediate-zone type, not from primary sensory centers.


These afferents are of four general classes, depending on the source of their activation: (1) olfactory from the hemisphere, ending in the diffuse neuropil ; (2) olf acto-somatic from the habenula, ending in the specific (spiral) neuropil; (3) olf acto-visceral from the hypothalamus, ending in diffuse neuropil; and (4) tegmento-interpeduncular from the overlying tegmentum and ending in both diffuse neuropil and glomeruli. Those of the last class come from somatic sensory fields of correlation with no appreciable olfactory component. The glomeruli provide apparatus for nonspecific summation and reinforcement.



INTERPEDUNCULAR NUCLEUS 203

Each one of the other classes of afferents may have its own specific type of trans-synaptic transmission, depending on the histological and chemical structure and the strength and timing of volleys delivered.


Provision is thus made for discriminative responses to a considerable range of different kinds of activity that may be in process elsewhere in the brain — olfactory, visceral, somesthetic, and so on. The number of interpeduncular neurons activated and the temporal rhythm of their discharge may be determined not only by the algebraic sum of afferent impulses received but also by their timing. It is believed that this sort of analysis is characteristic of the nervous adjusters.


The outflow from the interpeduncular nucleus is distributed in part to remote regions and in part within the interpeduncular field, the latter providing apparatus for summation and intensification of the discharge. The remote effects of this outflow will depend on the location, structure, and connections of the motor pools entered by the efferent fibers.


The entire interpeduncular field is a well-circumscribed motor pool, in Sherrington's sense, in which every intrinsic nervous element is subject to activation by a wide variety of nervous impulses coming from diverse sources, each of which has a specific form of synaptic junction. The different structural patterns of these synapses presumably involve physiological differences in the type of activation. The dominant member of this complex of afferents evidently is the spiral of habenulo-interpeduncular fibers; every fiber of this tract may activate (or inhibit) every neuron of the nucleus, and, indeed, in successive turns of the spiral it may act upon several dendrites of the same neuron. There is no provision for separate localization of specific function here. The nucleus inevitably acts as a whole. But there is provision for summation of the effect in a large way. The diffuse neuropil also may act upon these neurons. Volleys may be discharged into it from many sources, each of which may reinforce (or inhibit) the activity instigated by the spiral.


Though temporal summation at an individual synapse is regarded as practically inoperative (Fulton, '43, p. 58), there is ample provision in this pool for spatial summation. The glomeruli present a series of difficult problems. Each of these small nodes of denser neuropil contains tufted dendritic terminals and tufted axonal terminals, and the latter come from two sources, intrinsic and extrinsic. The horizontal glomeruli seem to be especially adapted for summation of the discharge from the nucleus. If one of these neurons is experiencing sul)liminal activation, a collateral discharge from an active neuron may impinge upon it in a horizontal glomerulus, and the sum of the two excitations may be sufficient to fire the element. In a similar way, terminals of intrinsic fibers in the vertical glomeruli may reinforce the action of the extrinsic fibers. The glomerular activities may reinforce (or inhibit) those of the spiral, or conceivably they may act independently if the habenular component is inactive.


The observed structure seems to favor the supposition that, whatever may be the functions performed by this complex, they are of generalized type. The apparatus which patterns the specific reflexes, both somatic and visceral, is located elsewhere in the brain stem. The intrinsic structure of the interpeduncular field seems to preclude any well-defined local representation of different functions within it. The nucleus apparently acts as a unit, and any functional specificity that may exist would be in terms of differential excitation at the various forms of synaptic junction or of the functions of the motor fields into which its efterent fibers discharge. The ventral bulbar tract descends in the primary motor column of Coghill, and this suggests that it is in some way concerned with mass movements of the musculature of the trunk and limbs. The tegmental and dorsal bulbar efferent tracts form an aftatomical unit, and they probably are similarly related physiologically. These fibers discharge into a motor field within which the reflexes of the head are organized, notably those concerned with feeding. The interpeduncular complex seems to be so organized as to act as a whole without intrinsic localization of function. What part, then, does this complex play in patterned behavior?

In the first place, it is a noteworthy feature of the interpeduncular system that it is connected "in parallel," as the electrician would say, with the major systems of fore-and-aft conduction of the brain stem. The olfactory, optic, and lemniscus systems on the afferent side and the strio-pedunculo-bulbar systems on the efferent side, with their related centers of adjustment, seem to provide the apparatus which patterns overt behavior. The interpedunculo-habenular system is in addition to this apparatus and in some way ancillary to it.


In noncorticated vertebrates the activators of the motor zone receive their nervous excitations mainly from three regions which are


INTERPEDUNCULAR NUCLEUS 20.5

physiologically distinctive: (1) the somatic sensory field of the tectum and dorsal thalamus, (2) the visceral field of the hypothalamus, and (3) the olfactory field of the cerebral hemispheres. The parts of this primary activator system of the cerebrum are interconnected, and all are represented in the higher adjusters of the hemispheres.


The interpeduncular complex is separate from this great system of activators except at its upper and lower ends. It is activated from the same three physiologically distinctive regions, and its efferent impulses are discharged into the same lower motor fields. These relationships are in some respects similar to those of the cerebellum with the activators of the skeletal musculature. The cerebellum does not pattern behavior, but it acts upon the motor systems as going concerns, facilitating the activity in process by reinforcement and inhibition, appropriately placed and timed. The structure and connections of the interpeduncular nucleus suggest that it is similarly related with the great descending (extra-pyramidal) systems of the cerebrum, and specifically with those concerned in the feeding reactions.


In the light of Coghill's definition of the amphibian reflex "as a total behavior pattern which consists of two components, one overt or excitatory, the other covert or inhibitory," and making application to a specific problem of the general discussion of reflex and inhibition in chapter vi, an attractive hypothesis might be framed along the following lines:

The brain stem may be conceived as a labile, equilibrated, dynamic system, within which the excitatory and inhibitory components are balanced against each other in reciprocal interaction at every successive phase of motor activity. The efferent activating systems of patterned behavior discharge through the basal forebrain bundles and the tegmental fascicles (fig. 6). Above and below this central core of activating fibers and parallel with them, there are inhibitory systems of fibers which have the same origins and terminations as do the activating systems but which pursue different courses with different connections. Dorsally there is an olfactosomatic inhibitory system centered in the habenula, and ventrally there is an olfacto-visceral inhibitory system centered in the mamillary region of the hypothalamus. Both these inhibitory systems converge into the interpeduncular nucleus (figs. 19 and 112). Here the inhibitory influences are integrated and distributed to lower motor centers as going concerns, in accordance with momentarily changing physiological requirements. Presumably not all components of the action system are under this sort of regulatory control by the interpeduncular complex, for, as pointed out below, there are other local fields with inhibitory functions.


The unusually large size of the interpeduncular nucleus of urodeles may be correlated with the fact that in the normal behavior of these animals total inhibition is a conspicuous feature (p. 77). Adopting Sherrington's terms, there is usually a long period of inhibition between the anticipatory and the consummatory phases of the reaction, notably in the feeding reflexes.


This hypothesis can have only academic interest unless and until it is confirmed by physiological experiment or clinical evidence, but it may have heuristic value as an indication of profitable points of attack in further exploration. From such scanty evidence as is now available, it seems to the writer to rest on an insecure foundation because it probably is an oversimplification of the problem. Our knowledge of the mechanism of central inhibition is so inadequate that inferences about physiological action from anatomical evidence alone are unsafe in this field, though such inferences are justifiable in many other fields where adequate physiological controls are available and have been made. Keeping these admonitions in mind, let us inquire further into the possible role that inhibition may play in the interpeduncular complex.


Though we do not know exactly how inhibition is effected in the central nervous system, we do know that in some parts of the brain the dominant function is activation, and in other parts it is inhibition. Current analysis of the primate cerebral cortex has revealed specific activating fields in the motor cortex and also specific inhibitory fields, the so-called "suppressor bands" (p. 79). In the stem the head of the caudate nucleus and the reticular formation of the upper medulla oblongata are also known to have inhibitory functions. That the interpeduncular nucleus is a similar field in which inhibition is dominant is a legitimate working hypothesis that can be tested experimentally. This, of course, does not imply that inhibition is the only function here localized, for inhibitory action is balanced against excitatory action in normal behavior. It follows that experimental disturbance of this normal balance may come to expression in so great a variety of ways depending on so great a variety of factors that it is difficult to devise crucial experiments.



INTERPEDUNCULAR NUCLEUS 207

This difficulty is illustrated by the published report of experiments by Aronson and Noble ('45) designed to localize in the brain the mechanisms which mediate specific components of the mating behavior of male frogs. In hundreds of experiments involving local ablations or transections, there were a few which gave clear-cut evidence of precise localization of function. In this connection Dr. Aronson writes me: "I believe that, while some activities, e.g., warning croak, ejaculatory response, and release of the female by the male at the termination of oviposition, are quite definitely localized, the mechanism controlling other phases of mating behavior are rather diffuse. In the latter cases more precise experiments w^ould, I think, demonstrate diffuseness rather than localization."

The authors' very conservative analysis of their protocols does reveal a few areas with local specificity of function. Among these is the clasp reflex, which is one of the most characteristic features of this mating behavior. Normally, the mating clasp is relaxed or inhibited immediately after oviposition, and this presumably involves inhibitory action. There is evidence that the primary center of this reflex is in the spinal cord and that the inhibitory phase is under some measure of cerebral control. Relaxation or inhibition of the clasp is quite consistently abolished by destruction of the preoptic nucleus or dorsal part of the hypothalamus, but this result does not consistently follow injury to the habenulae or interpeduncular nucleus. Though the clasp involves the use of skeletal muscles, the reflex as a whole is a visceral reaction related with oviposition and ejaculation of sperm. Cerebral control of the release of the clasp in the hypothalamus may be transmitted to the upper levels of the spinal cord by a direct pathway via tr. mamillo-tegmentalis and ventral tegmental fascicles of group (3), as described on page 278, without involvement of the interpeduncular complex. But Dr. Aronson would not limit the inhibitory influence resulting in release to this pathway. Again in personal correspondence he says: "While the spinal clasp center might be quite limited, the mechanism inhibiting or modifying the spinal clasp reflex is very diffuse, especially in the midbrain

and diencephalon The interpeduncular nucleus might well be

involved here."

An interesting series of experiments on the Japanese toad reported by Kato ('34) reveals an inhibitory center for the muscles of the contralateral limbs in a region described as "on the anterior end of the lamina terminalis," but not accurately defined. The diagrammatic figures seem to place it in the region of the anterior commissure between the septum and the preoptic nucleus. The efferent pathway from this center decussates in the vicinity of the interpeduncular nucleus. Further information about the connections and physiological significance of this center is desirable.


These experiments yield no crucial evidence of participation of the interpeduncular nucleus in the components of mating behavior under investigation or of movements of the limbs in general. This nucleus may play a more specific part in some other types of behavior; and some experiments indicate that this is true.


The only direct experimental evidence about the functions of the interpeduncular nucleus of mammals known to me is a recent note by Bailey and Davis ('42a) in which they describe a "syndrome of obstinate progression" in cats following destruction of this nucleus. This is a locomotor impulsion, which continues without remission until the death of the animal. Mettler and Mettler ('41) have produced similar symptoms by lesions involving the head of the caudate nucleus, and more recently Mettler ('45, p. 180) describes in primates a direct connection from the globus pallidus to the interpeduncular nucleus. The caudate is supposed to be one link in a chain of conductors with inhibitory functions (Fulton, '43, p. 456) and the syndrome following the destruction of the interpeduncular nucleus suggests to me that this nucleus also may be an inhibitor. The "obstinate progression" may then be interpreted as a release phenomenon.


In mammals efferent fibers from the interpeduncular nucleus have been described with distribution to neighboring parts of the brain stem, of which by far the most important is the large pedunculotegmental tract to the dorsal tegmental nucleus and the related dorsal longitudinal fasciculus of Schiitz. In the recent study of this fasciculus of the opossum by Thompson ('42), she describes this interpeduncular connection as an important component of it. In this marsupial the dorsal longitudinal fasciculus is a complicated system of fibers, chiefly descending, connecting the deeper areas of gray between the diencephalon and the spinal cord.


The structure and connections of the dorsal longitudinal fasciculus suggest general functions of some sort rather than control of local reflexes. Another series of experiments by Bailey and Davis ('42) shows that this fasciculus has some activating function which is essential for maintenance of normal motor behavior. They succeeded


INTERPEDUNCULAR NUCLEUS >0J)

in destroying by electrocautery the periaqueductal gray of cats without injury to surrounding parts. This lesion must have involved the dorsal longitudinal fasciculus and the dorsal tegmental nucleus. If the injury is slight, the cats on awakening from the anesthetic are very wild and active but after a few days recover normal behavior. This seems to be an irritative excitation. If the lesion is extensive, the cats "lie inert, silent and flaccid as a wet rag." They never again show any spontaneous activity, though after a few days placing reactions are elicited normally and, if stimulated, they may walk slowly.


If now it is assumed that the dorsal longitudinal fasciculus contains fibers which normally transmit some sort of continuous nonspecific activating or facilitating influence upon the entire motor field of the lower brain stem — an influence which is essential for effective co-ordination and integration of these local systems of synergic muscular activity — then the destruction of this pathway would leave the animal helpless. This is not a flaccid paralysis, because the animal ultimately regains control of righting and placing reactions and even co-ordinated locomotor movements. All initiative and spontaneous activity are permanently lost because the animal is deprived of some essential component of the integrating apparatus. The "obstinate progression" that follows destruction of the interpeduncular nucleus may result from loss of an essential inhibitory influence normally acting upon the dorsal tegmental nucleus and the f. longitudinalis dorsalis. Loss of this influence allows the remaining apparatus of facilitation in the f. longitudinalis dorsalis to "race" like a steam engine deprived of its governor, and in this particular setup of conditions the result is uncontrolled progression.


In Rabat's ('36) study of alterations in respiration resulting from electrical stimulation of cats, it was reported that stimulation of the periaqueductal gray and of diencephalic regions known to contribute fibers to the dorsal longitudinal fasciculus was followed by increase in rate and amplitude of respiration. Shallower and unusually slower breathing followed stimulation of the habenula and habenulointerpeduncular tract.


Detwiler has shown (p. 6'2) that in Amblystoma a nonspecific mesencephalic influence is essential for the maintenance of motor efficiency. This is in addition to the action of the apparatus which activates patterned behavior, the latter maturing earlier and quite independently of cerebral influence. This mesencephalic influence in the amphibian brain seems to be comparable, in a general way, with the activating function of the dorsal longitudinal fasciculus of mammals.


These fragmentary observations cannot justify any final conclusions, but they support the working hypothesis that the habenulointerpeduncular complex is balanced against the strio-pedunculobulbar and f. longitudinalis dorsalis systems, the latter being the activating members and the former the inhibiting member of a dynamic system adapted to insure co-ordinated action of specific synergic systems of muscles.


It is evident that in macrosmatic animals the olfactory system is dominant in the interpeduncular complex. It is the chief functional component of the stria medullaris thalami, habenula, and f. retroflexus and probably also of the f. longitudinalis dorsalis, though all these structures survive its absence in anosmic animals. In Ichthyopsida the olfactory system is dominant in the hypothalamus. Attention has been called (p. 99) to the double role played by the olfactory system in the cortical activities of macrosmatic mammals; that is, in addition to the specific olfactory functions, there is a nonspecific facilitation by activation or inhibition of other cortical activities. This conception of the nonspecific facilitating action of the olfactory cortex may be extended into the subcortical field also, and in the interpeduncular complex the dominant part played by olfaction is probably of this nonspecific nature.


CONCLUSION

The hypothesis here tentatively suggested divides the adjusting apparatus of the brain into two reciprocally interacting components, the one a system of activators, the other a system of inhibitors. In the normal patterning of behavior these systems are balanced one against the other. In the brains of primitive vertebrates the dominant highest center of correlation of the exteroceptive sensory systems lies in the tecto-thalamic sector and of the interoceptive systems in the hypothalamus. From both these regions the activating fibers converge into the ventral thalamus and motor zone of the jnidbrain and isthmus, where they are joined by olfactory fibers and others from the supra-sensory centers of the hemispheres. Above and below this central core of activating fibers are two systems of inhib


INTERPEDUNCULAR NUCLEUS 211

itors — the olfacto-somatic system in the habenula and the olfactovisceral system in the hypothahimus — both of which discharge into the interpeduncular nucleus. The efferents from this nucleus reach the same lower motor centers as do the efferents of the activating systems.


In all macrosmatic species, and especially in those of more primitive groups, the olfactory system is dominant in the forebrain, and in all groups it plays an important part in nonspecific activation and inhibition. In lower groups the deep neuropil of the gray substance is a diffuse activator, and this sytem survives in mammals as the dispersed periventricular fibers which converge to descend in the f. longitudinalis dorsalis of Schiitz. Associated with this fasciculus is the dorsal tegmental nucleus, which receives inhibitory fibers from the interpeduncular nucleus. This complex discharges nonspecific activating and inhibiting impulses into the bulbar tegmentum, and these play an essential part in the maintenance of the appropriate balance between activation and inhibition in all muscular activity. The destruction of any major component of this equilibrated dynamic system may result in pathological behavior which has no counterpart in the normal animal but which may furnish clues pointing the way toward successful analysis.


Though, as suggested above, this schematic outline is doubtless oversimplified, the hypothesis or some variant of it may suggest profitable lines of experiment. For this purpose the urodeles, with great enlargement and elongation of the interpeduncular nucleus, are favorable subjects. The larger species, like Necturus and Cryptobranchus, can be operated upon more conveniently than can the salamanders and frogs, though their more sluggish behavior may make interpretation more difficult.



CHAPTER XV MIDBRAIN

IN THE preceding chapters the structure of the rhombencephalon and the functions intrinsic to it were summarized. We now pass to the cerebrum, with a radically different type of organization, dominated by the optic and olfactory systems. In lower vertebrates the major optic terminals in the midbrain evidently have determined the course of its differentiation ; but, since optic fibers terminate also in the diencephalon, a separate chapter is devoted to the visual system as a whole.


DEVELOPMENT

Because the mesencephalon is the site of a remarkable series of changes in the shape of the brain in early developmental stages, it is appropriate to review here some features of this growth. Attention has been called (p. 117) to the significance of von Kupffer's anterior and posterior intraencephalic sulci as stable landmarks in the development of all vertebrate brains. These mark the loci of two strong cerebral flexures, which are extremely variable in different species and among individuals of the same species. These flexures so modify the topographic arrangement of parts as to make morphological analysis difficult. The posterior sulcus of von Kupffer (the development of which is described in chap, xiii) marks the posterior limit of the great mesencephalic flexure, which bends the neural tube ventralward and backward so that, at the close of the early swimming stage of Amblystoma tigrinum (Harrison's stages 36-38), the hypothalamus is closely appressed against the ventral surface of the peduncle and isthmus ('38, figs. 1-4, 15-18). The tuberculum posterius marks the ventral fulcrum around which this bend is made.


Beginning somewhat later than this ventral bending, there is a flexure in the reverse direction more anteriorly at the site of von Kupffer's anterior sulcus, which bends the cerebral hemispheres forward and upward. In the early feeding stage (Harrison's stages 44, 45) this movement is far advanced ('386, figs. 1, 2); and from this stage onward to the adult stage these flexures appear to be somewhat straightened and masked by intussusception of newer tissues.


212


MIDBRAIN >i;}

Baker and Graves (S*^) described the external form of the brain in six stages of A. jeffersonianum (3-17 'mm.). Their figures 1 and 2 show that in their 3-mm. embryo, which is 7 days and 16 hours old and in a premotile stage, the mesencephalic flexure has begun. In Harrison's stage 30 of A. punctatum, about 24 hours younger than Coghill's nonmotile stage, this flexure is far advanced ('38, fig. 1), and the reverse flexure leading toward the evagination of the cerebral hemisphere is incipient.


Correlated with these changes in external form, cells are proliferating internally, and the fibrous connections are established. CoghilFs charts show that differentiation is precocious in the peduncle, the chiasma ridge, and the olfactory area of the hemisphere. Other local areas of differentiation appear successively, and, before the early swimming stage is reached, confluence of these areas has begun, and fibers are descending from the peduncle and ventral thalamus in the primary motor tract, which is the precursor of the fasciculus longitudinalis medialis. This development is well advanced before connection is made between the sensory field of the tectum and the motor field of the peduncle (compare '37, figs. 1 and 2; '38, figs. 5 and 20). Fibers from the retina enter the brain in early swimming stages, but they do not reach the tectum until the close of this period. These optic fibers enter the tectum at its anterior end in the eminence of the posterior commissure, and the first connection between optic tectum and peduncle is by way of this commissure, in which fibers first appear in the coil stage (Harrison's stage 35), long before optic fibers reach the tectum.


The commissura tecti in the remainder of the midbrain roof appears much later, beginning as an extension spinalward from the posterior commissure. As late as the midlarval period (A. tigrinum of 38 mm.) this commissure has matured only in the anterior half of the tectum ('39a, figs. 1, 11, 12; '39/;, fig. 1). Full details of the development of the optic nerve and its connections have been published ('41), as well as some features of the development of the tectum ('42, p. 243).


In development the nucleus posterior tecti is retarded as compared with the tectum opticum, more so in A. tigrinum than in A. punctatum ('386, p. 413; '39a, p. 265). The ventricle is here dilated as the recessus posterior mesencephali, and in larval stages the deep sulcus isthmi ends dorsally in this recess ('38, fig. 18; '14a, fig. 2). The roof and side walls of the posterior recess remain relatively thin until late larval stages, and they are expanded so as to form a noticeable external eminence, which disappears in the adult except in the midplane.


SENSORY ZONE TECTUM OPTICUM: SUPERIOR COLLICULUS

Nothing need be added here to the statements in chapters iv and xvi and the description of the mesencephalic V nucleus in chapter x. For details my paper of 1942 may be consulted.


NUCLEUS POSTERIOR TECTi: INFERIOR COLLICULUS

The small nucleus posterior tecti is an undifferentiated primordium of the inferior colliculus, with perhaps some additions in higher animals from the posteroventral part of the optic tectum.


The most characteristic cells of the posterior nucleus are small elements with thick bushy dendrites directed forward into the tectum opticum. Slender axons arise from these dendrites and are directed forward to the thalamus as an important component of the brachium of the inferior colliculus ('42, p. 265 and figs. 51 and 79). Other efferent fibers go out in large numbers in tractus tecto-thalamicus et hypothalamicus cruciatus posterior (p. 297; '42, p. 221), of which the strongest component decussates in the postoptic commissure and distributes to the tegmentum in tegmental fascicles (8) and (6). Other shorter efferents reach all surrounding parts. The chief connections of this region are shown diagrammatically in figures 10-18 and 21.


In Amblystoma there is no sharp boundary between the optic tectum and the nucleus posterior. The distinguishing features of the latter are the absence of optic terminals and the presence of two special tracts — the primordial lateral lemniscus {tr.b.t.l.) and the efferent tr. tecto-bulbaris posterior (tr.t.b.p.). The spinal lemniscus and the general bulbar lemniscus terminate chiefly in the optic tectum, though the spinal lemniscus has a larger proportion of terminals in the nucleus posterior. The tecto-cerebellar tract arises chiefly in the nucleus posterior. These connections suggest proprioceptive functions (chap. xii). This suggestion is strengthened by the presence of many cells of the mesencephalic V nucleus here and by the undeveloped condition of the auditory apparatus. The dominance of the cochlear connection is a later acquisition in land animals, with resulting transformation of the nucleus posterior into the colliculus inferior and differentiation of the medial geniculate body. This trans


MIDBRAIN 215

formation is far advanced in the frog, for Aronson and Noble ('45) report that the warning croak uttered during mating is abohshed by extensive injury of the inferior colHcuH, though it is not lost after complete ablation of the hemispheres, diencephalon, superior colliculus, cerebellum, and anterior parts of the tegmentum.


In Anura the optic tectum is greatly enlarged so as to cover the dorsal convexity of the midbrain. The nucleus posterior is also enlarged under the influence of the well-developed lateral lemniscus, and it is folded into the aqueduct as the so-called "corpus posticum" or "torus semicircularis." In some reptiles and in all mammals the definitive inferior colliculus reappears on the dorsal surface.


INTERMEDIATE ZONE

The distinctive characteristics of the intermediate zone of Amblystoma are more clearly shown in the midbrain than elsewhere. This band of subtectal tissue is well differentiated from the overlying tectum and the underlying peduncle and isthmic tegmentum and is given a distinctive name.


TEGMENTUM DORSALE

The dorsal tegmentum is separated from the isthmus by the sulcus isthmi (s.is.) and from the peduncle by the limiting sulcus (s.) of the nucleus of the tuberculum posterius (fig. 2). On the external and ventricular surfaces there is no visible boundary between this zone and the overlying tectum, but the internal structure is sharply contrasted. The gray of the tectum is irregularly laminated, with plaques of cells separated by thin sheets of neuropil; that of the subtectal area is more homogeneous (figs. 34, 36, 93, 94). The texture of the white substance is even more distinctive.


The neurons are rather small and are imbedded in dense neuropil, which is continuous with that of adjoining areas. Typical illustrations are shown in figures 24, 48, and 49 (see also '42, figs. 29, 30, 44, 45, 48). Their dendrites ramify among dorsal and ventral tegmental fascicles and into the tectum, peduncle, and isthmic tegmentum. The axons arise from the dendritic tree, and most of them are short, arborizing locally and in adjoining areas. Many divide with ascending and descending branches. Axons of the larger elements (fig. 49) descend to the isthmic tegmentum in the dorsal tegmental fascicles (fig. 21, tr.tegds.), these fibers being the largest component of fascicles of group (7). All these neurons seem to be concerned with local correlation.


This is typical correlational tissue interpolated between the sensory zone above and the motor zone below, with both of which it is connected by numberless short fibers, by long dendrites, and by the intrinsic neuropil (fig. '24). In addition to these local connections, it receives longer fibers from the cerebral hemispheres, dorsal and ventral thalamus, hypothalanms, and tectum, many of these fibers decussating in the anterior and postoptic commissures. There are long efferent fibers which descend in the dorsal tegmental fascicles to the isthmic and trigeminal tegmentum.


This field evidently contains important components of the apparatus of sensori-motor adjustment, but how this mechanism operates is unknown. This relatively homogeneous area probably loses its identity in higher animals, in which its parts are specialized and dispersed in the brain stem or perhaps supplanted by the more efficient adjustors of the thalamus and hemispheres.


MOTOR ZONE


PEDUNCLE


At the sharp ventral cerebral flexure of the brain stem, a ventricular eminence marks the position of the nucleus of the tuberculum posterius, which is called the "peduncle" in a restricted sense (p. 21). This peduncular area, which is boundetl by a shallow and variable sulcus (fig. 2, s.), is an arbitrarily defined field at the ventral surface of the midbrain, bounded dorsally by the dorsal tegmentum and posteriorly by the isthmic tegmentum. It arises from the anterior end of the embryonic basal plate of the neural tube, and nerve fibers appear within it very early in embryogenesis. In the coil stage, fibers are added to this tract from the ventral thalanuis and dorsal tegmentum ('37, fig. 1), and, before the early swimming stage is reached, there are additions also from the tectum and other neighboring parts ('37, fig. 2). This basal sector of the midbrain is the nucleus of origin of the first efferent fibers of this region for lower motor centers, and, as early as the S-reaction stage, fibers go out from it to the periphery in the oculomotor nerve. This is before any afferent fibers are connected with this region. The motor mechanism develops autonomously.


In Necturus the f . longitudinalis medialis is definitely organized in the peduncle and is a well-defined anatomical structure from here caudad ('36, p. 349); but in Amblystoma the fibers which compose this bundle are mingled with others in tegmental fascicles numbered (1), (4), (5), and (6) in the peduncle and isthmus, and below the isthmus a residue of fibers from all these fascicles is assembled to continue spinalward in the f. longitudinalis medialis ('36, figs. 9-16).



MIDBRAIN 217

The gray layer is relatively thin, and in the mid-plane most of its cells are displaced by the massive decussations of the commissure of the tuberculum posterius (figs. 2C, 29, 30, 31, 94). More laterally, large and small cells are mingled in the gray, with little evidence of local segregation except for the nucleus of the oculomotor nerve at the posteroventral border.


The largest cells include dorsally the primordium of the nucleus of Darkschewitsch, which is related primarily with the posterior commissure (figs. 6, 18, '2'2, 24; shown but not named in fig. 32), more ventrally Cajal's interstitial nucleus of the f. longitudinalis medialis (figs. 18, 22, 104), and posteriorly the III nucleus (figs. 18, 22, 24, 93, 104). Larval forms of these cells have been illustrated ('396, figs. 42, 53, 72, 73). They resemble those of Necturus ('17, figs. 23, 24, 30, 33).


These cells evidently are collectors of a wide variety of nervous impulses brought into the peduncle by the numberless fibers which converge to end here and by fibers of passage with collaterals in this area. The axons of these cells descend in the ventral tegmental fascicles and seem to be primarily concerned with activation of mass movements, particularly of locomotion. Most of these thick fibers descend in ventral tegmental fascicles of group (5) and are uncrossed, but some of them decussate in the ventral commissure, mingled with the tecto-bulbar fibers of the ventral median fascicles of group (1) ('396, figs. 23, 24).


Small cells in the anteroventral part of the peduncle are intimately connected with the underlying hypothalamus and are doubtless concerned primarily with olfacto-visceral adjustments (figs. 18, 21, 53). There are numberless terminals here of the nervus terminalis, tr. hypothalamo-peduncularis from the ventral hypothalamus ('42, p. 226 and figs. 3, 22, 23, 68, 69), tr. mamillo-peduncularis from the dorsal hypothalamus ('396, fig. 22; '42, fig. 39), and tr. olfactopeduncularis from the anterior olfactory nucleus and primordial head of the caudate nucleus ('396, fig. 1). This area also receives terminals of the secondary and tertiary visceral-gustatory tracts, and there is a strong pedunculo-hypothalamic connection (fig. 8; '42, p. 227). Ventral tegmental fascicles (3) and (4) from the hypothalamus ('42, fig. 3) are related with this area ('42, figs. 39, 40, 44), and these fascicles, together with the related small cells, probably contain primordia of those components of the mammalian dorsal longitudinal fasciculus of Schutz which are related with the peduncle and hypothalamus (Thompson, '42, p. 249).


Cells of medium size at the posterior end of the peduncular gray in the vicinity of the nucleus of the III nerve (fig. 59) have the chief dendrites directed rostrad, some of them longer and more sharply bent forward than those shown in figure 59. These dendrites engage fibers from the basal optic tract, visceral-gustatory system, hypothalamus, tr. olfacto-peduncularis, and mesencephalic terminals of the f. retroflexus, among other systems. Their axons also are directed forward; how far they extend within the basal forebrain bundles is not evident, probably for a long distance because they are thick, smooth fibers similar to those of the long descending and ascending tracts. It is not improbable that they reach the primordial corpus striatum, where similar fibers end in wide terminal arborizations ('396, figs. 39, 41, 50, 72, 73; '42, fig. 69). These fibers may be comparable with those described in mammals as ascending from the entopeduncular nucleus and substantia nigra to the corpus striatum.


The neuropil of the alba resembles that of the ventral thalamus, but that of the gray substance is more dense and of different composition. In the white substance there are mingled terminals of axons derived from a great variety of sources — tectum, basal optic tract, posterior commissure and its nucleus, dorsal and isthmic tegmentum, the entire diencephalon, and the basal forebrain bundles (figs. 12, 14, 15, 17, 18, 20, 21, 22). There is little evidence of regional localization of these terminals except in two places.


Posteriorly, in the vicinity of the nucleus of the III nerve, the deep neuropil of the gray is very dense, with strands of fibers descending from the tectum and the dorsal and isthmic tegmentum (fig. 24). These clearly envelop the cell bodies of the III nucleus, and many of them decussate here in the ventral commissure. A second and more sharply localized area of neuropil is the area ventrolateralis pedunculi, extending superficially from the III root forward (fig. 23), as described in chapter iii.


The commissure of the tuberculum posterius includes all fibers of the ventral series of commissures which cross in the midbrain; for details of its connections see chapter xxi.


Summary. — The peduncle of Amblystoma as here defined receives fibers from practically all parts of the brain above the isthmus, and its primary function is control of mass movements of the trunk and limbs and conjugate movements of the eyes. Control of local reflexes is effected elsewhere. It is intimately connected with the hypothalamus, and various visceral motor adjustments are made here, though of these little is known.



CHAPTER XVI OPTIC AND VISUAL-MOTOR SYSTEMS

THE eyes in Amblystoma are much better developed than they are in Necturus, where they are degenerate, though functional. The retina of Triturus is more highly differentiated, and vision is evidently more efficient. In frogs this advance is carried much farther, with corresponding structural elaboration of the visual centers of the tectum opticum and thalamus.


OPTIC NERVE AND TRACTS

The optic tracts and their connections of Necturus have been described ('41a). An account of the development of the optic nerves and tracts of Amblystoma was published in the same year ('41), and subsequently ('42) a detailed analysis of the visual centers and their connections, together with commentary on related physiological problems. Some further theoretical considerations were published later ('44a). These details need not be repeated. The salient features were outlined in chapter iv, and some topics are amplified here.


In the coil stage (Harrison's stage 34), optic nerve fibers begin to grow from the ganglion cells of the retina through the optic stalk before the other retinal layers are histologically differentiated ('41, p. 477; Coghill, '16, Paper II, p. 274). In the S-reaction stage (Harrison's stages 35, 36) these fibers reach the brain; in early swimming stages (Harrison's stages 36, 37) they decussate in the chiasma ('41, p. 480); and shortly thereafter they reach the tectum (Harrison's stage 38) and the primordial basal nucleus in the peduncle. Up to this stage, rods and cones and other retinal layers are incompletely differentiated. In stages immediately preceding effective feeding reactions (about stage 44) the retina is functional, and the larva can orient itself with reference to an object moving in the field of view; and shortly thereafter well co-ordinated lurching and snapping movements efficiently capture moving prey. The first visual responses are probably activated by fibers from the tectum through the posterior commissure to the nucleus of Darkschewitsch and the fasciculus longitudinalis medialis and perhaps also by the basal optic tract, though the time of the appearance of these fibers has not been recorded. Tecto-bulbar and spinal fibers mature very soon after those of the posterior commissure, the uncrossed fibers preceding the crossed fibers. These and the tecto-peduncular fibers described below are apparently the pathways employed in the activation of the skeletal musculature of the trunk and head when they are employed in the capture of food. Early feeders (stage 46) have attained full larval status, and in subsequent stages the visual system is not radically changed except for an increase in the number of thin retinal fibers and an enormous increment of these at metamorphosis, the total number of fibers in adult Amblystoma tigrinum probably reaching about 8,000 ('41, p. 506).


Correlation of this developmental history with experimental work by others has led me to the hypothesis that in Amblystoma the thick retinal fibers, which appear first in ontogeny, activate generalized total behavior patterns and that the thin fibers are concerned with more refined analysis of visual experience. (It does not follow that this is true in mammals, in which the visual responses are under thalamo-cortical control.) What kind of mechanism is employed in this analysis is still obscure. My opinion ('41, p. 528) that endings of thin and thick fibers have separate locahzation in the tectum proved on further study to be unfounded ('42, pp. 284, 293), and there is little anatomical evidence of any other kind of visual localization here, though experimental evidence cited below indicates that retinal loci are projected locally upon the tectum.


The more important known connections of the visual system are shown in the simplified diagrams (figs. 6, 11-16, 18, 20-24, 93, 101; '36, figs. 2, 6, 7; '42, figs. 4, 5, 79). Fibers of the optic nerve are distributed to five quite separate fields in the brain stem: (1) tectum, (2) pretectal nucleus, (3) cerebral peduncle, (4) thalamus, and (5) hypothalamus, as shown schematically in figure 14. This wide distribution is in marked contrast with the tendency of the other functional systems of peripheral nerves to converge into a single primary central field. This dispersal appears to have been determined by the types of motor response to be evoked rather than by the sensory qualities of the visual excitations. These five receptive fields will now be examined.


1. The tectum opticum evidently is the most extensive of these fields and physiologically the dominant member (see below).


The pretectal nucleus may participate in the regulation of the intrinsic musculature of the eyeball and have other functions still unknown.


3. The basal optic tract is present in all classes of vertebrates from cyclostomes to man. It is evidently of major importance, though little attention has been given to it in current studies of the physiology of vision. Two possible functions have been suggested for it in Amblystoma. The more primitive of these is to act as a general activator of the ocular and somatic muscles in response to visual stimuli, manifested in such behavior as the "regarding reaction" described by Coghill (pp. 38, 78). In the second place, the structure and connections of the area ventrolateralis pedunculi, within which the basal optic fibers terminate, suggest that this neuropil may be part of the apparatus of conditioning of visual reflexes (p. 38).


4. Optic terminals are widely spread in the thalamus and are especially concentrated in its posterior part. This ill-defined area of neuropil receives the brachia of the superior and inferior colliculi and is regarded as the primordium of the lateral and medial geniculate bodies (p. 239; '42, p. 280). Large numbers of efferent fibers go out from this field to the peduncle and the dorsal and isthmic tegmentum. The fibers to the tegmentum take four courses. Two are superficial, the uncrossed tractus thalamo-tegmentalis rectus and a crossed path ('42, p. 224, tr.thieg.d.c.A.) which decussates in the postoptic commissure. Two other tracts take deep courses close to the gray, one of which decussates ('42, p. 224, tr.th.teg.d.c.B.) and the other descends uncrossed in dorsal tegmental fascicles of group (7). For a detailed description see the analysis of the postoptic commissures in chapter xxi. These four tracts end chiefly in the dorsal and isthmic tegmentum, two of them in the superficial neuropil and two at deeper levels. They provide pathways for visual control of local reflexes, particularly those employed in capturing food. There is no separate optic projection tract from the thalamus to the cerebral hemisphere, though visual influence may reach the hemisphere through the thalamo-frontal tract, which is a common pathway of sensory projection from the dorsal thalamus (pp. 95, 238, and fig. 19, tr.th.f.).


5. There is a small number of optic terminals and collaterals in the hypothalamus near the chiasma ('42, pp. 219, 233, 237) . They connect with the large cells of the preoptic nucleus, axons of which form the hypophysial tract (p. 245). This is doubtless the pathway for visual influence upon hypophysial endocrine activity. This connection is probably smaller in Ambly stoma than in some other vertebrates, and different opinions have been expressed about it. Eugen Frey's description ('38) of a large hypothalamic optic root in some amphibians was evidently based on inadequate material and faulty observation, as I have pointed out ('396, pp. 558-76; '41a, p. 498; '42, pp. 218, 233). The same is probably true of Geiringer's ('38) description of a passage of fibers from the optic nerve directly into the hypophysial tract. I have preparations that give this appearance, but more critical examination disproves it. There is some evidence that in fishes and amphibians efferent fibers go out from the preoptic nucleus through the optic nerve to the retina ('336, p. 254); but I have not been able to confirm this or the claim of some authors that efferent fibers go out to the retina from the tectum. Both these connections may exist. In the chiasma the coarser optic fibers are segregated from the finer, and most of them take deeper courses in the optic tracts — the axial bundles. As the optic tracts traverse the thalamus, they separate into medial (dorsal) and lateral (ventral) tracts, each of which contains both thick and thin fibers. In the tectum most of these fibers end by wide arborizations in a common pool of intermediate neuropil (fig. 93, layer 2).


TECTUM OPTICUM

In embryological development the first afferent fibers to reach the tectum are those of the optic tract. These terminate in its anterior part near the posterior commissure, and tectal structure matures in subsequent stages from this region posteriorly. In the adult the common pool of neuropil, to which reference was made above, also receives similar terminals of the spinal and bulbar lemnisci, the brachium of the superior colliculus from the thalamus, the strio-tectal and habenulo-tectal tracts, and some other fibers. This pool of neuropil is spread throughout the entire tectum, with little evidence of localization of function within it.


The optic tectum is structurally nearly homogeneous, with two exceptions: (1) a dorsal thickening on each side, which is related with the dorsal tectal commissure and probably corresponds (in some respects only) with the torus longitudinalis of fishes ('42, pp. 250, 287), and (2) the eminence of the posterior commissure at the anterior border of the tectum, likewise chiefly commissural in function. The stratification of the tectum, which is so conspicuous in the frog and many other vertebrates, is here notably absent. The only layers


OPTIC AND VISUAL-MOTOR SYSTEMS 223

that are clearly evident are the deep gray and the superficial white, though there are obscure indications of the incipience of further stratification, more evident in the alba than in the grisea. For descriptive purposes I have divided the optic tectum, somewhat arbitrarily, into eight concentric layers, as shown in figures 36 and 93; for their characteristics see page 244 of 1942.


Two features of tectal structure merit special emphasis. The first is that the convergence of most of the afferent fibers into a common pool of neuropil of almost homogeneous structure suggests a totalizing or integrative function for the tectum as a whole. The second point of major interest is that the efferent fibers from the tectum are extremely diversified in structure and distribution. There are all gradations from diffusely spread fibers in the deep neuropil to welldefined long myelinated tracts like the tecto-bulbar and tecto-spinal systems. This suggests that such localization of function as exists in this tectum is determined more by what is going on in the efferent side of the arc than in the afferent side, a point which has been emphasized by Crosby and Woodburne ('38).


Five strong systems of myelinated fibers leave the tectum. These take widely divergent courses. They are as follows:

1. Commissura posterior. — These are the first to appear in the embryo. The crossed fibers are joined by uncrossed fibers from the eminence underlying the commissure, and they spread widely in the peduncle and adjoining parts of the thalamus and dorsal tegmentum. The primary connection is with the nucleus of Darkschewitsch (p. 217 and figs. 6, 18, 22) and thence to the f . longitudinalis medialis, providing innervation of the trunk musculature in response to visual stimulation.


2. Tractus tecto-peduncularis rectus et cruciatus (p. 303 and figs. 18, 22, tr.t.p.c; '42, p. 267). — These thick fibers arise from the anterior part of the tectum, descend parallel with fibers from the posterior commissure, decussate (in part) in the commissure of the tuberculum posterius, and arborize in the alba of the peduncle of both sides more ventrally than those of the posterior commissure. Their course as seen in horizontal sections is shown in figures 30-36, tr.t.p.c.l. (see also '42, p. 267, and the figures there cited). They activate the interstitial nucleus of the f. longitudinalis medialis and probably also the oculomotor nucleus and appear to be adapted to regulate mass movements of trunk, limbs, and eyes in the orientation and movements of the body with reference to objects in the visual field. These fibers are joined by similar thick fibers from the pretectal nucleus and dorsal thalamus (tr.fh.p.c. of the figures).


Superficially of the connection just described, there is a large collection of uncrossed fibers passing between the pretectal nucleus (pars intercalaris diencephali) and the peduncle. These were described in Necturus ('17, p. 264; '336, p. 110 and fig. 57) under the name "tr. thalamo-peduncularis dorsalis superficialis." These fibers are spread close to the pial surface in the wide di-mesencephalic fissure and end in large numbei-s in the neuropil of the area ventrolateralis pedunculi. Some of them also reach the hypothalamus. The condition in Amblystoma is similar, thus bringing the pretectal nucleus into intimate relation with terminals of the basal optic tract and the oculomotor nucleus.


3. Tr actus tecto-thalamicus et hypothalamicus cruciatus (fig. 12, tr.t.th.h.c.a. and p.\ see also the section on com. postoptica in chapter xxi). — The anterior and posterior divisions of this complex were described in 1942 (p. 221). The anterior tract passes from the dorsal part of the superior colliculus to the hypothalamus, decussating in the postoptic commissure. It is here termed "tr. tecto-hypothalamicus anterior" (figs. 25-36, tr.t.hy.a.; see further on p. 296). It has connections with the thalanms and hypothalamus of both sides, and some of its crossed fibers may reach the peduncle and tegmentum.


The posterior division arises chiefly from the inferior colliculus, with accessions from the ventral border of the superior colliculus. It lies parallel with the anterior division and, more posteriorly, partially decussates in the postoptic commissure and has terminals in the thalamus and hypothalamus of both sides (figs. 25-35, tr.t.th.h.c.p.). Its thicker and more myelinated fibers continue as tr. tecto-tegmentalis cruciatus into tegmental fascicles of groups (6) and (8), to end in the peduncle and tegmentum as far back as the VII nerve roots. The distribution of these fibers suggests that they are primarily concerned with mass movements and local reflexes of the musculature of the head and probably also with conditioning of these movements. What seems to be the equivalent of this tract is described as connected with the nucleus isthmi in all animals in which that nucleus is well developed; but in urodeles this tract is very large, though the nucleus isthmi is vestigial.


4. Tecto-hulbar and tectospinal tracts. — These tracts have been described and illustrated in the papers of 1936 (p. 340) and 1942 (p. 268). They are the largest of the efferent tracts, and most of their fibers are well myelinated. There are two crossed tracts (figs. 12, 27-36, tr.t.b.c.l. and ir.i.b.c.2.) and an uncrossed tract (tr.t.h.c.r.) from the tectum opticum. The posterior crossed tract itr.t.h.c.2.) decussates transversely in the vicinity of the nucleus of the IV nerve. The anterior crossed tract {tr.t.b.c.l.) takes a peculiar course, its fibers entering the ventral medial tegmental fascicles, f.v.t.{l), within which they decussate obliquely. There are about a dozen of these anastomosing fascicles extending from the commissure of the tuberculum posterius backward to the level of the IV nucleus, where these crossed fibers join those of tr.t.b.c.2. and descend to the medulla oblongata and spinal cord (figs. 6, 12).


The uncrossed tecto-bulbar fibers — tr. tecto-bulbaris rectus — were seen in my earlier studies to arise only from the posterior part of the tectum, as drawn here in figure 12 (tr.t.b.r.); but later ('42, p. 269) other fibers of this system were found to arise more anteriorly and to pass backward by several pathways to join the posterior group. The most anterior of these fibers enter dorsal tegmental fascicles of group (7) , from which they separate to join the posterior tract. Others pass spinal ward in layers 3 and 4 of the tectum, at the posterior end of which they join the other fascicles. Here they send collateral branches into the isthmic visceral-gustatory nucleus, as shown in figure 23. Below the junction of these three divisions, the tr. tecto-bulbaris rectus descends superficially in the isthmus, then turns spinalward, under the lateral recess of the ventricle, passing through the ventral border of the auricle into the medulla oblongata, where it joins the crossed tecto-bulbar tract.


5. Brachia of superior and inferior colliculi. — These are large collections of fibers passing in both directions between the tectum and the thalamus. They are widely dispersed, and most of them discharge from tectum to dorsal thalamus. Those from the superior colliculus are spread throughout the alba and form a massive tract (figs. 34, 35, 36, tr.t.th.r.). These are joined by some fibers from the inferior colliculus, and from the latter area other strands of unmyelinated fibers take separate courses, some passing forward in the dorsal part of the tectum and some at its ventral margin; most of these fibers pass from inferior to superior colliculus, but some of them continue into the thalamus, particularly ventrally, to enter the geniculate neuropil (figs. 14, 15, 16, br.col.; '42, p. 264).


CONCLUSION

The internal structure of the optic tectum of Amblystoma and the wide spread of its fibrous connections, both afferent and efferent, suggest that in this animal the primary function of the tectal system is visual control over movements of the body as a whole and, in particular, the orientation of the body and conjugate movements of the eyeballs with reference to objects in the visual field. Such other local visual reflexes as the animal possesses are probably organized elsewhere.


TECTO-OCULOMOTOR CONNECTIONS

The visual location of food and enemies involves the innervation of the muscles of the eyeballs, and our knowledge of the apparatus employed in the conjugate movements of the eyes is scanty. The anterior and posterior divisions of the crossed tecto-bulbar tract decussate, respectively, in the vicinity of the nuclei of the III and IV nerves (fig. 12). The anterior division {tr.t.h.c.l.) crosses in the ventral medial tegmental fascicles which span the rather long distance between the III and the IV nuclei. When these ventral fascicles — group (1) of my analysis in 1936 — were first identified as tecto-bulbar fibers, it was supposed that this anomalous arrangement was a provision for activation of the eye-muscle nuclei by collaterals from these fascicles. No evidence of this has been found, but the myelinated fibers of both tr. tecto-bulbaris cruciatus 1 and 2 are accompanied by very many shorter, unmyelinated fibers, which pass from the tectum to the III and IV nuclei of both sides. This tr. tecto-peduncularis is shown diagrammatically in figures 18, 22, and 24 (and in more detail in '42, figs. 14, tr.t.p.2., and 45, tr.t.p.c.2.). In addition to the tectopeduncular fibers in the alba, there is a deep series (tr. tecto-peduncularis profundus, '42, p. 267 and fig. 14, ir.t.p.3.), which takes tortuous courses in the neuropil of the grisea. From the anterior part of the tectum and pretectal nucleus the well-defined tr. tecto-peduncularis cruciatus and the uncrossed tr. thalamo-peduncularis dorsalis superficialis may participate in the regulation of conjugate movements of the eyes.


If pupillary constriction is controlled from the pretectal nucleus, as in mammals, there are two available pathways to the III nucleus, first, by way of the pretectal components of the two tracts last mentioned and, second, by tr. pretecto-hypothalamicus (p. 296 and fig. 15, tr.pt.hy.) and thence to the III nucleus by tr. hypothalamopeduncularis (figs. 18, 23, tr.hy.ped.).



OPTIC AND VISUAL-MOTOR SYSTEMS 227

VISUAL FUNCTIONS

Most of our experimental knowledge of visual functions of amphibians has been derived from Amblystoma, Triturus, and frogs. In these animals there is no visual cortex, and in urodeles there is no specific optic projection tract to the cerebral hemisphere. Here the intrinsic features of the stem portion of the visual apparatus can be studied, simplified by absence of the cortical connections; and yet the complications of internal structure as just summarized present baffling physiological problems. Triturus has more efficient eyes than Amblystoma, and in frogs this differentiation is much further advanced, so that the Anura offer the most promising approach for experimental analysis, and for this we need more information about the histological structure of the anuran brain than is now available.


The electrical activity of the optic tectum of the frog has been investigated by Beritoff ('43, pp. 215-40) in co-operation with Tzkipuridze, who present oscillographic records of its spontaneous activity and of the effects of various kinds of stimuli. A distinction is drawn between the quick oscillations due to excitation of the ganglion cells of the retina and slow oscillations ascribed to activity of the neuropil. Similar studies with the aid of microelectrodes may profitably be made of the properties of the layers of the tectum and other central stations of visual activity.


Attention has been called ('41, p. 521) to evidence that the retina of Triturus is more highly, differentiated histologically than that of Amblystoma. That it is functionally superior is confirmed by the experiments of Stone and Ellison ('45), who exchanged eyes of adult A. punctatum and T. viridescens. The Triturus grafts degenerated, but the Amblystoma eyes on Triturus hosts followed a course of recovery as they do in homoplastic transplantations, retinal degeneration being followed by complete regeneration. Visual acuity appeared to be higher than in normal Amblystoma but lower than in normal Triturus.


In the feeding reactions of larval Amblystoma, the location and seizure of prey can be successfully done, in the absence of eyes, with the aid of the lateral-line organs (Scharrer, '32; Detwiler, '45). The lemniscus systems may discharge lateral-line impulses into the tectum, in company with those of other sensory systems; the mesencephalic V system may participate in the reaction; and the habenulotectal tract may transmit an olfactory influence. In addition to these, there is a large thalamo-tectal connection and a small strio-tectal tract. But there is no experimental evidence about the specific functions of any of these connections.


The tectum and thalamus are connected by fibers passing in both directions (figs. 11, 12). These are precursors of the mammalian brachia of the superior and inferior colliculus. It is evident that the tectum and the dorsal thalamus may act as an integrated unit in regulating the major activities of the body in adjustment to external situations. With the emergence of the cerebral cortex in higher animals, this pattern is radically changed, and for the understanding of these changes it is essential that the pre-existing organization be adequately known. The history of the transformation of this generalized amphibian structure to those of reptiles and mammals has been sketched by Huber and Crosby ('43).


Evidence has been published that in some vertebrates the several quadrants of the retina have local representation in the optic tectum. In a series of vertebrates, including several species of amphibians, Stroer ('39, '39a, '40) described continuous fascicles of fibers from the retinal quadrants to the tectum, where they end locally in an arrangement which projects the retinal quadrants upon the tectum in reversed orientation. I have made diligent search for evidence of such an arrangement in Necturus and in young and adult Amblystoma without success ('41, '41a, '42). In Amblystoma thick and thin fibers from the different retinal quadrants are inextricably intertwined in the optic nerve. At the chiasma the thick .fibers are segregated from the thin fibers, and in the tectum the terminals of both kinds of fibers mingle and seem to be distributed nearly uniformly to all parts. There is physiological evidence that Amblystoma can visually localize objects in the field of view; and, by analogy with other animals, the presumption is strong that retinal loci are projected upon the tectum in more or less definitely circumscribed areas.


In Triturus and Amblystoma good localization of objects in the visual field is restored after transplantation of the eyes (Stone, '44; Stone and Zaur, '40; Stone and Ellison, '45), and it is preserved after rotation of the eyeball through 180°, though in the latter case motor reactions to objects in the visual field are reversed from the normal (Sperry, '43). Similar physiological results follow operations which eliminate the chiasma and connect the retina with the tectum of the same side, and also after transplanting an eye to the opposite side of the head (Sperry, '45). These observations have been repeated upon seven species of anurans, all of which have much more highly


OPTIC AND VISUAL-MOTOR SYSTEMS 229

differentiated visual apparatus, with similar results. After the cutting of the optic nerve and its regeneration in the adult animal, the same results were obtained. Vision was normal in those animals whose retinas had been kept in normal position, but it was reversed about the optic axis in animals whose retinas had been rotated through 180° prior to nerve section (Sperry, '44, '45). In frogs, destruction of quadrants of the tectum of both normal and operated animals resulted in scotomas of local quadrants of the visual field in a pattern which is in agreement with the anatomical observations of Stroer ('39rt); and, in cases in which the eyeball was rotated, the scotomas were in arrangement reversed from that found when the eyes were in the normal position.


These experiments increase the probability that, in Amblystoma, retinal loci are in some way projected locally upon the tectum; but the mechanism employed in localization of visual functions can be clarified only by further experimentation. Sperry 's experiments demonstrate "the high degree to which the complex and precisely patterned neural mechanisms subserving adaptive visuomotor coordination are dependent upon inherently predetermined rather than upon functionally acquired neural adjustments."


CHAPTER XVII DIENCEPHALON

GENERAT. FEATURES

IN THE f()ll()win,t>- analysis tlie structures at the di-telenceplia ic junction are omitted, since they are fully described in the next chapter. The retina and the optic nerve are integral parts of the diencephalon. The retina itself is more than a simple receptor, for it contains elaborate apparatus for analysis and synthesis of v'sual excitations (Polyak, '41).


The diencephalon combines characteristics of the sensory, intermediate, and motor zones, and the tissues involved are so closely interwoven that toi)ographic subdivisions into zones is arbitrary and misleading. The lai-gest number of afl'erent fibers enter by the optic nerve and are distributed to the thalamus, pretectal nucleus, and hypothalanms, in addition to their mesencephalic connections. The nervus terminalis has widely spread terminals in the preoptic nucleus and the hypothalamus. The parietal nerve contains a few fibers of uncertain connections and functions. The hypo])hysial nerve carries many fibers to the ])ars nervosa of the hypophysis ami a smaller number to the j)ars distalis.


Between the central stations of these peripheral nerves and interpenetrating them is correlating tissue of intermediate-zone type. As repeatedly emphasized in several preceding contexts, the diencephalon as a whole (except for the retina) is a transitional sector of the brain, inter})olated between the olfactory field anteriorly and all other sensori-motor fields posteriorly. This apparatus of correlation is supplementary to the more direct paths of through traffic which organize the basic patterns of the stable action system characteristic of the species. In urodeles the most direct descending olfactory connections are by way of the tractus olfacto-peduncularis for somatic responses and to the hy})othalamus for visceral responses. Additional diencephalic olfactory connections are ancillary to these. The apparatus of other stereotyped patterns of sensori-motor adjustment is similarly organized in the stem below the diencephalic level. These


DIENCEPHALON 231

systems all have a collateral discharge into the thalamus, which is probably concerned with conditioning and other types of "inflection" of the stable components of behavior. In the most primitive vertebrates this primordial thalamic apparatus of higher synthesis and conditioning is small and at a low grade of differentiation. In Amblystoma it is notably larger, but still relatively unspecialized and probably concerned more with regulation of central excitatory state and the general physiological disposition and diathesis than with refined analysis. In mammals this tissue is elaborately specialized, with provision for precise localization of function correlated with the simultaneous differentiation of the neopallial cortex. But even in man the intrinsic functions of the thalamus are not supplanted by cortical development. The thalamo-cortical connections carry two-way traffic and all cortical activity may be profoundly affected by diencephalic influence.


The primary subdivisions of the adult diencephalon are more clearly evident in the gross preparation of the urodele brain than in most other vertebrates, as illustrated in figures IB and 2. I emphasized this in the paper of 1910, and the analysis there proposed has been widely accepted as applicable in all vertebrates. The epithalamus includes the habenular nuclei, the pars intercalaris, the epiphysis, and the membranous dorsal sac and paraphysis. The dorsal thalamus is an undifferentiated nucleus sensitivus, the ventral thalamus a field of motor co-ordination, and the hypothalamujs an olfactovisceral adjuster, including the pars nervosa of the hypophysis. These subdivisions and their more important connections have been described in previous publications (see the summary, '42, p. 204; also, '10, '21a, '27, '35a, '36, '39, '396, '41, '42; Necturus, '17, '336, '346, '34c, '41a; for other urodeles see the works cited on p. 11).


DEVELOPMENT

The larger part of the wall of the diencephalic sector of the early neural tube is evaginated to form the retina and the optic nerve. The details of this development, which begins at the anterior end of the medullary plate, have been recorded by Adelmann ('29, '36). I have described later stages of the development of these structures ('41). During these stages most of the dorsal median wall of this part of the neural tube remains membranous and is elaborately enlarged and folded to form the epiphysial vesicle, paraphysis, and diencephalic chorioid plexuses, the development of which has recently been described in several species of amphibians by Rudebeck ('45). The adult structure of these epithelial organs and the related meninges and blood vessels I described in 1935. The ventral medial wall of the adult (fig. 2) between the anterior commissure ridge and the tuberculum posterius is thin but nervous, except for the dorsal wall of the infundibulum. It has a massive thickening — the chiasma ridge — midway of its length. The arrangement of the chiasma ridge, the anterior commissure ridge, and the adjoining preoptic recess and precommissural recess is peculiar in the amphibian brain (fig. 2 and p. 291).


The massive side wall of the adult diencephalon is composed of four columns of gray substance, which project into the third ventricle as longitudinal ridges separated by deep sulci. As shown here in figures 1 and 2, these are epithalamus, dorsal thalamus, ventral thalamus, and hypothalamus. This adult configuration is achieved by a series of local pulses of proliferation and differentiation of cells, which shift in location and rate of growth from stage to stage, as indicated by changes in the sculpturing of the ventricular surface.


Rudebeck ('45) illustrates embryonic stages of Necturus, Triturus, and the anuran Pelobates which are younger than any of my models of Amblystoma. His pictures show that, in each of these species in stages preceding hatching of the eggs, the first ventricular markings to appear in the rostral part of the neural tube take the form of a series of transverse sulci. The most anterior of these, his "sulcus subpalliahs," separates the subpallial from the pallial parts of the hemisphere. Next follows von Kupffer's sulcus intraencephalicus anterior (p. 117), which separates the hemisphere and preoptic nucleus from the diencephalon. Posteriorly of this sulcus is a high transverse ridge, the di-telencephalic ridge, the adult derivatives of which apparently are the habenula and those structures ventrally of it which are listed in our next chapter as bed-nuclei of the di-telencephalic junctionnucleus of Bellonci, eminentia thalami, nucleus of tr. olfactohabenularis (p. 239), and part of the preoptic nucleus and hypothalamus. This ridge is bounded spinalward by his "sulcus diencephalicus ventralis," the dorsal part of which persists as the sulcus posthabenularis and the ventral part as the sulcus ventralis thalami plus the sulcus hypothalamicus. Posteriorly of the ditelencephalic ridge is an undeveloped area, giving rise to the adult pars intercalaris diencephali and the major parts of the dorsal thalamus, ventral thalamus, and hypothalamus.


After swimming is well established (about Harrison's stage 37),


DIENCEPHALON 233

differentiation is much further advanced, though there is no radical change in the topographic arrangement of the areas involved ('38, figs. 15-18). But, by the time the larvae begin to feed (Harrison's stage 46), the parts have shifted their positions toward the adult condition ('386, figs. 1, 2), which is fully attained in midlarval stages (38 mm, in length) of Amblystoma tigrinum ('39a, fig. 1).


These changes are brought about in two ways. There is, in the first place, a gradual straightening of the two great flexures of the neural tube, with a resulting rearrangement of the several areas in relation to the long axis of the brain and to one another. In the second place and accompanying these changes in the general shape of the forebrain, the regions of most rapid proliferation and difi'erentiation of cells shift their positions. There results a change in the pattern of the ventricular eminences and the intervening sulci. The original transverse ridges and sulci are broken up, and their parts are rearranged in the longitudinal series which we see in the adult brain.


The details of these changes have been described during recent years by a number of workers in Professor Holmgren's Zootomical Institute at Stockholm. Rudebeck ('45), in his admirable paper on the development of the forebrain of lungfishes, shows that the sequence of these changes is essentially the same in Dipnoi, Urodela, and Anura. In the youngest stages, proliferation of cells is most active along the lines of the primitive sulci. These active zones enlarge, and this results in obliteration of the primary sulci or a shift in their positions. The primary grooves, accordingly, do not mark the boundaries of the primordia of the different diencephalic centers, but, on the contrary, they are zones of more active growth from which these centers are developed. In some regions the original proliferation grooves disappear, in others they are gradually displaced and so become limiting sulci at the boundaries of the differentiated areas of the adult.


My interpretation of these changes as outlined above differs somewhat from that of Rudebek as summarized on page 63 of his monograph. Certainly, the details of these developmental processes vary from species to species, and examination of still more species of vertebrates along the lines of the critical studies made in the Stockholm Institute will doubtless clarify the questions which are still in controversy. One is impressed by the close resemblance of this development in lungfishes and amphibians, which suggests that these groups of animals are more intimately related phylogenetically than has hitherto been generally admitted.


EPITIIALAMUS

The epithalamus, like all other parts of the sensory zone, combines reception from the periphery with functions of correlation. It has two radically different parts that are separated by the pineal recess and the overlying pineal vesicle. The habenular nuclei comprise the anterior division. The posterior division is the pars intercalaris diencephali, which is relatively larger in urodeles than in higher animals. Because of this enlargement and the small size of the dorsal thalamus, the habenula lies farther forward than usual. The two habenulae are connected by the habenular or superior commissure and the pars intercalaris by the commissura tecti diencephali.


The pars intercalaris is sharply defined in front, but posteroventrally it merges with the posterior sector of the dorsal thalamus. There is a narrow subependymal layer of small nerve cells imbedded in dense periventricular neuropil, and externally of this lies a lentiform mass of crowded cells, mostly small, with some of larger size, which I term the "pretectal nucleus" (figs. ^B, 14, 15, 35, 36, nuc.pt.) without commitment about its homologies in other animals. It is probable that the pulvinar of mammals has been derived from the undifferentiated gray of the posterior part of this field.


As described in 1942 (pp. 205, 259, 279), the pretectal nucleus is permeated and covered with dense neuropil, which is continuous with that of all surrounding parts — tectum, habenula, and dorsal thalamus — with fibers passing in both directions. It receives terminals and collaterals from all chiasmatic bundles of the optic tracts (figs. 14, 22), tectum (figs. 12, 14, ir.t.pi.), posterior commissure, dorsal thalamus, habenula, and tr. strio-tectalis (figs. 14, 101, tr.st.pt.). Efferent fibers leave this nucleus for the tectum (figs. 11, 15, tr.pt.tec), thalamus (fig. 15), hypothalamus (figs. 15, 16, 25-36, tr.pt.hy.), and peduncle (tr. thalamo-peduncularis cruciatus and tr. thalamo-peduncularis dorsalis superficialis) . Some fibers accompany the fasciculus retroflexus to reach the area ventrolateralis pedunculi (fig. 23).


In my previous publications the tr. pretecto-hypothalamicus was not identified. Here its course is shown in figures 25-36. Many unmyelinated fibers, accompanied by a few with myelin sheaths, leave the pretectal nucleus (figs. 35, 36) and descend near the surface rostrally and internally of the marginal optic tract. Many of these fibers end in the thalamus (tr. pretecto-thalamicus, tr.pt. th.), and the remainder descend to decussate in the postoptic commissure and spread in the hypothalamus. These fibers and those here termed tr.



DIENCEPHALON 235

tecto-hypothalamicus anterior (tr.t.hy.a. of the figures) comprise the mixed system which in my previous papers was called "tr. tectothalamicus et hypothalamicus cruciatus anterior" (p. 296). The thickest myelinated fibers of this complex arise in the tectum and are clearly followed through the postoptic commissure into the hypothalamus.


The dominant afferent connections of the pretectal nucleus apparently are optic. The efferent connections to the hypothalamus and peduncle provide pathways for optic control of the intrinsic muscles of the eyeball, as in mammals, and also for possible conditioning or other influence upon the primary motor paths of the brain stem.


The pineal body appears in the premotile embryo as a low evagination from the roof of the diencephalon ('38, fig. 2). In early swimmers it is a hollow vesicle with a lumen, communicating with the third ventricle by a patent recessus pinealis ('38, figs. 3, 4, 18). Within three days after the early swimming stage this communication is closed ('38a, p. 18). The adult epiphysis is a small, flattened, simply lobulated epithelial vesicle, entirely detached from the brain except for a few fibers of the parietal nerve. No nervous elements other than these fibers have been found in it, and their connections are still unknown.


In Necturus there are more of these fibers, and their courses are more easily followed ('17, p. 236). These myelinated fibers are spread between the cells of the epithelial wall of the pineal vesicle, from which they pass into the brain in several small fascicles associated with the habenular commissure and posteriorly of it. They do not cross the mid-plane and they appear to have no functional connection with the epithalamus but to pass through it and join the f. retroflexus. They can be followed ventralward nearly to the cerebral peduncle, where they mingle with other myelinated fibers. The parietal nerve of Amblystoma has a smaller number of myelinated fibers than in Necturus (not more than 10), but their arrangement is similar ('42, p. 205). These certainly are not aberrant fibers of either the habenular commissure or the com. tecti, but their central connections and functions are uncertain. They resemble those associated with the IV nerve, which are distributed peripherally to the meninges and chorioid plexus (p. 181). It is not improbable that all these fibers have sensory functions, but for this there is no evidence. All chorioid plexuses are abundantly supplied with unmyelinated nerve fibers, probably of vasomotor function, but in the search for the sources of these fibers in Amblystoma I have been no more successful than I was with Necturus ('336, p. 15).


DOESAL THALAMUS

The dorsal thalamus is a well-defined part of the diencephalon and is intimately related by a web of connecting fibers with all contiguous parts and with many remote parts of the brain. It evidently is an important center of sensory correlation, but it is also the primary terminal station of a considerable fraction of the fibers of the optic nerve which come directly from the retina. The thalamic connection of optic fibers is not a late acquisition incidental to the differentiation of cerebral cortex, for it is present in all vertebrates, even the most primitive types. In corticated animals the larger part of the thalamus is clearly a cortical dependency, the neothalamus; but there is also a paleothalamus, which is a primordial part of the vertebrate brain, being present in all vertebrates, from the lowest to the highest. This paleothalamic component is not lost in mammals, as is shown by the surviving thalamic residue, which does not degenerate after total decortication.


The amphibian dorsal thalamus shows only vague outlines of an incipient subdivision into the local nuclei so characteristic of mammals. This is more evident in anurans than in urodeles, and in Amblystoma than in Necturus. In the urodeles, as in fishes, the undifferentiated dorsal thalamus may be regarded as a single "nucleus sensitivus," which acts, in the main, as a whole, without well-defined functional localization. Nevertheless, an early stage of local differentiation can be recognized. In Amblystoma there are three areas of gray, separated by shallow and variable ventricular sulci, namely :

1. Anteriorly the nucleus of Bellonci (fig. 2, nuc.B) produces a slight ventricular eminence wedged between the middle area, the ventral habenular nucleus, the eminentia thalami, and the ventral thalamus. Superficially of this gray is an area of very dense and complicated neuropil, which is intimately connected with that of all the surrounding areas. It receives terminals of the optic tract, and this justifies its inclusion within the sensory zone, though its other connections are those of correlating tissue of intermediate-zone type. The nucleus of Bellonci is primarily a bed-nucleus of the neighboring tracts — stria meduUaris, optic and postoptic systems, and others (chap, xviii).


2. The middle part of the dorsal thalamus is marked by a low ven


DIENCEPHALON 237

tricular eminence inclosed by the dorsal and middle thalamic sulci (fig. 2, s.d. and s.m.). It is the precursor of several of the sensory nuclei of the mammalian thalamus, though there is no visible local differentiation of its gray substance corresponding to these nuclei. This homology is clearly established by the location of these cells and their fibrous connections, which conform, as far as they are present, with the mammalian pattern (figs. 14, 15). This area receives terminals and collaterals of the optic tract, terminals of the spinal and bulbar lemniscus, strong tecto-thalamic tracts (precursors of the brachia of the superior and inferior colliculi), some fibers from the habenula, and probably also fibers from the hypothalamus by way of the postoptic commissure systems and from the cerebral hemisphere accompanying the tr. strio-pretectalis. These fibers all enter a common pool of neuropil, which is most dense in the intermediate alba but which also pervades the entire structure.


From this thalamic pool, efferent fibers go out to all surrounding parts of the brain, some in the diffuse neuropil and some as wellfasciculated tracts. An extensive and complicated system of efferents discharges into the underlying motor zone of the same and the opposite side, including direct and crossed tracts to the ventral thalamus, hypothalamus, cerebral peduncle, and isthmic and bulbar tegmentum. The more important of these are shown in figure 15. A large tract containing some thick myelinated fibers (tr.th.p.) joins similar fibers from the pretectal nucleus and tectum (fig. 18, tr.t.p.c), partially decussates in the commissure of the tuberculum posterius, and spreads through the alba of the peduncle on both sides, as illustrated in figures 30-36 {tr.th.p.c. and tr.t.p.c. 1.). Parallel with these deeper fibers there is a very large superficial uncrossed tract from the dorsal thalamus and pretectal nucleus to the peduncle (not shown on the figures) similar to that described in Necturus as tr. thalamo-peduncularis dorsalis superficiahs ('17, p. 264). Other similar fibers are dispersed in the neuropil, some of which go directly to the ventral part of the thalamus — tr. dorsoventralis thalami (shown but not labeled in figs. 15 and 16; see '17, p. 266). A small number of fibers go to the habenula (tr.th.hab.), and a larger number to the tectum in the brachia of the superior and inferior colliculus {br.coL). The large tr. thalamo-tegmentalis rectus includes deep and superficial fibers to the dorsal and isthmic tegmentum (figs. 32-34, 94, tr.th.teg.r.) ; for those which decussate in the postoptic commissure see chapter xxi and '42, p. 223.



?38 THE BRAIX OK THE TKJER SALAMANDER

The fibers wliich ascend to the heinispliere are in small, compact fascicles (tr. thalamo-froiitalis; see figs. 15, 19, oO-lU. 71. 7-2, 75. 95, 101, 1(H. 10:>, tr.fh.f.). which enter the lateral forebrain bundle. "Within this bundle the fibers pass to the strio-amygdaloid field in the ventrolateral wall of the hemisphere. Here they spread out and, so far as is now known, they end here, thus constituting a thalamostriatal system of projection fibers, which persists in mammals with the addition of the thalamo-cortical projection system. The thalamofrontal fibers arise as axons of cells of the middle part of the dorsal thalamus only, so that this "nucleus sensitivus thalami" may be regarded as the primordium of most of the mammalian thalamic nuclei which have cortical connections. This is not a new connection in the Amphibia, for it is pi-esent in various groups of fishes ('-2'-2a).


3. The posterior part of the dorsal thalamus is less differentiated and less clearly delimited. Its tissue is confluent with that of the pars intercalaris of the epithalamus. the eminence of the posterior commissure, the dorsal tegmentum, the peduncle, and the ventral thalamus. This suggests that it is functionally related with all these parts. All the connections of the middle part are more or less evident, except the thalamo-frontal tract. AYhat structures of higher brains have been derived from this area is not clear. Some of this tissue should probably be assigned to the ventral thalamus.


This posterior sector develops prtxxu^iously in company with the ventral thalamus vCoghill, "-2S, Paper VIII). Uncrossed impregnated fibers pass from it to the ventral thalamus and peduncle in the S-reaction stage — Harrison's stages 35, 36 (,'37, fig. 'i). In early swimmers these are more numerous, and similar fibers, which decussate in the post optic commissure, make their appearance (,'38, fig. 5: Coghill, '30, Paper IX, fig. -4). Differentiation of the anterior parts of the dorsal thalamus is relatively retarded.


The dense and intricate neuropil of the posterior part of the dorsal thalamus spreads to surrounding parts with no well-defineil boundaries, being intimately joined with that of the luiddle part and the ventral thalamus (fig. 73). This field (,figs. 14. 10, np.geu.) receives two systems of fibers which seem to have special phylogenetic significance, viz., abundant terminals of the optic tracts and the brachium of the inferior colliculus. The arrangement of these connections in Amblystoma is difi'erent from that seen in both Xecturus and the frog l,'-!'^, p. "278). Optic terminals are spread through the alba of the entire dorsal thalamus and most abundantly in this posterior neu


DIENCEPHALON >;{9

ropil. I have calleci tliis c'oiniiu)ii jxk)! of unditt'orentiated tissue of llie dorsal thalainiis and dorsal part of the ventral thalamus tlie "genieulate neuropil," on the assumption that out of it the medial and lateral geniculate bodies have emerged (]). '-I'-U). In the course of i)hylogeny the more sharp segregation of the optic terminals was accompanied by the differentiation of the lateral geniculate body, a process which is well advanced in the frog ('"25), and the differentiation of the cochlear apparatus and lateral lemniscus was accompanied by the emergence of the medial geniculate body from the same common pool. This hypothesis is held subject to revision, pending further study of the related species.


VENTRAL THALAMUS

On the ventricular surface (figs. 1, '2, 15>, 95) the ventral thalanuis is sharply delimited by the sulcus medius thalami (.s-.m.) above and the sulcus ventralis (s.r.) below, and it is separated into anterior and posterior parts by a depressed area containing in some si)ecimens a shallow sulcus. The two parts differ in embryological origin and connections, yet in the adult brain their most fundamental features are similar ('42, p. 207).


The posterior part of this field extends forward from the peduncle, and in early functional stages it has similar structure and connections, though it is clearly in diencephalic territory. In these stages the ])rimordium of the anterior i)art (area 7a of my analysis, ','??, p. 393) lies at the di-telencephalic junction ('38, p. 213 and fig. 18), and in prefunctional stages it is joined with area 7, which becomes the corpus striatum. This anterior part of the ventral thalamus may be genetically telencephalic, depending on how the arbitrary ditelencephalic boundary is defined.


In the adult ventral thalamus none of the nuclei of more highly differentiated brains are well defined, though some local differentiation is evident. Anteriorly, the eminentia thalami (figs. 2, 16, cni.fh.) belongs in a series of bed-nuclei related with important tracts of fibers at the di-telencephalic junction (chap, xviii). Both this area and the underlying nucleus of the olfacto-habenular tracts discharge fibers backward into the unspecialized gray of the anterior part of the ventral thalamus (fig. 17). Between the anterior part of the ventral thalamus and the dorsal (mamillary) part of the hypothalanuis there are fibers passing in both directions which seem to be precursors of the mammalian tr. mamillo-thalamicus ('396, p. 554 and figs. 22, 35).


The posterior part has a thicker gray layer, which produces a rather high ventricular eminence, the structure of which is similar to that of the peduncle. As illustrated by previously published figures ('396, p. 544 and figs. 23, 28), the large neurons of the ventral thalamus have widely spread dendrites and are evidently collectors of impulses from many sources (figs. 16, 17). Spinalward of the bednuclei at the anterior end, the two parts receive the strong tr. striothalamicus, which descends from the strio-amygdaloid field by way of the lateral forebrain bundle (f.lat.t.). Short fibers descend into both parts from the dorsal thalamus (tr. dorsoventralis thalami) and from the pretectal nucleus by tr. pretecto-thalamicus et hypothalamicus. There are much larger connections from the tectum, some uncrossed in the brachia of the superior and inferior colliculi (br.col.) and some decussating in the postoptic commissure (tr.t.th.h.c.p.).


Fibers stream backward from the ventral thalamus into the motor zone of all lower parts of the brain stem and into the hypothalamus, some uncrossed and some decussating in the postoptic commissure and in the ventral commissure of the isthmus. These have been seen especially clearly in larval brains ('39, fig. 2; '396, p. 546 and fig. 23). Most of these fibers end in the cerebral peduncle and tegmentum of the midbrain and isthmus ; many of them extend into the trigeminofacialis region; and relatively few which enter the f. longitudinalis medialis may go into the spinal cord.


The connections just described indicate that the ventral thalamus collects fibers from all parts of the cerebrum, and chiefly those concerned with somatic responses to exteroceptive stimuli. Its efferent fibers descend to those motor fields which supply the skeletal musculature. These fibers take a surprising variety of courses, the details of which need not be recounted here ('396, p. 544), but all the more important pathways are links in the chain of conductors which activate the skeletal muscles. For this reason this field is regarded as part of the motor zone, though it has no direct connections with the periphery.


This ventral thalamus corresponds in all essential respects with the subthalamus of mammalian neurology. In contrast with the mammals, it is here far larger than the dorsal thalamus, as is evident upon inspection of figure 2. This difference is probably due not to shrinkage of the ventral thalamus in mammals but to the great enlargement of the dorsal thalamus, i.e., to the addition of the neothalamus to the primordial paleothalamus as the latter is seen in Amblystoma.



DIENCEPHALON 241

HYPOTHALAMUS

The liypothalamus as here defined inchides the ventral part of the brain stem between the anterior commissure ridge and the tuberculum posterius. At its anterodorsal border the nucleus of the olfactohabenular tract might be included or assigned to the ventral thalamus. Its fibrous connections, like those of the preoptic nucleus, are mainly of hypothalamic type. Posteriorly of the chiasma ridge the deep sulcus hypothalamicus separates the dorsal from the ventral part of the hypothalamus, and each of these parts is further subdivided. A shallow and variable sulcus hypothalamicus dorsalis separates the dorsal hypothalamus again into dorsal and ventral lobes, both of which are confluent with the ventral thalamus and peduncle (fig. 2). These parts contain the primordium of the mamillary body, but this structure is not recognizably differentiated. The ventral hypothalamus is obscurely separable into posteroventral and anterodorsal parts. All these subdivisions are more clearly seen in Necturus than in Amblystoma ('35a, p. 253).


The sulcus preopticus separates the preoptic nucleus into anterior and posterior lobes. The lateral preoptic recess at the anteroventral border of the chiasma ridge is a remnant of the lumen of the hollow epithelial optic stalk of the embryo ('41), which persists in the endocranial part of the optic nerve of adult Necturus ('41a, p. 494).


The sequence of differentiation of the nervous connections of the hypothalamus has been summarized elsewhere, together with a description of these connections in midlarval stages ('396, p. 550). The adult structure and connections of the hypothalamus of Necturus have been fully described ('336, '346, '41a), and those of Amblystoma are essentially similar, though with considerable advance in differentiation throughout ('42, pp. 211 ff.). None of the hypothalamic nuclei described in mammals are here clearly segregated, though some of them are recognizable in more dispersed arrangement. The preoptic nucleus is an exception to this. It is very large, and its boundaries are well defined except anterodorsally, where it merges with the nucleus of the olfacto-habenular tract, and posterodorsally, where it merges with the ventral part of the hypothalamus.


AFFERENT CONNECTIONS

The nervus terminalis, as described by McKibben ('11), has endings distributed throughout the hypothalamus both before and behind the chiasma ridge. The only other peripheral fibers which reach the hypothalamus are a few which separate from the optic tracts and ramify in the vicinity of the chiasma (p. 221).


By far tlie larger part of the white substance of the hypothalamus is occupied by the great medial forebrain bundles (f.med.t. of the figures); and, since these bundles are composed chiefly of fibers descending from the olfactory field of the hemispheres, the implication is that olfactory functions are dominant here. The olfacto-peduncular tract arises chiefly from the head of the caudate nucleus and distributes some of its fibers to the dorsal part of the hypothalamus. Fibers descend in the dorsal fascicles of the medial forebrain bundle from the septum and primordium hippocampi and distribute chiefly to the preoptic nucleus and dorsal hypothalamus. Included here are the precommissural fornix and part of the stria terminalis system. The ventral fascicles have descending fibers from the olfactory bulb and from the ventral and medial sectors of the anterior olfactory nucleus, which spread throughout the preoptic nucleus and the ventral part of the hypothalamus. Strong collaterals of both dorsal and ventral fascicles enter the stria medullaris thalami. Many fibers descend from the preoptic nucleus in diffuse formation, to spread throughout the hypothalamus posteriorly of the chiasma ridge. One component of this preoptico-hypothalamic connection is the compact tr. preopticus, as described below.


The dorsal olfactory projection tract (ol.p.tr.) is a compact fascicle of unmyelinated fibers, which descend from the dorsal striatal nucleus and in larger number from the amygdala. This fascicle accompanies the lateral forebrain bundle, and above the chiasma it turns ventrad to connect with its nucleus at the posterior border of the chiasma ridge. This tract also contains ascending fibers. It was first described in Ambly stoma and the frog ('21a) and subsequently in Necturus ('336, p. 158) ; its course was fully illustrated in 1921 and is shown in many other figures (figs. 19, 25, 26, 27, 95, 96, 97, 101, 102, 103; '27, figs. 11-20, 24-32; '36, fig. 5).


Afferent fibers of the visceral-gustatory system enter the dorsal part of the hypothalamus (fig. 8) . The ventral part receives fibers from the pretectal nucleus by tr. pretecto-hypothalamicus (figs. 15, 16), and much larger numbers by the accompanying tr. thalamohypothalamicus, and also by tr. thalamo-hypothalamicus cruciatus (fig. 16). The entire tectum is connected with the hypothalamus by two large tracts — (1) the anterior division of tr. tecto-thalamicus et


DIENCEPHALON 243

hypothalamicus cruciatus (tr. tecto-hypothalamicus anterior) from the superior colliciilus and {%) the posterior division of this tract from both inferior and superior colhcuh (fig. l'^).


Through these manifold afferent connections nervous impulses are discharged into the hypothalamus from almost all parts of the brain, directly or indirectly. All sensory systems are represented here, but the olfactory connections evidently are preponderant. The significance of this convergence will appear after comparison with similar pools in the habenula and several other places (p. 252).


INTRINSIC HYPOTHALAMIC CONNECTIONS

All parts of the hypothalamus are interconnected by dense and intricately woven neuropil in both gray and white substance. Though all activities of the hypothalamus may thus be integrated, the structure is diversified, indicating the inception of specialization of the local nuclei as these are seen in higher animals. The most complicated and interesting of these local fields of neuropil surrounds and permeates the chiasma ridge, of which I have given a detailed description ('42, p. 214). Neuroblasts of this area are differentiated very early in embryogenesis ('37, '38), and some of their axons form tr. hypothalamo-peduncularis (figs. 18, 23) and tr. hypothalamo-tegmentalis (fig. 21), as elsewhere described ('42, p. 226). This area seems to be the primary focus into which converge all tracts of the ventral part of the hypothalamus for discharge through the two tracts just mentioned. Among these afferents are the intrinsic fibers of tr. preoptico-hypothalamicus from in front and of tr. infundibularis ascendens from the postero ventral lobe.


The last-mentioned ascending tract divides. One moiety is directed dorsally, providing a broad connection from the ventral part of the hypothalamus to the dorsal, and a larger moiety is directed forward into the medial forebrain bundle. Most of these ascending fibers apparently end in the postchiasmatic neuropil and preoptic nucleus, but some may go farther into the hemisphere ('336, p. 250; '346). This tract is quite independent of the ascending fibers of the olfactory projection tract previously mentioned. There are probably other fibers which ascend from the hypothalamus in the basal forebrain bundles, but our preparations have not revealed them.


From the anterior part of the ventral thalamus, fibers stream backward into the dorsal part of the hypothalamus. These thalamomamillary fibers are accompanied by fibers passing in the reverse direction — tr. mamillo-thalamicus — the combined tract being the probable precursor of the mammalian mamillo-thalamic bundle of Vicq d'Azyr.


EFFERENT CONNECTIONS

From the whole extent of the preoptic nucleus fibers pass dorsad to enter the stria medullaris thalami. Others probably ascend to the hemisphere in the basal bundles. The dorsal (mamillary) part of the hypothalamus is connected with the anterior part of the thalamus, as just mentioned. There is probably also a mamillo-cerebellar connection (p. 170). The other efferents from the mamillary region are the very extensive and complicated systems of mamillo-tegmental, peduncular, and interpeduncular fibers illustrated crudely in figures 18 and 21 and mentioned on page 278, where some references to literature are given (for fuller description see '396, p. 551).


The efferents from the ventral part of the hypothalamus include the ascending fibers and the hypothalamo-peduncular and tegmental systems already mentioned. Fibers may go out with some components of the postoptic commissure to the thalamus and tectum, but these have not been recognized.


The pars magnocellularis of the preoptic nucleus gives rise to the large hypophysial tract, which is one of the major features of the ventral hypothalamus. In most lower vertebrates, including some amphibians, these large cells are aggregated as a well-defined nucleus, homologous with the supraoptic nucleus of mammals. But in Amblystoma, as in Necturus, they are dispersed, being most numerous above and anteriorly of the chiasma ridge. These widely scattered large cells in the aggregate are here termed the nucleus magnocellularis. Their long dendrites are widely branched and may be activated by practically all nerve fibers of the preoptic and epichiasmatic areas. In this mixed collection of fibers there are two systems which seem to be specifically related with these large cells.


The first of these systems is the tr. preopticus (fig. 2C, 96, 97, tr.po.). These fibers are axons of cells at the postero ventral border of the anterior commissure ridge and the anterior part of the preoptic nucleus, which descend in the thin floor of the long preoptic recess, then recurve dorsally along the anterior face of the chiasma ridge, where they spread in the alba among dendrites of the cells of the nucleus magnocellularis. In the catfish, Ameiurus, the course is simi


DIENCEPHALON 245

lar, and here the connection of these fibers specifically with the compact nucleus magnocellularis is unmistakable ('41&). This, accordingly, is one source of activation of the hypophysial tract, the course of which in Ameiurus has been described by Palay ('45). His observations also indicate that in these fishes the nucleus magnocellularis itself is a source of endocrine secretion, as is clear also from the work of Scharrer and Scharrer ('40).


A second specific source of excitation of the nucleus magnocellularis is the small number of fibers which separate from the optic tracts near the chiasma and arborize in the area where these cells are most abundant (p. 221), thus providing visual control of endocrine secretion. This, however, is minimal in Amblystoma, and the very extensive hypophysial innervation must be concerned in the main with other functions.


The very numerous unmyelinated axons of the cells of the nucleus magnocellularis are diflficult to follow, for they descend in dispersed arrangement, penetrate the chiasma ridge, and then converge into the compact tr. hypophysius in the thin floor of the infundibulum. Their further course is clear. We have many brilliant Golgi impregnations of their distribution, some of which have been illustrated ('42, figs. 55-65). In the pars nervosa of the hypophysis they form a very dense neuropil, are less abundant in the pars intermedia, and a few of them penetrate the pars distalis. They terminate in tiny end-bulbs resting upon the epithelial cells. The pars nervosa is a rather thin sheet of convoluted epithelium which forms the posterodorsal wall of the wide infundibulum. These lobules take the form of irregular villi, in the axis of each of which is a capillary loop.


CONCLUSION

In most lower vertebrates the hypothalamus is relatively much larger than in the higher groups, and this seems to be correlated with the dominance of the olfactory system in the organization of the forebrain of the lower forms. All other functional systems have extensive representation here, and in this respect the hypothalamus resembles the habenular complex, but with the difference that in the latter the afferents converge into a very small compact area with one major efferent path, the f . retroflexus, while the much larger hypothalamus is diversified in structure and has a wider range of distribution of its efferent fibers. Though both these complexes are evidently concerned


24(5 THE BRAIN OP^ THE TIGER SALAMANDER

with the correlation of olfactory with a wide variety of other types of sensory experience, it is equally evident that the type of adjustment made is radically different. To this topic we shall return (p. 252).


The intricate connections described above are very different from those of the specialized fishes. How they are related with those of mammals remains uncertain, pending further study of intermediate species. The lack of differentiation of a localized mamillary body here is probably explained by the failure of the fibers of the primordial postcommissural fornix to reach the hypothalamus (p. %55) ; and this, in turn, is correlated with the primitive structure of the hippocampal formation. The analysis of the intricate system of postoptic commissures (chap, xxi) probably provides an instructive point of departure for further study of these connections, leading up to a solution of still controversial problems about the mammalian supraoptic commissures.


The nervous connections of the hypophysis in urodeles are very large, and they are so arranged as to be easily accessible for experimental study. It is hoped that advantage will soon be taken of this favorable material for investigation of some problems of endocrinol

ogy.



CHAPTER XVIII THE HABENULA AND ITS CONNECTIONS

IN ALL vertebrates with fully evaginated cerebral hemispheres the deep transverse stem-hemisphere fissure separates the dorsal parts of the hemisphere from those of the diencephalon. Ventrally of the floor of this fissure there are two great systems of fore-and-aft conduction — the basal forebrain bundles and the stria medullaris thalami. These form a superficial ventrolateral eminence, which is more conspicuous in the frog, where it was termed by Gaupp the "prominentia fascicularis." Attention has been called to the fundamental difference between these two great systems of fibers. The basal bundles are concerned primarily with the patterning of behavior, and, accordingly, their composition and connections vary widely from species to species in conformity with diverse modes of life. The composition of the stria medullaris, on the other hand, is remarkably constant throughout the vertebrate series, and its functional role is obscure. This chapter is devoted to these fibers and their widely dispersed connections, together with a few other systems of fibers which are associated with them.


THE DI-TELENCEPIIALIC JUNCTION

All afferent fibers to the habenula, with the exception of the tectohabenular tract, enter the stria medullaris thalami for longer or shorter parts of their courses. These fibers are assembled from all parts of the cerebral hemisphere and preoptic nucleus and, in smaller number, from the thalamus. They form a massive fasciculus, which ascends vertically in the posterior lip of the stem -hemisphere fissure; and associated with them is a series of bed-nuclei at the di-telencephalic junction, which will next be described. These nuclei have diverse and complicated connections, and all have this in common, that they are in functional relations with the stria medullaris. The following are listed in this series; for their arrangement see figure 2B.


1. Preoptic nucleus. — Among the connections of this nucleus, as described in the preceding chapter, are terminals and collaterals of descending fibers of the nervus terminalis, medial forebrain bundle, precommissural fornix, stria terminalis, and ascending fibers of the medial bundle. Fibers ascend from all parts of this gray to the habenula in the stria medullaris, some externally of the lateral forebrain bundle (tractus olfacto-habenularis lateralis) and some medially of it (tr. olfacto-habenularis medialis) . Among these fibers are strong collaterals from both dorsal and ventral fascicles of the medial forebrain bundle ('396, p. 538).


2. Bed-nuclei of the anterior commissure (figs. 25, 26, 27, 97). — These are cells sparsely scattered in the anterior commissure ridge and massed laterally of it. They are related with tracts of the basal forebrain bundles decussating here, the stria terminalis, and the tr. septo-habenularis .


3. Nucleus of tr. olfacto-habenularis (figs. 27, 28, 29, nuc.tr. ol.h.; '35a, p. 250; '396, p. 538).— This gray is interpolated between the anterior commissure ridge, the anterodorsal part of the nucleus preopticus, the anterior end of the ventral thalamus, and the eminentia thalami. It is penetrated by fibers of the tr. olfacto-habenularis medialis, with which it is related by terminals and collaterals. It is also connected with the amygdala by dispersed thick fibers, some of which are myehnated, termed "tr. amygdalo-thalamicus," though the direction of conduction is unknown (shown but not named on figs. 19 and 30; also shown on fig. 96, where the nucleus of the tr. olfacto-habenularis medialis is marked, p.v.th.). This tract is probably a part of the complicated stria terminalis system. It was described in Necturus ('336, p. 218) as connected with the nucleus of Bellonci, but its main connection is with the nucleus of the olfactohabenular tract, as shown in figure 40 of the paper cited. Impregnated neurons of the posterior part of this nucleus are shown in figure 24 of the paper of 1942. Their unmyelinated axons take widely divergent courses — to the stria medullaris, the amygdala, medial and lateral forebrain bundles, and the ventral thalamus.


In describing the early development of both dipnoan and amphibian brains, Rudebeck ('45) recognizes two parts of the nucleus preopticus: (1) an inferior "nucleus preopticus proper," most of which is developed rostrally of the sulcus intraencephalicus anterior, and (2) a pars superior, developed posteriorly of this sulcus in the ventral part of the di-telencephalic ridge. The superior part he regards as the primordium of the amygdala of the adult amphibian. This part is larger in dipnoan embryos than in amphibians and may be the source of a more extensive area of the adult brain. In Amblys


THE HABENULA AND ITS CONNECTIONS 249

toma the posterior lobe of the inferior or "proper" nucleus preoptions develops posteriorly of the sulcus intraencephalicus anterior; and, more dorsally, Rudebeck's pars superior of this nucleus is clearly the primordium of my nucleus of the olfacto-habenular tract, including, perhaps, also the posterior end of the amygdala. In any case the two nuclei last mentioned are very intimately related.


4. Eminentia thalami. — This name was given ('10, p. 419) to a prominent ventricular eminence lying behind the interventricular foramen between the anterior commissure ridge and the habenula, which has been variously interpreted by authors. It is appropriately named "nucleus commissurae hippocampi" by Addens ('46). I have regarded it as the anterior end of the ventral thalamus, differentiated as bed-nucleus of several large tracts which converge here. It is well defined in early swimming stages (area 7a) , when the first fibers of the stria medullaris are visible ('38, p. 213 and figs. 18, 19) ; and in early feeding stages, with good development of stria medullaris and hippocampal commissure, it has acquired essentially adult form ('386, p. 401, figs. 1, 2, 6). It is derived from the middle of the di-telencephalic ridge of Rudebeck's ('45) pictures of the embryonic brain.


This eminence receives collaterals from the stria medullaris and in larger numbers from the fibers of the hippocampal commissure (figs. 31, 32, 71, 76; '27, p. 294; '35a, p. 250). As the latter fibers swing downward behind the interventricular foramen toward their crossing in the anterior commissure ridge, they are accompanied by similar fibers which pass from the primordium hippocampi to the ventral thalamus — tr. cortico-thalamicus medialis (figs. 31, 32, 33, 72, 75). Like the fibers of the hippocampal commissure, these are mostly unmyelinated, with a few thick, well-myelinated fibers scattered among them. Many of them end in the eminentia thalami, and others continue into the anterior part of the ventral thalamus. This tract is probably the precursor of the mammalian column of the fornix, though here none of its fibers have been seen to reach the hypothalamus.


Numberless short, thin, unmyelinated axons descend from the small cells of this eminence to spread in the adjoining ventral thalamus ('21a, figs. 26-28; '27, p. 269), and thicker axons enter the complex crossed and uncrossed system of tr. thalamo-tegmentalis ventralis ('39, p. 120 and fig. 2; '396, p. 546; '42, p. 225).


5. Nucleus of Bellonci. — This name is applied to a group of cells which form a low ventricular eminence between the anterior parts of the dorsal and the ventral thalamus. I have regarded it as part of the dorsal thalamus, though its morphological status is uncertain. This nucleus was not recognized by Bellonci ('88), hut he did describe in the frog a peculiar field of very dense neuropil related with it, which I term the "neuropil of Bellonci" in preference to his name, nucleus anterior superior corporis geniculati thalami." This nucleus and associated neuropil I first described in Necturus ('17, p. 243) under the name of "pars optica thalami," but this term also is inappropriate and has been discarded. In Necturus and Ambly stoma this peculiar neuropil receives terminals and collaterals from the optic tract, stria medullaris, tr. cortico-thalamicus medialis, the anterior tectal fasciculus (p. 297), and other sources. Axons of the cells of the nucleus go to the habenula and to the ventral thalamus and peduncle. This neuropil has received much study ('25, p. 454; '27, p. 295; '33&, p. 216; '34, p. 107), but, as recently remarked ('42, p. 279), many details of its structure and connections are still obscure. The probable equivalent of this nucleus has been found in all major groups of vertebrates by Addens ('46 and papers there cited).


6. Habenula. — The general structure and chief connections of the amphibian habenular nuclei are well known and have been summarized by Ariens Kappers, Huber, and Crosby ('36). There are two habenular nuclei on each side, different in position from that of mammals but with similar fibrous connections. They are arranged one above the other at the anterodorsal border of the diencephalon, dorsally of the eminentia thalami. The taenia thalami, or line of attachment of the membranous dorsal sac, extends from the eminentia thalami along the anterior border of both nuclei and backward along the dorsolateral border of the dorsal nucleus as far as the habenular commissure. Posteriorly of this commissure the roof is again membranous as far as the recessus pinealis. The dorsal and ventral nuclei are separated on the ventricular surface by a shallow sulcus intrahabenularis and bounded posteriorly by the sulcus posthabenularis, which is the surviving dorsal end of the sulcus diencephalicus ventralis of Rudebeck ('45). Below the ventral nucleus is the deep sulcus subhabenularis. This is essentially the embryonic arrangement, in contrast with that of adult man, where the great enlargement of the dorsal thalamus and reduction of the membranous dorsal sac and paraphysis result in great changes in the relative positions of parts.


The ventral and dorsal habenular nuclei of Amblystoma are probably the equivalents, respectively, of the lateral and medial nuclei of


THE HABENULA AND ITS CONNECTIONS 251

higher forms, but the details of their connections are not exactly comparable. Both nuclei receive terminals of the stria medullaris thalami (apparently all components of it), and the chief efferent path of both nuclei is the tr. habenulo-interpeduncularis, which is the largest component of the fasciculus retroflexus (of Meynert). The dorsal nucleus is connected with the tectum and pretectal nucleus by fibers running in both directions, and the ventral nucleus is similarly connected with the dorsal thalamus, the ventral border of the tectum, and the dorsal tegmentum.


In the ventral habenular nucleus the nerve cells are relatively few in the usual arrangement at the ventricular side. The alba is almost completely filled by the massive stria medullaris, internally of which is neuropil, which invades the gray substance. Most of the cell bodies are widely separated within this dense neuropil. The larger dorsal nucleus has very numerous small cells, densely crowded on all sides except dorsally of the commissure. This layer of cell bodies incloses a central core of dense neuropil, which connects with the stria medullaris below and the habenular commissure above. The thick contorted dendrites ramify in the central neuropil (fig. 73; '42, figs. 77, 79).


Primitively, these nuclei had direct connection with the parietal eye, as is the case in some still living species. It is possible that in some early ancestor of the vertebrates, now extinct, the dorsal parietal eye was better developed than the lateral eyes and that for this reason the primordial olfacto-visual correlation was made in the epithalamus rather than in the tectum of the midbrain.


The habenular system is one of the most conservative parts of the vertebrate brain. Edinger's statement ('11, p. 370) that this is perhaps the only part of the brain the organization and connections of which show no alterations during the whole course of vertebrate phylogeny requires some qualification; yet, as we pass from cyclostomes to man, with revolutionary changes in all surrounding parts, the chief habenular connections show a surprising uniformity.


In urodeles afferent fibers enter the habenula from all parts of the cerebral hemisphere, except perhaps the olfactory bulb (some authorities would not make this exception) and also from the preoptic nucleus, thalamus, and tectum. The largest of these tracts come from areas which are under the strongest olfactory influence, and the habenular complex is generally regarded as primarily concerned with olfactory adjustments. That this is not its only function is evident from the fact that in anosmic animals, like some birds and cetaceans, the habenular system, though reduced in size, retains most of its components. The associational connections of the habenula with the tectum, dorsal thalamus, and adjacent regions justify the conclusion that the complex, viewed as a whole, is adapted to insure the correlation and integration of the activities of all parts of the olfactory field with those of all other exteroceptive systems.


The sense of smell is both interoceptive and exteroceptive. Olfactory nervous impulses of exteroceptive type are somehow sorted out from the entire olfactory field and converged into the epithalamus, and those of interoceptive type are similarly converged into the hypothalamus. In the former of these areas they may be correlated with all relevant somatic sensory experience, and in the latter with the sum total of visceral experience. The habenula discharges into that portion of the motor field known to control the activities of the skeletal musculature and chiefly into the interpeduncular nucleus, which is imbedded within this motor field and articulated with it in very complex patterns (chap. xiv). The olfacto-visceral correlations of the hypothalamus may come to expression in overt action of the skeletal muscles or in less obvious visceral changes. The chief nervous pathways involved in the overt responses can be identified, but those of visceral activities are still obscure. There is a strong connection from the hypothalamus to the interpeduncular nucleus.


Since both the hypothalamic and the epithalamic olfactory systems are large in all vertebrates, it was suggested by Edinger ('11, p. 371) that these regulate the feeding activities, or muzzle reflexes {Oralsinnapparat) — dorsally the exteroceptive components and ventrally the interoceptive. This attractive hypothesis seems to explain the different courses and connections of the medial forebrain bundle and stria medullaris, but it leaves out of account the equally large lateral forebrain system. Moreover, the chief connections of the medial and lateral forebrain bundles are localized in the hemisphere, respectively medially and laterally of the ventricle; but the habenular connections are drawn equally from both sides of the hemisphere. The habenular connections seem to be of a different kind, something added to an existing adequate provision for both visceral and somatic adjustments.


The chief efferent discharge from the habenula is to the interpeduncular nucleus, and there is a large efferent tract from the hypothalamus to this nucleus (figs. 18, 21). It is suggested in chapter xiv that the habenulo-interpeduncular system is the inhibitory com


THE IL\BENULA AND ITS CONNECTIONS 253

ponent of an equilibrated dynamic system, of which the basal forebrain bundles comprise the activator component. On this hypothesis the activation of all olfacto-motor systems goes out through the basal forebrain bundles, and this activation is accompanied by an inhibition of all conflicting activities, the inhibitory component of the reaction being centered in the interpeduncular nucleus. This implies that inhibitory influences derived from exteroceptive fields are transmitted to the interpeduncular nucleus by the f . retroflexus and from interoceptive fields by the mamillo-interpeduncular tract. This hypothesis seems to be consistent with the known structure, but it lacks experimental proof. If there is factual basis for it, the activating and inhibitory systems must not be regarded as independent units of structure; they are everywhere interconnected, and all their activities are balanced one against the other in an integral dynamic system. The literature contains many fragmentary accounts of the habenular connections of Amphibia, with some conflict of observation and confusion of nomenclature. When these descriptions are assembled, together with additional observations here reported, the salient features of the habenular connections may be grouped into three classes. These are:

1. There are two groups of commissural connections between the two cerebral hemispheres. Group 1 comprises the commissura superior telencephali, with fibers from the anterior olfactory nucleus, piriform area, and amygdala (components 3, 4, 5, and 6 of the list given on pp. 257-60). This is the larger part of the habenular commissure. It was seen by van Gehuchten in Salamandra ('97rt), though its true origins were not recognized and his belief that none of these fibers have connection with the habenular nuclei needs confirmation. Group 2 is the com. pallii posterior (component 8 of the list). These fibers come from the primordium hippocampi and are probably comparable with the com. aberrans of some reptiles.


2. The habenula is an important way-station for through traflSc from all parts of the cerebral hemisphere and preoptic nucleus to the brain stem below this level, particularly to the peduncle and interpeduncular nucleus. All these conduction pathways except the commissural fibers are interrupted by synaptic junctions in the habenula, and the habenular synaptic field is connected with neighboring regions by fibers passing in both directions.


3. The through traffic is influenced by these local connections with areas of the sensory zone concerned primarily with exteroceptive functions. The habenula, accordingly, may be regarded as a bednucleus interpolated in one of the main pathways from higher centers of correlation, where olfaction as an exteroceptive function is integrated with other functions of this type and transmitted to lower centers in the motor zone, where patterns of response are organized. The hypothalamus is a similar way-station for through traffic from the same higher centers to lower motor fields, with olfacto- visceral functions dominant.


Most of the afferent connections of the habenula are in the stria medullaris, and most of the efferent connections in the f. retroflexus. Before describing this chief thoroughfare of through traffic, mention should be made of two systems of fibers which are in intimate functional relation with the stria medullaris, though not component parts of it. These are the fornix and stria terminalis.



In mammals this name is given to a complicated system of fibers which descends medially from the hippocampal formation to the underlying brain stem in two groups separated by the anterior commissure. Both groups descend within or adjacent to the lamina terminalis dorsally and rostrally of the interventricular foramen. The postcommissural fornix, commonly called columna fornicis, passes downward between the foramen and the anterior commissure and thence across the thalamus to end chiefly in the mamillary body. The precommissural fornix is a complex system of more loosely arranged fibers descending in front of the anterior commissure to the gray of the septal nuclei and neighboring parts and continuing spinalward in the medial forebrain bundles to preoptic nucleus, hypothalamus, and (according to some descriptions) as far as the cerebral peduncle.


Primordia of the two main divisions of the mammalian fornix are obviously present in the amphibian brain, but the topography of the region of the lamina terminalis is here so different that these tracts take peculiar courses. The anterior and hippocampal commissures cross, not in the lamina terminahs, but in a commissural ridge behind the interventricular foramen (fig. 2; '27, p. 235; '35, p. 299). To reach this crossing, the fibers of the massive hippocampal commissure swing downward behind the foramen (figs. 72, 76), and here they are accompanied by the fibers of tr. cortico-habenularis medialis and tr. cortico-thalamicus mediahs (fig. 20, nos. 8 and 9), which, as pointed


THE HABENULA AND ITS CONNECTIONS ^55

out in the preceding description of the eminentia thalami, marks the beginning of the differentiation of the mammahan columna fornicis. Some of these fil)ers are collaterals of those of the com. hippocampi and tr. cortico-habenularis medial is. In Ambly stoma none of these fibers have been followed as far as the dorsal (mamillary) part of the hypothalamus, so that the homology with the mammalian postcommissural fornix is incomplete. In other urodeles it has been described as reaching the hypothalamus (Rothig, '24, pp. 10, 15; Salamandra, Kreht, '30, p. 252). The precommissural fornix is an ancient system, being well developed in many fishes. In Ambly stoma it is large and of typical pattern. These fibers descend from the hippocampal formation rostrally of the foramen to the medial forebrain bundle (figs. 98, 99), within which they descend for an undetermined distance.


' The transverse sections described in 1927 revealed clearly the relations of the precommissural fornix ('27, p. 311), but the postcommissural connections were obscure. In Necturus these connections are perfectly clear, and the morphological relations of precommissural and postcommissural fornix fibers have been discussed ('336, p. 188). Amblystoma resembles Necturus. Accompanying the unmyelinated fibers of the hippocampal commissure there are similar fibers of tr. cortico-habenularis medialis and tr. cortico-thalamicus medialis, and among these are a few myelinated fibers. In the horizontal Cajal sections (figs. 28-32) the very numerous thin, fibers of this complex are not impregnated, but the thick axons are deeply stained. Few of the thick fibers decussate in the hippocampal commissure (figs. 28, 29, 30). Many of them ascend in the stria medullaris (figs. 32, 33, 34, tr.c.h.m.), and most of the others pass backward and downward through the eminentia thalami into the ventral thalamus and (probably) the preoptic nucleus (figs. 31, 32, tr.c.th.ni., 72, 75). These thick fibers and the much more numerous thin fibers by which they are accompanied I have termed "tr. cortico-thalamicus medialis." An extension of this tract into the hypothalamus would give the connection described by Kreht ('30) in Salamandra and by Loo ('31, p. 84) as tr. cortico-hypothalamicus in the opossum. The mammalian tr. cortico-mamillaris is a further addition to this system.


STRIA TERMINALIS

In mammals the ventromedial septal field is connected with the ventrolateral amygdala by fibers, passing probably in both directions, some ventrallv of the basal forebrain bundles (diagonal band of Broca) and some dorsally of them (stria terminalis, stria semicircularis, or stria cornea). Here the enormous enlargement of the basal forebrain bundles, particularly the internal capsule system of fibers, displaces the dorsal stria so that its course between medial and lateral ends is a wide arc, in some species almost a complete circle. At both of these terminal areas the stria makes complicated connections with the preoptic nucleus and hypothalamus, some of which were termed by Cajal the "olfactory projection tract." In Amblystoma, with smaller basal forebrain bundles, these fibers take more direct courses, and they are more dispersed so that there is no compact stria terminalis as in mammals. For description and illustrations of these tracts the reader is referred to my paper of 1927 (p. 302). The following connections have been described:

1. The com. amygdalarum, crossing in the anterior commissure

(fig. 97). '

2. The stria terminalis sensu strido (fig. 96), a massive connection passing downward from the amygdala medially of the basal forebrain bundles. Some of its fibers are directed rostrad toward the septal area, but most of them turn posteriorly in the dorsal fascicles of the medial forebrain bundle to reach the preoptic nucleus and dorsal part of the hypothalamus. The rostrally directed fibers of this group and the commissural fibers are comparable with the fascicles of the stria which in mammals follow the taenia semicircularis and tail of the caudate nucleus, arching over the internal capsule and connecting the amygdala with the septal field. None of the fibers of this tract are drawn in the figures of horizontal sections. At the levels of figures 27-30 the descending fibers accompany the tr. olfacto-habenularis medialis and then turn spinalward in the dorsal fascicles of the medial forebrain bundle. Accompanying these are shorter myelinated fibers which pass between the amygdala and the nucleus of the olfactohabenular tract— the tr. amygdalo-thalamicus (p.- 248).


3. The ventral olfactory projection tract, descending from the amygdala externally of the basal forebrain bundles. These fibers are shown but not named in figure 16 of 1927.


4. The dorsal olfactory projection tract (p. 242, figs. 19, 25, 26, 27, ol.p.tr.) .


STRIA MEDULLARIS TIIALAMI

In cyclostomes, with incomplete evagination of the cerebral hemisphere, the connections between the dorsal parts of the telencephalon


THE HABENULA AND ITS CONNECTIONS 257

and of the diencephalon are massive and direct ('22«, fig. 8), and the habenular tracts are not fasciculated to form a stria medullaris. With further evagination of the hemisphere in amphibians and the appearance of a deep stem-hemisphere fissure, the habenuhir connections of the palHal parts of the hemisphere must turn downward to pass under the floor of this fissure. Here they are joined by other habenular tracts to form the stria medullaris thalami. These tracts of Amblystoma were described in 1910 (and further details, '27, p. 284 and figs. 15-18; '396, p. 538), and now some corrections and additions can be contributed. Compare also my description of the habenular connections of Necturus ('336, pp. 204-14).


The components of the stria which have been identified in Amblystoma are shown in figure 20, where they are projected upon the median plane. The courses of four of the more lateral components are shown as projected upon the lateral aspect of the brain in figure 85. Their arrangement as seen in horizontal sections is shown in figures 25-36 and in sagittal sections in figures 74-78. Some components are electively impregnated in our Golgi preparations, and these specimens have been especially useful in demonstration of the truly commissural connections of several of the tracts listed below. Components 3, 4, 5, and 6 are known to decussate in com. superior telencephali and component 8 in com. pallii posterior. On the figures the components of the stria are numbered as in the following list:

1. Tractus olfacto-habenularis medialis.


2. Tractus olfacto-habenularis lateralis.


Most of the fibers of these two tracts are axons of cells of the preoptic nucleus, ascending, respectively, medially and laterally of the basal forebrain bundles (figs. 25-30, 74-77; '27, fig. 16; '42, figs. 18, 24), but some of them are collateral branches of axons of the medial forebrain bundle ('396, p. 538). This is doubtless true in Necturus also, though I was not able to demonstrate it ('336, p. 208). In the passage last cited, references are given to descriptions of this connection in other urodeles. The fibers of these tracts are assembled from all parts of the preoptic nucleus, ascending on both sides of the sulcus preopticus. The fibers of the medial tract pass through the gray of the nucleus of this tract, with many terminals and collaterals, and here the tract receives accessions from the nucleus. More dorsally both tracts have collateral connections with the eminentia thalami and neuropil of Bellonci. They form the most posterior component of the stria medullaris, and within the habenular nuclei they spread out in the neuropil. Wliether any of them cross in the habenular commissure has not been determined. A few of these fibers are myehnated.


3. Tractus olfacto-habenularis anterior, ventral division.


4. Tractus olfacto-habenularis anterior, dorsal division.


As shown in figure 85, these fibers arise chiefly in the medial sector of the nucleus olfactorius anterior, from which they diverge ventrally and dorsally in a medial fascicle of fibers, which I have called the "f. postolfactorius" (fig. 100; '27, p. 283, figs. 2-4). The fibers of this system which are directed ventrally (no. 3) arise also in the ventral sector of the anterior nucleus and the septum. They take a posterolateral course along the ventral surface of the hemisphere, turning dorsally at the level of the anterior commissure to enter the stria medullaris under the posterior pole of the hemisphere ('27, p. 284 and figs. 5-17, tr.ol.hah.ant.). In the stria they lie anteriorly of the tr. olfacto-habenularis (1 and 2) and posteriorly of the tr. corticohabenularis lateralis (5), as shown in sagittal sections (figs. 75, 76, 77, tr.ol.h.a.v.). In the horizontal sections (figs. 25-34) this tract is marked ir.ol.h.a. Some of these fibers terminate in the habenular neuropil, and some cross at the posterior end of the habenular commissure in the com. superior telencephali.


The dorsal fibers of this system (no. 4) cross from the medial to the lateral side of the hemisphere in the f. postolfactorius between the olfactory bulb and the anterior olfactory nucleus (fig. 5) and then turn posteroventrally across the lateral aspect of the hemisphere in company with the large tr. olfactorius dorsolateralis ('27, p. 284). Over the piriform area these fibers are joined by those of tr. corticohabenularis lateralis (no. 5), and the mixed fascicle enters the stria medullaris under the posterior pole of the hemisphere. In the stria the mixed fascicle 4 + 5 ascends anteriorly of no. 3, and at the boundary between ventral and dorsal habenular nuclei it splits into lateral and medial bundles. The lateral fibers are dispersed in the habenular neuropil, and the medial fibers decussate in the com. superior telencephali (figs. 74-77, tr.c.h.l.). This component is shown in horizontal sections in figures 30-34, and its entire course is seen in transverse sections in figures 6-17 of 1927 {tr.c.hah.L).


The anterior olfacto-habenular tract is the first component of the stria medullaris to mature in ontogeny, its fibers appearing in early swimming stages ('38, pp. 222, 238, figs. 10, 19); and its relations in cyclostomes indicate that it is very ancient phylogenetically ('336,


THE HABENULA AND ITS CONNECTIONS 259

p. 212). In Necturus the arrangement of these tracts is somewhat different ('336, pp. 210-13).


5. Tractus cortico-habenularis lateralis. These unmyelinated fibers, as just described, pass from the piriform and dorsal pallial areas in company with those of no. 4 (fig. 20) to the habenula, where the majority of them decussate in the com. superior telencephali. Although there is no differentiated cortex in the amphibian brain, this tract is so evidently homologous with the one so named in mammals that this name is preferable to Ariens Kappers' term, "tr. olfacto-habenularis lateralis," because the latter term is commonly applied to a different tract, no. 2 of the present list. In my paper of 1910 (p. 428) the tr. olfacto-habenularis anterior was not recognized as a separate entity but was regarded as a forward extension of the tr. olfacto-habenularis from the preoptic nucleus. This usage has some justification in cyclostomes and other most primitive vertebrates and is still employed by some authors, but the distinction between the fibers which arise from the preoptic nucleus and those from the hemisphere is significant and should be recognized in the nomenclature. In most species tr. cortico-habenularis lateralis is well defined. To avoid confusion, the terms "tr. olfacto-habenularis lateralis et medialis" should be restricted to fibers which enter the stria medullaris between the anterior commissure and the optic chiasma, arising either in the preoptic nucleus or as collaterals from the medial forebrain bundle.


6. Tractus amygdalo-habenularis. These are thick fibers, some of which are myelinated, passing from the gray of the amygdala to the habenula. Some of them may connect farther forward with the primordial corpus striatum — tr. strio-habenularis. These fibers ascend in the stria between nos. 5 and 8, and many of them end in the habenular neuropil (figs. 31-34, 74-78; '27, p. 302 and figs. 13, 14). As they approach the habenular commissure, these thick fibers mingle with those of the medial cortico-habenular tract, and these components can be separated only in electively impregnated specimens. In one such preparation of adult Amblystoma (no. 2257) the sections are inclined about 45° to the sagittal plane and the tr. amygdalo-habenularis is heavily and electively impregnated on both sides. The only other component of the stria stained is a small number of fibers of tr. cortico-habenularis medialis on one side only. The large tract from the amygdala is clearly followed through the commissure to the amygdala of the opposite side.


7. Tractus septo-habenularis. These fibers arise from the septal area and the bed-nuclei of the anterior commissure and ascend to the stria medullaris between this commissure and the interventricular foramen (fig. 75) . They spread in the ventral habenular neuropil and have not been followed farther. In some preparations there is evidence that these fibers are accompanied by others from the anteroventral part of the corpus striatum complex (primordial nucleus caudatus) — a strio-habenular connection.


8. Tractus cortico-habenularis medialis (figs. 32, 33, 34, 71, 76; '27, figs. 11-17, here marked, str.med.). This is a large component of unmyelinated fibers and a few thick fibers with myelin sheaths. They arise from all parts of the primordium hippocampi, leaving its posteroventral border in company with similar fibers of the hippocampal commissure and tr. cortico-thalamicus medialis. Under the stemhemisphere fissure they turn sharply dorsad, to ascend as the most anterior component of the stria medullaris. They spread widely in the habenular neuropil, and some of them decussate at the anterior end of the habenular commissure, thus forming the com. pallii posterior.


9. Tractus cortico-thalamicus medialis (figs. 31, 32, 72, 75). These fibers are regarded as primordia of the columna fornicis. They accompany the tr. cortico-habenularis medialis, and some of their fibers are collaterals from that tract and the hippocampal commissure. They cross the stria medullaris obliquely and are not integral parts of it except for the collateral connection mentioned.


10. Tractus olfacto-thalamicus. This name was given to a collateral connection from the stria medullaris to the neuropil of Bellonci in Necturus ('336, p. 205). It is present also in Amblystoma.


11. Tractus thalamo-habenularis. These fibers arising from the dorsal thalamus and regions posteriorly of it enter the stria medullaris within the ventral habenular nucleus (fig. 76).. They are accompanied by fibers passing in the reverse direction, the tr. habenulothalamicus ('396, p. 539; '42, p. 261 and fig. 77). This connection has been described in mammals by several authors, recently by Marburg ('44, p. 220) in man, where its fibers arise from the anterior nucleus and pulvinar.


The tr. strio-habenularis is probably present in Amblystoma but has not been clearly seen. If so, its fibers may accompany those of the tr. amygdalo-habenularis and tr. septo-habenularis, as mentioned above. The former of these probable striatal connections is compa


THE HABENULA AND ITS CONNECTIONS 261

rable with the pallido-habenular connection described in man by Marburg ('44).


Number 12 of figure 20 is an afferent tract to the habenula from the tectum and nucleus pretectaHs, which has no connection with the stria medullaris — tr. tecto-habenularis. These fibers are accompanied by others which pass in the reverse direction — tr. habenulo-tectaUs.


FASCICULUS RETROFLEXUS

Efferent fibers leave the habenula, so far as is known, in three tracts: (1) from the dorsal nucleus to the tectum by the tr. habenulotectalis, (2) from the ventral nucleus to the dorsal thalamus by the tr. habenulo-thalamicus accompanying the tr. thalamo-habenularis, and (3) between these from both nuclei in the much larger f. retroflexus of Meynert. There may be other efferent fibers, e.g., to the cerebral hemispheres accompanying those of com. superior telencephah and com. pallii posterior, but these have not been observed. One habenula of the newborn rabbit was destroyed by von Gudden ('81), and subsequently Meynert's bundle of the same side was found to be atrophied.


The chief component of the f. retroflexus is the habenulo-interpeduncular tract, the fibers of which arise from both dorsal and ventral habenular nuclei (figs. 20, 71, 73, 77, 103). This tract is composed chiefly of thin unmyelinated fibers, which form the central core of Meynert's bundle. Surrounding this core, other thicker fibers are loosely arranged, some of these being myelinated. Weigert sections show that a few myelinated fibers arise from both dorsal and ventral habenular nuclei and that these are joined by a few others from the dorsal thalamus, pretectal nucleus, and eminence of the posterior commissure. This fasciculus is also accompanied for part of its course by the few myelinated fibers of the parietal nerve (p. 235). No fibers have been seen to enter it from the tectum. The myelinated fibers leave the fasciculus and scatter in the alba of the nucleus of the tuberculum posterius; none of them enter the interpeduncular nucleus. In addition to these myelinated and other thick axons which terminate in the alba of the peduncle, there are many fine fibers from the axial core which take similar courses ; some of these are probably of habenular origin.


The fasciculus descends across the thalamus at the outer border of the gray and is partly imbedded within it. Its course as seen in horizontal sections is shown in figures 30-35 (for the course in transverse


^6» THE liRAIX OF THE TIGER SALAMANDER

sections see '"io, ^gs. -2-7. and 'il , figs. '■28-oo, tr.hab.pcd.). As it enters the peduncle, it turns outward, passing ventrolaterally thriiugh tlie alba to reach the surface near the superficial origin of the III nerve root. Passing ventrally of these emerging root fibers, it turns medially at the fovea isthmi, immediately spinahvard of which it enters its decussation in the ventral commissure (figs. 50, 51, 53). The spiral endings of these fibers are described in chapter xiv. In the alba of the peduncle it passes internally of the area ventrolateralis pedunculi and well separated from this superficial neuropil {fig. 94). Here many thin axons separate from the fasciculus to enter the superficial neuropil. A little farther spinalward. at the level of the III nerve root, a larger number of fine fibers leave the fasciculus and turn forwartl into the posterior end of the neuropil, as shown in a favorable (lolgi section ('4^, fig. 40, fibers from, f.retr. to a.I.t.).


This is the usual arrangement. That the connection of the f. retroflexus with the area ventrolateralis pedunculi is important physiologically is evident from atypical courses of the fasciculus which have been observed in a number of specimens, always on one side only. One such anomaly is well illustrated in figures 56. 57. and 58. In this specimen (horizontal Golgi sections) the right f. retroflexus is unstained but can be seen to take the usual course; the left is abundantly impregnated from the habenula to its decussation. As it enters the alba of the peduncle it divides into two bundles. The medial bundle, containing about one-third of the fibers, takes the typical course: the larger lateral bundle passes along the inner border of the area ventrolateralis pedunculi with numberless terminals or collaterals entering this neuropil and spreading in the surrounding alba. Both bundles pass spinalward under the emerging III root fibers and converge to the decussation below the fovea isthmi. Many of these fibers spread in the alba rostrally and caudally of the fovea without decussation.


Another adult specimen (^no. "2'-217) from the same. lot and similarly prepared is like that just described except that the lateral bundle is smaller than the medial. This smaller bundle enters the area ventrolateralis and here breaks up into several slender fasciculi. Posteriorly of this area of neuropil, these fasciculi, somewhat diminished in volume, i-ejoin the medial bundle at the level of the III nerve root and both bundles then enter the decussation. This arrangement is found. on one side only, in the series of sagittal Golgi sections from which figures 7'2 and 73 were drawn. A somewhat different atypical feature


THE HABENULA AND ITS COXNECTIOXS 263

is seen in the series of sagittal Cajal sections from which figures 74-78 were drawn. The fasciculus is well impregnated on both sides, and on one side it takes the typical course. At the decussation the thinnest fibers cross in the usual way, and some fibers of thicker caliber descend uncrossed at the lateral border of the interpeduncular neuropil. Whether they ultimately decussate and enter the interpeduncular spiral is unknown, because the impregnation fails below the decussation. On the opposite side the f . retroflexus divides as it enters the peduncle. The larger medial bundle takes the txT^ical course. A small compact fascicle separates from it and passes more laterally to reach the area ventrolateralis pedunculi, where its fibers spread in the neuropil and disappear. Modifications of the arrangements just described have been seen in several other series of Golgi sections of late larv^ae and adults. The lateral bundle may be larger or smaller than the medial bundle, and in the latter instances it may end in the peduncle or rejoin the medial bundle at the decussation.


From these observations it is concluded that in Amblystoma the f. retroflexus is a mixture of fibers of different physiological characteristics. The main axial bundle of fine unmyelinated fibers is the habenulo-interpeduncular tract. Most of these fibers decussate below the fovea isthmi and enter a peculiar elongated spiral in the interpeduncular neuropil within which they end. A smaller number, including some thicker axons, descend uncrossed along the lateral border of the interpeduncular neuropil. These endings are in isthmic territory. Rostrally of the fovea many thin unmyelinated fibers, evidently of habenular origin, end uncrossed in the posteroventral border of the cerebral peduncle near the level of the nucleus of the III nerve. This area may be regarded as a mesencephalic sector of the interpeduncular nucleus, though there is no cellular differentiation here, like that of this nucleus below the fovea. Other similar fibers, doubtless also from the habenula, leave the f. retroflexus farther forward and terminate in the superficial neuropil termed "area ventrolateralis pedunculi." In some specimens these are very numerous. The few myelinated fibers of f . retroflexus derived from the habenula end in the cerebral peduncle. This fasciculus receives accessions of fibers, a few of which are myelinated, from the dorsal thalamus, pretectal nucleus, and eminence of the posterior commissure. All these fibers probably end in the cerebral peduncle; there is no evidence that any of them reach the interpeduncular nucleus. In a series of sagittal Golgi sections (no. 2215), in which the tr. habenulointerpeduncularis is impregnated, a few thick axons are seen to arise in the gray of the dorsal thalamus and join the f . retroflexus, which they accompany almost to the ventral surface. Here some of them end by wide arborization in the area ventrolateralis pedunculi. This area of neuropil, accordingly, receives fibers by way of the f . retroflexus from both the habenula and the dorsal thalamus.


The connections just described are a simplified version of the complicated structure of the f. retroflexus of Ceratodus described by Holmgren and van der Horst ('25, p. 105). The fact that the atypical arrangements occur on only one side in Amblystoma probably is explained by the asymmetry of the habenular system in Ceratodus and other primitive species.



CHAPTER XIX THE CEREBRAL HEMISPHERES

SUBDIVISIONS OF THE HEMISPHERE

LITTLE need be added here to the general descriptions given in ^ chapters iv and vii. For details the reader is referred to the paper of 1927. That description was based mainly on a survey of a small number of well-preserved specimens cut in the transverse plane. There are in our collection many more instructive Golgi sections cut in longitudinal planes which have not been critically studied, though preliminary surveys have been made. It is deemed unprofitable at this time to continue the study of these sections because their interpretation should be based on physiological experiments correlated with the anatomical analysis.


At the present time our knowledge of the details of the internal structure of the cerebral hemispheres of Necturus ('336) is more complete than of any other amphibian. This brain is not only larger than most of the others, but it is less compact and its great elongation is favorable for accurate localization of experimental studies by a wide variety of methods. This generalized brain provides a norm or basic pattern for the vertebrate phylum as a whole. The other urodeles and the anurans present a series of progressively more differentiated brains, and the sequence of stages of this process of specialization can readily be followed. That such a program of correlated histological and experimental work is practicable was demonstrated by Coghill in a restricted field of embryological research. With the refined experimental methods now at our disposal and with some reorientation in the fields of developmental mechanics, localized experimental extiipations, and electrical excitations, supplemented by oscillographic records, the steps in progressive phylogenetic differentiation of structure can be correlated with changes in overt behavior. For the completion of such a program frogs will probably prove to be more serviceable animals than the more sluggish salamanders (p. 98). These data will enable the comparative psychologists to identify and interpret prodromal stages of some of the basic patterns of human mentation.


Comparison of the amphibian cerebral hemisphere with the human shows a common plan of organization, and in the amphibian brains we find evidence of the beginning of differentiation of some mammalian structures at the earliest stages of their emergence from an undifferentiated matrix. The formative agencies which are operating to produce this local specialization are open to inspection and experimental investigation.


On the basis of position, internal structure, and connections the following mammalian organs have been identified in the brain of Amblystoma. First, the pallial field is distinguishable from the stem, and within this field primordia of hippocampal and piriform cortical areas are unmistakable. Some connections are suggestive of influences which may be precursors of neopallial differentiation, but these are vague and uncertain. Most of the mid-dorsal pallial area is probably represented in higher brains at the margins of archipallial and paleopallial cortex — such transitional cortex as the subicular and perirhinal areas.


In the subpallial part the lateral and medial walls of the hemisphere are organized essentially as in mammals. Laterally, the strioamygdaloid complex is well defined, though its subdivisions are not clearly separable. Of these, the amygdala is definitely organized, with connections very similar to those of mammals. In the corpus striatum the nucleus accumbens septi is present as in lower mammals, and associated with it is an area which probably corresponds with the head of the caudate nucleus. The remainder of the corpus striatum is an undifferentiated lentiform nucleus, within which large and small cells are mingled. The connections of these cells suggest that the dorsal part of this area becomes the putamen and the ventral part the globus pallidus (p. 96).


On the medial side of the hemisphere the structure and connections of the septum conform with the mammalian arrangement, and below this is an undifferentiated area which gives rise in some of the fishes and in mammals to the tuberculum olfactorium.


THE OLFACTORY SYSTEM

As outlined in chapter vii, the olfactory nerve and its connections have played the dominant role in the morphogenesis of the cerebral hemispheres of lower vertebrates. The brief summary of the structure at the end of chapter iv is here supplemented by further description of the distribution of the olfactory tracts.



THE CEREBRAL HEMISPHERES 207

Nervns terminalis. — These unmyelinated fibers enter the brain in small compact fascicles mingled with those of the olfactory nerve. Their peripheral and central courses can be accurately followed only in elective Golgi impregnations, which, fortunately, are frequently obtained. All fibers of the olfactory nerve end in the olfactory bulb, but none of the terminalis fibers do so. The latter enter the brain at the ventral border of the olfactory bulb and course backward in several small fascicles, which terminate in the septum, preoptic nucleus, and hypothalamus. In Necturus some of them reach the interpeduncular nucleus, and this may be true in Amblystoma also. This nerve is present in vertebrates generally, from fishes to man, but our knowledge is incomplete about its terminal connections and functions (McKibben, '11; for Necturus see my '336, p. 120, and '346, '34c; for the frog, '09). It is regarded here as a sensory nerve, but even this is a debatable question.


Olfactory bulb. — In the olfactory bulb, as in the retina, the peripheral receptors discharge into a field which receives few afferent fibers of other functional systems. In the other primary sensory centers there is a common pool of neuropil within which terminals of peripheral fibers of diverse sensory modality are mingled, and to these there are added terminals of other correlating fibers of central origin. The bulbar formation of Amblystoma receives an enormous number of fibers of the olfactory nerve and no others from the periphery. There are also terminals of fibers ascending from other parts of the brain:

(1) many of these are collaterals from the secondary olfactory tracts;

(2) some are axons of cells of the anterior olfactory nucleus; (3) some may be commissural fibers by way of the anterior commissure; (4) some may come from more remote parts of the hemisphere. By far the larger number of these ascending fibers belong to the first two classes, in which olfactory influence is clearly dominant. From this it follows that the impulses conducted by the secondary olfactory tracts are influenced relatively little by other functional systems. In this respect they differ from the lemniscus systems of the lower brain stem; and the fact that these almost purely olfactory tracts reach all parts of the cerebral hemisphere is probably the reason why this hemisphere remains at a low level of structural differentiation and physiological specificity.


Necturus and Amblystoma exhibit two well-defined stages in the histological differentiation of the olfactory bulb, but the mammalian type of structure has not been attained ('246, '31). Throughout the bulbar formation, except for the accessory bulb, the structure is nearly homogeneous, with little evidence of localization of function. The sense of smell lacks any provision for localizing in external space the source of odorous excitations. In the retina there is very complicated mechanism for analysis of the components of visual excitation (Polyak, '41). The analysis of olfactory sensibility for discrimination of odors is evidently a much simpler process. Judging by analogy with Polyak's description of the retina, there is little provision for this in the olfactory bulb. It is possible that the periglomerular cells may perform this function, but the structural organization of the bulbar formation gives clear evidence that the dominant activity here is not analysis but summation and intensification. The correlation of olfaction with other sensory systems is effected throughout the cerebral hemisphere, hypothalamus, and epithalamus, beginning in the nucleus olfactorius anterior.


Anterior olfactory nucleus. — This nucleus was first defined in Ambly stoma ('10, p. 497) as undifferentiated olfactory tissue of the second order, closely associated with the olfactory bulb and extending backward a longer or shorter distance between the bulbar formation and the more specialized parts of the hemisphere. It is of large extent in the amphibian brain, and to it considerable attention has been given ('246; '31; '27, p. 288; '336, p. 133; '34, p. 99). In higher brains it shrinks in size as progressively more of this generalized tissue is specialized. Its comparative anatomy was discussed in connection with a detailed description of it in the Virginia opossum ('24c?). In the amphibian brain it is a broad ring of gray bordering the bulbar formation on all sides. This cylinder is divided topographically into ventral, medial, dorsal, and lateral sectors, each of which has its own distinctive connections with other parts of the hemisphere. The ventral sector and the lower part of the medial contain the primordium of the tuberculum olfactorium. The arrangement of these sectors in Necturus is shown in figures 111 and 112, and their structure and connections have been described in detail ('336, p. 133). Transverse sections through this region of Ambly stoma are in the paper of 1927 (p. 288 and figures 2-5). The neurons and neuropil of this nucleus are illustrated in figures 105, 108, and 109 and in 1934, figures 1 and 2.


Typical neurons of the anterior nucleus have widely spread thorny dendrites, and axons which enter the olfactory tracts. Many other forms of cells are seen, some of which are transitional to those of the olfactory bulb. The axons of some of its cells are directed peripheral


THE CEREBRAL HEMISPHERES 3G9

Iv, to end with wide arborizations in the granular layer of the bulb. Most of the smaller cells have short, much branched axons, which participate in the formation of the dense axonic neuropil of this region.


Olfactory tracts. — Strictly defined, a tractus olfactorius includes axons of olfactory neurons of the second order only, that is, axons of mitral cells ; but practically all these fibers are mingled with those of higher order from the anterior olfactory nucleus and other parts of the hemisphere, so that the tracts so designated on the figures are all mixed })undles. These axons of mitral cells stream backward from all margins of the olfactory bulb. Only the shorter fibers to the anterior nucleus are drawn in figures 111 and 112. As shown in figure 6, these are accompanied by longer fibers, which join the tracts descending from the anterior nucleus. Olfactory tracts from the lateral and ventral borders of the bulb take direct courses backward in three series. The more dorsal fibers enter tr. olfactorius dorsolateralis for distribution to the dorsolateral olfactory nucleus, which is primordium piriforme. This is the largest of the olfactory tracts and is comparable with the lateral olfactory stria of mammals. Other lateral fibers pass to the corpus striatum and amygdala. Some of these fibers join tr. olfacto-peduncularis, most of the fibers of which arise in the anterior nucleus and primordial caudate nucleus. Fibers from the ventral border of the bulb enter tr. olfactorius ventralis and descend for an undetermined distance in the medial forebrain bundle.


As shown by figure 4, the lateral ventricle extends forward almost to the anterior end of the olfactory bulb. Many of the longer fibers from the bulb take tortuous courses to reach their terminal stations. They accumulate in the medial sector of the anterior olfactory nucleus and primordium hippocampi, where they form a very large compact sheet of fibers termed "fasciculus postolfactorius" (fig. 100, f.po.; '27, figs. 2, 3). These fibers run vertically around the tip of the lateral ventricle, some directed ventrally to enter tr. olfactorius ventralis and some dorsally to enter tr. olfactorius dorsolateralis (fig. 5).


Olfactory tracts of the third and higher orders, i.e., those separated by two or more synapses from the periphery, are generally designated by hyphenated compound words, as tr. olfacto-peduncularis; but, as mentioned above, many of these tracts are mixtures containing some axons of mitral cells. For further details of these connections see the summaries ('33&, p. 124; '27, p. 282).



CHAPTER XX THE SYSTEMS OF FIBERS

THE principles of classification and nomenclature of nerve fibers here employed are explained on page 9. A systematic account of them is very difficult because of their dispersed arrangement, their many deviations from the familiar mammalian pattern, and a cumbersome and confused nomenclature; yet gratifying success has attended efforts to resolve the amphibian tissue and discover mammalian homologies or their primordia. It is not our purpose in this chapter to give a comprehensive list of the tracts of Amblystoma that have been identified and named. References are given to the literature in which incomplete lists have been published and to pages of this book where particular systems of fibers are described. To these there are added descriptions of a few other systems of fibers of special importance and complexity.


Lists of tracts of Necturus have been published for the medulla oblongata ('14rt, '30), midbrain and thalamus ('17), and forebrain ('336). At the end of the latter paper is an alphabetical list of abbreviations which includes many tracts, with references to previous descriptions and synonyms. For Amblystoma a similar list of abbreviations ('36, pp. 309-12) includes many tracts of the cerebral peduncle, with page references to the text. Other lists of tracts of Amblystoma have been published ('27, '36, '39, '396, '42), and Bindewald ('14) gave in tabular form a useful summary of previous descriptions of the parts of the amphibian forebrain and related fiber tracts, with homologies. The tracts of several species of urodeles have been described in the literature cited on page 11. The following systems of fibers of Amblystoma have been more or less completely described: peripheral nerves (Coghill, '02), olfactory tracts ('27, p. 282; Bindewald, '14; compare Necturus, '336, p. 124, '346), optic tracts and tectal connections ('25, '41, '42; compare Necturus, '41a), postoptic (supra-optic) commissures ('42, p. 219; Necturus, '41a, p. 513), the visceral-gustatory system ('44a, '446), cerebellum (Larsell, '20, '32), tegmental fascicles ('36).


In this work under appropriate headings there are lists of tracts

270


THE SYSTEMS OF FIBERS 271

related with the several parts of the brain, accompanied by diagrams, and some classified lists, including the lemniscus and visceralgustatory systems (chap, xi); connections of the interpeduncular nucleus (chap, xiv) ; some tectal connections (chap, xvi) ; stria medullaris thalami, fasciculus retroflexus, stria terminalis, and fornix (chap, xviii); some olfactory connections (chap, xix); and in the next following sections the basal forebrain bundles, tegmental fascicles, f. tegmentalis profundus, and the commissural systems.


THE BASAL FOREBRAIN BUNDLES

In a survey of the forebrains of fishes, attention was called to the important part played in morphogenesis by these great systems of longitudinal conduction ('22a, p. 175), and here this theme has received further consideration in chapter vii. These bundles of Necturus were described in comparison with those of other amphibians ('336, p. 166), and their arrangement in Amblystoma has been illustrated ('27, p. 285; '36, p. 335 and figs. 5, 6; '396, p. 533 and fig. 1). In sections of these brains the basal bundles are the most obvious landmarks, extending from the hemispheres backward through the brain into the medulla oblongata. Some of their fibers, in other species, have been described as extending into the ventral funiculi of the spinal cord, but our material yields no evidence that in Amblystoma any fibers which descend from the hemispheres in these bundles go back without interruption farther than the level of the roots of the VIII nerve. These bundles contain ascending and descending fibers, many of the latter decussating in the anterior commissure. In addition to these main lines of through descending traffic, many other kinds of fibers enter these bundles, and these connections will now be summarized.


The fibers of these bundles are arranged in three groups of fascicles as indicated in figures 19, 20, 21, 101; but these groups are not sharply separated, for there is much interchange of fibers among them. Their descending fibers are roughly comparable with the subcortical components of the mammalian extra-pyramidal systems, and the more dorsal ascending fibers of the thalamo-frontal tract (figs. 19, 101, tr.th.f.) correspond with the thalamo-striatal system. The dorsal group of fascicles, the lateral forebrain bundle (i. lateralis telencepha\\,f.lat.t.), connects with the lateral wall of the hemisphere and has dorsal and ventral components with different connections. The ventral group, the medial forebrain bundle (f. medialis telencephali, f.med.t.), connects with the medial and ventral walls and also has dorsal and ventral components related, respectively, with the dorsal and ventral parts of the hypothalamus. The third group of fascicles comprises the tractus olfacto-peduncularis (tr.ol.ped.), which lies between the lateral and the medial bundles, connecting the anterior olfactory nucleus and the head of the caudate with the hypothalamus, peduncle, and interpeduncular nucleus.


Lateral forebrain bundle. — Most of these fibers are connected with the primordial corpus striatum and amygdala, and some of them with the piriform area. The descending fibers correspond with the human ansa lenticularis, and most of them are well myelinated. They appear earlier in ontogeny than do the ascending fibers. The dorsal and ventral fascicles of this bundle interchange fibers and in Necturus are not clearly separable ('336, p. 170). They are connected mainly, though not exclusively, with the dorsal and ventral nuclei, respectively, of the corpus striatum, as described in chapter vii. When first studied, this relationship could not be demonstrated ('27, p. 286), but subsequent examination of sagittal sections ('36, p. 335) convinces me that this incomplete separation marks the beginning of differentiation of the globus pallidus. In 1927 the ventral fascicles were named tr. strio-peduncularis ('27, p. 287) and the dorsal fascicles tr. strio-tegmentalis, but these terms are inappropriate because both of them terminate in the peduncle and also in the tegmentum, though in different areas and evidently with different physiological import. The dorsal fascicles enter group (9) of the tegmental fascicles and the ventral fascicles enter group (10), as described below.


The descending fibers of the ventral fascicles {f.lat.t.v. and (10) of the figures) arise from the large cells distributed throughout the striatal gray. Their terminals are widely spread in the ventral parts of the peduncle and isthmic tegmentum. They d6 not extend so far spinal ward as do some fibers of the dorsal fascicles. In the analysis of the tegmental fascicles ('36) these fibers comprise group (10). They make their chief synaptic connections with the large cells of the peduncle and tegmentum, the axons of which enter the ventral tegmental fascicles of groups (4), (5), and (6), descending to the medulla oblongata, and some of which continue in the f . longitudinalis medialis into the spinal cord. This connection is regarded as provision for cerebral control of mass movements of the trunk, limbs, and eyeballs.


The descending fibers of the dorsal fascicles (f.lat.t.d. and tegmen


THE SYSTEMS OF FIBERS 273

tal fascicles of group (9)) extend farther spinalward with different distribution. They give numberless collaterals to the thalamus, the dorsal part of the peduncle, and the dorsal and isthmic tegmentum; and posteriorly they turn laterally into the trigeminal tegmentum dorsally of its motor zone (figs. 29, 30). The chief synaptic connections of these fibers are not with the primary motor column but with areas of intermediate-zone type. They appear later in embryonic development than do the ventral fascicles, and their functions are believed to be cerebral control and conditioning of bulbar reflexes.


Another important descending component of these dorsal fascicles is the recently described tr. strio-tectalis (figs. 11, 101, tr.st.tec; '42, p. 262) and the associated tr. strio-pretectalis (figs. 14, 101, tr.st.pt.). These fibers probably arise from the corpus striatum, though this has not been demonstrated. They end by wide arborizations in the optic tectum, pretectal nucleus, and geniculate neuropil of the thalamus.


Some fibers descend from the dorsal striatal nucleus in company with more from the amygdala in that component of the stria terminalis complex known as the dorsal olfactory projection tract (ol.p.tr.), as described on page 242.


The most noteworthy ascending system of fibers of the lateral bundle is tr. thalamo-frontalis {tr.th.f.). These slender unmyelinated axons arise from cells of the middle sector of the dorsal thalamus and descend in several small compact strands to the dorsal fascicles of the lateral forebrain bundle, within which they turn forward. As in Necturus ('336, p. 170), they probably all end in the striatal neuropil, though their terminals in Ambly stoma have not been described. These are precursors of the ascending thalamic radiations of mammals. They arise from the undifferentiated nucleus sensitivus of the thalamus, and no evidence has been seen of any separation among them of projection tracts related with different functional systems. The course of these fibers, so far as known, has been fully illustrated (figs. 3, 15, 30-34, 75, 95, 101, 102, 103; '396, figs. 1, 7, 8, 13-17).


Medial forebrain bundle. — Most of the fibers of this large system are unmyelinated, passing in both directions between the olfactory bulb and the medioventral parts of the olfactory area of the hemisphere and the preoptic nucleus and hypothalamus. Many of these fibers decussate in the ventral part of the anterior commissure ridge. Analysis of this complex is possible only with the aid of elective Golgi impregnations, and there are few of these in our collection of Amblystoma sections. The Necturus material has been more instruc


274 THE BRAIN OF THE TIGER SALAM.\NDER

live, and the following summary is based on data from both species (Amblystoma, figs. 6, ^25, 71, 75; '^27, p. 285; '396, p. 534 and figs. 1, 79; Necturus, '336, pp. 173, 261 and fig. 14; '346). The medial bundle is incompletely separable into dorsal and ventral fascicles connected, respectively, with the dorsal and ventral parts of the hypothalamus. Both fascicles have extensive connections with the preoptic nucleus, and both contain descending and ascending fibers. Posteriorly of the anterior commissure ridge, thick collaterals of fibers of both groups of fascicles ascend to the habenula in tr. olfacto-habenularis of the stria medullaris (p. 257; '396, fig. 79).


The dorsal fascicles contain descending fibers arising in the septum (tr. septo-hypothalamicus) and in the primordium hippocampi (precommissural fornix, p. 254), also some components of the stria terminalis system (p. 256). These are accompanied by some ascending fibers the connections of which are not clear.


The ventral fascicles contain secondary olfactory fibers from the bulb and some of higher order from the ventral and medial sectors of the anterior olfactory nucleus. These fibers are distributed to the preoptic nucleus and ventral part of the hypothalamus and are accompanied by many preoptico-hypothalamic fibers. Slender filaments of the nervus terminalis are spread among these fascicles for their entire length. The large hypophysial tract arises from neurons which are widely scattered throughout the hypothalamic field reached by both the ventral and the dorsal fascicles of the ventral bundle. In the floor of the preoptic recess there is a median fascicle of unmyelinated fibers, among which are a few with myelin sheaths. This is tr. preopticus (p. 244).


Very slender unmyelinated fibers arise from small cells at the extreme posterior end of the ventral hypothalamus and ascend for undetermined distances in the ventral fascicles. There are doubtless other ascending fibers, but their courses have not been recorded.


Olfacto-peduncular tract. — This is a well-defined round bundle, lying between the lateral and the medial bundles and less myelinated than the former (figs. 18, 21, 25-30, 53, 54, 59, 72, 95-99, 101; '27, p. 286; '36, p. 336; '396, p. 534, figs. 1, 79, 80). Its fibers arise from the head of the caudate nucleus and neighboring parts and distribute to the dorsal part of the hypothalamus, ventral part of the peduncle, and interpeduncular nucleus. The hypothalamic connection allies this tract with the dorsal fascicles of the medial forebrain bundle, the peduncular connection with the ventral fascicles of the


THE SYSTEMS OF FIBERS 275

lateral forebrain bundle. The connection with the interpeduncular nucleus is extensive and physiologically distinctive. This tract is the most direct connection between the olfactory area and the primary motor field of the peduncle and the interpeduncular nucleus. It is an internuncial of intermediate-zone type, with the olfactory influence predominant.


THE TEGMENTAL FASCICLES

The strio-thalamic and strio-peduncular components of the lateral forebrain bundle activate large neurons of the ventral thalamus and peduncle, and the axons of these cells transmit impulses downward to the tegmentum from the isthmus to the spinal cord. These descending fibers form the most conspicuous components of a series of ventral tegmental fascicles, some of which at lower levels are assembled in the f. longitudinalis medialis (figs. 6, 18, 91). Ventrally and dorsally of these are many other tegmental fascicles, composed chiefly of descending fibers; and still farther dorsally are the ascending lemniscus systems (fig. 14). The analysis of the composition of these longitudinal fascicles has been very diflficult, yet this knowledge is essential for an understanding of the brain stem.


The topographical analysis of the brain stem of 1935 was followed in 1936 by a reconnaissance survey of the related fibers in the peduncle and tegmentum. The more obvious and constant bundles of longitudinal fibers were enumerated as ten fascicles or groups of fascicles ('36, p. 303), and the specific tracts represented in these and some other bundles, so far as then known, were Hsted ('36, pp. 33446). These fascicles are divided into a ventral and a dorsal series, the former including fascicles (1) to (6) and part of (10); the latter, (7) to (10). Figures 91, 92, 94, and 102-4 are here reproduced from the paper of 1936; compare figures 30-33 and 101.


This classification was arbitrary for descriptive purposes only and included only the most clearly defined and constant fascicles bordering the gray substance. The symbol for each group is an Arabic number inclosed in parenthesis. Externally of these ten groups, in the intermediate alba, there are other less well-defined fascicles with a larger proportion of unmyelinated fibers. These are imbedded in much neuropil and are more variable than the deeper fascicles. Superficially in the subpial neuropil there is another series of fascicles, chiefly of unmyelinated fibers, some of which form recognizable tracts. Dorsally of all these fascicles are the lemniscus systems; these and the ascending secondary visceral-gustatory tract are described elsewhere in this work. Further study has yielded additional details about the composition of the ten groups of tegmental fascicles. These and some other fascicles and tracts of this region are here analyzed as far as their composition is now known. Most of these fibers are descending. There are ascending fibers also, but our material has yielded little information about them.


During the preliminary study it was anticipated that each group of tegmental fascicles would prove to be composed chiefly or wholly of fibers of a single tract or related group of tracts, as are the lemniscus systems. This proves not to be the case, for most of these bundles are mixtures of fibers of diverse sorts from unexpectedly widely separated sources. The reasons for their fasciculation in the pattern observed are not clear. Some specific tracts, like the mamillointerpeduncular and strio-tegmental, are fairly clearly segregated (in groups (2) and (9) in the cases mentioned), but most of the bundles are heterogeneous mixtures. The pattern of fasciculation seems to be determined more by the ultimate destination of the fibers than by their nuclei of origin. In Weigert and especially in reduced silver preparations these fascicles, particularly those bordering the gray, are clearly defined for long distances; but Golgi sections show that there is much anastomosis among them and that there are numberless unmyelinated fibers which are not fasciculated but spread diffusely in the alba.


As the lateral forebrain bundles recurve dorsally at the anterodorsal border of the peduncle (figs. 6, 16, 101, 102; '36, figs. 5, 6), sagittal Weigert sections show that the compact bundles of myelinated fibers disintegrate, with diffuse spread of the fibers in the alba of the posterior part of the thalamus, dorsal tegmentum, and peduncle. Some of these fibers are reassembled farther spinal ward in fascicles (9) and (10). Golgi sections show a similar dispersal of the thinner unmyelinated fibers. Reduced silver preparations, however, reveal many slender fascicles of unmyelinated fibers which traverse this region without loss of their individuality. It is evident that these bundles, like most of the other tegmental fascicles, are mixtures of fibers of diverse distribution and physiological significance. In transverse Golgi sections of adult brains, in which the myelinated fibers of the deeper fascicles are slightly darkened and the unmyelinated fibers are electively impregnated, it is seen that the proportion of unmyelinated fibers is greater in the dorsal fascicles than in the ven


THE SYSTEMS OF FIBERS 277

tral and that almost all fibers bordering the gray in the ventral fascicles are myelinated.


Ventromedian fascicles (1). — These are limited to the midbrain and isthmus, extending spinalward from the commissure of the tuberculum posterius. Most of their fibers decussate obliquely in the ventral commissure. They comprise 5 to 12 well-defined bundles of rather thick myelinated fibers arranged close to the gray under the ventral angle of the ventricle, with some admixture of unmyelinated fibers. These fibers are derived from various sources, and they are variously distributed. The bundles as definite anatomical entities are assembled only in the space bounded approximately by the levels of the nuclei of the III and IV cranial nerves. This distance is greater in urodeles than in most other vertebrates, and the arrangement of these fascicles found in Amblystoma has not been described in any other species. In Necturus reduction of the optic system involves corresponding shrinkage of these median fascicles ('36, p. 348), though the general plan is similar. Three components of these fascicles have been distinguished.


a) Tractus tecto-bulbaris et spinahs cruciatus, pars anterior (fig. 12, tr.t.h.c.l.). — This is by far the largest component of the group. Its fibers, chiefly myelinated, descend from the tectum and turn spinalward in the mid-plane, here decussating in component 4 of the commissure of the tuberculum posterius, and posteriorly of this very obhquely in the ventral commissure ('36, pp. 303, 330, figs. 2, 7). At the level of the nucleus of the IV nerve they turn laterally and descend in tr. tecto-bulbaris et spinalis within the ventromedial alba of the medulla oblongata. The course of this tract, as seen in horizontal sections, is shown in figures 27-36 (for complete description see '36, p. 340, and '42, p. 268). The posterior division of this tract (figs. 27-36, tr.t.h.c.2.) does not enter the tegmental fascicles but decussates transversely, at the level of the nucleus of the IV nerve.


b) Tractus pedunculo-bulbaris ventralis cruciatus ('36, figs. 2, 7, f.m.t.{l)hr, '39&, p. 582 and figs. 23, 24).— These fibers arise from the ventral thalamus and peduncle and enter ventral fascicles (1). Here they decussate obliquely, mingled with those of the crossed tectobulbar tract; and after crossing they separate from the latter to descend in the ventrolateral fasciculi of the medulla oblongata, and some of them in the f. longitudinalis medialis, as described in the references cited.

c) Tractus pedunculo-tegmentalis cruciatus ('36, figs. 2, 7, fjn.t.{l)c.). — These fibers enter the ventral fascicles from the peduncle and perhaps also from the ventral thalamus. At the posterior end of these fascicles they do not turn laterally with the others but continue posteriorly and dorsally as one of the components of the f. tegmentalis profundus. They arborize in the deeper layers of neuropil of the isthmic tegmentum.


V entromedian fascicles {2) . — These fascicles of thin unmyelinated fibers comprise tr. mamillo-interpeduncularis, lying laterally and ventrally of those of group (1). These are shown in figure 19 and separating from tr.mam.teg. in figure 21. They assemble in the periventricular neuropil of the ventral lobe of the dorsal part of the hypothalamus. Figure 27 {tr.mam.inp.{2)) shows them converging into a ventricular protuberance of this lobe, which, immediately dorsally of this level, extends across the mid-plane to join the corresponding structure of the opposite side at the attenuated ventral border of the commissure of the tuberculum posterius (compare the median section, fig. 2C). Here some of these fibers decussate as component 1 of this commissure, as shown in figure 28. The crossed and uncrossed fibers continue dorsalward and then spinalward, recurving around the cerebral flexure at the extreme ventral surface (figs. 29, 30), to end in open arborizations within the interpeduncular neuropil. Their entire course in the larva is shown in a published diagram ('396, fig. 22; other details are also illustrated in that paper — figs. 6-9, 35, 41, 42, 57-61). These fibers comprise the whole of group (2) of the tegmental fascicles, described in 1936. They do not form a compact bundle but are rather loosely spread, and their courses can be followed only in favorable Golgi sections. Unlike the other connections of the dorsal part of the hypothalamus, these fibers are aggregated in the deep periventricular neuropil, and evidently their physiological properties are radically different, from those of the mamillo-peduncular and mamillo-tegmental tracts.


V entral fascicles {3). — The chief component is tr. mamillo-tegmentalis (fig. 21; '36, figs. 3, 8; '42, fig. 3). These thin myelinated and unmyelinated fibers pass, probably in both directions, between the dorsal hypothalamus and the tegmentum in close association with similar fibers related with the peduncle and thalamus. The mammalian equivalents of the complex are found in the mamillo-thalamic and mamillo-peduncular tracts and the mamillary peduncle.



THE SYSTEMS OF FIBERS 279

Aniblystoma has no differentiated corpus inamillare; its primordiuni is in the dorsal part of the hypothahiniiis, from which efferent fibers go out dorsalward to the peduncle, forward to the ventral thalamus, and backward to the tegmentum. They are dispersed in the alba of the hypothalamus, and as they leave it those for the thalamus and tegmentum accumulate rostrally of those for the peduncle. Some of them decussate in components 1 and 2 of the commissure of the tuberculum posterius ('36, fig. 2). Afferent fibers to the hypothalamus are known to be present in the mamillo-peduncular and mamillo-thalamic tracts, and this may be true also of the mamillo-tegmental.


Mamillo-peduncular fibers have wide distribution in the alba of the nucleus of the tuberculum posterius, including the neuropil of the area ventrolateralis pedunculi (figs. 6, 18, 27-30, tr.mam.ped.; '36, p. 338, figs. 3, 8; '39/;, p. 338, figs. 6-12, 22, 35, 42, 89; '42, fig. 39; Necturus, '336, p. 246; '346, p. 422). The accompanying pedunculomamillary fibers transmit visceral sensory, gustatory, and optic impulses received by the area ventrolateraHs pedunculi to the hypothalamus. The mamillo-thalamic tract and the accompanying thalamo-mamillary tract put the mamillary region into reciprocal relations with the anterior part of the thalamus, as has been fully illustrated ('396, figs. 22, 35; '42, fig. 39; Necturus, '336, p. 247; '346, p. 423, figs. 2, 8, 9).


The fibers of tr. mamillo-tegmentalis arise in company with those of the two preceding systems and form a loose fascicle rostrally of those of tr. mamillo-peduncularis (figs. 27-31; '36, p. 338, figs. 3, 8-21; '396, p. 552, fig. 43; '42, fig. 3). Most of these fibers enter ventral tegmental fascicles of group (3), as shown in figures 92 and 94, some of them first decussating in the two ventral components of the commissure of the tuberculum posterius. These fascicles lie ventrolaterally of those of group (4), which contains thicker and more heavily myelinated fibers from^ the ventral hypothalamus. They terminate in the alba of the isthmic tegmentum, chiefly through the f. tegmentalis profundus. Mingled with these hypothalamic fibers are some thicker well-myelinated fibers of tr. pedunculo-tegmentalis. Some of the latter, and probably some of the hypothalamic fibers also, descend for long distances in the f . longitudinalis medialis.


Ventral fascicles (4). — These are mixed bundles derived chiefly from the postoptic commissure. The largest component is tr. hypothalamo-peduncularis et tegmentalis (figs. 21, 23), which contains both descending and ascending fibers. In the first description these were designated simply as "fibers from the postoptic commissure" ('36, p. 304, figs. 3, 8, 19, po.{Jt-)); but now their connections are better known ('42, p. 226 and fig. 3), thanks to the fact that they mature very early in ontogeny and so can be seen in young stages, despite their dispersed arrangement. A few elective Golgi impregnations confirm these findings in the adult.


Fibers related with the entire ventral part of the hypothalamus converge into the commissura tuberis in the caudal part of the postoptic commissure complex. After crossing here in diffuse arrangement, they recurve around the tuberculum posterius, where most of them spread and end in the peduncle (tr. hypothalamo-peduncularis) . Others descend in ventral tegmental fascicles of group (4) as tr. hypothalamo-tegmentalis. Only the longer fibers of the latter tract are entered on the drawings of the horizontal sections (figs. 25-32, marked tr.hy.teg.{Jf.), or simply (4)). Figure 32 cuts these fibers (^) at the most dorsal level of their arched course through the peduncle; compare their projection on the sagittal plane ('36, fig. 3, po.{4-)). Other similar fibers extend dorsally from both ventral and dorsal parts of the hypothalamus to enter the midbrain without decussation in the chiasma ridge ('42, figs. 22, 23, 39). Some of these decussate in component 2 of the commissure of the tuberculum posterius, but most of them are uncrossed.


This system of fibers is evidently the main descending pathway from the ventral hypothalamus and neuropil of the chiasma ridge to the motor field of the peduncle and tegmentum. The associated fibers connected with the dorsal part of the hypothalamus probably shoiild be classed with the mamillo-peduncular system, though they are not included in the bundles so designated. This more dorsal system, and perhaps the ventral system also, contain some fibers which are afferent to the hypothalamus, though most of them evidently are efferent. Most of the longer fibers of this system which reach the tegmentum descend in bundles of group (4), and here they are joined by thalamo-tegmental and pedunculo-tegmental fibers ('42, p. 225 and fig. 4). The longest fibers enter the f. longitudinalis medialis (figs. 27, 28), and it is uncertain whether these are of hypothalamic or of peduncular origin.


Most of the hypothalamic fibers of groups (3) and (4) pass into f. tegmentalis profundus, and most of these end in the neuropil


THE SYSTEMS OF FIBERS 281

related with the central nucleus of the isthmic tegmentum. These are doubtless precursors of the hypothalamic component of the mammalian f. longitudinalis dorsalis of Schiitz.


Tegmental fascicles of groups (5) to (10) contain fibers of diverse origin, and there is much anastomosis of their finer fibers. The coarser fibers, however, are well fasciculated in an arrangement which seems to be determined primarily by the terminal distribution of the descending systems. Their analysis has been clarified by the previously published embryological studies; see particularly the general survey ('396) and for the decussating systems the description of the postoptic commissure of the adult ('42). The data upon which this summary of the composition of the several fascicles is based are to be found mainly in the two papers just cited.


Ventral fascicles (5). — This is a group of large fascicles mainly composed of fibers (chiefly myelinated) descending from the ventral thalamus and peduncle. Most of them are uncrossed, but some decussate in company with those of tr. tecto-bulbaris cruciatus in the ventromedial fascicles of group (1) ('396, p. 546 and fig. ^S,f.v.t.{5)). They arise from all parts of the ventral thalamus and peduncle, and most of them terminate in the alba of the isthmic and trigeminal tegmentum, where they are in synaptic contact with the large cells of this region. The thicker and more heavily myelinated fibers are closely fasciculated laterally of the ventral angle of the ventricle, and many of these descend for undetermined distances in the f . longitudinalis medialis. The origin of these fibers from the ventral thalamus is shown in figures 31 and 32. At those levels similar thick fibers enter fascicles (5) from large cells of the peduncle, but they are not drawn in these figures. Figures 6 and 18 show fibers entering the f. longitudinalis medialis from large cells of the dorsal and ventral parts of the peduncle. The former (shown but not named in fig. 32) are in synaptic connection with terminals of the posterior commissure and doubtless correspond with the mammalian nucleus of Darkschewitsch (p. 217). The more ventral large cells of the peduncle (fig. 31) correspond with the interstitial nucleus of Cajal. In Amblystoma, unlike Necturus and most other vertebrates, the f. longitudinalis medialis is definitely organized only spinalward from the isthmus ('36, p. 334).


The fascicles of group (5) also receive some fibers from other systems, as will appear below. For their courses as seen in horizontal sections see figures 27-32; in transverse sections, figures 91-94 and '36, figures 9-16; in sagittal sections, '36, figures 3, 18-21. The thalamoand pedunculo-tegmental fibers of this group are similar in many respects to the crossed and uncrossed fibers of tr. thalamo-tegmentalis ventralis of groups (4), (6), and (8). Those of group (5) arise chiefly from the peduncle, the others from the thalamus. In the aggregate these thick descending paths comprise the chief final common paths from the cerebrum to the peripheral neuromotor apparatus of the primary activities of the skeletal musculature, notably those of locomotion and feeding. In early larval development these long fibers from the peduncle and ventral thalamus are among the first to appear in the cerebrum. Their adult distribution in the several tegmental fascicles seems to be determined primarily not by the arrangement in space of the groups of cells from which they arise but by the lower motor fields into which they discharge their nervous impulses. Those in groups (4), (5), and (6) descend more medially and ventrally, the longest in the f. longitudinalis medialis. These longer fibers evidently activate the trunk and limbs. Collaterals of these fibers and accompanying shorter fibers end throughout the isthmic and bulbar tegmentum ('396, figs. 84, 93), thus insuring coordination of head movements with those of the trunk and limbs. These more ventromedial fibers, accordingly, comprise the final common paths of fundamental mass movements, total patterns of behavior.


The more dorsal fibers of group (8), accompanied by the striotegmental fibers of group (9), descend along the outer border of the isthmic gray (figs. 29, 30) to end far laterally in the white substance of the isthmic and trigeminal tegmentum. Here the fundamental reflexes of the musculature of the head, concerned primarily with feeding, are organized. These more dorsal fibers, crossed and uncrossed, have innumerable terminals and collaterals in the peduncle, dorsal tegmentum, and lower tegmental fields ('396, figs. 79-82).


Ventral fascicles (6). — These fascicles, like those of the fourth group, were originally designated as fibers from the postoptic commissure ('36, figs. 4, 8, 18, po.{6)). They are now known to be composed chiefly of two systems of thick myelinated and unmyelinated fibers which decussate in the postoptic commissure, derived, respectively, from the tectum and the ventral thalamus — tr. tecto-thalamicus et hypothalamicus cruciatus posterior ('42, p. 221 and fig. 5) and tr, thalamo-tegmentalis ventralis cruciatus ('42, p. 225 and fig. 4). The coarsest fibers of these fascicles, probably including some of both


THE SYSTEMS OF FIBERS 283

the systems just mentioned, traverse the whole length of the peduncle and isthmic tegmentum, and some of them enter the f . longitudinalis medialis. Figure 3^2 cuts these fascicles a few sections ventrally of their most dorsal course as they arch across the peduncle; compare the sagittal section (fig. 103).


The thickest fibers of tegmental fascicles (4), (6), and (8) decussate in the dorsal and posterodorsal part of the postoptic commissure, those of group (8) being the most dorsal fibers of this complex at their crossing in the mid-plane. Most of the latter enter group (8), but some enter group (6). These are fibers of tr. tecto-thalamicus et hypothalamicus cruciatus posterior of my former descriptions, but these longer fibers are properly called tr. tecto-peduncularis et tegmentalis cruciatus. Posteriorly and ventrally of these at the decussation are similar coarse fibers of tr. thalamo-tegmen talis ventralis cruciatus, which cross in more dispersed arrangement posteriorly and ventrally of the preceding system. These enter the three groups of fascicles, (4), (6), and (8). In the horizontal sections here illustrated the courses of these fibers are in some places not very clearly seen, for they are not separately fasciculated. Their courses are indicated on the drawings as determined (where doubt arises) by comparison with sections in other planes and with the early larvae, where their courses are clear. It should be noted that this crossed system of ventral thalamo-tegmental fibers is accompanied by similar uncrossed fibers, most of which enter fascicles of group (8). These were first described in the larva ('39&, p. 546, figs. 13-16, 81) and in the adult are seen here in figures 30-33 as fibers joining bundle (8) from the thalamus. There is also a broad uncrossed connection from both ventral and dorsal parts of the thalamus to the tegmentum by tr. thalamotegmentalis rectus (figs. 30-34, tr.th.teg.r.). These fibers pass from the thalamic gray to the pial surface of the tegmentum and end here in the superficial neuropil and obviously have physiological properties different from the deeper fibers of groups (4), (6), and (8), which make synaptic contacts with the large cells of the tegmentum in the intermediate and deep neuropil.


Dor ml fascicles (7). — Two systems of fibers have been identified in these fascicles: {a) one from the dorsal thalamus and dorsal tegmentum and (6) one from the tectum.


a) Tractus tegmento-isthmialis (fig. 21). — This tract was first described as the chief component of fascicles of group (7) under the name tr. tegmento-bulbaris ('36, p. 334, figs. 4, 6, f.d.t.{7); '39&, p. 584). That name is inappropriate for two reasons, first, because few of its fibers reach the bulb and, second, because there is a large tegmento-bulbar tract more ventrally (figs. 27, 28, 29, tr.tecj.h.). The tract here under consideration passes from the posterior part of the dorsal thalamus, eminence of the posterior commissure, and dorsal tegmentum into fascicles of group (7). Its fibers enter f. tegmentalis profundus and end in relation with the small cells of the central nucleus of the isthmic tegmentum. Its entire course is shown in figures 29-33, here marked (7). The isthmic tegmentum is regarded as the pool within which the bulbar reflexes concerned with feeding are organized, and this tract probably plays a critical role in l)ringing to this center the appropriate aft'erents from the intermediate zone. This tract carries only part of the efl^erent fibers from the dorsal tegmentum. Many of these fibers pass directly ventrally to enter the peduncle and lower tegmental levels uncrossed or with decussation in the ventral commissure. Others ascend or descend in various other tegmental fascicles.


b) Tractus tecto-bulbaris rectus. — It has recently been found that these fibers leave the tectum by various courses and that many of those from the anterior part of the tectum enter fascicles of group (7), from which they separate in the isthmus to enter the medulla oblongata in company with other tectal fibers (p. 225). These are the fibers which in 1936 were followed from fascicles (7) into the bulb.


Dorsal fascicles (8). — This conspicuous group contains thick and thin fibers, many of which are well myelinated. Those which come from the postoptic commissure ('36, pp. 304, 338, figs. 4, 8, 17, 18, ■po.{8)) comprise two quite separate systems, («) from the tectum and {})) from the ventral thalanms. The latter are joined by many uncrossed fibers of the same system.


a) Tractus tecto-thalamicus et hypothahinueus cruciatus i)osterior. — The coarsest and most heavily myelinated fibers of the complex so named in earlier papers, after decussation in the postoptic commissure, enter fascicles of group (8) and distribute to the dorsal, isthmic, and trigeminal tegmentum ('42, p. 221). Other shorter fibers enter fascicles (6) to reach the isthmus, and others end in tlie peduncle and hypothalamus. These longer and coarser fibers to the tegmentum were variously interpreted and named in my earlier papers, but good elective impregnations have now clarified their connections as tr. tecto-tegmentalis cruciatus.


b) Trjictus thalamo-tegmentalis ventralis.— As previously men


THE SYSTEMS OF FIBERS 285

tioned, this large and complicated system of crossed and uncrossed fibers was overlooked until early larval stages revealed its essential features ('396, p. 546). These thick fibers arise in all parts of the ventral thalamus. Some of them decussate in the postoptic commissure and enter fascicles (4), (6), and (8). Others descend uncro sed in all these groups and also in group (5). Golgi sections show that these tracts have numberless terminals and collaterals throughout their lengths. The thicker fibers of group (8) course more dorsally and terminate more laterally in the isthmic and trigeminal tegmentum.


The composition of group (8) as it leaves the postoptic commissure is shown in figures 25 and 26 and its further course in figures 27-33. Figures 30-33 show uncrossed fibers from the ventral thalamus joining the bundles of crossed fibers (compare '39&, fig. 2, and see the concluding comment about fascicles of group (9)).


Dorsal fascicles (.9). — These are composed chiefly of tr. strio-tegmentalis (fig. 101; '27, p. 287; '36, pp. 304, 335; '396, fig. 79), which is the dorsal component of the lateral forebrain bundle, passing from the dorsal nucleus of the primordial corpus striatum to the trigeminal tegmentum. Some of these fibers decussate in the anterior commissure, and a small number descend as far as the level of the VII nerve roots, with terminals and collaterals along the way.


These fascicles are easily followed in both horizontal and transverse sections (figs. ^S-SS,f.lat.t.d. and (9)), and especially clearly in sagittal sections (figs. 16, 19, 21, 72, 102; '36, fig. 5; '39, figs. 3, 7-12; '396, figs. 1, 79). There is interchange of fibers between these fascicles and those of groups (7a) and (8), and the thicker fibers of both groups turn laterally in the trigeminal tegmentum close to the gray and spray out dorsally of the V and VII nuclei, where they apparently activate both large and small tegmental cells. As the fascicles of group (9) reach the posterior end of the ventral thalamus, they turn sharply dorsad, parallel witlv the limiting sulcus, s, around the anterodorsal border of the nucleus of the tuberculum posterius. Here those of (9) lie ventrolaterally of (8) (figs. 16, 91-94; '36, fig. 5), with much anastomosis between them. The finer fibers of both groups spread widely in the neuropil of the peduncle and dorsal and isthmic tegmentum, and some descend in fascicles of group (7). The primary function of groups (7), (8), and (9) appears to be the control of the feeding musculature of the jaws and hyoid from the corpus striatum, thalamus, and tectum.


Fascicles {10). — These comprise the greater part of the ventral fascicles of the lateral forebrain bundles (figs. 26-32, 101, 102, f.lat.t.v. and (10)). In 1927 (p. 287) this was termed tr. strio-peduncularis, but it now appears that these fibers arborize not only in the peduncle but also throughout the length of the isthmic tegmentum, though they do not extend as far spinalward as those of group (9). This tr. strio-peduncularis et tegmentalis arises from the ventral nucleus of the corpus striatum, decussates partially in the anterior commissure, traverses the ventral thalamus, and in the peduncle breaks up into a number of small fascicles which are widely distributed among the ventral and dorsal tegmental fascicles. Its fibers are less myelinated than those of group (9), and, in general, they distribute in the alba more ventrally and laterally (figs. 92, 102, 103; '36, pp. 304, 336, fig. 6; '396, fig. 79; '42, figs. 2, 17); compare the preceding description of the lateral forebrain bundles.


Most of the numbered groups of fascicles which have just been listed lie in the deeper layers of the alba close to the gray. Superficially of these and in part mingled with them are many dispersed fibers, loosely fasciculated in variable arrangements. These include many of the shorter tracts mentioned in the descriptions of the several regions, such as thalamo-tegmental, pedunculo-tegmental, tectotegmental, and tegmento-bulbar tracts. The well-defined tr. olfactopeduncularis also belongs in this series. The thalamo-tegmental fibers are very numerous, and some of them are well fasciculated. From both dorsal and ventral thalamus uncrossed fibers stream backward to the isthmus, some deeply and some superficially — tr. thalamotegmentalis rectus (figs. 15, 31-34, tr.th.teg.r.). These are in addition to the thick fibers from the ventral thalamus, which enter tegmental fascicles (6) and (8), some uncrossed and some decussating in the postoptic commissure (fig. 17, tr.th.teg.v.r. and c. '3.96, p. 546). Analysis of the fibers from the dorsal thalamus which decussate in this commissure is very difficult. The earlier accounts are confused and inaccurate, but the essential features are now clear, as described in the next chapter.


FASCICULUS TEGMENTALIS PROFUNDUS

This name has been given to a loosely arranged sheet of fibers which borders the gray in the tegmental region of the midbrain and the isthmus. Most of these fibers are unmyelinated and are part of an


THE SYSTEMS OF FIBERS 287

extensive tegmento-peduncular and tegmento-bulbar system of internuncials, passing obliquely from the tegmental to the motor field; but fibers of various other systems are mingled with these. They comprise important components of the periventricular, deep and intermediate tegmental neuropil, with a dorsoventral trend. At the outer border of the gray they are especially numerous, and here some of the components are fasciculated as anatomically separate tracts for part of their courses. Other more dispersed components also are named as tracts in cases in which their terminal connections are revealed by elective impregnations. Many of these fibers decussate in the ventral commissure in the tuberculum posterius and spinalward of it through the medulla oblongata. In the midbrain and upper rhombencephalon there are numberless short crossed and uncrossed fibers, some strictly commissural between the tegmental fields and larger numbers passing from the tegmentum obliquely rostrad or caudad, with or without decussation, to other parts of the motor field.


The list below includes all the tracts which have been identified in this complex, and in addition to these there are diffuse connections through the neuropil with all neighboring regions. This summary is based upon what has been seen in both Amblystoma and Necturus.


Brachium conjunctivuni (p. 176 and fig. 10). — This is one of the largest and most compact components of the complex, though rarely impregnated in our material. From the cerebellum to its decussation it accompanies the isthmic sulcus, as is well shown in figure 71.


Tertiary visceral-gu.s-tatory tract (p. 169 and figs. 8, *23). — These fibers accompany those of the brachium conjunctivuni and are indistinguishable from them except in elective impregnations.


Tegmento-inter peduncular and inter pedunculo-teginental fibers (pp. 199, 201 and figs. 19, 60-68, 79, 80, 83, 84).— These very numerous fibers, passing in both directions between the interpeduncular nucleus and the isthmic and trigeminal tegmentum, are diffusely spread in the neuropil of the gray and the deeper layers of alba.


Tractus tegmento-pedimcularis (p. 215 and fig. 18). — These fibers pass from the dorsal and isthmic tegmentum to the peduncle, dispersed at all depths of gray and white substance. Some of them decussate in the ventral commissure in the vicinity of the fovea isthmi. They accompany similar fibers of tr. tecto-peduncularis (fig. 24).


Tractus tegmento-isthmialis (p. 283 and fig. 21).— These fibers from the dorsal to the isthmic tegmentum form a large part of the dorsal tegmental fascicles of group (7), but many of them are mingled with more dispersed fibers of tr, tecto-tegmentalis.


Tractus tecto-tegmentalis (fig. 21).— Fibers from all parts of the superior and inferior colliculus stream backward into the isthmic tegmentum. Most of the former enter dorsal tegmental fascicles of group (7) in company with longer fibers of tr. tecto-bulbaris rectus (p. 225), but many of them descend dispersed in the neuropil (fig. 79).


Tractus pedunculo-tegmentalis crticiatus.— These, as described above (p. 278), enter the ventral median tegmental fascicles of group (1) from the peduncle and, after decussation here, pass to the isthmic tegmentum by way of f . tegmentalis profundus.


Tractus 7namillo-tegmentalis .—^ome fibers from the dorsal hypothalamus reach the isthmic tegmentum accompanying the superficial tr. mamillo-cerebellaris (p. 170). Others take a deeper course in ventral tegmental fascicles of group (3) and f . tegmentalis profundus, as described on page 279.



CHAPTER XXI THE COMMISSURES

GENERAL CONSIDERATIONS

THE commissures are in two series, dorsally and ventrally of the ventricles. The fibers which cross the mid-plane are of two sorts: (1) some are strictly commissural, connecting corresponding regions of the two sides; (1) most of them are decussations, like the optic chiasma, connecting dissimilar regions. Those of the dorsal series include fibers of correlation which arise and terminate within the sensory zone and also connections between the sensory zone and other zones. Those of the ventral series are concerned in the main with motor co-ordination on the two sides of the bod3% some passing from sensory and intermediate zones to the motor zone, others lying wholly within the motor zone, and both sorts being accompanied by uncrossed fibers.


In addition to the crossed systems of correlation and co-ordination to which reference has just been made, it is a noteworthy fact that the main lines of fore-and-aft ascending and descending conduction in the brains of all vertebrates decussate, so that the adjusting centers for organs on the right side of the body are in the left side of the brain and vice versa. This applies especially to the apparatus of somatic adjustment, but not so generally to the visceral systems. Most of the fibers of the ascending secondary visceral-gustatory tract {tr.v.a.) and of the olfacto-hypothalamic tracts are uncrossed. The proprioceptive systems are both crossed and uncrossed.


The reason for the decussation of the major conduction pathways of the somatic sensori-motor systems has puzzled neurologists for many years, and fantastic theories have been expressed, some bearing the names of great masters — Wundt, Flechsig, Cajal, and others. This extensive literature has been reviewed by Jacobsohn-Lask ('24), who, instead of elaborating a theory in terms of the highly specialized human brain, like his predecessors, reviews the entire history of the evolution of the nervous system, from its first appearance in coelenterates, in relation to the bilateral symmetry of the body. None of these speculations have yielded satisfying conclusions.


A survey of the commissures and decussations of all vertebrate brains shows that some of them, like the posterior commissure and optic chiasma, hold constant positions throughout the series, while others vary widely in position and composition. It is evident that in the latter cases the site of crossing of a particular system of fibers is determined by the topographic arrangement of the parts to be connected. These arrangements are widely diversified in the lower groups of vertebrates and the crossing fibers tend to take the shortest available pathways. But when the course of a particular decussating tract has been established in an ancestral species, the original site of the decussation may be retained, despite great changes in the relative sizes of parts of the brain in phylogenetic descendants of the ancestral form, so that the tract may take a circuitous course to reach its decussation, as illustrated by some components of the postoptic complex in fishes. This conservatism may be accounted for by the fact that the decussating fibers may have collateral connections at any part of the course both before and after crossing. These considerations suggest that great caution should be observed in using the actual sites of crossing of the various systems of fibers as criteria of homology. Some of them are very stable; in other instances fibers with similar origins and terminations may cross in surprisingly different places, as illustrated by the various courses taken by fibers of the hippocampal commissure.


Primitively the roof plate of the brain, unlike the floor plate, was membranous, and the locations of commissures which invade it are determined by the functional requirements of the organs differentiated below it in the various species of animals. In the telencephalon the topographic arrangements of parts are extremely variable. In all cases there is a wide interruption of the commissures at the site of the stem-hemisphere fissure, paraphysis, dorsal sac, and their derivatives. In the diencephalon the habenular commissure is separated by the pineal recess from the commissura tecti, and the latter is continuous spinal ward with the posterior commissure and the com. tecti of the mesencephalon. In cyclostomes the middle part of the mesencephalic tectum is a membranous chorioid plexus, and the com. tecti is here interrupted.


At the posterior end of the tectum its commissure is continuous with a thin sheet of crossed and uncrossed fibers in the anterior medullary velum, and this, in turn, with the massive cerebellar commissures. Between the cerebellum and the calamus scriptorius the


THE COMMISSURES 291

roof is a membranous chorioid plexus except in some fishes, in which the large acousticolateral areas fuse over the ventricle, with some decussating fibers. In the region of the calamus there is a dorsal crossing of visceral sensory fibers (com. infima of Haller) and of somatic sensory fibers between the funicular nuclei. This dorsal commissure, reduced in size, is present through the entire length of the spinal cord.


The ventral series of commissures is concentrated in the anterior commissure of the telencephalon, the chiasma ridge of the diencephalon, the so-called ansulate commissure of the midbrain; and posteriorly of this it extends continuously as the ventral commissure of the rhombencephalon and spinal cord. The ventral decussations of Necturus from the tuberculum posterius to the spinal cord were described in 1930 ('30, p. 89) and the other commissures in papers before and after that date.


The arrangement of the telencephalic commissures of the Amphibia differs radically from that of all animals higher in the scale. As seen in median section (fig. 2), the anterior and hippocampal commissures do not cross in or above the lamina terminalis but in a high anterior commissure ridge which projects upward from the floor posteriorly of the interventricular foramina. This ridge is separated from the lamina terminalis by the wide precommissural recess, or aula, which is the vestibule of the interventricular foramina (fig. IB). Both pallial and subpallial parts of the hemispheres contribute fibers to the commissures in this ridge, and many fibers from the olfactory and hippocampal areas also cross in the habenular commissure.


The peculiar arrangement of the anterior and hippocampal commissures, the lamina terminahs, and the related chorioid plexuses of Ambly stoma has been described and illustrated ('35). It is explained by the topography at the di-telencephalic junction and the presence of unusually wide interventricular foramina. The explanation of this topography, in turn, is to be^ sought in the phylogenetic ancestors of existing Amphibia. In Protopterus as described by Rudebeck ('45), pallial formation is confined to the lateral wall of the hemisphere, and the hippocampal commissure passes down behind the foramen to cross in close association with the anterior commissure in essentially the same way as in amphibians. In view of the close similarity of development of the brains of amphibians and lungfishes, it is probable that the ancestral amphibian resembled Protopterus in this region.


The arrangement of myelinated fibers in all the commissures of Amblystoma as seen in the median plane is shown in figure 2C. The components of these commissures will now be summarized, with references to more detailed descriptions.


THE DORSAL COMMISSURES

Commissura hippocampi. — These fibers converge from all parts of the hippocampal area to its ventral border, from which they descend behind the foramen to cross in the dorsal part of the anterior commissure ridge (figs. 96-99). Many of these fibers join a mixed longitudinal fascicle, termed "fimbria," which borders the primordium hippocampi for its entire length at the ventromedial margin. Another fascicle, composed exclusively of commissural fibers, assembles at the ventrolateral border of the primordium. Most of these fibers are unmyelinated. Weigert sections show a few brilliantly stained myelinated fibers scattered among them.


Accompanying this commissural bundle as it leaves the hippocampus are other similar fibers of tractus cortico-thalamicus medialis and tr. cortico-habenularis medialis. Some of the latter cross in the habenular commissure and return to the hippocampal area of the opposite side, this being the com. pallii posterior (figs. 32, 34, 71, 72, 76).


Commissura hahenularum {com. superior). — The analysis of the stria medullaris as detailed in chapter xviii reveals the following components of the habenular commissure: (1) The most anterior member is the com. pallii posterior (fig. 20, no. 8). (2) Posteriorly of this the com. superior telencephali includes components 3, 4, 5, and 6 of the diagram, viz., tr. olfacto-habenularis anterior, tr. cortico-habenularis lateralis, and tr. amygdalo-habenularis. (3) Still more posteriorly are crossing fibers of uncertain connections, some of which may come from other components of the stria medullaris. There are doubtless strictly commissural fibers connecting the habenulae of the two sides. The commissural fibers from the hemisphere are accompanied by uncrossed fibers. It is possible that these tracts are accompanied by crossed and uncrossed fibers, which pass from the habenular nuclei forward into the hemispheres, but this has not been demonstrated. Commissura tecti dieiicephali.— Most of these fibers are commissural, connecting the two pretectal nuclei. Some tecto-habenular and habenulo-tectal fibers decussate here.



THE COMMISSURES 293

Commissura posterior. — This is the primary pathway from the anterior part of the optic tectum to the motor zone of the midbrain. Its decussating fibers appear very early in embryogenesis, accompanied by uncrossed fibers from the tectum and eminence of the posterior commissure. In the adult many of the commissural fibers end in this eminence. Crossed and uncrossed fibers spread widely in the posterior part of the thalamus, the nucleus of the tuberculum posterius, and the dorsal tegmentum. The largest fascicles connect with the big cells of the nucleus of Darkschewitsch (p. 217).


Commissura tecti mesencephali. — This thin sheet of crossing fibers is continuous between the posterior commissure and the anterior medullary velum. It contains thin and thick fibers (many of the latter myelinated) which spread widely through all layers of the tectum. Most of these seem to be commissural between the tecti of the two sides, but no satisfactory analysis has been recorded. Some fibers of the mesencephalic root of the V nerve apparently decussate here, but, if so, the number is small.


Commissures of the anterior medullary velum. — Most of the fibers in the velum are longitudinal — tr. tecto-cerebellaris — and some of these may decussate here. The most constant and noteworthy component is the decussation of the IV nerve roots ('36, p. 342; '42, p. 255). The velum contains cells and fibers of the mesencephalic V root, and some of these may decussate here. Our preparations give no clear evidence of crossed fibers of this root; if present, the number is certainly not large.


Cerebellar commissures. — Larsell's analysis of these commissures is confirmed. The two systems are quite distinct. (1) The com. cerebelli is related with the median body of the cerebellum, including decussating fibers of tr. spino-cerebellaris, sensory root fibers of the V nerve, secondary trigeminal fibers from the superior sensory V nucleus in the auricle, and commissural fibers between these nuclei and between the two sides of the corpus cerebelli. (2) The com. vestibulo-lateralis cerebelli is a more dispersed system of fibers related with the vestibular and lateral-line centers of adjustment in the auricles. It is composed of root fibers of the VIII nerve and secondary fibers of both vestibular and lateral-hne systems. There are doubtless also commissural fibers between the two auricles. None of these fibers make significant connections with the median body of the cerebellum through which they pass. Their terminal relations are with auricular tissue which is the primordium of the floccular part of the mammahan flocculonodular lobe.


Commu-sura infima Halleri. — This is a decussation of the fascicuH soHtarii at the calamus scriptorius, containing both root fibers and secondary fibers of the visceral-gustatory system and doubtless also commissural fibers between the two commissural nuclei of this system.


Commifisure of the funieular nuclei. — Intimately associated with the preceding are commissural fibers between the nuclei of the dorsal funiculi in the calamus region, with which decussating fibers are mingled. This dorsal commissure of somatic sensory fibers is extended, reduced in size, downward through the entire length of the spinal cord. There is some evidence that the visceral sensory com. infima is also represented in the cord (p. 125).


This completes the summary of the dorsal commissures. We now turn to the ventral series, beginning, as before, at its anterior end.


THE VENTRAL COMMISSURES COMMISSURA ANTERIOR

The complex com. anterior occupies the entire anterior commissure ridge except its dorsal border. Its largest components are the partial decussations of the medial forebrain bundles below and the lateral forebrain bundles above (figs. 25-28). Associated with these fibers are others, including the com. amygdalarum (which is part of the stria terminalis system, p. 256), some fascicles of the nervus terminalis, and a dispersed decussation between the olfactory fields of the anterior parts of the hemispheres ('396, fig. 21).


The crossing fibers of the anterior commissure ridge are enveloped by a thin layer of gray which expands laterally as the large bednuclei of the anterior commissure. The thin floor of the long preoptic recess between the anterior commissure ridge and the chiasma ridge contains the longitudinal fibers of tr. preopticus, some of which decussate here.


CHIASMA OPTICUM

All fibers of the optic nerves decussate in the chiasma opticum. The crossing occupies the anterior border of the chiasma ridge (figs. 2B and 2C). At the posterior border of the chiasma there is some mingling of optic fibers with those of the postoptic commissure, but in some of our Golgi preparations the optic fibers are electively impregnated and can be separated from the others ('41, '42).



THE COMMISSURES 295

COMMISSURA POSTOPTICA

The complex com. postoptica is represented in mammals by the supra-optic commissures, but its composition is so different in urodeles and mammals that exact homologies cannot be established. Further analysis of intervening species is requisite before these relationships can be clarified.


In Amblystoma this complex includes decussating fibers derived from the superior and inferior colliculi, the entire diencephalon, the amygdala, and the subpallial olfactory field of the cerebral hemisphere. Many of these tracts have collateral connections along their courses, both before and after crossing, and are accompanied by uncrossed fibers. There are also strictly commissural fibers connecting some of these regions (for evidence of such fibers in the frog see '25, p. 480). These decussating systems connect, after crossing, with extensive fields of the intermediate and motor zones, including the strio-amygdaloid area, preoptic nucleus, hypothalamus, ventral thalamus, geniculate neuropil, dorsal tegmentum, peduncle, isthmic tegmentum, and bulbar tegmentum. The posterior part of the postoptic commissure is comparable with the com. tuberis of some other vertebrates and consists mainly of decussating fibers from the ventral part of the hypothalamus to the peduncle and interpeduncular nucleus. The direction of conduction of most of these fibers has not been clearly determined, though most of the larger systems evidently converge from the sensory zone into the motor field.


The fibers of few of these components are assembled in wellorganized tracts ; most of them are so dispersed and commingled that analysis is very difficult. Some of them are myelinated, and these tend to be assembled in recognizable tracts. These myelinated tracts, as seen in Weigert sections, were the first to be described, but their distribution after crossing baffled analysis. These tracts are accompanied by much larger numbers of unmyelinated fibers in more dispersed arrangement. The courses of some of these have been revealed by elective Golgi impregnations, and other systems have been clarified by study of the sequence of their development as published in a series of papers from 1937 to 1941. In view of the complexity of these connections and the technical difficulties encountered in their study, it is not surprising that the earlier descriptions were incomplete and not free from error. It is believed that now it is possible to present an analysis of this complex which, though still incomplete, reveals its major features.


This generalized arrangement as seen in urodeles is probably primitive and may be taken as the point of departure in the study of the postoptic systems of more specialized brains of both fishes and higher vertebrates. These commissures have been analyzed in Necturus ('41a, p, 513), where they are still more generalized. Our present knowledge of these systems of Amblystoma was summarized on pages 219-28 of the paper of 1942, with diagrams illustrating the connections of the principal tracts. All known components are assembled in the following list. Here some of the earlier names of tracts have been replaced by more accurate terms; some others are retained, though now known to be inappropriate or inadequate. In this list there are included, first, the groups of fibers which descend to the chiasma ridge from the tectum and pretectal nucleus, followed by those descending from the thalamus, next, the systems arising in the hypothalamus, and, finally, a heterogenous group with hypothalamic connections. For additional details about some of these tracts in preceding chapters consult the Index.


1. Tr actus tecto-thalamicus et hypothalamicus cruciatus anterior (fig. 12, tr.t.th.h.c.a.). — This anterior tectal fasciculus is a mixture of myelinated and unmyelinated fibers from the dorsomedial part of the tectum opticum and the pretectal nucleus which descend across the thalamus in company with the more anterior fascicles of the optic tract. After partial crossing in the antero ventral part of the chiasma ridge, its fibers spread in the neuropil of the chiasma ridge and hypothalamus. Some of them may reach beyond this region. This cumbersome name was applied in my earlier papers to a mixed fascicle which had not been analyzed. It can now be replaced by the names of the several tracts of which it is composed. Some of these are uncrossed (notably tr. pretecto-thalamicus) , and some fibers of two of them decussate in the chiasma ridge, nos. 2 and 3 below. The arrangement of these components as seen in horizontal sections is shown in figures 25-36.


2. Tractus tecto-hypothalamicus anterior (tr.t.hy.a.). — This is the tectal component of the preceding fasciculus; it is evidently an optic pathway to the hypothalamus (p. 224). It passes through the pretectal nucleus and is accompanied by tecto-pretectal fibers. It probably is physiologically related with no. 3.


3. Tractus pretecto-hypothalamicus (fig. 15, tr.ptJiy.). — This tract as


THE COMMISSURES 297

it leaves the pretectal nucleus is accompanied by a large tr. pretectothalamicus (p. 234 and figs. 35, 36).


4. Tr actus tecto-ihalamicu.s ei hypothalamicus cruciatus posterior (fig. 12, tr.t.th.h.c.p). — This, like no. 1, is a mixed fascicle which has been analyzed. Both these names may now be discarded in favor of shorter terms — anterior and posterior tectal fascicles. The posterior fascicle arises chiefly from the nonoptic nucleus posterior tecti and the adjoining ventrolateral margin of the optic tectum, the latter region in Necturus receiving few terminals of the optic tract ('41a, p. 516). This fascicle is probably activated primarily by lemniscus, rather than optic, fibers or by a combination of the two. The tracts of which it is composed have collateral connections with the geniculate neuropil and ventral thalamus both before and after crossing. These fibers descend from the tectum parallel with those of the lateral optic tract and internally of them. So far as known they have a common origin in the tectum, but, after crossing, they take widely divergent courses. The tracts, consequently, are named according to their terminal distribution. The most important components are the following, nos. 5 and 6.


5. Tractus tecto-hypothal amicus posterior. — These are finer fibers which terminate in the postoptic neuropil and neighboring regions of the hypothalamus. They are more clearly seen in Necturus ('41a, p. 516) than in Ambly stoma.


6. Tractus tecto-tegmentalis cruciatus (fig. 12, tr.t.teg.c). — The course and distribution of the thicker fibers of the posterior tectal fascicle were not clarified until they were identified in larval stages, in which they were electively impregnated because they mature precociously. These were first recognized in early feeding larvae ('39, p. 106) and later in adult Necturus ('41a, p. 516) and Amblystoma ('42, p. 222). Some erroneous descriptions of these fibers in my earlier papers have been corrected ('39, p. 110).


The more heavily myelinated fibers of this tract decussate in the dorsal part of the chiasma ridge (figs. 2C, 12, 16, 25, tr.t.th.h.c.p.), and after crossing they enter tegmental fascicles of groups (8) and (6), and in smaller numbers they are dispersed in other fascicles. The dispersed fibers spread in the peduncle. Those which enter fascicles numbered (8) take a longer and more dorsal course, distributing to the dorsal, isthmic, and trigeminal tegmentum, some of them extending as far as the level of the V nerve roots. The entire course of these fibers can be followed in the horizontal sections, figures 25-35, where they are marked tr.t.th.h.c.p. before their decussation and (8)

beyond the crossing.


There are two strong systems of crossed tecto-peduncular and tecto-tegmental fibers, both of which are drawn in figure 12. The system just described descends chiefly from the nonoptic part of the tectum {tr.t.th.h.c.p.) and, after crossing, spreads in the peduncle by way of tegmental fascicles (6) and throughout the tegmentum by way of fascicles (8). The second system is tr. tecto-peduncularis cruciatus (tr.t.p.c), which arises in the optic part of the tectum, crosses in the commissure of the tuberculum posterius, and then spreads out in the peduncle. This well-myelinated tract is accompanied by similar fibers from the pretectal nucleus and dorsal thalamus. These two tectopeduncular systems evidently have quite different physiological significance.


Attempts to analyze the postoptic components arising in the thalamus were unsuccessful until the sequence of development of these fibers was revealed by embryological studies. These findings were then confirmed by elective Golgi impregnations of older larvae and adults. Some errors in the earlier descriptions have been corrected, and now it is possible to give a fairly complete account of both the crossed and the uncrossed fibers which diverge from the thalamus. Since the direct and crossed fibers are evidently intimately related physiologically, both series are included in the following description. Efferent fibers from the thalamus are arranged in two sharply contrasted series, which arise, respectively, from the dorsal thalamus and the ventral thalamus.


The efferent series from the dorsal thalanuis includes uncrossed fibers to the tectum, habenula, cerebral hemisphere, ventral thalamus, hypothalamus, and peduncle which need not be further considered here; but some of the other uncrossed tracts, which evidently are in reciprocal physiological relation with the crossed tracts, should be specifically mentioned. Decussating fibers from the dorsal thalamus are in two groups. The first includes thick myelinated fibers of tr. thalamo-peduncularis cruciatus (tr.th.p.c), which, as already mentioned, joins tr. tecto-peduncularis cruciatus {tr.t.p.c.) to decussate in the commissure of the tuberculum posterius as described under that caption below. The second group is a much larger number of thin unmyelinated or lightly myelinated fibers which decussate in the postoptic commissure and will next be described.



THE COMMISSURES 299

7. Tractus thalamo-hypothalamicus et peduncnlaris cruciatus {tr.ih.h.p.c.). — In the earlier descriptions of both Aniblystoma and Necturus this name was given to a large collection of unmyelinated and lightly myelinated fibers which descends in disj^ersed arrangement from the dorsal thalamus to the postoptic commissure. Their distribution beyond the decussation could not be clearly followed, and some of those descriptions now require correction. As elsewhere pointed out ('42, p. 223), this name should now be discarded because at least three quite distinct tracts are represented here and relatively few of these fibers have any connection with the peduncle. The three tracts represented in this complex (nos. 8, 9, and 10) have a common origin in the dorsal thalamus, chiefly its middle sector, and, after crossing, take widely different courses. Their decussation is posterior to that of the tectal components of the commissure in a band which is narrow dorsally and spreads ventrally through a wide area of the neuropil of the chiasma ridge (fig. 2C, tr.th.h.d.c).


8. Tractus thalamo-hypothalamicus dorsalis cruciatus (tr.th.h.d.c). — These fibers, most of which are unmyelinated, cross in the middle region of the chiasma ridge (figs. 2C, 15, 25) and then spread in the hypothalamus. In figures 27-33, 95, 102, and 103 the symbol tr.th.h.d.c. refers to the mixture of fibers of tracts 8, 9, and 10 in their descending course from the thalamus to the commissure. The fibers of nos. 9 and 10 cross dorsally of those of no. 8, though there is mingling of the fibers of the three tracts with one another and with those of surrounding decussations.


The remaining fibers of this complex, after crossing, are distributed to the tegmentum in two tracts which take parallel courses, one superficially, the other at the border of the gray. These are designated components A and B, respectively. In figures 15 and 21 the components A and B are not separately designated.


9. Tractus thalamo-tegmentalis dorsalis cruciatus A {tr.th.teg.d.c.A.). — This tract was first described in the early feeding larva ('39, p. 116) and later in the adult ('42, p. 224). After decussation, these unmyelinated fibers ascend from the chiasma ridge parallel with the course of the descending uncrossed limb of this commissure and more superficially. Their further course is shown in figures 26-34, here marked A . In the dorsal tegmentum they turn spinal ward along the ventrolateral border of the tectum. Here they lie close to the pial surface and immediately ventrally of the lateral optic tract (fig. 94, A). In this part of their course they join an uncrossed tract with similar origin from the dorsal thalamus and similar distribution in the tegmentum — tr. thalamo-tegmentalis rectus (figs. 15, 31-34, 94, tr.th.teg.r.). This latter tract arises from both dorsal and ventral thalamus, but only the dorsal component of it is under consideration here (fig. 21, tr.th.teg.d.r.). In some Golgi preparations there is evidence that axons from the dorsal thalamus may divide, with branches entering both the uncrossed and the crossed thalamo-tegmental tracts ('42, p. 224). Evidently, the crossed and uncrossed tracts are reciprocally related physiologically.


10. Tractus thalamo-tegmentalis dorsalis cruciatus B. — This is a deep component of the same system as the preceding, receiving nearly all the myelinated fibers of no. 7 of this list. In the chiasma ridge these fibers separate from the others of the group and cross at the dorsal border of the postoptic commissure (marked B in figs. 26 and 95). Beyond the decussation they scatter widely, and most of them enter the well-myelinated tegmental fascicles of group (8), within which they may descend as far as the V nerve roots.


The two crossed tracts to the tegmentum, nos. 9 and 10 of this list, seem to have the same origin and about the same field of distribution, except that one of them {A) terminates in the superficial tegmental neuropil and the other {B) arborizes in the deep neuropil. The physiological properties of these zones of neuropil evidently are different.


11. Tractus thalamo-tegmentalis ventralis cruciatus (figs. 2C, 17, tr.th.teg.v.c). — These thick fibers (many of them well myelinated) converge into the postoptic commissure from all parts of the ventral thalamus. In this part of their course they are not fasciculated but are scattered among other similar fibers so that their courses could not be followed until they were studied embryologically. They mature early and may be impregnated with reduced silver at stages when few other fibers respond to this treatment. We also have good elective Golgi impregnations of them in later larval stages ('39, pp. 98, 120; '396, p. 546; '42, p. 225).


These fibers cross in the posterodorsal part of the chiasma ridge, mingled with those of other systems. The thickest of the myelinated fibers are crowded together at the dorsal margin of the postoptic complex. After crossing, they spread widely in the peduncle and tegmentum. Most of those from the posterior part of the ventral thalamus enter ventral tegmental fascicles of group (4), and some of these may descend in the f . longitudinalis medialis. Many fibers from the


THE COMMISSURES 301

middle and anterior parts of the ventral thalamus reach the tegmentum through fascicles of groups (6) and (8) . These crossed fibers are accompanied by many others that take similar courses without decussation.


The hypothalamic components of the postoptic commissure include, in addition to the terminals of extrinsic tracts already described, a few well-defined tracts and several other less-well-known connections.


12. Tractus hypothalamo-peduncularis et tegmentalis (figs. 18, 23, tr.hy.ped.; fig. 21, tr.hy.ieg.). — This is the most noteworthy component originating within the hypothalamus. Its fibers assemble from the whole of the ventral part of the hypothalamus and comprise the main pathway from this region to the motor field of the peduncle and tegmentum. Some of them decussate in the posterior part of the chiasma ridge ; others are uncrossed ; still others decussate in the commissure of the tuberculum posterius. They connect by terminals or collaterals with the dorsal part of the hypothalamus, ventral part of the peduncle (including the neuropil of the area ventrolateralis pedunculi), and isthmic tegmentum. The last-mentioned group includes thick fibers, some of which are well myelinated, which enter ventral tegmental fascicles (4); and some of these may take long courses in the f. longitudinalis medialis (for further description see p. 280, and '42, p. 226).


13. Olfactory projection tract (fig. 19, oLp.tr.). — This name has been given to a thin strand of unmyelinated fibers which pass in both directions between the strio-amygdaloid area and a specific nucleus at the posterior border of the chiasma ridge ('21, p. 247; '27, p. 304; '36, fig. 5). Some of these fibers decussate here.


14. Tractus pedunculo-hypothalamicus. — A large component in the posterodorsal part of the postoptic commissure was provisionally given this name, though the exact connections of its fibers could not be determined ('42, p. 227).


15. Medial forebrain bundle. — The fibers of the medial forebrain bundle are interlaced with all components of the postoptic commissure laterally of the chiasma ridge. Many of these fibers of both descending and ascending systems enter the postoptic neuropil and participate in its formation. This implies a transfer of more or less of this activity to the opposite side of the brain, but no details of specific decussational or commissural pathways have been revealed.


Postoptic neuropil. — In the mid-plane, the postoptic decussations occupy most of the chiasma ridge, and these fibers are enveloped on all sides except ventrally by a gray layer, the bed-nucleus of the postoptic commissure. This nucleus is expanded laterally. From this gray layer, richly arborized ependymal elements and dendrites of neurons are spread among the decussating fibers. Similar long dendrites enter it from all surrounding parts, including the nucleus magnocellularis, from which tr. hypophysius arises ('42, fig. 51). The entire chiasma ridge is also permeated with dense axonic neuropil which is continuous with that of surrounding parts. This neuropil receives terminals and collaterals of axons of hypothalamic neurons, medial forebrain bundles, tractus preopticus, and most of the decussating systems. The chief outflow from it seems to be by axons of the nucleus of the postoptic commissure directed into tr. hypothalamo-peduncularis. Other fibers enter the medial forebrain bundle (for further details see '42, p. 219; Necturus, '336, p. 251; 'Mb, p. 383).


This neuropil is clearly one of the major adjusting centers of the urodele brain. Situated in the center of the great olfacto-visceral field, its connections indicate that it may be activated from every correlation center of the cerebrum. Undoubtedly it plays an important part in all general visceral activities. It is equally evident that there is no provision here for localization of specific functions. This is probably the undifferentiated primordium from which some of the specialized hypothalamic nuclei of mammals have been elaborated.


COMMISSURA TUBERCULI POSTERIORIS

The commissure of the tuberculum posterius was defined in 1917 (]). 224) as the ventral mesencephalic decussations between the infundibulum and the fovea isthmi. Posteriorly of the latter the ventral tegmental commissure extends backward without interruption through the rhombencephalon and spinal cord. In Necturus there is a very short interruption of the ventral commissural system at the fovea isthmi ('30, p. 89), but in most urodeles this gap does not appear.


The composition and arrangement of these ventral commissures are very diversified in different vertebrates. The hypothalamic connections at the anteroventral end of the commissure of the tuberculum posterius of Amblystoma are in some other vertebrates widely separated as the retro-infundibular decussations, or decussatio hy


THE COMMISSURES 303

pothalamicus posterior. The remainder of the commissure of the tubercuhim and the anterior part of the ventral tegmental commissure comprise the ansulate commissure of the literature.


In Amblystoma I have recognized four chief components of the commissure of the tubercuhim posterius ('36, p. 305 and fig. 2), arranged as shown in sagittal section in figures 81, 104; in horizontal sections in figures 28, 29, 30, 31 (component 4 being here marked, tr.t.b.c.l.); and in transverse section in figure 94 (components 2 and 3 shown here, components 1 and 4 being shown in a neighboring section, '36, fig. 11). The connections of these components are as follows :

1. This is the partial decussation of unmyelinated fibers from the dorsal (mamillary) part of the hypothalamus to the interpeduncular nucleus ('36, p. 338 and figs. 3, 8, 11-13). Most of them enter ventral tegmental fascicles of group (2) as marked tr.mam.inp., on figures 19, 27-30, 71, 81, 92, 103.


2. This component contains partly myelinated fibers from the dorsal hypothalamus to the peduncle and tegmentum, accompanied by many uncrossed fibers and by fibers conducting in the reverse direction from the peduncle to the hypothalamus (p. 278; figs. 8, 21, tr.mam.feg., 23, tr.mam.ped.). Most of them spread in the alba of the peduncle or enter ventral tegmental fascicles of group (3). Another system is tr. hypothalamo-peduncularis et tegmentalis from, the ventral part of the hypothalamus to the peduncle and tegmentum (fig. 21, ir.hy.teq. and fig. 23, tr.hy.ped.). There is interchange of all these fibers with those of component 1 and with other tegmental fascicles. They are marked tr.mam.ped., tr.mam.teg., and (3) on figures 18, 21, 23, 27-31, 79, 82, 92, 94, 103.


3. The third component contains heavily myelinated fibers from the tectum, pretectal nucleus, and dorsal thalamus to the peduncle, accompanied by many uncrossed fibers — tr. tecto-peduncularis cruciatus {tr.t.p.c.) and tr. thalamo-peduncularis cruciatus (tr.th.p.c), as elsewhere described (p. 223; '42, p. 267). These are shown on figures 12, 15, 18, 22, 29-36, 94, 103.


4. The dorsal component of this commissure is larger than the others and is composed chiefly of tecto-bulbar and tecto-spinal fibers from the anterior part of the tectum opticum, marked tr.t.b.c.l. on the figures. Its fibers after oblique decussation enter ventral tegmental fascicles of group (1) (see p. 277 and figs. 12, 27-36, 93, 94).


In addition to these four well-defined components, there are unmyelinated fibers which descend from the tectum and dorsal tegmentum through the gray and deeper layers of the alba to the region of the nucleus of the III nerve (fig. ^2, tr.t.ped.). Some of these tectopeduncular fibers decussate in the ventral commissure both before and behind the fovea isthmi. In Necturus some fibers of the nervus terminalis probably decussate in the commissure of the tuberculum posterius (McKibben, '11), and this may be true in Amblystoma.


COMMISSURA VENTBALIS

The ventral decussations of Necturus were analyzed in the paper of 1930 (p. 89), and those of Amblystoma are similar. For the details the reader is referred to that description and to a later contribution on larval Amblystoma ('396).



BIBLIOGRAPHY


BIBLIOGRAPHY

This list includes onh' the works cited in the text. It is not a systematic bibliography of the amphibian nervous system. Additional references are cited in the author's papers, in the comprehensive work by Ariens Kappers, Huber, and Crosby ('36), and in numerous other publications.


Addens, J. L. 194(). The nucleus of Bellonci and adjacent cell groups in selachians.


II, Proc. kon. Akad. Wetensch., Amsterdam, 49: 94-100. Also earlier papers cited. Adelmann, H. B. 1929. Experimental studies on the development of the eye. I,

J. Exper. Zool., 54:249-90.


. 1936. The problem of cyclopia. Quart. Rev. Biol., 11: 161-82; 284-304.


Agar, W. E. 1943. A contribution to the theory of the living organism. Melbourne and London: Melbourne University Press and Oxford University Press. Ariens Kappers, C. U. 1929. The evolution of the nervous system in invertebrates,

vertebrates and man. Haarlem: Bohn. Ariens Kappers, C. U.; Huber, G. Carl; and Crosby, E. C. 1936. The comparative anatomy of the nervous system of vertebrates including man. New York:

Macmillan Co. Aronson, Lester R., and Noble, G. K. 1945. The sexual behavior of Anura.


II. Neural mechanisms controlling mating in the male leopard frog, Rana pipiens,

Bull. Am. Mus. Nat. Hist., 86:83-140, article 3. Bagley, Charles, Jr., and Langworthy, O. R. 1926. The forebrain and midbrain

of the alligator, etc., Arch. Neurol. & Psychiat., 16: lo4-66. Bailey, P., and Bonin, G. von. 1946. Concerning cytoarchitectonics. Trans. Am.


Neurol. Assoc, 71st Meeting. Pp. 89-93. Bailey, P., and Davis, E. W. 1942. Effects of lesions of the periaqueductal gray

matter in the cat, Proc. Soc. Exper. Biol. & Med., 51:305-6.


. 1942a. The syndrome of obstinate progression in the cat, ibid., p. 307.


Baker, R. C. 1927. The early development of the ventral part of the neural plate

of Amblystoma, J. Comp. Neurol., 44:1-27. Baker, R. C, and Graves, G. O. 1932. The development of the brain of Amblystoma (3 to 17 mm. body length), J. Comp. Neurol., 54:.501-59. Barnard, J. W. 1936. A phylogenetic study of the visceral afferent areas, etc.,

J. Comp. Neurol., 65:. 503-602. Bellonci, J. 1888. Ueber die centrale Endigung des Nervus opticus bei den Verte braten, Ztschr. f. wissensch. Zool., 47:1-46. Benedetti, E. 1933. II cervello e i nervi cranici del Proteus anguineus Laur.,

Mem. 1st. ital. di speleologia, ser. biol., Mem. Ill, pp. 1-80. Benzon, a. 1926. Die markhaltigen Faserziige im Vorderhirn von Cryptobranchus

japonicus, Ztschr. f. mikr.-anat. Forsch., 5:285-314. Beritoff, J. S. (ed.). 1943. Trans, of the J. Beritashvili Physiologica Institute, No.


5. Pp. xiv+532. Tbilisi (Tiflis), Georgia, U.S.S.R. (with abstracts in English). BiNDEWALD, C. A. E. 1914. Das Vorderhirn von Amblystoma mexicanum, Arch. f.


mikr. Anat., Abt. 1, 84:1-74. Bishop, S. C. 1943. Handbook of salamanders: the salamanders of the United

States, of Canada and of Lower California. Ithaca, N.Y.: Comstock Pub. Co. BoDiAN, D. 1937. The structure of the vertebrate synapse, J. Comp. Neurol., 68:117-59.


BoDiAN, D. 1942. Cytological aspects of synaptic function, Physiol. Rev., 22: 146-69. BoNiN, G. VON. 1945. The cortex of Galago; its relation to the pattern of the primitive cortex. Illinois Monog, M. Sc. Vol. 5, No. 3. Urbana: University of Illinois

Press. Brickner, Richard M. 1930. A new tract in Herrick's gustatory system in certain

teleosts, J. Comp. Neurol. 50:153-57. Brodal, a. 1940. Experimentelle Untersuchungen uber die olivo-cerebellare Lokali sation, Ztschr. f. d. ges. Neurol, u. Psychiat., 169:1-153. Brodal, A., and Jansen, J. 1946. The ponto-cerebellar projection in the rabbit

and cat, J. Comp. Neurol., 84:31-118. BucY, Paul C. (ed.). 1944. The precentral motor cortex. Illinois Monog. M. Sc,

Vol. 4, Urbana: University of Illinois Press. Burr, H. S. 1922. The early development of the cerebral hemispheres in Am blystoma, J. Comp. Neurol., 34:277-301. Calderon, Luis. 1928. Sur la structure du ganglion interpedonculaire, Trav. du

lab. de recherches biol. de I'Univ. de Madrid, 25:297-306. Campion, George G., and Smith, G. Elliot. 1934. The neural basis of thought.


New York: Harcourt, Brace & Co. Chezar, H. H. 1930. Studies on the lateral-line system of Amphibia. II, J. Comp.


Neurol., 50:159-75. Child, C. M. 1941. Patterns and problems of development. Chicago: University of

Chicago Press. Clark, W. E. Le Gros. 1933. The medial geniculate body and the nucleus isthmi,

J. Anat., 67:536-48. . 1943. The anatomy of cortical vision, Tr. Ophth. Soc, 62:229-45 (for

1942). Coghill, G. E. 1902. The cranial nerves of Amblystoma tigrinum, J. Comp. Neurol.,

12:205-89. . 1913. The primary ventral roots and somatic motor column of Amblystoma,

ibid., 23:121-43.


1914-36. Correlated anatomical and physiological studies of the growth of


the nervous system of Amphibia, Papers I-XII, ibid.. Vols. 24^64.


1929. Anatomy and the problem of behaviour. Cambridge: Cambridge


University Press.


1930. The structural basis of the integration of behavior, Proc. Nat. Acad.



Sc, 16:637-43. . 1933. The neuro-embryological study of behavior: principles, perspective


and aim, Science, 78:131-38.


1936. Integration and motivation of behavior as problems of growth.



J. Genet. Psychol., 48:3-19.


1940. Early embryonic somatic movements in birds and in mammals other


than man. Monog. Soc. Research in Child Development, Vol. 5, No. 2. Washington, D.C.: National Research Council.


1943. Flexion spasms and mass reflexes in relation to the ontogenetic


development of behavior, J. Comp. Neurol., 79:463-86. CoNEL, J. LeRoy. 1929. The development of the brain of Bdellostoma stouti.


I. External growth changes, J. Comp. Neurol., 47:343-403. . 1931. The development of the brain of Bdellostoma stouti. II. Internal

growth changes, ibid., 52:365-499. CoRBiN, K. B. 1940. Observations on the peripheral distribution of fibers arising

in the mesencephalic nucleus of the fifth cranial nerve, J. Comp. Neurol., 73:

153-77.



BIBLIOGRAPHY 309

Craigie, E. Horne. 1938. The vascularization of the hypophysis in tailed amphibians, Tr. Roy. Soe. Canada, ser. 3, 32:43-50.


. 1938a. The blood vessels of the brain substance in some amphibians, Proc.


Am.Phil.Soc, 78:615-49.


1939. Vascularity in the brains of tailed amphibians. I. Amblystoma


tigrinum (Green), ibid., 81: "21-27. . 1940. The cerebral cortex in palaeognathine and neognathine birds, J. Comp.



Neurol., 73:179-234. . 1945. The architecture of the cerebral capillary bed, Biol. Rev., 20:133


46. Crosby, Elizabeth C. 1917. The forebrain of Alligator mississippiensis, J. Comp.


Neurol., 27:325-402. Crosby, Elizabeth C, and Woodburne, R. T. 1938. Certain major trends in the

development of the efferent systems of the brain and the spinal cord, Univ.


Hosp. Bull., Ann Arbor, Mich., 4:125-28. . 1940. The comparative anatomy of the preoptic area and the hypothalamus,

A. Research Nerv. & Ment. Dis., Proc, 20: 52-169. CusHiNG, H. 1903. The taste fibers and their independence of the n. trigeminus,

Bull. Johns Hopkins Hosp., 14:71-78. Dempster, W. T. 1930. The morphology of the amphibian endolymphatic organ,

J. Morphol., 50:71-126. Detwiler, S. R. 1945. The results of unilateral and bilateral extirpation of the

forebrain of Amblystoma, J. Exper. Zool., 100:103-17. . 1946. Experiments upon the midbrain of Amblystoma embryos. Am. J.


Anat., 78:115-38. Dewey, John. 1896. The reflex arc concept in psychology, Psychol. Rev., 3:357 70. See also Dewey's later statement in J. Phil., 9:664-68, 1912, especially the

footnote on p. 667. Dow, R. S. 1942. The evolution and anatomy of the cerebellum, Biol. Rev., 17:

179-220. EcoNOMO, C. 1926. Ein Koeffizient fUr die Organizationshohe der Grosshirnrinde,

Klin. Wchnschr., 5 Jhrg., I Halbjahres, pp. 593-95. . 1929. The cytoarchitectonics of the cerebral cortex. Translated by S. Parker. Oxford University Press. Edinger, L. 1911. Vorlesungen iiber den Bau der nervosen Zentralorgane, Vol 1.


8th ed. Leipzig: F. C. W. Vogel. Emerson, Alfred E. 1942. The modern naturalist. Bull. Transylvania Coll., 15:

71-77.


, 1943. Ecology, evolution and society. Am. Nat., 77:97-118.


Est ABLE, C. 1924. Terminaisons nerveuses branchiales de larve du Pleurodeles

waltlii et certaines donnees sur Kinnervation gustative, Trav. du lab. de re cherches biol. de I'Univ. de Madrid, 22:369-84. Evans, F. Gaynor. 1944. The morphological status of the modern Amphibia

among the Tetropoda, J. Morphol., 74:43-100. Ferraro, a., and Barrera, S. E. 1935. Posterior column fibers and their termination in Macacus rhesus, J. Comp. Neurol., 62:507-30. FoREL, A. 1877. Untersuchimgen iiber die Haubenregion u.s.w.. Arch. f. Psychiat.,

7:393-495. Fox, C. A. 1941. The mammillary peduncle and ventral tegmental nucleus in the

cat. J. Comp. Neurol. 75:411-25. Francis, E. T. B. 1934. The anatomy of the salamander. Oxford University Press.


Frey, Eugen. 1938. Studien liber die hypothalamische Opticuswurzel der Am phibien. II. Proteus anguineus, Proc. kon. Akad. Wetensch., Amsterdam,

41:1015-21. Fulton, J. F. 1943. Physiology of the nervous system. 2d ed. Oxford Univ. Press. Ganser, S. 1882. Vergleichend-anatomische Studien liber das Gehirn des Maul wurfs, Morphol. Jahrb., 7:.591-72o. Gaupp, E. 1899. Anatomie des Frosches, Abt. 2, Auf. 2. Braunschweig. Gehuchten, a. van. 1894. Contribution a I'etude du systeme nerveux des tele osteens, Cellule, 10:255-95. . 1895. Le faisceau longitudinal posterieur, Bull. Acad. roy. de med. de

Belgique, 9:1-40.


1897. La moelle epiniere des larves des batraciens (Salamandra maculosa),


Arch, de biol., 15:251 . 1897a. Le ganglion basal et la commissure habenulaire dans I'encephale de

la salamandre, Bull. Acad. roy. de Belgique, ser. 3, 34:38-07.


1900. Anatomic du systeme nerveux de Thomme. 2 vols. Louvain.



Geiringer, Martha. 1938. Die Beziehungen der basalen Optikuswurzel zur

Hypophyse und ihre Bedeutung fur den Farbwechsel der Amphibien, Anat.


Anz., 86: 202-7. Gesell, Robert, and Hansen, E. T. 1945. Anticholinesterase activity of acid as a

biological instrument of nervous integration. Am. J. Physiol., 144:120-63. GiLLiLAN, LoiB A. 1941. The connections of the basal optic root (posterior accessory

optic tract) and its nucleus in various mammals, J. Comp. Neurol., 74:367 408. Glasser, Otto (ed.). 1944/ Medical physics. Chicago: Year Book Publishers. Gregory, William K. 1943. Environment and locomotion in mammals. Nat.


Hist., 51:222-27. Griggs, L. 1910. Early stages in the development of the central nervous system of

Amblystoma punctatum, J. Morphol., 21:425-83. GuDDEN, B. VON. 1881. Mitteilung liber das Ganglion interpedunculare. Arch. f.


Psychiat., 11:424-27. Hamburger, V., and Keefe, E. L. 1944. The effects of peripheral factors on the

proliferation and differentiation in the spinal cord of chick embryos, J. Exper.


Zool., 96:223-42. Herrick, C. Judson. 1899. The cranial and first spinal nerves of Menidia: a contribution upon the nerve components of the bony fishes, J. Comp. Neurol., 9:

153-455. . 1900. A contribution upon the cranial nerves of the cod fish, il)id., 10:

265-310. . 1903. The organ and sense of taste in fishes, Bull. U.S. Fish Com., 22:237


(2.


1903a. On the morphological and physiological classification of the cu


taneous sense organs of fishes. Am. Nat., 37:313-18.


-. 1903fc. On the phylogeny and morphological position of the terminal buds


of fishes, J. Comp. Neurol., 13:121-38. . 1905. The central gustatory paths in the brains of bony fishes, ibid., 15:


375-456.



-. 1908. The morphological subdivision of the brain, ibid., 18:393-408.


. 1909. The nervus terminalis (nerve of Pinkus) in the frog, ibid., 19:175-90.


. 1910. The morphology of the forebrain in Amphibia and Reptilia, ibid.,

20:413-547.


1913. Brain, anatomy of the, in Reference handbook of the medical sciences.



2:274-342. 3d ed. New York: William Wood & Co.



BIBLIOGRAPHY 311

Herrick, C. Judson. U)13a. Some reflections on the origin and significance of tlie

cerebral cortex, J. Anim. Behavior, 3:222-36. 1914. The cereVieUum of Necturus and otlier urodcle Anipliihia, J. ('omp.


Neurol., 24: 1-29.


1914a. The medulla oblongata of larval Amblystoma, ibid., pp. 343-427.


1917. The internal structure of the midbrain and thalamus of Necturus,


ibid., 28:215-348.


19^20. Irreversible dift'erentiation and orthogenesis. Science, 51:()21-2o.


\ 1921. A sketch of the origin of the cerebral hemispheres, J. Comp. Neurol.



32:429-54. 1921a. The connections of the vomeronasal nerve, accessory oltactory


bulb and amygdala in Amphibia, ibid., 33:213-80.


. 1922. What are viscera? J. Anat., 56: 167-76.


. 1922a. Functional factors in the morphology of the forebrain of fishes,

Libro en honor de D. Santiago Ramon y Cajal, 1:143-204. Madrid. . 19226. Some factors in the development of the amphibian nervous system.



Anat. Rec, 23:291-305. , . t, u- . . 1924. Origin and evolution of the cerebellum. Arch. Neurol. & 1 sychiat.,


11:621-52. , ^ ^ ^

1924a. The amphibian forebrain. I. Amblystoma, external form, J. C.omp.



Neurol., 37:361-71. , „ . . ,, .


-. 19246. The amphibian forebrain. II. The olfactory bulb of Amblystoma,


ibid., pp. 273-396. ^. , XT . 1924c. Neurological foundations of animal behavior. New lork: Henry


Holt & Co. ^ ^ ^^ , . 1924(/. The nucleus olfactorius anterior of the opossum, J. Lomp. Neurol.,


37' 317—59.


1925. The amphibian forebrain. III. Tlie optic tracts and centers of Am


blystoma and the frog, ibid., 39:433-89. ' 1925a. Morphogenic factors in the differentiation of the nervous system,


Physiol. Rev., 5:112 30. . 19256. The innervation of palatal taste buds and teeth of Amblystoma, J.



Comp. Neurol., 38:389-97.


1926. Brains of rats and men. Chicago: University of Chicago Press.


1927. The amphibian forebrain. IV. The cerebral hemispheres of Ambly


stoma, J. Comp. Neurol., 43:231-325.


. 1929. Anatomical patterns and behavior patterns, Physiol. Zool., 2:439-48.


. 1930. The medulla oblongata of Necturus, J. Comp. Neurol., 50: 1-96.


. 1930a. Localization of function in the nervous system, Proc. Nat. Acad. Sc,


16:643-50,

1931. The amphibian forebrain. V. The olfactory bulb of Necturus, J.



Comp. Neurol., 53:55-69. . 1931a. An introduction to neurology. 5th ed. Philadelphia: W. B. Saunders


. 1933. The functions of the olfactory parts of the cerebral cortex, Proc. Nat.



Acad. Sc, 19:7-14.


. 1933a. Morphogenesis of the brain, J. Morphol., 54:233-58.


. 19336. The amphibian forebrain. VI. Necturus, J. Comp. Neurol., 58: 1-288.


1933c. The amphibian forebrain. VH. The architectural plan of the brain.



ibid., pp. 481-505. . mvsd. The evolution of cerebral localization patterns, Science, 78:439-44.


Herrick, C. Judson. 1933^. The amphibian forebrain. VHI. Cerebral hemispheres

and palhal primordia, J. Comp. Neurol., 58:737-59. . 1934. The amphibian forebrain. IX. Neuropil and other interstitial nervous

tissue, ibid., 59:93-116.


1934a. The amphibian forebrain. X. Localized functions and integrating


functions, ibid., pp. 239-66.


. 19346. The hypothalamus of Necturus, ibid., pp. 375-429.


. 1934c. The interpeduncular nucleus of the brain of Necturus, ibid., 60: 111


35. . 1934rf. The endocranial blood vascular system of Amblystoma, Ztschr. f.



mikr.-anat. Forsch., 36:540-44. . 1935. The membranous parts of the brain, meninges and their blood vessel'


in Amblystoma, J. Comp. Neurol., 61:297-346.


1935a. A topographic analysis of the thalamus and midbrain of Ambly


stoma, ibid., 62:239-61.


. 1936. Conduction pathways in the cerebral peduncle of Amblystoma, ibid.,

63:293-352.


1937. Development of the brain of Amblystoma in early functional stages.



ibid., 67:381-422.


1938. Development of the cerebrum of Amblystoma during early swimming


stages, ibid., 68:203-41.


1938a. Development of the brain of Amblystoma punctatum from early


swimming to feeding stages, ibid., 69:13-30.


-. 19386. The brains of Amblystoma punctatum and A. tigrinum in early


feeding stages, ibid., pp. 391-426. . 1939. Internal structure of the thalamus and midbrain of early feeding


larvae of Amblystoma, ibid., 70:89-135. . 1939a. The cerebrum of Amblystoma tigrinum in midlarval stages, ibid.



pp. 249-66. . 19396. Cerebral fiber tracts of Amblystoma tigrinum in midlarval stages,


ibid., 71:511-612.


. 1941. Development of the optic nerves of Amblystoma, ibid., 74:473-534.


-. 1941a. Optic and postoptic systems of fibers in the brain of Necturus, ibid..



75:487-544.


. 19416. The eyes and optic paths of the catfish, Ameiurus, ibid., 75:255-86.


1942. Optic and postoptic systems in the brain of Amblystoma tigrinum.



ibid., 77:191-353. . 1943. The cranial nerves: a review of fifty years, Denison Univ. Bull., J.



Sc. Lab., 38:41-51.


. 1944. The incentives of science, Scient. Monthly, 58, 462-66.


1944a. Apparatus of optic and visceral correlation in the brain of Ambly


stoma, J. Comp. Psychol., 37:97-105. . 19446. The fasciculus solitarius and its connections in amphibians and


fishes, J. Comp. Neurol., 81:307-31. . 1947. The proprioceptive nervous system, J. Nerv. & Ment. Dis., 106:


355-58.


-. 1948. George Ellett Coghill, naturalist and philosopher. To be published by


University of Chicago Press. Hoagland, H. 1933. Electrical responses from the lateral-line nerves ot catfish.


I, J. Gen. Physiol., 16: 695-714. . 19336. Quantitative analysis of responses from lateral-line nerves of fishes.



II, ibid., pp. 715-31.



BIBLIOGRAPHY 313

Holmgren, Nils. 1922. Points of view concerning forebrain morphology in lower

vertebrates, J. Comp. Neurol., 34:391-459.


. 1946. On two embryos of Myxine glutinosa, Acta Zool., 27:1-90.


Holmgren, Nils, and van der Horst, C.J. 1925. Contribution to the morphology

of the brain of Ceratodus, Acta Zool., 6:59-165. Holtfreter, J. 1945. Differential inhibition of growth and differentiation by

mechanical and chemical means, Anat. Rec, 93:59-74. Hooker, Davenport. 1944. The origin of overt behavior. Ann Arbor: University

of Michigan Press. Howell, A. Brazier. 1944, Speed in animals, their specialization for running and

leaping. Chicago: University of Chicago Press. HuBER, G. C, and Crosby, Elizabeth C. 1933. A phylogenetic consideration

of the optic tectum. Proc. Nat. Acad. Sc, 19: 15-22. . 1934. The influences of afferent paths on the cytoarchitectonic structure of

the submammalian optic tectum, Psychiat. en Neurol. Bl, pp. 459-74. . 1943. A comparison of the mammalian and reptilian tecta, J. Comp.


Neurol., 78:133-68. Humphrey, Tryphena. 1944. Primitive neurons in the embryonic human central

nervous system, J. Comp. Neurol., 81: 1-45. Jacobsohn-Lask. 1924. Die Kreuzung der Nervenbahnen und die bilaterale

Symmetric des tierschen Korpers. Berlin: S. Karger. Jansen, Jan. 1930. The brain of Myxine glutinosa, J. Comp. Neurol., 49:359-507. Jeserich, Marguerite W. 1945. The nuclear pattern and the fiber connections of

certain non-cortical areas of the telencephalon of the mink (Mustek vison),

J. Comp. Neurol., 83:173-211. Johnston, J. B. 1901. The brain of Acipenser, Zool. Jahrb., 15:59-260.


. 1902. The brain of Petromyzon, J. Comp. Neurol., 12:1-86.


. 1916. Evidence of a motor pallium in the forebrain of reptiles, ibid., 26:

475-79 Jones, D. S. 1945. The origin of the ciliary ganglia in the chick embryo, Anat. Rec,

92:441-47. Kabat, H. 1936. Electrical stimulation of points in the forebrain and midbrain: the

resultant alterations in respiration, J. Comp. Neurol., 64:187-208. Kato, G. 1934. The microphysiology of nerve. Tokyo: Maruzen Co., Ltd. Kingsbury, B. F., 1895. On the brain of Necturus maculatus, J. Comp. Neurol.,

5:139-205. . 1930. The developmental significance of the floor-plate of the brain and

spinal cord, ibid., 50:177-207. KoDAMA, S. 1929. Ueber die sogenannten Basalganglien. B. Ueber die Faserver bindungen den Basalganglien u.s.w., Schweiz. Arch. f. Neurol, u. Psychiat.,

23:179-265. KosTiR, W. J. 1924. An analysis of the cranial ganglia of an embryo salamander,

Amblystoma jeffersonianum (Green), Ohio J. Sc, 24:230-63. Kreht, Hans. 1930. Ueber die Faserzuge im Zentralnervensystem von Salamandra

maculosa L., Ztschr. f. mikr.-anat. Forsch., 23:239-320. . 1931. Ueber die Faserzuge im Zentralnervensystem von Proteus anguineus

Laur., ibid., 25:376-427. Kuhlenbeck, H. 1921. Zur Morphologic des Urodelenvorderhirns, Jenaische

Ztschr. f. Naturwissensch., 57, N.F., 50:463-90.


. 1922. Zur Morphologic des Gymnophionengehirns, ibid., 58:453-84.


KuPFFER, C. VON. 1906. Die Morphogenie des Zentralnervensystems, Hertwig's

Handb. f. Entw. d. Wirbeltiere, 2, Teil 3, 1-272. Jena.



Landacre, F. L. 1921. The fate of the neural crest in the head of the urodeles, J. Comp. Neurol., 33:1-43.


. 1926. The primitive lines of Amblystoma jettersonianum, ibid.. 40:471-9.5.


Larsell, O. 1920. The cerebellum of Amblystoma, J. Comp. Neurol., 31:259-82.


. 1923. The cerebellum of the frog, ibid., 36:89-112.


. 1924. The nucleus isthmi of the frog, ibid., 36:309-22.


. 1925. The development of the cerebellum in the frog (Hyla regilla) in relation to the vestibular and lateral-line systems, ibid., 39:249-89.


. 1929. The nerve terminations in the lateral-line organs of Amblystoma,

ibid., 48:465-70.


. 1931. The cerebellum of Triturus torosus, ibid., 53: 1-54.


. 1932. The development of the cerebellum in Amblystoma, ibid., 54:357


435.


1934. The differentiation of the peripheral and central acoustic apparatus


in the frog, ibid., 60:473-527.


1934a. Morphogenesis and evolution of the cerebellum, Arch. Neurol. &


Psychiat., 31:373-95.


. 1937. The cerebellum: a review and interpretation, ibid., 38:580-607.


. 1945. Comparative neurology and present knowledge of the cerebellum,

Bull. Minnesota M. Foundation, 5:73-85.


1947. The cerebellum of myxinoids and petromyzonts, including develop


mental stages in the lampreys, J. Comp. Neurol., 86:395-445.


1947a. The nucleus of the IV nerve in petromyzonts, ibid., pp. 447


Lashley, K. S. 1934. The mechanism of vision. VII, J. Comp. Neurol., 59:341-73.


. 1934a. The mechanism of vision. VIII, ibid., 60:57-79.


. 1941. Thalamo-cortical connections of the rat's brain, ibid., 75:67-121.


Lashley, K. S., and Clark, George. 1946. The cytoarchitecture of the cerebral

cortex of Ateles: a critical examination of architectonic studies, J. Comp. Neurol.,

85: 223-305. Liggett, J. R. 1928. An experimental study of the olfactory sensitivity of the white

rat. Genet. Psychol. Monog., 3: 1-64. Lillie, Ralph S. 1945. General biology and philosophy of organism. Chicago:

University of Chicago Press. Loo, Y. T. 1931. The forebrain of the opossum, Didelphis virginiana. II, J. Comp.


Neurol., 52:1-148. McKiBBEN, Paul S. 1911. The nervus terminalis in urodele Amphibia, J. Comp.


Neurol., 21:261-309.


. 1913. The eye-muscle nerves in Necturus, ibid., 23:153-72.


Magoun, H. W. 1944. Bulbar inhibition and facilitation of motor activity. Science,

100:549-50. Marburg, O. 1944. The structure and fiber connections of the human habenula.


J. Comp. Neurol., 80:211-33. Mettler, Fred A. 1945. Fiber connections of the corpus striatum of the monkey

and baboon, J. Comp. Neurol., 82:169-204. Mettler, Fred A., and Mettler, Cecilia C. 1941. Role of the neostriatum.


Am. J. Physiol., 133:594-601. MoNCRiEFF, R. W. 1944. The chemical senses. London: Leonard Hill, Ltd. Murphy, J. P., and Gellhorn, E. 1945. Multiplicity of representation versus punctate localization in the motor cortex. Arch. Neurol. & Psychiat., 54:256-273. Neimanis, Emma. 1931. Individual variation of form of the brain of Triton cristatus

Laur. and its relation to the specific variation of the brain of Urodela, Bull. Soc.


biol. de Lettonie, 2:67-92.



BIBLIOGRAPHY 315

Noble, G. Kingsley. 1931. The biology of the Amphibia. New York: McGrawHill Book Co. O'Neill, H. M. 1898. Hirn- und Ruckenmarkshiillen bei Amphibien, Morphol.


Arb., Sehwalbe, 8:47-6-1. OsBORN, H. F. 1888. A contribution to the internal structure of the amphibian

brain, J. Morphol., 2:51-96. Palay, S. L. 1944. The histology of the meninges of the toad (Bufo), Anat. Rec,

88:257-70. . 1945. Neurosecretion. VII. The preoptico-hypophysial pathway m fishes.


J. Comp. Neurol., 82:129-43. Papez, James W. 1936. Evolution of the medial geniculate body. J. Comp. Neurol.,

64:41-61. . 1944. Structures and mechanisms underlying the cerebral functions. Am. J.


Psychol., 57:291-316. Parker, G. H. 1912. The relation of smell, taste, and the common chemical sense

in vertebrates, J. Philadelphia Acad. Nat. Sc, ser. 2, 15: 219-34. . 1918. A critical survey of the sense of hearing in fishes, Proc. Am. Phil. Soc,

57:1-30. . 1922. Smell, taste and allied senses in the vertebrates, Philadelphia: J. B.


Lippincott Co. Parker, G. H., and van Heusen, A. P. 1917. The reception of mechanical stimuli

by the skin, lateral-line organs and ears in fishes, especially in Amiurus, Am. J.


Physiol, 44:463-89. Pearse, a. S. (ed.). 1936. Zoological names: a list of phyla, classes and orders.


Durham, N.C.: Duke University Press. Pearson, A. A. 1945. Further observations on the intramedullary sensory type

neurons along the hypoglossal nerve, J. Comp. Neurol., 82: 93-100. Piatt, Jean. 1945. Origin of the mesencephalic V root cells in Amblystoma, J. Comp.


Neurol., 82:35-53. . 1946. The influence of the peripheral field on the development of the mesencephalic V nucleus in Amblystoma, J. Exper. Zool., 102:109-41. PoLYAK, S. L. 1941. The retina. Chicago: University of Chicago Press. Ramon y Cajal, P. 1922. El cerebro de los batracios. Libro en Honor de D. S.


Ramon y Cajal, 1:13-150. Madrid. Ramon y Cajal, S. 1911. Histologic du systeme nerveux de I'homme et des verte bres. Vol. 2. Paris. Rasmussen, G. L. 1946. The olivary peduncle and other fiber projections of the

superior olivary complex, J. Comp. Neurol., 84:141-219. Reiser, Oliver L. 1946. The world sensorium. New York: Avalon Press. RoTHiG, Paul. 1911. Zellanordnungen und Faserzuge im Vorderhirn von Siren

lacertina, Abh. kgl. Preuss. Akad. Wissensch., Anhang, pp. 1-23. . 1911a. Beitrage zum Studium des Zentralnervensysteins der Wirbeltiere.


IV. Die marklialtigen Faserzuge im Vorderhirn von Necturus maculatus, Arch. f.


Anat. (u. Physiol.), pp. 48-56. . 1912. Beitrage V. Die Zellanordnungen im Vorderhirn der Amphibien,

Verhandl. Kon. Akad. Wetensch., Amsterdam, sec. 2, Deel 17, pp. 1-23. 1923. Beitrage VIII. Ueber das Zwischenhirn der Amphibien, Arch. f.


mikr. anat., 98:616-45. . 1924. Beitrage IX. Ueber die Faserzuge im Zwischenhirn der Uro delen, Ztzchr. f. mikr .-anat. Forsch., 1:5-40. . 1927. Beitrage XI. Ueber die Faserzuge im Mittelhirn, Kleinhirn und

der Medulla oblongata der Urodelen und Anuren, ibid., 10:381-472.


RoMER, A. S. 1946. The early evolution of fishes, Quart. Rev. Biol., 21:33-69.


RooFE, P. G. 1935. The endocranial blood vessels of Amblystoma tigrinum, J. Comp. Neurol., 61:257-93.


. 1937. The morphology of the hypophysis of Amblystoma. J. Morphol.,

61:485-94.


. 1938. The blood vascular system of the hypophysis of Amblystoma tigrinum,

J. Comp. Neurol., 69:249-54.


RuDEBECK, BiRGER. 1945. Contributions to forebrain morphology in Dipnoi, ActaZool., 26:9-156.


ScHARRER, E. 1932. Experiments on the function of the lateral-line organs in the larvae of Amblystoma punetatum, J. Exper. Zool., 61:109-14.


ScHARRER, E., and ScHARRER, B. 1940. Secretory cells within the hypothalamus, A. Research Nerv. & Ment. Dis. Proc, 20:170-94.


ScHARRER, E.; Smith, S. W.; and Palay, S. L. 1947. Chemical sense and taste in the fishes, Prionotus and Trichogaster, J. Comp. Neurol., 86:183-98.


ScHRiEVER, H. 1935. Aktionspotentiale des N. lateralis bei Reizung der Seitenorganie von Fischen, Arch. f. d. ges. Physiol., 235:771-84.


Shanklin, W. M. 1933. The comparative neurology of the nucleus opticus tegmenti with special reference to Chameleon vulgaris. Acta Zool., 14: 163-84.


Sheldon, R. E. 1909. The reactions of the dogfish to chemical stimuli, J. Comp. Neurol., 19:273-311.


Sherrington, C. S. 1906. The integrative action of the nervous system. New York: Charles Scribner's Sons.


SoDERBERG, Gertie. 1922. Contributions to the forebrain morphology in amphibians, Acta Zool., 3:65-121.


SosA, Julio Maria. 1945. Collateral nerve fibers within septum dorsale of the spinal cord and medulla oblongata and their connections, J. Comp. Neurol., 83: 157-71.


Speidel, C. C. 1946. Prolonged histories of vagus nerve regeneration patterns, sterile distal stumps, and sheath cell outgrowths (abstr.), Anat. Rec, 94:499.


Sperry, R. W. 1943. Effect of 180 degree rotation of the retinal field on visuomotor coordination, J. Exper. Zool., 92:263-79.


. 1944. Optic nerve regeneration with return of vision in anurans, J. Neuro physiol., 7:57-70.


. 1945. Restoration of vision after crossing of optic nerves and after contralateral transplantation of eye, ibid., 8: 15-28.


. 1945a. Centripetal regeneration of the 8th cranial nerve root with systematic restoration of vestibular reflexes, Am. J. Physiol., 144:735-41.


. 19456. The problem of central nervous reorganization after nerve regeneration and muscle transposition, Quart. Rev. Biol., 20:311-69.


Stensio, E. a. 1927. The Downtonian and Devonian vertebrates of Spitsbergen. I. Family Cephalaspidae. Det Norske Vidensk.-Akad. i Oslo.


Stone, L. S. 1922. Experiments on the development of the cranial ganglia and the lateral-line sense organs in Amblystoma punetatum, J. Exper. Zool., 35:421-96.


. 1926. Further experiments on the extirpation and transplantation of mesec toderm in Amblystoma punetatum, ibid., 44: 95-131.


1944. Functional polarization in retinal development and its reestablish


ment in regenerating retinae of rotated grafted eyes, Proc. Soc. Exper. Biol.


Med., 57:13-14. Stone, L. S., and Ellison, F. S. 1945. Return of vision in eyes exchanged between

adult salamanders of different species, J. Exper. Zool., 100:217-27. Stone, L. S., and Zaur, I. S. 1940. Reimplantation and transplantation of adult

eyes in the salamander (Triturus viridescens) with return of vision, J. Exper.


ZooL, 85:243-69.



BIBLIOGRAPHY 317

Stroer, W. F. H. 1939. Ueber den Faserverlauf in den optischen Bahnen bei

Amphibien, Proc. kon. Akad. Wetensch., Amsterdam, 42:649-56. . 1939a. Zur vergleichenden Anatomie des primaren optischen Systems bei

Wirbeltieren, Ztschr. f. Anat. u. Entwcklngsgesch., 110:301-21. . 1940. Das optische System beim Wassermolch (Triturus taeniatus), Acta

neerl. morphol., 3:178-95. Strong, O. S. 1895. The cranial nerves of the Amphibia, J. Morphol., 10: 101-230. SuMi, R. 1926. Ueber die Morphogenese des Gehirns von Hynobius nebulosus,

Folia Anat. Japon., 4: 171-270. , 1926a. Ueber die Sulci und Eminentiae des Hirnventrikels von Diemictylus

pyrrhogaster, ibid., pp. 375-88. Taylor, A. C. 1944. Development of the innervation pattern in the limb bud of the

frog, Anat. Rec, 87:379-413. Thompson, d'Arcy Wentworth. 1944. On growth and form. New York: Mac millan Co. Thompson, Elizabeth L. 1942. The dorsal longitudinal fasciculus in Didelphis vir giniana, J. Comp. Neurol., 76:239-81. TuGE, H. 1932. Somatic motor mechanisms in the midbrain and medulla oblongata

of Chrysemys elegans (Wied), J. Comp. Neurol., 55:185-271. Wallenberg, A. 1931. B,eitrage zur vergleichenden Anatomie des Hirnstammes,

Deutsche Ztschr. f. Nervenh., Vols. 117, 118, 119, pp. 677-98. Warren, John. 1905. The development of the paraphysis and the pineal regipn in

Necturus maculatus. Am. J. Anat., 5: 1-27. Weiss, Paul. 1936. Selectivity controlling the central -peripheral relations in the

nervous system, Biol. Rev., 11:494-531.


. 1939. Principles of development. New York: Henry Holt & Co.


. 1941. Self-differentiation of the basic patterns of coordination, Comp.


Psychol. Monog., 17, No. 4, 1-96. Whitman, C. O. 1892. The metamerism of Clepsine, Festschr. f. Rudolf Leuck harts, pp. 384-95. Leipzig: Wilhelm Engelmann. . 1899. Animal behavior. Biological lectures from the Marine Biological

Laboratory, Woods Hole, for 1898, pp. 285-338. Boston. Woodburne, R. T. 1936. A phylogenetic consideration of the primary and secondary centers and connections of the trigeminal complex in a series of vertebrates,

J. Comp. Neurol., 65:403-501. . 1939. Certain phylogenetic anatomical relations of localizing significance

for the mammalian central nervous system, ibid., 71:215-57. Yntema, C. L. 1937. An experimental study of the origin of the cells which constitute the Vllth and Vlllth cranial ganglia and ner-ves in the embryo of Am blystoma punctatum, J. Exper. Zool., 75:75-101. . 1943. An experimental study on the origin of the sensory neurones and

sheath cells of the IXth and Xth cranial nerves in Amblystoma punctatum,

ibid., 92:93-119. . 1943a. Deficient efferent innervation of the extremities following removal

of neural crest in Amblystoma, ibid., 94:319-49. Youngstrom, K. a. 1938. Studies on the developing behavior of Anura, J. Comp.


Neurol., 68:351-79.


. 1940. A primary and a secondary somatic motor innervation in Ambly • stoma, ibid., 73:139-51. . 1944. Intramedullary sensory type ganglion cells in the spinal cord of

human embryos, ibid., 81:47-53.



ILLUSTRATIONS


ILLUSTRATIONS

GENERAL STATEMENT

All of the 113 figures are of adult or late larval Amblystoma tigrinum, except figures 86B, 86C, 111, 112, and 113 of Necturus and figure 107 of Rana pipiens. The indicated magnifications of the drawings show wide diversity in actual sizes of the specimens due to two things: in the first place, the fresh specimens vary greatly in size, and, in the second place, shrinkage during preparation may amount to as much as one-fourth of the original linear dimensions. Figures 2-85 are original drawings, and figures 1 and 86-113 are selected from previous publications, with minor alterations in some of them.


The internal structure of the brain of Necturus has been illustrated by sections drawn at close intervals in the three conventional planes ('30, '336) and by many drawings of detail in several other publications. The general plan of the brain of Amblystoma is so similar to that of Necturus that comparisons are readily made. The sections of Amblystoma have been under investigation for nearly forty years, and the exigencies of the study and of publication have required concentration upon particular topics rather than description of the brain as a whole, and no comprehensive atlas of sections has been prepared. It is the aim of this book to supply a general view, but, unfortunately, the figures necessary to illustrate it are scattered in many publications. Some figures of especial value for general orientation are included here, and references are given to others in the literature.


Three series of Weigert sections of adult A. tigrinum, prepared by the late Dr. P. S. McKibben in 1910, were selected as standards of reference. These are his numbers IIC, transverse. IC, horizontal, and C, sagittal. All were fixed in a formalinbichromate mixture, imbedded in paraffin, cut serially at 12 n, and stained on the slides. These show minimum distortion of form, excellent histological preservation, and brilliant stain of myelinated fibers, with decolorization arrested at a stage which leaves all cell bodies clearly stained. The transverse series was chosen as the type specimen, and this has been quite fully illustrated, though these drawings were published at different times. To assist the reader who may wish to assemble an atlas of this specimen, a list of all available figures of it was published ('35a, p. 241), with serial numbers of the sections pictured, and to this list the following numbers may now be added:

Section

98 Olfactory bulb. '246, fig. 5

570. . . . Nucleus posterior tecti. '36, fig. 15

595 ... . Cerebellum. Figure 91

635 .... Posterior border of V nerve roots. Figure 90

730 .... Level of IX nerve roots. Figure 89

929 .... Anterior root of first spinal nerve. Figure 88

956. . . . Calamus scriptorius. '446, fig. 4

975 .... Commissura infima. Figure 87

990 .... Nuclei of dorsal funiculi. '446, fig. 6

Figure 2C of this work is the median section of the middle part of the brain stem, reconstructed from this series of sections; and figures 1-4 of the paper of 1935 are similar reconstructions, showing chorioid plexuses and blood vessels.


A paper model of this transverse series was prepared, X 75, each section being drawn on cardboard seventy-five times the thickness of the section and then cut out along external and ventricular surfaces. These sheets, when properly stacked, show the external and ventricular configuration, and upon them details of internal structure were drawn in colored inks. The relevant data about this specimen were published ('35a) in connection with the preparation of the diagram of the median section, shown here as figure 2C (cf. fig. IB). Scales accompanying the diagram give the section numbers, so that any section figured can be accurately located on the projection. This outline has been used as the basis for many diagrams of internal structure in this work and previous papers. Figures 2A and '2B are similar diagrams, made from the series of horizontal sections illustrated in figures 25-36. A wax model was made from the sagittal Weigert series C (figs. lA and 85).


Two series of transverse sections of the adult forebrain based mainly on Golgi sections were published in 1927; six of these are here reproduced as figures 95-100. Six sagittal sections of the middle part of the adult brain stem by Rogers' reduced silver method were published in 1936, figures 17-22; and three of these are shown here as figures 102, 103, and 104. Seven pictures of the chiasma region from the sagittal Weigert series C are in figure 11 of 1941. A series of twelve horizontal Cajal sections of the adult follows in this book (figs. 25-36). Serial sections of larvae of 38 mm., prepared by Cajal's reduced silver method, have been illustrated (horizontal, '39b, figs. 2-20; transverse, '14a, figs. 4-14, and '39a, figs. 2-14). Somedetails of the development of the external form of the brain and the internal structures are reported in the papers of 1937-41. For meninges and blood vessels see pages 24-27, my papers ('34c^ and '35), Roofe ('35), and Dempster ('30).


FIGURES AND DESCRIPTIONS

Figure lA, B, and C'.^Lateral and median views of the brain of adult A. tigrinum, reproduced from '24a, figures 1, 2, and 3.


A. — Drawn from a wax model made from sagittal Weigert sections, no. C, collected at Chicago, 111. X 15. The lateral wall of the cerebral hemisphere has been cut away to open the lateral ventricle; compare figure 85 drawn from the same model.


B. — Median section of a specimen from Colorado. X 10. The dissection was prepared by Dr. P. S. McKibben and drawn by Katharine Hill. Brains of several specimens were exposed, fixed in situ in formalin-Zenker, then removed from the head, washed in water, and hardened in graded alcohols before being cut in the mid-sagittal plane. The shrinkage in alcohol accentuates the ventricular sculpturing.


C. — Key drawing to accompany figure IB.


Figure 2 A, B, and C. — Three drawings of the median section of the brain stem of the adult prepared by graphic reconstruction from sections.


A. — This median section is reconstructed from the series of horizontal Cajal sections ilhistrated in figures 25-36. X 23. Compare figure 2C made from a specimen differently prepared.


The shape of the two specimens is somewhat different, owing chiefly to slight dorsoventral compression of the midbrain and other distortions of the Cajal specimen. Despite these defects, the correspondence of the two median sections is fairly close. The figures of the two specimens are drawn on the same plan, with the important exception that in the drawing of 2C tlie ventricular sulci are projected upon the median plane, while in 2A and B the outhiies of the chief cellular areas are thus projected. In general, the sulci mark tlie boundaries of these areas, but this correspondence is not exact, and there is great individual variation in both these features. In the posterior part of the tegmentum isthmi a vertical dotted line marks, somewhat arbitrarily, tlie boundary between its central nucleus and a posterior sector composed of larger cells and transitional to the large-celled component of the trigeminal tegmentum.


Two drawings based on this reconstruction are shown, with emphasis on different features. In figure 2A the larger subdivisions of this part of the brain stem are demarcated so as to assist


ILLUSTRATIONS


the reader in comparing the amphibian topography with the conventional mammalian analysis. Median structures, that is, the cut surfaces of the section, are outlined with full lines, paramedian structures in dotted lines (compare the gross section, fig. IB; Necturus, '17, fig. 63; and McKibben, '11, fig. 5). Heavy broken lines mark the boundaries of the larger divisions as here defined, and dot-and-dash lines mark the subdivisions of the diencephalon.


B. — The same outline, bearing names of some of the smaller subdivisions. The scales at right and left indicate the approximate planes of the sections shown in figures 25-36. These levels are not exact, for unequal shrinkage during preparation produced some irregularities which have been smoothed in the diagram. The scales at top and bottom show section numbers of the type specimen, no. IIC, from which figure 2C was drawn, as indicated on that figure. Here again the correspondence is not exact, though sufficiently close to facilitate orientation of the published transverse sections with reference to the two median sections.


C- — This is the median section of the type specimen, IIC (p. 321), copied from 1935a, figure 1, with some changes in the lettering. X 30.


Figure 3. — Outline drawing of the dorsal surface of the adult brain. X 10. The paraphysis and chorioid plexus of the fourth ventricle have been removed. The courses of the .sensory fibers of the trigeminal and dorsal spinal nerve roots and of the spino-bulbar, spino-cerebellar, and spinal lemniscus secondary tracts are diagrammatically indicated. The general bulbar lemniscus {Im. of the other figures) arises from the entire sensory zone of the medulla oblongata and ascends approximately parallel with the spinal lemniscus.


Figures 4 and 5. — Two diagrams illustrating the extent of the sensory and motor zones as seen in horizontal sections of the adult brain (chap. v). X 15. The outlines are from a series of Cajal sections cut in a slightly different plane from those illustrated in figures 25-36, with the anterior end more dorsal.


Fig. 4. — This passes through the hippocampal commissure and shows on the left side the extent of the motor zone, as here defined, by oblique hatching and the sensory zone in the olfactory area by hatching in the reverse direction. The plane of section is about the same as that of figures 28 anteriorly and 27 posteriorly. It passes below the tuberculum posterius at the cerebral flexure, above which the motor zone of the peduncle and isthmus is continuous between the medulla oblongata and the ventral thalanms (fig. 30).


Fig. 5. — This shows the extent of the sensory zone at a more dorsal level, the plane being about that of figures 35 anteriorly and 34 posteriorly. The undifferentiated anterior olfactory nucleus encircles the base of the olfactory bulb, and through it passes the very large fasciculus postolfactorius (cf. fig. 105), from wliich tlie various olfactory tracts of the liemisphere are distributed (p. 55; '27, p. 283).


Figure 6. — Selected examples of long pathways of conduction leading toward the skeletal musculature of the trunk and limbs, seen as projected upon the lateral aspect of the brain. X 10. The only afferent systems drawn are the olfactory and optic, and from these receptive fields only a few lines of descending conduction are indicated as typical representatives of through fore-and-aft transmission.


Figure 7. — Diagram of the central courses of the sensory components of cranial nerves V to X of larval Amblystoma seen as projected upon the lateral surface of the medulla oblongata. X 36. The drawing is based on figure 3 of 1914. The general cutaneous component is drawn in dashed lines, the vestibular in thick unbroken lines, the visceral-gustatory in dotted lines, the three lateral-line VII roots in thick dash lines, the two lateral-line X roots in dot-and-dash lines, and correlation tracts a and b in thin continuous lines. Some fibers of the ascending roots of the general cutaneous and vestibular systems decussate in the cerebellum, and some visceral root fibers of the f. solitarius decussate in the commissura infima at the commissural nucleus of Cajal.


Figure S.— Diagram of the central connections of the visceral-gustatory system seen as projected upon the lateral surface of the adult brain. X 10. A probable direct connection from the superior visceral-gustatory nucleus (nuc.i'is.ti.) to the hypothalamus is indicated by the dotted line.


Figure 9. — Diagrammatic transverse section of the larval medulla oblongata near the level of the IX nerve roots, showing four types of neurons of the sensory zone. X 100. Drawn from preparations illustrated in 1914. Figure 89 shows a section of the adult brain in about the same plane. The arrangement of the fascicles of nerve root fibers is indicated. Neuron 1 is in synaptic relation with all components of these cranial nerves. Neuron 2 makes its chief connection with visceral-gustatory fibers of the f. solitarius and less intimate connection with vestibular and trigeminal fibers. Neuron 3 connects only with root fibers of the trigeminus. Neuron 4 connects with the trigeminus and also with the reticular formation and motor zone. Similar elements have been seen to connect also with fascicles of the VIII and other nerve roots. Axons of all these types decussate in the ventral commissure and may ascend in the general bulbar lemniscus. The axon of neuron 2 divides, one branch ascending in the secondary visceral tract {tr.r.a.) of the same side and the other crossing to the opposite side. Some similar neurons connect only with the f. solitarius and have unbranched axons entering tr.r.a. only. Numberless permutations of the various types of connection here shown have been observed.



Figure 10. — Diagram of the chief aflFerent connections of the body of the cerebellum and of the brachium conjunctivum seen as projected on the median section of the brain (chaps. iv and xii). X 18. The more lateral vestibular connections are not drawn. The outline is that of figure 2C, and dotted lines mark ventricular sulci and the boundaries of the chief subdivisions of the brain wall.


Figure 11. — Diagram of the chief afferent tracts to the tectum (pp. 48, 220).


Figure 12. — Diagram of the chief efferent tracts from the tectum. Many shorter connections are omitted (p. 223).


Figure IS. — Diagram of the connections of the mesencephalic nucleus of the V nerve and of some tecto-bulbar and tegmento-bulbar tracts probably concerned with feeding reflexes (p. 140).


Figure H.—DiagTaxn of the chief afferent tracts to the dorsal thalamus and of some other connections of the optic tracts (pp. 49, 236).


Figure 15. — Diagram of the chief efferent tracts from the dorsal thalamus (pp. 49, 237) and pretectal nucleus (p. 39).


Figure 16. — Diagram of the chief afferent tracts to the ventral thalamus (p. 239).


Figure 17.- — Diagram of the chief efferent tracts from the ventral thalamus (p. 240).


Figure 18. — Diagram of the chief connections of the "peduncle" (nucleus of the tuberculum posterius, pp. 50, 217). Many shorter connections are omitted. For the connections of the ventrolateral neuropil see chapter iii and figure 23.


Figure 19. — Diagram of the diencephalic connections of the amygdala (pp. 52, 248) and of some of the connections of the interpeduncular nucleus (chap. xiv).


Figiire 20. — Diagram of the connections of the habcnula. In the stria medullaris thalami the components are arranged in the fore-and-aft order in which they ascend, as also are the decussations in the habenular commissure (compare the horizontal sections, figs. 25-36). Efferent fibers from the habenida enter three tracts (fr./iab.t.,f.refr., and tr.hab.th.). The afferent fibers are numbered as in the analysis of the stria medullaris in chapter xviii:

1. Tr. olfacto-habenularis medialis

2. Tr. olfacto-habenularis lateralis

3. Tr. olfacto-habenularis anterior, ventral division

4. Tr. olfacto-habenularis anterior, dorsal division

5. Tr. cortico-habenularis lateralis

6. Tr. amygdalo-habenularis

7. Tr. septo-habenularis

8. Tr. cortico-habenularis medialis

9. Tr. cortico-thalamicus medialis

10. Tr. olfacto-thalamicus

11. Tr. thalamo-habenularis

12. Tr. tecto-habenularis

Not shown in this figure is the tr. pretecto-habenularis and a probable tr. strio-habenularis.


Figure 21. — Diagram of the chief afferent connections of the tegmentum isthmi (p. 179). Most of the tracts here indicated end in both the isthniic and the trigeminal tegmentum. For the connections of the interpeduncular nucleus and its neuropil see chapter xiv.


Figure 22. — Diagram of the most direct connections between the retina and the peduncle. Optic tracts and efferent paths from the peduncle are drawn in full lines, internuncial connections in broken lines, and thick myelinated fibers in heavier lines (chap. xvi).


Figure 23.- — The chief nonvisual afferent connections of the ventrolateral peduncular neuropil (chap, iii) seen as projected upon the median section. Most of these fibers are unmyelinated, of thin or medium size. The visceral-gustatory system has more fibers than any of the others.


Figure 2!^. — Diagrammatic thick transverse section through the middle of the optic tectum and the oculomotor nucleus, illustrating typical connections between the tectum and the cerebral peduncle. The gray is outlined by a broken line, and the ventral border of the tectum by dotted lines. On the right side a typical neuron of the peduncle is drawn, with axon entering the ventral tegmental fascicles (f.r.t.). Four outlying cell bodies are seen at the border of the ventrolateral peduncular neuropil (a.rl.p.). On the left side is a neuron of the oculomotor nucleus, the dendrites of which connect with terminals of the basal optic tract in the ventrolateral neuropil. Compare figure 93.


Figures 25 to J6.— These semidiagrammatic drawings are made from sections selected from a horizontal series of the adult brain prepared by P. S. McKibben, Cajal's reduced silver method after fixation in alcohol. X 35. For approximiite planes of section see figure 2B. The sections are exactly horizontal right and left.


In this specimen the cell bodies are not blackened, and no dendrites are visible. All nuclei of cells are stained, and the gray pattern is clearly shown. There is scanty impregnation of the neuropil. In the cerebral hemispheres only the thickest axons are impregnated; elsewhere the thick and medium fibers are brilliantly differentiated. In some places where the boundaries of fascicles and tracts are obscure, interpretation has been aided by comparison with other specimens by methods of Cajal, Rogers, Golgi, and Weigert.


Horizontal sections are not so favorable for analysis of the tegmental fascicles (chap, xx) as are those cut in transverse and sagittal planes, and it is difficult to follow individual bundles as they recurve around the tuberculum posterius in the peduncle and tegmentum; but, by comparison with sections prepared by other methods and cut in various planes, the courses of most of the tracts and of the tegmental fascicles of groups (1) to (10) can be followed. The limits of the numbered groups of fascicles are not always clear, but their identification by number on the drawings is believed to be substantially correct.



Fig. 25. — Through the ventral part of the anterior commissure ridge and the dorsal border of the chiasma ridge. Only a few of the thicker fibers of the medial forebrain bundle are impregnated. These thread their way through the decussating fascicles of the chiasma ridge and spread out in the ventral part of the hypothalamus. The dense neuropil of these regions is not impregnated. All the optic fibers decussate ventrally of this level, except the most dorsal fibers of the axial tract (tr.op.ax.). The lightly stippled area in the postoptic commissure, marked tr.th.h.d.c, contains thin fibers from the dorsal thalamus, which decussate more ventrally (fig. 2C, tr.th.h.d.c). Farther dorsally (fig. 26) these fibers, after crossing, separate into fascicles A and B (p. 299).


Fig. 26. — Ten sections more dorsally, the section passes through the dorsal fascicles of the medial forebrain bundle, tr. olfacto-peduncularis, and above the chiasma ridge. The mixed system of thalamo-hypothalamic and tegmental fibers (fig. 25, tr.th.h.d.c.) has separated into superficial (.4) and deep {B) tracts for the tegmentum, as described on page 299. The course of the superficial tract (A) can be followed in figures 27-34; most of the deeper fibers (B) join the dorsal tegmental fascicles of group (8).


Fig. 27.— This cuts the decussation of the dorsal fascicles of the lateral forebrain bundle in the anterior commissure and the motor roots and nucleus of the V cranial nerve. The fibers descending from the dorsal thalamus to the postoptic commissure {tr.th.h.d.c.) lie deep in the alba. The same fibers after decussation appear in the superficial tract, A, and in the deeper fascicles of group (8). Rostrally of the V roots, most of the thick fibers of the ventral commissure belong to tr. tegmento-bulbaris cruciatus. Among these the f. longitudinalis medialis is assembling from fascicles of groups (4), (5), and (6) (p. 281). Medially and dorsally of these are finer fibers of the dorsal division of tr. interpedunculo-bulbaris (tr.inp.b.d.).


Fig. 28. — Through the hippocampal commissure and the middle of the sensory root of the V nerve. The section cuts the ramus communicans posterior of "the arterial system (r.c.p.) and the decussation of the posterior division of tr. tecto-bulbaris cruciatus {tr.t.h.c.2.), also the most ventral fibers of the commissure of the tuberculum posterius.


Fig. 29.— This passes immediately below the cerebral flexure at the tuberculum posterius and the ventral border of the peduncle. Mingled with the crossing fibers of tr.f.b.c.2. are the most ventral crossing fibers (unmarked) of the anterior division of tr. tecto-bulbaris cruciatus. The central nucleus of the isthmus (nuc.i.s'.c.) is here well defined, and laterally of it are terminals of the dorsal tegmental fascicles of groups (7) to (10).


Fig. 30. — This section cuts the ventral border of the junction of the peduncle with the isthmus, a locus marked by the superficial origin of the III nerve. The anterior division of tr. tecto-bulbaris cruciatus (tr.t.b.c.l.) is decussating within the ventral tegmental fascicles of group (1) (p. 277). More posteriorly, the section cuts the dorsal border of the sensory V root and the visceral sensory roots of the VII and IX nerves, showing the entire prevagal part of the f. solitarius (pp. 148, 166 and figs. 37, 38).


Fig. 31. — Through the nucleus of the III nerve and the most dorsal component of the commissure of the tuberculum posterius containing decussating fibers of tr. tecto-b'jlbaris cruciatus 1. Farther forward the section cuts the ventral thalamus at its widest part and passes through the ventral border of the eniinentia thalami. Large numbers of thick fibers pass from the ventral thalamus to the peduncle and isthniic tegmentum {tr.th.teg.r.), both superficially and deeper in the alba. Similar fibers (not drawn) arise from the large cells of the peduncle, which comprise a primordium of the mammalian interstitial nucleus of the f. longitudinalis niedialis. Many of the thalamic and most of the pedimcular fibers enter ventral tegmental fascicles of group (5) (see p. 281). Tiiis section cuts tegmental fascicles of group (3) at the dorsal convexity of their course as they recurve around the tuberculum posterius.


Fig. 32. — Section passing just above the interventricular foramen and through the middle of the eminentia thalami. Accompanying fibers of the hippocampal commissure are those of tr. cortico-habenularis medialis for the stria medullaris and tr. cortico-thalamicus niedialis (primordium of the column of the fornix). Thick fibers stream backward from the ventral thalamus in tr. thalamo-tegmentalis rectus, some superficially, some at middle depth, and some at the inner border of the alba, the latter entering dorsal tegmental fascicles of group (8). Similar fibers arise from the larger cells of the peduncle, but to simplify the pictures these have not been drawn in figures 31 and 32. Most of them enter tegmental fascicles of group (5), and some of these continue spinalward in f. longitudinalis medialis. The more ventral large cells of the peduncle (fig. 31) correspond with the interstitial nucleus (Cajal) of f. longitudinalis medialis, the more dorsal of these cells (fig. 32) to the nucleus of Darkschewitsch (p. 217 and figs. 6, 18). For further details see the analysis of the tegmental fascicles in chapter xx and references there given.


At the level of figure 32, tegmental fascicles of groups (4), (5), and (10) are cut at the dorsal convexity of their courses through the peduncle, and here they are insinuated among the fibers of the anterior and posterior divisions of tr. tecto-bulbaris cruciatus, descending from the tectum toward their decussations. Farther back in the posterior isthmic neuropil the bulbar and spinal lemniscus systems {Im. and Im.sp.) are turning upward and forward to ascend in the dorsal tegmentum (figs. 33, 34). The wide auricle contains terminals of vestibular and lateral-line root fibers, other bulbar connections, and nucleus cerebelli.


Fig. 33. — This passes through the floor of the stem-hemisphere fissure, cutting the most dorsal fibers of tr. cortico-habenularis medialis as they enter the stria medullaris and the dorsal border of the eminentia thalami. Fibers of tr. thalamo-frontalis are seen emerging from the ventral border of the dorsal thalamus. Large cells of the nucleus of Darkschewitsch occupy the dorsal border of the peduncle. Laterally of these, tegmental fascicles (7) and (9) are recurving over the peduncle, and two sections farther dorsally, fascicles of group (8) are cut at the top of their convexity. More posteriorly the section passes through the dorsal border of the central nucleus of the isthmic tegmentum, laterally of which a few fibers of the brachium conjunctivum are impregnated. Some of these are seen to arise from the nucleus cerebelli. Externally of this nucleus is the posterior neuropil of the isthmus containing dendrites from the superior visceral-gustatory nucleus, terminals of the ascending secondary visceral tract {tr.v.a.), and many other components. Behind the auricle the section passes through the ventral part of the lateral recess of the ventricle and spinalward of this through the "dorsal island" of Kingsbury.


Fig. 34. — Ten sections farther dorsally, all components of the stria medullaris can be identified in the ventral habenular nucleus. The tr. habenulo-thalamicus (tr.hab.ih.) contains fibers passing in both directions between this nucleus and the thalamus and regions posteriorly of it. The bulbar lemniscus (Im.) and the spinalJemniscus (Im.sp.) traverse the dorsal tegmentum to reach the dorsal thalamus. More posteriorly the section passes through the gray of the superior visceral-gustatory nucleus {nuc.vis.s.) and the junction of the body of the cerebellum {c.ch.) with the auricle.


Fig. 35. — Through the ventral borders of the dorsal habenular nucleus, nucleus pretectalis, and tectum. Posteriorly it passes through the superior medullary velum, showing the decussation of the IV nerve roots, and through the body of the cerebellum dorsally of the cerebellar commissure. Fibers of com. vestibulo-lateralis cerebelli are crossing at this level. Compare 1942, figures 78 and 79, cut in about the same plane, Weigert and Golgi.


Fig. 36. — Through the com. posterior and the tectum at its widest part. In figures 35 and 36 the brachia of the superior and inferior colliculi are designated tr. tecto-thalamicus rectus {tr.t.th.r.).



Figures 37, 38, and 39. — These semidiagrammatic sketches show some details of the medulla oblongata from an advanced larva. The sections are obliquely horizontal, with the right side and the anterior end more dorsal. X 30. The series contains 27 thick Golgi sections, numbered from dorsal to ventral surfaces, and the data here assembled are selected from sections 11-24. The outlines of sections 13, 17, and 20 are drawn, and upon these there are sketched some details from these and neighboring sections. The impregnation is scanty on a clear ground, with long courses of individual fibers clearly shown. There is no impregnation of the lemniscus systems and but little of fibers descending from above the isthmus, so that the intrinsic bulbar structures impregnated are well defined. In figures 37 and 38 the boundary between gray and white layers is marked by a broken line; in figure 39 the gray of the interpeduncular nucleus is stippled.


Fig. 37. — On the left side this section cuts through the base of the auricle and the floor of the lateral recess. The right side is more dorsal and includes the superior visceral-gustatory nucleus and its neuropil, within which are collaterals from tr. tecto-bulbaris rectus (p. 225). On the left side, V root fibers arborize in the ventral part of the isthmic neuropil, and some of them continue into the cerebellar commissure. Almost the entire course of the postfacial f . solitarius is projected from several sections.


Fig. 38. — This section on the left cuts the superficial origins of the V, VII, and VIII nerve roots and, farther forward, of the III root. The sensory V fibers are seen to bifurcate with more slender ascending branches, most of which arborize in a neuropil, wliich is the primordiuni of the superior sensory V nucleus of mammals. The prefacial f. solitarius fibers are smooth and unbranched to their terminals rostrally and internally of the entering V root. Some impregnated fibers of the vestibular root (r. VIII.) extend forward to end within and rostrally of the superior V neuropil, but these are not drawn. Internally of tlie trigeminal neuropil and confluent with it are terminals of ascending crossed and uncrossed fibers termed tr. bulbo-isthmialis (fr.b.is.), which is continuous with tr. spiuo-bulbaris {tr.sp.b.). These fibers ascend and descend from all levels of the medulla oblongata, many of them first decussating as internal or external arcuate fibers. Many of them bifurcate into ascending and descending branches with or without decussation. They accompany the spinal lemniscus and comprise a mixed bulbo-spinal and spino-bulbar system.


Fig. 39. — This section, parallel with the ventral surface, illustrates the courses of some of the external arcuate fibers, many of which are impregnated on the right side and none on the left. Useful landmarks are provided by the decussation of Mauthner's fibers (fib.M.) and two blood vessels, which, farther dorsally, are related with the roots of the V (b.r.V.) and VII (b.r.VlI.) nerves. No fibers of lemniscus and secondary visceral systems are impregnated. The visible fibers evidently are concerned chiefly with bulbar and spinal adjustments.


Figure .1,0. — Detail from a horizontal Golgi section of a late larva, showing the entrance of the sensory V root and three neurons of the motor V nucleus. X 50. The dendrites extend laterally to engage collaterals of the sensory root fibers, and in the adjoining section ventrally other dendrites of these cells ramify downward into the ventral alba, engaging collaterals and terminals of the descending and ascending fibers of the ventral funiculus, here unstained. Nothing except the elements drawn is impregnated in this region. Presumably, the axons of these cells enter the motor V root, the unimpregnated myelinated fibers of which are here darkened by the Golgi fluid.


Figures 41 to 44- — These semidiagrammatic sketches are made from horizontal Golgi sections of an advanced larva, with elective impregnation of the spinal lemniscus and a few other details. X 30. The series contains 26 thick sections, numbered from ventral to dorsal. Sections 2, 3, and 4, illustrating innervation of the hypophysis, have been published ('42, figs. 56, 57, 58). Here sections 7, 10, 12, and 14 are drawn, with some additions in each case from intervening sections.


Fig. 41. — The section passes through the V and VHI roots, with impregnation of superficial fascicles of descending fibers of both roots and ascending VHI fibers, some of which extend as far as the cerebellum. No ascending sensory V fibers are impregnated. The loci of the III, IX, and X roots farther dorsally are indicated by blood vessels which accompany these roots. The thick arcuate fibers from the region of the calamus scriptorius enter a mixed tract marked Im.sp., which here contains bulbo-spinal, spino-bulbar, spino-cerebellar, and spinal lemniscus fibers. Most of the thin unmyelinated fibers here impregnated come from lower levels of the spinal cord, and their further ascending course is shown in figures 42, 43, and 44. A few thick fibers of the ventral funiculi are impregnated, some of which decussate in the ventral commissure below the auricle.


Fig. 42. — At this level, thick arcuate fibers in the vicinity of the first spinal roots descend in the calamus region and after crossing join the spinal lemniscus. Some of these bifurcate with a thick ascending branch and a slender descending. Farther forward under the auricle the lemniscus fibers turn dorsalward in the ventral part of the isthmic neuropil, and here many of them end. This is tr. bulbo-isthmialis of figures 38 and 39. The lemniscus fibers which continue rostrad have collateral endings here. Other elective Golgi preparations from this lot show the origin of spinal lemniscus fibers from the nuclei of the dorsal funiculus ('446).


Fig. 43. — This section cuts the III nerve root and its nucleus and, in the isthmic neuropil, dispersed fibers of the spinal lemniscus and ascending VIII root. Medially of these are axons emerging from the gray of the trigeminal tegmentum, which turn forward and arborize within the gray of the isthmic tegmentum (p. 185).


Fig. 44. — At this level the few remaining impregnated fibers of the spinal lemniscus are turning forward across the dorsal tegmentum to reach the dorsal thalamus. The crude drawing gives a very inadequate picture of the delicacy of these widely branched terminal arborizations. The more numerous fibers of this tract, which arborize in the tectum, are not impregnated in this preparation (compare fig. 101 and '396, fig. 26). For the patterns of vascular supply sketched in the left thalamus see page 27.


Figure 45. — Terminals of the dorsal lateral-line root of the facial nerve, from the right side of a horizontal Golgi section of a late larva. X 50. Only three fibers of this root are impregnated, and each of these ramifies through almost the entire extent of the "dorsal island" of the area acusticolateralis immediately spinalward of the auricle, which is the exclusive central field reached by fibers of this root (compare fig. 33).


Figures ^6 and ^7. — Details from horizontal Golgi sections of an advanced larva. X 75.


Fig. 46. — A neuron of the anterodorsal part of the trigeminal tegmentum in the ventral part of the auricle (p. 185). The plane of section is approximately that of figure 31.


Fig. 47. — A small neuron at the anteroventral border of the body of the cerebellum, where it merges with the nucleus cerebelli. The slender dendrites extend outward into the dorsal part of the isthmic neuropil, and the axon is directed forward and downward into the brachium conjunctivum. Fibers of the trigeminal component of the cerebellar commissure (com.cb.) are approaching their decussation immediately dorsally of this level, as also are more scattered fibers of the vestibulo-lateral component (com.cb.L).


Figure ^8. — Detail from a transverse Golgi section previously pictured ('42, fig. 43). X 50. The small neuron of the dorsal tegmentum has a short axon, which ramifies in the deep neuropil of the gray near the cell body and also dorsalward in the alba of the anterior end of the nucleus posterior tecti. On the vascular loops {b.v.) seen here see page 27.


Figure 49. — A typical neuron of the dorsal tegmentum from a horizontal Golgi section of an advanced larva. X 75. The ependymal surface is marked by a thin line at the left, the pial surface by a thicker line at the right, and the outer border of the gray by a broken line. The axon arises from the dendrite and is directed spinalward (downward in the figure) and probably enters tegmental fascicle (7). A similar neuron >Ji s-itu is shown in figure 44 of the paper of 1942.


Figure 50. — Horizontal section through the decussation and spiral endings of the f. retroflexus (p. 197). Golgi method. X 50. The semidiagrammatic drawing is based on several adult specimens, in each of which the fasciculus of both sides is massively impregnated from the habenula to the interpedvuicular neuropil. All details drawn are visible in two sections of the specimen outlined (no. 2229), wliich is, however, poorly preserved. Several otlier series of sections show the same relations more clearly, though the oblique planes of some of the sections are less convenient. The considerable lateral obliquity of one of these gives an interesting view of the interior of the spiral. In these specimens the spiral endings are impregnated for about twothirds of the distance between the decussation and the level of the V'roots.


Figures 51 and 52. — Two adjoining horizontal Golgi sections of a late larva, illustrating the decussation and spiral terminals of the f . retroflexus, which is impregnated on only one side. The right side is more ventral. X 50.


Figures 53 and 54. — Two adjacent thick horizontal Golgi sections from an advanced larva. X 50.


Fig. 53. — At the level of the fovea isthmi (f.i.) and the decussation of the f. retroflexus. Laterally of the interpeduncular neuropil are impregnated fibers of tr. olfacto-peduncularis, with endings by open arborizations in this neuropil; compare figure 59, cut in a similar plane a little more dorsally.


Fig. 54. — Ventrally and posteriorly of the decussation only a few isolated fibers of f. retroflexus are impregnated, and tr. olfacto-peduncularis is terminating in the ventral interpeduncular neuropil in the same area as the spiral. The tufted endings seen here are not derived from either of these tracts.


Figure 55. — Detail of the spiral course of fibers of the f . retroflexus below the decussation, from a horizontal Golgi section of a late larva. X 50. In the interpeduncular neuropil nothing but these three fibers is impregnated, so there is no possibility of confusion.


Figures 56, 57, 5S.- — Three horizontal Golgi sections of an adult brain, in which the right f . retroflexus is unstained and the left fasciculus is abundantly impregnated from the habenula to its decussation. X 50. Some details of the decussation are added to figure 58 from the section adjoining it ventrally. The impregnation fails spinalward of the decussation. The left fasciculus shows an atypical division into two bundles as it enters the alba of the peduncle (p. 262).


Figure 59.- — An obliquely horizontal section through the superficial origins of the III roots, advanced larva. X 50. Here it is in about the plane of figure 30, but sharply inclined to the horizontal plane, with the posterior end more ventral and the anterior end more dorsal than that level. Impregnated fibers of the olfacto-peduncular tract pass the region of the fovea isthmi and then turn medially. Most of them end in the interpeduncular neuropil with open arborizations, though some extend farther spinalward (compare figs. 53, 54). From this neuropil slender axons of the dorsal interpedunculo-bulbar tract descend near the mid-plane and soon turn laterally to end in a dense axonic neuropil at the ventral border of the caudal end of the tegmentum isthmi and rostral end of the tegmentum trigemini. Here they engage dendrites of the smaller neurons of this region. Other preparations show that some of these fibers extend spinalward as far as the IX nerve roots.


Figures 60 to 6^. — These drawings present additional details from a series of transverse Golgi sections (no. 2246), which has already been quite fully described and illustrated in the sector of the brain stem between the interventricular foramen and the nucleus of the IV nerve ('27, pp. 271, 278, figs. 22-40). Those figures were drawn from sections selected from nos. 55-97 of the series. Section 88 was subsequently drawn on a larger scale to show some details of structure, including the arrangement of the tegmental fascicles, at the level of the III nerve roots ('42, fig. 44). Figures 60-64 extend the series spinalward (X 50), with special reference to the interpeduncular connections of the tegmentum, which are here well impregnated. These tegmento-interpeduncular fibers comprise one component of the complex f. tegmentalis profundus (p. 286), and the incomplete references to them made in 1927 can now be clarified and rectified. In this specimen there is no impregnation of the f . retroflexus or tr. olfacto-peduncularis. A few neurons of the interpeduncular nucleus are incompletely stained, but apparently no axons from these cells are impregnated. The dense interpeduncular neuropil here seen is composed almost exclusively of axons from the tegmentum and tr. mamillo-interpeduncularis. Terminals of the latter, when separately impregnated, are seen to be of more open texture than the dense vertically arranged terminal tufts of the tegmento-interpeduncular fibers (fig. 60). The two types of terminals are closely interwoven with each other and with tufted terminals derived from axons of neurons of the interpeduncular nucleus (not here impregnated) .



Fig. 60. — This section lies between the nuclei of the III and IV nerves, i.e., between the levels of figures 92 and 93, at about the level of figure 13 of 1936 and two sections rostrally of figure 39 of 1927. The drawing is a composite, containing some details from the two adjoining sections. Axons from both dorsal and isthmic tegmentum descend in the f. tegmentalis profundus, some terminating in the underlying interpeduncular neuropil and some of thicker caliber decussating in the ventral commissure (fr.ieg.b.), where collaterals separate from them to end in tufts of the interpeduncular neuropil. In figure 39 of 1927 these decussating fibers are called tr. tegmento-peduncularis, and the same fibers are here called tr. tegmento-bulbaris. Both designations are correct. The larger number of these fibers after decussation turn spinalward, some sending collaterals forward also. Others turn rostrad into the peduncle.


Fig. 61.— This section passes through the nucleus of the IV nerve, and the three neurons impregnated probably belong to this nucleus. Approximately the same plane is illustrated in several published figures ('25, figs. 9, 19; '36, fig. 14; '42, fig. 43). At this level the myelinated fibers of the ventral median fascicles (1) are turning laterally to enter tr. tecto-bulbaris cruciatus in the position of the broken lines under the gray of the interpeduncular nucleus. The thick fibers of tr. tegmcnto-bulbaris shown in figure 60 are here reduced in number, and the thin fibers of tr. tegmento-interpeduncularis are more numerous. These terminals and those of tr. mamillo-interpeduncularis enter into the dense axonic interpeduncular neuropil.


Pig 62. This detail is from the lower part of the interpedimcular nucleus rostrally of the

V nerve roots, not far from the plane of figure 91. There is impregnation of a typical large neuron of the trigeminal tegmentum and a smaller element at the border of the interpeduncular nucleus. In the interpeduncular neuropil the visible fibers are nearly all derived from the overlying tegmentum, with perhaps some terminals of tr. mamillo-interpeduncularis. At this level the ventral fascicles (4), (5), and (6) have united to form a compact f. longitudinalis medialis, and the dorsal fascicles spread out laterally into the isthmic neuropil and bulbar tegmentum (compare the horizontal sections, figs. 29-32).



Fig. 63. — Detail of floor-plate structure immediately caudad of the V nerve roots. Though the interpeduncular nucleus is not considered to extend as far spinalward as this level, the lateral and dorsal interpeduncular axonic neuropil is still rather dense (compare figs. 79, 81). The fibers impregnated here are derived chiefly from the overlying tegmentum, and some of them descend as far as the VII roots accompanying the dorsal and ventral tr. interpedunculobulbaris, which is not impregnated in this specimen. Two typical ependymal elements of the bulbar tegmentum are drawn.


Fig. 64. — This is a similar sketch of the ventral median raphe at the level of the VII nerve roots. A remnant of the tegmento-interpeduncular neuropil persists, and this extends no farther spinalward, though other preparations show that axons from the interpeduncular nucleus reach at least as far as the IX nerve roots. The ependyma here is more compact and mossy than at the level of the trigeminus.


Figures 65 and 66.— Two transverse sections through the interpeduncular region at the level of transition between the isthmic and the trigeminal tegmentum of the adult. X 50. Each figure is a composite of two adjacent sections, so that four consecutive sections are represented in the two drawings.


Sections taken farther forward from this well-impregnated series have been shown ('42, figs. 45-47). Between the levels of the III and IV nuclei, tr. tegmento-interpeduncularis is selectively stained, with tufted endings in the interpeduncular neuropil similar to those shown in figures 61 and 6:2. At the level of figure 65 only a few of these fibers are impregnated and also a few neurons of the interpeduncular nucleus with dendrites extending downward into the interpeduncular neuropil.


In figure 66 the neurons of the interpeduncular nucleus are seen to have thick dendrites extending laterally to ramify widely in the alba of the isthmic and trigeminal tegmentum and thinner dendrites directed ventrally into the interpeduncular neuropil. Both dendritic and axonic arborizations enter the glomerulus-like tufts; but, in order to clarify their relations, only the dendritic component is drawn on the left side and the axonic component on the right (compare figs. 83, 84).


Figure 67.- — Transverse section through the middle of the interpeduncular nucleus of a halfgrown larva. X 75. The broken line marks the outline of the gray substance. The Golgi impregnation is scanty, showing in this region only a few decussating fibers of tr. tegmento-bulbaris on a clear ground. The more dorsal of the two fibers shown is sketched from the adjoining section spinalward. These thick axons probably arise from unimpregnated cells of the isthmic tegmentum, as shown in figiu-e 68. In the ventral commissure, where these fibers decussate, slender collaterals leave them to ramify in the interpeduncular nucleus and its neuropil.


Figure 6S. — A diagram based on Golgi sections of a larval Arablystoma from the same lot as figure 67. X 75. The sections are obliquely transverse. The section outlined is in about the same plane as figure 67 on the right and passes through the auricle on the left. The details are from this and the two adjoining sections. Everything drawn was observed, but the assembly is schematic. Thick axons of neurons of the isthmic tegmentum {tr.feg.b.) converge into the ventral commissure, where slender collaterals separate from them to arborize in tufted form in the interpeduncular neuropil.


Figures 69 a/id 70. — These two drawings illustrate details of the interpeduncular region at and immediately rostrally of the nucleus of the IV nerve (compare the diagram, fig. 68). X 50. The impregnation of the transverse Golgi sections of this adult brain is exceptionally good, and it has been quite fidly illustrated (';*7. figs. ^4-30; U, figs. 24-31).


Fig. 69. — This section is adjacent posteriorly to the one shown in figure 30 of 1942, to which reference may be made for the topographic relations. In both figures the bundles of myelinated fibers of the ventral and ventromedian tegmental fascicles are outlined with broken lines. In the interpeduncular region of the section here shown, there is no impregnation of any nervous elements except a few fibers of tr. tegmento-bulbaris at their decussation in the ventral commissure. Within the commissure slender collaterals descend into the interpeduncular neuropil, where some of them have tufted endings.


Fig. 70.^This is drawn from the second section spinalward of figure 69, through the nucleus of the IV nerve, one neuron of which is impregnated. Thick myelinated axons of the IV nerve root (darkened by the Golgi fluid) ascend from this nucleus in the position indicated by broken lines. Typical ependymal elements of the ventral raphe are drawn.


Figure 71. — Obliquely longitudinal section of the adult brain taken not far from the midsagittal plane, with the dorsal and anterior sides inclined somewhat laterally, so that almost the entire length of the f. retroflexus appears in a single thick Golgi section. Golgi method. X 30. The drawing is semidiagrammatic, with some details added to the section outlined from neighboring sections and from the opposite side.


At the point where tr. cortico-habenularis medialis joins the stria medullaris, numberless fine collaterals separate from it to enter the eminentia thalami. The section cuts the optic nerve at its junction with the brain. Peripherally, each optic nerve between the foramen and the brain shows several hundred impregnated unmyelinated fibers, most of which lose the stain before entering the brain. The surviving impregnated fibers decussate in chiasmatic bundles 1 (ch.l.) at the extreme anteroventral angle of the chiasma and ascend toward the tectum as the most rostral fibers of the marginal optic tract, losing the stain before reaching the tectum ("42, p. 232).


The abundant impregnation of the f . retroflexus extends only as far as the decussation, below which its fibers are unstained. As these fibers turn medially toward the decussation, many of them separate and descend uncrossed along the lateral margin of the interpeduncular neuropil. The locus of the sidcus isthmi on the ventricular siu-face is projected upon the drawing as a thick broken line. In the posterior hp of this fissure there is massive elective impregnation of the brachium conjunctivura and the locus of the decussation of these fibers more medially is marked by three crosses. Before and after crossing, many of these fibers turn spinalvvard in the superficial alba of the isthmic tegmentum.


Figures 72 and 7S. — Two semidiagrammatic drawings from an obliquely longitudinal series of Golgi sections of an adult brain. The sections are cut at an angle of about 30° from the sagittal plane, with dorsal and posterior sides more lateral.


Fig. 72. — These details are assembled from several consecutive sections to illustrate some components of the complex of fibers at the di-telencephalic boundary. X 25. Only some of the thicker unmyelinated axons are impregnated in this preparation. The ventral border at the



left is near the median plane, and posteriorly of this place the section is inclined laterally. At the posteroventral border of the primordium hippocampi some thick fibers are seen to converge at the thalamic junction, and Weigert sections show that a few of them are myelinated. The larger number of these fibers descend to the hippocampal commissure, and a smaller number a.scend in the stria medullaris thalami as tr. cortico-habenularis medialis. Others descend obliquely posteroventrad into the ventral thalamus, passing through and posteriorly of the gray of the eminentia thalami^ — tr. cortico-thalamicus medialis. This is the precursor of the mammalian columna fornicis, but here apparently none of these fibers reach the mamillary region of the hypothalamus. These thick fibers are accompanied by far more numerous very thin fibers, some of which arise as collaterals from the decussating fibers of the hippocampal commissure (fig. 71). An impregnated neuron of the bed-nucleus of the anterior commissure ridge is sketched from the most lateral of the sections here represented. Its dendrites spread dorsally and ventrally among the hippocampal fibers; the course of the axon is not revealed.


The well-impregnated fibers of tr. thalamo-frontalis arise from the cells of the dorsal thalamus shown in figure 73. In the chiasma ridge, chiasmatic bundles 2 and 3 of optic fibers are impregnated, and their terminals in the lateral part of the tectum are seen at the top of the figure. The details of this optic connection have been pubHshed ('42, p. 235). The fibers of f . tegmentaHs profundus seen at the right are ascending from their decussation in the ventral commissure to spread superficially over the lateral aspect of the isthmic tegmentum. Most of these fibers belong to the brachium conjunctivum.


Fig. 73. — Section more medially, showing details of the origin of the f. retroflexus from the habenula. X 50. The drawing is a combination of three adjoining sections. Only the thicker axons of the fasciculus are here impregnated. These surround a dense fascicle of thinner unstained fibers. The fibers here impregnated arise from neurons of the habenula, from others more posteriorly in the pars intercalaris, and probably also from some of the impregnated neurons of the pars dorsalis thalami. None of these stained fibers reach the interpeduncular nucleus. They spread out in the alba of the posteroventral part of the peduncle, some turning forward into area ventrolateralis pedunculi. The dendrites of the impregnated neurons of the dorsal thalamus extend dorsally among terminals of tr. tecto-thalamicus rectus and also for long distances posteroventrally, accompanying the f . retroflexus into the area of the primordial geniculate bodies (pp. 221, 238). Sections taken more laterally show abundant impregnation of axons from this group of cells which enter tr. thalamo-frontalis (fig. 72).


Figures 74 to 78. — Five sagittal sections of the adult brain, illustrating the composition of the stria medullaris thalami (p. 256 and fig. 20). X 37. The sections were prepared by the reduced silver method of Cajal and cut at 10 n slightly oblique to the sagittal plane, with the dorsal and rostral sides more lateral (compare "36, figs. 17-79, where the dorsoventral obliquity is in the reverse direction). In these sections the thickest axons, both myelinated and unmyelinated, are black, and the thinner fibers range through gray to brown or yellow, so that components of the stria are well dififerentiated.


Fig. 74. — This cuts through the di-telencephalic junction near its lateral border, passing through the olfacto-peduncular tract below, and above this through the dorsal and ventral fascicles of the lateral forebrain bundle. The artery (r.h.m.) lies in the floor of the stem-hemisphere fissure. Here four components of the stria medullaris converge, two of them- — tr. olfactohabenularis anterior and lateralis — passing up from the ventral surface externally of the lateral forebrain bundles.


Fig. 75. — In the region of the stem-hemisphere fissure the section passes medially of the lateral forebrain bundles, showing the connection of tr. olfacto-habenularis medialis and tr. septo-habenularis with the stria. Ventrally the plane of section is not far from that of figure 103; dorsally it is much more lateral.


Fig. 76. — At this level tr. cortico-habenularis medialis enters the stria and ascends as its mo.st rostral member. All components of the stria are distributing fibers in the habenular neuropil, posteriorly of which several of the tracts retain their identity. Fibers of tr. habenulothalamicus which recurve spinalward over the gray of the dorsal thalamus have been seen to connect only with the ventral habenular nucleus. This tract includes also thalamo-habenular fibers. The boundary of the gray of the eminentia thalami is marked by a broken line.


Fig. 77. — A large proportion of the fibers of all components are here dispersed in the habenular neuropil. A residue of fibers is accumulated posteriorly in a commissural bundle, within which the following components can be identified: tr. cortico-habenularis lateralis and medialis, tr. olfacto-habenularis anterior, tr. amygdalo-habenularis.


Fig. 78. — In this section, 300 ^ from the mid-plane, the limits of the gray of the dorsal habenular nucleus (corresponding with the medial nucleus of mammals) and of the pretectal nucleus are marked by broken lines. Most of the habenidar neuropil has disappeared. Some fibers from this residue enter the habenular commissure, but most of the crossing fibers are derived from recognizable tracts. The thickest fibers cross at the anterior end of the commissure. These apparently are derived chiefly from tr. amygdalo-habenularis, with additions from other components. The thinner fibers of the commissural bundle include clearly some residue of tr. cortico-habenularis lateralis and medialis and tr. olfacto-habenularis anterior.


Figures 79 and SO.— Two adjoining obliquely sagittal Golgi sections of the adult brain, illustrating some components of the interpeduncular neuropil. In this specimen the finest axons are richly impregnated, especially the deep neuropil of the gray substance, which is continuous throughout the tectum, peduncle, and dorsal, isthmic, and bulbar tegmentum. In the area illustrated it receives fibers from the fields dorsally and anteriorly, and from it axons descend to the isthmic and bulbar tegmentum. All cell bodies are imbedded in this dense fibrillar reticulum. On these small-scale drawings it is impossible to portray the delicacy and complexity of this tissue, and the attempt is made to indicate the general trend of its coarser fibers.


Fig. 79. — This section is median in the floor of the ventricle from the fovea isthmi to the level of the IX nerve roots. X 40. The ventricular floor from the fovea to the tuberculum posterius is projected from two adjacent sections, as indicated by a dotted line. This section is very thick and is so inclined dorsoventrally as to include above the ventricular floor a thin slice of the nucleus of the tubercidum posterius, tegmentum isthmi, and rostral end of the trigeminal tegmentum of one side, ^'entrally of the ventricular floor in the isthmus region the section cuts the interpeduncular nucleus and neuropil close to the mid-plane on the opposite side. The level of the VIII nerve roots is indicated by the decussation of Mauthner's fibers, from which one of the fibers is seen descending. The other fiber appears in the adjacent section 24. From the latter section tlie portion of tr. interpedunculo-bulbaris dorsalis (ir.inp.b.d.) spinalward of this decussation is added to the drawing, showing that some fibers of this tract descend as far as the level of the IX nerve roots. None of its fibers have been seen to extend farther spinalward. The ventral division of this tract {Ir.inp.h.r.), on the contrary, goes much farther, perhaps as far as the spinal cord.


At the fovea isthmi several ependymal elements are impregnated, and a small artery here enters the foveal pit at the ventral surface. These ependymal elements are crossed by unimpregnated myelinated fibers of the ventral tegmental fascicles {f.m.f.{l)), as indicated by dashed lines. A single small neuron of the nucleus of the tuberculum posterius is impregnated. Its dendrites spread downward among slender tortuous axons of tr. mamillo-peduncularis, which form a dense neuropil in the gray of the peduncle. In this plane no fibers of tr. mamillo-interpeduncularis are stained, but they are abundant farther laterally.


Unimpregnated cell bodies of the interpeduncular nucleus are clearly visible, arranged as indicated by the dotted outlines. Among these and ventrally of them is dense neuropil, very inadequately shown in the drawing. These fibers are derived in part from tr. mamillo-interpeduncularis and from the overlying tegmentum, and in larger part they are axons of the cells of the interpeduncular nucleus. These axons take tortuous courses, mainly directed dorsalward into the isthmic and trigeminal tegmentum and spinalward into the dorsal and ventral interpedunculo-bulbar tracts. From them arise numberless collaterals which ramify in the interpeduncular neuropil and enter the glomeruli (fig. 83).


Fig. 80.— The adjoining section. X 75. The ependymal floor of the ventricle is slightly to one side of the mid-plane, ventrally of which the section is inclined laterally. Heavy broken lines mark the upper and lower limits of the zone of unimpregnated cell bodies of the interpeduncular nucleus. Scattered cells of this nucleus are distributed in the underlying neuropil except its ventral part. Above this zone of den.sely crowded cells are the less crowded cells of the isthmic and trigeminal tegmentum, none of which are impregnated. One neuron of the interpeduncular nucleus is impregnated with dendrite directed ventrally into the neuropil. Fragments of other dendrites, with tufted terminals in glomeruli, are spread throughout tlie neuropil, some of which are drawn (compare figs. 65, 66, 81-84). There is no other impregnation in the ventral interpeduncular neuropil containing the unimpregnated spiral endings of the f. retroflexus, but dorsally of this there are stained axons of various sorts among the dendrites. These are not drawn, except those lying between the cells of the interpeduncular nucleus and the myelinated fibers of the ventral commissure, the most dorsal bundles of which are outlined with dotted lines. Most of the.se immyelinated fibers enter tr. interpeduncvdo-bulbaris dorsalis. The interpeduncular neuropil is broadly connected with the overlyiiig tegmental neuropil by fibers passing in both directions. In this plane (not drawn) there are a few fibers of tr. niamillo-interpeduncularis and more of them farther laterally.


Figure ,Si .—Obliquely sagittal Golgi section of an adult brain taken close to the mid-plane. X 60. Some other sections of this specimen have been published (fig. 101; '25, figs. 14, 40). The plane of section of this figure is similar to that of figure 79, but somewhat more oblique. The ventral surface is nearly median in the isthmus, and dorsally and anteriorly of this the plane is inclined laterally. Since the sections are very thick, the dorsal ventricular border of the tuberculum posterius (indicated by the dotted line), and a slice of the lateral wall dorsally of it are included in the section outlined. The detail is drawn from this section and the adjacent section laterally. The four components of the commissure of the tuberculum posterius (p. 302) and the dorsal fascicles of the ventral commissure posteriorly of the fovea isthmi are outlined with dotted lines. The decussation of the anterior division of tr. tecto-bulbaris cruciatus in the ventral median tegmental fascicles {J.m.t.{l)) is indicated by dashed lines (compare '36, fig. 2).


A neuron of the dorsal (mamillary) part of the hypothalamus and the dendrite of another are impregnated. Their axons branch, and one branchlet enters tr. mamillo-interpeduncularis. This tract partially decussates in component 1 of the commissure of the tuberculum posterius (the retroinfundibular commissure of the literature), and its slender unmyelinated fibers arborize in the rostral part of the interpeduncular neuropil chiefly laterally of the plane here shown (compare '36, figs. 3, 8, 20). Two ependymal elements are drawn. These lie near the ventral median raphe under the cerebellum (compare figs. 63, 64, 70).


Unimpregnated cell bodies of the interpeduncular nucleus are scattered under the ependymal surface and among the myelinated median fascicles and ventral commissure bundles. A well-impregnated dendrite of one of these is drawn; it has a tufted terminal near the ventral surface (compare figs. 62, 65, 66, 80, 82). There is scanty and very clear impregnation of some of the coarser fibers of the interpeduncular neuropil. Those in its anterior part are derived chiefly from tr. mamillo-interpeduncularis and probably some from tr. olfacto-peduncularis, which is partially impregnated more laterally (fig. 101). More posteriorly the visible fibers all seem to be axons of cells of the interpeduncular nucleus which assemble to descend in the dorsal and ventral interpedunculo-bulbar tracts.


Figure 82.— k semidiagrammatic drawing of a thick Golgi section from an obliquely sagittal series of an adult brain. X 37. The section is nearly median posteriorly (at the right), and anteriorly it is much more lateral, including the f. retroflexus in the peduncle. Some fibers of this fa.sciculus descend uncrossed into the interpeduncular neuropil. Mingled with the uncros.sed fibers of the f. retroflexus are thicker and smoother fibers of tr. olfacto-peduncularis, some of which enter the interpeduncular neuropil. Dorsally of these is the mamillo-peduncular tract, comprising ventral tegmental fascicle (3), and still farther dorsally are fibers of tegmental fascicle (6), which come from the postoptic commissure. These tegmental fascicles contain both myelinated and unmyelinated fibers, and both sorts are here darkened by the Golgi treatment. Between the two tegmental fascicles just mentioned are the darkened thick myelinated fibers of the anterior division of tr. tecto-bulbaris cruciatus (f.m.1.{l)), some of which decussate near the fovea isthmi dorsally of the olfacto-peduncular fibers and others descend in the ventral median fascicles. One neuron of the interpeduncular nucleus is well impregnated and dendrites of several others, all of which exhibit the characteristic dendritic tufts. No ependyma, spiral fibers, or axonic tufts are impregnated in the interpeduncular neuropil. The scanty axonic neuropil is probably composed exclusively of terminals of uncrossed fibers of the f. retroflexus and olfacto-peduncular tract.


Figure SJ.— Diagram of a neuron of the interpeduncular nucleus, seen as projected upon the median sagittal section. X 40. This is a composite drawing from observations made on many sections cut in various planes. Compare figures 19 and 84.


Figure 84. — Diagram of the composition of the interpeduncular glomeruli. X 45. This drawing, like figure 79, shows the median section of the floor plate and dorsally of this an oblique slice of the overlying tegmentum. Two neurons of the interpeduncular nucleus are drawn. Tufted collaterals from the axon of the anterior element engage dendritic glomeruli of the posterior element. Tufted axonic terminals enter glomeruli from small cells of the isthmic and trigeminal tegmentum and also from collaterals of tr. tegmento-bulbaris, which arises from large cells of the same tegmental areas. Compare figures 19 and 83.


Figure 85. — The lateral aspect of the adult brain, drawn from the same wax model as figure lA. X 15. This drawing shows the courses of the four most superficial components of the stria medullaris thalami seen as projected upon the lateral surface (fig. 20 and chap, xviii). Tractus olfacto-habenularis lateralis (2) goes directly dorsally from the preoptic nucleus. Tractus olfacto-habenularis anterior arises in ventral (3) and dorsal (4) divisions from the anterior olfactory nucleus. Tractus cortico-habenularis lateralis (5) arises from the primordial piriform lobe and joins no. 4 before entering the stria medullaris.


Figure 86A, B, and C— The brains of A. tigrinum and Necturus drawn to approximately the same scale.


A.— Lateral aspect of the brain of A. tigrinum and its arteries. After Roofe ('35). X 7.5. Drawn after fixation in 10 per cent formalin for six weeks.


B. — Lateral aspect of the cerebrum of Necturus. This and figure 86C are copied from 1933c, figures 3 and 4. X 8.


(2 — Median section of the brain of Necturus (compare fig. 2). In the thalamus and midbrain the section is median, and the dotted lines indicate ventricular sulci which mark the boundaries of the chief cellular areas. In the cerebral hemisphere the section is slightly to one side of the mid-plane, and the dotted lines indicate the boundaries of cellular areas with limits marked internally by ventricular sulci.


The section pictures which follow are here reproduced from previous publications (with minor alterations in some instances), and more complete descriptions will be found in the references cited.


Figures 87 to 91.- — Five transverse sections of the adult medulla oblongata. Method of Wiegert. X 37. These are copies, respectively, of figures 5, 3, 2, and 1 of 19446 and figure 16 of 1936, drawn from the type specimen, no. IIC.


Fig. 87. — Section taken immediately below the calamus scriptorius through the commissural nucleus and the com. infima of Haller.


Fig. 88. — Through the lower vagus region and the most anterior ventral rootlet of the first spinal nerve.


Fig. 89 .^At the level of the IX nerve roots. A rootlet of the \T nerve emerges in this section. The two lateral-line roots of the vagus enter immediately rostrally of this level.


Fig. 90. — Section taken immediately below the superficial origin of the V nerve roots, including the posterior part of the motor V nucleus.


Fig. 91. — Section through the cerebellum, auricle, and rostral end of the trigeminal tegmentum.


Figure 92. — Transverse section immediately spinalward of the nucleus of the IV nerve. X 37. Ventrally it passes through the posterior end of the infundibulum and dorsally through the junction of superior and inferior colliculi. The Arabic numbers in parentheses refer to the tegmental fascicles described in chapter xx. This is a copy of figure 14 of 1936, drawn from a reduced silver preparation, method of Rogers.


Figure 93. — Semidiagrammatic transverse section at the level of the III nerve roots. X 37. Copied from figure 14 of 1942. The layers of the tectum and some of its afferent tracts are shown on the left side and efferent tracts on the right.


Figure 9^. — Section through the rostral border of the commissure of the tuberculum posterius. X 37. Copy of figure 10 of 1936 and drawn from the same specimen as figure 92.


Figures 95 to 100.- — These are obliquely transverse sections (ventral side inclined spinalward), copied, with some alterations, from the paper of 1927, where full descriptions will be found. X 25. They are from the brain of a recently metamorphosed adult from Colorado, prepared by the Golgi method. The drawings are composite, the outlines and much of the detail being derived from the specimen mentioned, supplemented by additional details selected from 20 other specimens cut in the transverse plane and stained by the methods of Golgi, Weigert, and Cajal.


Fig. 95. — Through the habenular commissure and the posterior border of the po.stoptic commissure ('27, fig. 20).


Fig. 96. — Through the rostral border of the chiasma ridge and the middle of the eminentia thalami. By reason of the obliquity of the section, it passes rostrally of the habenulae ('27, fig. 15). The cellular area marked p.v.ih. is the nucleus of the olfacto-habenular tract (p. 248).


Fig. 97.- — Through the hippocampal commissure and the decussation of the lateral forebrain bundles in the anterior commissure ("27, fig. 12).


Fig. 98. — Immediately in front of the lamina terminalis. On the right side is a characteristic impregnation of the neuropil of the corpus striatum. In the preparation drawn, only the dendritic component of the neuropil is impregnated ("27, fig. 9).


Fig. 99. — Section about 0.1 mm. anterior to the last, through the mid-septal region. The head of the caudate nucleus is here at its maximum size. In the ventrolateral wall on the right side the neuropil of the corpus striatum, in the preparation here drawn, has only the axonic component impregnated (cf. fig. 108). These contorted and branched axons are interlaced with the dendrites shown in figure 98, and the boundary of this area of dense neuropil is sharply defined ('27, fig. 8).


Fig. lOO.^Section through the middle of the olfactory bulb and the extreme anterior border of the primordium hippocampi. It includes also the posterior border of the f. postolfactorius ("27, fig. 4).


Figure 101.- — Semidiagrammatic sagittal section of the adult brain, Golgi method. X 35. The section is oblique, with the dorsal side somewhat more lateral, cut in a plane which includes almost the whole length of the lateral forebrain bundles and tr. olfacto-peduncularis. In the tectal neuropil there are typical endings of the spinal lemniscus and tr. strio-tectalis (copied from '42, fig. 17).


Figures 102, 103, /O^.^Three sagittal sections from the brain of a specimen from Colorado prepared about two weeks after metamorphosis by the reduced silver method of Rogers. The sections are slightly oblique, with the dorsal surface and caudal end more medial. X 37. In this series of sections the courses of the tegmental fascicles can be clearly followed. These pictures are selected from a series published in 1936.


Fig. 102. — The section passes through the middle of the gray substance of the tectum opticum at its widest part and shows the courses of the lateral forebrain bundles and lemniscus systems ('36, fig. 17).


Fig. 103.- — Section taken 0.17 mm. more medially and showing the arrangement of the more medial tegmental fascicles ('36, fig. 18).


Fig. 104. — Section taken 0.16 mm. more medially and about 0.1 mm. from the mid-plane at the tuberculum posterius. It includes the interstitial nucleus of the f . longitudinalis medialis {nuc.tub.p.) and the nuclei of the III and IV nerves ('36, fig. 22).


Fig. 105. — Semidiagrammatic horizontal section through the dorsal border of the olfactory bulb and dorsal sector of the anterior olfactory nucleus of the adult. X 50. The medial border (at the right) is slightly inclined ventrally. The diagram is based on sections of a single specimen. The details pictured have all been observed, but the assembling is a .schematic composition ('34, fig. 1). The histological structure of the olfactory bulb as seen in horizontal sections is shown in figure 110. Figure 105 is a similar diagram at a more dorsal level. Two mitral cells (1) are drawn and a typical granule cell (2). Two elements of transitional type (3) in the stratum granidare send slender axons to the anterior olfactory nucleus, three neurons (//) of which are drawn. Farther back there are three typical neurons of the primordium hippocampi (6') and one of the primordium pirifornie (-5).


Fig. 106.^ — Detail of the neuropil of the gray substance of the dorsal sector of the anterior olfactory nucleus in the plane of figure 105 and from the same specimen ('34, fig. 2). Golgi method. X 100. Only the axonic component of the neuropil is here impregnated. The positions of the unimpregnated cell bodies are indicated by the dotted outlines.


Fig. 107. — This is a detail of the neuropil of the gray substance of the rostral part of the corpus striatum of the adult frog, Rana pipiens ('34, fig. 3). Golgi method. X 142. The section pictured is obliquely longitudinal, inclined about 45° from the horizontal plane, witli the ventricular border more ventral. The rostral end is above. The impregnation is similar to that shown in figure 106. One small neuron of the striatal gray is impregnated.


Figure 108. — Neuropil of the anterior olfactory nucleus and corpus striatum, as seen in an obliquely horizontal section of the adult brain. The right side is more dorsal ('42, fig. 15). Golgi method. X 50. The section pa.sses through the right hemisphere near the ventral surface, including the ventral sector of the anterior olfactory nucleus and the ventral border of the gray of the corpus striatum {c.s., here unimpregnated). Laterally of this gray is the striatal neuropil {c.s.n.), in which only the axonic component is impregnated (cf. fig. 99). Rostrally of the striatal gray is the extreme ventral border of the neuropil of the caudate nucleus, the structure of which is .shown in figure 109.



Figure 109.— An obliquely longitudinal section through the ventrolateral border of the adult hemisphere ('42, fig. 16). (lolgi method. X 80. The section is inclined to the sagittal plane, with the anterior and dorsal sides much more lateral. It cuts the accessory olfactory bulb and the anterior end of the caudate nucleus, whose gray {nucxaud.) is unimpregnated. Above and externally of this is a dense impregnation of the caudate neuropil with sharply defined borders. Some unimpregnated cell bodies are enmeshed within this neuropil, only the axonic component of which is impregnated. This neuropil is continuous with that of the remainder of the striatal complex shown in figure 108. Anteriorly of it, six neurons of the anterior olfactory nucleus are impregnated.




Figure llOA, B.— Diagrams of the structure of the olfactory bulb as seen in horizontal section of the adult brain.


A. The outline is taken from a section at the widest part of the olfactory bulb and the elements are drawn to scale from several Golgi sections, most of them from the same specimen ('246, fig. 1). X 50. Glomeruli are outlined with dotted lines. The approximate locus of figure hob' is indicated. The following types of neurons have been described: (1) periglomeridar cells; (2) mitral cells; (3) subglomerular tufted cells; (4) granule cells; (5) cells intermediate between granules and neurons of nucleus olfactorius anterior; (6) transitional cells of nucleus olfactorius anterior.


B._The layers of the olfactory bulb as seen in horizontal section after fixation in formalinZenker followed by Mallory's stain ('246, fig. 4). X 150.


Figure 111. — Lateral aspect of the forebraiii of Necturus (cf. fig. 86B), upou which some olfacto-somatic pathways are projected (*33e, fig. 1). X 8. Olfactory tracts are drawn in broken lines, descending tracts in solid lines, and the thalamic radiations and their connections are in red lines (p. 102). Connections with the epithalamus and hypothalamus are omitted.


Figure 112. — Median section of the brain of Necturus (cf. fig. 86C), upon which some of the descending pathways from the cerebral hemisphere to the hypothalamus and epithalamus are projected (p. 268) ('33p, fig. 2). X 8.


Figure 113. — Diagram illustrating the relations of projection and association fibers in the cerebral hemispheres of Necturus as seen in transverse section a short distance rostrally of the lamina terminalis (p. 102). The gray substance is outlined by a broken line. Descending fibers arc drawn on the left, ascending and pallial association fibers on the right ("34, fig. 7). X 16.



The following list includes all abbreviations on the figures and a few others used in previous publications. Reference may also be made to the list appended to the monograph on Necturus ('336, pp. 280-88), where synonyms of many of these terms are given, with references to the literature. For reference to published lists of tracts see page 270, and for nomenclature of blood vessels see the paper of 1935 and Roofe ('35).


c.s.v.n., neuropil of c.s.v. cb., cerebellum cell M., cell of Mauthner ch., chiasma opticum col.inf., celliculus inferior = nuc.p.t. col.swp., colliculus superior; tectum opticum com.amg., commissure of amygdalae coni.ant., anterior commissure com.cb., commissura cerebelli com.ch.L, commissura vestibulo-lateralis cerebelli com.cb.l.L, lateral-line component of com.cb.l. com.cb.V., trigeminal component of com.cb. com.cb.VIII., vestibular component of

com.cb.l. com.hab., habenular commissure com. hip., hippocampal commissure com.i., commissura infima Halleri com.po., commissura postoptica com.post., commissura posterior com.t.d., commissura tecti diencephali com.t.m., commissura tecti mesencephali com.tub., commissura tuberis coni.tub.p., commissura tuberculi posterioris com.v., commissura ventralis


A., division A of tr.th.ieg.d.c.


a.a., arteria auditiva

a.ac., area acusticolateralis

a.b., arteria basilaris

a.c.c, arteria carotis cerebralis

a.c.c.a., arteria carotis cerebralis, ramus

anterior a.c.c. p., arteria carotis cerebralis, ramus

posterior a.c.i., arteria carotis interna a.ch.r.a., arteria chorioidea rhombencephali

anterior a. gen., area geniculata = np.gen. a.l.t. = a.vl.p. a.o., arteria ophthalmica a.s.a., arteria spinalis anterior a.s.l., arteria spinalis 1 a.s.l.r.a., arteria spinalis 1, ramus anterior a.s.l.r.p., arteria spinalis 1, ramus posterior a.p., angulus ventralis

a.vl.p., area ventrolateralis pedunculi = a.l.t. amg., amygdala amg.n., neuropil of amygdala as.amg.pal., amygdalo-pallial association as.pal.amg., pallio-amygdaloid association as.pal.st., pallio-striatal association as.st.pal., strio-pallial association aur., auricle

B., division B of tr.th.ieg.d.c.


b.ol., bulbus olfactorius

b.ol.ac, bulbus olfactorius accessorius

6.«'.,blood vessel

b.v.III., v., VII., IX., X., blood vessels

associated with cranial nerve roots br.coL; br.col.s.; br. col.inf., brachia of superior

and inferior colliculus = tr.th.r. br.conj., brachium conjunctivum

c.cb., corpus cerebelli

C.S., corpus striatum

c.s.d., corpus striatum, pars dorsalis

c.s.d.n., neuropil of c.s.d.


c.s.v. , corpus striatum, pars ventralis


d.b., diagonal band of Broca d.f.l.t., decussation oi f.lat.t. d.f.m,.t.; d.f.med.t., decussation oi f.med.t. d.f.retr., decussation oi f.retr. d.fib.M., decussation of Mauthner's fibers d.isL, dorsal island of Kingsbury d.r.IV., decussation of IV nerve roots d.rinf., decussatio retroinfundibularis; components 1 and 2 of com.tub.p. d.tr.st.ped., decussation of tr.st.ped. d.tr.st.t., decussation of tr.st.t.


em.com,.p., eminentia commissurae posterioris em.s.t., eminentia subcerebellaris tegmenti

= nuc.cb. + teg.is. em.th., eminentia thalami em.V., eminentia trigemini ep., epiphysis


F., foramen interventriculare

f.d.t., fasciculi dorsales tegmenti

f.i., fovea isthmi

f.l.m., fasciculus longitudinalis medialis

f.lat.t., fasciculus lateralis telencephali; lateral

forebrain bundle f.lat.t.d., dorsal fascicles oi f.lat.t. = tr.st.t. f.lat.v., ventral fascicles oi f.lat.t. = tr.st.pcd. f.m.t., fasciculus medianus tegmenti f.med.t., fasciculus medialis telencephali;

medial forebrain bundle f.med.t.d., dorsal fascicles oi f.med.t. f.med.t. v., ventral fascicles oi f.med.t. f.jpo., fasciculus postolfactorius f.retr., fasciculus retrofiexus f.sol., fasciculus solitarius f.sol.pf., prefacial fasciculus solitarius f.teg.p., fasciculus tegmentalis profundus f.v.t., fasciculi ventrales tegmenti fib.arc, arcuate fibers jib.M., fiber of Mauthner jim., fimbria fiin.d., funiculus dorsalis

g.I., ganglion cells of nervus olfactorius gl.ol., glomeruli olfactorii

hab., habenula

hah.d., nucleus dorsalis habenulae

hah.d.n., neuropil of hah.d.


hab.v., nucleus ventralis habenulae

hab.v.n., neuropil of hab.v.


hem., cerebral hemisphere

hyp., hypophysis

hyp.d., hypophysis, pars distalis

hyp.g., hypophysis, pars glandularis

hyp.i., hypophysis, pars intermedia

hyp.n., hypophysis, pars nervosa

hyp.i., hypophysis, pars tuberalis

hyth., hypothalamus

7 to .Y, cranial nerve roots inf., infundibulum inp.n., interpeduncular neuropil is., isthmus rhombencephali

lam.t., lamina terminalis Im., general bulbar lemniscus Im.sp., lemniscus spinalis

M., mouth of paraphysis

m.c, mitral cells

m.l.e., membrana limitans externa

m.l.i., membrana limitans interna

n.Il.; n.op., nervus opticus n.ol., nervus olfactorius n.vn., nervus vomeronasalis 111., neurilemma nuclei np.gen., geniculate neuropil


np.pt., pretectal neuropil nuc.ac, nucleus accumbens septi nuc.amg., nucleus amygdalae mic.amg.d.l., nucleus amygdalae dorso lateralis nuc. B., nucleus of Bellonci nuc.B.n., neuropil of Bellonci nuc.caiid., nucleus caudatus nuc.cb:, nucleus cerebelli nuc.covi.C, nucleus commissuralis of Cajal nnc.d.f.l.t., bed-nucleus of decussation of

f.lat.t. nuc.d.f.m.t., bed-nucleus of decussation of

f.med.t. nuc.Dark., nucleus of Darkschewitsch nuc.ecmam. = a.ii.p. nuc.f.sol., nucleus of fasciculus solitarius II lie. 1 1 1., nucleus of oculomotor nerve niic.inp., nucleus interpeduiicularis nnc.inp.n., neuropil of nuc.inp. nuc.is.c, nucleus centralis isthmi nuc.IV., nucleus of trochlear nerve nuc.l.s., nucleus lateralis septi nuc.m.s., nucleus medialis septi nup.ol.a.; nuc.ol.ant., nucleus olfactorius

anterior nuc.ol.a.d.; nuc.ol.ant.d., nucleus olfactorius

anterior, pars dorsalis nuc.ol.ant.m., the same, pars medialis nuc.ol.ant.v., the same, pars ventralis nuc.ol.d.L, nucleus olfactorius dorsolateralis

= p.pir. nuc.oLp.tr., nucleus of olfactory projection

tract nuc.p.t., nucleus posterior tecti nuc.po., nucleus preopticus nuc.po.a., nucleus preopticus, pars anterior nuc.po.p., nucleus preopticus, pars posterior nuc.pt., nucleus pretectalis nuc. Sep., nucleus septi nuc.sep.m. = nuc.m.s. nuc.tr.ol.h., nucleus of tractus olfacto habenularis nuc.tub.p., nucleus tuberculi posterioris;

pedmicle nuc.v.l, nucleus ventrolateralis of cerebral

hemisphere; amygdala nuc.V.m., motor nucleus of trigeminus nuc.V.mes., mesencephalic nucleus of trigeminus nuc.V.s., superior nucleus of trigeminus nuc.vis.n., neuropil of nuc.vis.s. nuc.iris.s., nucleus visceralis superior; superior

gustatory nucleus

o.s.c, organon subcommissurale ol.p.tr., olfactory projection tract


P., paraph ysis

p.d.hi/th., pars dorsalis hypotliakmi

p.d.hytk.d., dorsal lobe of p.d.hijth.


■p.d.hyth.v., ventral lobe of p.d.hyfh.


p.d.th., pars dorsalis thalami

p.d.th.m., pars dorsalis thalami, area medialis

p.d.th.p., pars dorsalis thalami, area posterior

p.g.c, periglomerular cells

p.hip., primordium hippocampi

p.i.d.; p.i.th., pars intercalaris diencephali

p.p.d., primordium pallii dorsalis

p.pir., primordimii piriforme = nuc.ol.d.l.


p.v.hyth., pars ventralis hypothalami

p.v.hyth.a., anterior lobe of p.v.hyth.


p.v.hyth.p., posterior lobe of p.v.hyth.


p.v.th., pars ventralis thalami

p.r.fh.a., pars ventralis thalami, area anterior

p.r.th.p., pars ventralis thalami, area posterior

par., paraphysis

pars.d.L, pars dorsolateralis of hemisphere = p.pir.


pars.v.l., pars veutrolateralis of hemisphere = c.s.


ped., pedunculus cerebri

pl.c.r., plexus chorioideus rhombencephali

po. (4),(6),(5), fibers from postoptic commissure entering tegmental fascicles

prorn.l., prominentia lateralis

prom.r., prominentia ventralis

r.c.c.b., ramus communicans cum arteria

basilaris r.c.p., ramus communicans posterior r.h., ramus hypothalamicus r.h.m., ramus hemisphaerii medialis r.h.v., ramus hemisphaerii ventralis r.III., root of oculomotor nerve r.IV., root of trochlear nerve r.IX., root of glossopharyngeal nerve r.IX.m., motor root of IX nerve r.IX.v.s., visceral sensory root of IX nerve r.L; reel., recessus lateralis rhombencephali r.mes.s., ramus mesencephali superior r.p.m., recessus posterior mesencephali r.po., recessus preopticus r.sp., root of spinal nerve r.sp.v.l., ventral root of first spinal i'lerve r.V., root of trigeminal nerve r.V.asc, ascending fibers of sensory V root r.V.m.S., posterior motor V root T.V.mes., mesencephalic root of V nerve r.V.mes.do., dorsal division of r.V.mes. r.V.mes.v., ventral division of r.V.mes . r.V. mot., motor root of V nerve r.V.sen., sensory root of V nerve r.V.sp., spinal root of V nerce r.V I., root of abducens nerve


r.V II., root of facial nerve r.VILl.L, lateral-line roots of VII nerve r.VII.l.l.d., dorsal lateral-line VII root r.VII.l.l.m., middle lateral-line VII root r.VII.l.l.v., ventral lateral-line VII root r.VII.m.; r.V II. mot., motor root of VII

nerve r.VII.m.2., posterior motor VII root r.VII.v.s., visceral sensory root of VII nerve r.V 1 1 1., root of VIII nerve r.VIII.d., dorsal root of VIII nerve r.VIII.v., ventral root of VIII nerve r.X., root of vagus nerve r.X.l.l., lateral-line roots of vagus nerve r.X.l.l.d., dorsal lateral-line root of vagus r.X.l.l.r., ventral lateral-line root of vagus r.X.v.s., visceral sensory root of vagus reed., recessus lateralis rhombencephali rec.op., recessus preopticus rec.op.l., recessus opticus lateralis rec.p.m., recessus posterior mesencephali rec.pcm., recessus precommissuralis rec.pin., recessus pinealis

s., limiting sulcus of nuc.tnb.p.


s.d., sulcus dorsalis thalami

s.erh., sulcus endorhinalis

s.hyth., sulcus hypothalamicus

.s.hyth.d., sulcus hypothalamicus dorsalis

s.hyth.p., sulcus hypothalamicus posterior

s.ih., sulcus intrahabenularis

s.is., sulcus isthmi

s.l.b.o., sulcus limitans of olfactory bulb

s.l.h., sulcus limitans hippocampi

sd.s., sulcus limitans septi

s.lat.mes., sulcus lateralis mesencephali

s.m., sulcus medius thalami

S.O., sinus obliquus

s.po., sulcus preopticus

s.rh., sulcus rhinalis

s.shab., sulcus subhabenularis

s.st., sulcus striaticus

s.st.c, sulcus strio-caudatus

S.V., sulcus ventralis thalami

s.v.a., sulcus ventralis accessorius thalami

sac.d., saccus dorsalis

sac.v., hypophysis, pars nervosa

Sep., septum

sep.ep., septum ependymale

sep.m., septum mediale

sep.t.p., septum transversum paraphysis

st.amg.f., strio-amygdaloid field

st.ep., ependymal layer

st.glom., layer of glomeruli

st.gr., granular layer

st.m.c, layer of mitral cells

st.mol., molecular layer = st.plx.

st.n., layer of nerve fibers st.plx., plexiform layer str.med., stria meduUaris thalami str.t., stria terminalis

t.c, transitional cells

t.f., taenia fornicis

t.th., taenia thalami

t.v.q., taenia ventriculi quarti

teed., dorsal thickening of tectum

tect., tectum mesencephali

teg.d., tegmentum dorsale mesencephali

teg. is., tegmentum isthmi

teg.is.m., tegmentum isthmi, pars magnocellu laris teg. v., tegmentum trigemini teg.VII., tegmentum facialis fh., thalamus

tr.a., dorsal correlation tract of Kingsbury tr.amg.hab., tractus amygdalo-habenularis tr.amg.th., tractus amygdalo-thalamicus tr.b., ventral correlation tract of Kingsbury tr.b.in., tractus bulbo-isthmialis tr.b.sp., tractus bulbo-spinalis tr.b.t.L, tractus bulbo-tectalis lateralis tr.c.h.L; tr.c.hab.l., tractus cortico-habenu laris lateralis tr.c.h.m.; tr.c.hab.m., tractus cortico-habenu laris medialis tr.c.th.m., tractus cortico-thalamicus medialis tr.cb.teg., tractus cerebello-tegmentalis tr.hab.t., tractus habenulo-tectalis tr.hab.th., tractus habenulo-thalamicus tr.hy.ped., tractus hypothalamo-peduncularis tr.hy.teg., tractus hypothalamo-tegmentalis tr.hyp., tractus hypophysius tr.inj.asc, tractus infundibularis ascendens tr.inp.b., tractus interpedunculo'-bulbaris tr.inp.b.d., tractus interpedunculo-bulbaris

dorsalis tr.inp.b.v., tractus interpedunculo-bulbaris

ventralis tr.mam.inp., tractus mamillo-interpeduncu laris tr.mam.ped., tractus mamillo-peduncularis tr. mam. teg., tractus mamillo-tegmentalis tr.mam.th., tractus mamillo-thalamicus tr.ol., tractus olfactorius tr.ol.d., tractus olfactorius dorsalis tr.ol.d.l., tractus olfactorius dorsolateralis tr.ol.h.a.; tr.ol.hab.ant., tractus olfacto habenularis anterior tr.ol.h.a.v., ventral division of tr.ol.h.a. tr.ol.h.l.; tr.ol. hab.lat., tractus olfacto-habenu laris lateralis tr.ol.h.m.; ir.ol.hab.vied., tractus olfacto habenularis medialis


tr.ol.hip.m., tractus olfacto-hippocampalis medialis

tr.ol. pal.d. = tr.ol.d.


tr.ol. pal.l. = tr.ol.d.l.


tr.ol.ped., tractus olfacto-peduucularis

tr.ol. s., fasciculus olfactorius septi

tr.ol.r., tractus olfactorius ventralis

tr.ol.v.l., tractus olfactorius ventrolateralis

tr.op., tractus opticus

tr.op.ac.p. = tr.op.b.


tr.op.ax., axial bundle of tractus opticus

tr.op.b., basal bundle of tractus opticus

ir.op.l; tr.op.lat., tractus opticus lateralis, or ventralis

tr.op.m.; tr.op. vied., tractus opticus medialis, or dorsalis

tr.op., mar., tractus opticus marginalis

tr.ped.mam., tractus pedunculo-mamillaris

tr.pc, tractus preopticus

tr.pt. hy., tractus pretecto-hypothalamicus

tr.pt.tec, tractus pretecto-tectalis

tr.pt.th., tractus pretecto-thalamicus

tr.sep.c, tractus septo-corticalis

tr.sep.hab., tractus septo-habenularis

tr.sp.b., tractus spino-bulbaris

tr.sp.cb., tractus spino-cerebellaris

tr.sp.t., tractus spino-tectalis = Im. sp.


tr.st.ped., tractus strio-peduncularis = f.lat.t.v.


tr.st.pt., tractus strio-pretectalis

tr.st.t.; tr.st.teg., tractus strio-tegmentalis = f.lat.d.


tr..st.tec., tractus strio-tectalis tr.st.th., tractus strio-thalamicus

tr.t.b., tractus tecto-bulbaris tr.t.b.c, tractus tecto-bulbaris cruciatus tr.t.b.p., tractus tecto-bulbaris posterior tr.t.b.r., tractus tecto-bulbaris rectus tr.t.cb.; tr.tec.cb., tractus tecto-cerebellaris tr.t.hab., tractus tecto-habenularis tr.t.hy.a., tractus tecto-hypothalamicus anterior tr.t.p.; tr.t.ped., tractus tecto-peduncularis tr.t.p.c, tractus tecto-peduncularis cruciatus tr.t.pt., tractus tecto-pretectalis tr.t.sp.; tr.tec.sp., tractus tecto-spinalis tr.t.teg., tractus tecto-tegmentalis ir.t.teg.c, tractus tecto-tegmentalis cruciatus tr.t.th.h.c.a., tractus tecto-thalamicus et

hypothalamicus cruciatus anterior tr.t.th.h.c.p., tractus tecto-thalamicus et

hypothalamicus cruciatus posterior tr.th.r., tractus tecto-thalamicus rectus =

br.col. tr.teg.b., tractus tegmento-bulbaris tr.teg.inp., tractus tegmento-interpedimcularis


tr.tcg.is., tractus tegmento-isthmialis

tr.teg.p.; tr.teg.ped., tractus tegmeiito-peduncularis

tr.th.h., tractus thalamo-bulbaris

tr.th.f., tractus thalamo-frontalis

tr.th.h.d.c, tractus thalamo-hypothalamicus dorsalis cruciatus

tr.th.hab., tractus thalamo-habenularis

tr.th.mam., tractus thalamo-mamillaris

tr.th.p.; tr.tk.ped., tractus thalamo-peduncularis

tr.th.p.c; tr.th.ped.c., tractus thalamo-peduncularis cruciatus

tr.th.p.d., tractus thalarao-peduncularis dorsalis

tr.th.t., tractus thalamo-telctalis

tr.th.teg.d.c, tractus thalamo-tegmeutalis dorsalis cruciatus, with divisions A and B

tr.th.feg.r., tractus thalamo-tegmentalis rectus


tr.th.teg.v.c, tractus thalamo-tegmentalis ven tralis cruciatus tr.th.teg.r.r., tractus thalamo-tegmentalis ven tralis rectus tr.v.a., tractus visceralis ascendens tr.r.d., tractus visceralis descendens tr.v.t., tertiary visceral tract tub. p., tuberculum posterius

v.az.sep., vena azygos septi r.l., ventriculus lateralis r.m.a., velum meduUare anterius r.par., paraphysial veins r4-, fourth ventricle

z.lim.lat., zona limitaus lateralis z.lim.med., zona limitans medialis

(1) to (10), tegmental fascicles