Revisiting the neuron doctrine Catherine Stinson Abstract The neuron doctrine is the dominant theory of the structure and function of the nervous system. Standard histories of neuroscience celebrate Cajal, who first formulated this theory, and overlook Golgi, who resisted adopting it. A few historians have turned up evidence that challenges this story, suggesting that Golgi accepted the main claim that defines the neuron doctrine. The revised history says that Golgi and Cajal agreed on the facts, but differed in their theoretical preconceptions. I add further details to this revised history. First, I show that the standard history was largely manufactured by Cajal. Second, I argue that Golgi and Cajal agreed only on the fact of there not being anastomosis between cells, but not on anatomical independence. Third, I investigate the reasons why Golgi refused to accept anatomical independence, while Cajal did. I show how the particular brain areas Golgi and Cajal each took as their model system played a part in the construction of their theories. 1. Introduction By the middle of the 19th century, anatomists and physiologists had begun to use the microscope to explore the structure and function of the nervous system. Two issues had held back progress in this area, as compared to microscopic investigation of other cell types. First, nervous tissue has a peculiar structure, where branches from cell bodies travel long distances along complicated routes, interweaving with the branches from many other cells, as opposed to having a compact cell body easily visualized in a single microscopic slide, and easily distinguished from neighbouring cells. On top of the complicated structure of neurons1, two other factors conspired to make the staining 1 A typical neuron has three main structural features: the cell body, the axon, and multiple dendrites. The axon and dendrites are all thin protuberances from the cell body that can 1 techniques that worked well on other cell types ineffective in making the neural tissue visible under the microscope. These are the fact that the axons are covered in myelin, which resists staining by the agents known at the time; and the soft, delicate consistency of neurons, which requires hardening before they can be sliced and mounted on slides. These problems were overcome through the work of a number of researchers spread over most of Europe. By the turn of the century the basic structure of the neuron was understood, and had been synthesized into “the neuron doctrine.” In 1906 two researchers, Camillo Golgi and Santiago Ramón y Cajal, jointly received the Nobel Prize for having played essential roles in the development of this theory. It is generally believed that Golgi did not agree with the neuron doctrine, however. In his Nobel Prize lecture Golgi defended reticular theory, which had been the neuron doctrine’s main early competitor. Golgi's actions at the ceremonies, and his refusal to agree with Cajal's “correct” theory are seen as somewhat mysterious. The standard explanation offered is that Golgi was bitter that his rival, Cajal, was busily using Golgi's great discovery of the Golgi stain to disprove and correct his observations. This standard history is somewhat unsatisfactory, and a few historians of neuroscience have provided more nuanced accounts. Gordon M. Shepherd's 1991 book, Foundations of the Neuron Doctrine, provides a detailed history of the development of the neuron doctrine, including translations and excerpts of a number of the primary stretch for either short or long distances, and give off branches in various directions. The axon is structurally different from the dendrites in that it usually runs for a fair distance before giving off branches, and over this distance, it is covered by an insulating tissue called myelin. Axon collaterals also branch out at right angles from the axon, while dendrites branch at acute angles. There are usually many dendrites per cell, but typically a single axon. The basic function of dendrites is as input channels that communicate to the cell the activity of neighbouring cells with which they are in contact. The axon is the output channel that transmits signals from the cell to its neighbours downstream. 2 sources. The materials in this book reveal that the neuron doctrine, rather than being a well-defined theory, included different sets of claims according to different people, and at different times. There are also passages in Shepherd's book suggesting that Golgi may have been widely misinterpreted. Key claims that are attributed to Golgi are contradicted by what Golgi wrote. Shepherd concludes by pointing out ways in which Golgi's reticular theory has stood the test of time. Several Italian scholars have also dug deeper into this history, providing accounts more sympathetic to Golgi than the standard view. Paolo Mazzarello's 1996 biography of Golgi investigates what Golgi's scientific reasons for denying the neuron doctrine were. Giovanni Berlucchi and Guido Cimino's 1999 papers explore more details of the relationship between research in anatomy and physiology at the time, and Golgi's intellectual background and aims. The picture that emerges from these accounts is that whether Golgi disagreed with the neuron doctrine is a complex question, and that on those points where he disagreed, his reasons had more to do with theoretical interpretations than with facts. My goal here is to add further clarifications of the mystery of Golgi's reputed refusal to accept the neuron doctrine. I argue that the view that Golgi and Cajal agreed on the facts but disagreed on the theory is too simple. On the question of anatomical independence, it is not clear whether they agreed on the facts. I show that what Golgi meant by anatomical independence was something more than just the issue of whether axons anastomose. Although Golgi clearly held some false views about neural anatomy, Cajal was importantly wrong about the bigger picture in ways that Golgi was not. Cajal focused on local circuits, and thought that brain function could be understood on their basis alone. Golgi saw that looking only at local circuits couldn’t tell us much about brain 3 function, because of the massive scale of the brain’s complexity. I argue that although Cajal is now seen as one of the founding fathers of neurophysiology, he did not appear to be interested in questions of neural function early in his career. Finally, I argue that the emphasis on the theoretical differences between Golgi and Cajal overlooks an important part of the story. Differences in the evidence each of them considered as representative of brain anatomy also played a role in their disagreement. 2. Historical context The controversy between Golgi and Cajal sprang from disagreements over the fine details of their theories, which are best understood within the context of the neuroanatomy of their time. 2.1 Late 19th century developments in neuroanatomy Before Golgi began his studies in the 1860's, it was known that nervous tissue contained both nerve cell bodies as well as nerve fibres, but the relationship between the two structures was unknown. It was unclear whether the fibres arose from the cell bodies, or whether they were part of an independent organ. It was also unclear whether the fibres had definite endings, or whether they were all connected together into a continuous network. It was known that electrical signals travel down the fibres, and this was assumed to be the brain's mode of functioning. The dominant theory at the time was Joseph von 4 Gerlach's reticular theory, which claimed that dendrites anastomose2 together, forming a network (Shepherd 1991, p. 61). Albrecht von Kölliker supported this theory in his authoritative textbook, Handbuch der Gewebelehre des Menschen, saying that protoplasmic connections between dendrites are “the simplest solution to the enigma of these extremely numerous extensions, and also the best explanation for the large conduction capacity of the grey matter and the reflexes” (Shepherd 1991, p. 53). Thus cell theory, the claim that the elementary parts of all tissues are formed of cells, which Robert Hooke first suggested for plants in 1667, and Jan Purkinje and Theodor Schwann extended to animals in the late 1830's, had not yet been firmly established for nervous tissue. Golgi's most celebrated contribution to neuroscience was developing the silver nitrate stain (also called the Golgi stain), which he announced in 1873, and that proved to be an essential tool for making nervous structures visible under a microscope. In addition, he discovered that axons branch off into many collaterals, much like dendrites; demonstrated that dendrites do not anastomose; and distinguished between motor and sensory neurons. Golgi elucidated the structure and connections between numerous parts of the nervous system, and discovered a number of anatomical features, such as the Golgi apparatus, the function of which has only recently been discovered. Golgi supported a reticular theory in which axons alone perform neural functions, with dendrites playing only a nutritive role. He thought that the terminal branches of axon collaterals connect 2 Anastomosis is the fusing together of cells or tissues such that the protoplasm of the cells is continuous, or the tissues form a continuous body. This occurs in tubular structures like blood vessels and intestines. 5 together, forming a “diffuse nervous network”3 (Golgi, 1883, p. 306). Golgi's hypothesis that dendrites play a purely nutritive role turned out to be wrong, and he was criticized for suggesting that axon collaterals anastomose, although as we will see, he was willing to concede this point. Golgi's impressive stained specimens, and his preliminary investigations of various brain areas using his new staining method were the springboard for further research. August Forel's demonstration that nerve cells are independent nutritive units was done using the Golgi stain (Forel, 1991[1937]). Fridtjof Nansen's work establishing the nerve cell as an individual functional unit was also based on experiments using the Golgi stain, which Nansen learned on a visit to Golgi's lab (Nansen, 1991[1887]; Mazzarello, 1999, p. 155). These two studies, together with Wilhelm His's demonstration that nerve cells are individual embryological units (His, 1991[1886]) were the basic building blocks from which Cajal went on to establish the neuron doctrine. When Cajal first saw tissue stained with silver nitrate in 1888, he launched into an extraordinarily fruitful period of research. His key contributions were making improvements to Golgi's staining technique, and discovering the basket endings of stellate cells on the Purkinje cells of the cerebellum, which convinced him that axon collaterals make contact with dendrites, but without anastomosis--a proposal that overturned the popular reticular theories of the day. He went on to elucidate the structure and connections between numerous parts of the nervous system. Cajal brought his initial findings to Berlin in 1889 for the meeting of the German Anatomical Society, where he succeeded in impressing Kölliker, one of the giants of the field at the time. In 1890 3 I follow the translation of this phrase appearing in Shepherd's excerpt (p. 99). All other quotes from this paper I translate myself. 6 Kölliker tested Cajal's results and published a paper confirming what he called the His- Forel-Cajal theory of the individuality of the nerve cell (Shepherd, 1991, p. 177). The next year, another giant of the field, Wilhelm von Waldeyer, wrote a series of articles that reviewed these findings, and synthesized them into a new theory that became known as “the neuron doctrine” (Shepherd, 1991, p. 181). In the final paper of this series, he drew two conclusions: that axons originate directly from cells, and do not connect to or originate from a network of fibres (Waldeyer, 1991[1891], p. 181), and that “nerve fibers terminate freely with ‘end arborizations’... without a network or anastomotic formation” (Waldeyer, 1991[1891], p. 182). He united these two points into a law: “The nervous system consists of numerous nerve units (neurons), anatomically and genetically independent. Each nerve unit consists of three parts: the nerve cell, the nerve fiber and the fiber arborizations (terminal arborizations)” (1991[1891], p. 183). With both Kölliker and Waldeyer's support, Cajal's findings were established. Within about fifteen years of Golgi's invention of the silver nitrate stain, the basic anatomy of nerve cells had been discovered, a topological map of the nervous system had been more or less completed, and the manner of communication between neurons was beginning to be understood. Cajal, and Arthur van Gehuchten, added to this understanding the idea that conduction in nerves travels from dendrites to axon terminals, which Cajal called “the law of dynamic polarization” (Shepherd, 1991, p. 209), but a more complete understanding of interneuronal communication would have to wait until 1906 for Sherrington's explanation of his concept of the synapse. 2.2 Context within 19th century science 7 This work was happening at a time when the methods for studying nature and the human body were undergoing change. In the 1860s and 1870s, when Golgi was studying medicine and beginning his career, “positivistic materialism” was spreading in Italy through an influx of Dutch and German academics (Mazzarello, 1999, pp. 33). This approach involved the idea that biology could be reduced to chemical or physical principles, and brought along with it new methods in experimental physiology, microscopy and histology (Shepherd, 1991, pp. 30-32). Golgi's early influence in this direction was Cesare Lombroso, the director of the Psychiatric Clinic at the University of Pavia, whose approach to psychiatry was as a “positivistic science” (Shepherd, 1991, p. 82). Golgi later worked alongside Giulio Bizzozero, who was Professor of General Pathology, and in whose lab the new experimental approach to medicine was enthusiastically embraced by Bizzozero's many disciples, much to the annoyance of Pavia's old guard. At this time, the fields of psychology, medicine, and biology were not as neatly compartmentalized as they are today. Questions that would now belong to neurophysiology were not considered to belong to a separate field from questions that would now belong to clinical neurology. It was not unusual for the same people to deal with both. Psychology was on the verge of becoming a separate science. People moved easily between academic departments, often ending up where they did based on who they knew, and the political climate. For example, Golgi went from Chief Physician in a hospital, to Professor of Anatomy, then Professor of Histology, and finally Professor of General Pathology, with strings needing to be pulled at each stage (Mazzarello, 1999, pp. 101-105). Cajal had to navigate similar obstacles despite Spain's professorships being 8 based on an examination system. This fluidity can also be seen in how Cajal naturally switched gears to study bacteriology, since it fell, with histology, within microbiology (Cajal, 1937, pp. 283-291). The result was not only that careers often spanned several fields, but also that people could be working in what appeared to be the same field, and yet be asking quite different research questions. There were differing views at the time as to what the relationship between anatomy and physiology should be. Cimino remarks both on Claude Bernard's establishment of “a physiology that was autonomous and separate from anatomy,” and on the revival among anatomists of the idea that “the study of functions ought to depend upon... histological knowledge” (1999, p. 445). Which approach was taken varied across academic communities. It will become clear below that although Golgi and Cajal both performed histological studies of the brain, their approaches were quite different. 3. Controversy With this background in mind, we come to the controversy between Golgi and Cajal, which came to a head when they shared the Nobel Prize in 1906. That they were awarded the prize together suggested that they had contributed to the same body of work, but in his speech, Golgi challenged the neuron doctrine. The story most often told about this controversy is that Golgi stubbornly refused to admit that he'd been wrong about anastomosis between nerve cells, and that in his Nobel Prize lecture, he “made a display of pride and self-worship so immoderate that they produced a deplorable effect upon the assembly... Retzius was in consternation... all the Swedish neurologists and histologists looked at the speaker with stupefaction” (Cajal, 9 1937, p. 552). This description of the events, as recalled by Cajal some 30 years later, perhaps with a bit of bias, seems to have been swallowed whole by a number of commentators. In an account that quotes extensively from Cajal's autobiography, Hellman describes the event as “embarrassing in the extreme” (2001, 96). Brazier calls Golgi's speech an “unfortunate attack” (1988, p. 145). Katz-Sidlow refers to the “amazement of the audience” (1998, p. 239). Hellman speculates as to why Golgi behaved this way, “Was he just being pigheaded? Was he upset that Cajal, who had made his name on Golgi’s technique, had outshown [sic] him? Did he really continue to believe that he was right and Cajal was wrong?” (2001, p. 103). Hellman even goes so far as to suggest that Golgi's mistake of believing in a diffuse nerve network “held back the field’s progress for years” (2001, p. 91). Further aspersions cast on Golgi's work are the suggestion that he came across the silver nitrate stain by chance (Cajal, 1937, p. 306); that he altered his drawings, later adding dendritic spines in order to claim priority in this discovery over Cajal (Robinson 2001, 14); that he received the Nobel Prize merely for his discovery of the silver nitrate stain (Jones, 1999, p. 177); and that he had lost touch with advances in the field (Jones, 1999, p. 177). This story, however, is not particularly convincing. First, Cajal's description of Golgi's Nobel lecture is wildly unfair. The only source cited in support of this story by Hellman, Brazier and Katz-Sidlow is Cajal’s autobiography. There are other sources, however. Berlucchi cites a passage from Retzius's autobiography in which he gives quite a different description of his reaction to Golgi's behaviour at the Nobel Prize ceremony than that described by Cajal. Retzius thought that Golgi “behaved extremely nice [sic] and with dignity to Cajal” (Berlucchi, 1999, p. 199). Second, as Mazzarello argues, the 10 invention of the silver nitrate stain was not a chance occurrence, but the result of many years of work (1999, pp. 63-68). Furthermore, Golgi did not take credit for the discovery of dendritic spines, but rather noted that they had been seen early on4, and remarked that he still wasn't convinced that they had any neural function. Instead, there is ample evidence that Cajal did not always acknowledge the work of others. E.G. Jones points out Cajal's unseemly competitiveness and obsession with priority over the issue of nerve regeneration (1999, p. 176). Cajal has also been accused of not giving due credit to the other contributors to the law of dynamic polarization, including van Gehuchten, Sherrington, and William James (Shepherd, 1991, p. 209; Berlucchi, 1999, p. 197). As mentioned earlier, Golgi's contribution did not consist solely in the invention of the silver nitrate stain. Golgi is referred to as “the pioneer of modern research into the nervous system” in the Nobel Prize presentation speech (Mörner, 1967[1906], ¶13). That Golgi supported a reticular theory as late as 1906 may not indicate that he had lost touch with current research. Reticular theory was in fact seeing a resurgence in popularity at the time (Shepherd, 1991, p. 226; Anderson and Anderson, 1992, p. 182; Cimino, 1999, p. 471), so his claim that the neuron doctrine was “going out of favour” (Golgi, 1967[1906], p. 189) was not entirely untrue. Finally, the most serious mistake attributed to Golgi, that he insisted on anastomosis, is also a misrepresentation, as will be discussed below. 4. Facts of observation 4 Mazzarello describes how dendritic spines caused controversy in 1888, when Rossbach and Sehrwald published a criticism of the black reaction, in which one of their complaints was that small drop-like processes appeared on the dendrites of Golgi stained Purkinje cells, suggesting to them that lymphatic vessels and not nerve cells were being stained (1999, p. 190). 11 More serious scholarship on the subject has yielded more plausible explanations for Golgi's obduracy. Shepherd, Mazzarello, and Cimino5 all agree on one important point: the mistake attributed to Golgi of remaining fixed on the idea of anastomosing nerve fibres is not quite true. 4.1 The question of anastomosis Shepherd mentions that, “over the years there had been hints that Golgi might be softening his belief in continuity between nerve terminals” (1991, p. 260). Cimino makes the point more forcefully, claiming that while “Golgi and Cajal appeared to disagree over the ‘factual’ level, in reality their dispute arose more from the ‘ideal’ level” (1999, p. 466). Cimino explains that the one fact to which Cajal clung most tenaciously was that the prolongations make contact by contiguity but never continuity (Cimino, 1999, p. 460). Golgi's “incontrovertible fact of observation” was the existence of a diffuse nervous network. Cimino stresses, however, that Golgi's fact did not stand in opposition to Cajal's. Cimino cites as evidence Golgi's paper from 1885, in which Golgi suggests anastomosis only as a probability, but also admits “the possibility (and this was before the advent of the neuron theory) that the filaments were only intertwined” (Cimino, 1999, p. 449; see also Mazzarello, 1999, p. 214). 5 The reference lists in the Italian works suggest that there is much more scholarship available only in Italian that the English-speaking world has not taken notice of. This is ironic given that the seed for Golgi and Cajal's rivalry came about because Golgi and Cajal were, in the early days, only publishing in Italian and Spanish, so the wider scientific community didn't have access to their work. After Golgi bristled at Cajal claiming to have discovered axon collaterals in the spinal cord several years after Golgi had announced this discovery, Cajal complained that he should not be expected to know the contents of “a regional medical review unknown to the scientific world” (Cajal 1937, p. 379). 12 There is ample evidence beyond what Cimino cites showing that Golgi's idea of a network did not require anastomosis. In a paper from 1883, Golgi says that it is probable that axon collaterals anastomose, but that nothing positive can be said on the subject (pp. 297-298). He provides Cajal-like illustrations of individual cells, showing dendritic trees, cell body, and axon with branching collaterals, as shown in Figure 1. Golgi states more explicitly in 1891 that, “axons could overlap without actually forming anastomoses” (Mazzarello, 1999, p. 214). Kölliker confirmed this as having always been Golgi's position, but Waldeyer took it as evidence that Golgi had “changed his opinion” (1991[1891], p. 187). In his Nobel Prize lecture, Golgi again affirms his acceptance of transmission by indirect contact, “I have stated that there was no reason to think that direct contact between fibrils of different origin was an indisputable condition for the transmission of the impulse from one to the other. On the contrary, I thought that these contacts were more than sufficient for an impulse to be transmitted in any direction” (1967[1906], p. 202). 13 Figure 1. Golgi (1883) drawing of a Purkinje cell. Golgi’s unwillingness to categorically state that there are no direct connections is more circumspect than Cajal’s insistence that the terminals do not anastomose, but this was the more correct statement to make at the time. The microscopic evidence available to them could not decide the question between direct and indirect contact, because the gap junction is much smaller than can be detected with a light microscope. Nevertheless, it is clear that Golgi was never a strong adherent of anastomosis, so Cimino's claim that Golgi and Cajal did not disagree on the facts holds up, at least with regard to this fact. However, Golgi and Cajal took their disagreement as having to do with the facts; Cajal never recognized Golgi's acquiescence on the question of anastomosis, and Golgi 14 repeatedly claimed that he could not accept the neuron doctrine. Cimino's elaboration of the differences between Golgi and Cajal's theoretical preconceptions, discussed below, might explain why they interpreted the facts differently, but it does not fully explain why they misdiagnosed the nature of their disagreement. To sort this out, we’ll look in more detail at which facts the neuron doctrine encompassed. 4.2 The neuron doctrine The neuron doctrine as it is understood today states that “the neuron is the anatomical, physiological, genetic and metabolic unit of the nervous system” (DeFelipe, 2002, p. 483), but in 1906, the physiological picture was still quite unclear, and actively debated. Neither Golgi nor Cajal went into much detail about what sort of connections they thought were made between neurons, beyond espousing relationships of close contiguity over protoplasmic connections. Cajal offers the vague suggestion that nervous conduction might be passed by "induction or influence from a distance” (1967[1906], p. 221). Jones complains that developments were occurring in physiology that “could have substantially contributed to [the neuron doctrine] and to its acceptance as the fundamental basis of neural function” (1999, p. 170). Berlucchi notes that, “neither Golgi nor Cajal paid much heed to Sherrington’s findings and to neurophysiological studies in general” (1999, p. 191). The neuron doctrine did not acquire its physiological aspect until later. The 1896 edition of Kölliker's textbook stated that “the question of the nerve cell as independent anatomical unit was settled” (Shepherd, 1991, 222) but makes no mention of the nerve cell as physiological unit. Jones remarks that, “the Neuron Doctrine was completely 15 anatomical in its conception” (1999, p. 170). Despite this, Golgi lists in his Nobel Prize lecture “The neuron is a physiological unit” as one of the “ideas on which the neuron theory is based,” and one of his targets for attack (1967[1906], p. 190). Much of Golgi's disagreement then, was with a claim that was not yet an established part of the neuron doctrine. This can be seen in two ways: as Golgi misunderstanding the nature of the theory, or as Golgi looking ahead to the physiological implications. Golgi does not claim to disagree only with the idea of the neuron as a physiological unit, however. He also launches attacks against the idea of the neuron as an independent anatomical unit. This seems strange given that he was willing to give up anastomosis. Others took anatomical independence and “free terminations” to be synonymous. There are many examples of this. In 1889 Cajal describes as “the facts” that, “nerve cells are independent elements which never are anastomosed” (p. 158). In 1891 Cajal says that, “Retzius has recently proved the cellular independence and the free termination of axonal arborizations in the ganglia of crustacea” (p. 24). If the claim that there are no anastomoses, and the claim that neurons are anatomically independent are taken to be the same fact, then it is not clear whether Golgi agreed to the facts. Cajal and Golgi agreed on the observable data, insofar as they had both looked at slides of neural tissue stained with silver nitrate, and neither of them saw evidence of anastomosis.6 6 It is debatable whether not seeing anastomoses should count as observable data. Two instruments were needed in order to make these observations, neither of which was theoretically unproblematic. First, light microscopy does not have the resolution to distinguish gaps as small as synaptic clefts, as mentioned already. Synaptic clefts were not really "seen" until the 1950's with the advent of electron microscopes. Second, it was unclear whether the silver nitrate stain affected the entirety of nerve cells, and nothing beyond. It is still unknown today how this stain works. 16 Whether a cell is independent may involve more than just the fact of having clear boundaries. Golgi took anatomical independence to mean something beyond the fact of no anastomosis. For him independence had to do with the density of connections between cells, and not just the nature of these connections. In his 1883 article he says, The peripheral fibres, far from each being in individual, isolated contact with one central cell, on the contrary are linked in great numbers, to vast groups of cells-- and inversely every cell is in contact with a large number of fibres having different destinations and functions. As a result, the fundamental disposition of central elements is a tendency to effect the most widespread and complicated communications, and not restrained, isolated contacts. (p. 306) He echoes this point in his Nobel Prize lecture, “I have always stressed, that nerve cells, instead of working individually, act together, so that we must think that several groups of elements exercise a cumulative effect on the peripheral organs through whole bundles of fibers” (1967[1906], p. 216). As Cimino describes it, “for Golgi the diffuse nervous network constituted the main anatomo-physiological structure of the brain, and this ran counter to the neuron theory, which considered the nerve cell the basic anatomo- physiological element” (1999, p. 464). 5. Guiding ideals Golgi and Cajal agreed on the fact of anastomosis, despite appearances to the contrary, but, as I have shown, nevertheless disagreed on the fact of anatomical independence. Cimino explains their disagreement as a difference in guiding ideals. He 17 describes Golgi's ideals, but says very little about Cajal's. Below, I first review Golgi's ideals, then provide more details about Cajal's. 5.1 Golgi Golgi's guiding ideal was a “holistic” concept, according to Cimino (1999, p. 467). Cimino traces Golgi's concern with the function of the brain as a whole back to his academic roots. As mentioned earlier, Golgi began his academic career working with the psychiatrist Lombroso. In his graduation thesis of 1865, Golgi discussed the causes of mental disorders and the necessity of classifying them according to anatomical and aetiological factors (Shepherd, 1991, p, 82). Golgi drew away from psychiatry not because of a lack of interest in the subject, but because he sought more rigorous experimental methods. In Bizzozero's Laboratory of Experimental Pathology, Golgi acquired his “passion for histological research as the most direct method for penetrating the mysterious nervous system” (Mazzarello, 1999, p. 38). He felt that neurohistology was more fact-driven than psychiatry, and therefore “potentially capable of providing an understanding of the causes of mental disease” (Cimino, 1999, p. 443). Golgi was not concerned with the “’facts’ as an end in itself, but only as the path to ‘deduce’ their physiological and physiopathological significance with the greatest degree of certainty” (Mazzarello, 1999, p. 45). This view may stem from Golgi's anatomy professor, Bartolomeo Panizza, who taught that, “Physiology and pathology are reducible to anatomy plus some hypotheses” (Mazzarello, 1999, p. 20). In 1880 Golgi wrote that the chief task of Anatomy is: 18 “placing itself in the position to be able to answer the most pressing questions asked by Physiology” (Berlucchi, 1999, p. 192). Golgi's focus on extracting information about function from studies of morphology is evident in his earliest papers. In the brief paper announcing his first results with the newly invented stain, Golgi makes room to comment on ways in which the facts presented are”probably very important for the physiology of the cerebral organ” (1991[1873], p. 87). In his first major paper reviewing his results, Golgi begins by stating that his purpose is to determine whether ganglion cells have any characteristic properties that might reveal their “nervous nature” (1883, p. 285). In addition to describing many morphological details, Golgi discusses the physiological implications of these details at length. To give a rough idea of how much he discussed function, I counted 9 mentions of “physiologie,” and 24 mentions of “fonction,” in this paper of approximately 11,000 words. A similar analysis of Cajal will be provided below for comparison. 5.2 Cajal Cajal, in contrast, did not approach anatomy as a tool for solving physiological, and psychiatric puzzles, but as an end in itself. Cajal's first passion in the medical field was anatomy, and his early career shows little evidence of other interests. His autobiography describes how he was an apathetic student prone to wild behaviour, which was finally tamed only after his father moved to Zaragosa, where Cajal was studying medicine, and became professor of dissection (1937, p. 169). Cajal relates how they spent long hours together at the dissection table, working through manuals of anatomy, and how his father finally came to appreciate Cajal's artistic skill through his anatomical 19 sketches (p. 170). After Cajal left the army, he began to study microscopy, competed unsuccessfully for a professorship in anatomy having impressed the judges only with his knowledge of descriptive anatomy (p. 255), then learned to read German so that he might improve his knowledge of anatomy and histology, succeeded in becoming Director of Anatomical Museums in Zaragosa (p. 259), then finally got the chair of Anatomy in Valencia (p. 260). Despite this history, Shepherd claims that, “from the very start of his work on nerve cells... Cajal was obsessed with understanding how they functioned” (1991, p. 197). The only evidence Shepherd cites is that starting in his early papers, Cajal contested Golgi's idea that dendrites perform only nutritive functions. Cajal also encourages the view of himself as a budding physiologist in his autobiography. He claims that after his first experiments with the Golgi stain in 1888, “the new truth... rose up suddenly in my mind like a revelation” (1937, p. 322). The new truth he refers to is the law of dynamic polarization, which he calls the “physiological corollaries” of his anatomical findings. It is hard to say whether this claim is true, but it is certain that the autobiography includes other egotistical flourishes. One piece of evidence suggesting that Cajal reinvented his younger self as a budding physiologist later in life comes from some comments he made about his prospects for getting the Nobel Prize. Cajal seemed quite sure at the time that his work was not physiological in nature. In a letter dated 1904, Cajal explains to a colleague who wanted advice on how to nominate him for the Nobel Prize the reasons why past nominations had not been successful. He writes, “This nomination has previously been made by other Spanish and foreign Universities with little success, due no doubt to the 20 following reason: the Prize is given for medical (or rather, pathological) and physiological works whereas my own work ... relates to anatomy and histology, which sciences do not seem to enter into the competition” (Corral et al., 1998, 43). The Royal Caroline Institute seems to have changed that habit and awarded the 1906 Prize for work in anatomy. The first words of the speech presenting Golgi and Cajal the Nobel Prize were, “This year's Nobel Prize for Physiology or Medicine is presented for work accomplished in the field of anatomy. It has been awarded to Professors Camillo Golgi of Pavia and Ramón y Cajal of Madrid in recognition of their work on the anatomy of the nervous system” (Mörner, 1967[1906]). The idea of Cajal as a budding physiologist is also contradicted by the evidence from his early papers, in which he seems to fit much more neatly into the role of anatomist and histologist. He talks endlessly about layers, thickness, shape, terminations of collaterals, and directions of branching. He spills very little ink discussing the implications for physiology of his morphological findings, in contrast to Golgi who talks constantly about how structure relates to function. In DeFelipe and Jones's (1988) collection of Cajal's papers, the three earliest ones together mention physiology once and function not at all. In a paper from 1889, Cajal describes physiology as imposing prejudices (p. 157) and says, “it is time to separate from histology all physiological obligations” (p. 158). Cajal's first major article reviewing his work provides an apt comparison to that of Golgi discussed above. The article identifies as its topic the “general connections of nerve cells” (1988[1891], p. 23). This paper contains 1 mention of physiology, 4 mentions of function, and 3 mentions of purpose in its approximately 21 22,000 words (compared to 33 overall in Golgi's paper of half that length). All of this evidence suggests that Cajal's interest in physiology was not yet well developed. When Cajal does discuss nerve function in later papers, his focus is on single cells, and his descriptions of pathways suggest very sparse connectivity between cells. Cimino explains that, “what Cajal meant by <<functional individuality>> was that every neuron carried out its own function autonomously, and that all the activities of the nervous system are the sum of the single neuronal actions” (1999, p. 454). Describing the relationship between Purkinje cells and basket cells, Cajal says, “each element is an absolutely autonomous physiological canton” (1991[1888], p. 147). Figures 2, 3, and 4 show the progression of his depictions of this relationship. In Figure 2, he shows the descending baskets at the same level as the Purkinje cell bodies, although their relationship is unclear. In Figures 3 and 4, he shows two basket cells each making contact with several Purkinje cells. Figure 2. Cajal (1888) drawing of chick cerebellum (A=Purkinje cell, S=basket endings). 22 Figure 3. Cajal (1894b) drawing of the cerebellum (a=Purkinje cell, d=basket ending, b=basket cells). Figure 4. Cajal (1894a) drawing of Purkinje cells with baskets. 23 It is not clear whether Golgi interpreted Cajal literally as claiming that neurons are sparsely interconnected, or whether he could not conceive of how complex mental phenomena could arise from the summing together of these simple functional units. Either way, Golgi had good reason for disagreeing with this view of neural function, as more recent work in neuroscience has shown that mental phenomena often arise from the dynamic activity of whole populations of neurons (Shepherd, 1991, Ch. 20). Others working around the time read Cajal in similar ways. Lewellys Barker thought it pertinent to point out that, “Nerve conduction paths may, and probably usually do, in higher animals at least, involve more than one neurone, the neurones being, as it were, superimposed upon one another to make simple or more complex neurone chains or chains of neurone groups, one individual neurone through its various processes being in a position to be affected by and in turn to affect several or many other neurons” (1968[1899], pp. 41-42). In the 1918 edition of Gray's Anatomy of the Human Body, the cerebellum is depicted with only one stellate cell providing basket endings over Purkinje cells. The caption indicates, “Diagrammatic, after Cajal and Kölliker” (fig. 706). Cajal first appears interested in matters of neural function in his Croonian Lecture of 1894, which Kölliker had advised him to give a “physiological slant” (Shepherd, 1991, 239). Cajal's description of the terminal arborizations of basket cells adds a new emphasis, “one should also know that this connection is not individual; it is collective; that is to say, every pericellular plexus contains ramifications coming from several stellate cells” (1991[1894], p. 252). The illustration accompanying this description, shown in Figure 4, still shows only two basket cells connecting to each Purkinje cell. Cajal sounds almost like Golgi for the first time in this lecture, when he mentions “the 24 increasing diffusion of the currents as they reach the most central organs” (p. 250). By 1906, his Nobel Prize speech shows much more interest in function where he says, “the recurrent collaterals serve to associate in a dynamic ensemble the neurons of the same kind from the same area of the grey matter” (p. 229). It is not until his 1911 book that he shows denser interconnections diagrammatically by adding to his picture of the cerebellum significantly more collaterals descending to form each basket, as shown in Figure 5. Figure 5. Cajal (1911) drawing of the cerebellum (A=Purkinje cell, B=basket ending, G=axons of basket cells, with collaterals descending). This summary of Cajal's theoretical predispositions demonstrates that, despite Cajal's reminiscences to the contrary, he began with quite different aims than Golgi, and came to be concerned with physiology and how multiple neurons simultaneously act on 25 one another quite a long time after the period in dispute. Cajal thought that individual connections between small numbers of neurons were the basis on which brain function depended. Golgi understood that these individual connections could not do the job, and that to understand brain function, you had to take into account the massive scale of interconnectedness of large numbers of neurons working in concert. This is the point on which Golgi disagreed with what he understood as the neuron doctrine, and rightly so. 6. The importance of evidence While it is true that Golgi and Cajal had different guiding ideals, Cimino's thesis that these differences were responsible for Golgi and Cajal's theoretical choices leaves out part of the story. Golgi and Cajal each had a favourite piece of evidence that they examined early on, and returned to repeatedly in their writing. These served for each of them as representative models of the key point at the center of their theory. While their theoretical convictions may have conditioned them to see different phenomena as manifest in these bits of brain, the evidence each of them depended on also differed in important ways. An example is provided by Golgi's holding fast to the conviction that dendrites play only a nutritive role, and the corollary that dendrites and axons do not communicate. From his earliest papers he observes in the hippocampus that, “dendrites traverse the grey matter of the fascia dentata, ending on its surface, which is covered with blood vessels and connective tissue” (1883, p. 294). He reasons from this that the dendrites must serve a nutritive function, since there are no axons with which to make contact to be seen in the 26 area. Kölliker rejected this conclusion, not for theoretical reasons, but because he had different data available to him; he confirmed the existence of axons in the area using a different staining method (1991[1887], p. 168). It is also the hippocampus that provides Golgi's key observations of the diffuse nerve network. As Jones points out, the hippocampus is a “densely intertwined mass of axons” (1999, p. 173) that functions as a single entity with “nerve cells operating in collective mass action and not as individual elements” (p. 174). These characteristics of the hippocampus are what impressed upon Golgi the idea that the “reticular structure, whether a network or a simple web, was an indisputable fact and not a hypothesis” (Cimino, 1999, p. 450). Golgi's interpretations are based on the example he had before him under the microscope. In contrast, Cajal's favourite example was the cerebellum, in which he first saw the axons of basket cells enveloping Purkinje cell bodies (1991[1888], p. 143). For Cajal, this was the key piece of evidence demonstrating that cellular communication passes from axon collateral to dendrite and soma through contact, but not continuity. Jones suggests that Cajal's choice of brain area was crucial, “it is likely that Cajal’s early success lay in his choice of highly laminated structures, such as the cerebellar cortex and retina, as the objects of his investigations” (1999, p. 173). Shepherd further notes that these basket endings on the Purkinje cells are “one of the most specialized type of terminal in the entire nervous system” (1991, p. 150). The cerebellum and hippocampus differ in structure in ways that are important for the neuron doctrine. In the cerebellum, “A single basket cell may give rise to as many as ten baskets, to as many Purkinje cells” (Shepherd, 1979, p. 220). These connections are 27 much more sparse than in the hippocampus, where “one basket cell... establishes contact with 200-500 pyramidal cells... This is more than an order of magnitude greater than the comparable estimate for the basket cells in the cerebellum” (Shepherd 1979, 315). If Golgi and Cajal each took their preferred area as representative of the entire brain, it is not surprising that the one emphasized the complexity of neural architecture, and the other emphasized its orderliness. Neither area is entirely representative. The cerebellum is particularly idiosyncratic in terms of the types of intercellular connections represented. The cerebellum has a “limited stock of synaptic types... in contrast to the greater variety of connections in many other parts of the brain” (Shepherd 1979, p. 226). Specifically, it possesses almost exclusively connections from axon to soma or dendrite. Cajal's law of dynamic polarization decreed that this was the only direction in which neural communication could flow. Much later it was discovered that connections from axon to axon, and dendrite to dendrite also exist, as well as the anastomotic connections that Cajal so strongly denied (Shepherd, 1991, Ch. 20). That Cajal took so much guidance from what he saw in the cerebellum allowed him to develop a simplified view of neural communication, which proved to be quite fruitful as an early guide to research, despite the existence of exceptions. 7. Conclusion Standard histories of neuroscience celebrate Cajal, and overlook the contributions of Golgi, who is seen as having clung too long to his reticular theory in the face of Cajal's findings. As we saw here, a few historians have begun to remedy this misapprehension by pointing out that Golgi did not insist that cells anastomose, and that his holistic view of 28 neural function was more forward-looking than Cajal's doctrine of neurons as individual physiological units. The thesis that Golgi and Cajal agreed on the facts, but differed in terms of theoretical preconceptions was seen to be basically correct, but in need of refinement in three ways. First, Golgi and Cajal agreed only on the fact of there not being anastomosis between cells, but not on whether this implied anatomical independence, nor on the physiological implications of sparse interconnection. Second, the influence of Cajal's academic background on his theorizing needed elaboration. Cajal did not appear to be interested in questions of function early in his career. Third, the emphasis on differences in theoretical predispositions overlooks another important factor. Golgi and Cajal did not just look at the same evidence through the lens of different theories. They also looked at different evidence, and this evidence played a part in the construction of their theories. 29 References Anderson, C.G. & Anderson, B. (1993). Koelliker on Cajal: Translated excerpts from Erinnerungen aus meinem Leben. International journal of neuroscience, 70, 181-192. Barker, L. (1968). The nervous system and its constituent neurons. In E. Clarke & C. D. O'Malley (Eds.), The human brain and spinal cord: A historical study illustrated by writings from antiquity to the twentieth century (pp. 40-42). San Francisco: Norman Publishing. (First published 1899) Berlucchi, G. (1999). Some aspects of the history of the law of dynamic polarization of the neuron: From William James to Sherrington, from Cajal and van Gehuchten to Golgi. Journal of the history of the neurosciences, 8(2), 191-201. Brazier, M.A.B. (1988). A history of neurophysiology in the 19th century. New York: Raven Press. Cimino, G. (1999). Reticular theory versus neuron theory in the work of Camillo Golgi. Physis, 36(2), 431-472. Corral, I. C., Corral, C. C., & Castanedo, A. C. (1998). Cajal's views on the Nobel prize for physiology and medicine (October 1904). Journal of the history of the neurosciences, 7(1), 43-49. 30 DeFelipe, J. (2002). Sesquicentenary of the birthday of Santiago Ramón y Cajal, the father of modern neuroscience. Trends in neurosciences, 25(9), 481-484. DeFelipe, J. and E.G. Jones (1988). Cajal on the cerebral cortex: An annotated translation of the complete writings. New York: Oxford University Press. Downes, S.M. (1992). The importance of models in theorizing: A deflationary semantic view. PSA: Proceedings of the biennial meeting of the Philosophy of Science Association, 1, 142-153. Finger, S. (2000). Minds behind the brain: A history of the pioneers and their discoveries. Oxford: Oxford University Press. Forel, A. (1991) Out of my life and work. In G.M. Shepherd (Ed.), Foundations of the neuron doctrine (pp. 115-116). New York: Oxford University Press. (First published 1937) Golgi, C. (1991). On the structure of the gray matter of the brain. In G.M. Shepherd (Ed.), Foundations of the neuron doctrine (pp. 84-88). New York: Oxford University Press. (First published 1873) 31 ------- (1883). Recherches sur l'histologie des centres nerveux. Archives italiennes de biologie, 3, 285-317. ------- (1991). On the fine structure of the central organs of the nervous system. In G.M. Shepherd (Ed.), Foundations of the neuron doctrine (pp. 96-100). New York: Oxford University Press. (First published 1886) ------- (1967). The neuron doctrine – theory and facts. Nobel lectures, physiology or medicine 1901-1921 (pp. 189-217). Amsterdam: Elsevier Publishing Company. (Presented in 1906) Gray, H. (1918). Anatomy of the human body (20th ed. revised and edited by Warren H. Lewis). Philadelphia: Lea & Febiger. http://www.bartleby.com/107 (Accessed 8 September 2007) Hellman, H. (2001). Great feuds in medicine: Ten of the liveliest disputes ever. New York: John Wiley & Sons. His, W. (1991). On the structure of the human spinal cord and nerve roots. In G.M. Shepherd (Ed.), Foundations of the neuron doctrine (pp. 106-110). New York: Oxford University Press. (First published 1886) 32 Jones, E.G. (1999). Golgi, Cajal and the neuron doctrine. Journal of the history of the neurosciences, 8(2), 170-178. Katz-Sidlow, R.J. (1998). The formulation of the neuron doctrine. Archives of neurology, 55, 237-240. Kölliker (1991). Golgi's studies of the finer structure of the central nervous system. In G.M. Shepherd (Ed.), Foundations of the neuron doctrine (pp. 167-169). New York: Oxford University Press. (First published 1887) Mazzarello, P. (1999). The hidden structure: A scientific biography of Camillo Golgi (H.A. Buchtel and A. Badiani, Trans.). New York: Oxford University Press. Mörner, K.A.H. (1967). Presentation speech. Nobel lectures, physiology or medicine 1901-1921 (pp. 220-253). Amsterdam: Elsevier Publishing Company. (Presented in 1906). http://nobelprize.org/nobel_prizes/medicine/laureates/1906/press.html (Accessed 18 September 2007) Nansen, F. (1991). The structure and combination of the histological elements of the central nervous system. In G.M. Shepherd (Ed.), Foundations of the neuron doctrine (pp. 120-124). New York: Oxford University Press. (First published 1887) 33 Ramón y Cajal, S. (1991). Structure of the nervous system of birds. In G.M. Shepherd (Ed.), Foundations of the neuron doctrine (pp. 141-148). New York: Oxford University Press. (First published 1888) ------- (1991). General connections of the nervous elements. In G.M. Shepherd (Ed.), Foundations of the neuron doctrine (pp. 157-162). New York: Oxford University Press. (First published 1889) ------- (1988) On the structure of the cerebral cortex of certain mammals. In J. DeFelipe and E.G. Jones (Eds.), Cajal on the cerebral cortex: An annotated translation of the complete writings (pp. 23-54). New York: Oxford University Press. (First published 1891) ------- (1991). The fine structure of the nervous centers (Croonian lecture). In G.M. Shepherd (Ed.), Foundations of the neuron doctrine (pp. 239-253). New York: Oxford University Press. (First published 1894a) ------- (1990). New ideas on the structure of the nervous system in man and vertebrates. Neely Swanson and Larry W. Swanson (Trans.). Cambridge, MA: The MIT Press. (First published 1894b) ------- (1955). Studies on the Cerebral Cortex (Limbic Structures). (Lisbeth M. Kraft Trans.). Chicago: The Year Book Publishers, Inc. (First published in 1901-2) 34 ------- (1967). The structure and connexions of neurons. Nobel lectures, physiology or medicine 1901-1921 (pp. 220-253). Amsterdam: Elsevier Publishing Company. (Presented in 1906) ------- (1995). Histology of the nervous system. Oxford: Oxford University Press. (First published 1911) ------- (1937). Recollections of my life (E.H. Craigie, Trans.). MIT Press. Robinson, J.D. (2001). Mechanisms of synaptic transmission: Bridging the gaps (1890- 1990). Oxford University Press. Shepherd, G.M. (1979). The synaptic organization of the brain. New York: Oxford University Press. Shepherd, G.M. (1991). Foundations of the neuron doctrine. New York: Oxford University Press. Waldeyer, W. (1991). A review of new research on the anatomy of the central nervous system . In G.M. Shepherd (Ed.), Foundations of the neuron doctrine (pp. 181-193). New York: Oxford University Press. (First published 1891) 35