Regarding Several Points of Doubt of the Structure of the Olfactory BulbAs Described by T. Blanesкод для вставкиСкачать
THE ANATOMICAL RECORD 291:751–762 (2008) Regarding Several Points of Doubt of the Structure of the Olfactory Bulb: As Described by T. Blanes CATHERINE LEVINE* AND ALEXANDER MARCILLO The Miami Project to Cure Paralysis, University of Miami Miller School of Medicine, Miami, Florida ABSTRACT In order to complement the studies on the ﬁla olfactoria completed by Dr. Santiago Ramón y Cajal, we have included a translation of ‘‘Sobre Algunos Puntos Dudosos de la Estructura del Bulbo Olfatorio’’ by T. Blanes, a student of Dr. Cajal. This work describes in stunning detail additional morphological aspects of the olfactory pathway, including what was at the time the modestly studied neuroglia. The neuroglia of the olfactory system has been revisited in the last several decades for its importance in the ﬁeld of regenerative neuroscience. Olfactory ensheathing glia has the unique quality of providing ensheathment to neurons which traverse from the central to the peripheral nervous system and are being used as a candidate in present-day transplantation studies to mimic this phenomenon at the dorsal root entry zone after a central nervous system injury. Although this ﬁne work has passed its centennial anniversary since initial publication, it has been widely cited throughout the years, and of recent when Pressler and Stowbridge reported Blanes cell electrophysiological recordings (Neuron V 49, 6; p 889–904, 2006). An English translation the details of what Blanes initially documented with unduplicated precision can now be made available to a wider audience in the ﬁeld of neuroscience, and is especially important now that more and more present-day studies require a precise and complete understanding of the anatomical structures contained within the olfactory system. Anat Rec, 291:751–762, 2008. Ó 2008 Wiley-Liss, Inc. Key words: olfactory system; mitral cells; granules; neuroglia The original title, ‘‘Sobre Algunos Puntos Dudosos de la Estructura del Bulbo Olfatorio’’ by T. Blanes, Pupil of Medicine at the Medical Faculty of Madrid was published in 1898 (Revista Trimestral Micrográﬁca 3:99– 127). Here, it has been translated and interpreted from the 19th Century Spanish Original by Catherine Levine and Alexander Marcillo, from The Miami Project to Cure Paralysis, University of Miami Miller School of Medicine, Lois Pope LIFE Center, 1095 NW 14th Terrace, Miami, Florida. After the fundamental and important research of Golgi (1) and Cajal (2), the ﬁne anatomy of the olfactory bulb seemed completed along general lines; the subsequent work carried out by Pedro Ramón (3), van Gehuchten and Martı́n (4), Retzius (5), Calleja (6), and Kölliker (7) has served most of all to establish the morphology of the diverse bulbar elements in animals, and Ó 2008 WILEY-LISS, INC. perfect, in some secondary aspects, our notion about the structure and functionalism of the olfactory system in man and superior mammals. The ﬁgures were reproduced and modiﬁed for clarity with the permission of, and are copyrighted to Herederos de Santiago Ramón y Cajal. *Correspondence to: Catherine Levine, The Miami Project to Cure Paralysis, University of Miami Miller School of Medicine, Lois Pope LIFE Center, 1095 NW 14th Terrace, 5th Floor, Miami, FL 33136. Fax: 305-243-2691. E-mail: email@example.com Received 17 May 2007; Accepted 11 January 2008 DOI 10.1002/ar.20680 Published online 2 April 2008 in Wiley InterScience (www. interscience.wiley.com). 752 C. LEVINE AND A. MARCILLO When it comes to the physiology of the bulb, almost all scholars, without reservation, accept the dynamic schema thought up by Cajal, because he is the only one whose ﬁndings are in agreement with the facts. For the creation of this schematic, our Professor solely took into account the examination of microscopic preparations; all bias of teaching and all theoretic commitments were dismissed beforehand. Therefore, it is not strange that the most valuable authorities of neurology, such as Retzius, Lenhossék, van Gehuchten, His, Edinger, and even Kölliker himself, (who previously refuted the conductive role of dendrites), enthusiastically accept conduction as a mechanism for olfactory excitation, based on the new facts contributed by Cajal, among these, the demonstration that the ﬁbers received from the olfactory mucosa do not pass the olfactory glomeruli, where they create a relationship through contact with the tufts of the dendritic terminals of the mitral and tufted cells. It is clear that Cajal did not have the illusion that his would be the ﬁnal word on the subject; he knows a great deal, and through unfortunate experience, he knows how far opposing resistance to new ideas can reach because of excessive narcissism, and because of culture, eminently egotistical toward the scientiﬁc consequence; moreover, it is well known that the most sovereign talents are precisely the ones that are most hostile to all innovation. Thus, he never had the pretentiousness to try and convince those illustrious scholars, which, like Golgi, are noted for passionately and prematurely defending scientiﬁc hypotheses that were completely irreconcilable with the ideas of the histologist of Madrid. Nor did he ever expect to attract to his team the Italian neurologist’s enthusiastic disciples L. Sala and Monti, both of whom had a perfectly justiﬁed affection and admiration toward their teacher, which one can presume may have allowed them to forego the serenity of the impartial judgment necessary to examine reformist doctrines. But objections themselves, wherever from, are always respectable and should all be the focus of an attentive exam. Therefore, we will try to examine in the present work all of the open questions regarding the structure and activity of the olfactory bulb, clearly and calmly studying the facts alleged by the various authors, without being persuaded by any other purpose than the enlightenment of the truth. To better achieve this commitment, we have dedicated 2 years to the study of the structure of the olfactory bulb, having, therefore, completed an inﬁnite number of preparations, as much as with Ehrlich’s method as with Golgi’s Method. To compare them with ours, we have had at our disposal all of Cajal’s and Calleja’s collection of preparations concerning the anatomy of the aforementioned organ, and we believe that, at least for the explanation of some of the problems waiting to be solved, our opinion will not be lacking in value. The material of study has consisted of rats, mice, and most of all cats, from the moment of their birth to adulthood. NATURE OF THE GRANULES OF THE OLFACTORY BULB It is known that between fascicles of anteroposterior nerves, the region of the white matter of the bulb contains numerous islets made up of small triangular or ovoid corpuscles, originating from central and peripheral expansions, among which, as already recognized by Golgi, it is not possible to ﬁnd any that possess axonal characteristics. Cajal completed a detailed study of these cells, demonstrating their perfect radial orientation and revealing an important fact: the peripheral expansion, almost always the thickest, proceeds continuously to the molecular layer and forms an elegant tuft in its thickest portion, whose appendices, notably spiny, come in contact with the dendrites of the medial and inferior mitral and tufted cells. Because this expansion possesses a ﬁxed relationship, and the body and central appendices seem to be in contact with centrifugal ﬁbers, Cajal is inclined to consider, from a functional nature, the aforementioned prolongation, according to the theory of dynamic polarization, must function like an axon and not like a dendrite. All of which naturally involves the idea of considering the granule as a species of nervous corpuscle, in a whole comparable with amacrine or retinal spongioblast cells, elements that are undoubtedly nervous, although void of a functional expansion, or rather, of differentiation into two categories of expansions. Concerning this opinion of Cajal, and any of the facts from which it is founded, several objections have been raised, of which we will demonstrate their total lack of foundation. 1. It has been said by van Gehuchten and Kölliker (7), that the peripheral tuft of the granules extends, in effect, in the vast majority the cases, as Cajal asserts, to the molecular layer; but in a few corpuscles, are observed to end on the mitral cells, that is, in the granule layer. We have carefully examined a large number of embryonic and adult granule preparations, and have become completely convinced of the reality of Cajal’s ﬁndings; the peripheral expansion always sends its tuft to the molecular layer. Certainly, in some granules, this stem does not pass the level of the mitral cells, but the careful observation with a 1.30 apochromatic objective, proves in an impervious way that, in these corpuscles, the stem has been impregnated in an incomplete mode: the end of the aforementioned expansion ends up brusquely without ramiﬁcation; and the error of van Gehuchten and Kölliker, has involved considering the set of collateral branches, as a deﬁnite terminal arborization, emitted from the stem before its artiﬁcial interruption. In our own preparations, and even in Kölliker’s drawings, we can suspect the cause of this error (see Fig. 763 in this scholar’s text), although it may have also happened that the aforementioned author may have identiﬁed a dislocated epithelial corpuscle as a granule cell. This brings us to a point to add some details regarding the morphology of the granules, details that must be kept in mind to avoid all mistakes between the granules and the epithelial cells. Body The body of the granule presents in one of two modes: smooth and bristly with small spines or short and irregular outgrowths. The smooth soma, appears in the thickness of the islets (Fig. 1A), in contact with other granules, and it is not rare to see it faceted with reciprocal pressure. The smoothness of the body of these central granules, and the scarce quantity of protoplasm that REGARDING SEVERAL POINTS OF DOUBT OF THE STRUCTURE OF THE OLFACTORY BULB 753 tory, until it reaches the mitral cells, presents quite a spiny contour during its passage through the nervous plexuses, and a smooth surface as it crosses the islets. As it passes the mitral cells, it either loses its spines, or it notably reduces them, recovering them and extraordinarily exaggerating them at the level of the terminal tuft. The collateral branches, that at a sharp angle supply some stems for the granular zone, do not end in the islets, but in the nervous bundle, and also display a spiny contour. Central Expansions Fig. 1. Granules of the olfactory bulb of the 20-day-old cat. Golgi’s Method. A: Smooth body of a granule resting in the center of a cell grouping. B: Peripheral granule with spines. C,D: Spiny granules. E: Voluminous granule. F: Peripheral granule of the mitral cell layer. G: Tiny granule located in the mitral cell layer. H: Dislocated mitral cell. surrounds the nucleus, is an indication that pericellular arborizations do not exist in the thickness of the islets. Aside from this, Cajal in his extensive work on the Nervous Centers (8), underscores this particularity, and cites the bipolar cells of the retina, in acknowledgment that somas poor in protoplasm do not tend to possess nervous nests (shells). On the other hand, all bordering granules or those lateral of the islets, like those separately grouped in the nervous plexuses, exhibit numerous points of roughness. In the immediate bordering cells, the surface that faces the interior of the islet is also smooth, and rough at the border of the nervous plexuses of the white matter (Fig. 1B). A worthy fact to also note, which conﬁrms Cajal’s aforementioned hypothesis on the importance of conduction of the perisomatic protoplasm, it is that the marginal granules and the isolated granules possess more protoplasm than the central ones belonging to the islets. There is, however, an exception. Let us ﬁnish this brief description of the soma, adding that its size can be quite variable, and in light of these large differences in diameter, it would be pertinent to separate the granules into small, medium, and large. Peripheral Expansion As it has already been well described by Cajal, Retzius, van Gehuchten, and Kölliker, we will add merely a minutia to the facts set forth by these scholars: The aforementioned expansion, in its complete trajec- In number, these appear as one, two, three, and rarely four. These expansions are also spiny, as van Gehuchten indicates, and they diverge in a fan-shape, splitting and getting lost in the thickness of the nervous plexuses. According to our observations, these branches never end at the islets, noting that, as the branches cross the islets, they present a smooth contour equal to the peripheral expansion. There is a relation between the position of the granule and the length of these expansions: the granules situated close to the mitral cells, exhibit long and richly branched central expansions, while those that are more profound, bordering the zone of the epithelial corpuscles, exhibit shorter and more rudimentary central expansions. However, these expansions are never as short, branched, and numerous as those characteristic of the epithelium or the neurological elements. Because of not attentively considering these differences, Kölliker incorrectly identiﬁed both categories of elements, that is, the epithelial cell and the granule. All of our efforts to ﬁnd a central expansion that can morphologically be considered as an axon have failed; regarding this point, therefore, we ﬁnd ourselves completely in agreement with Golgi, Cajal, Retzius, van Gehuchten, and P. Ramon. Therefore, it is no surprise that in a recent work by Hill (9), he believes to have come across the axon of the granules. The simple inspection of the ﬁgures annexed to this author’s text plainly reveal that the elements that have intruded on the axon, are not legitimate granules, but short axis–cylinder cells (also called inferior tufted cells), previously described by Golgi, Cajal, and van Gehuchten. In this way, of the two types of granules with an axon, the one that Hill cites as type 1 (Fig. 2 of his work), undoubtedly Fig. 2. Small periglomerular nerve cells of the 1-month-old cat (superﬁcial granules of Kölliker). a: Bifurcating axon. b: Cell whose axon emits collaterals descending toward the interglomerular spaces. c: Description for cell c was omitted in the original. d: Cell whose ﬁne axon appears to end at a short distance in two branches. e: A cell with three descending branches. 754 C. LEVINE AND A. MARCILLO Fig. 3. Biglomerular nerve cells of a 1-month-old cat. A: Cell that seems to branch out into three glomeruli. B,D: Biglomerular cells. C,E: Monoglomerular cells. a. Axon. correspond to short axon cells described by Cajal (see upcoming description), and those reproduced in Figure 3, under the name type 2 granules, represent stained dislocated mitral cells described by van Gehuchten. 2. Is the granule a neurological element? In light of the morphologic resemblance of the granule with the epithelial cell, of the transitions between both classes of elements, of the absence of the functional expansion, and so on, Kölliker considers, taking into account Cajal’s opinion, the olfactory bulb’s granules as legitimate neuroglial cells. With reference to histogenesis, this scholar afﬁrms that the deepest granules may originate from the epithelium and represent epithelial dislocated and somewhat modiﬁed cells, while the most superﬁcial granules would be the progeny of ectodermal cells, destined to transform directly into spider like elements or Deiters’s cells. On the origin of the granules, our studies are not sufﬁciently complete to release a sound opinion; regarding the identiﬁcation of the granule with the epithelial or neuroglial corpuscle, we will illustrate several reasons that we expect will contribute to clariﬁcation of the question at hand: 1.a It is known than Ehrlich’s method (methylene blue by injection, Dogiel’s or Bethe’s ﬁxation, and so on), does not color neuroglia, only cells and nerve ﬁbers. We ﬁnd authentic testimony of this truth in the memoirs of Arnstein, Smirnow, and, above all, of Dogiel. Nor has Cajal been able to stain the aforementioned elements with methylene blue (pure Ehrlich’s method); only with a special staining method (method of anaerobic action of methylene blue) were the epithelials of the retina and encephalons of certain vertebrates impregnated. On the other hand, mammalian Deiters’ cells always gave the appearance of being refractory, regardless of the method of methylene blue staining used. However, the aforementioned Ehrlich’s method, consistently and with great ease stains granules in the dog, rabbit, and cat, animals that we have made numerous tests in. The soma, acquires an intense blue shade, as do the expansions, a singularly descending expansion can continue on to the molecular zone, where it arrives at its terminal tuft. Only in this tuft, as Cajal has demonstrated in other cells stained with methylene blue, the spines cease to appear clearly, disappearing, without a doubt, because of the intense varicose alteration provoked by air or other conditions that remain enigmatic. The protoplasm of these granules behave, in the presence of the methylene blue, like that of nerve cells, and although they are stained much easier than the tufted cells, according to this point of view, they are comparable to retinal spongioblasts, singular, avid elements of methylene blue. 2.a There is a complete discrepancy between the neuroglial and epithelial granule. In Figure 8 (current work), the epithelial cells and neuroglia of the olfactory bulb are shown, and in Figure 1, the principal types of granules are revealed. Compare both ﬁgures and it will be evident that granules cannot be considered as neuroglial or epithelial elements. Noting the differences of general form, orientation, and connections, there are several details that defy all confusion: (a) The soma of the granule always possesses ascending expansions, long and divided into branches; the epithelial cell does not tend to possess these; (b) The granule is smooth, or at the most it exhibits some short spines, rarely divided into branches; meanwhile, the soma of the epithelial and neuroglial cell appear bristly with a vast amount of varicose appendices, very ﬁne, divided into branches, distributed throughout in partitions or small interstitial sheets, the whole of which forms about the soma, an extremely thick atmosphere of granules (Fig. 8c); (c) The granule’s peripheral stem only possesses certain short, unbranched spines, like the protoplasmic stems of the nerve cells; while the stem of the epithelial cell that shares the same name exhibits, at the level of the white matter plexuses, many laminar prolongations similar to winding shrouds divided into branches, which form about the nerve ﬁbers a system of alveoli, comparable in all its parts to those formed by the epithelial corpuscles of the retina and Bergman ﬁbers of the molecular layer of the cerebellum; (d) ﬁnally, the granule has a tiny soma, poor in protoplasm, so much so that it is translucent allowing the light brown stained nucleus to show through the cell, unless the epithelial or neuroglial cell possesses a voluminous cell body copious in protoplasm, through which the nucleus would always be imperceptible. Many other reasons challenge the opinion of Kölliker: The perfect orientation of the granules, the continuity of its relationship with the dendritic plexus of the molecular layer, and, most of all, the large number of these corpuscles existent in the olfactory bulb of all vertebrates. In fact, the granules are not missing, as Pedro Ramón has demonstrated, neither in ﬁshes, reptiles, and batrachians, where it is doubtful that genuine neuroglial corpuscles exist at any level; and indeed in these animals, instead of looking like epithelial elements (the only representation of neuroglia in the inferior vertebrates), they differ from these more so than in mammals, because they lack any type of central appendix and have a perfectly smooth body. Calleja has also seen their smooth appearance possessing only peripheral expansions in the urodelan (Pleurodeles waltl). All of the expressed facts are of great importance for the appreciation of the physiological role of the granule, and our teacher’s foresight is credited with considering this element as a homolog of retinal spongioblasts. Because granules are lacking a central expansion in ﬁshes, batrachians, reptiles, and birds, there is no other remedy than to consider them as amacrine cells, of REGARDING SEVERAL POINTS OF DOUBT OF THE STRUCTURE OF THE OLFACTORY BULB centrifugal conduction, unless the doctrine of dynamic polarization is disregarded; if it is reasoned that they conduct from the terminal tuft to the soma, it would be necessary to admit that impulses emerge from the soma and pass to neighboring nerve ﬁbers, an idea that struggles against everything that modern studies teach us concerning the dynamism of the nerve cells. The mammalian granule represents an amacrine corpuscle, which, because of reaching a more elevated phase of perfection and differentiation, has augmented its collecting apparatus, adding to the soma and principal stem a collection of dendrites or expansions of centripetal conduction. If the granule possessed a genuine axon, the problem of its physiology would have an easy solution. It would simply involve admitting that the central expansions as well as the peripheral expansions have receptor-like character and that the recollected impulses come from the inﬁnite collaterals that the axons of the mitral cells and tufted cells scatter about the molecular layer and nervous plexuses of the granules. But the absence of an axon compels us to explore other explanations. One of these explanations, which is quite rational and has been mentioned in various writings by Cajal, assigns these cells to the category of retinal spongioblasts, and assumes that granules orient themselves, in the same way as these corpuscles do, in the same relation to the soma as the expansions, with centrifugal ﬁbers from which a current is propagated to the dendrites of the mitral and tufted cells. In this way, a motor impulse could be projected onto the glomeruli, an action represented by an intimate relationship between the terminal tufts and the arborization of the olfactory ﬁbers, perhaps increase the strength of the discharge of the tufted cell dendrites, with a resulting discharge brought about, having a more energetic and diffuse reception of the olfactory impulse. These are all conjectures that cannot be proven or disproven, because unfortunately the physiological methods are too unreﬁned to allow us to attempt experimental veriﬁcation. Sustained in mere analogies, its value would increase tremendously if the action of the brain on the nerve endings and dendrites of the ﬁrst junctions of the sensory nerves, recently deﬁned by Duval (10) and Manouelian (11) under the name of the nervi-nervorum theory, could be demonstrated in other nervous systems. Kölliker’s Superﬁcial Granules According to Kölliker, tiny corpuscles rest around the olfactory glomeruli. These corpuscles are pear-shaped, triangular, or stellate, and have branches that penetrate into the glomeruli, contributing even further to the intricacy of their existing plexus. Additionally, the lack of an axon in one area, and the compactness of the soma in another area, has served to incline the aforementioned scholar to consider these cells as true granules, and consequently as a variety of neuroglial corpuscles. Furthermore, these cells were also seen by Cajal, who regarded them as nervous system cells and even believed to have distinguished a functional expansion in them, which would target nearby glomeruli (2). We have been able to perfectly stain these cells in many preparations of the cat, dog, and rabbit. Additionally, our teacher Cajal has recently provided us with a 755 series of slices of the bulb of the 1-month-old cat, in which these aforementioned tiny corpuscles appear by the hundred, these were perfectly stained, with the exclusion of the tufted cells. From the study of these preparations, we have concluded that the superﬁcial granules cited by Kölliker are true nervous elements, belonging to Golgi’s sensory type, designated by Cajal as short axon neurons. These cells are quite small, they do not exceed 6 to 8 microns in diameter; their shape is ovoid, spherical, or polyhedral; they preferentially rest in the contour of the glomeruli in the interglomerular spaces. They are also dispersed above and below the glomeruli, very close to the most peripheral tufts. Few reside in the glomerulus itself, as Cajal already observed, these being preferentially in the cortical layers. From the point of view of its dendrite distribution, two types of periglomerular cells can be distinguished: the monoglomerular type and the bi- and polyglomerular type. The monoglomerular type is the most common. Its dominant form is stellate, but it is not uncommon to ﬁnd it somewhat pear-shaped, semilunar, and even fusiform. The elements that are fully stellate with multiple stems tend to reside in the thickest area of the glomerulus, preferentially in the cortical regions, while the pyriform variety lies externally in the interglomerular spaces. The scarcity of protoplasm is so much so that the nucleus can be perfectly distinguished. The dendrites, which alternate between ﬁve or six in number, penetrate and spread throughout the glomerulus, they branch out proliﬁcally, and generate a dense plexus of ﬁne, nonspiny varicose bunches. Some, however, present expansions, like villi, that is to say, are covered with relatively long appendices, sprouted at an acute angle. Each glomerulus can possess branches containing four, ﬁve, or more of these elements (Fig. 3C). There are cells whose branches are moderate in secondary and tertiary branching, only occupying a limited area of the glomerulus; other elements possess more intricate ramiﬁcations. The biglomerular type tends to reside between the glomeruli. They generally consist of stellate or fusiform cells, which can sprout into two stems, each of which can branch out into a nearby glomerulus (Fig. 3D), or three stems, of which two can distribute themselves in one glomerulus and the third in another (Fig. 3B). The mode of ramiﬁcation is analogous to the aforementioned monoglomerular cells; all of its branches are destined to the thickest area of the glomerulus, which is an important detail that allows the separation, at ﬁrst glance, of certain elements of the inferior tufted cells which were well described by Cajal. The axon is very difﬁcult to stain and even more difﬁcult to track, not only because of its thin structure, but also because of its irregular and contoured trajectory. The rarity of the axon being stained and the difﬁculty of recognizing it, in many cases among the mare magnum (Great Ocean) of the ﬁne dendrites sprouting from the soma, have been without a doubt the reasons for which Kölliker has not managed to distinguish it. We must, however, confess that, even in the best preparations, a good number of these corpuscles do not show even the slightest sign of an axis–cylinder; but because this expansion appears clearly in cells that are exactly the same in size, position, and behavior as those that do not 756 C. LEVINE AND A. MARCILLO Fig. 4. Peripheral tufted cells and periglomerular cells. A,B: Axons of peripheral tufted cells with peripheral collaterals. C: Axon of a periglomerular cell. D: An axon that delivered initial branches toward the periphery of the glomerulus of origin. E: Cell whose axon appears to end up inside a distant glomeruli. exhibit it, is impossible to dismiss the idea that the lack of an axon is simply a defect of staining. Figures 2–4, reveal the details of the pathway and branching of the axon of the superﬁcial granules. In general, the axon sprouts from the peripheral or smooth end of the soma, makes one or two initial turns and then becomes more or less horizontal, contouring through the glomeruli, and frequently underneath these, at times reaching very long distances. In these axons, we have never come across the tendency to penetrate into the molecular layer, which is an important ﬁnding that separates them from the sprouts of the tufted cells, which as we know, possess a pathway that is resolutely central. At its horizontal trajectory through the granules, the ﬁne axons of the periglomerular elements supply an intricate course of delicate collaterals (Figs. 2, 4), the majority of which connect with the glomeruli, spreading and branching out in its interstices. We have thought that, in some cases, these collaterals made their way into the glomeruli and arborized in them (Fig. 4D). We would not, however, afﬁrm this resolutely, as it is difﬁcult to determine, in Golgi’s preparations, when the glomeruli appear without staining, if a ﬁber passes on top or through them. The collaterals as well as the terminal branch make multiple turns, and after doing so end up at a varicosity, an arborization deprived of large branches, well situated in the glomerular interstices, either under or on top of these. Some axons bifurcate near their origin, as occurs in Figure 2a, and theses branches follow a path in the opposite direction of each other, supplying collaterals for groups of distant glomeruli. Other times, the division of the axon generates a branch situated inside and another outside of the glomerulus, both of equal direction and endowed with collaterals. Regarding the axon and its multitude of collaterals, it would be pertinent to distinguish two types of cells: the moderately robust, whose very long and horizontal axon supplies collaterals that branch into a row of glomeruli (Fig. 2b,e); and the smaller sort, which have delicate axons that are difﬁcult to stain, which abandon themselves in the immediate interglomerular interstices, after having emitted only two or three collaterals. What dendrites do the nervous arborizations of the aforementioned cells connect with? For us, it is without a doubt that the majority of the mentioned nervous branches ﬁnd themselves destined toward the body and dendrites of the inferior tufted cells, which, although in small number, inhabit the area between the glomeruli, and even between these and the olfactory ﬁber layer. In this way, the excitation gathered by the peri- or intraglomerular cells in the glomerular layer, would be able to spread out transversely reaching long distances, involving in the conduction many tufted elements that are very far apart from each other. In summary, the superﬁcial granules of Kölliker, correspond to what Cajal has called association cells (cerebellum and brain) and what may be better termed dispersion cells or dissemination of impulse cells. Glomerular interstices possess other nerve ﬁbers as well. In Figure 4, we have reproduced some of these, emanating from the axons of the most peripheral tufted cells. These collaterals, which can be designated as the peripheral collaterals of the tufted cells, sometimes rise from the border of the molecular layer to the interglomerular spaces, and pass transversely thorough these during long trajectories, giving off few small branches, resembling arborization, in the aforementioned interstices. At times, the axon of the tufted cells initially runs horizontally, passing two or three glomeruli, and before turning into the molecular layer, generates a peripheral or interglomerular collateral (Fig. 4A,B). To summarize, the inter- or intraglomerular spaces contain a plexus of nerve ﬁbers, the majority being nonmyelinated, composed of three factors: (1) peripheral collaterals of the most superﬁcial tufted cells; (2) collaterals and terminals of the granules, or small periglomerular corpuscles, both small and large; and (3) a horizontal trajectory, which is at times something recurrent when stemming from the initial portion of the axon of the tufted cells. Because these horizontal trajectories are observed mostly under the glomeruli, between the glomeruli and the molecular layer, and because we also ﬁnd the long axons of certain periglomerular corpuscles in this area, a nervous plexus is constructed here, which is rich, and partially medullated, as we see in the Weigert-Pal preparations, whose dominant ﬁbers advance parallel to the cortex, in all directions, but always in the same plane. NERVE CELLS OF THE GRANULAR LAYER According to the evident research of Golgi and Cajal, the nervous plexus of the granular layer accommodates various nervous corpuscles, of which, we will be able to distinguish by the behavior of the axon in several varieties of cells: (1) Golgi cells, that is, corpuscles whose axon arborizes in the nervous plexuses of the white matter; (2) Cajal cells, namely, those whose axons reach the periphery, branching out into the molecular layer; and (3) dislocated mitral cells or cells of van Gehuchten. All of these cell types appear in our preparations. We will not circumstantially report on these cells to avoid repetition of explanations that have been masterfully presented by the aforementioned scholars. At this time, we will simply present a succinct review. REGARDING SEVERAL POINTS OF DOUBT OF THE STRUCTURE OF THE OLFACTORY BULB Fig. 5. Cells of the olfactory bulb of the 20-day-old cat. a–c: Dislocated mitral cells. d: Short axon cells with numerous dendritic expansions. e: Conventional mitral cell whose axon emanates from a dendrite. f: Small mitral cell that supplies a delicate tuft into the glomerulus. g,h: Small tuft emanating from a dislocated mitral cell. Dislocated Mitral Cells These cells were very concisely mentioned by van Gehuchten without detail; these cells deserve to be described in detail, because in our opinion, its morphology can shed light on the functional value of the spiny appendices found on dendrites, at the same time they conﬁrm the law of saving space established by Cajal, which offers us noteworthy information about the retina and the cerebellum. The ﬁrst thing that calls one’s attention about these cells (which lie in various planes of the internal molecular layer, and even in the nearby white matter plexuses), is that they are not found in all mammals, or at least their distribution varies appreciably among diverse species. We have yet to ﬁnd them in the rabbit, rat, or mouse. In Cajal’s collection on these animals, not even a corpuscle of this sort is seen. On the other hand, they are quite abundant in the cat, to the point of almost touching laterally in certain areas of the olfactory bulb (Fig. 5b,c). They must be quite scarce in the dog, as, in many preparations, only two have been stained, and even these rest not far from the row of legitimate mitral cells. The dislocated mitral cells faithfully conserve the disposition and connections of the dendrites of the con- 757 ventional mitral cells. One of the expansions of these cells, generally the most robust, directs itself to the periphery, crosses the mitral cell and molecular layers, and ends at an intraglomerular tuft. In its journey through the superﬁcial molecular layer, it can extend collateral branches, which are connected with the tufts of the deep granules, throughout this zone. (1) The internal or lateral dendritic expansions often follow an arch of variable radius reaching the mentioned molecular layer and distributing themselves in the same way as the conventional mitral horizontal dendrites. (Those who do not believe in the conductive property of dendrites must ponder upon this fact that conﬁrms, even in the dislocated cells, that the connections do not change.) No protoplasmic branch seems to be destined to gather impulses at the deep molecular layer or in the nervous compartments, which proves to be particularly interesting. Additionally, no dendrite possesses spiny collateral appendices. This last circumstance is all the more notable because all of the corpuscles at the granular layer, including the Golgi and Cajal stellate cells, are bristly with innumerable spines. What must we infer from this? From our point of view, this structure teaches us two things: (1) The dislocated cells have not physiologically adapted to the chance environment in which they live, that is, they do not possess the characteristic spines of all of the cells in the granular layer; (2) Because the aforementioned dislocation does not contribute a functional adaptation, its signiﬁcance can only be to economize space, which accounts for the dislocation itself, because there is already an excess of mitral corpuscles making it impossible to arrange all of them in a single row, and because there is also a deﬁciency in the number of granules and stellate cells in the deep molecular layer. In our opinion, this last alternative is the most probable, and precisely in the cat, the layers that are most proximal to the tractus, that is, the ones with the least glomeruli, granules, and stellate cells, are the most copious in dislocated mitral cells. Nature, thus, in this case, has ﬁlled a gap and obeyed the law of economy of space, and according to what the research of Cajal and Lenhossék has taught us, this is fervently shown in the retina. The complete lack of spines, in the dislocated mitral cells, as in the conventional mitral cells, leads us to believe that dendrites do not need spiny appendices to carry a current. We, therefore, consider unacceptable Berkley’s conjecture, which states that the nervous surge would exclusively penetrate the collecting apparatus by means of the cited appendices. It would be wiser to ascertain, according to the accepted facts, that, all dendrites, even those without spines, possess receiving capability. The spines would only expand the surface area of the entrance of the nervous stimuli, permitting a protoplasmic stem to connect with far away nervous ﬁbers. In this manner, the range of action of dendrites would increase notably, and more cells would be able to receive the nerve impulse. The axon of the dislocated mitral cells behaves entirely like the axon of the conventional mitral cells; it almost always sprouts from an internal dendrite, of acriform density, conﬁrming the law of economy of protoplasm indicated by Cajal, and proceeds toward the axis of the bulb, getting lost in the nerve ﬁber layer, not without previously having supplied collaterals for the inter-granular plexuses. 758 C. LEVINE AND A. MARCILLO Fig. 6. Small nerve cells situated in the granular layer not far from the epithelium. A: Cell whose axon is distributed amid the deepest granules. B: Cell whose axon is inwardly directed, although it branches out in the same zone. a Axon. Cajal Cells or Descending Axon Cells These appear in our preparations consistent with Cajal’s description. They consist of cells that are generally found in the third peripheral area of the granular layer. Its form is spindle shaped or triangular, and some of its expansions, which are quite spiny, proceed toward the peripherals, while others proceed toward the center. The peripherals principally distribute themselves in the deep molecular layer, occasionally reaching the mitral corpuscle layer. The axon often proceeds from the principal peripheral expansion in accord with the law of economy of the trajectory, discovered by Cajal, it advances undivided until reaching the deep molecular layer, in which it branches a great deal. Ultimately, the branches that are formed develop into varicose arborizations that surround the mitral cell bodies and connect with the dendritic plexus of the superﬁcial molecular layer. plays, and its notably dentate appearance. Nevertheless, the Belgian scholar has not managed to impregnate the axon, and thus, has not been able to establish the morphologic character of these nerve cells. We have been more fortunate, being able to pursue this expansion in many cells, and within a considerable area. As it can be seen in Figures 7 and 5d, this axon generally proceeds in a parallel manner extending to the layers of white matter, or bifurcates supplying numerous branches, ﬁnalizing within the nervous intergranular plexuses. In the axon shown in Figure 7, the majority of the nervous collaterals were going to these plexuses; only one branch descends to the deep molecular layer, although it does not pass the inferior border of this layer; in cell C, Figure 5d, the axon was very long and proceeded horizontally close to one third of a millimeter, spreading out into various branches. In Figure 5, only a portion of its trajectory has been reproduced. In summary: All of the short axis–cylinder cells of the granular layer can be classiﬁed in three groups: (1) cells whose axon connects with the deepest granules; (2) cells whose axon is associated with the external and medial granules; (3) cells (Cajal cells), whose axon forms relationships in the molecular layer with the protoplasmic plexus formed by the mitral cells, the tufted cells, and the terminal arborization of the granules. Do the above-mentioned short-axon cells receive impulses from centrifugal ﬁbers or from interbulbar associated axons? It is actually impossible to answer this question. The issue at hand is, anyway, very possible, because the dendrites of these cells are spiny and they come in contact with the nervous intergranular plexuses; but nowadays, there is no method to delimit, between the ﬁbers of these plexuses, the ones that are coupled to the short axon corpuscles. If it were possible to discover which ﬁbers these were, then nothing would be easier than to explain the ﬂow of current: The nerve Golgi Cells Two types of these short axon corpuscles show up in our preparations. A small corpuscle, with few expansions, that resides in the superior third of the granular layer, often bordering the epithelial layer; and a much more voluminous cell, with spiny expansions, housed in the middle or inferior third of the aforementioned layer. The small type or deeply located appears in Figure 6A,B. Spherical, triangular, or fusiform, it is characterized by the frugality of its dendrites, its small stem, slightly larger than the thick granules, and by the varicose and minimally spiny aspect of its protoplasmic branches. The soma is bare and seems to inhabit within the assemblage of the granules. Its axon is ﬁne, and proceeds in various directions in short distances, it intricately branches, generating a plexus of ﬁne and granulose nervous branches that seem associated to the soma and central stems of the deep granules. The large type, or peripheral, type has been seen positively by van Gehuchten, who took notice of the extraordinary number of the peripheral dendrites that it dis- Fig. 7. A large short-axon cell situated not far from the deep molecular layer. a: Axon. b: Row of mitral cells. REGARDING SEVERAL POINTS OF DOUBT OF THE STRUCTURE OF THE OLFACTORY BULB 759 impulse that spreads out from the cerebral cortex or from another part of the encephalon would go to the soma and the dendrites of the previously cited corpuscles, from these it would propagate to the granules, and ﬁnally, it would go up to their peripheral tufts, to the mitral and tufted cells, with the purpose of causing an indescribable action (Inhibition? Current loading? Protoplasmic contractions?). NEUROGLIA OF THE OLFACTORY BULB The neuroglia of the olfactory bulb has been studied by Golgi, and in particular, by van Gehuchten and Kölliker, but the descriptions given by these scholars are rather incomplete and cannot serve to appreciate the role that neuroglia play in the function of the bulb. The nervous–neuroglial connection was also of interest, especially in light of the hypothesis thought up by the Cajal brothers regarding the function of insulation of the neuroglial appendices, and in light of Weigert’s conjecture that neuroglia is exclusively an organic ﬁlling, which is the role it plays in the bulb, where the ﬂow of current is apparent, and where special zones exist that are destined for contact connections, that is the neuron–glia connection. The purpose of the present study is to review the nervous–neuroglial connection. It is understood that neuroglia may not be completely developed in newborn animals and that its arrangement may be different in the adult; therefore, we have impregnated olfactory bulbs of the adult, and the 2- and 3-month-old dog and cat. These studies have convinced us that, in fact, neuroglia, like epithelial cells, become modiﬁed in the adult stage; because of this detail, the ﬁgures that Kölliker, van Gehuchten, and even Golgi have drawn of the interstitial structure of the bulb weakly correspond with the results that we have obtained; the aforementioned authors have preferentially worked with newborn animals or postnatal animals of just a few days old. To study the neuroglia, we have divided the bulb into the following layers: olfactory ﬁber layer, glomerular layer, molecular layer, mitral cell layer, granular layer, and the subependymal layer or layer of epithelial cells. Olfactory Fiber Layer This region contains an enormous quantity of neuroglial cells, as Golgi has shown in his experiments. They are large, stellate, or fusiform, and rest between the strands of olfactory ﬁbers in the thickness of this layer preferentially close to the pia mater (Fig. 8A). The most peripheral of these tend to take on a conical or pyramidal shape with a superﬁcial base and an internal apex. There are two classes of expansions (cell processes): (1) The basal or peripheral, which are robust, granulose, and contain intervals of plate-like crests, which penetrate into the nervous bundles to separate and isolate the olfactory ﬁbers. Among all of these expansions, occasionally one is found that is more robust than the others, which terminates below the pia, amid an enlargement. (2) The central expansion, thick at its base, conforming into a bunch of unbranched smooth ﬁbers, distributed within the thickness of the molecular layer, which pass in between the glomeruli (Fig. 8A). This cell type most certainly belongs to neuroglial cells Fig. 8. Neuroglial cells of the olfactory bulb of a 2-month-old cat. A: Neuroglial cells of the nerve ﬁber layer. D: Interglomerular neuroglia. B,C: Neuroglial cells whose expansions arborize within the glomerulus. E: Neuroglial cells of the molecular layer. F: Neuroglia of the mitral cell layer. G: Neuroglia of the deep molecular layer. H,I: Dislocated epithelial cells. J: Epithelium proper. of Kölliker with long outgrowths, the only type that appears to stain, as demonstrated by Cajal, using Weigert’s method. Perhaps it may have been better to consider these two as a mixed type comparable to the description given by Terrazas (12) in the granular layer of the cerebellum, because its internal expansions are smooth (except at its base, which can have small plates or collateral ﬁns) and external varicosities supply laminar expansions. Glomerular Layer This region encompasses two types of neuroglia: intraglomerular cells and peri- or interglomerulares cells. The intraglomerular cells (Fig. 8C) are robust; stellate; lie in diverse segments of the thickness of the glomeruli; have ﬁve, six, or more robust expansions; are bristly with spines and branched plates; and are exclusively distributed in the glomerular territory. With a good apochromatic objective, there is evidence that these collateral appendices constitute a kind of sponge or system of 760 C. LEVINE AND A. MARCILLO alveoli, in which portions of nerve ﬁbers and intraglomerular protoplasmic extensions, which should not enter this area and form a reciprocal relationship, house themselves. It is without a doubt that the gaps that are seen in the intraglomerular nervous plexus that are completely impregnated positively correspond to the bodies of these neuroglial cells. These have been seen but incorrectly described by Golgi, as there is no mention of their essential characteristic, which is the presence of the laminar collateral expansions. The periglomerular cells are very robust, elongated, and for the most part reside outside of the glomeruli, between these and the olfactory ﬁbers (Fig. 8B,D). Some dwell between the glomeruli and even below the molecular layer. They possess the following: basilar expansions, which are few in number, notably fuzzy, and destined to separate among themselves the periglomerular nerve cells and even the deepest strands of olfactory ﬁbers; a thick central stem that quickly adapts into a terminal tuft of spongy or alveolar ramiﬁcations. This tuft penetrates into the thickness of the glomeruli sometimes extending one, two sometimes, and even three ramiﬁcations (Fig. 8B); in other corpuscles, it can be seen crossing an interglomerular interstitium, passing the subjacent molecular layer, spilling out within, and generating a very thick plexus with expansions of equal nature from native neuroglial cells, or better yet, a system of alveoli in which the ﬁbers and cells of this molecular layer would be partially blocked-in (Fig. 8D). Molecular Layer This region possesses stellate neuroglial cells of the same variety as previously mentioned neuroglial cells, the only difference is that they are considerably smaller, and often their size affects their ideal stellate shape (Fig. 8E). Some lie in the thickness of the molecular zone; other cells inhabit the external contour of this layer; additionally, some are in close proximity of the mitral corpuscle zone. Moreover, as indicated by van Gehuchten, the molecular region is also where atrophic tufts from epithelial corpuscles are located. Mitral Cell Layer This region scarcely contains neuroglial elements, which are stellate and somewhat radially stretched. They are generally in the same plane or at the borders of the layer. The expansions, which are covered with appendices and laminar outgrowths, are distributed in a complex plexus in the two contiguous molecular layers; only a few laminar expansions, which are slightly extended, seem destined to the mitral cell interstices, and their reciprocal contacts are avoided because of the presence of the peripheral stems of the epithelial cells. Granular Layer This region encompasses two types of neuroglial corpuscles: The dislocated epithelial cells, and the short outgrowth neuroglial corpuscles. The dislocated epithelial cells have been seen by all of the authors who have applied Golgi’s method, and they were described in most detail by van Gehuchten, but he, without a doubt, worked with newborn animals and was thus unable to observe the adult phase of these elements, which differs in some essential features, which correspond to the ﬁrst stages of postnatal development. In the Figure 8H, we have reproduced some of these corpuscles that were found in the 2- to 3-month-old cat, that is to say, an age when the neuroglial elements are completely or almost completely formed. The cell body lies in various planes of the layer that we are studying, it is principally located in the central region of the granular layer, lacks an internal prolongation, and is surrounded by a halo of granulose appendices, which are divided into branches that are so close to each other that it is difﬁcult to tell them apart, even at a high magniﬁcation. From the peripheral portion of the layer, a thin, but initially thick, stem sprouts out and crosses the entire granular layer and the mitral cell layer to ultimately end in the various molecular layers, with tufts of robust and varicose branches (Fig. 8) that resemble webbed feet. This endpoint, which is established in newborn animals above and between the glomeruli, seems to be retracted in the adult, as it has been found in the external limits of the molecular layer and even in the actual granular layer (Fig. 8); although in certain cells a defect of impregnation cannot be excluded. The feature that is primarily found in the dislocated epithelial adult cell is the quantity of branched collateral expansions, destined for the nerve ﬁber layer. These collateral branches often proceed at a right angle, they integrate themselves between the nerve ﬁbers and nearby granular expansions in the aforementioned plexuses, and give off a spongy frame, similar the one described by Cajal in the molecular layer of the cerebellum. From time to time, and generally corresponding to the central limit of a given group of granules, the appendices multiply and lengthen, generating a sort of rippled tuft whose appendices run within the aforementioned limit. Additionally, other appendices, which are more or less smooth and radial, pass in between the cell bodies of the granules to impede contact between them. Each epithelial cell can form two or more ornamental borders of this type, the latter of which inhabits the internal molecular layer. The stems of the epithelial cells, are smooth as they pass through the granular cells, and expand their ramiﬁcations in the nervous plexuses; this is done in such a way that the vast majority of the expansions seem destined to ﬁll the existent intervals between the inﬁnite protoplasmic and unmedullated nervous branches that rest in the stratum of the granular and molecular layers. The other cell type corresponds to the neuroglial category of short hair-like radiations: These preferentially live in the intergranular layer (Fig. 8G) and when its expansions pass through the granules, they are more smooth and more moderately branched than when they integrate themselves between the prolongations of the granules. Subependymal Layer or Layer of Epithelial Cells In mammals, below the ependyma, exists a thick zone exclusively inhabited by the bodies of the epithelial cells. These exhibit an ovoid body, which is either smooth or contains small outgrowths, a central expansion with some crests and accommodation protuberances, and a peripheral prolongation, which is initially curved, to accommodate the soma of analogous corpuscles, and REGARDING SEVERAL POINTS OF DOUBT OF THE STRUCTURE OF THE OLFACTORY BULB which, once it arrives at the granular layer, behaves in the same manner as the soma of the dislocated epithelial cells (Fig. 8J). In summary, the presence of the central expansion extended to the ependyma, the low number of outgrowths of the soma, peripheral stem structure, and subependymal disposition, are features that, at ﬁrst glance, allow the distinction between these and the epithelial dislocated cells. Additionally, in the adult dog and cat, the thick, deep strands of myelinated nerve ﬁbers contain stellate cells with smooth and long radiations, a ﬁnding that is equivalent to ﬁndings in man. The ﬁndings regarding neuroglial disposition that we have just described seem to favor the opinion of the brothers Cajal more so than of Weigert; however, with present-day science, it is impossible to formulate a deﬁnite tenet. We see, in fact, in the area where the olfactory ﬁbers lie, the quantity of neuroglial ﬁbers is enormous, existing to impede the contacts between the ﬁbers, there is a multitude of crests and collateral appendices of accommodation. The same occurs in the glomeruli, interglomerular spaces, and molecular layers. The contacts between the innumerable dendritic expansions that are concentrated in these plexuses and the unmedullated nervous ﬁbers, in a similar manner as the aforementioned neuroglial intersticial plates, consisting, as previously described, by a sort of sponge that groups together these ramiﬁcations. Brieﬂy, the insulating neuroglia is present in all of the areas of the bulb where conductive ﬁbers are, and, where the risk of an uncontrolled diffusion of the olfactory impulse is greater. Neuroglia, therefore, has in this case, the task of channeling currents, isolating from one another all of the expansions that pertain to systems of independent conduction. On the other hand, the corpuscles of long outgrowths would dominate at the regions formed exclusively by medullated tubes and where the aforementioned channeling is not necessary. DO NERVE FIBERS OF BULBAR ORIGIN THAT PENETRATE INTO THE GLOMERULI EXIST? The doctrine of dynamic polarization, created by Cajal, based on the way the sensory nerves terminate, has a great line of reasoning in the olfactory bulb, according to van Gehuchten, His, Edinger, Lenhossék, Retzius, and Kölliker. The fundamental fact, favorable to said doctrine, is the following: In the glomeruli, there is exclusive branching by the olfactory nerve ﬁbers without passing further within the structure, and only the protoplasmic expansions of the mitral and tufted cells arrive to gather the olfactory impulse deposited in these territories by the ﬁbers. Because no nerve ﬁber, collateral or terminal, coming from the torso of the olfactory bulb penetrates the glomerular area, there is no other recourse than to allow transmission by contact between the olfactory ﬁbers and the aforementioned protoplasmic appendices. The asserted facts, and the explanation founded in them, are so very clear and easy to demonstrate, that all authors have given them credence, less those compromised by the defense of personal theories. Consequently, is it no surprise to see that Golgi and his disciple A. Monti are still adamant that the protoplasmic expansions that penetrate into the glomeruli, do not serve for impulse conduction, and the olfactory current is delivered exclusively by long axons and nervous col- 761 laterals. In fact, Golgi (13), in a minimally impartial, critical article, reviewing new work on the olfactory bulb, recalls the existence of recurrent nervous collaterals, which sprout from the axon of tufted corpuscles and from other cells deep in the olfactory tract, have their endpoints in the glomeruli, where they would anastomose with the olfactory ﬁbers. These collaterals would not have been, according to the Italian scholar, stained in Cajal’s preparations. A. Monti (14), a devoted disciple of Golgi, in a brief composition on the subject, refutes the contact theory, asserting to have seen outgrowths from the axons of small paint brush-like cells (inferior tufted Cajal cells) arrive at the recurrent glomerular collaterals. These collaterals form a scaffold within the glomerulus, which continues peripherally with the olfactory ﬁbers. With such insistence, in the presence of a great determination to defend at any cost, a doctrine that has been unanimously rejected by the most notable histologists, we would simply be able to refrain from the subject, and state only that we have never managed to see the referred to recurrent ﬁbers, despite having examined some hundreds of high-quality preparations, in which, uncountable olfactory ﬁbers showed up completely impregnated. Moreover, as we cannot but assume that Golgi and Monti proceed with good faith, and they have positively seen possible elements that would induce them to ponder the existence of the aforementioned connection, we have tried to ﬁnd the ﬁbers that these scholars have identiﬁed as nerves and recurrent ﬁbers. These ﬁbers are varied and of diverse nature. They are as follows: Collateral Fibers From Inferior Tufted Cells In fact, there exists, as we have said previously, certain recurrent collaterals of tufted cells, but these ﬁbers arborize exclusively between the glomeruli and never within them. An analysis of ﬁne slices in alcohol or in xylol to avoid such excessive transparency that the balsam or the Damar resin imparts onto the glomeruli, allows the true topography of the aforementioned collaterals to be distinguished. Axon collaterals of Mitral Cells These can go up to the molecular layer, but they never enter the glomeruli nor do they reach the border of the glomerular layer. Terminal Tufts of the Epithelial Cells In newborn or postnatal mammals of just a few days old, radial ﬁbers are sometimes found that end on or within the glomeruli by means of a tuft of strongly varicose branches. These ﬁbers, when thin or do not stain completely until its arrival at its cell of origin, could give rise to errors of interpretation. Fine Dendritic Appendices Emanating From Dislocated Mitral Cells and From Those in Their Usual Position Also the aforementioned stems, which branch out frugally within the glomeruli, could appear, upon a superﬁcial examination, like centrifugal nerve ﬁbers (Fig. 5g,h). 762 C. LEVINE AND A. MARCILLO In summary, a nerve ﬁber that originates in the bulb, or other major center, would never end up within the glomeruli. Although a comparable termination is occasionally found, we have gone through hundreds of preparations, asking: ‘‘How does this ﬁnding oppose the contact theory and the doctrine of dynamic polarization?’’ Within this doctrine, these ﬁndings are easily explained by simply visualizing that certain tufted cells propagate electric current in two modes, directly and indirectly. That is, directly, by means of the axon, and indirectly, or through its collaterals, which at the level of the glomeruli, enters into a contact relationship with the dendritic tufts of other corpuscles of the same type. Golgi and Monti’s doctrine demonstrates its weak foundation by not giving any importance to the most essential facts of the structure of the bulb. To believe in these scholars would lead to the following outcome: (1) that millions of mitral cell, medial and inferior tufted cell, dendritic expansions end in the glomeruli in intimate contact with the olfactory ﬁbers, and are only an adornment with no other function than to serve as an eternal enigma for neurologists. Moreover, the wellknown mitral cells would not serve a function, despite generating the largest ﬁbers from the external root of the olfactory nerve, because—as Kölliker points out— with Golgi’s doctrine they would be excluded of all olfactory conductive function. (2) Millions of olfactory ﬁbers come to an end within the glomeruli and then the rare and scarce recurrent Golgi and Monti ﬁbers emerge from the glomeruli (because it is impossible from all points of view that they could have found many ﬁbers that appeared to have an endpoint in the glomeruli), resulting in the current arriving at the olfactory mucosa, being discontinuous at the borders of the olfactory apparatus. Let us point out, that the aforementioned recurrent ﬁbers are completely missing in the inferior vertebrates, in which, as demonstrated by the brothers Cajal, Calleja, Edinger, and others, glomeruli exist and represent the only juncture between the olfactory ﬁbers and the protoplasmic tufts of the bulbar cells; which reveals, the great importance of the nervous–protoplasmic connection that occurs in the glomeruli, present in all vertebrates. In conclusion, Golgi and Monti’s opinion does not have concrete facts to support it, but false interpretations. In order for this doctrine to be formally discussed, it would be necessary that these scholars get preparations where these two ﬁndings appear: (1) the existence of glomeruli in a vertebrate, where no dendritic expansion that comes from a mitral or tufted cell, is penetrating the glomeruli; (2) the presence of olfactory ﬁbers that, piercing the borders of the glomeruli, continue without interruption by way of axons or nervous collaterals of tufted cells. Without trying to be prophetic, we believe that we are not mistaken when predicting that neither Golgi nor his disciples would have the opportunity to contradict the referred facts. Recently Manouelian (15), while defending Duval’s ingenious theory concerning the nervi nervorum, believed to have found centrifugal nerve ﬁbers in the thickness of the olfactory bulb, possibly coming from the brain, and that terminated in the glomerulus. By way of these centrifugal ﬁbers, which Manouelian compares to the ones found by Cajal in the retina, the nerve centers would act on the nervous–dendritic articulation of the glomeruli, regulating the intensity of, or facilitating the ﬂow of current. We will not refute Manouelian’s assertion, which has the merit to ﬁt well within the existing doctrine about the functionality of the senses; but, we must confess that we have not managed to come across the aforementioned ﬁbers. Could it be that the aforementioned scholar has mistaken ﬁne dendrites of dislocated mitral cells for nerve ﬁbers, like, for example, the ones that we have depicted in Figure 5g? To pursue this issue, it would be necessary to not only consider the description and drawings given by this author, but also examine his preparations and compare them to ours. Therefore, we reserve our opinion on this point, in anticipation of new reports. ACKNOWLEDGMENTS Excerpts from this work was previously presented in a poster entitled ‘‘The origin of the ﬁrst cranial nerve: The ﬁla olfactoria according to Santiago Ramón y Cajal and T. Blanes’’ by C.L. at the 2006 International Meeting of the Cajal Club in Stockholm, Sweden, at The Nobel Forum, Karolinska Institutet in celebration of the Centennial Anniversary of the Nobel Prize jointly awarded to Santiago Ramón y Cajal and Camillo Golgi in Physiology or Medicine (1906) for their work in the ﬁeld of neuroscience. Funding for attendance at the International Meeting of the Cajal Club was provided to C.L. by the International Brain Research Organization (IBRO). Illustrations by T. Blanes have been digitally enhanced for color and quality. Original Title: Sobre algunos puntos dudosos de la estructura del bulbo olfatorio. T. Blanes, Revista Trimestral Micrográﬁca 3, 99–127 (1898). LITERATURE CITED 1. Golgi. Sulla ﬁna struttura dei Bulbi olfatorii. Reggio-Emilia, 1875. 2. Cajal. Origen y terminación de las ﬁbras nerviosas olfatorias. Diciembre, 1890. Barcelona. 3. P.Ramón. Estructura de los bulbos ópticos de las aves. Gaz. Sanit. De Barcelona. Julio 1890. El encéfalo de los reptiles, Septiembre de 1891. Barcelona. 4. Van Gehuchten y Martı́n. Le bulbe olfatif de quelques mammiferes. La Cellule, t. VIII, 2 fasc. 1891. 5. Retzius. Biol. Untersuchungen Neue Folge, 1892. 6. C.Calleja. La región olfatoria del cerebro. Madrid, 1893. 7. Kölliker. Lehrbuch del Gewebelehre, 6 Auﬂ. Bd. II, 1896. 8. Cajal. El sistema nervioso de los vertebrados. Fasci. I, 1897. 9. A.Hill. Notes on granules. Brain. Vol. 77 y 78. 1897. 10. Mathias Duval. L’amoeboisme du systeme nerveux. La théorie histologique du sommeil. Les nervi nervorum. Revue Scientiﬁque, 12 Mars., 1898. 11. Manouelian. Societé Biologie, 19 Fevr. 1898. 12. Terrazas. Notas sobre la neuroglia del cerebelo, etc. Rev. trimestral micrográﬁca. Tomo 11, 1897. 13. Golgi. Ergebnisse der Anatomie und Entwickelungsgeschichte, 11 Band, 1892. Articulo Nervensystem. Wiesbaden, 1893. 14. Monti. Sulla ﬁna Anatomia del Bulbo olfattorio. Fatti vecchi e nuovi che contradicono alla teroria dei Neuroni. Paris, 1895. 15. Manouelian. Societé de Biologie, 19 fevrier 1898. Also see L’anne psychologique, Paris, 1898. p. 447, and Deyber’s Thesis. Etat actuel de la question de l’amoeboidisme. Steinheil. Paris, 1898.