The Inferior Olivary Nucleus of the Opossum (Didelphis marsupialis virginiina), Its Organization and Connections GEORGE F. MARTIN,' R. DOM,2 J. S. KING,3 M. R o B A R D S 4 AND C. R. R. W A T S O N 5 D e p a r t m e n t of A n a t o m y , T h e Ohio State University College of Medicine, C o l u m b u s , Ohio 43210; 2 Department of A n a t o m y , Medical University of S o u t h Carolina, 80 Barre Street, Charleston, S o u t h Carolina; 4 Department of Neurological Surgery, T h e University of Virginia School of Medicine, Charlottesville, Virginia a n d 5 School of A n a t o m y , University of N e w South W n l e s , K e n s i n g t o n , N . S . W. 2033, Australia

ABSTRACT

Although the inferior olivary nucleus of the opossum is small, sections stained either for Nissl substance, normal axons or cholinesterase activity reveal distinct medial, dorsal and principal nuclei. The medial nucleus contains three major subdivisions (labelled a, b, c after Bowman and Sladek, '73) and a group of neurons which is comparable to the c a p of Kooy. In contrast to the cat and monkey, the major portion of the "medial" nucleus (subgroup a) lies lateral to the principal nucleus in rostral sections. The dorsal nucleus can also be subdivided, as can the principal nucleus which contains distinct dorsal and ventral lamellae. A small area is identified which based on position and connections may conform to the dorsal medial cell group. The experimental portion of the study provides evidence for an olivary projection from the motor-sensory cortex and a massive input from the midbrain (red nucleus, pretectum, midbrain tegmentum). I n addition, the opossum inferior olive receives fibers from the deep cerebellar nuclei (cerebellar feedback loops), the spinal cord and the dorsal column nuclei. Of particular interest is the finding that fibers from the nucleus cuneatus and nucleus gracilis have distinctly different olivary targets and that those from the nucleus gracilis, but not the cuneate nucleus, overlap (in part, a t least) with the direct spinal fibers. Other examples of overlapping fields of terminal degeneration are present and are discussed. In general our results reveal that although certain relationships between the nuclear divisions are different, the opossum olive conforms well to that of placental mammals and provides a basic mammalian model for future experimental electron microscopic and physiological studies.

The mammalian inferior olivary nucleus is divided generally into medial and dorsal accessory nuclei and a principal nucleus. Although Kooy ('17) thought he could identify such divisions in the opossum, subsequent investigators have used either a somewhat different terminology (Oswaldo-Cruz and Rocha-Miranda, '68) or chose not to make comparisons with other mammals (Bowman and King, '73). The primary goal of this report is to present architectural, histochemical and connectional data which clarifies the organization of the opossum olive and provides a basis for future ultrastructural (King et al., '75) and physiological studies. J. COMP. NEUR.,160: 507-534.

The connections of the inferior olivary nucleus have been studied most extensively in the cat (see review of the literature by Armstrong, '74), although contributions have been made using other species (Mehler, '69; Jane and Schroeder, '71 ; Schroeder and Jane, '71; Courville and Otabe, '73; Miller and Strominger, '73; Mizuno et al., '73), including the opossum (Mehler, '69; Hazlett et al., '72; Martin, '73; Dom et al., '73). However, except for the early work on the cat (Brodal et al., '50; Walberg, '56, 'SO), the terminations of multiple olivary inputs have not been compared and contrasted in a single species. For that reason and in light of certain new findings in our 50 7

508

MARTIN, DOM, KING, RoBARDS AND WATSON

own laboratory (Martin, '73; Dom et al., '73) and that of others (Graybiel et al., '73; Mizuno et al., '73; Edwards, '74; Graybiel, '74), it seemed that a re-evaluation of these connections was in order. MATERIALS AND METHODS

The conformation of the opossum inferior olivary complex was examined in Nissl stained sections from several normal brains cut in frontal, sagittal and horizontal planes. Additional brains were stained either for cholinesterase activity (the Lewis, '61, modification of the Koelle technique, '50) or for normal axons by a non-reduced variation of del Rio Hortega's silver carbonate method (Scharenberg and Liss, '69). In order to gain information concerning its descending inputs, the inferior olivary nucleus was examined for axonal degeneration in brains with lesions of the cerebral cortex (25 animals), the caudate nucleus (4 animals), various areas of the midbrain (40 animals), the pontine reticular formation ( 3 animals) and the deep cerebellar nuclei (44 animals). In three specimens, pon tine hemisec tions were attempted to provide information regarding the degeneration present after interruption of all fibers from sources rostral to the lesion. Additional brains were studied with lesions in either the spinal cord (14 animals) or the dorsal column nuclei (7 animals). Since many of the above mentioned cases were used for prior studies, several silver methods were employed to demonstrate axonal degeneration. In most instances, however, modifications of the FinkHeimer method ('67) were used and every fifth section was stained for Nissl substance in order that the terminal target of the fibers in question could be interpreted in light of the nuclear conformation. Postoperative survival times varied from two days to three weeks depending, in large part, on the system being studied. In this report terminal degeneration is used to denote aggregates of argyrophilic spheroids or in some cases, fragmented axons which have left their main trajectory and course in an obviously random fashion within the neuropil. As a control for our interpretation of terminal degeneration with the light microscope the distribution of spinal fibers was verified by experimental electron microscopy (King et al., '75).

RESULTS

In order that the targets of its various inputs can be described clearly, the conformation of the opossum inferior olivary nucleus will be discussed prior to the account of its afferent connections. A s indicated by Bowman and King, '73, the identification of subnuclei in the opossum cannot be made readily without information concerning their connectivity. For this reason, brief reference is made to a few key connections in the initial account, although they are described in more detail in the second part of the results.

The organization of the opossum inferior olivary nucleus The shape of the opossum inferior olivary nucleus was described by Bowman Abbreviations a, subnucleus a of medial accessory ofivary complex b, subnucleus b of medial accessory olivary complex bc, brachium conjunctivum bp, brachium pontis c, subnucleus c of medial accessory olivary complex CAUD., caudal CcD, dorsal cochlear nucleus d, dorsal accessory olivary nucleus Fast, fastigial nucleus Hg, hypoglossal nucleus inp, interpositus nucleus K, c a p of Kooy LAT., lateral LES, lesion MED., medial pr, principal olivary nucleus pr.d., dorsal lamella of principal olivary nucleus Prt, pretectal nucleus pr.v., ventral lamella of principal olivary nucleus Pyr, pyramidal tract RN, red nucleus ROST., rostral Fig. 1 An artist's drawing of the dorsal aspect of the opossum inferior olivary nucleus made from a scaled reconstruction. T h e arrows A-L demonstrate the level of t h e sections d r a w n and comparably labelled in figure 3. Fig. 2 Drawing of t h e lateral aspect of the reconstruction shown in figure 1. Fig. 3 Drawings of selected transverse sections of t h e opossum olive. T h e level of each section (A-L) is indicated on the reconstruction in figure 1. Fig. 4 Photomicrograph of t h e caudal olive at the level shown by line C i n figure 1. T h e bar measures 0.5 m m in this and all subsequent photomicrographs. Nissl preparation. Fig. 5 Photomicrograph of a section comparable to that in figure 4, but stained for cholinesterase activity.

THE OPOSSUM INFERIOR OLIVARY N U C L E U S

509

510

MARTIN, DOM, KING, RoBARDS AND WATSON

and King ('73) and drawings of their reconstruction are used for orientation to the transverse sections (figs. 1, 2, 6, 7, 14, 15). For clarity, the nucleus is described as seen in transverse sections progressing from caudal to rostral. At its caudal end the inferior olivary nucleus first appears as a small aggregate of neurons just dorsal to the medial lemniscus. As the first group enlarges, a second comes into view medial to it (fig. 3A) and still a third population appears laterally (figs. 3 B , 4). Using the terminology of Olszewski and Baxter ('54) and Bowman and Sladek ('73), the three subnuclei just described will be referred to as subnuclei a, b and c (from lateral to medial) of the medial accessory nucleus. They were considered collectively as the ventral lamella by Bowman and King ('73). Their identification as parts of the medial accessory complex is borne out by the existence of spinal input to the lateral two cell groups (fig. 30A, (B), since material in our laboratory (unpublished) shows degeneration in similarly positioned nuclei of the rhesus monkey (cell groups a, b of Bowman and Sladek, '73) after cervical hemisection. A fourth group of neurons (the dorsal cell column of Bowman and King, '73) soon appears (d, figs. 3C, 4) which is closely related to the emerging fascicles of the hypoglossal nerve. Its dorsal position suggests that it is a part of the dorsal accessory nucleus (Kooy, '17). Sections stained for cholinesterase (Che) activity (fig. 5) reveal that the most lateral division of the medial accessory nucleus (group a) is strongly positive for the reaction (both neurons and neuropil) and that three additional divisions can be identified: (1) the intermediate area described above (group b) which stains somewhat lighter (neurons and neuropil) than group a, (2) a medial region (subnucleus c) in which both neurons and neuropil show only a moderate staining for the reaction product and (3) a small dorsomedially situated cell group which stains intenseIy and appears to be equivalent to the dorsal cap of Kooy ('17) described in a number of species. The dorsal accessory nucleus is also strongly positive for Che activity, a characteristic which helps to identify i t throughout its caudal to rostral extent (figs. 5, 10, 13, 17, 21). In progressively more rostral sections

the medial accessory complex begins to bend (fig. 3D) assuming the shape of a shepard's crook (figs. 8, 9). A s in more caudal sections, subnucleus b is traversed by fibers (fig. 9) and its neurons are generally scattered and light staining (fig. 8). At such levels an area of compact neurons appears between subnucleus b (as just defined) and subnucleus c. For purposes of this account these cells (asterisk, fig. 8) will be considered as a dorsolateral part of subnucleus c, although they could just as readily be described as a dorsomedial extreme of subnucleus b. The cap of Kooy (part of the dorsal medial extension of Bowman and King, '73, figs. 3D,E, 8) can be distinguished from subgroup c, especially in sections stained for Che activity. The dorsal accessory nucleus is separated from the rest of the complex in most sections, but occasionally shows continuity with subnucleus a and/or b. In more rostral sections the dorsal nucleus becomes totally separated from subnucleus a, and subnucleus c becomes replaced by the principal nucleus (figs. 3E, 10). Another division of the dorsal accessory nucleus appears medially which soon expands to meet the portion described above (fig. 3F,G). Even before the two parts form a continuum, as seen in the reconstruction (figs. 1 , 6 , 14), they are occasionally bridged Fig. 6 Artist's reconstruction of t h e dorsal aspect of the opossum inferior olive made from a scaled reconstruction. T h e numbered lines indicate the levels shown in figures 8-13. Fig. 7 Drawing of t h e lateral aspect of t h e reconstruction shown in figure 6. Fig. 8 Photomicrograph of a transverse section through the opossum olive rostral to that shown in figures 4, 5 (see fig. 6 for level). This i s the level of the shepard's crook described in the text and the star indicates t h e dorsolateral division of subnucleus c. Nissl preparation. Fig. 9 Photomicrograph of a section comparable to that shown in figure 8, but stained by a modification of del Rio Hortega's silver carbonate method for nerve fibers. Fig. 10 Photomicrograph of a section rostral to that shown in figures 8, 9 and stained for cholinesterase activity. Note the presence of a principal nucleus. Fig. 11 Photomicrograph of a transverse section through the olive j u s t rostral to that shown in figure 10. Nissl preparation. Fig. 1 2 Photomicrograph of a section at t h e level shown in figure 11, but stained by a modification of del Rio Hortega's silver carbonate method. Fig. 1 3 Photomicrograph of a section comparable to that in figures 11 and 12, but stained for cholinesterase activity.

THE OPOSSUM INFERIOR OLIVARY N U C L E U S

51 1

512

MARTIN, DOM, KING, RoBARDS AND WATSON

by cells and at such levels the entire olivary complex assumes a more or less trilaminar appearance in transverse sections (fig. 11). The dorsal lamina is composed entirely of the dorsal accessory nucleus (often appearing as two or three neuronal aggregates) which can usually be distinguished from remnants of the cap of Kooy (figs. 3F, 11, 12,13). The ventromedial lamina is entirely co-extensive with the C shaped principal nucleus in which separate dorsal and ventral lamellae become identifiable (figs. 3F, 11-13). At the same level, most of the area intermediate between the dorsal and ventromedial laminae (intermediate lamina of this account) is continuous with subnuclei a and b of the medial complex (a conclusion supported by all techniques, figs. 1113). However, in some sections the dorsomedial end of the intermediate lamina has somewhat different cell density and staining characteristics (fig. 1l), appearing similar to the dorsolateral part of subnucleus c more caudally (compare regions indicated by asterisk in figs. 8 and 11). In rostral sections, the inferior olivary complex assumes the “C” shaped configuration emphasized by Bowman and King (‘73) (figs. 3H, I, 16, 17). Based on position, histochemistry and input, the dorsal lamina of Bowman and King is comparable to the dorsal accessory nucleus and their ventral lamina includes a portion of the medial nucleus laterally (rostral continuation of subnucleus a, fig. 16) as well as the dorsal and ventral lamellae of the principal nucleus. Because of its small size the principal nucleus is recognized best by the input from the red nucleus to its dorsal lamella (fig. 19, see Edwards, ’72; Miller and Strominger, ’73 for review) and by its histochemistry (fig. 17). The dorsal accessory nucleus is particularly large at this level and in some sections four divisions can be discerned. A dorsolateral portion is readily identified by the presence of degeneration after spinal lesions, whereas its ventromedial area is clear (fig. 18). The ventromedial extension of the dorsal nucleus just described (figs. 3G,H,I, 16, 17) appears to be comparable to a similar area described in the monkey by Bowman and Sladek (‘73). Occasionally this portion of the dorsal nucleus widens and connects with the dorsal lamella of the principal nucleus (fig. 17). In some sections, neurons appear ventral

to the ventromedial bend of the dorsal nucleus (figs. 3H, I, 17; asterisk) which are continuous with the ventral lamella of the principal nucleus, and based on their position, may be comparable to the dorsomedial cell column (see Bowman and Sladek, ’73 for review). A second neuronal aggregate appears ventral to the one just mentioned (a, fig. 31; open arrow, fig. 17) which is continuous by a thin strand of cells (solid arrow, fig. 17) with the part of subnucleus a that lies lateral to the principal olive. In view of this continuity and because its connections differ somewhat from those of the proposed dorsal medial cell column (fig. 28F), we have tentatively labelled the second group (fig. 17, open block arrow) as a part of subnucleus a. The entire dorsal accessory complex is intensely positive for cholinesterase activity (neuronal somata and neuropil). On the other hand, the neuropil of the principal nucleus typically stains very lightly, causing the positively reacting neurons to stand out in relief (pr.d. and pr.v., fig. 17). Subnucleus a and its ventromedial extension stain about equally and both have a neuroFig. 14 Drawing of the dorsal aspect of the opossum inferior olive made from a scaled reconstruction. The numbered lines indicate the level of the sections in figures 1 6 2 1 . Fig. 15 Drawing of the lateral aspect of the reconstruction shown in figure 14. Fig. 16 Photomicrograph of a transverse section through the olive (level shown in fig. 14) at the level of the c shaped configuration. Nissl preparation. Fig. 17 Photomicrograph of a section comparable to that in figure 16 (slightly more rostral), but stained for cholinesterase activity. The arrow on the left indicates the cell strand between the two divisions of subnucleus a and the arrow on the right shows the medial representation of subnucleus a. The asterisk indicates the extension of the ventral lamella beneath the dorsal nucleus. Fig. 18 Photomicrograph of the degeneration impregnated in the lateral part of the dorsal accessory nucleus after a C-2 spinal lesion. The level is comparable to that shown in figures 16, 17. Fink-Heimer technique. Fig. 1 9 Photomicrograph of a section through the opossum olive at the level of figures 16, 17 showing the impregnation of axonal degeneration in the dorsal lamella of the principal olive after a large lesion of the red nucleus. Fink-Heimer technique. Fig. 20 Photomicrograph of a transverse section through the opossum olive at its rostral end. Nissl preparation. Fig. 21 Photomicrograph of a section comparable to that in figure 20, but stained for cholinesterase activity.

T H E OPOSSUM INFERIOR OLIVARY NUCLEUS

513

514

MARTIN, DOM, KING, RoBARDS AND WATSON

pi1 which is darker than that of the principal nucleus (a and open block arrow, fig. 17). At the rostral tip of the olive only portions of the dorsal and principal nuclei can be recognized with certainty (figs. 3J-L, 20, 21). Afferent connections of the inferior olivary nucleus Descending connections. In cases subjected to unilateral destruction of the neocortex numerous degenerating fibers leave the pyramidal tract rostral to the ipsilateral inferior olivary nucleus and sweep over its rostral pole to more dorsal and caudal targets. Although some of them course through the lateral part of the dorsal accessory nucleus, their terminal character is questionable. A large number of degenerating fascicles course medially and laterally through the rostral olive and some of them terminate within the lateral and medial portions of subnucleus a and within the extension of the ventral lamella beneath the dorsal accessory nucleus (fig. 22G). The terminals within subnucleus a are not apparent in P-342, but are present in another specimen (P-152) at levels comparable to those shown in figure 22D,E,F. More caudally, thin degenerating fibers leave the pyramidal tract and invade the dorsomedial extreme of the “intermediate lamina,” terminating more caudally within the lateral part of subnucleus c and the adjacent part of subnucleus b. Although considerable terminal degeneration is present, many fibers continue dorsally and laterally on their way to extraolivary targets. The caudal end of both subnucleus b and c are relatively free of degeneration in spite of the large size of the lesion (fig. 22A). Only minimal degeneration is present in the contralateral olive. The inferior olivary nuclei do not contain degeneration in brains with lesions restricted to either visual, auditory or preorbital areas, although some is present subsequent to involvement of the motor-sensory cortex as defined by Lende (’63a,b). Such degeneration is so sparse after small lesions, however, that it is difficult to circumscribe terminal fields. Although four brains with caudate lesions are available, little degeneration can be traced caudal to the midbrain which cannot be accounted for by cortical contamination.

Lesions limited to either the superior or inferior colliculi usually fail to produce unequivocal degeneration within the inferior olivary nucleus, but those which extend into either the pretectum or the deeper midbrain tegmentum consistently provide positive results. It was intended that the lesion in P-92 be limited to the red nucleus; however, secondary vascular involvement destroyed a large portion of the tegmentum lateral to the periaqueductal gray as well as the midbrain portion of the pretectal complex (fig. 23, insert). In this case, terminal degeneration is located within the opposite medial accessory nucleus - particularly within the area described herein as the dorsolateral part of subnucleus c and the cap of Kooy (minimal) (fig. 23B, reader’s left). In caudal sections, however, no degeneration is present within either area (fig. 23A, reader’s left). On the side of the lesion, heavy degeneration is located within the dorsal lamella of the principal nucleus, and additional terminal debris is present within portions of the ventral lamella and its extension beneath the ventromedial bend of the dorsal nucleus (fig. 23C,D, reader’s right). More caudally, coarse degenerating axons disperse among the neurons of the cap of Kooy and that portion of subnucleus c (ventromedial part) which is relatively free of degeneration on the opposite side (compare in fig. 23A,B). All of our cases with lesions involving the midbrain tegmentum and/or the red nucleus show some degeneration within the opposite medial accessory nucleus (particularly the dorsolatera1 part of subnucleus c and the rostra1 cap of Kooy, figs. 23B, 24B), but only those with extensive pretectal involvement show ipsilateral degeneration within the ventromedial extreme of subnucleus c. In two brains the lesion is limited to the periaqueductal gray (dorsal and ventral) and in each, the inferior olivary nuclei appear to be free of axonal debris. In most brains with damage to the red nucleus there is extensive terminal degeneration within the dorsal lamella of the ipsilateral principal nucleus, although in some Fig. 22 Series of drawings of transverse sections through the inferior olivary nucleus on the side of a decortication (insert). Degenerating axons, as demonstrated by the Fink-Heimer technique, are illustrated as broken lines (preterminals) and dots (terminals) and the sections are arranged in a caudal to rostTal sequence from A to G .

THE OPOSSUM INFERIOR OLIVARY NUCLEUS

W

, -

Figure 22

515

516

MARTIN, DOM, KING, RoBARDS A N D WATSON

P-92 LES

THE OPOSSUM INFERIOR OLIVARY NUCLEUS

51 7

Fig. 24 Series of stacked, selected sections through the medulla (caudal A to rostra1 E) of a brain with a lesion of the lateral red nucleus and adjacent tegmentum (insert). The side of the lesion is to the reader’s right and the method of illustrating degeneration is the same a s described for figure 22.

518

MARTIN, DOM, KING, RoBARDS AND WATSON

Fig. 25 Drawings of transverse sections through the inferior olivary nucleus (caudal A to rostra1 G) on the side of the rubral, tegmental lesion shown in the insert. Degeneration is illustrated as described for figure 22.

THE OPOSSUM INFERIOR OLIVARY NUCLEUS

cases additional degeneration is present within the ventral lamella (including its extension beneath the dorsal nucleus) and in parts of subnucleus a (figs. 23C,D, 24D,E, 25E-G, 26B-F). On the side opposite the lesion degeneration is usually noted within the dorsolateral part of subnucleus c, as described above, and occasionally within the cap of Kooy. As might be expected, the amount of degeneration and its exact position varies with the size of the lesion as well as its placement (compare figs. 23-25). Since rubral lesions either encroach upon the adjacent tegmentum or undercut fascicles emanating from it, the precise origin of degenerating fibers in areas other than the dorsal lamella of the principal nucleus (Edwards, '72) awaits analysis by other techniques. Several brains are available with extensive damage to ventromedial midbrain areas (fig. 26, insert). A s expected, degeneration is present (particularly ipsilateral, fig. 26B-F) within the cap of Kooy, subnucleus c (level between figs. 26A and B), the dorsal and ventral lamellae of the principal olive and the extension of the latter beneath the dorsal nucleus. However, additional terminal degeneration is located within regions which are relatively clean after more dorsally placed lesions. These areas include: most of the intermediate lamina (both a and b) at the level of the trilaminar configuration (fig. 26B,C), the continuation of subnucleus (a) rostral to that level (fig. 26DF) and the caudal part of the dorsal accessory nucleus (fig. 26A). Such degeneration may result from damage to neurons of the ventromedial tegmentum,6 although recent autoradiographic results (Edwards, personal communication, Conrad et al., '74) suggest that this is not so. It is also possible that some of the extra debris is due to damage of fiber tracts spared to some extent by the more dorsal lesions. Although several brains with rostral pontine lesions were examined, results from them can be summarized by referring to the case illustrated in figure 27. In that animal a large lesion was made by passing a scalpel through the lateral pons producing ipsilateral degeneration within the dorsal (heavy) and ventral lamellae of the principal olive (fig. 27C-G), the medial extension of the ventral lamellae (fig. 27E,F), cell groups a and b of the medial

519

nucleus (fig. 27A-F), the dorsolateral part of subgroup c (fig. 27B), the cap of Kooy (fig. 27B,C) and the caudal end of the dorsal accessory nucleus (fig. 27A). The fine grain degeneration plotted in subnuclei a and b after five days survival is not present in another case with a comparable lesion which survived for two weeks postoperatively. Degeneration in all of the above areas can be explained by interruption of fibers arising at midbrain levels, although a pontine contribution cannot be ruled out. Terminal degeneration is also present on the side opposite the lesion, but it appears to be due to undercutting of the superior cerebellar peduncle prior to its crossing (see following account of cerebellar afferents). In one specimen (not illustrated) we successfully hemisected the pons including the pyramidal tract and the descending brachium conjunctivum. In that brain the entire ipsilateral olive contains terminal debris - except for those portions of the rostral dorsal accessory nucleus which receive spinal input (see below). The heaviest degeneration is located within the dorsal lamella of the principal nucleus, whereas the lightest scattering of axonal debris is found in the ventromedial part of subnucleus c and the caudal end of the dorsal accessory nucleus. A large series of brains are available with destruction of various parts of the deep cerebellar nuclei. In several of them degeneration can be followed in the crossed descending brachium conjunctivum to its termination within the contralateral inferior olivary nucleus. Although in cases with large lesions there is shrinkage of the contralateral olive due to undercutting of olivocerebellar fibers, the axonal debris is obviously orthograde in character and can be traced from the descending brachium conjunctivum. In case P-177 (which serves as a representative of several cases, fig. 28, insert) axonal debris is extensive within the ventromedial part of the dorsal accessory nucleus (particularly its medial extreme, fig. 28&G) and in portions of the subjacent ventral lamella of the principal 6 Neurons in the nucleus linearis contain horseradish peroxidase after iqjections of the enzyme into the inferior olivary nucleus (Henkel, King and Martin, unpublished results). However, in each case there is spill over of horseradish into paramedian medullary areas so that we can not be positive as to the exact location of the terminals incorporating the protein.

520

MARTIN, DOM, KING, RoBARDS AND WATSON

Fig. 26 Drawings of transverse sections through the inferior olivary nucleus on the side of the largest part of the lesion shown in the insert (caudal A to rostra1 F). Degeneration is illustrated as described for figure 22.

THE OPOSSUM INFERIOR OLIVARY NUCLEUS

52 1

IP.267

Fig. 2 7 Drawings of transverse section through the inferior olivary nucleus on the side of the lesion shown in the insert (caudal A to rostra1 G ) . Degeneration is illustrated as described for figure 22.

522

MARTIN, DOM, KING, RoBARDS A N D WATSON

P-I77

Fig. 28 Drawings of sections through the inferior olivary nucleus (caudal A to rostra1 G ) on the side opposite the lesion shown in the insert. Degeneration is illustrated a s described for figure 22.

THE OPOSSUM INFERIOR OLIVARY NUCLEUS

nucleus (fig. 28F). Only a few scattered degenerating axons are present within the dorsal lamella of the principal nucleus (fig. 28E,F),but degeneration is also extensive within the dorsomedial end of the intermediate lamina (fig. 28C). Injured fibers arborize within subnucleus b of the medial complex (fig. 28A,B), but rather than coursing within the descending brachium conjunctivum, they appear to reach the olive by traversing the reticular formation. In caudal sections, additional degeneration is located within subnucleus a (fig. 28A), but it is minimal. On the side of the lesion, degenerating fascicles filter out of the reticular formation and terminate in subnucleus b with some spill over into subnucleus a. Degeneration is present within comparable caudal areas of the medial nucleus (subnucleus b, particularly) in brains with damage restricted to the caudal fastigial nucleus indicating that at least some of the fibers in that area which degenerate after larger lesions are from that source. In brains with lesions restricted to the dentate nucleus there is little evidence of degeneration in the inferior olivary nucleus. It should be noted, however, that all of the dentate nucleus was not destroyed in those experiments7 As mentioned above, knife cuts of the rostral pons (fig. 29, insert) also transect the superior cerebellar peduncle and elicit degeneration of cerebello-olivary fibers on the opposite side (fig. 29A-G). In such brains, the areas containing degeneration, are comparable to those in the cerebellar cases (compare fig. 29 with fig. 28). Ascending and local connections. In cases subjected to cervical, thoracic or lumbar spinal lesions which interrupt the ventrolateral white matter, terminal degeneration is present throughout the length of the ipsilateral inferior olivary nucleus. No such degeneration is present after sacral lesions, however. Caudally axonal debris is minimal in the dorsal accessory nucleus, but fairly extensive within the medial accessory complex (subdivisions a and b, fig. 30A-C). In a section through the trilaminar configuration, dense terminal debris fills most, if not all, of the dorsal nucleus (fig. 3OD-F) and in brains with more involvement of the ventral funiculus than that shown in figure 30, degeneration is still present within the intermediate lamina -

523

reflecting its continuity with subnuclei a and b caudally. More rostrally, the degeneration becomes restricted to the dorsal accessory nucleus stopping short of its ventral medial extension. The above targets of spinal fibers have been substantiated in experimental electron microscopic material (King et al., '74, '75) which shows that they end as small synaptic profiles containing round vesicles and that they make asymmetric contacts primarily on dendrites measuring 0 . 5 2 f i in diameter. Terminal degeneration is also present within the inferior olivary nucleus after damage to the dorsal column nuclei. In the case illustrated in figure 31 (insert), both the nucleus cuneatus and the nucleus gracilis are destroyed caudal to the obex and the lesion extends beyond the midline to involve the opposite nucleus gracilis. In sections through the caudal olive (contralateral to destruction of both dorsal column nuclei), scattered and relatively coarse degeneration is present within portions of both subnucleus a and b of the medial nucleus (fig. 31A,B particularly). At the trilaminar level, finer terminal debris is located within the dorsomedial part of the intermediate lamina (fig. 31E) and in progressively more rostral sections it fills the enlarging ventromedial extension of the dorsal accessory nucleus (fig. 31F). At its rostral end, however, the comparable part of the dorsal nucleus is relatively clear, although coarse degeneration outlines its dorsal lateral portion (fig. 31G). The latter fibers reach their target by passing around the lateral side of the olive from the medial lemniscus. In sharp contrast, on the side opposite the lesion limited to the nucleus gracilis, little, if any, terminal degeneration is present within areas other than the dorsal, lateral tip of the dorsal accessory nucleus. The brain shown in figure 32 (insert) sustained damage only to that portion of the cuneate and accessory cuneate nuclei rostral to the obex. Degenerating terminal fibers are located bilaterally within subnucleus b of the medial complex (fig. 32B-D) and more rostrallv fine terminal 7 Placements of horseradish in the inferior olive result in the presence of the enzyme in cells of the contralateral interpositus and dentate nuclei as well as within the fastigial nuclei bilaterally (Henkel, King and Martin, unpublished results). I t should be emphasized, however, that the horseradish spilled over into the paramedian medulla and perhaps accounting, at least in part, for its presence in fastigial neurons.

524

MARTIN, DOM, KING,

Fig. 29 Drawings of transverse sections through the inferior olivary nucleus (caudal A to rostra1 G ) on the side opposite t h e lesion shown in the insert. Degeneration is illustrattd a s described for figure 22.

222

S f l T I 3 f l N AXVAITO XOIXZdNI MIflSSOdO 3 H d

526

MARTIN, DOM, KING, RoBARDS AND WATSON

P-319

U

....

Fig. 31 Drawings of transverse sections through the inferior olivary nucleus (caudal A to rostra1 G ) on the side opposite the lesion shown in the insert. Degeneration is illustrated as described for figure 22.

THE OPOSSUM INFERIOR OLIVARY NUCLEUS

527

Fig. 32 Drawings of transverse sections through the inferior olivary nucleus on the side opposite the Iesion shown in the insert (caudal A to rostra1 G ) . Degeneration is illustrated a s described for figure 22.

528

MARTIN, DOM, KING, RoBARDS A N D WATSON

debris is present within the dorsomedial part of the intermediate lamina (fig. 32E) and within the ventral medial part of the dorsal accessory nucleus (fig. 32F). In contrast to the case referred to above (fig. 31), the remaining parts of the dorsal nuclei are free of degeneration. In several brains the degeneration within the ventromedial part of the dorsal nucleus was either more extensive or limited to a smaller area than that described above, suggesting that some topography exists in the cuneo-olivary projection. Although cases with damage to the reticular core of the medulla and the vestibular nuclei were examined, they are not particularly helpful because of either encroachment on the inferior olivary nuclei or cerebellar contamination. Meaningful study of possible olivary inputs from those areas must wait the application of other techniques.

only a small medial representation. In contrast, in the monkey and cat all of subnucleus a lies ventrally and medially in rostral sections. Subnucleus b extends into the area referred to herein as the intermediate lamina, but more rostrally i t is "squeezed out" by the principal nucleus. In the opossum subnucleus c can be further broken down into ventromedial and dorsolateral sectors, but it is probable that the ventromedial group is comparable to the B nucleus of the cat - based both on its position (Brodal, '40) and the presence of degeneration after lesions which include the pretectal complex (Mizuno et al., '73). It is assumed that previous investigators have included our dorsolateral part of subnucleus c with either nucleus b (monkey, Bowman and Sladek, '73) or the dorsal division of the medial nucleus (Walberg, '56; Taber, '61). Although most of the opossum medial accessory olive was referred to as "the venDISCUSSION tral inferior olive" by Oswaldo-Cruz and Organization of t he opossum inferior Rocha-Miranda, '68, they apparently recogolivary complex nized its continuity with the intermediate As suggested by Kooy, '17, medial, dor- lamina (P 7.80, Oswaido-Cruz and Rochasal and principal nuclei can be identified Miranda, '68) and, more rostrally, its posiin the American opossum. However, if fig- tion lateral to the bulk of the principal nuure 3 of this communication is compared cleus (P-7.00). The opossum inferior olive with table 26 of Kooy's published report it also contains a small group of neurons is obvious that our descriptions differ mark- which according to their position appear edly. It should be noted, however, that comparable to the cap of Kooy ('17) as deKooy dealt with a large number of verte- scribed for the cat (Waiberg, '56; Taber, brate species and did not have the advan- '61) and the monkey (cell groups d, e; Bowtage of histochemical and experimental man and Sladek, '73). data. Our delineation of the dorsal accessory In caudal sections the medial accessory olive of the opossum is not markedly differnucleus can be divided into three major ent from that of either Kooy, '17, or Ossubnuclei (exclusive of the cap of Kooy) Waldo-Cruz and Rocha-Miranda, '68, alwhich differ in their neuronal density, his- though it should be noted, that the dorsal tochemistry and connectivity. In keeping nucleus is not a uniform cell mass, either with the terminology used for similar but on the basis of its cellular groupings or its more obvious divisions in the monkey (Bow- inputs, and that it has a pronounced venman and Sladek, '73) and man (Olszewski tromedial bend in rostra1 sections, conformand Baxter, '54), they have been referred to ing to the comma shaped area described as subnuclei a, b and c from lateral to for the monkey by Bowman and Sladek medial. Similar divisions of the cat medial ('73). This part of the dorsal nucleus was accessory olive have been recognized (re- included as part of both the medial and ferred to as ventral, dorsal and B subnu- principal nuclei by Kooy, '17. One feature clei, Walberg, '56; Taber, '61) and they common to the entire dorsal nucleus, howare present in 18 other species of marsu- ever, is its strong reaction for Che activity. pials we have examined. In the opossum, The principal nucleus of the opossum subnucleus a continues throughout most olive is small and, in some sections, difficult of the olive, but rostrally it is positioned to resolve from adjacent cell groups. It is mainly lateral to the principal nucleus with not surprising that Kooy, '1 7, overextended

T H E OPOSSUM INFERIOR OLIVARY NUCLEUS

its boundaries and included areas which are revealed by contemporary methods to be part of the medial and dorsal nuclei. The dorsal lamella of the principal nucleus is particularly well outlined by degeneration when the red nucleus and/or its efferent axons are destroyed, providing one of the major criteria used for its identification (see Edwards, '72, for review of literature on the cat and Miller and Strominger, '73, for the monkey). In general, the principal nucleus was correctly delineated by Oswaldo-Cruz and Rocha-Miranda, '68. In spite of its small size and the lack of distinctness to many of its subdivisions, the opossum olive is generally comparable to that of the cat (Walberg, '56; Taber, '61) and the monkey (Bowman and Sladek, '73). Since the available data in the cat suggests that the principal nucleus projects to the lateral zones of the cerebellar cortex (Brodal, '40; Armstrong et al., '74) and these portions of the cerebellar cortex are relatively small in the opossum, a diminutive principal nucleus might be expected. Examination of marsupial material from our own laboratories and that of Dr. John Johnson (Michigan State) reveals that the relative proportions of different subnuclei show some variation, probably correlated with the development of the cerebellar cortex and the motor "specialization" of the species under consideration. A detailed study of these variations is in progress.

529

(arising within face, forelimb and hindlimb cortices) and ending within the dorsal accessory nucleus, part of the medial nucleus and the rostra1 ventral lamella of the principal nucleus (Sousa-Pinto, '69; Sousa-Pinto and Brodal, '69). Although such fibers distribute bilaterally in the feline olive, the main projection is to the side opposite their origin. In agreement with Sousa-Pinto and Brodal, '69, evidence for cortico-olivary fibers was found in the opossum only after lesions of motor-sensory face or limb cortex. We could not determine any somatotopic organization as described for the cat, but that may simply reflect the undercutting of white matter even in cases with small lesions and the small amount of degeneration present even after large lesions. Such fibers are thin in the opossum and in agreement with the physiological literature (review by Armstrong, '74), they distribute mainly to the ipsilateral olive. It should be emphasized that the greatest number of cortico-olivary fibers end within the dorsolateral part of subnucleus c (as we define it) and the adjacent part of subnucleus b where they potentially overlap with those from the midbrain tegmentum. Although it can not be considered definitive, our material provides no evidence for direct caudato-olivary connections. As emphasized by Walberg, '56 and subsequently by others (review by Armstrong, '74), the cat inferior olive receives a large number of fibers from the midbrain. Our Afferent connections of the inferior material suggests the midbrain projects to olivary nucleus all areas of the olive except for part of the Since the scope of the present study pre- dorsal accessory nucleus and provides evicludes use of a broad spectrum of survival dence for: (1) an ipsilateral projection from times for every lesion, it is possible that the pretectal complex to the cap of Kooy some connections have not been demon- and that portion of the medial accessory strated. However, our material provides in- nucleus which is apparently comparable to formation as to the pattern of olivary inputs, the B nucleus of the cat, (2) a strong conproviding a necessary base for future exper- tralateral input to the dorsolateral part of iments using the opossum, and has revealed subnucleus c which arises from dorsal tegcertain details not previously emphasized mental areas, (3) an extensive ipsilateral for any species. projection from the red nucleus to the dorIt is generally accepted that stimulation sal lamella of the principal nucleus, (4) of the sensorimotor cortex evokes climbing fibers of unknown origin which distribute fiber response in the cerebellum anterior to the ventral lamella of the principal nulobe and paramedian lobule (review by cleus and to various areas of the medial Armstrong, '74) and that the impulses are and dorsal accessory nuclei. The distribution of degeneration in the carried to the inferior olive by the pyramidal tract. In the cat, cortico-olivary fibers ipsilateral olive after pretectal damage is are described as somatotopically organized remarkably similar to that reported by

530

MARTIN, DOM, KING, RoBARDS AND WATSON

Mizuno et al. (’73). A pretecto-olivary pathway has been described by physiological techniques (Maekawa and Simpson, ’69) and has been suggested as being responsible for the climbing fiber responses in the flocculus after optic nerve stimulation. The presence of silver grains over the olive after large pretectal placements of either tritiated leucine or proline (Graybiel, ’74) is direct evidence that at least some of the olivary degeneration produced by pretectal damage is not a result of injury to fibers e n passage. It may be worth noting that the degenerating axonal fragments in subnucleus c are coarse and relatively large, a finding which suggests they may be comparable to the type B terminals described in Golgi impregnations of kitten olive (Scheibel et al., ’56). All of our cases with damage to the lateral tegmentum contain some degeneration within the contralateral inferior olive particularly within the small area we label the dorsolateral part of subnucleus c. Although Walberg, ’56 tended to negate an olivary projection from midbrain areasother than the ventral periaqueductal grey and red nucleus, subsequent reports (Mabuchi and Kusama, ’70; Mizuno et al., ’73) have reopened the question and, most recently, Edwards (’74 and personal communication) reports autoradiographic evidence for a crossed midbrain-olivary path to “. . . a discrete area in the medial accessory olive . . .” from the tegmentum. In a recent re-evaluation of his earlier work Walberg (’74) has also verified the existence of tegmento-olivary fibers. Reports such as those of Ogawa, (’39), Mabuchi and Kusama (’70) and, most recently, Walberg (’74), suggest the possibility of additional midbrain-olivary paths which they claim arise either within the nucleus of Darkschewitsch, the interstitial nucleus of Cajal or from surrounding areasH Fibers from such sources would likely be interrupted by our ventromedial lesions, perhaps accounting for the degeneration in the olive not seen after damage limited to more dorsal and lateral areas. Obviously, the question of midbrain-olivary fibers is complex and deserves careful study with some of the newer techniques available. One midbrain-olivary projection that can be accepted with reasonable certainty is that from the red nucleus to the dorsal

lamella of the principal olive (Edwards, ’72). In the opossum, such fibers arise from large-medium neurons in the rostra1 twothirds of the nucleus (Martin et al., ’74), some of which may be comparable to the “parvocellular” neurons of the primate brain (King et al., ’71; Miller and Strominger, ’73). The density of olivary degeneration present after rubral damage suggests that at least some of the terminal ramifications conform to the bushy arbors seen in kitten Golgi material (Scheibel et al., ’56). As reported previously for the opossum (Martin, ’73; Dom et al., ’73) and cat (Graybiel et al., ’73) fibers from the deep cerebellar nuclei project to the inferior olive. In the opossum, one contingent traverses the descending brachium conjunctivum, whereas the other filters through the reticular formation to its olivary targets. Although the distribution of such fibers could be described only in general terms in our previous descriptions, it appears from the present study that most of those within the descending brachium conjunctivum, distribute rostrally to the ventromedial bend of the dorsal accessory nucleus, to the adjacent portion of the ventral lamella of the principal nucleus and, more caudally, to a zone in the dorsal medial extreme of the “intermediate lamina.” Some fibers may project to more caudal portions of the medial accessory nucleus, but many of them do not course in the descending brachium conjunctivum. Although our results are in general accord with those obtained by autoradiography in the cat (Graybiel et al., ’73), there are minor differences. For example, in some of their cases (e.g., CRA-12), there is evidence for a strong projection to the dorsal lamella of the principal olive. In contrast, only a little degeneration was seen in the comparable area of the opossum olive, even after transection of the brachiurn conjunctivum (see Note added in proof). Although our previous description of direct spino-olivary fibers (Hazlett et al., ’72) did not identify their terminal sites by accepted terminology, the results of the present investigation make it obvious that they Neurons in these areas contain horseradish after placements of the enzyme within the inferior olivary nucleus (Henkel, King and Martin, unpublished results) although spread from the injection site militates against stating the exact location of terminals which incorporate the protein.

THE OPOSSUM INFERIOR OLIVARY NUCLEUS

distribute to the caudal part of the medial accessory nucleus (subnuclei a and b) and to most of the dorsal accessory nucleus ( e x cluding the part receiving cerebellar and cuneate inputs). These spinal targets conform to those described in other species (the hedgehog, Jane and Schroeder, '71 ; the tree shrew, Schroeder and Jane, '71; the rat and rabbit, Mehler, '69; the cat, Brodal et al., '50; Mehler, '69; and the rhesus monkey, Mehler, '69). Interestingly enough, such a connection has not been documented for the chimpanzee or man (Mehler, '69). In all the forms studied to date, there are non-spinal portions of both medial and dorsal accessory nuclei. Although the physiological literature documents several indirect spino-olivary pathways which supposedly utilize brainstem areas other than the dorsal column nuclei (Grant and Oscarsson, '66; Larson et al., '69; Oscarsson, '68, '69) the location of the relay neurons is yet to be determined. A direct dorsal column-olivary projection has been reported for several species (the cat, Hand and Liu, '66, Morest, '67, Ebbesson, '68; the opossum, Hazlett et al., '72; the hedgehog, Jane and Schroeder, '71; the tree shrew, Schroeder and Jane, '71) and in each the m a i n target of such fibers is described as the dorsal accessory nucleus. Our current results are comparable to those reported previously, but further suggest that fibers from the nucleus cuneatus and gracilis end in different areas of the dorsal accessory nucleus. Fibers from the nucleus cuneatus end in the rostral, ventromedial bend of the dorsal accessory nucleus, whereas those from the nucleus gracilis distribute to a smaller, more lateral (and perhaps rostral) portion of the same nucleus. I t is interesting to note that gracile fibers overlap in their termination (at least in part) with those directly from the spinal cord and that the latter arise mainly within caudal thoracic and lumbar segments (Hazlett et al., '72, present results). The cuneate target conforms to that portion of the cat dorsal accessory nucleus determined by electrophysiological techniques to receive input from the forelimbs via the dorsal funiculus-spino-olivary circuit (Armstrong et al., '74) and although not yet proven, the portion of the opossum olive receiving a projection from the nucleus gracilis probably corresponds to the

53 1

dorsal funiculus hindlimb region (discussion by Armstrong et al., '74). It is noteworthy that the dorsal funiculus forelimb area of the cat olive projects, for the most part, to the forelimb areas of the paraverma1 zone (including the paramedian lobule, Brodal, '40; Armstrong et al., '74) and that the proposed hindlimb region also relays to the appropriate cortex. When considering the role of the dorsal column nuclear input to the inferior olive it should be kept in mind that they receive spinal projections not only from primary dorsal column fibers, but also by non-primary afferents which traverse both the dorsal funiculus (the cat, Rustioni, '73) and lateral or ventral lateral funiculi (the opossum, Hazlett et al., '72; the cat, Rustioni, '73). These latter fibers in the opossum end in those regions of the dorsal column nuclei which receive pyramidal inputs (compare Hazlett et al., '72 with Martin and West, '67) and are topographically organized (Hazlett et al., '72). In our material at least, one major cerebello-olivary feedback distributes to that portion of the olive which receives indirect input from the forelimb via the nucleus cuneatus. Relatively few cerebellar fibers terminate in those regions of the dorsal accessory nucleus which receive fibers from caudal body areas - either directly or indirectly through the nucleus gracilis. Since this may not be true for the cat (compare Graybiel et al., '73 with Brodal et al., '50; Ebbesson, '68), it is either a reflection of technique or the generally poor representation of the opossum hindlimbs throughout the neuraxis (Martin and West, '67; Martin and Fisher, '68; Hazlett et al., '72) and their subservient role to the forelimbs in development (Coghill, '38) (see Note added in proof). There is considerable overlap of projections into most olivary nuclei of the opossum (e.g., the overlap of cuneate and cerebellar fibers in the ventral medial part of the dorsal accessory nucleus), suggesting the possibility of complex informational processing prior to relay into the cerebellum. Further insight into that complexity is provided by certain electron microscopic observations: (1) glomeruli (complex synaptic islands) are present, Bowman and King, '73 and (2) a large number of synapses remain normal in the direct spinal area of the olive, even after C-2 hemisec-

532

MARTIN, DOM, KING, RoBARDS AND WATSON

tion (King et al., ’75). Although the nucleus gracilis can be considered the origin for some of the normal terminals in the direct spinal area (present study), the small size of that nucleus in the opossum and the great number of normal fibers remaining suggest the presence of still other inputs to that portion of the olive. Potential sources include: (1) either recurrent collaterals of olivo-cerebellar axons or the terminals of inhibitory interneurons (Llinas et al., ’74) either of which could provide the basis for the 80-120 ms “inhibition” observed in the olive after its activation (see Armstrong, ’74 for review) and (2) projections from areas of the medulla other than the dorsal column nuclei which receive spinal input and apparently relay to the olive (Miller and Oscarsson, ’70; Armstrong et al., ’73). The techniques are now available to unravel the complex anatomy of the olive and it is our intent that the results reported herein provide the basis for such studies in the opossum. ACKNOWLEDGMENTS

This investigation was supported by the United States Public Health Service, grants NS-07410 to Dr. Martin and NS-08798 to Dr. King. The authors wish to thank Mrs. Kit Franklin and Mrs. Mary Ann Jarrel for technical help and Ms. Malinda Amspaugh for typing the manuscript. The photographic help of Mr. Gabriel Palkuti is also greatly appreciated. LITERATURE CITED Armstrong, D. M. 1974 Functional significance of connections of the inferior olive. Physiological Review, 54: 3 5 8 4 1 7 . Armstrong, D. M., R. J. Harvey and R. F. Schild 1973 Spino-olivary pathways to the posterior lobe of the cat cerebellum. Exptl. Brain Res., 1 8 : 1-18. 1974 Topographical localization in the olivo-cerebellar projection: A n electrophysiological study of the cat. J. Comp. Neur., 154: 287302. Bowman, J. P., and J . R. Sladek 1973 Morphology of the inferior olivary complex of the rhesus monkey (Mncnca mulotto). J. Comp. Neur., 152: 299-316. Bowman, M. H., and J. S. King 1973 The conformation, cytology and synaptology of the opossum inferior olivary nucleus. J. Comp. Neur., 148: 491-524. Brodal, A. 1940 Experimentelle Untersuchungen iiber die olivo-cerebellare Localisation. Ztschr. ges. Neurol. Psychiat., 169: 1-153.

Brodal, A., F. Walberg and T. H. Blackstad 1950 Termination of spinal afferents to inferior olive in cats. J . Neurophysiol., 1 3 : 4 3 1 4 5 4 . Coghill, G . E. 1938 Early movements of the opossum with special reference to the walking gait. Proc. of the SOC.for Experimental Biology and Medicine, 39: 31-35. Conrad, L. C., C. M. Leonard and D. W. Pfaff 1974 An autoradiographic and degeneration study of the projections of the median and dorsal raphe nuclei in the rat. Anat. Rec., 178: 334. Courville, J., and S. Otabe 1973 Rubro-olivarp projections in the macaque. Anat. Rec., 175: 297-298. Dom, R., J. S. King and G. F. Martin 1973 Evidence for two direct cerebello-olivary connections. Brain Res., 57: 498-501. Ebbesson, S. 0. E. 1968 A connection between the dorsal column nuclei and the dorsal accessory olive. Brain Res., 8: 393-397. Edwards, S B. 1972 The ascending and descending projections of the red nucleus in the cat. An experimental study using a n autoradiographic tracing method. Brain Res., 48: 4 5 6 3 . 1974 Fiber tracts descending from the midbrain reticular formation of the cat: an autoradiographic study. Anat. Rec., 1 7 8 : 349. Fink, R. P., and L. Heimer 1967 Two methods for selective silver impregnation of degenerating axons and their synaptic endings in the central nervous system. Brain Res., 4 : 36S-374. Grant, G., and 0. Oscarsson 1966 Massdischarges evoked in the olivo-cerebellar tract on stimulation of muscle and skin nerves. Exptl. Brain Res., 1 : 329-337. Graybiel, A M. 1974 Some efferents of the p r e tectal region in the cat. Anat. Rec., 178: 365. Graybiel, A. M., H. J. W. Nauta, R. J. Lasek and W. J. H. Nauta 1973 A cerebello-olivary pathway in the cat: A n experimental study using autoradiographic tracing techniques. Brain Res., 58 : 205-221. Hand, P., and C. N. Liu 1966 Efferent projections of the nucleus gracilis. Anat. Rec., 154. 353-354. Hazlett, J. S., R. Dom and G. F. Martin 1972 Spino-bulbar, spino-thalamic and medial lemniscal connections in the American opossum. J. Comp. Neur., 146: 95-118. Jane, J. A., and D. M. Schroeder 1971 A comparison of dorsal column nuclei and spinal afferents in the European hedgehog (Erinnceus europeaits). Exptl. Neur., 30: 1-17. King, J . S . , G. F. Martin and M . H . Bowman 1974 A light and electron microscopic study of the direct spinal receiving area of the inferior olivary nucleus. Anat. Rec., 178: 392. 1975 The direct spinal receiving area of the inferior olivary nucleus. A n electron microscopic study. Exp. Br. Research, 22: 13-24. King, J. S . , R. C. Schwyn and C. A. Fox 1971 The red nucleus in the monkey (Macnca nulattcc): A Golgi and a n electron microscopic study. J . Comp. Neur., 142: 75-108. Koelle, G. B. 1950 The histochemical differentiation of types of cholinesterases and their location in tissues of the cat. J. Pharmacol., 100: 158-179. Kooy, F. H. 1917 The inferior olive in vertebrates. Folia Neurobiol., 10: 205369. Larson, B., S. Miller and 0. Oscarsson 1969 Termination and functional organization of the dor-

T H E OPOSSUM INFERIOR OLIVARY NUCLEUS solateral spino-olivocerebellar path. J. Physiol. (London), 203: 6 1 1 4 4 0 . Lende, R. A . 1963a Sensory representation in the cerebral cortex of the opossum ( D i d e l p h i s u irg inia n u ) . J. Comp. Neur., 121 : 395404. 1963b Motor representation in the cerebral cortex of the opossum (Didelphis virginia n a ) . J. Comp. Neur., 121 : 405415. Lewis, P. R. 1961 The effect of varying the conditions in the Koelle technique. Bibl. Anat. (Basel), 2: 11-20. Llinas, R., R. Baker and C. Sotelo 1974 Electrotonic coupling between neurons in cat inferior olive. J . Neurophysiol., 37. 560-571. Mabuchi, M., and T. Kusama 1970 Mesodiencephalic projections to the inferior olive and the vestibular and perihypoglossal nuclei. Brain Res., 17: 133-136. Maekawa, K., and J. I. Simpson 1973 Climbing fibre responses evoked in vestibulo-cerebellum of rabbit from the visual system. J. Neurophysiol., 36: 649466. Martin, G. F. 1973 Projections of the cerebellum to pre-cerebellar relay nuclei in the opossum. Anat. Rec., 1 7 5 : 384. Martin, G. F., R. Dom, S. Katz and J. S. King 1974 The organization of projection neurons in the opossum red nucleus. Brain Res., 78: 17-34. Martin, G. F., and A. M. Fisher 1968 A further evaluation of the origin, the course and the termination of the opossum corticospinal tract. J. Neurol. Sci., 7: 177-187. Martin, G. F., and J. H. West 1967 Efferent neocortical projections to sensory nuclei in the brainstem of the opossum, D i d e l p h i s uirginianu. J. Neurol. Sci., 5: 287-301. Mehler, W. R. 1969 Some neurological species differences - a posteriori. Ann. N. Y. Acad. Sci., 167: 424-468. Miller, R. A., and N. L. Strominger 1973 Efferent connections of the red nucleus to the brainstem and spinal cord of the rhesus monkey. J. Comp. Neur., 152: 327-346. Miller, S., and 0. Oscarsson 1970 Termination and functional organization of spino-olivocerebellar paths. In: The Cerebellum in Health and Disease. W. S. Fields and W. D. Willis, eds. Warren H. Green, St. Louis, pp. 172-200. Mizuno, N., K. Mochizuki, C. Akimoto and R. Matsushima 1973 Pretectal projections to the inferior olive in the rabbit. Exptl. Neur., 39: 498-506. Morest, D. K. 1967 Experimental study of the projections of the nucleus of the tractus solitarius and the area postrema in the cat. J. Comp. Neur., 130: 277-299.

533

Ogawa, T. 1939 The tractus tegmenti medialis and its connections with the inferior olive in the cat. J . Comp. Neur., 70: 181-190. Olszewski, J., and D. Baxter 1954 Cytoarchitecture of the human brain stem. J. B. Lippincott Co., Philadelphia. Oscarsson, 0. 1968 Termination and functional organization of the ventral spino-olivocerebellar path. J. Physiol. (London), 196: 453478. 1969 Termination and functional organization of the dorsal spino-olivocerebellar path. J . Physiol. (London), 200: 129-149. Oswaldo-Cruz, E., and C. E. Rocha-Miranda 1968 The brain of the opossum ( D i d e l p h i s m a r s u p i alis). Instituto de Biofisica, Universidade Federal do Rio de Janeiro, Rio de Janeiro, 99 pp. Rustioni, A. 1973 Non-primary fierents to the nucleus gracilis from the lumbar cord of the cat. Brain Res., 51 : 81-95. Scharenberg, C., and L. Liss 1969 Techniques for silver carbonate impregnation of nervous tissue. In: Neuroectodermal Tumors of Central and Peripheral Nervous System. Williams and Wilkens, Baltimore, pp. 21&226. Scheibel, M. E., and A. B. Scheibel 1955 The inferior olive-A Golgi study. J. Comp. Neur., 102: 77-132. Scheibel, M., A. Scheibel, F. Walberg and A. Brodal 1956 Areal distribution of axonal and dendritic patterns in inferior olive. J. Comp. Neur., 106: 2 1 4 9 . Schroeder, D. M., and J. A. Jane 1971 Projection of dorsal column nuclei and spinal cord to brainstem and thalamus in the tree shrew, tupoicr glis. J. Comp. Neur., 142: 309-350. Sousa-Pinto, A. 1969 Experimental anatomical demonstration of a cortico-olivary projection from area 6 in cat. Brain Res., 16: 7&83. Sousa-Pinto, A., and A . Brodal 1969 Demonstration of a somato toDical uattern in the corticoolivary projection in the cat. Exptl. Brain Res., 8 : 364-386. Taber, E. 1961 The cytoarchitecture of the brain stem of the cat. J . Comp. Neur., 11 6 : 2 7 4 9 . Walberg, F. 1956 Descending connexions to the inferior olive. J . Comp. Neur., 104: 77-1 73. 1960 Further studies on the descending connexions to the inferior olive: reticulo-olivary fibers: an experimental study in the cat. J. Comp. Neur., 114: 79437. 1974 Descending connections from the mesencephalon to the inferior olive: An experimental study in the cat. Exp. Brain Res., 21: 145156.

Note added in proof: Since submission of this manuscript, tritiated leucine h a s been injected into the d e e p cerebellar nuclei of several opossums a n d evidence has been obtained for projections to t h e dorsal lamella of the principal nucleus, the spinal p a r t of t h e dorsal accessory nucleus a n d subnucleus a of t h e medial accessory complex. The distribution of silver grains varied somewhat dependent upon t h e location of t h e most heavily labelled n e u r o n s a t t h e injection site suggesting some topographical organization to t h e projection.