Volume 291, Issue 7 p. 741-750
Free Access

Origin and Endpoint of the Olfactory Nerve Fibers: As Described by Santiago Ramón y Cajal

Catherine Levine,

Corresponding Author

Catherine Levine

The Miami Project to Cure Paralysis, University of Miami Miller School of Medicine,Lois Pope LIFE Center, Miami, Florida

Fax: 305-243-2691

The Miami Project to Cure Paralysis, University of Miami Miller School of Medicine, Lois Pope LIFE Center, 1095 NW 14th Terrace, Miami, FL 33136Search for more papers by this author
Alexander Marcillo,

Alexander Marcillo

The Miami Project to Cure Paralysis, University of Miami Miller School of Medicine,Lois Pope LIFE Center, Miami, Florida

Search for more papers by this author
First published: 27 March 2008
Citations: 3

The figures were reproduced and modified for clarity with the permission of, and are copyrighted to Herederos de Santiago Ramón y Cajal.


In the late Nineteenth Century, Santiago Ramón y Cajal was able to reproduce an exceptional illustration of the Olfactory Nerve pathway and its myriad of cells, by using the Golgi Method. Dr. Cajal focused intense study on the histology of the nervous system and published a treatise on the olfactory nerve fibers and the distinct peripheral origin and central nervous system endpoint of this unique pathway. The original title of this work is “Origen y terminación de las fibras nerviosas olfatorias” published in 1890. As the original publication is in Spanish, here we provide an English translation allowing present-day English speakers to read these writings. Cajal followed the trajectory of the olfactory nerve fibers as they transitioned between the peripheral and central nervous system and was able to assert that these fibers were not continuous from the olfactory bulb to the bipolar cells that relinquish into the olfactory epithelium, but that the olfactory system was made up of various cell types each having distinct morphologies and functions. This may very well be the first definitive description of the olfactory receptor neurons and the first illustrations of the continuity of these cells throughout the olfactory pathway. These meticulous histological preparations were created by first using Camillo Golgi's potassium dichromate and silver nitrate impregnation method known as “reazione nera” or “black reaction,” where nerve cells, nerve fibers, and neuroglia could be visualized. This study exhibits the structural and functional organization of the mammalian fila olfactoria as it was investigated in centuries past. Anat Rec, 291:741-750, 2008. © 2008 Wiley-Liss, Inc.

The original title “Origen y terminación de las fibras nerviosas olfatorias” by Santiago Ramón y Cajal was published by Gaceta Sanitaria de Barcelona on 11 October 1890 and was composed of three parts: Part I (133–139 ), Part II (174–181 ), Part III (206–212 ). 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.


As it is well known, almost all authors, from the memorable research of M. Schultze ((1)), are inclined to acknowledge that the fibers of the olfactory nerve end in the nasal mucosa, continuing along with the inferior extremes of the bipolar cells, without making a connection with the epithelial or support corpuscles.

Nevertheless, the methods used to verify these nerve endings are not very demonstrative because of the fact that neither the dissociation, osmic acid, or gold chloride method used by Schultze, Krause, Brun, Ranvier, Schwalbe, Lustig, and so on, definitively reveal the continuity of an olfactory fiber with a bipolar cell.

The first author who in our opinion has tested this continuity was Arnstein ((2)), using intra vitam staining with methylene blue. But because of the short time that it takes for the selective coloration to fade away, and the difficulties that this caused, there was a need to apply another procedure that could supply preparations which were demonstrative and could be preserved over time. For this reason, in the past year, we have set out to test Golgi's method repeatedly, in the anticipation of impregnating the olfactory fibers and revealing the particularities of its pathway and endpoint. The obtained results were published with a detailed drawing in September of 1889 ((3)).

Golgi's method stains the bipolar cells of the olfactory mucosa black, allowing the nucleus that appears as brown color, to be seen behind a layer of impregnated protoplasm (Fig. 1a). Of the appendices of those cells, the ascending process is robust, flexible, and reaches up to the level of the uninhibited surface; the descending process is quite fine, varicose, exhibits a semitranslucent brown color, and crosses the basal portion of the cell, and continues along with one of the infinite varicose fibers that stem from the olfactory nerve, subsequently progressing along the subepithelial connective tissue. Often these fibers trace a curve above the bordering area, contouring around the basal corpuscles, later meandering their way to their target cells. We have not managed to discover, neither the plexus nor the ramifications that Ranvier ((4)) indicates in the intraepithelial path of the olfactory fibers. For us, a fact that is absolutely true, and devoid of all doubt, is certainty that of the perfect individuality of each olfactory fiber, as it courses through the mucosa and the nervous bundles. Additionally, another fact is that all olfactory fibers end their trajectory in the same way, that is, at a bipolar cell.


Ramón y Cajal's 1890 Illustration of the Olfactory Pathway—Anteroposterior cut of the olfactory bulb and nasal mucosa of the newborn rat. Golgi Method. A: Olfactory epithelium situated below the cribriform plate. a: Bipolar cell. b: Epithelial supporting cell. B: Dermis of the mucosa and bundle of olfactory fibers. C: Ethmoidal cartilage. D: Layer of olfactory fibers. E: Layer of olfactory glomeruli. F: Inferior molecular layer [external plexiform layer]. G: Zone of mitral cells. H: Superior molecular zone [internal plexiform layer]. I: Granular layer. c: Cartilage formation of the cribriform plate. e: Olfactory nerve. f: Arborization of the olfactory cells (glomeruli). g: Axis-cylinder of an inferior tufted cell. h: Mitral cell. i: Granule cells. j: Epithelial cells. n: Inferior granule. m: Terminal filament of epithelial cells. o: Large stellate cells.

There are occasions where silver chromate does not completely stain the bipolar cell, or only stains the cell body and ascending process, or the descending filament. In this last case, it would be said that fibers ended without obstruction within the thickness of the epithelium.

But it is essential to keep in mind that this occurs frequently in all of the preparations of the nervous system stained using Golgi's method. Whenever the preliminary induration is excessive, silver chromate deposits almost exclusively in the axis-cylinder itself, concentrating the staining at the start of the cells. On the other hand, in all of the cases of true endpoints of the nervous fibers, the branching axis-cylinder becomes enlarged and varicose, ending in a rounded enlargement. In the fibers that end freely without obstruction, this type of thickening is not seen. Similar circumstances will allow us to distinguish the extreme end of a cut or incompletely stained fiber from the true terminal end of the nervous arborizations.

Support cells also periodically stain, showing the same characteristics as in dissociation preparations. Above, a thick prism is formed that ends in the unrestricted surface plane; and underneath, they extend into a stem sculptured of winding sheets where the bipolar cells rest. The inferior extremity, which never unites with the olfactory fibers, comes to an end at the basal level by way of an enlargement, in accord with a bifurcation whose void houses a small basal cell.

Additionally, the inside of Bowman's glands are also stained, often giving the appearance of an upright sack with delicate lateral processes that penetrate between the glandular cells, others are shaped like a bifurcated, even trifurcated conduit, with the same collateral ramifications ((5)).

The data that we have just presented comprises a summary of what has been published in our memoirs over the previous year. After our work, a recollection surfaced of B. Grassi and Castronuovo ((6)), which was on the same topic. These histologists, who used the same method, arrived at results that concur with our own in very few points indeed.

According to what is inferred from their published drawings, they have seen bipolar cells along with their descending filaments, which are anastomosed and divided into branches (see Lamina XXII, Fig. I of Grassi's work). Additionally, between the intraepithelial fibers, the existence of certain thick and arborized filaments are noted, making it impossible to identify among them the fine individual olfactory fibers.

In light of these surprising results, we returned to reexamine our preparations and carry out new impregnations in the rabbit, cat, mouse, and so on. This new knowledge convinced us that Grassi and Castronuovo had fallen into a mistake that is actually difficult to avoid when dealing with incomplete preparations, and more so if observed at low magnifications. In our opinion, the aforementioned authors have regarded pairs of parallel confluent fibers, intimately joined by staining, as individual fibers, therefore considering the simple separation of these fibers as a case of branching. Occasionally, the Golgi reaction spreads from one fiber to another one when two fibers cross, resembling examples of branching or anastomosis. This error can be avoided by examining the preparation under a high objective, and not acknowledging other branches that merely look like those that appear at the same plane and level as the triangular enlargements which are always present in cases of legitimate bifurcation.

In the case of the robust, richly arborized nervous fibers, deemed as olfactory fibers and described by Grassi, these are, in my opinion, undoubtedly fibers of the sensory system. This conclusion is drawn because apart from having a stem that is of an origin much thicker than that of the olfactory fibers (in its entire length), they are only found in the respiratory regions of the nasal fossa. These fibers are absent in the olfactory or yellow portion, and for that reason we did not observe them last year. Grassi himself affirms that he found them in the bordering epithelium, that is, between the olfactory and the respiratory boundary, and not in the olfactory epithelium proper.

Our new work has allowed us to observe that the olfactory nerve fiber always conserves its original diameter and continuity throughout its trajectory (Fig. 1). This evidence of the complete trajectory of the olfactory fibers is not difficult if one operates on small newborn mammals. For this preparation, the olfactory bulb must harden for a period of time along with a segment of the lamina cribrosa and its mucosal covering. If the reaction is a success, as shown in Figure 1a, the pathway of the fibers can be followed from the glomeruli where they end, to their start at the epithelial cells of the mucosa. The cartilaginous state of the ethmoides easily permits thin and demonstrative cuts.

Figure 1a illustrates two thick groups of nerve fibers that cross the perforations of the lamina. Note that each descending nerve receives fibers from different glomeruli, fibers that after traveling the cartilaginous conduits together disperse in secondary bundles diverging in the region of the sub-mucosal tissue.

2. Central Endpoint of the Olfactory Fibers

The endpoint of the axis-cylinder cells that emanate from the bipolar corpuscles of the olfactory mucosa has its place in the thickness of the olfactory bulbs, at the level of the enlargements called olfactory glomeruli.

The olfactory bulb exhibits an interesting quality that has been subject of numerous investigations. The most important studies have been from Leydig ((7)), Clarke ((8)), Owsjannikow ((9)), Walter ((10)), Meinert ((11)), Henle ((12)), Golgi ((13)), Broca ((14)), Shwalbe ((15)), and Obersteiner. Incidentally, we have visited the subject in an investigation on the general connection of the nervous elements ((16)). My brother (Perdo Ramón) has recently studied the olfactory bulb of birds ((17)).

Conventional coloration methods supply incomplete results, as the contradictions of the works of the authors from the past two decades have demonstrated. Our preferred method is that of Golgi, to which this author owes notable discoveries in the structure of the olfactory bulb. We have also used Weigert-Pal method, which is very useful to correct and to complete the disposition of the medullary fibers.

The authors distinguish various alternating layers of gray and white matter in vertical and tranversal cuts of the olfactory bulb. Schwalbe counts six strata, which are—from the periphery to the center: (1) Olfactory layer, (2) Layer of the olfactory glomeruli, (3) Gelatinous layer, (4) Ganglion cell layer, (5) Granule and nervous plexus layer, (6) The bulb's ventral medullary substance (Broca's medullary layer, and [7] Henle's layer).

Golgi reduces them to three layers: (1) Superficial or fibrillary layer, (2) Medial layer, or that of the gray matter, and (3) Deep layer of white matter.

The number of layers accepted by Golgi seems to be excessively few to us; therefore, with only minimal variation, we will follow the stratification indicated by Schwalbe, which we found reliable to describe. From the outer to the inner portion the zones are:

1.a Layer of the Olfactory Nerve Bundles

The nervous bundles that cross the lamina cribrosa reach the border of the surface of the bulb, they disperse within the cortex of the bulb, interlacing with each other at this point (Fig. 1D). Golgi's method consistently stains these bundles, demonstrating that the strands that form these nerves are fine, varicose, unbranched and of equal thickness as the ones that are emitted from the epithelium. Between them there is a small amount of homogenous cement, and, separating the parcels, are stellate neuroglial cells which are extremely irregular and well described by Golgi.

2.a Layer of the Olfactory Glomeruli

Concentric to the preceding layer with an irregular border, exists a layer formed by one or two rows of granulose pear-shaped masses. The pedicle, which generally directs itself obliquely to the periphery, continues with one or more nerve bundles from the anterior zone; the wide part appears free and quite contoured, proceeding toward the center of the bulb. The thickness of the glomeruli, in carmine stained preparations, appears to consist of a fibrillary plexiform substance, in between which, numerous tiny corpuscles can be found.

The authors generally agree that the glomeruli represent the territory where the olfactory fibers come to an end, but they disagree notably when explaining the connections that are established in the region. For Walter and Owsjannikow, the olfactory fibers continue in the glomeruli with the peripheral expansions of the nerve cells of the bulb. Meinert and Krause reason that the glomerulus is constituted by a ball of olfactory fibers which are intercalated with fusiform cells within this complex meander. In the opinion of Broca, to which Schwalbe is inclined, the aforementioned organs consist simply of a large ball of small nerve cells.


Only Golgi, in favor of his valuable method, has been able to shine a light on this dark problem. This author has proven that the glomeruli are converging points of two orders of expansions: (1) On one hand, the olfactory fibers that repeatedly branch off in the aforementioned bodies; (2) On the other hand, a multitude of protoplasmic prolongations of nerve cells resting in proximal zones. These expansions are long, and once they arrive at the interior of the glomerulus, they divide into an elegant fringe or terminal tuft, perfectly free, but restrained within the limits of the glomerulus. The mesh of this plexus is full of certain small cells that Golgi considers to be of a neuroglial nature.

The existence of branching in the intraglomerular olfactory fibers, and the penetration of these and of the protoplasmic expansions of the immediate nerve cells (large pyramidal and small or fusiform) into the thickness of the glomeruli, constituting a dense plexus, are facts that are easily demonstrable (see Figs. 1f, 2, and 3) in mammals as well as in birds.


The glomerulus of a 1-month-old rat, depicting terminal olfactory fiber arborizations. a: Small varicose branches of the arborization. b: Afferent nerve stems.


Anteroposterior section of the olfactory bulb of the duck. A: Arborization of the olfactory fibers inside the glomerulus. B: Descending stems of the large tufted cells ending in a granulose fringe en the glomerulus. C: Large superior tufted cells. E: Inferior granule.

These facts shed light on a very interesting phenomenon: The transmission of the nervous action potential can be accomplished between arborizations of nerve fibers and arborizations of protoplasmic expansions, because the arborizations of the olfactory fibers never exit the glomeruli, nor does anything else enter these except protoplasmic expansions. Hundreds of consistent and demonstrative preparations have allowed us to establish this affirmation upon concrete tenets.

But how do the olfactory fibers come to an end inside the glomerulus? Golgi asserts, after repeated sectioning of this structure, that inside the glomerulus a tight continuous net is formed, not with protoplasmic expansions, but with nerve fibers originating from periglomerular cells and of the same tractus olfactorius. That is, afferent nerve fibers on the peripheral side correspond with fibers protruding from the central side, which establish a dynamic relationship supporting an intermediate network.

In our opinion, Golgi, influenced by the nerve network hypothesis, and because of the bias that the protoplasmic expansions do not represent organs of transmission, but purely nutritive consumers, an error was made here. The olfactory fibers do not constitute a net inside the glomerulus, but a perfectly free arborization like the ones of all nervous fibers (Fig. 2a). It is clear that to become convinced of such disposition, we should not resort to complete impregnations (dog, lamb, cat, rabbit, and so on.), but to the incomplete impregnations where one can often see the trajectory and arborization of a single fiber. It would even be easier still to observe the aforementioned nervous intraglomerular arborization if we select small animals for these preparations (mouse, white rat, and so on) that are either newborn or a few days old; in these animals branching is of a more simple nature.

Note that the branches of the olfactory fibers successively divide into secondary branches that are flexible, relatively robust, varicose, and ending in an olive-shaped or rounded knot. Similar to the motor arborizations, the secondary branches often proceed at a right angle, and among them there are such short branches, that they seem to be mere spines. These filaments have never come out of the glomerulus nor have they become continuous with a nerve fiber or protoplasmic expansion. Nor have we found nerve fibers from the tractus or from the immediate ganglion cells, penetrating into the glomeruli, for this reason I am inclined to presume that the nerve fibers cited by Golgi are a result of an error, or, in the case that they do exist, they are quite scarce. Neither our numerous impregnations with Golgi's three methods, rapid, slow, and semislow, nor the preparations done using the Weigert-Pal method have allowed us to see these fibers.

The final arborization of the olfactory nerve fibers is evident in the glomeruli of birds, as my brother has demonstrated and we have confirmed many times in the duck, hen, pigeon, and sparrow. The glomeruli are small, particularly in birds (its diameter is from 0.01 to 0.03 mm), and are available in 3, 4, or more irregular layers. The glomerular contour is at times completely inappreciable because of its inexact borders.

Figure 3 represents a section of the glomerular layer of the olfactory bulb of the duck. In some glomeruli, the protoplasmic arborization of the large pyramidal cells have been drawn as short, robust, and varicose, in the form of a tree penetrating into the superior layer. In others, only the branches of the olfactory fibers are represented which is simple and allows the path of the secondary branches to be followed until their endpoint at elongated protuberances. The assessment is made easier due to the small number of fibers that border a glomerulus, which penetrate and arborize two or three nerve fibers. In birds, where the glomeruli are even smaller, they often exhibit a single olfactory terminal, supplied with a short and varicose arborization like the one belonging to the cylinder-axis in a slide preparation of Rouget.

The same observation can be made in mammals with respect to the behavior of the protoplasmic expansions, which emanate from the nerve cells, and penetrate into the glomeruli: As the figure illustrates, these branches finish in tufts of varicose scattered filaments. In birds, it is not rare to see stems that split in its path supplying two or more glomeruli with endpoint tufts (Fig. 3B).

In view of the absence of an anastomosis between the protoplasmic and nervous arborizations, an observation already noted by Golgi, there is no other recourse than to admit to a transmission by multiple contacts between the olfactory fibers and the aforementioned protoplasmic branches. Similar communication, comparable to a phenomenon of electric induction, is facilitated by the interlacing and intimate mesh between both orders of fibers (this meshwork has been revealed through direct observation) and perhaps because of the presence of any type of intercalary conducting material.

Regarding the small cells resting in the thickness of the glomeruli, our studies are not so conclusive. Unfortunately, these corpuscles rarely stain, and when they do appear to stain, the intraglomerular nerve fibers stain as well, which makes an analysis notably difficult. Despite the findings of Golgi, we are not going to assume these to be neuroglial in nature. Our observations are as follows: The quantity of these cells is at times considerable in the large mammalian glomeruli; their location is at a point that could bring about the movement of olfactory excitations; they have a stellate shape with a relatively voluminous achromatic nucleus and robust and rough protoplasm stretched out into flexible expansions, without the characteristic reddish color of genuine neuroglial elements; they are present in certain expansions that appear to be of a nervous nature, and exit the glomerulus to advance toward the immediate structures, and finally, the fact that the Weigert-Pal method occasionally stains fine medullary fibers, originating in the thickness of the glomerulus, that are directed downward and to the sides (never upward), to reach the same level as nearby glomeruli. These are observations that incline us to think that here we are dealing with true nerve cells that are quite small, destined perhaps to make contacts with the nearby glomeruli. Regarding this matter, it is essential to undertake new investigations.

Fusiform nerve cells (Fig. 4b,c)

Some very small cells inhabit the glomerular layer and the spaces between the glomeruli, they are ovoid or fusiform and have been well described by Golgi. They all emit a thick protoplasmic expansion that after penetrating into the thickness of the neighboring glomeruli, end in a tuft of thick varicose fibers. To abbreviate, we will call all of the cells that collaborate in the construction of the intraglomerular plexus tufted or plumed cells. The ones that we have just cited are the most peripheral ones of this type, which is why we distinguish them as peripheral or external, reserving the term of medial and deep or profound for those cells that are in other layers.5


Anteroposterior section of the olfactory bulb of the rat. A: Glomerular layer. B: Inferior molecular layer. C: Mitral cell layer. D: Superior molecular layer. E: Granular and nerve fiber layer. F: Islet of granules. a: Small cells inside of the glomeruli. b: Inferior tufted cell. c: Inferior tufted cell atop a glomerulus. d: Medial tufted cells. e: Mitral cells. f: Granule with superior expansions. g: Inferior spiny arborization of the granule.


Two large stellate cells of the newborn dog. A: Inferior molecular layer. B: Mitral cell layer. a: Peripheral axis-cylinder.

The inferior tufted cells exhibit a fine axis-cylinder that emerges from the upper portion of the cell body, it draws around and above the glomeruli, and then crosses the successive layers to penetrate into the region of the granules. These axis-cylinders supply numerous collateral processes, but conserves its individuality until reaching the white matter. We will later expand on this issue.

3.a Inferior Molecular Layer (Schwalbe's gelatinous layer Henle's layer 6.a)

This layer is named for its granulose appearance, comparable to the molecular layer of the cerebellum or the retina. It is made up of protoplasmic arborizations of the mitral cells (pyramidal for the authors) and the inferior external of the descending expansions of the granules (Figs. 1F, 4B).

It also contains specialized nerve cells (medial tufted), that can be pyramidal or fusiform in shape. This last type is usually the most peripheral. An assessment, using a silver chromate preparation as with a carmine preparation, demonstrates the rarity of these elements, as well as its disorganized and disperse arrangement. They contain various lateral protoplasmic expansions, which take a more or less horizontal path and end according to the aforementioned manner in a nearby glomerulus (Fig. 4d).

The axis-cylinder of these cells departs from its profound extreme, crossing the inferior molecular layer and the mitral cell layer in almost a straight line once it reaches the superior molecular layer, it supplies 3, 4, or more fine collaterals at a right angle, that extend themselves to very large distances in the plane of the same layer, preserving a certain parallelism among themselves and with the free surface of the bulb. These collateral fibers appear cut crosswise in the bulb's cross sections and longitudinally in the anteroposterior sections (Fig. 4D). After this branching, that seems to end freely by means of fine swellings, the axis-cylinder continues its central path, arching to go to the back, and enter into one of the nerve fiber bunches that crosses the granular layer. Often, at the point of inflection, it supplies a collateral that heads in the opposite direction, that is, forward, by means of the same fascicle of medullated fibers (Figs 1, 4).

4.a Mitral Cell or Superior Tufted Layer (Schwalbe's Ganglion cell layer part of Golgi's Medial layer, and so on)

This layer consists of a defined stretch of cells, which appear somewhat tightly packed with two or three rows of corpuscles. These corpuscles are large cells comparable to the giant nerve cells of the brain, and corpuscles belonging to the category of the granules.

Mitral or pyramidal cells

In mammals, they are distributed in one row, which is more or less even. In birds, which has been demonstrated by P. Ramón, they align in two or more chains. The body of these corpuscles is triangular with almost straight or concave perimeters, and often the triangle is delimited by two superior convex borders and one inferior concave border. This particular configuration, similar to that of a bishop's cap, is exaggerated in preparations from newborn or young mammals (Fig. 1h)

The axis-cylinder emerges from the superior angle and the protoplasmic expansions from the lateral angles. These last expansions can be divided into lateral and descending.

The descending expansion is a robust stem, materializing from the cell body, already a protoplasmic lateral branch. This stem crosses the entire inferior molecular layer without branching and enters into the thickness of the glomeruli to construct a thick and intricate tuft (Figs. 1, 3).

In small mammals, and probably also in larger mammals, each cell possesses one stem with a tuft. In birds, my brother has confirmed that one cell supplies 18 or 20 tufted stems that end in various glomeruli, and it would not be rare to see one stem supplying two or three of these organs.

The protoplasmic lateral branches are robust, proceed toward the sides, and, after dichotomizing several times, they end freely inside the molecular layer at a considerably far distance from its origin. The assembly of this multitude of these parallel branches at the surface is what gives the molecular layer its plexiform appearance.

The axis-cylinder is thick and ascends across the superior molecular layer, stopping at different heights and proceeding anteroposteriorly at the granular layer. In the olfactory bulbs of newborn animals, this trajectory can be followed for a large distance from its origin (close to a millimeter), noting that once it changes direction (becoming anteroposterior) it supplies fine collaterals to the superior and inferior molecular layer through unbound arborizations. These collaterals were not shown by Golgi, possibly due to incomplete impregnations, and were first demonstrated by P. Ramón in the bulb of birds, where they appear in numerous amounts and stain easily.

5.a Superior Molecular Layer (Figs. 1H, 4D)

The mammalian olfactory bulb reliably demonstrates a layer with a molecular appearance, comparable to the layer of the same name in the retina. The superior molecular layer is situated above the mitral cells and below the granular layer.

Good impregnations reveal that the aforementioned layer consists essentially of fine nerve fibers, typically parallel, and aimed in the same direction as the bulbar surface. Almost all of these fibers are collateral branches, originating at a right angle from the axis-cylinder of the medial and inferior tufted corpuscles, not to mention from various terminal arborizations of nerve fibers from the brain. Additionally, we find the presence of another granule, that belonging to certain large stellate or fusiform cells that we will later discuss.

6.a Granular Layer

This is the thickest layer, spanning from the center to the epithelium of the ventricle of the bulb. It consists of nerve fibers that interlace and congregate at many points, leaving areoles or oblong, fusiform spaces, that are flattened from the outside in, where the granules and large stellate cells dwell. We have to consider three things in this layer: granules, large stellate cells, and nerve fibers.


This is the name given to diminutive corpuscles of spherical appearance in simple carmine sections. They are arranged in islets in the spaces left by bundles of nerve fibers

Golgi's method reliably and perfectly stains the granules, revealing that, despite the rough resemblance that they have with the granules of the cerebellum, they are not similar to these cerebellar granules in properties. As Golgi has demonstrated, they are frequently triangular and divergent, supplying central and peripheral expansions.

The peripheral expansion (Fig. 4g) is unique and extremely interesting for two reasons: one is that it ends in several divergent and flexible branches within the thickness of the inferior molecular layer, without leaving the margins of this region; another reason is that the aforementioned terminal branches are covered with spines, emerging at a right angle almost at a right angle and ending in small round enlargements. A similar spiny disposition also appears, as demonstrated by my brother, in the olfactory bulb of birds and reptiles.


The central expansions are short, varicose, thinner than the peripheral expansions and extremely flexible (Figs. 1, 4). They appear in quantities of two, three, and rarely four; they diverge from the cell body and slowly become thinner and thinner until almost invisible, fine, and granulose. In the peripheral expansion, there is a fixed point of termination, here the opposite occurs, here, where the point of termination varies based on the elevation where each granule finds itself. The deepest granules emit and crisscross its expansions close to the ventricular epithelium; the lowest expansions touch the filaments of the superior molecular layer (Fig. 4). The granules that are more inferior rest in line with the mitral cells and have short inferior expansions that divide to terminate along with the other expansions.

Golgi was not able to find an axis-cylinder in the granules, in light of which, he doubted the nature of these cells. This uncertainty becomes more rational knowing that which characterizes a nerve cell is the existence of a fine expansion of considerable length. And in the granules this did not exist, in effect, no cell has the characteristic of a nerve cell. Our efforts to find these expansions in newborn, young, and adult mammals, as well as in birds and reptiles have utterly failed. We are unable to blame incomplete impregnations, because the granule is the corpuscle that best and most reliably appears stained even in medial preparations.

I understand that, in this and other analogous cases, the determination of the nature of the cell cannot be determined by morphological criteria alone, but the criteria of the type of connection is much more conventional and promising. What do we see in all axis-cylinders, apart from their delicate nature and contours? It does have a considerable length, plus the fact that it arborizes constantly, over a long trajectory between certain other cellular elements.

Examining the expansions of the granule in light of this principle, one realizes that the prolongation functions, it is represented by the peripheral stem, which always ends between the protoplasmic branches of the mitral cells (that is, in the thickness of the inferior molecular layer), despite the orientation of the granule. And, we will add, that in certain preparations, it is not difficult to see that these protoplasmic branches lie in the spaces that are left by the spines of the aforementioned terminal arborization, establishing an anatomic connection by contact, like the one that exists at the superficial zone of the brain between the nerve fibers of the first layer, and the dentate peripheral arborization of the pyramidal cells.

An identical situation is noted in the granules of birds. Their corpuscles are shorter than those of mammals, and are supplied in its peripheral extremity with two or more spiny expansions that end in the inferior molecular layer (Fig. 3E).

Large stellate cells

These nervous corpuscles, discovered by Golgi, are scarce, and they rarely stain. We have seen them most often in the bulb of the newborn dog.

The carmine preparations confirm this peculiarity, showing these cells at a great distance from one another, in the most inferior part of the medial zone of the granular layer. Occasionally you see them in the thickness of the superior molecular layer.

These are easily distinguished from the granule by the magnitude of its body and the volume of the nucleus. The silver chromate preparations show that its shape is stellate or fusiform with ascending, descending, and lateral expansions. The nervous prolongations proceed in various directions from the body toward the periphery, and, when it reaches the inferior molecular layer, it gives off an extensive and intricate arborization that intertwines with the protoplasmic branches of the mitral cells. The ends of the arborization are strongly varicose terminating in granulose protuberances. Often, between the last branches there are some that regress to the mitral cell layer and terminate in ascending varicose fringes.

Golgi represents the axis-cylinder of these cells with a central pathway, allowing it to end in a net or intricate plexus situated on the granulose layer. For our part, we have seen such a disposition represented this way.

Nerve fibers

The fiber bundles that cross the granular layer, have, for the most part, an anteroposterior orientation, and are of two types: (1) Axis-cylinders of the mitral cells and of the medial or inferior tufted corpuscles. The first ones are thick and form arches with posterior concavity and gather various bundles of white matter. Axis-cylinders of the medial or inferior tufted cells are finer and supply delicate collaterals and are of an anteroposterior orientation running along the various bundles mentioned, possibly reaching the brain; they never seem to lose their individuality, contrary to the opinion of Golgi, who has them terminating in a plexus or fine net in the granular layer, and among recurrent collaterals, within the thickness of the glomeruli. (2) Nerve fibers from the brain. Thick and thin fibers of this type exist. The first are particularly abundant in the central zones, and proceed in a serpentine manner, in large undulations, backward, forward, and at different elevations of the bulb. Close to its anterior end, it concludes into an elegant and extensive arborization. The greater part of the secondary branches get lost between the nerve bundles and the plaid of granules, but some penetrate into the thickness of the superior and even inferior molecular layers, although they never reach the glomerular layer. The fine fibers branch equally between the granules and the molecular layers, but their delicate nature does not allow for a detailed study of its course and endpoint

7.a Epithelial Layer

Bordering the ventricular cavity of the bulb, there exists an epithelium of elongated cells, which, like the ependymal cells of the young medulla, branch out in various directions. These cells can be easily studied in the olfactory bulb of the newborn dog and mouse. They possess a round body, formed almost entirely of the nucleus, a central short expansion at the edge of the epithelial surface, and a large thin, dentate peripheral expansion which is spiny or varicose at selected intervals (Fig. 1j,m). This expansion, after crossing the first few strata of granules, ends freely among the granules at a sharp point. We have not managed to see them in newborn animals at the bulbar surface.

Myelin fibers

These easily stain using the Weigert-Pal method, and their trajectory and orientation are similar to the axis-cylinders stained using Golgi's method. Beginning at the surface and going toward the center (Fig. 6) we see that medullated fibers exist with relative abundance around and within the glomeruli. Periglomerular fibers, which are generally very fine, correspond to the axis-cylinders of the inferior tufted cells. It is not uncommon to be able to track one of these fibers along the inferior molecular layer to the granular layer where it takes an anteroposterior orientation. Intraglomerular medullated fibers are difficult to distinguish. These fibers frequently advance from the proximity of an intraglomerular nucleus, and after emerging from the glomerulus, they proceed transversally, to rest in the periphery, or even in the thickness of another glomerulus (Fig. 6c). In general, these fibers do not have the tendency of rising up with the others nor do they descend to the layer of the olfactory fibers. What we judged to be similar fibers turn out here to be fine axis-cylinders that rarely stain, coming from intraglomerular cells, and ending in the glomerular layer


Section of the olfactory bulb of the one month old rat. Weigert-Pal and aluminous carmine staining methods. A: Glomerular layer. B: Inferior molecular layer. C: Mitral cell layer. D: Superior molecular layer. E: Granular layer. a: Myelin fiber that corresponds to the axis-cylinder of the mitral cells. b: Axis-cylinder of the inferior tufted cells. c: Medullary fiber from the inside of the glomeruli. d: Fascicles of fibers from the granular layer. e: Fine horizontal filaments from the superior molecular layer. f: Mitral cells. g: Glomerulus (pictured adjacent to A).

Axis-cylinders of the medial tufted cells are supplied with myelin sheathing, as one can argue with a comparison of Figures 4 and 6, looking at the nerve fibers that cross the inferior molecular and mitral cell layer. Note that often in these fibers we find areas of constriction (Fig. 6b).

The nervous expansions of the mitral cells appear perfectly medullated, and its visual identification proves that it has an expansion that is more robust than all of the others that cross the same territory. Note that the myelin sheath does not begin adjacent to the cell, but slightly above the cell, in the thickness of the superior molecular layer. This path agrees with that of the axis-cylinders stained using the Golgi method in the aforementioned corpuscles (Fig. 6a). The bundles of white matter consist of an infinite number of medullated tubes whose origin is not easy to determine, although for the most part, they should be fibers of cerebral origin, and axis-cylinders of the tufted mitral corpuscles.

As is well known, the olfactory fibers and the expansions of the granules lack myelin. On the other hand, they present fine collaterals of the axis-cylinders, which cross the superior molecular layer (Fig. 6e). We have not been able to determine whether the descending axis-cylinders of the large stellate cells have myelin; we can only make known that its final arborizations at the inferior molecular layer do not have it.

Connection of the Bulbar Elements

We have shown in previous works that, in all likelihood, the nervous elements transmit their activity by contacts, between protoplasmic expansions, between protoplasmic prolongations, and nerve fiber arborizations or axis-cylinders. Regarding this concept, let us put forth an explanation of the propagation of olfactory impulses.

Excitation is conducted at the glomeruli, where numerous olfactory fibers end. Here, the motion is transmitted along several currents directed along the path of the tufted cells (mitral or superior, medial, and inferior), from the intraglomerular tufts, to the axis-cylinders and their cerebral endpoints in the olfactory centers.

From what is inferred, the transmission is not individual, that is, from the olfactory fiber to the tufted nerve cell, but from a group of olfactory fibers to a collection of nervous elements. Because, as we have previously discussed, the numerous cell tufts have their conclusion in each glomerulus.

This diffusion of transmission was previously mentioned by Golgi, but he stated that, in the glomerulus, nerve networks were communicating with fibers of the tractus on one side and with olfactory fibers on another side, excluding the protoplasmic expansions of the tufted cells of all participation.

In addition to this centripetal direct current, one can accept the existence of another centrifugal current, within the hypothesis that the nervous fluid changes from the cells to the arborizations. This current would get to the bulb by those fibers divided into branches from the point of the tractus, whose endpoint is in the region of the granular layer and in part of the molecular layers. Because these arborizations touch the fine central expansions of the granules at several points, the theory seems likely that these corpuscles carry the action [potential] to the tufted cells, whose lateral protoplasmic expansions maintain a peripheral dentate arborization of the granules and an intertwined network. With this path, the glomeruli would also be influenced. One should not say that the direction of the current is indeterminable, because the same issue is up for conjecture supposing a pathway from the tufted cells and granules to the white matter, or the other way around.

If, as we feel is most probable, all of the protoplasmic expansions have a role in propagation, they would be able to be useful for putting a group of analogous corpuscles. This could occur between the large mitral cells accompanying one another and between the stellate nerve cells of the granular layer.

We do not want to, nor should we go too far along this purely hypothetical road. It is useless for the anatomist to establish connections if the physiology cannot illustrate the nature and direction of the currents that cross them, or the varied roles that they play and the roles that other factors hold as they intervene in the act of transmission.


The preservation of the accuracy of the original text and the flavor of the unique and picturesque language used was of foremost importance when translating this text. Present day anatomical terms have been used when necessary, and original titles, subtitles, and opinions within the text have been preserved. Notes and opinions were also maintained from the original text and references. Illustrations by Cajal have been digitally enhanced for color and quality. It is additionally worth noting that, despite the many technological advances in microscopy, Cajal's illustrations are still considered a benchmark for anatomical studies. Future research includes a continuing analysis of developments regarding the structure and function of the olfactory system. Through accurate and detailed translations of original scientific research a historical appreciation of the most intricate details of the material can be attained and made available to a greater portion of the community in various disciplines of neurobiology including olfactory processing and regenerative studies. 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 S. Ramón y Cajal have been digitally enhanced for color and quality.