THE NEURON DOCTRINE (Translated and Annotated)

PART I, Fifty Years of an Open Question

1. Modern English Translation (of the 1891 German)

In 1841, J. Henle wrote in his General Anatomy: “What becomes of the processes of the ganglion globules [ganglion cells] is still not settled. It is unlikely that the ends one sees lie free within the substance of the central organs, since they are of such irregular shape and length that one must suspect they were torn during preparation. Just as little can they be regarded as mere commissures between the globules, as occur in the ganglia; this is contradicted by the direction of the processes in the cortex of the brain, which point precisely toward the side where no globules lie. For the same reason, the view that suggests itself at first glance, that they connect with, or pass over into, the primitive tubes [nerve fibres], is also refuted.”

Exactly fifty years have passed since the great master wrote this, and, speaking quite generally, we must still today make the same confession: “what becomes of the processes of the ganglion globules is still not settled”, even though we now know it for some cases. And we command entirely different resources than the microscopists of that time, who lacked staining technique, embedding procedures, and usable microtomes. Nevertheless, these means have helped us to a quite considerable expansion of our knowledge, and precisely in recent times very substantial progress has been achieved in answering the questions concerning the fate of the processes of the ganglion cells within the central organs, and concerning their connection with the anterior and posterior nerve roots that enter and exit. These advances shall be laid out here in overview.

(Footnote: After a lecture given to the Berlin Medical Society, June 1891.)

2. Historical Commentary

This opening page establishes the entire rhetorical and scientific architecture of Waldeyer’s essay. He begins not with his own claim, but with a quotation from Jacob Henle, one of the most authoritative German anatomists of the mid-nineteenth century. Henle’s 1841 statement captures the central unsolved problem of neuroanatomy at that time: nerve cells (“Ganglienkugeln,” literally “ganglion globules,” the contemporary term for what we now call neurons or nerve cell bodies) were known to possess processes (Fortsätze), but nobody could say with certainty where those processes went, whether they ended freely, whether they connected cell to cell, or whether they merged into nerve fibres.

Waldeyer’s rhetorical move, “exactly fifty years have passed… and we must still today make the same confession”, is deliberate. He uses Henle’s honest admission of ignorance as a baseline against which to measure the progress of the intervening half-century, while acknowledging that the fundamental question is still not fully resolved even in 1891. This is characteristic of Waldeyer’s method throughout the essay: he is not claiming to have solved the problem himself. He positions his own essay as a survey and evaluation of work done by others, a synthesizer’s opening, not a discoverer’s.

He explicitly credits the technical revolution that made new answers possible: improved Färbetechnik (staining technique), Einbettungsverfahren (embedding procedures), and brauchbare Mikrotome (usable microtomes), none of which existed in Henle’s day. This is a crucial historiographical point: the neuron question was not resolved by better theorizing alone, but by better tools, above all Golgi’s silver impregnation method (developed in the 1870s–80s) and Ramón y Cajal’s refinements of it, which allowed individual nerve cells to be visualized in something close to their entirety for the first time, rather than as an indistinguishable tangled mass.

The footnote confirms the essay’s origin: a lecture delivered to the Berlin Medical Society in June 1891, later published in the Deutsche Medicinische Wochenschrift. This is significant, the piece was conceived as a spoken review for working physicians and anatomists, not as a primary research report. This confirms from the outset that Waldeyer’s role in this history is that of rapporteur and synthesizer, not original discoverer.

3. Neuroscience Commentary: 1891 Compared With Today

Henle’s 1841 uncertainty reads, from a modern vantage point, as a remarkably clear-eyed statement of the actual limits of microscopy at the time. He correctly suspected that the visible endings of cell processes were probably artifacts of tissue damage during preparation (“zerrissen,” torn) rather than natural free endings, a caution that would later prove partly wrong, since better fixation eventually showed that some processes genuinely do end freely (axon terminals, dendritic tips), and partly right, since many apparent “endings” in poorly prepared nineteenth-century tissue were indeed torn.

What Waldeyer is setting up, though he does not yet name it on this page, is the debate that would culminate in the Neuron Doctrine: are nerve cells discrete, individual units connected only by contact, as Cajal, and, following him, Waldeyer, would argue? Or do their processes fuse into a continuous physical network, a reticulum, as Golgi’s reticular theory maintained? Henle’s passage shows that this question predates the specific Golgi–Cajal controversy by decades; it was inherent in microscopic neuroanatomy from its earliest days, simply because the tools to resolve it did not yet exist.

Today we know, via electron microscopy, that neurons are indeed discrete cells separated by synaptic clefts, vindicating the side of the debate that Waldeyer would ultimately champion. But we also know that some cells do communicate through direct cytoplasmic or electrical continuity in specific contexts (electrical synapses via gap junctions), and that glial networks show extensive gap-junction coupling that in some respects resembles the syncytial connectivity Golgi imagined for neurons, simply relocated to a different cell population. So the nineteenth-century binary, discrete cells versus fused network, turns out to be more nuanced than either camp fully anticipated.

4. Conceptual Reflection

This opening page is worth pausing on precisely because it is not yet about the neuron doctrine, but about the deeper epistemic condition that made the doctrine necessary: a fifty-year-old admission of not knowing where a cell’s processes go. It is a reminder that major conceptual reorganizations in biology, like the shift from “protoplasmic continuity” to “discrete cellular units” that the neuron doctrine represents, typically follow, rather than precede, improvements in observational technology. The idea of the neuron as a bounded, individual unit could not be seen until staining and microtome technique made individual cells visible against the tangle of surrounding tissue.

This has a broader resonance for how scientific paradigms shift generally: conceptual clarity is often blocked not by lack of intelligence or theory, but by lack of resolution, literal and figurative. It is a caution against assuming that the “correct” framework should have been obvious earlier; Henle’s uncertainty was not confusion, it was appropriate epistemic humility given the instruments available to him. Waldeyer’s essay, as the following pages will show, is precisely the record of what became visible once the instruments changed, and of how a change in visibility becomes, eventually, a change in concept: from tissue, to cell, to network, and, much later, beyond Waldeyer’s own horizon, to systems that extend past the nervous system altogether into hormonal, glial, and whole-body communication.

1. Modern English Translation (of the 1891 German)

To better clarify what has newly been established, and what has been examined more precisely, a picture of the state of our knowledge in these matters up to Golgi’s well-known publication in Pavia on the central organs, “Studii istologici sul midollo spinale” (Proceedings of the Third Italian Phreniatric Congress, held in Reggio Emilia in September 1880, Archivio italiano per le malattie nervose, etc., 18th year, fasc. I, 1881), together with a brief historical overview, may be useful here.

The ganglion cells in the spinal ganglia were first seen by Ehrenberg. They were first correctly described by Valentin. The ganglion cells in the brain were discovered by Purkyně. That the processes of ganglion cells in part pass over into nerve fibres was first stated by Remak, for the sympathetic nerve fibres of vertebrates; soon afterward by Helmholtz, in his doctoral dissertation, and by Hannover, for invertebrates. For the spinal ganglia, Kölliker supplied the definite proof of an origin of dark-bordered, medullated nerve fibres from unipolar ganglion cells, as the spinal ganglion cells were long regarded to be. Later (1847), Bidder, Reichert, Robin, and Rud. Wagner simultaneously made the important discovery that the nerve cells in the spinal ganglia of fishes are bipolar and continue on both sides into a medullated nerve fiber, thus being, as it were, inserted into its course. A long time passed before one was able, after considerable improvement of investigative methods, to state anything definite about the connection of the ganglion cells of the central nervous system, in man and the higher vertebrates, with the conducting pathways. The most pioneering step in this respect was taken by Deiters in Bonn, who generalized a finding made by R. Wagner on the electric lobe of the brain of Torpedo, which Remak (Deutsche Klinik No. 27, 7 July 1855) later put forward as valid for all motor nerve cells. According to Deiters, all central ganglion cells have two kinds of processes: 1) axis-cylinder processes [Axencylinderfortsätze], which stand in connection with dark-bordered nerve fibres by passing over into their axis cylinder, and 2) protoplasmic processes [Protoplasmafortsätze], which branch abundantly and finely, but for which a connection with nerve fibres cannot be reliably demonstrated, and whose ultimate fate Deiters left entirely undetermined.

2. Historical Commentary

Here Waldeyer explicitly declares his method: he will use Golgi’s 1880/81 publication, Studii istologici sul midollo spinale (“Histological Studies on the Spinal Cord”), as the pivot point of his historical review, everything before it belongs to the “old” era of cell staining without process-tracing; everything after it belongs to the era that his essay is really about. This is telling: Waldeyer treats Golgi’s introduction of silver impregnation, not any single anatomical discovery, as the true turning point of the field, because it was the technique, not a particular finding, that reopened the fifty-year-old Henle question in a tractable form.

The genealogy he then supplies is a compressed history of nineteenth-century neurohistology: Ehrenberg (first sighting of ganglion cells in spinal ganglia), Valentin (first accurate description), Purkyně (discovery of brain ganglion cells, what we now call Purkinje cells of the cerebellum, among others). He then traces the gradually strengthening evidence that cell processes become nerve fibres: Remak for sympathetic fibres, Helmholtz (yes, the same Hermann von Helmholtz later famous for optics and acoustics, this was his doctoral dissertation) and Hannover for invertebrates, Kölliker for unipolar spinal ganglion cells, and the joint 1847 discovery by Bidder, Reichert, Robin, and Rudolph Wagner that fish spinal ganglion cells are bipolar, sitting inline within a fibre’s course rather than at its end.

The centerpiece of this page, however, is Otto Deiters of Bonn, whom Waldeyer credits with the “most pioneering step”: the generalization, from a specific finding by Rudolph Wagner on the electric lobe of the torpedo ray’s brain (later extended by Remak to all motor neurons), that every central ganglion cell possesses exactly two kinds of processes, the axis-cylinder process (Axencylinderfortsatz, i.e., what we call the axon) which continues into a nerve fibre, and the protoplasmic processes (Protoplasmafortsätze, i.e., what we call dendrites) which branch finely but whose connectivity and function Deiters could not determine. This Deiters dichotomy, one process type becoming the conducting fibre, the other remaining of uncertain, perhaps nutritive, function, is the direct conceptual ancestor of the modern axon/dendrite distinction, though the modern functional understanding (dendrites receive, axons transmit) was not yet available to Deiters or to Waldeyer in 1891.

3. Neuroscience Commentary: 1891 Compared With Today

Modern neuroscience confirms the core of the Deiters dichotomy with striking fidelity: neurons genuinely do have one axon (in Deiters’s terms, the axis-cylinder process) and typically multiple dendrites (protoplasmic processes), and this remains one of the most basic morphological facts taught in every introductory course. What nineteenth-century anatomists could not yet know is the functional logic behind the distinction, that dendrites integrate incoming synaptic signals while the axon actively propagates an all-or-none action potential outward to the axon terminal. That functional understanding required twentieth-century electrophysiology (Hodgkin and Huxley’s work on the squid giant axon, for instance), which lay entirely beyond the reach of fixed, stained, dead tissue under a nineteenth-century microscope.

It’s also worth noting how much of this “genealogy of discovery” reflects the fragmented, cross-national character of early neuroscience, German, Italian, Scandinavian, and later Spanish researchers, often unaware of or only partially citing one another’s work, converging piecemeal on pieces of what would become a single picture. Waldeyer’s essay is itself an act of integration across these national research traditions, which is part of why a synthesizing review like his was so valuable to the field in 1891, independent of any new observation it contained.

4. Conceptual Reflection

This page shows the neuron doctrine being built, brick by brick, well before anyone had the concept of “the neuron” to organize the bricks around. Ehrenberg saw a cell; Valentin described it properly; Purkyně found more of them; Remak, Helmholtz, and Kölliker connected some of them to fibres; Deiters proposed a general structural rule. Each contributor solved a local problem without necessarily grasping the larger architecture that would later organize all of their work into a single doctrine. This is a common pattern in the history of science: a unifying concept frequently arrives only after a critical mass of partial, locally-motivated observations has accumulated, and the person who supplies the unifying concept (in this case, Waldeyer) often contributes no new data at all, only the recognition that the data, taken together, describe one coherent kind of object. The “neuron” as an idea was, in this sense, assembled from spare parts that had been lying around German, Italian, and Scandinavian laboratories for fifty years.

1. Modern English Translation (of the 1891 German)

Deiters further divides the axis-cylinder processes into two kinds: the coarse principal axis-cylinder processes, and the finer ones. Of the former, each cell possesses only one; of the latter, it may possess several; these arise from the protoplasmic processes and pass over into the axis cylinders of the fine, central, medullated nerve fibres. The axis-cylinder processes are supposed always to be unbranched, and thereby to differ essentially from the protoplasmic processes. Deiters assumed, as I would particularly like to emphasize, such axis-cylinder processes both for the motor and for the sensory nerve root fibres of the spinal cord and brain. One must nevertheless admit that, in the work published after his death in 1863 by Max Schultze, Investigations on the Brain and Spinal Cord of Man and Mammals, Braunschweig 1865, he had not produced a fully satisfactory proof of the connection between a ganglion cell of the central organ and a nerve fibre. The uncertainty that still remained here is also evident from the words of two of our best masters in the investigation of the central nervous system, J. v. Gerlach and A. v. Kölliker, of whom the latter, in his world-renowned work Handbook of Human Histology, fifth edition, 1867, expresses himself with great caution.

[…] As late as 1871, in the then-published Handbook of Histology, edited by Stricker, v. Gerlach expressed himself as follows, p. 681: “As regards the first question (concerning the behaviour of the nerve process), it was decidedly a fortunate stroke on Deiters’s part to bring the nerve process of the cells into relation with the origin of the spinal nerves. Even though the true relations are not as simple as Deiters imagined them, he had the fibres of the anterior roots arise from the cells of the anterior horns, and those of the posterior roots from the cells of the posterior horns, still, one can say with approximate certainty that Deiters hit upon the truth as regards the origin of the anterior roots.”

2. Historical Commentary

Waldeyer continues his historiographical build-up by refining the Deiters picture: the axis-cylinder process (our axon) is subdivided into a single coarse “principal” branch and several finer branches, all unbranched in Deiters’s scheme, an early, partially incorrect version of what we now know as axon collaterals (which Golgi and later Ramón y Cajal would prove do branch extensively). Crucially, Waldeyer stresses that Deiters applied his single-axon rule to both motor and sensory nerve cells alike, an early gesture toward a unified structural law for all nerve cells, which is exactly the kind of generalization the neuron doctrine would later formalize.

But Waldeyer is careful, in the true spirit of a critical historian, to note the limits of Deiters’s evidence: Deiters died in 1863, and his manuscript, edited posthumously by Max Schultze, never actually delivered an airtight proof that a central ganglion cell process truly continues into a nerve fibre. This honesty is reinforced by quoting two towering contemporary authorities, Joseph von Gerlach and Albert von Kölliker, both of whom, writing years later (1867 and 1871), still hedged their claims. Gerlach’s 1871 quotation is particularly important: he credits Deiters with a “fortunate stroke” of insight regarding anterior (motor) roots specifically, while conceding that Deiters’s picture of posterior (sensory) roots was probably wrong. This distinction between confidently-understood motor pathways and poorly-understood sensory pathways recurs throughout the essay and will resurface dramatically when Golgi and Ramón y Cajal’s silver-stain findings are discussed.

3. Neuroscience Commentary: 1891 Compared With Today

Modern neuroanatomy validates part of Deiters’s scheme and corrects another part in a very specific way. The “one axon per neuron” rule (Deiters’s Hauptaxencylinderfortsatz) remains essentially correct and is still taught as a defining structural feature of the neuron. However, the claim that axis-cylinder processes are “always unbranched” is now known to be wrong: axons commonly give off collateral branches, a fact that Golgi’s staining method would soon reveal in Waldeyer’s own subsequent pages, and which Ramón y Cajal would document extensively. The asymmetry Gerlach notes, motor (anterior root) connectivity being better understood than sensory (posterior root) connectivity, reflects a real biological and technical asymmetry that persisted for decades: motor neurons have large, easily stained cell bodies with a clear single exit point (the ventral root), while sensory pathways involve more complex, harder-to-trace circuitry through dorsal root ganglia and multiple synaptic relays.

4. Conceptual Reflection

This page is a valuable corrective to any simplified narrative in which nineteenth-century anatomists “already knew” what the neuron doctrine would state. What we see instead is scientists of the highest caliber, Gerlach, Kölliker, refusing to commit beyond what their evidence actually supported, even while crediting a colleague’s insight as directionally correct. This is good scientific practice: distinguishing a fruitful hypothesis (Deiters’s structural rule) from a proven fact (the actual continuity of cell process into fibre). The neuron doctrine, when it later crystallizes in this same essay, would not have carried authority if it had simply repeated Deiters’s hunch, it needed exactly the kind of rigorous, technically-enabled proof that the next several pages describe.

1. Modern English Translation (of the 1891 German)

Gerlach’s schema establishes a definite distinction in the mode of origin of the anterior and posterior nerve roots. The anterior motor root fibres are supposed, in their axis cylinders, to be simple continuations of the principal axis-cylinder processes of the large multipolar ganglion cells of the anterior horns, whereas the posterior roots arise from a network into which the finely branching protoplasmic processes of the anterior and posterior nerve cells of the spinal cord enter. According to Gerlach, by no means all cells of the spinal cord have nerve processes; they are lacking, for example, in those of the so-called Clarke’s columns; two “morphologically distinct kinds of nerve cells” must therefore be assumed, according to him, of which one kind stands in direct connection with the anterior root fibres and with the nerve network of the grey matter, while the other kind stands in connection only with the latter.

In its simplicity and clarity, Gerlach’s schema had a compelling effect, and it has retained widespread following right up to the most recent times; we shall see that it has been confirmed, to a certain degree, by the more recent research.

Figure 1, reproduced from the original 1891 offprint (public domain).

2. Historical Commentary

Gerlach’s schema, as Waldeyer presents it, was the dominant working model of spinal cord wiring in the decades before Golgi’s silver method. It is important because it already contains the seed of a network theory (Nervennetz), the idea that sensory (posterior root) input enters a diffuse, fused mesh of finely branching processes rather than connecting to individual, traceable cells. This is the direct ancestor of what would become Golgi’s reticular theory, the chief rival to the neuron doctrine that Waldeyer will ultimately endorse. Gerlach’s scheme also proposes a division of spinal cord cells into two anatomically distinct types (with and without nerve processes), an early, cruder version of the functional cell-typing that Ramón y Cajal would later carry out with vastly more precision using Golgi’s stain.

Waldeyer’s closing remark, that the schema “has been confirmed, to a certain degree, by the more recent research”, is a small but important signal of his overall argumentative strategy. He is not going to present the newer, Golgi/Cajal-based findings as a total repudiation of the older German tradition (Gerlach, Deiters, Kölliker); rather, he frames the newer work as a refinement and partial vindication of earlier insight, combined with correction of its errors (particularly the reticular/network assumption for sensory input, which Cajal would argue against).

3. Neuroscience Commentary: 1891 Compared With Today

Gerlach’s distinction between motor roots (traceable, cell-to-fibre) and sensory roots (network-like) reflects a genuine biological difference that nineteenth-century anatomists were correctly detecting, even if they explained it incorrectly. Modern neuroscience confirms that motor neurons in the ventral horn do send a single, identifiable axon out through the ventral root to innervate muscle, essentially validating Gerlach’s motor picture. But the “network” appearance of sensory input into the dorsal horn, which Gerlach attributed to fused protoplasmic processes, is now understood very differently: it reflects the genuinely dense, multi-synaptic connectivity of interneuron circuits in the dorsal horn, where many discrete neurons synapse onto one another in complex, overlapping patterns, dense, but not fused. The visual impression of a “network” was real; the anatomical interpretation (physical continuity rather than a dense field of discrete contacts) was the part that later research would correct.

4. Conceptual Reflection

Gerlach’s schema illustrates a recurring feature of scientific transition periods: an intermediate model that gets part of the picture right (the categorical distinction between motor and sensory circuitry) while getting the underlying mechanism wrong (fusion versus contact). Such intermediate models are not failures to be dismissed; they are necessary scaffolding. They correctly identify that there is a difference to be explained, even before anyone has the tools to explain why. The neuron doctrine’s eventual triumph would not erase Gerlach’s basic motor/sensory distinction, it would re-explain it in terms of discrete cells and their specific patterns of contact, rather than differential participation in a continuous mesh.

1. Modern English Translation (of the 1891 German)

What was known in the year in question [1880] concerning the finer anatomy of the medulla oblongata yields, as regards the relation of ganglion cells to nerve fibres, and in particular to the root fibres of the exiting spinal-type cranial nerves, nothing different from what has just been set out for the spinal cord. Just as little definite clarification concerning these matters do we obtain from the finer anatomy of the small brain [cerebellum]. The shape and branching of the great Purkyně nerve bodies had indeed long been known, but the most varied conceptions circulated regarding their relation to nerve fibres and to the other cells of the cerebellar cortex. The central process of the Purkyně cells was held to be an undivided axis-cylinder process; but how uncertain the opinion still was regarding its transition into the axis cylinder of a medullated nerve fibre is evident, for example, from the wording in the aforementioned Quain’s handbook.

2. Historical Commentary

Having established the state of knowledge for the spinal cord, Waldeyer now systematically extends the same “as of 1880” survey to the medulla oblongata (brainstem) and the cerebellum, before moving on (in the pages that follow) to the cerebral cortex. This is a deliberate structural choice: Waldeyer is building region-by-region toward the most complex and clinically important structure, the cerebrum, and wants the reader to appreciate that uncertainty about cell-to-fibre connectivity was not confined to the spinal cord but pervaded the entire central nervous system as of Golgi’s starting point.

The Purkyně cells (Purkinje cells) of the cerebellum, first identified by Jan Evangelista Purkyně, credited earlier in the essay, are singled out because their large size and dramatic, richly branching dendritic tree made them the most visually striking neurons known to nineteenth-century microscopists, and therefore a natural test case for theories about process continuity. Waldeyer notes that even for these famous, easily observed cells, there was no secure proof that their central process (axon) actually became a myelinated nerve fibre, he cites Quain’s Anatomy, the standard English-language reference handbook of the era, as evidence of how equivocal the standard textbook treatment still was.

3. Neuroscience Commentary: 1891 Compared With Today

The choice of Purkinje cells as a touchstone is apt and durable, they remain one of the most-studied neuron types in modern neuroscience, prized for their large size, singular and unmistakable morphology, and central role in cerebellar motor coordination and learning. Modern cellular neuroscience fully confirms that the Purkinje cell axon (the “central process” Waldeyer mentions) does indeed become a myelinated fibre projecting to the deep cerebellar nuclei, exactly the connection nineteenth-century anatomists suspected but could not prove. What nineteenth-century science could not yet appreciate is the extraordinary computational significance of the Purkinje cell’s dendritic tree, now understood as receiving well over 100,000 synaptic inputs from parallel fibres, making it one of the most heavily-innervated single neurons anywhere in the nervous system, a fact utterly invisible to 1880s silver-free staining methods.

4. Conceptual Reflection

This brief but pointed page underscores how uneven scientific knowledge was across brain regions in the pre-Golgi era, not because some regions were intrinsically harder to study, but because visibility, technique, and researcher attention were unevenly distributed. It is a useful reminder that “the state of knowledge in 1880” was not a single uniform level of ignorance or insight, but a patchwork, with some structures (like the readily visible Purkinje cell body) generating confident description of gross form long before anyone could resolve the finer question of connectivity, form and function, in other words, advanced on different timelines.

1. Modern English Translation (of the 1891 German)

For the cerebral cortex, the treatises addressing the topic under discussion here read similarly to those for the cerebellar cortex. More precisely known, at the time taken here as our starting point (1880), were only the large so-called pyramidal cells; for the remaining very numerous cells of the cerebral cortex, occurring in various forms, a connection with other elements was merely surmised, there existed no definite statement whatsoever. On the pyramidal cells, Koschewnikoff claims to have demonstrated the axis-cylinder process (Arch. f. mikrosk. Anat., vol. V, p. 374). He found, in agreement with Meynert’s conjecture, that it was the middle basal process, not the apical process that Rud. Arndt had shortly before held to be the axis-cylinder process. But it must be considered that Koschewnikoff, despite much searching, found only a single cell in which he saw the process in question pass into a medullary sheath, and further, that the illustration of this transition point leaves room for all manner of doubt.

2. Historical Commentary

Waldeyer now reaches the cerebral cortex, the structure of greatest interest to physicians and, ultimately, to psychology and psychiatry, since it was increasingly understood (following Theodor Meynert’s influential localization work) as the seat of higher mental functions. The pyramidal cells were, by 1880, the best-characterized cortical neuron type by virtue of their large size and distinctive triangular shape, paralleling the privileged status of Purkinje cells in the cerebellum.

The passage on Koschewnikoff is a small but instructive case study in the evidentiary standards of the period. Koschewnikoff claimed to have traced the pyramidal cell’s axis-cylinder process to its origin, settling a dispute with Rudolph Arndt over whether the apical dendrite (the tall process reaching toward the cortical surface) or the basal process was the true axon, and sided with Meynert’s prediction that it was a basal (not apical) process. This is, in modern terms, actually correct: the pyramidal cell axon does arise from the base of the soma, not from the apical dendrite. Yet Waldeyer, in the spirit of a critical reviewer, immediately qualifies the finding: Koschewnikoff, despite extensive searching, found only a single cell in which he could see the process transition into a myelin sheath, and even that one illustration “leaves room for all manner of doubt.” This is Waldeyer modeling careful scientific skepticism, a single, ambiguous observation, even a directionally correct one, does not constitute proof.

3. Neuroscience Commentary: 1891 Compared With Today

Modern neuroanatomy fully confirms Meynert’s prediction and Koschewnikoff’s tentative finding: the axon of a cortical pyramidal neuron does originate near the base of the cell body (from the axon hillock, typically near the base opposite the apical dendrite), while the tall apical dendrite reaching toward the pial surface is a dendritic, receptive structure, not the conducting axon. What nineteenth-century science lacked was any way to move from a single lucky, ambiguous silver-stained or otherwise-prepared specimen to a general, statistically confident rule, precisely the gap that Golgi’s more reliable and reproducible staining method, discussed on the following pages, would close.

4. Conceptual Reflection

The Koschewnikoff episode is a small parable about the difference between an isolated observation and a scientific finding. A single well-stained cell, however striking, cannot by itself overturn fifty years of uncertainty; it can only nominate a hypothesis for further testing. What changes the status of a claim from “suggestive anecdote” to “established fact” is not the boldness of the interpretation but the reproducibility of the observation across many specimens, precisely the capability that the Golgi method, and above all Ramón y Cajal’s disciplined and extensive application of it, would soon supply. Waldeyer’s careful, page-by-page accounting of exactly how thin the pre-1880 evidence really was serves a rhetorical purpose: it makes the subsequent flood of Golgi-era findings look all the more decisive by contrast.

1. Modern English Translation (of the 1891 German)

Improvements in investigative method have always had an outstanding share in achieving progress in the understanding of natural-scientific matters. For this reason, those technical resources that have most advanced our uncovering of the structure of the central nervous system may briefly be mentioned here. The first notable advance came with the founding of microscopic staining by J. v. Gerlach (carmine staining) in 1858. There followed the introduction of osmic acid by Max Schultze, the mordant-staining procedure by C. Weigert, then the investigative methods of Golgi and S. Ramón y Cajal, as well as in-vivo staining by means of methylene blue by P. Ehrlich, a procedure we may well describe as one of the most significant advances in our investigative methods.

Nor should the improved hardening and embedding procedures be forgotten, I mean the introduction of chromic acid by A. Hannover, embedding in paraffin (Klebs, His, see Klebs in the Archiv für mikroskopische Anatomie, vol. V), and in celloidin (Merkel and Schiefferdecker), the mounting of stained preparations, after dehydration and clearing in volatile oils, in balsam, by Lockhart Clarke and Reissner., The moistening of the cutting knife with alcohol, which to my knowledge Benedikt Stilling first practiced and recommended, and the improvement of the microtome, may likewise claim their share in the effective advancement of our knowledge of the central nervous system.

2. Historical Commentary

This page is a compact history of nineteenth-century histological technique, and it functions as the essay’s technological hinge, Waldeyer explicitly frames scientific progress here as method-driven rather than idea-driven. He credits Joseph von Gerlach (the same Gerlach whose spinal cord schema was discussed earlier) with founding microscopic staining via carmine dye in 1858; Max Schultze with introducing osmium tetroxide fixation; Carl Weigert with a mordant-staining technique important for myelin visualization; and, crucially, Camillo Golgi and Santiago Ramón y Cajal with the silver-based methods that make the rest of this essay possible. He singles out Paul Ehrlich’s methylene blue vital staining (staining living, unfixed tissue) as perhaps the single most significant methodological leap, since it allowed structures to be visualized without the distorting effects of chemical fixation.

Equally important, though less glamorous, are the embedding and sectioning improvements he lists: paraffin embedding (associated with Edwin Klebs and Wilhelm His, who will reappear later as a major embryological authority), celloidin embedding, and mechanical improvements to microtomes and blade handling. These unglamorous procedural refinements were what made it possible to cut consistent, thin, undistorted sections of nervous tissue at all, a precondition for any staining method to be useful.

3. Neuroscience Commentary: 1891 Compared With Today

This page reads almost as a manifesto for tool-driven discovery, a principle that remains centrally true in neuroscience today. Just as Golgi’s silver stain and Ehrlich’s methylene blue opened the nineteenth century’s view of the neuron, modern neuroscience’s major advances continue to track technological innovation: green fluorescent protein and optogenetics for visualizing and controlling specific neurons, patch-clamp electrophysiology for measuring individual synaptic currents, and large-scale connectomic electron microscopy for mapping every synapse in a piece of tissue (the approach that finally, unambiguously settled the discrete-versus-network question in the twentieth century). Waldeyer’s insistence that “improvements in investigative method have always had an outstanding share” in scientific progress is as valid a description of contemporary neuroscience as it was of 1891 histology.

4. Conceptual Reflection

This page is a useful antidote to any narrative that treats scientific revolutions as purely conceptual events, moments when a genius simply “figures it out.” Waldeyer’s own account insists otherwise: the neuron doctrine became thinkable only because dye chemists, embedding specialists, and instrument-makers solved a series of unglamorous practical problems first. The lesson generalizes well beyond 1891: many of the deepest shifts in biological understanding, cellular theory itself, the germ theory of disease, the structure of DNA, were unlocked by a new way of seeing, not merely a new way of thinking. Concepts and instruments co-evolve; neither leads the other indefinitely.

1. Modern English Translation (of the 1891 German)

The works which, essentially since 1882, have advanced our knowledge in the field under discussion are those of Camillo Golgi in Pavia, Santiago Ramón y Cajal in Barcelona, Wilhelm His in Leipzig, Albert von Kölliker in Würzburg, and Gustaf Retzius in Stockholm.

Golgi arrives, by a method resting essentially on impregnation, at first he used a sublimate impregnation, then the formation of precipitates of chromate of silver, and more recently he hardens [tissue] in a mixture of three parts of a 2% potassium bichromate solution with one part of a 1% osmic acid solution, then transfers the pieces into a ½–1% solution of silver nitrate for two days, then into approximately 40–50% alcohol; the tissue is embedded in celloidin and cut immediately upon reaching cuttable consistency; the sections are mounted, without cover-glass, in Canada balsam, at the following results:

Above all, two kinds of nerve cells must be distinguished, which Golgi designates by the names: cells of Type I and Type II. Those of Type I show an axis-cylinder process, a nerve process, as I shall henceforth call it, following the more recent authors, and protoplasmic processes; the latter branch abundantly, at once, in the manner long known and described in particular by Deiters and J. v. Gerlach; the nerve process, which all authors had hitherto held to be unbranched, also gives off side branches, though in small number, without thereby losing its individuality, and ultimately becomes the axis cylinder of a medullated nerve fibre.

The side branches of the nerve process now, according to Golgi, pass over into a network of nervous fibres, which is essentially formed by the nerve processes of the cells of Type II. The cells of Type II have protoplasmic processes and nerve processes; the latter, however, branch abundantly without leaving the grey matter of the central organs, and form, together with the branches of the nerve processes of all cells of the same type, and the few branches of the nerve processes of the Type I cells, the aforementioned network of nervous threads, which is continuous throughout the entire central nervous system.

2. Historical Commentary

This is one of the most important pages in the entire essay. Waldeyer now names the five researchers whose work, essentially since 1882, constitutes the real substance of his review: Camillo Golgi, Santiago Ramón y Cajal, Wilhelm His, Albert von Kölliker, and Gustaf Retzius. This list is worth pausing on, because it makes explicit what has been implicit throughout: everything of scientific substance in this essay belongs to these five men (plus, later, others such as Fritjof Nansen and August Forel). Waldeyer’s own contribution, still to come, is to weigh their claims against one another and draw a unified conclusion, not to add a sixth original finding of his own.

Golgi’s staining recipe is given in full technical detail, Waldeyer clearly regards the precise chemistry (potassium bichromate, osmic acid, silver nitrate, celloidin embedding) as important enough to reproduce exactly, underscoring how central the reagent chemistry itself was to the entire scientific breakthrough. This is the famous “black reaction” (reazione nera), still known today by that name, which stains only a small, apparently random subset of neurons in a tissue slice completely and darkly, against a pale, transparent background, a stroke of chemical fortune that made single-cell visualization possible for the first time.

The heart of the page is Golgi’s distinction between Type I cells (with a long, cable-like nerve process, what we would call a projection neuron with a long axon, eventually contributing to a myelinated tract fibre) and Type II cells (with a short, richly branching nerve process that never leaves the grey matter, what we would call a local interneuron). This typology, still broadly recognizable today, is a genuine and lasting Golgi contribution. But note the crucial theoretical claim embedded in the description: Golgi holds that the branches of these nerve processes, from both cell types, merge into a single, anatomically continuous network (Nervennetz) that runs throughout the entire central nervous system. This is the explicit statement of Golgi’s reticular theory, the direct rival to what would become the neuron doctrine.

3. Neuroscience Commentary: 1891 Compared With Today

Golgi’s Type I/Type II distinction survives today in modified form as the distinction between projection neurons (long axons connecting distant regions) and local interneurons (short axons confined to a local circuit), a fundamental and still-used organizing principle of neural circuitry. The chemistry of the Golgi stain itself remains in active laboratory use more than 140 years later, prized precisely for the same property Waldeyer describes: it stains only a sparse, seemingly random subset of neurons in their near-entirety, allowing individual cell morphology to be studied without the visual clutter of every neighboring cell also being stained.

Where Golgi was wrong, and where the drama of this entire essay ultimately turns, is in the claim of anatomical continuity: that the branching processes of different neurons physically fuse into one continuous nervous network. Electron microscopy in the twentieth century definitively established that neurons remain discrete cells at every point of apparent contact, separated by a narrow synaptic cleft rather than fused cytoplasm. Golgi’s own staining method, ironically, was too coarse at the level of individual contact points to distinguish “very close apposition” from “true fusion”, the resolving power needed to settle that question, electron microscopy, would not exist for another sixty years. It is a striking historical irony that Golgi, who provided the essential tool for the neuron doctrine’s eventual triumph, remained a lifelong opponent of the doctrine itself, and shared the 1906 Nobel Prize with Ramón y Cajal despite their diametrically opposed interpretations of what their shared staining method showed.

4. Conceptual Reflection

This page crystallizes a profound and recurring situation in the history of science: the same observational tool, applied by different researchers with different theoretical commitments, can be read to support opposite conclusions. Golgi looked at silver-stained tissue and saw continuity; Ramón y Cajal, using the same basic method with modifications, would look at similar tissue and see discreteness. The instrument alone did not settle the question, interpretation, technique refinement, and theoretical framework all mattered enormously. This is a caution against any simple story in which “better microscopy” mechanically produces “better theory.” Golgi had superb microscopy and, on this specific point, arrived at what most now regard as the wrong conclusion. Progress required not just seeing more, but seeing differently, which is precisely what the next several pages of Waldeyer’s essay, describing Ramón y Cajal’s departures from Golgi, will show.

1. Modern English Translation (of the 1891 German)

Since the cells of Type I lie within the territory of the anterior root fibres and, in part, can be shown with certainty to send their nerve process down into the fibres of an anterior (motor) root, Golgi designates them as motor; the cells of Type II, by contrast, as sensory. From the nervous network of the Type II cells (into which fibrils of the nerve processes of the Type I cells also, however, pass) the sensory axis cylinders arise, namely through the union of a number of fine network fibres.

In these propositions lies the core of Golgi’s views. […] The protoplasmic processes are not nervous; rather, they are connected with what R. Virchow termed the neuroglia fibres, and with the blood vessels of the central nervous system, and probably represent a nutritive apparatus for the ganglion cells.

2. Historical Commentary

Waldeyer here extracts what he calls “the core of Golgi’s views” as a compact, three-part doctrine: (1) Type I (“motor”) cells send a long nerve process directly into a motor root fibre; (2) Type II (“sensory”) cells feed into a diffuse network from which sensory fibres somehow emerge by the union of many fine network threads, rather than from any single traceable cell; and (3) the protoplasmic processes (dendrites) are not nervous tissue at all, Golgi assigns them a purely nutritive role, linking them instead to Rudolf Virchow’s neuroglia (the supporting, non-neuronal cells of the nervous system, now understood as astrocytes, oligodendrocytes, and other glial types) and to blood vessels.

This third claim is historically significant and easy to overlook: Golgi, despite being the great pioneer of neuron visualization, denied that dendrites carry nervous signals at all, treating them instead as a kind of feeding structure analogous to roots drawing nourishment from soil. This view would later be overturned, dendrites are now understood as the primary receptive structures of the neuron, essential to nervous signaling, not accessory nutritive appendages. Waldeyer, in citing this claim without yet endorsing or rejecting it, is setting up a debate that Ramón y Cajal and Kölliker will directly engage on subsequent pages.

3. Neuroscience Commentary: 1891 Compared With Today

Golgi’s motor/sensory dichotomy for Type I/Type II cells is now understood to be an oversimplification, many Type I (long-axon, projection) neurons are neither purely motor nor purely sensory in the specific sense Golgi meant, and sensory information is not generated by “union of network fibres” but by the graded, discrete synaptic transmission of signals along dedicated chains of individual sensory neurons, exactly the kind of pathway Ramón y Cajal would later trace cell by cell. Golgi’s claim about the non-nervous, purely nutritive role of dendrites is now known to be incorrect in its main thrust, dendrites are unambiguously part of the neuron’s signal-receiving apparatus, studded with synapses and, in many neuron types, capable of complex active electrical processing in their own right (dendritic spikes, a phenomenon entirely unknown in 1891). Golgi was, however, not simply wrong to link glial and vascular elements to nutritive support of neurons, modern neuroscience does recognize an essential metabolic and trophic relationship between neurons, astrocytes, and the vasculature (e.g., the neurovascular unit, astrocyte-mediated glucose and lactate shuttling), he simply misassigned this genuine biological relationship to the wrong anatomical structure (dendrites rather than glial processes proper).

4. Conceptual Reflection

Golgi’s “dendrites are not nervous” claim is a useful case study in how a brilliant observer, working with genuinely revolutionary tools, can still be led into error by an overly tidy theoretical scheme. Having established a clean distinction between “nerve processes” (axons, actively signaling) and “protoplasmic processes” (dendrites, passively nourishing), Golgi’s reticular theory required dendrites to play a subordinate, non-signaling role, because if dendrites were themselves conducting structures participating in a fused network, the entire architecture of his theory would become far more tangled. The error here is instructive: theoretical elegance was allowed to guide interpretation of ambiguous staining results, in a direction that later, more direct physiological evidence would not support. This is a recurring hazard in science generally, the pull toward parsimony can sometimes anticipate nature correctly, and sometimes lead a very capable scientist astray.

1. Modern English Translation (of the 1891 German)

As new and particularly outstanding, we find in Golgi’s work the following three things: 1. The branching of the nerve processes of the motor root fibres, 2. The branching, by side branches toward the nervous network, of the axis cylinders of the white columns of the spinal cord, 3. The non-nervous nature of the protoplasmic processes.

From this it must be concluded, a) that, broadly speaking, the posterior and anterior roots arise in the same way, in that both maintain connections with the nervous network; only the anterior roots have a more immediate additional relation to the ganglion cells, and their relation to the nervous network appears to be the secondary matter, whereas for the sensory roots it is the principal matter; b) that a completely isolated conduction exists in no nerve fibre.

2. Historical Commentary

Waldeyer pauses here to distill Golgi’s contribution into a tight three-point summary, a rhetorical technique he will use repeatedly throughout the essay to keep a complex, footnote-heavy survey navigable for his lecture audience. The three points are: axon branching in motor fibres, axon collateral branching in white matter tracts, and the non-nervous status of dendrites. From these he draws two consequences that state Golgi’s reticular theory in its strongest and most consequential form: that motor and sensory pathways are fundamentally similar in relying on the shared nervous network, and, most importantly, “a completely isolated conduction exists in no nerve fibre.”

This final sentence is the crux of the entire Golgi–Cajal dispute, rendered as plainly as possible. If true, it would mean the nervous system cannot be decomposed into separately-functioning individual units at all, signal conduction would be inherently a property of the network as a whole, not of any single traceable cell-to-cell pathway. This is the single claim that the rest of Waldeyer’s essay, drawing on Ramón y Cajal and Kölliker, will ultimately overturn.

3. Neuroscience Commentary: 1891 Compared With Today

Point 1 and 2 of Golgi’s three discoveries, that axons branch, both near their origin and via collaterals along their length, are entirely confirmed by modern neuroscience and remain basic, uncontested facts of neuronal structure. Point 3, the non-nervous nature of dendrites, is now understood to be incorrect, as discussed on the previous page. The overall conclusion, that “a completely isolated conduction exists in no nerve fibre”, is the point on which the entire twentieth century’s electron microscopy and electrophysiology would ultimately rule against Golgi: neurons are discrete conducting units, and while their axons do branch extensively (validating points 1 and 2), branching is not the same as anatomical fusion with other cells. A single axon’s branches remain the branches of one neuron’s cytoplasm, contacting, but not merging with, the many downstream targets it reaches.

4. Conceptual Reflection

This page shows how a scientist’s genuinely correct, empirically well-supported observations (axon branching) can be marshaled in service of a broader theoretical conclusion (network fusion, no isolated conduction) that later proves mistaken. The branching findings did not, by themselves, require the reticular interpretation, they were compatible with a very different reading, in which extensively branching axons make many discrete point-contacts with many discrete target cells, without ever fusing with any of them. That alternative reading is exactly what Ramón y Cajal would propose. The lesson is that data typically underdetermines theory: the same set of solid observations can support more than one overarching interpretation, and choosing between them requires additional evidence, or, sometimes, decades of additional evidence.

1. Modern English Translation (of the 1891 German)

S. Ramón y Cajal now comes into conflict with Golgi on some points, while in others he considerably extends our knowledge. As regards the most important differences between him and Golgi: S. Ramón y Cajal does not accept the network-like connection of the branchings of the nerve processes (i.e., an anastomosis of these processes with one another). The stem fibre of the nerve process continues into an axis cylinder, which may run either centripetally or centrifugally (anterior roots). The branches of this nerve process that still branch within the grey matter of the central organs all run out freely, without anastomosing with the corresponding processes of other nerve cells. S. Ramón speaks only of two exceptional cases: in the sympathetic plexuses of vertebrates, and in the nerve cells of insects, anastomoses between the processes of different cells are said to occur.

As a further consequence of what has been developed above, it follows that one cannot distinguish two kinds of cells as motor and sensory (Type I and Type II according to Golgi) in the manner Golgi does, and herein lies a second major difference from Golgi. One can only say that there are cells with long nerve processes (Golgi’s Type I) and with short nerve processes (Golgi’s Type II).

A third decisive difference between the conception of S. Ramón y Cajal and that of Golgi lies in the fact that the former declares the protoplasmic processes (dendrites, His) to be likewise nervous processes.

2. Historical Commentary

This page marks the true turning point of the essay: the introduction of Santiago Ramón y Cajal’s counter-reading of the silver-stained material. Waldeyer lays out three precise disagreements with Golgi, each of enormous consequence:

First, and most fundamentally, Cajal denies that the branching nerve processes of different cells fuse into a network at all. Instead, every branch, no matter how finely it divides within the grey matter, “runs out freely” (läuft frei aus), ending without physical union with any other cell’s process. This is the direct denial of Golgi’s reticular theory and the empirical seed of what would become the neuron doctrine. Waldeyer notes Cajal allows only two narrow exceptions (sympathetic plexuses of vertebrates, insect nerve cells), a telling detail, because it shows Cajal himself was not dogmatic; he was prepared to report genuine exceptions where his own observations indicated them, rather than forcing every case into a single rule.

Second, Cajal rejects Golgi’s clean motor/sensory (Type I/Type II) dichotomy as a functional distinction, retaining only the morphological distinction between long-process and short-process cells, a much more modest, purely descriptive claim, stripped of the physiological interpretation Golgi had attached to it.

Third, and perhaps most far-reaching for later neuroscience, Cajal declares that the protoplasmic processes (dendrites) are also nervous, directly contradicting Golgi’s claim (from Page 9 above) that dendrites are mere nutritive appendages linked to glia and blood vessels. Waldeyer notes here, in passing, the important terminological credit: it is Wilhelm His who introduced the term “dendrite” for these processes, a naming contribution parallel to, and just prior to, Waldeyer’s own naming of “neuron.”

3. Neuroscience Commentary: 1891 Compared With Today

All three of Cajal’s departures from Golgi have been vindicated by modern neuroscience, in each case decisively. The “no anastomosis” claim is confirmed by electron microscopy: synaptic contacts are close appositions with a physical gap (the synaptic cleft), not fused continuities, except at the specific sites of electrical synapses (gap junctions), a real but comparatively rare exception in vertebrate systems that gives Cajal’s own cautious exception-making an odd, partial vindication (he was right that most connections are not fused, and right that a minority of specific cases show tighter coupling, though the specific mechanism of gap junctions was unknown to him). The rejection of a strict motor/sensory dichotomy for Type I/Type II cells anticipates the modern, much richer taxonomy of neuron types, which cuts across many functional and morphological axes rather than a single binary. And the recognition that dendrites are genuinely nervous, signal-relevant structures is one of the cornerstones of modern cellular neurophysiology, dendrites are now known to receive the overwhelming majority of a neuron’s synaptic inputs and to perform substantial computation of their own.

4. Conceptual Reflection

This page is the empirical heart of the neuron doctrine, and it is worth noting explicitly, in line with this edition’s guiding distinction, that none of this is Waldeyer’s own discovery. He is reporting Cajal’s claims, evaluating their coherence, and, as later pages will show, ultimately endorsing them and giving the resulting synthesis a name. The scientific labor of looking through the microscope and refusing to see what Golgi saw belongs entirely to Cajal. Waldeyer’s labor is different but not trivial: it is the labor of recognizing, from a position of broad comparative knowledge across many researchers’ work, that Cajal’s reading is the one best supported by the weight of the combined evidence, and of making that judgment persuasively legible to the wider medical community.

1. Modern English Translation (of the 1891 German)

The nerve process then runs on as the axis cylinder of a motor nerve fibre, over a great distance, up to a muscle fibre, and only here, at the motor endplate, does the antler-like (Kühne) branching begin (Ranvier’s “arborisation terminale”), with freely running-out endings. The behaviour of the long sensory axis-cylinder processes is no different: these too stand in connection with only a single ganglion cell (in the spinal ganglion), but toward the periphery they run out, according to S. Ramón y Cajal, into a freely ending arborization.

2. Historical Commentary

Having established that nerve processes do not fuse with one another inside the central organs, Waldeyer now extends the same principle to the periphery, the point where nerve fibres actually meet their targets: muscle fibres (motor) and sensory receptor structures. He cites Wilhelm Kühne’s description of the “antler-like” terminal branching at the motor endplate, and Louis-Antoine Ranvier’s term arborisation terminale (“terminal arborization”) for the same phenomenon, both pre-existing observations that Cajal’s broader framework now incorporates as instances of a single general rule: nerve fibres everywhere end freely, in branching terminal structures, rather than fusing continuously with their targets or with other nerve fibres.

This is an important structural move in the essay’s argument: Waldeyer is showing that the free-ending principle is not confined to the exotic terrain of central grey matter, visible only with Golgi’s temperamental stain, it is also consistent with decades of prior, independently-established knowledge about the neuromuscular junction, one of the most extensively studied peripheral structures in nineteenth-century physiology.

3. Neuroscience Commentary: 1891 Compared With Today

The motor endplate, now more commonly called the neuromuscular junction, remains one of the best-characterized synapses in all of biology, and modern electron microscopy and physiology have fully confirmed the free-ending, non-fused nature of the nerve terminal there: the presynaptic terminal releases acetylcholine across a genuine synaptic cleft onto the postsynaptic muscle membrane, with no cytoplasmic continuity. The “antler-like branching” Kühne and Ranvier described corresponds to what is now understood as the terminal presynaptic bouton structure, from which neurotransmitter-containing vesicles are released. This page shows nineteenth-century physiologists arriving, via purely peripheral and comparatively easy-to-study tissue, at essentially the correct structural picture well before the central nervous system’s far more difficult-to-resolve connectivity could be settled.

4. Conceptual Reflection

It is notable that the neuromuscular junction, arguably the easiest synapse to study, given its large size and accessibility, had already yielded a broadly correct “free ending” picture well before the central controversy over reticular versus discrete connectivity was resolved. This illustrates a common pattern in biological discovery: peripheral, accessible systems often serve as proving grounds and analogical templates for principles that are then extended, with appropriate caution, to less accessible central systems. Waldeyer’s citation of Kühne and Ranvier alongside Cajal is a deliberate strategy of triangulation, showing that the free-ending principle is not an artifact of one exotic new stain, but a pattern independently observed by multiple techniques in multiple contexts.

1. Modern English Translation (of the 1891 German)

S. Ramón y Cajal now provides, for the spinal cord, to dispose of this matter here directly, a classification of the cells into different groups on the basis of the course of their nerve processes. He distinguishes: 1. Commissural cells, 2. Tract cells [Strangzellen], 3. Nerve root cells, 4. Cells with richly branched axis-cylinder processes of the posterior horns.

The commissural cells are found in every one of the known cell groups of the spinal cord (Golgi, Edinger, and others); their nerve processes cross through the anterior commissure to the opposite side, where they connect with one or more of the longitudinal fibres of the antero-lateral columns, either by T-shaped junction or by simply bending into such a fibre. The tract cells, likewise belonging to all parts of the grey matter, similarly let their nerve processes become the axis cylinder of one or more of the white fibres of all the spinal columns, but without crossing. The nerve root cells send out their strong nerve process as the axis cylinder of an anterior nerve root.

Figure 2, reproduced from the original 1891 offprint (public domain).

2. Historical Commentary

Here Waldeyer reports Cajal’s systematic functional-morphological classification of spinal cord neurons based on the destination of their nerve processes, a considerably more refined taxonomy than Gerlach’s earlier binary (cells with versus without processes) or even Golgi’s binary (Type I versus Type II). Cajal distinguishes commissural cells (crossing to the opposite side of the cord), tract cells (staying on the same side, feeding into ascending or descending white matter tracts), nerve root cells (sending fibres directly out via the ventral root to the periphery), and a fourth category of richly branching dorsal horn cells. Waldeyer credits this classification explicitly to Cajal, while noting that the commissural cell group itself had already been recognized, in a more limited form, by Golgi and by Ludwig Edinger (another leading German neuroanatomist and author of an influential textbook cited elsewhere in the essay).

This page demonstrates the transition from broad, binary theoretical categories (Golgi’s motor/sensory split) to a granular, empirically grounded cell taxonomy organized purely by anatomical trajectory, where does a given cell’s process actually go? This is a methodological shift of real significance: it treats cell identity as an empirical, traceable property rather than a theoretical inference from function.

3. Neuroscience Commentary: 1891 Compared With Today

Cajal’s spinal cord classification anticipates, with remarkable durability, categories still used in modern spinal cord neuroanatomy: commissural interneurons (which cross the midline, essential for coordinating bilateral movement and studied intensively in modern developmental neuroscience, for instance in the classic work on axon guidance at the floor plate), propriospinal (tract) neurons (connecting different spinal segments on the same side), and motor neurons proper (Cajal’s “nerve root cells,” sending axons directly to the periphery). The T-shaped and simple-bending patterns of fibre insertion that Cajal described for commissural and tract cells correspond to real, still-recognized patterns of axon collateral branching within the white matter tracts. This page shows Cajal’s classificatory scheme functioning almost as a modern circuit diagram in embryonic form, organized entirely by connectivity, which remains the primary organizing principle of circuit neuroscience today.

4. Conceptual Reflection

The shift illustrated on this page, from theory-driven binary categories to trajectory-based empirical classification, foreshadows a durable methodological commitment in neuroscience: that a neuron’s identity is best defined by where it connects, not merely by its presumed function or its gross morphological family. This connectivity-first approach to classification is the conceptual ancestor of the modern connectome project, which seeks to map, exhaustively and empirically, exactly which neurons connect to which, precisely the kind of project Cajal’s hand-drawn, cell-by-cell tracings represent in miniature, a century before electron microscopy made whole-circuit reconstruction feasible at scale.

[Edition continues in Part I, Page 14, “Kölliker’s Independent Confirmation, and the Question of the Dendrites”]

1. Modern English Translation (of the 1891 German)

Kölliker, in his most recent treatise, states, in confirmation and extension of the results of Golgi and S. Ramón, as the actual behaviour of the protoplasmic processes, that in all nerve cells of the spinal cord these branch most finely over great distances in all directions, often penetrate into the white substance, and do not anastomose. Kölliker weighs carefully and thoroughly the reasons speaking for and against Golgi’s view of the protoplasmic processes, but arrives as yet at no conclusive opinion.

2. Historical Commentary

Waldeyer now introduces Albert von Kölliker, already familiar from earlier pages as one of the great, cautious German authorities. Kölliker’s role in this essay is distinctive: he is not a discoverer of the silver-staining method (that credit belongs to Golgi) nor its most systematic and prolific new-fact generator (that role belongs to Cajal), but rather an independent, extremely careful verifier, a German anatomist of the highest standing, working with his own preparations, checking whether the extraordinary claims coming out of Pavia and Barcelona actually replicate.

His confirmation matters enormously for the essay’s argument, because it addresses a basic scientific worry: were Golgi’s and Cajal’s results perhaps artifacts of unusual staining conditions, seen only by them? Kölliker’s independent replication, confirming that protoplasmic processes (dendrites) branch extensively without anastomosing, substantially strengthens the evidentiary basis for the anti-reticular, discrete-cell position. Yet Waldeyer is careful to note that even Kölliker, for all his confirmation of the non-anastomosis finding, remains genuinely undecided on the separate question of whether dendrites are nervous or non-nervous in function, showing that “confirmed by multiple independent observers” and “theoretically resolved” are two different scientific milestones, and the essay has reached only the first.

3. Neuroscience Commentary: 1891 Compared With Today

The pattern shown here, independent replication by a scientist not affiliated with the original discovery, is one of the bedrock practices of the scientific method, as true today as in 1891. Kölliker’s meticulous, non-partisan checking of both Golgi’s and Cajal’s claims exemplifies good scientific citizenship: rather than simply picking a side based on reputation or national scientific tradition, he re-examined the material himself. Modern neuroscience has, of course, resolved definitively the question Kölliker left open, dendrites are now firmly established as nervous, signal-processing structures, but the manner in which that resolution eventually came, through accumulating, converging, independently replicated evidence rather than any single decisive experiment, mirrors exactly the process visible on this page.

4. Conceptual Reflection

This page is a valuable illustration of the distinction between consensus-building and truth-finding as two related but separable scientific activities. Kölliker’s independent confirmation of the “no anastomosis” finding is a genuine step toward consensus, it means the finding is not an artifact of one lab’s technique, but consensus on one point (non-fusion) does not automatically deliver consensus on a related point (nervous versus non-nervous function of dendrites). Waldeyer’s essay, itself an act of consensus-building, is careful to preserve this distinction rather than blur it for rhetorical convenience, modeling exactly the kind of intellectual honesty that makes his eventual synthesis, several pages later, credible rather than merely assertive.

1. Modern English Translation (of the 1891 German)

Speaking for their nervous nature, according to Kölliker, are: 1) Their texture and chemical behaviour, neither of which differs essentially from that of the nerve processes, and which agree closely with that of the cell body. 2) The fact, established by Golgi himself and confirmed by both S. Ramón and Kölliker, that nerve processes not infrequently arise from the protoplasmic processes, and indeed at a considerable distance from the cell body. 3) The extraordinary size, abundance, and branching of these processes. 4) The behaviour of the so-called “brush cells” [Federbuschzellen].

Speaking against the nervous nature of the dendrites, according to Kölliker, are: 1) The very frequent penetration of branches of protoplasmic processes into the white substance, often almost to its surface. 2) The advance of protoplasmic processes into regions where there are no nerve fibres at all. 3) Golgi’s statement that the protoplasmic processes attach themselves to glial cells and to blood vessels. 4) The circumstance that nowhere does one see relations of the protoplasmic processes either to other protoplasmic processes or to the nerve processes.

2. Historical Commentary

This page shows Kölliker doing something Waldeyer clearly admires and wants to model for his audience: laying out a genuinely balanced, itemized weighing of evidence on both sides of an unresolved question, rather than simply announcing a verdict. The stakes of the question, are dendrites nervous or not?, bear directly on whether the neuron doctrine, once formalized, would need to include dendrites as functional signaling structures or treat them as accessory nutritive appendages, following Golgi.

The “for” list draws on chemical and morphological similarity between dendrites and known nervous tissue, the striking observation that nerve processes (axons) sometimes arise directly from dendrites rather than from the cell body itself (a genuine and still-recognized phenomenon called axon-bearing dendrites in specific cell types), and the peculiar behavior of “brush cells” (Federbuschzellen), a term referring to olfactory bulb cells with tuft-like dendritic endings that appeared, even to contemporary observers, functionally active rather than merely nutritive.

The “against” list is equally serious: dendrites penetrate into white matter regions where there are no nerve fibres at all (suggesting a non-signaling role), Golgi’s own observation that dendrites attach to glial cells and blood vessels (suggesting a nutritive role), and, significantly, the observation that dendrites never show connections to other dendrites or to axons, which Kölliker takes as evidence against active signaling participation (an inference that would later prove incorrect, since dendrites do receive synaptic contact from axons, just not through anastomotic fusion).

3. Neuroscience Commentary: 1891 Compared With Today

Modern neuroscience has resolved this debate decisively in favor of the nervous, signal-relevant nature of dendrites, but it is worth appreciating how reasonable Kölliker’s “against” list actually was given available evidence. The observation that dendrites don’t show visible connections to axons or other dendrites was, in a sense, technically correct under nineteenth-century light microscopy, synaptic contacts are submicroscopic structures that cannot be resolved at the light-microscope level; what Kölliker was looking for (a visible fusion or connection) was never going to be seen, because the actual mechanism (a narrow synaptic cleft with specialized pre- and postsynaptic membrane machinery) requires electron microscopy, developed in the mid-twentieth century, to visualize directly. Meanwhile, the “for” observations were prescient: axon-bearing dendrites are a real, recognized phenomenon in certain neuron types, and the olfactory “brush cells” Kölliker cites do indeed play an active signaling role in odor processing, as modern olfactory neuroscience confirms.

4. Conceptual Reflection

Kölliker’s balanced list is a model of scientific reasoning under genuine uncertainty, and it is worth noting explicitly that both sides of his list contain real, correctly-observed facts, the disagreement was not about what was seen, but about how to interpret consistently-observed patterns given the available theoretical vocabulary. This is a recurring feature of unresolved scientific debates: the observational record is often not actually contested; what’s contested is which observations should be treated as decisive and which as incidental. Kölliker’s inability to reach a firm conclusion here is not indecisiveness, it is an accurate reflection of the genuine limits of 1891 technology to settle a question that would ultimately require an entirely different order of magnification to resolve.

1. Modern English Translation (of the 1891 German)

Also important for the question in hand are the investigations of G. Fritsch. In the two large giant ganglion cells of the electric catfish, the peripherally situated protoplasm resolves into a kind of net-like perforated plate, which then radiates ever further outward into the protoplasmic processes. The totality of the dendrites thus appears as a gradually progressing unravelling of the protoplasmic body of the ganglion cell, whereby, however, as emerges unmistakably from Fritsch’s investigations, near the cell body the individual dendrites are connected in net-like fashion with one another. Into the meshes of this network penetrate loops of blood vessels, which bore deep into the great body of the ganglion cell. From this it follows, in my view, that one must conceive of the protoplasmic processes just as one conceives of the protoplasm of the ganglion cell itself; the name originally given to them by Deiters is therefore entirely apt. They are of nervous nature, just as the protoplasmic body of the ganglion cells is, and must be judged functionally exactly as the latter is.

2. Historical Commentary

Waldeyer now brings in a strikingly different line of evidence: Gustav Fritsch’s work on the giant ganglion cells of the electric catfish (Malopterurus electricus) and, elsewhere in his broader body of work, the electric rays (Torpedinidae), the same family of “electric fish” whose electric lobe had already yielded Deiters’s foundational observations decades earlier (recall Rudolph Wagner’s finding on Torpedo, generalized by Deiters, discussed in Page 3). Electric fish were prized nineteenth-century research subjects precisely because their electric organs are innervated by unusually large, easily visualized giant neurons, offering, in effect, a naturally magnified specimen for studying cellular structure that ordinary mammalian neurons could not provide at that resolution.

Fritsch’s key observation is that the dendrites of these giant cells appear, near the cell body, as a continuous, net-like “unravelling” of the cell’s own protoplasm, not a separate class of structure grafted onto the cell, but literally more of the same substance as the cell body itself, simply thinning and branching outward. This is exactly why Waldeyer immediately endorses Deiters’s original terminology (“protoplasmic processes”) as “entirely apt”, since Fritsch’s observation shows these structures genuinely are continuous, functionally unified extensions of the cell’s own protoplasm, not a distinct anatomical category. Waldeyer draws the direct conclusion, his own first clearly stated personal judgment in the essay so far, that dendrites must be “of nervous nature… judged functionally exactly as” the cell body is. This is Waldeyer explicitly siding with Cajal and against Golgi on the question left open by Kölliker on the previous two pages, and doing so on the strength of independent evidence from a third research tradition (comparative/electric-fish anatomy) rather than repeating either Cajal’s or Kölliker’s own reasoning.

3. Neuroscience Commentary: 1891 Compared With Today

Fritsch’s observation, that dendrites are continuous with, and structurally the same substance as, the cell body’s protoplasm, is fully consistent with modern cell biology: dendrites are indeed simply specialized extensions of the neuronal cytoplasm, containing the same organelles, cytoskeletal elements, and much of the same biochemical machinery as the soma, differentiated by local protein trafficking and cytoskeletal organization rather than by being a separate kind of tissue. The image of blood vessel loops boring into the “meshes” of the near-cell-body dendritic network also anticipates, in a crude nineteenth-century way, the modern understanding of the tight anatomical and metabolic coupling between neurons and the surrounding vasculature, though modern neuroscience attributes the active mediation of that neurovascular coupling primarily to astrocytes, not to the neuron’s own dendrites directly.

4. Conceptual Reflection

This page is a good example of convergent evidence strengthening a scientific conclusion: Waldeyer does not rest his endorsement of the “dendrites are nervous” position solely on Cajal’s silver-stain interpretations or Kölliker’s careful weighing of pros and cons, but adds a third, quite different line of evidence, comparative anatomy of an extreme, naturally-occurring specimen (giant electric-fish neurons), that happens to make the same structural point through an entirely different observational route. When independent methods, applied by independent researchers to independent organisms, converge on the same conclusion, that convergence carries more evidentiary weight than any single line of evidence alone. This is precisely the kind of argument-by-convergence that would soon justify Waldeyer’s confidence in stating the neuron doctrine as a general law, in the pages still to come.

1. Modern English Translation (of the 1891 German)

The investigations of S. Ramón y Cajal further yield the following new results: 1. The posterior root fibres, arising from a ganglion cell of the spinal ganglia, divide dichotomously upon entering the spinal cord, namely at a rather acute angle (, <), into a longitudinally running ascending and descending branch. These branches form the entire mass of the posterior column fibres, both of Burdach’s and of Goll’s columns. 2. From the root fibres themselves, as well as from their two longitudinal branches, side branches, collaterals, S. Ramón’s term, go off, mostly at a right angle, which enter the grey matter and there come to an end with free, tree-shaped branched endings (arborizaciones finales, S. Ramón; “end-trees” [Endbäumchen], Kölliker). As already mentioned, these end-trees resemble the endings of the motor nerves on the striated muscle fibres, which, as is well known, Ranvier designated by the name “arborisation terminale”; S. Ramón’s naming follows from this same designation. 3. What has hitherto been described in the anatomy of the spinal cord as posterior root fibres directly penetrating into the grey matter are, in fact, only their above-mentioned collaterals, either the stem-collaterals or the branch-collaterals.

Figure 3, reproduced from the original 1891 offprint (public domain).

2. Historical Commentary

This page reports what is arguably Ramón y Cajal’s single most important structural discovery for spinal cord anatomy: the collateral branch. Before this finding, anatomists such as Gerlach (Page 4) had described posterior (sensory) root fibres as entering the grey matter directly and somehow dissolving into a diffuse network. Cajal’s silver-stained preparations revealed a far more orderly picture: each incoming sensory root fibre first splits (dichotomizes) into an ascending and a descending longitudinal branch upon entering the cord, these branches together constitute the well-known dorsal columns (named for Karl Friedrich Burdach and Friedrich Goll, two earlier anatomists whose named fibre tracts had been known for decades without a clear cellular origin). From these longitudinal branches, and from the root fibres themselves, smaller collateral branches peel off at roughly right angles, entering the grey matter and terminating in free, tree-like endings, which Cajal calls arborizaciones finales and Kölliker calls Endbäumchen (“end-trees”), explicitly modeled on Ranvier’s already-established term for the motor endplate’s “arborisation terminale” (see Page 12).

Point 3 is the crucial reinterpretation: what earlier anatomists had mistaken for the root fibres themselves diving directly into the grey matter were, Cajal shows, actually just their collateral branches, smaller offshoots of fibres whose main trunks continue on, undivided, into the long ascending/descending columns. This single reinterpretation resolves a great deal of the earlier confusion (going all the way back to the Page 1 uncertainty and the competing schemas discussed on Pages 4–6) about where sensory fibres “go.”

3. Neuroscience Commentary: 1891 Compared With Today

This page describes a genuinely durable and still-standard piece of spinal cord anatomy. The dichotomizing branch pattern of dorsal root ganglion axons entering the dorsal columns, and the existence of collateral branches terminating in the dorsal horn grey matter, remain textbook facts of modern neuroanatomy, Burdach’s and Goll’s (now more commonly called the cuneate and gracile) fasciculi are still named and studied, and Cajal’s basic wiring diagram for primary sensory afferents has been confirmed and extended by every subsequent technique, from degeneration studies to modern viral tracing. The concept of the axon collateral, a branch given off along an axon’s course, distinct from its main terminal arbor, remains one of the most basic organizing concepts of circuit neuroscience, essential to understanding how a single sensory afferent can simultaneously contribute to a long ascending pathway (carrying, e.g., touch information to the brainstem) while also engaging local spinal reflex circuits via its collaterals.

4. Conceptual Reflection

The reinterpretation described on this page, recognizing that what looked like one kind of structure (a root fibre “diving into” grey matter) was actually a different kind of structure (a collateral branch of a fibre that itself continues elsewhere), is a clean example of how improved resolution doesn’t just add new facts, it can reclassify previously known facts. The observation itself (fibres entering grey matter) had been made by many anatomists before Cajal; what changed was the ability to trace that same fibre far enough, in both directions, to see that it was a branch rather than a terminus. This is a durable lesson for interpreting any partial dataset: an apparent “ending” is often simply the edge of what the available technique can resolve, not necessarily the edge of the actual structure.

1. Modern English Translation (of the 1891 German)

S. Ramón y Cajal has now given still more precise data on these various points, some of which may be highlighted here. As regards the collaterals of the various columns, those of the anterior columns are somewhat stronger; their end-trees lie in the anterior horns, some also in the posterior horns of the same side; part of the anterior-column collaterals cross in the anterior commissure. The collaterals of the posterior column fibres are the finest and most numerous; some end only within the region of the anterior horns; part cross in the posterior commissure and then end, in part in the posterior horn, in part in the anterior horn of the other side.

The posterior root fibres, according to S. Ramón y Cajal, insert into the white column fibres in three ways: a) They insert wholly, at a right angle, onto a longitudinal nerve fibre, and at the point of connection there is always found a node of Ranvier. b) They first divide into two fibres, each of which inserts, in the manner just described, at a node of Ranvier of a longitudinal fibre. In both cases, a and b, we would thus have T-connections, and this would correspond perfectly to the behaviour of a posterior root fibre which, arising as the nerve process of a cell of the spinal ganglion, divides T-shaped upon entering the spinal cord into a longitudinally descending and ascending branch. c) The nerve process bends directly into the axis-cylinder process of a white column fibre. This is the case, for example, with the cells of Clarke’s columns, whose nerve process, running centripetally, bends longitudinally upward in order to continue in the cerebellar lateral tract.

Figure 4, reproduced from the original 1891 offprint (public domain).

Figure 5, reproduced from the original 1891 offprint (public domain).

2. Historical Commentary

This page provides the fine-grained detail behind the collateral discovery reported on the previous page, and introduces a structural relationship that would become a key piece of evidence for the discrete-connectivity view: the observation that collateral branches consistently insert onto longitudinal fibres precisely at a node of Ranvier (Ranvier’sche Einschnürung, the periodic gaps in the myelin sheath, already known from Louis-Antoine Ranvier’s earlier work). This precise, reproducible spatial relationship, always at a node, always in a T- or Y-shaped pattern, is presented as too regular and too orderly to be an artifact of a diffuse fused network; it looks like a designed, discrete point of connection.

Waldeyer also introduces the cells of Clarke’s columns (named for the British anatomist Jacob Augustus Lockhart Clarke, already mentioned in earlier method-history pages), whose axons bend directly upward to form the cerebellar lateral tract (spinocerebellar tract), a pathway carrying proprioceptive information to the cerebellum. This is presented as a clean, traceable example of Cajal’s third connection type, in which a cell’s nerve process becomes, without any intervening network, the origin of a major ascending white matter tract.

3. Neuroscience Commentary: 1891 Compared With Today

Clarke’s column (now more often called the nucleus dorsalis or Clarke’s nucleus) and its projection via the dorsal spinocerebellar tract remain standard, textbook neuroanatomy, essential to understanding how proprioceptive information from muscles and joints reaches the cerebellum for motor coordination. The observation that axon collaterals insert at nodes of Ranvier reflects a genuine anatomical regularity, though modern neuroscience explains its functional significance rather differently than nineteenth-century anatomists could: nodes of Ranvier are now understood primarily as the sites of saltatory conduction (where the action potential regenerates along a myelinated axon), and while collateral branch points often do occur near nodal regions, the deeper explanation involves axonal cytoskeletal and membrane specializations unknown to Cajal or Waldeyer.

4. Conceptual Reflection

The recurring, precise geometric regularity described here, T-shaped junctions, consistent association with nodes of Ranvier, functioned rhetorically and scientifically as evidence against the reticular theory: a genuinely fused, syncytial network would not be expected to show such crisp, repeatable, rule-governed junction patterns across many different cell types and pathways. Regularity of this kind is itself a form of evidence: structures built by a small number of consistent rules (discrete cells making discrete, patterned contacts) tend to produce visibly repeated motifs, whereas an undifferentiated fused mesh would not obviously generate such motifs. Waldeyer, in accumulating example after example of this kind of structural regularity, is building an argument from pattern as much as from any single decisive observation.

1. Modern English Translation (of the 1891 German)

A very remarkable finding is the one, already mentioned above, made by M. v. Lenhossék and S. Ramón y Cajal, according to which there exist, in the posterior roots of the chick, fibres that, in the manner of anterior roots, arise from one of the lateral ganglion cells of the anterior horn, run posteriorly, then suddenly bend at a right angle into the posterior root and pass through the spinal ganglion, without entering into connection with any cell of the latter. Only a few cells of this kind can be present, and thus far they have been found only in the chick. S. Ramón designates them plainly as “cellules motrices de la racine postérieure” [motor cells of the posterior root]; Lenhossék prefers to describe them only as “centrifugally conducting.”

2. Historical Commentary

This page reports a genuine anomaly, a finding that does not fit the tidy motor-anterior/sensory-posterior division that had organized the discussion up to this point. Mihály (Michael) von Lenhossék, a Hungarian anatomist working closely with Kölliker, and Ramón y Cajal, independently observed a small population of cells in the chick embryo whose axons behave like motor fibres (arising from anterior-horn-type ganglion cells) but travel out via the posterior (sensory) root rather than the anterior root, passing straight through the spinal ganglion without forming any connection with the sensory ganglion cells there. Waldeyer reports the researchers’ differing interpretive caution: Cajal is willing to call them outright “motor cells of the posterior root,” while Lenhossék, more conservatively, will commit only to describing them functionally as “centrifugally conducting” (i.e., carrying signals away from the centre, like a motor fibre, without asserting they are truly motor in function).

This is a valuable moment in the essay because it shows Waldeyer resisting the temptation to smooth over an inconvenient exception in service of a tidier narrative. Real anatomy, he is showing his audience, contains genuine oddities that do not fit the cleanest possible rule, and honest science reports them rather than suppressing them.

3. Neuroscience Commentary: 1891 Compared With Today

The general phenomenon Lenhossék and Cajal describe, an atypical population of cells with axons exiting via an unexpected route, anticipates a category of findings that modern developmental neuroscience recognizes much more broadly: various embryonic neuron populations show transient or atypical axon trajectories during development that do not always match their adult, “textbook” wiring pattern, and some populations of efferent (centrifugally conducting) fibres genuinely do run alongside afferent sensory fibres in mixed nerve roots in ways that complicate a simple anterior-motor/posterior-sensory dichotomy (which itself, embodied in the “Bell-Magendie law,” was itself a nineteenth-century generalization with real, if limited, exceptions). This specific chick-embryo finding is a narrow, somewhat dated observation by modern standards, but the broader pattern, atypical, minority cell populations complicating clean anatomical rules, remains a live and recurring theme in circuit neuroscience.

4. Conceptual Reflection

The inclusion of this admittedly minor, exception-reporting finding illustrates something valuable about how scientific rules and their exceptions relate to one another. A rule like “anterior roots are motor, posterior roots are sensory” (the Bell-Magendie law) is useful and largely correct, but nature routinely produces small, genuine exceptions to even well-established rules, and a mature scientific literature reports and investigates those exceptions rather than dismissing them as noise. Waldeyer’s willingness to include this oddity, even while building toward a grand unifying doctrine of his own (the neuron doctrine, still a few pages away), shows an appropriate scientific humility: a general law can be true and useful while still admitting documented exceptions at its margins.

1. Modern English Translation (of the 1891 German)

The investigations of Kölliker, to which I now turn, concern the spinal cord, the cerebellar cortex, and the medulla oblongata. They claim a particular importance for two reasons: first, because they investigate the regions named in a very thorough and extensive manner, and second, because they have secured for the hitherto only little-regarded statements of Golgi and S. Ramón y Cajal the attention that rightly belongs to them, at least among us [German-speaking anatomists]. Kölliker was able to confirm, to the fullest extent, the findings obtained by Golgi and in particular by Ramón y Cajal, and to extend them in a number of points.

2. Historical Commentary

Waldeyer here makes an important sociological observation about the international transmission of scientific knowledge in the 1880s: Golgi’s Italian-language publications and Cajal’s Spanish-language publications had, by Waldeyer’s own account, received only limited attention within the German scientific community before Kölliker took them up. Kölliker’s extensive, independent replication across three major CNS regions functioned not only as scientific verification but as a translation and legitimation event, bringing findings published in less widely-read languages and journals into the mainstream of German (and by extension, broader European) neuroanatomy, where Waldeyer himself was now positioned to further synthesize and popularize them via this very essay. This page makes explicit something implicit throughout: language barriers and differing national publication cultures were a real obstacle to scientific progress in this period, and figures like Kölliker (and Waldeyer himself) played a crucial bridging role.

3. Neuroscience Commentary: 1891 Compared With Today

The pattern described here, an idea from one national research tradition requiring a respected local authority to “vouch” for it before being broadly accepted elsewhere, is less visible in today’s globalized, English-dominant scientific publishing environment, but analogous dynamics persist: findings published in less widely indexed journals, or in languages with smaller international readership, can still take longer to enter mainstream awareness and citation networks, even in an era of instant electronic access. The episode is a reminder that scientific consensus-formation is a social as well as an empirical process.

4. Conceptual Reflection

Kölliker’s role as replicator-and-legitimator, more than originator, mirrors Waldeyer’s own role in this essay at one level removed: just as Kölliker made Golgi’s and Cajal’s findings credible and visible to German anatomists through his own independent confirmation, Waldeyer is making the entire body of findings, Golgi’s, Cajal’s, Kölliker’s, Retzius’s, and others’, credible and visible to the wider medical community through synthesis and naming. Neither man discovered the neuron; both were essential to its acceptance as scientific fact.

1. Modern English Translation (of the 1891 German)

Golgi taught us, by means of his method, an unimagined richness of branching of the distal processes of the Purkyně cells directed toward the cerebellar surface, these are their protoplasmic processes, and I must confess that, although the beautiful preparations of J. v. Gerlach, and those of G. Denissenko, who worked in my Strassburg laboratory in 1877, had shown me an extremely rich branching, I was nevertheless surprised when Dr. W. Zimmermann, assistant at the anatomical institute here, presented to me the cerebellar preparations he had made using Golgi’s method. Golgi further described more precisely the branching of the axis-cylinder processes, which had first been noted by myself. The protoplasmic processes, notwithstanding their extraordinarily rich branchings, enter into no connections demonstrable up to now; Golgi, however, in accordance with his views on their significance, lets them attach in part to the connective-tissue boundaries of the cerebellum and to the walls of the blood vessels.

Figure 6, reproduced from the original 1891 offprint (public domain).

2. Historical Commentary

This page contains one of the very few moments in the essay where Waldeyer speaks in the first person as an active researcher rather than a synthesizer, noting that he himself had “first noted” the branching of the axis-cylinder processes in the cerebellum, and recounting his own personal, first-hand experience of surprise upon seeing a Golgi-stained cerebellar preparation from his own assistant, Dr. W. Zimmermann, at his own institute. This is a valuable moment for the reader of this critical edition, since it shows that Waldeyer was not a purely passive compiler, he had genuine, direct laboratory experience with these staining techniques and had made minor original contributions of his own (the description of axon branching), even as the essay’s larger arc remains one of synthesis rather than primary discovery of the central doctrine.

The description of the Purkinje cell’s dendritic tree as showing “unimagined richness of branching” is historically significant: this cell, with its extraordinarily elaborate, fan-like dendritic arbor, became (and remains) one of the most iconic and instantly recognizable images in all of neuroscience, precisely because Golgi’s stain rendered its complexity visible for the first time. Note again, however, Golgi’s persistent theoretical commitment: even faced with this staggering dendritic complexity, he continues to interpret the dendrites as attaching to connective tissue and blood vessels, a nutritive rather than signaling role, showing how tenaciously a theoretical framework can persist even as the sheer scale of a phenomenon (here, dendritic branching complexity) seems to invite reconsideration.

3. Neuroscience Commentary: 1891 Compared With Today

The Purkinje cell dendritic arbor remains, to this day, used as a textbook example of dendritic architecture, and modern quantitative studies confirm what Golgi’s stain first suggested: a single Purkinje cell dendritic tree can receive well over 100,000 individual synaptic contacts, primarily from cerebellar granule cell parallel fibres. Waldeyer’s personal astonishment at seeing this structure captures something real about the phenomenon, it remains, even to modern researchers encountering it for the first time, a visually striking demonstration of the sheer combinatorial complexity that a single neuron’s receptive surface can achieve. Golgi’s insistence on a merely nutritive, vascular/connective-tissue role for these dendrites stands as one of the clearest instances in the essay of theoretical commitment outpacing what the observation itself actually required.

4. Conceptual Reflection

This page’s brief first-person aside is worth valuing for what it reveals about the texture of scientific practice: Waldeyer was not simply reading journal articles at a desk; he was replicating techniques in his own institute, training assistants in new staining methods, and reacting with genuine surprise to what those methods revealed. The neuron doctrine’s eventual acceptance rested on exactly this kind of distributed, hands-on replication across many laboratories, Golgi’s method had to actually work, reliably, in Strassburg and Berlin as well as in Pavia, before its findings could be trusted as general truths about nervous tissue rather than local curiosities of one laboratory’s technique.

1. Modern English Translation (of the 1891 German)

As is well known, three concentrically arranged layers can be distinguished in the cerebellar cortex: the central medullary layer, which is surrounded by the granular layer, this in turn bounded externally by the grey cortex or molecular layer, upon which the pia with its subpial layer rests. […] The number of side branches of the axis-cylinder process directed toward the granular layer is small, and the branches themselves are, as I had already stated, very fine; some of them run back to the molecular layer of the cerebellum (Golgi, Ramón, Kölliker). […] The others are larger, and show a very remarkable behaviour, investigated in particular by S. Ramón y Cajal and Kölliker: their protoplasmic processes run, rather richly branched, in the molecular layer; their axis-cylinder process, arising very fine from the cell body, gradually thickens and runs in a transverse-tangential direction closely above several Purkyně cells, to which it then sends branches which, with numerous fine ramifications, in the manner of a basket-weave, envelop the Purkyně cell bodies. Kölliker directly designates these formations as “fibre-baskets” [Faserkörbe].

2. Historical Commentary

This page describes the classic three-layer architecture of the cerebellar cortex (molecular layer, Purkinje cell layer implicit between them, granular layer, and the deep white matter) and introduces one of the most visually distinctive cell types in the entire nervous system: what Kölliker names “basket cells” (Faserkörbe, literally “fibre-baskets”), whose axons form dense, woven, basket-like terminal structures wrapping tightly around the cell bodies of the Purkinje cells. This is presented as a “very remarkable” (sehr merkwürdiges) structural finding, jointly credited to Cajal and Kölliker’s independent, converging investigations.

The basket cell finding is important to Waldeyer’s larger argument because it is an unusually vivid, hard-to-misinterpret example of precise, targeted cell-to-cell contact: a specific axon, from a specific identifiable cell type, forming a specific, elaborately structured terminal arrangement around specific target cells. This kind of exquisitely organized, geometrically specific contact pattern is very difficult to reconcile with the idea of an undifferentiated fused network, it looks, instead, like precision-engineered point-to-point wiring.

3. Neuroscience Commentary: 1891 Compared With Today

The cerebellar basket cell remains a standard, actively studied inhibitory interneuron type in modern neuroscience, and its “basket” of axon terminals wrapping around the Purkinje cell soma, including, as later research would show, around the axon initial segment specifically, is now understood in precise molecular and physiological detail: basket cells release the inhibitory neurotransmitter GABA onto Purkinje cells, providing powerful, spatially precise inhibitory control over Purkinje cell firing, essential to the temporal precision of cerebellar output. This represents one of the clearest possible illustrations of how a purely structural observation from 1891, a basket-like axonal wrapping, anticipated, by more than half a century, a specific and functionally crucial synaptic arrangement that modern electrophysiology would later characterize in full mechanistic detail.

4. Conceptual Reflection

The basket cell is a particularly striking early example of what would become a defining feature of the mature neuron doctrine: that precise, structured, and reproducible patterns of connectivity between identifiable cell types are the basic organizational currency of the nervous system. A fused, undifferentiated network could not, even in principle, produce something as specific and consistently reproducible as a basket-shaped terminal arrangement targeting one particular downstream cell type. The discovery of the basket cell is, in this sense, not just supporting evidence for the neuron doctrine, it is an early glimpse of what circuit-level neuroscience would eventually become: the study not just of whether neurons are discrete, but of the specific, meaningful architectural patterns by which discrete neurons connect to one another.

1. Modern English Translation (of the 1891 German)

Concerning the cerebral cortex, which of all parts of the central nervous system must naturally interest us the most, we possess, as is well known, an accurate and now foundational knowledge of the layers, owed to Meynert, to whom we indeed owe the opening of the new pathways, since trodden with such great success, in the finer anatomy of the brain. […] The statements of Ramón go the furthest. According to him, essentially three cell forms are to be assumed: a) the well-known large pyramidal cells, b) spindle-shaped, triangular, and polygonal forms in the upper cortical layers, c) more deeply situated, triangular or spindle-shaped cells. The large pyramidal cells have, as Retzius has particularly shown recently, an extraordinarily rich branching of their protoplasmic process toward the cortex, where they advance as far as the subpial layer.

Figure 7, reproduced from the original 1891 offprint (public domain).

2. Historical Commentary

Waldeyer now brings his region-by-region survey to its culmination: the cerebral cortex, which he explicitly flags as the structure of greatest natural interest, given its association (through Theodor Meynert’s highly influential localizationist framework) with higher psychological function. He credits Meynert with establishing the foundational layered structure of the cortex that subsequent Golgi-era researchers, especially Ramón y Cajal and Gustaf Retzius, would populate with detailed cell-type descriptions.

Cajal’s three-part classification of cortical neurons (large pyramidal cells; spindle/triangular/polygonal forms of the upper layers; deeper triangular/spindle forms) represents an early attempt at what would become, over the following century, an enormously elaborated project of cortical cell-typing, a project still very much ongoing in modern neuroscience via transcriptomic and morphological classification schemes. Retzius’s specific contribution highlighted here, showing the extraordinary richness of the pyramidal cell’s apical dendrite, reaching all the way to the subpial (outermost) layer of the cortex, parallels the Purkinje cell finding from Page 21: another instance of a Golgi-revealed dendritic tree turning out to be far more extensive and elaborate than anyone had previously suspected.

3. Neuroscience Commentary: 1891 Compared With Today

Cajal’s and Retzius’s basic three-part division of cortical neuron morphology, into pyramidal cells and various non-pyramidal (spindle, triangular, polygonal) forms, anticipates, in embryonic form, the modern division of cortical neurons into excitatory pyramidal projection neurons and a highly diverse population of inhibitory interneurons (many of them non-pyramidal in shape), a distinction now refined through molecular markers, electrophysiological firing patterns, and connectivity mapping into dozens of recognized subtypes. The observation that pyramidal cell apical dendrites extend all the way to the cortical surface remains a basic and still-taught fact of cortical anatomy, essential to understanding how a single pyramidal neuron can integrate synaptic input arriving at very different cortical depths and, correspondingly, from very different sources.

4. Conceptual Reflection

This page shows the neuron doctrine beginning to generate not just a general principle (cells are discrete) but a research program: once it becomes clear that individual, identifiable cell types can be reliably distinguished and studied, cataloguing and classifying those types becomes a natural and enormously fruitful next step. The three-way cortical classification offered here by Cajal is modest by the standards of what would follow, but it represents the first concrete steps of a project, systematic cortical cell-typing, that remains a major and unfinished frontier of neuroscience more than a century later.

[Edition continues in Part I, Page 24, “Comparative Anatomy: Haller, Retzius, Nansen, and the Crustacean and Invertebrate Nervous System”]

1. Modern English Translation (of the 1891 German)

The facts here emphasized are now also supported by several outstanding recent works in the field of comparative anatomy. […] Béla Haller’s results agree in many respects with those of Golgi. He assumes, like Golgi, a twofold origin of nerve fibres: on the one hand directly from the nerve processes of the ganglion cells (motor fibres), but also, for the sensory roots, from a network of anastomosing, finely branched processes of all the ganglion cells. […] Nansen, who is equally skilled with the microscope as with the alpenstock and snowshoes, has already, since 1885–1887, observed in Myxine and various invertebrates the division of the posterior root fibres entering the cord. As regards the protoplasmic processes, he agrees with Golgi’s conception. Fusari’s statements agree, as regards the origin and course of the nerve processes, with those of Nansen. Similar results, on most points, emerge from the thorough investigations of G. Retzius: almost all the ganglion cells of the crustaceans investigated (Astacus and Palaemon) have only a single process, the stem process, Retzius’s term, and are thus monopolar.

2. Historical Commentary

Waldeyer broadens his evidentiary base beyond vertebrate central nervous systems to comparative anatomy across a striking range of species: the invertebrate researcher Béla Haller, and, in a charming aside characteristic of the essay’s personal touches, Fridtjof Nansen, better known to history as the great Arctic explorer, who in his earlier scientific career had done serious comparative neuroanatomical work on the hagfish (Myxine) and various invertebrates. Waldeyer’s remark that Nansen is “equally skilled with the microscope as with the alpenstock and snowshoes” is one of the essay’s few moments of light personal color, and a reminder that many nineteenth-century scientists moved fluidly between what today would be considered entirely separate careers.

Crucially, this page shows that the debate was not resolved uniformly: Haller sides with Golgi, maintaining the reticular network view for sensory roots even in invertebrates, while Nansen also aligns with Golgi’s view on the non-nervous nature of protoplasmic processes. This is an important corrective to any narrative in which the Cajal/Kölliker position simply steamrolled all opposition, as of 1891, serious, competent researchers working on different organisms were still finding results that supported Golgi’s reticular interpretation, not Cajal’s. Retzius’s crustacean work, by contrast, introduces the concept of the monopolar neuron (a single process serving as the cell’s sole outgoing connection) as a cross-species structural motif, extending the discussion to invertebrate nervous systems built on a different basic plan than the vertebrate CNS.

3. Neuroscience Commentary: 1891 Compared With Today

Modern comparative neuroscience confirms that invertebrate nervous systems do indeed often rely heavily on monopolar neurons, a genuine and important structural difference from the typically multipolar neurons of the vertebrate central nervous system, Retzius’s crustacean observation holds up well. On the reticular-versus-discrete question, however, modern electron microscopy has established, across essentially all animal groups studied, that neurons remain discrete cells connected by synapses, not fused into networks, meaning that Haller’s and Nansen’s Golgi-aligned interpretations for invertebrates, like Golgi’s own interpretation for vertebrates, would eventually be overturned by higher-resolution methods, even though their specific observations (branching patterns, process division) often remained valid data points once reinterpreted within the discrete-cell framework.

4. Conceptual Reflection

This page is valuable precisely because it complicates a simple, triumphalist history of the neuron doctrine. As of mid-1891, the reticular theory still had capable defenders working on diverse species with legitimate observations of their own; the doctrine had not yet “won” in any final sense, and Waldeyer, writing in real time, could not have known with certainty how the debate would ultimately resolve (though the essay’s later summary pages show him placing his own bet clearly). Comparative anatomy in this period functioned as an important testing ground precisely because different species’ nervous systems, with their differing scales and structural conventions, offered independent opportunities to check whether a proposed general rule (discreteness, or its absence) actually held up across the full diversity of animal life, a test that, as this page shows, was still very much in progress.

1. Modern English Translation (of the 1891 German)

Further, the relations at the higher sense nerves can be adduced in support of the views of Golgi (in part), P. and S. Ramón, and Kölliker. Pedro and Santiago Ramón, further Dogiel, van Gehuchten and Castranuovo, have shown the direct connection of olfactory fibres with Max Schultze’s olfactory cells. Recently v. Brunn, whose beautiful preparations I have had the opportunity to examine, has placed this beyond all doubt for man. From the epithelial olfactory cells, these nerve fibres run to the glomeruli olfactorii, where they end freely, with terminal end-trees. The fine investigations of His supplement this finding, in that they show that the olfactory fibres develop from the olfactory epithelium toward the brain. Further, in the glomeruli, the processes of ganglion-like cells also end with free tufts, which in turn send fibres toward the tractus that run into the brain.

In the retina, the tract, and the optic lobe of birds, S. Ramón found the following relations: the rod and cone fibres do not hang continuously with any other retinal element; their ends, turned toward the inner granular layer, run there, branched and free. The bipolar ganglion cells of the inner granular layer likewise have such free endings, and such are also shown in the molecular layer, where free endings of the optic ganglion cells are also found.

2. Historical Commentary

Waldeyer now extends the discreteness argument to the peripheral sense organs, the olfactory system and the retina, showing that the free-ending, non-anastomosing principle established for spinal and cortical neurons also holds at the sensory periphery, where a nerve fibre meets a specialized receptor cell. The olfactory work he cites (the Ramón brothers Pedro and Santiago, Dogiel, van Gehuchten, and von Brunn) establishes a direct, traceable connection between the axons of olfactory sensory neurons and specialized structures called glomeruli, where these fibres terminate in free branching tufts alongside the dendrites of second-order neurons, an arrangement Waldeyer treats as structurally parallel to everything else described in the essay.

He also credits Wilhelm His with a complementary embryological finding: that olfactory fibres grow outward from the developing olfactory epithelium toward the brain during development, rather than the reverse, an observation that will become central to the essay’s later discussion of neuronal growth and embryology. For the retina, he reports Cajal’s finding that photoreceptor cells (rods and cones), bipolar cells, and ganglion cells all show free, non-continuous endings at each stage of the visual pathway, extending the free-ending principle across yet another entirely distinct sensory system.

3. Neuroscience Commentary: 1891 Compared With Today

The olfactory glomerulus remains, more than a century later, a cornerstone structure of olfactory neuroscience: it is now understood in precise molecular detail as the site where axons from olfactory sensory neurons expressing a specific odorant receptor converge onto the dendrites of second-order mitral and tufted cells, forming a discrete functional unit, one glomerulus per receptor type, an organizational logic entirely unknown in 1891 but structurally consistent with the free-terminal-convergence picture described here. The retinal findings are likewise fully confirmed: the modern retina is understood as a precisely layered, synaptically organized relay of discrete cell types (photoreceptors → bipolar cells → ganglion cells, with horizontal and amacrine cells providing lateral modulation), connected by genuine chemical synapses rather than any fused network.

4. Conceptual Reflection

Extending the discreteness principle to sensory transduction systems matters because it shows the neuron doctrine is not a special-case rule confined to central “processing” tissue, but a general property of how the nervous system is built from its very point of contact with the outside world. That the same free-ending, point-to-point architecture appears in the retina, the olfactory bulb, the cerebellum, the spinal cord, and the cerebral cortex, structures with wildly different functions, embryological origins, and evolutionary histories, is precisely the kind of cross-system generality that justifies treating discreteness as a fundamental architectural law of the nervous system, rather than an incidental feature of any one region.

1. Modern English Translation (of the 1891 German)

If we now cast a rapid glance at the histogenetic development of the central organs, we shall find here too much that seems suited to support the views that have been developed. From these works [of Vignal, His, Ramón y Cajal, and von Lenhossék], the following may suffice to be emphasized, in which the named investigators agree: 1) All the elements of the central nervous system first arise from a single, uniform ectodermal primordium; from this the spinal ganglia also arise. 2) One can distinguish, already within this primordium, two kinds of cells: the spongioblasts and the germ cells (His). […] The white nerve fibres grow out from the germ cells and appear at first as cone-shaped elongations of the same; these grow forward to their destination, carrying at their free, forward-growing end a thickening, from which small processes emerge, the growth cone, S. Ramón. This growth cone later forms the terminal end-tree. […] The sensory fibres grow in two directions at once, from a single spinal ganglion cell, so that the spinal ganglion cells appear at first bipolar in all creatures.

2. Historical Commentary

Waldeyer now shifts registers entirely, turning from mature adult anatomy to embryology, drawing on the work of Vignal, Wilhelm His, Ramón y Cajal, and von Lenhossék. This section serves a distinct evidentiary purpose: if fully mature nerve cells and fibres are difficult to trace to their ultimate origin using static, fixed adult tissue, embryonic development offers a second, independent line of evidence, since a developing structure can be caught in the act of forming, its process of construction directly observed rather than merely inferred from a finished product.

The page reports genuine consensus among these investigators on two foundational points: first, that all neural elements, neurons and their supporting glial cells alike, arise from a single, undifferentiated ectodermal embryonic layer, differentiating early into two lineages Wilhelm His terms spongioblasts (the glial precursor lineage) and germ cells (the neuronal precursor lineage). Second, and most strikingly, the page describes Cajal’s discovery of the growth cone (Wachsthumskegel), a specialized, actively motile structure at the growing tip of a developing nerve fibre, putting out small processes as it extends toward its eventual target and ultimately becoming the mature terminal arborization.

3. Neuroscience Commentary: 1891 Compared With Today

The growth cone remains, to this day, one of the central objects of study in developmental and regenerative neuroscience, now understood in exquisite molecular detail: it is a dynamic, actin-and-microtubule-rich structure that senses guidance cues in its environment (chemoattractant and chemorepellent molecules, extracellular matrix signals) and steers the extending axon toward its correct target with remarkable precision. Cajal’s original 1890s description of the growth cone, cited here by Waldeyer within a year of its first publication, stands as one of the most durable and prescient observations in the entire essay, modern axon guidance research, including Nobel Prize-winning work on netrins, slits, and semaphorins in the late twentieth century, is a direct continuation of the phenomenon first visualized here. The His/Cajal division of a common neural precursor into neuronal and glial lineages is likewise substantially confirmed by modern developmental neurobiology, though the specific molecular and lineage-tracing tools (transcription factor cascades, genetic fate-mapping) that establish this today were of course entirely unavailable in 1891.

4. Conceptual Reflection

This embryological interlude functions as a second, independent axis of evidence supporting discreteness: not only do mature nerve cells fail to show anatomical fusion, but their very mode of construction, a single fibre extending outward from a single cell body, tipped by a single actively exploring growth cone, seeking out its own specific path and target, is intrinsically a story about individual cellular agency, not network-wide, undifferentiated growth. Watching a structure being built, rather than only inspecting the finished product, offers a distinct and often more decisive kind of evidence about what that structure fundamentally is. This principle, that developmental biology can resolve questions that adult anatomy alone cannot settle, remains a foundational methodological insight across all of modern biology, extending well beyond neuroscience.

1. Modern English Translation (of the 1891 German)

Since Golgi’s investigative method has also given us many valuable insights into the behaviour of the supporting and framework substance of the central nervous system, I would like to devote a few words to this as well. That such a framework substance exists at all (apart from the blood vessels) was first shown for the spinal cord, on Reil’s suggestion, by G. G. Th. Keuffel, in a very careful piece of work. […] Through R. Virchow’s thorough studies, knowledge of this formation was substantially advanced. Although Virchow still regarded the tissue in question as a connective tissue, its differences from ordinary connective tissue had not escaped him, and he accordingly introduced the generally adopted name “nerve-cement substance” [Nervenkittsubstanz], neuroglia, or in short, “glia.” […] After the researches of Deiters, Frommann, Jastrowitz, Golgi, Boll, Ranvier, W. Müller, and Schiefferdecker, we arrived at more accurate conceptions of the structure of this remarkable tissue.

Figure 8, reproduced from the original 1891 offprint (public domain).

2. Historical Commentary

Waldeyer now takes up a subject that runs as a quiet parallel thread throughout the entire essay: the non-neuronal supporting tissue of the nervous system, which he traces from its earliest identification by G. G. Theodor Keuffel (working on Johann Christian Reil’s suggestion) through Rudolf Virchow’s foundational studies. Virchow, one of the towering figures of nineteenth-century pathology and cell theory, is credited with coining the term “neuroglia” (nerve-glue, from Greek glia, “glue”), even though Virchow himself still classified this tissue as a form of ordinary connective tissue, a classification that subsequent researchers would gradually revise.

This page matters for the larger arc of the essay because it shows Waldeyer keeping a second, related question consistently in view alongside the primary neuron question: if nerve cells are discrete units, what exactly is the material that surrounds, supports, and, as Golgi insisted, perhaps nourishes them? The glial question and the neuron question turn out to be genuinely intertwined throughout the essay, since Golgi’s argument for a non-nervous, nutritive role for dendrites depended specifically on linking dendrites to neuroglia and blood vessels (see Page 9). Understanding what neuroglia actually is, and is not, is therefore not a side issue but a necessary companion to resolving the central debate.

3. Neuroscience Commentary: 1891 Compared With Today

Virchow’s term “neuroglia” survives intact in modern neuroscience, though the understanding of what it names has been transformed almost beyond recognition. Nineteenth-century researchers debated whether glia were a form of connective tissue, an epithelial derivative, or something else entirely; modern neuroscience recognizes several molecularly and functionally distinct glial cell types, astrocytes, oligodendrocytes, microglia, and ependymal cells, each with highly specific roles: astrocytes regulate the extracellular ionic and metabolic environment and participate directly in synaptic signaling (the modern concept of the “tripartite synapse”); oligodendrocytes produce the myelin sheath essential for rapid axonal conduction; microglia serve as the resident immune cells of the central nervous system, actively surveilling and remodeling synaptic connections throughout life. The nineteenth-century intuition that glia might serve a “nutritive” function, dismissed as an overreach when Golgi tried to extend it to dendrites themselves, turns out to contain a genuine kernel of truth when correctly assigned to astrocytes and their role in metabolic support of neurons, one more example, as with Fritsch’s electric fish observations (Page 16), of a correct biological insight attached to the wrong anatomical structure.

4. Conceptual Reflection

The parallel, intertwined histories of the neuron question and the glia question in this essay foreshadow a genuinely important theme for the concluding sections of this critical edition: the neuron doctrine, however conceptually powerful, was never a complete theory of the nervous system by itself. From its very inception, the doctrine required, and generated, a companion account of the non-neuronal cells that make neuronal function possible. The twentieth and twenty-first centuries would go on to show that glia are not passive scaffolding but active, indispensable participants in nervous system signaling, development, and disease, a discovery that substantially complicates, without contradicting, Waldeyer’s neuron-centered synthesis, and to which this edition will return in its closing conceptual reflections.

1. Modern English Translation (of the 1891 German)

According to the now-available newer investigations of S. Ramón y Cajal, Kölliker, von Lenhossék, Oyarzun, and Retzius, we must now conceive the structure and arrangement of the glial framework as follows. One must distinguish: 1) the ependymal or neuroepithelial cells; 2) the superficial or peripheral glial cells; 3) the deep-lying glial cells. The ependymal cells draw out, from the ventricular wall, long branched processes radially through the entire thickness of the brain or spinal-cord wall. […] The superficial and deep-lying glial cells do not connect with the ventricular wall; according to S. Ramón y Cajal and Lenhossék they are said to be, for the most part, ependymal cells that have been displaced out of the epithelial arrangement. […] Rohde considers it noteworthy that, in invertebrates (leeches), two kinds of supporting substance must likewise be distinguished within the central nervous system, one of ectodermal origin, the same source as the nervous elements, Rohde calls these cells the “supporting cells.” The supporting cells hang together with the spongioplasm of the ganglion cells and nerve fibres.

2. Historical Commentary

This page completes Waldeyer’s survey of glial cell types by reporting the tripartite classification then emerging from Cajal, Kölliker, von Lenhossék, Oyarzun, and Retzius: ependymal (neuroepithelial) cells lining the ventricles, and two further categories of superficial and deep-lying glial cells distributed throughout brain and spinal cord tissue, which several researchers suspected to originate developmentally from displaced ependymal cells. He also brings in a comparative note from Rohde’s work on invertebrate (leech) nervous systems, describing similarly dual supporting-tissue populations there.

The overall picture assembled across Pages 27–28 is of a nervous system built from (at minimum) two broad cellular populations, signaling neurons and supporting glia, both arising, as the previous embryology page showed, from a shared ectodermal origin but diverging early into structurally and functionally distinct lineages. This sets up, cleanly, the culminating synthesis of the essay’s next several pages, in which Waldeyer will draw together everything covered so far, spinal cord, cerebellum, cortex, sensory periphery, embryology, and glia, into a single, tightly stated set of general conclusions.

3. Neuroscience Commentary: 1891 Compared With Today

The ependymal-cell-as-glial-precursor hypothesis reported here anticipates, in a rough and not entirely accurate way, aspects of modern developmental neurobiology’s understanding of neural stem cell niches: the ventricular and subventricular zones of the developing (and, to a more limited extent, adult) brain are indeed understood today as important sources of new glial and neuronal cells, arising from radial glial cells that share some morphological resemblance to the ependymal cells described here, though the detailed lineage relationships are now known to be considerably more complex than a simple “displaced ependymal cell” model would suggest. The three-way glial classification offered here (ependymal, superficial, deep) is a reasonable, if crude, early approximation of what would later resolve into the modern distinction between ependymal cells, astrocytes, oligodendrocytes, and microglia, though nineteenth-century researchers had no way to distinguish, for instance, astrocytes from oligodendrocytes by any means other than gross location and morphology, since the molecular and functional distinctions that separate these cell types were entirely inaccessible to 1891 technique.

4. Conceptual Reflection

Having now traced both the neuronal and the glial side of the nervous system’s cellular population from their shared embryonic origin through to their differentiated adult forms, Waldeyer has assembled essentially the complete cast of characters his summary will need. It is worth noting, looking ahead, that the neuron doctrine Waldeyer is about to formalize is explicitly a doctrine about neurons, it says nothing, by itself, about the functional significance of glia, leaving that question for future generations of researchers to take up. The parallel but separate treatment of neurons and glia in this essay is an early instance of a disciplinary division of labor, a focus on neurons as the primary explanatory unit of nervous system function, that would shape neuroscience for the following century, and that only began to be seriously revised once modern tools revealed just how actively involved glial cells are in signaling processes previously thought to be the exclusive domain of neurons.

1. Modern English Translation (of the 1891 German)

Summary and Physiological Remarks.

From the findings reported, the following principal propositions may be derived:

I. The axis cylinders of all nerve fibres (motor, secretory, sensible, and sensory; centrifugally or centripetally conducting) have proven to be, directly, outgoing from cells. A connection with a fibre-network, or an origin from such a network, does not occur.

II. All these nerve fibres end freely, with “end-trees” (Kölliker), without network- or anastomosis-formation.

[…] Setting aside, then, the fluctuating concepts of “origin” and “end,” all those authors who assume no net-like connection of the nerve fibres within the central organs (Ramón y Cajal, Kölliker, His, Nansen, Lenhossék, Retzius) may easily unite propositions I and II into one general law of far-reaching scope. This would read:

“The nervous system consists of numerous nerve units (neurons), which are neither anatomically nor genetically connected with one another. Each nerve unit is composed of three parts: the nerve cell, the nerve fibre, and the fibre-tree (end-tree). The physiological process of conduction can proceed both in the direction from the cell to the fibre-tree, and in the reverse direction. The motor conductions proceed only in the direction from the cell to the fibre-tree; the sensible ones now in the one, now in the other direction.”

2. Historical Commentary

This is the single most consequential page in the entire essay, and one of the most consequential single pages in the history of neuroscience. Waldeyer here does exactly what a synthesizer does at their best: he takes the accumulated, independently-obtained findings of Golgi (for the branching structure, even while rejecting Golgi’s own reticular conclusion), Ramón y Cajal (for the free-ending, non-anastomosing principle and the collateral discoveries), Kölliker (for the term “end-tree,” Endbäumchen, and for extensive independent confirmation), Wilhelm His (for the embryological growth-cone evidence and the term “dendrite”), Fridtjof Nansen, Mihály von Lenhossék, and Gustaf Retzius, and distills them into two crisp general propositions (I and II), then fuses those two propositions into a single unified law.

It is in stating that unified law that Waldeyer writes the sentence that would permanently change the vocabulary of biology: “Das Nervensystem besteht aus zahlreichen untereinander anatomisch wie genetisch nicht zusammenhängenden Nerveneinheiten (Neuronen)”, “The nervous system consists of numerous nerve units (neurons), which are neither anatomically nor genetically connected with one another.” This is the coining of the word “neuron” in its modern scientific sense. It is worth stating plainly, in line with the guiding principle of this entire critical edition: Waldeyer did not discover anything new on this page. Every empirical claim compressed into this paragraph belongs to the researchers he names explicitly. What Waldeyer contributes is the act of synthesis and naming, recognizing that propositions I and II, when combined, constitute a single, general, nameable law, and supplying that law with a term simple enough, and memorable enough, to be adopted permanently by the entire field.

The definition itself is precise and three-part: a “nerve unit” (Nerveneinheit), or neuron, consists of (1) the nerve cell (soma and dendrites), (2) the nerve fibre (axon), and (3) the terminal fibre-tree (axon terminal arborization), a definition that maps with remarkable fidelity onto the modern structural definition of a neuron. Waldeyer also states a physiological rule distinguishing motor and sensory conduction direction, showing that the anatomical synthesis (discreteness) is immediately paired with a functional consequence (directionality of signal flow).

3. Neuroscience Commentary: 1891 Compared With Today

Modern neuroscience fully validates Proposition I (axons arise directly from cells, not from a network) and Proposition II (nerve fibres end freely, via synaptic terminals, without anatomical fusion) as basic, foundational facts, confirmed beyond doubt by twentieth-century electron microscopy. The three-part structural definition of the neuron, cell body/dendrites, axon, and terminal arborization, remains essentially the standard structural description taught in every introductory neuroscience course today, more than 130 years later, a remarkable testament to how well this 1891 synthesis captured the essential architecture despite working entirely from light-microscope, silver-stained, fixed tissue. The directional rule for motor versus sensory conduction anticipates, in a structural idiom, what twentieth-century electrophysiology would establish mechanistically: that neurons are functionally polarized, typically receiving input via dendrites and transmitting output via the axon, though the modern understanding recognizes far more complexity and exception (dendro-dendritic synapses, retrograde signaling, axo-axonic synapses) than this 1891 statement could anticipate.

What the passage does not yet resolve, and could not resolve with 1891 technology, is the precise physical nature of the point of contact between one neuron’s terminal arborization and the next neuron’s receiving surface. Waldeyer’s law establishes that fibres end “freely,” but the term “synapse” itself would not be coined for another six years, by the British physiologist Charles Sherrington in 1897, specifically to name and characterize this point of functional but non-continuous contact. The neuron doctrine, in other words, established that neurons are discrete before science could fully characterize how discrete neurons communicate with one another.

4. Conceptual Reflection

This page deserves to be read slowly, because it is a near-perfect illustration of the distinction this entire critical edition has been building toward. Waldeyer’s genius here is not observational, he is not describing anything he personally discovered down a microscope. His genius is integrative and rhetorical: recognizing that a scattered, multi-national, multi-lingual, decades-long accumulation of partial findings had, without anyone quite noticing, reached a critical mass sufficient to support one clean, general, nameable law, and then having the clarity and confidence to state that law plainly, and the linguistic instinct to give it a name simple enough to survive.

This is precisely how many of the most consequential concepts in the history of science come into being: not as a single discovery by a single genius at a single moment, but as an act of conceptual crystallization, performed by someone positioned, by breadth of reading, institutional standing, and rhetorical skill, to see the shape of the whole field at once and name what the whole field, taken together, had already shown. The neuron doctrine belongs, in this specific and important sense, to dozens of researchers across Italy, Spain, Germany, Switzerland, Norway, and Sweden. The word “neuron” belongs to Wilhelm von Waldeyer alone.

1. Modern English Translation (of the 1891 German)

For the sensible nerves there exist, however, still certain difficulties, when one wishes to enter into a more precise determination of the concepts “origin” and “end” of a nerve. One can distinguish here three determining considerations: a) the physiological, b) the topographic-anatomical, c) the genetic. Physiologically, the origin of a nerve is that place from which the conduction process, taking place in the normal way, has its beginning; the end is that place where it, following this direction, ceases. The topographic, or descriptive, anatomy places the origin in the central organ, the end at the periphery. Genetically, the place is the origin from which a nerve fibre develops, the place where it remains at rest during its further development, its end. For motor and psychomotor fibres, everything is now simple and clear, since here all three considerations coincide. If we let what has been said hold good for the determination of the origin of the sensible fibres, then great differences appear here, for the sensory fibres seem physiologically to have all their peripheral points of origin, while the ordinary sensible fibres, for example those of the cornea and the other mucous membranes, of the senses of touch, temperature, and pressure, with their free fibre-endings, also end in the brain and spinal cord with such (end-trees), while at a certain point of their course they stand in connection with an interposed cell, the spinal ganglion cell, which behaves embryologically and anatomically like an origin-cell.

2. Historical Commentary

Immediately after stating his sweeping general law, Waldeyer pauses to acknowledge, with real intellectual honesty, that the law does not resolve every question cleanly, particularly for sensory pathways. He distinguishes three different senses in which a nerve fibre might be said to “begin” or “end”: the physiological sense (where does the signal-conduction process actually originate?), the topographic-anatomical sense (the traditional textbook convention of treating the central organ as origin and the periphery as endpoint), and the genetic sense (where does the fibre first grow from during embryonic development?). For motor fibres, he notes, all three senses agree perfectly, the cell body is unambiguously the origin in every sense. For sensory fibres, however, these three senses can point in different directions simultaneously: physiologically, signal conduction genuinely originates at the peripheral receptor (a touch receptor in the skin, for instance); but anatomically, and to a significant degree genetically, the spinal ganglion cell functions as an “origin-cell,” interposed partway along the pathway.

This is a genuinely sophisticated conceptual point, and it shows Waldeyer resisting the temptation to let his own elegant, unifying formula paper over a real complication in the underlying biology. The neuron doctrine, as stated, is structurally accurate, but applying ordinary directional language (“origin,” “end”) to it requires care, because sensory neurons in particular do not map neatly onto a simple, single-direction, cell-body-to-terminal picture.

3. Neuroscience Commentary: 1891 Compared With Today

This distinction anticipates a genuinely modern and still-relevant issue in sensory neuroscience: the pseudounipolar structure of the dorsal root ganglion sensory neuron, in which a single process leaves the cell body and then bifurcates into a peripheral branch (extending to the sensory receptor) and a central branch (extending into the spinal cord), with the cell body itself sitting off to the side of the main conduction pathway rather than at either functional “end” of it. Waldeyer’s careful three-way distinction between physiological, anatomical, and genetic “origin” is, in effect, an early conceptual grappling with exactly this structural peculiarity, a peculiarity that modern developmental neurobiology now explains through the specific migratory and growth behavior of neural crest-derived sensory neuron precursors during embryogenesis, tracing directly back to the growth-cone-driven, bidirectional fibre outgrowth Cajal and His had described (see Page 26).

4. Conceptual Reflection

This page is a valuable corrective to any tendency to treat great unifying scientific statements as though they resolve every subtlety at a stroke. Waldeyer’s neuron law is genuinely general and genuinely correct as an anatomical principle, but he immediately shows, in the very next breath, that correctly applying general principles to specific, biologically idiosyncratic cases (like the pseudounipolar sensory neuron) requires ongoing conceptual care. Good scientific synthesis, this page suggests, is not simply about achieving maximal generality, it is about achieving generality while remaining honest about where that generality strains against genuine biological complexity.

1. Modern English Translation (of the 1891 German)

If we now consider the pathways that lead from the seat of conscious sensation and conscious willing to the periphery, we find, so far as things are known at present, always several neurons, at least two, along this stretch. Only two, perhaps, are needed for the pyramidal pathway: one nerve unit extends, crossed, from a ganglion cell of the cerebral cortex to the end-tree in the grey matter of the spinal cord; the second, from a motor anterior-horn cell to the end-tree in a muscle fibre. Probably all sensible nerve conductions consist of more than two neurons. […] Setting aside the wavering concepts of “origin” and “end,” we can easily unite propositions I and II into one general law of far-reaching scope, provided we assume, as do all authors who accept no net-like connection of nerve fibres within the central organs, that no anastomosing nerve networks occur, but rather only a nerve-felt [Nervenfilz] or nerve-net [Neuropilema, His’s term]. If we take Golgi and B. Haller’s nerve-nets with us, the conception is modified somewhat, though we can still retain the concept of nerve units; the boundary between two nerve units would then always lie within a nerve-network, and would, at least with our present resources, not be exactly determinable.

Figure 9, reproduced from the original 1891 offprint (public domain).

2. Historical Commentary

Having stated the doctrine in its purest anatomical form on Page 29, Waldeyer now works out its physiological consequences, introducing the concept that would later become fundamental to circuit neuroscience: the chain of neurons. He observes that even the seemingly simple pyramidal (voluntary motor) pathway, running from the cerebral cortex to a muscle, requires at minimum two separate neurons in series, an upper motor neuron descending from cortex to spinal cord, and a lower motor neuron carrying the signal onward to the muscle, with sensory pathways typically requiring even more than two neurons in sequence. This is the conceptual birth of the multi-neuron reflex/relay pathway, a way of thinking about nervous function as a series of discrete, serially connected units rather than a continuous, undifferentiated conducting medium.

Notably, Waldeyer closes this page by explicitly, generously acknowledging the continuing minority position: even if one follows Golgi and Béla Haller in retaining some notion of a genuine nerve-network (Nervennetz), the neuron concept could still be partially preserved, with the boundary between adjacent neurons simply becoming harder to pin down precisely. This is a mark of real intellectual generosity, Waldeyer does not declare total victory over the reticularist position; he shows how his own synthesis could, in a modified form, coexist even with partial concessions to it.

3. Neuroscience Commentary: 1891 Compared With Today

The concept of the multi-neuron pathway or “chain of neurons,” introduced here for the pyramidal (corticospinal) tract and sensory pathways, remains one of the most basic organizing frameworks of modern systems neuroscience: sensory and motor pathways are still described, in every neuroscience textbook, as sequences of first-order, second-order, and third-order neurons relaying information through a series of discrete synaptic stations, precisely the picture Waldeyer sketches here in its earliest form. The upper motor neuron/lower motor neuron distinction for voluntary movement, in particular, remains directly clinically relevant today, forming the basis for localizing neurological lesions in modern medical practice. Waldeyer’s closing hedge about Golgi and Haller’s persistent network view would, of course, ultimately be resolved decisively against them by twentieth-century electron microscopy, but his willingness to accommodate their position, even provisionally, shows a scientist reasoning in good faith under genuine uncertainty rather than dogmatically insisting on total consensus.

4. Conceptual Reflection

The idea of the nervous system as built from chains of discrete relay units, rather than either a single continuous cable or an undifferentiated mesh, is arguably the single most consequential functional implication of the neuron doctrine, distinct from (though closely related to) its purely structural claim about cellular discreteness. Once the nervous system is understood as a series of separable, serially-connected units, entirely new kinds of questions become askable: how is information transformed at each relay point? Can a lesion at one link in the chain be localized and diagnosed independently of damage elsewhere? Can individual links be studied, manipulated, or even replaced independently? This relay-chain conception of neural function, quietly introduced on this page as a natural physiological consequence of Waldeyer’s anatomical law, would prove just as generative for the future of neuroscience and clinical neurology as the underlying structural claim itself.

1. Modern English Translation (of the 1891 German)

Aside from Nansen, no one, up to now, has ventured to shake the conception of ganglion cells as nervous functioning apparatus. He has done so with a bold hand, and not without justification, and regards the nerve cells only as nourishing organs of the entire nervous apparatus. According to him, the psychomotor and psychosensory functions would have their course solely in the nerve-felt of the cerebral cortex; in the spinal cord, the transmission of a motor excitation from a pyramidal fibre to a motor axis-cylinder would take place in the felt there, and be conveyed onward to the motor root fibre. Conversely, in a sensible centripetal conduction, this would spread in the peripheral reception station up to the spinal-cord felt, there be carried over onto a centripetal path of second order, and conveyed onward from this to the cerebral nerve-felt, where the conscious sensation would come about. The cells would thus not need to function immediately in conduction. […] Kölliker has expressed himself with decisiveness against this conception at the Munich Anatomical Congress.

Figure 10, reproduced from the original 1891 offprint (public domain).

2. Historical Commentary

This page shows Waldeyer reporting the most radical dissenting position in the entire essay, that of Fridtjof Nansen, already introduced as a comparative anatomist on Page 24. Nansen proposed that nerve cells might not function as active signaling elements at all, but rather as purely nutritive organs, with all actual signal conduction occurring instead within the diffuse nerve-felt (Nervenfilz) formed by overlapping fibre terminals. This is, in effect, an extreme version of Golgi’s earlier claim about dendrites (Page 9), now extended to the entire cell body itself. Nansen suggests that cells may simply feed the conducting apparatus rather than participate in conduction directly.

Waldeyer reports this position with evident respect (“with a bold hand, and not without justification”) even while noting that Kölliker had already publicly and forcefully rejected it at the Munich Anatomical Congress. This is another instance of the essay’s consistent intellectual honesty. Rather than dismissing an extreme minority view out of hand, Waldeyer takes the time to explain its logic clearly, credits its proponent’s boldness, and only then reports the counter-argument. This models exactly the kind of fair-minded treatment of dissenting views that a serious intellectual history requires.

3. Neuroscience Commentary: 1891 Compared With Today

Nansen’s proposal that nerve cells are “merely” nutritive organs, with all functional signaling occurring in a diffuse felt-like network, was decisively refuted by subsequent electrophysiology. The cell body (soma) and its associated machinery are now understood to be essential, active participants in neuronal signaling: generating and integrating electrical potentials, housing the nucleus and protein-synthesis machinery on which axonal and dendritic function ultimately depends, and, in many neuron types, directly initiating action potentials at the axon hillock adjacent to the soma. That said, Nansen’s underlying intuition, that a great deal of important neural computation happens locally, at synapses and in fine neuropil, rather than solely “in” the cell body as a kind of central command post, has a distant, much more sophisticated echo in modern findings about local dendritic computation and even local translation of proteins far from the soma. Nansen was wrong about cells being non-functional, but not entirely wrong that a great deal of the interesting computational action in a neuron happens away from the soma itself.

4. Conceptual Reflection

The Nansen episode is a useful reminder that even a fundamentally sound emerging consensus (the neuron doctrine) can and should tolerate serious internal challenge from committed, competent dissenters, and that recording such dissent honestly, rather than editing it out of the historical record, is part of what makes a scientific synthesis trustworthy. A synthesis that only reports the evidence supporting its own conclusion, while suppressing serious counter-proposals, would be propaganda rather than scholarship. Waldeyer’s essay, in giving real space to Nansen’s radical alternative before reporting its rebuttal, exemplifies the harder, more honest version of consensus-building.

1. Modern English Translation (of the 1891 German)

Kölliker first emphasizes that side branches (collaterals), as Nansen’s hypothesis requires, do indeed occur at the motor roots, but so rarely that one cannot take them seriously into account. Further, one must think of the olfactory fibres, which begin at the periphery with cells, as well as of the cases where the endings of nerve fibres are corbeille-like entwined with nerve cells (spiral fibres of the sympathetic ganglion cells), and of the endbaskets discovered by S. Ramón and Kölliker around the Purkyne cells. For Nansen’s view, the overwhelming majority of unipolar cells of invertebrates, which lie almost entirely outside the nerve-felt and in which no ensheathing fibre-baskets occur either, would seem to speak. Kölliker refers, however, to the fact emphasized by Gad, that every stimulus striking a motor fibre is propagated toward both sides, so that it reaches the motor cells as well; it is therefore not excluded that these participate in the coming about of reflexes. Before the hand, it seems to me, from these grounds, whose weight I recognize, premature to deny the cells participation in the nervous function, and thus I believe we must also leave the protoplasmic processes, which are, in a manner, only branchings of the body-substance of the cells, this same function.

2. Historical Commentary

Waldeyer now presents Kölliker’s detailed counter-argument to Nansen’s cell-nutrition hypothesis, and it is worth noting the argument’s structure carefully. Kölliker does not simply dismiss Nansen. He grants a genuinely difficult point, that the sparse population of unipolar invertebrate neurons lying outside any dense nerve-felt is a real challenge for the standard, cell-centered view of function, since these cells would seem to have little opportunity to participate meaningfully in a network-mediated signaling process. But Kölliker counters with a physiological argument from John Gad, showing that any electrical stimulus applied to a motor fibre propagates in both directions along that fibre, reaching the cell body as well as the periphery, which is at least consistent with (though it does not conclusively prove) the cell’s active functional participation.

Waldeyer’s own closing judgment on this page is stated in the first person and with careful, hedged language: he finds it “premature” to deny cellular participation in nervous function, “whose weight I recognize” regarding the countervailing evidence. This is Waldeyer at his most personally engaged, rendering his own considered verdict on a genuinely contested question rather than simply reporting others’ positions. He extends this judgment explicitly to the protoplasmic processes (dendrites) as well, tying this final physiological question back to the structural argument he had already made on Page 16 using Fritsch’s electric fish evidence.

3. Neuroscience Commentary: 1891 Compared With Today

Modern neurophysiology has settled this question decisively in the direction Waldeyer favored, though on grounds neither he nor Kölliker could have anticipated: intracellular recording, patch-clamp electrophysiology, and calcium imaging all demonstrate directly and unambiguously that the neuronal cell body is an active, essential participant in signal processing, not merely a nutritive support structure. Gad’s bidirectional propagation observation, cited here as suggestive but inconclusive evidence, corresponds to a real phenomenon still recognized today (antidromic conduction, in which an action potential can travel backward along an axon toward the soma under certain stimulation conditions), though its interpretation and significance are now understood in a vastly more detailed biophysical framework involving ion channels, membrane potentials, and active dendritic and somatic integration.

4. Conceptual Reflection

This page captures something valuable about how scientific judgment under uncertainty actually works in practice. Waldeyer does not claim certainty he does not possess. He states plainly that he finds the case against cellular function unproven, while explicitly acknowledging the genuine weight of the evidence on the other side. This is a model of intellectually honest synthesis: reaching a considered, stated conclusion while being transparent about the strength of the case for the alternative. It is worth noting that this is the closest the essay comes to Waldeyer rendering a truly personal scientific verdict, distinct from his role elsewhere as reporter and namer of others’ findings, and it is fitting that this personal verdict concerns the deepest and most consequential question the whole essay has been building toward: not just whether neurons are structurally discrete, but whether they are functionally alive as active participants in thought and sensation, rather than mere scaffolding for a diffuse conducting felt.

1. Modern English Translation (of the 1891 German)

Kölliker summarizes the chief results for the spinal cord in seven points: 1) The antero-lateral columns consist in part of fibres given off by nerve cells (tract cells) of all parts of the grey horns. 2) These stem fibres send collaterals into the grey substance, as do the nerve processes of the tract cells themselves, so far as they still run within the grey substance. 3) The posterior columns consist of the dividing, longitudinally bending posterior root fibres. 4) The motor roots come, without exception, from cells of the anterior horn, as their nerve processes. 5) The sensible roots come from the cells of the spinal ganglia. 6) Into the anterior commissure enter nerve processes of cells of the entire grey substance, which pass into the contralateral antero-lateral columns, as well as crossing collaterals of the antero-lateral columns and crossing dendrites of the medial anterior-horn cells. 7) Into the posterior commissure belong crossing collaterals of the sensible root fibres.

As for the physiological performances of the spinal cord: the voluntary movements come about through the pyramidal pathways, so that an impulse passes from a pyramidal cell of origin in the brain onto a pyramidal fibre, in this crosses to the contralateral side, and there springs over into a collateral, through whose end-tree it is transmitted onto a motor cell, and from this onto a motor fibre in a muscle fibre. The conscious sensations are certainly conducted from the periphery, through the posterior roots, into the posterior columns and, in these, further cerebralward. In the medulla oblongata the posterior-strand fibres must all first end, where and how (the nuclei gracilis and cuneati being here, since Deiters, most named) is still uncertain.

2. Historical Commentary

This page shows Waldeyer performing his final major act of synthesis for the spinal cord specifically: distilling Kölliker’s extensive and thorough investigations into a tight, numbered set of seven findings that, taken together, form a virtually complete wiring diagram of spinal cord cellular connectivity as understood at the very end of 1891. This numbered summary functions as a kind of proof of concept for the entire essay’s method: dense, scattered, technically demanding original research, compressed into a form clear and memorable enough for a physician or student to actually retain and apply.

He then pivots to physiological function, tracing the voluntary movement pathway (the pyramidal tract, already introduced on Page 31) in explicit step-by-step detail, from cortical origin cell, across the midline, through a collateral branch, to a spinal motor cell, and finally out to muscle. Notice how naturally the neuron doctrine, once established, generates this kind of precise, step-by-step circuit description; a reticular theory of diffuse network conduction could not support nearly this level of pathway-specific detail. He closes this page candidly acknowledging genuine remaining uncertainty (where exactly the posterior column fibres terminate in the medulla oblongata), once again modeling honest acknowledgment of the limits of 1891 knowledge even within a triumphant closing synthesis.

3. Neuroscience Commentary: 1891 Compared With Today

Kölliker’s seven-point summary, as reported here, corresponds closely to the modern textbook description of spinal cord grey and white matter organization, and the specific pyramidal tract pathway description (crossing at the pyramidal decussation, synapsing on anterior horn motor neurons) remains, in its essential outline, exactly how the corticospinal tract is taught in medical neuroanatomy today. The nuclei gracilis and cuneatus, whose precise termination pattern Waldeyer honestly flags as still uncertain in 1891, are now fully characterized as the first synaptic relay station for the dorsal column medial lemniscus pathway, carrying fine touch and proprioceptive information, a gap in 1891 knowledge that was closed within a few subsequent decades through further silver-stain and degeneration studies.

4. Conceptual Reflection

This page shows how quickly a newly consolidated general principle (the neuron doctrine) can generate enormous specific, practically useful detail once it is available as an organizing framework. Before the doctrine, a description like “the pyramidal tract carries voluntary motor commands” had to remain vague about mechanism, since nobody could specify exactly how the signal moved from brain to muscle. After the doctrine, that same description could be rendered as a precise, cell-by-cell, synapse-by-synapse account. This is a recurring pattern in the history of science: a good unifying framework does not just explain what was already known, it makes previously unaskable questions askable, and previously unanswerable questions tractable.

1. Modern English Translation (of the 1891 German)

Surveying the main gain that the discussed investigations have secured for us, it lies, in my opinion, above all in the achieved sharper delimitation of the anatomical and functional elements of the nervous system, which we must regard as nerve units (neurons), as well as in the discovery of the collaterals with their end-trees, through Golgi and S. Ramón y Cajal. These things make possible for us the understanding of isolated conduction, both as to whether such actually occurs, and as to the spreading of conductions over wide stretches. But, new facts, new tasks! This consoling and gratifying side of natural inquiry has become evident to us here too, as I already suggested in the introductory words, at every step. To the solution of these tasks, however, the physiologists and pathologists will be called just as much as the anatomists.

2. Historical Commentary

Waldeyer’s closing sentences are worth reading with the same care given to his opening ones. He returns explicitly to the framing established on Page 1, the fifty-year-old Henle quotation about the unsettled fate of ganglion cell processes, and states, in his own words, what he believes has genuinely been achieved: a “sharper delimitation of the anatomical and functional elements of the nervous system,” now nameable as neurons, together with the specific discovery of collaterals and their terminal end-trees. He credits this discovery explicitly, one final time, to Golgi and Ramón y Cajal by name, a fitting final gesture from an essay that has, throughout, been scrupulous about assigning credit to the actual discoverers rather than claiming the achievement for itself.

His final sentence, “new facts, new tasks,” strikes a note of open-ended optimism rather than triumphant closure. Waldeyer explicitly states that the questions raised by this new anatomical clarity will require the combined efforts of physiologists and pathologists, not anatomists alone, to resolve; the essay ends not by declaring the subject closed, but by handing the baton onward to the next generation and the next set of disciplines.

3. Neuroscience Commentary: 1891 Compared With Today

Waldeyer’s closing call for physiologists and pathologists to take up the tasks the new anatomy had opened was answered, over the following decades, in exactly the way he anticipated. Charles Sherrington’s physiological characterization of the synapse (a term he coined in 1897, just six years after this essay), Santiago Ramón y Cajal’s own continuing anatomical work culminating in his shared 1906 Nobel Prize with Golgi, and the eventual development of intracellular electrophysiology in the twentieth century all represent exactly the kind of cross-disciplinary follow-through Waldeyer was calling for. The pattern he describes, that new anatomical clarity generates new physiological and pathological questions rather than closing off inquiry, has proven durable and continues to describe the relationship between structural and functional neuroscience today, for instance in the way that modern connectomic mapping (structural) continually generates new questions for functional imaging and electrophysiology to answer.

4. Conceptual Reflection

It is fitting that Waldeyer’s essay, having opened with an admission of fifty years of unresolved uncertainty, closes not with a claim of final victory but with an honest acknowledgment that new certainty simply relocates the frontier of uncertainty rather than eliminating it. This is, in miniature, an accurate portrait of how scientific progress actually works: the neuron doctrine did not end inquiry into the nervous system, it reorganized that inquiry around a newly clarified unit of analysis, the neuron, and thereby made possible an entirely new generation of questions (about synapses, about circuits, about development, about disease) that could not even have been coherently framed before the doctrine existed. This closing note, humble and forward-looking rather than self-congratulatory, is perhaps the most enduring lesson this 1891 essay offers to any reader encountering it today.

[This concludes the page-by-page critical edition of Waldeyer’s 1891 essay. The edition closes with a final synthesizing chapter, Part III, addressing the broader arc from tissue to neuron to network to whole-body systems.]

PART III: Closing Synthesis

What Waldeyer Actually Did

It is worth restating, plainly, at the close of this edition, the distinction that has organized every page of commentary: Wilhelm von Waldeyer did not discover the neuron. He did not develop silver impregnation, he did not trace a single collateral branch through spinal grey matter, he did not observe a growth cone extending toward its target, and he did not settle, through his own experimental work, whether nerve fibres fuse or merely touch. Every one of those achievements belongs to Golgi, to Ramón y Cajal, to Kölliker, to His, to Retzius, to Fritsch, to Lenhossék, to Nansen, and to a wider circle of researchers scattered across Pavia, Barcelona, Würzburg, Stockholm, and Christiania.

What Waldeyer did was read all of it, in multiple languages, across a decade of scattered publication, and recognize that it added up to one thing: a general, statable law about the basic architecture of the nervous system. He stated that law in two crisp propositions, fused those propositions into one, and gave the resulting unit a name simple enough that a physician, a student, or a future Nobel laureate could carry it forward without effort. That act, synthesis, naming, and confident public statement in front of the Berlin Medical Society, is a real and distinct contribution to science, even though it added no new observation to the record. Fields advance through both kinds of labor: the labor of seeing something new, and the labor of recognizing what the accumulated new things, taken together, actually mean.

How Focusing on the Neuron Changed Biology

Before 1891, the nervous system could only be discussed in terms that mixed uncertainty about its most basic unit of organization into every higher-level claim. Was a reflex the product of a specific, traceable chain of cells, or the product of a diffuse, undifferentiated conducting medium reacting as a whole? Was a lesion in one part of the brain expected to produce a localized, specific deficit, or a general, non-specific disturbance spread throughout a fused network? These were not academic questions. They shaped how physicians reasoned about paralysis, about sensory loss, about the very possibility of localizing function within the brain at all.

The neuron doctrine settled the question decisively enough, even in 1891, to license an entirely new style of reasoning: reasoning by circuit. If the nervous system is built from discrete, individually traceable units connected in specific, patterned ways, then damage to a specific link in a specific chain should produce a specific, predictable deficit, and repair or compensation should, in principle, be analyzable link by link. This circuit-based style of reasoning became the foundation of clinical neurology as a rigorous diagnostic discipline, and it became the foundation of experimental neuroscience as it developed across the twentieth century, from Sherrington’s reflex physiology, to Hodgkin and Huxley’s ionic basis of the action potential, to modern optogenetic dissection of specific neural circuits governing specific behaviors.

Focusing on the neuron as the basic unit of analysis also enabled a research program of classification and cataloguing that continues, essentially unbroken, from Golgi’s Type I and Type II cells, through Cajal’s spinal cord taxonomy and cortical cell types, through twentieth-century electrophysiological cell-typing, to today’s transcriptomic atlases attempting to catalogue every distinct neuron type in the mammalian brain. None of that cataloguing project would have been conceptually possible under a reticular theory, in which the very idea of a discrete, individually classifiable “cell type” is somewhat incoherent, since cells are not truly separate entities to begin with.

What Focusing on the Neuron Potentially Hid

But every framework that illuminates something also casts a shadow, and the neuron doctrine is no exception. By making the individual neuron the primary unit of biological and clinical explanation for the following century, neuroscience risked, and in some periods genuinely did, underweight everything that does not fit neatly into the neuron-as-discrete-signaling-unit picture.

The clearest case, already visible in embryonic form within Waldeyer’s own essay, is glia. Waldeyer devotes real attention to neuroglia (Pages 27 and 28 of this edition), but treats it, following the consensus of his sources, primarily as supporting or nutritive tissue, a passive scaffold for the “real” business of neuronal signaling. This framing persisted for the better part of a century. Only in recent decades has neuroscience fully absorbed the fact that astrocytes actively participate in synaptic signaling, that microglia continuously remodel synaptic connections throughout life, and that oligodendrocyte myelination is itself an activity-dependent, plastic process that shapes circuit function. Glia are not a footnote to the neuron doctrine. They are a co-equal cellular population whose functional significance the neuron-centered framework, for structural and historical reasons, was slow to recognize.

A second, related blind spot concerns everything that communicates through means other than the fast, point-to-point, synaptic signaling that the neuron doctrine so successfully described. Hormonal signaling, carried through the bloodstream rather than along a traceable fibre, operates on entirely different spatial and temporal principles, diffuse rather than discrete, slow rather than instantaneous, and it interacts constantly and consequentially with neuronal circuits (the entire field of neuroendocrinology exists precisely at this intersection). Immune signaling within the nervous system, mediated by cytokines, microglia, and increasingly recognized gut-brain and body-brain immune axes, likewise operates through mechanisms that a purely neuron-centered picture does not naturally accommodate. And the metabolic coupling between neurons, astrocytes, and the vasculature, the neurovascular unit that supplies the enormous and constant energy demands of neural computation, is a whole-system property that no single neuron, considered in isolation, can explain.

The Arc This Edition Has Traced

This edition has followed a specific historical arc, visible across its thirty-five pages: from an undifferentiated, poorly understood tissue (Henle’s 1841 uncertainty, Page 1), to a clearly delineated, individually identifiable cell (the Golgi stain and Cajal’s discrete-cell interpretation, Pages 8 through 17), to a named unit with a precise structural definition and a stated general law (the neuron, Page 29), to a network of such units connected in specific, traceable, and increasingly well-characterized patterns (the collaterals, the reflex chains, the cortical and cerebellar circuit motifs, Pages 17 through 34).

The story does not end there, and Waldeyer himself, in his closing paragraph (Page 35), seemed to sense that it would not. The twentieth and twenty-first centuries have extended this arc one step further still: from network, to whole-body system. The modern understanding of the nervous system increasingly treats it not as an isolated, self-contained signaling apparatus but as one deeply integrated component of a larger physiological whole, continuously exchanging information and mutual influence with the endocrine system, the immune system, the gut microbiome, and the metabolic state of the entire organism. A modern account of, for instance, chronic stress, or depression, or neurodegenerative disease, routinely and necessarily reaches beyond the neuron and even beyond the synapse to encompass cortisol signaling, systemic inflammation, mitochondrial function, and the felt, lived experience of the person whose nervous system is being described, precisely the kind of whole-system, ecosystem-level thinking that treats neurons, hormones, glia, and lived experience not as separate domains but as interacting facets of a single, integrated biological and personal reality.

The Enduring Value of Waldeyer’s Synthesis

None of this diminishes what Waldeyer accomplished. Quite the opposite: it is precisely because the neuron doctrine so successfully and durably clarified the nervous system’s basic cellular architecture that later generations were able to discover, investigate, and eventually integrate everything the doctrine itself did not originally address. You cannot productively study how glia modulate synaptic transmission until you first know, with confidence, what a synapse is and what a neuron is. You cannot productively study the gut-brain axis until you have a clear enough picture of neural circuitry to identify exactly where and how a hormonal or immune signal is exerting its influence upon it. Waldeyer’s 1891 synthesis was not a final answer. It was, in the fullest and best sense, a foundation: sturdy enough to build on, precise enough to be extended, and honest enough, in its own carefully hedged closing pages, to acknowledge that it was raising as many new questions as it was answering.

That is, perhaps, the deepest lesson this document holds for anyone encountering it today, whether historian, working neuroscientist, educator, or simply a curious reader: the value of a great synthesis is not that it closes a subject, but that it opens it, cleanly enough, and durably enough, that the next 130 years of inquiry could be built directly on top of it.