ABSTRACT
‘Mauthner’s cells’ are a pair of very large neurones found in the hind-brain in all main groups of fishes and amphibia, except in sharks and rays. Mauthner cells always have some dendrites in synaptic relation with incoming VUIth-nerve fibres and have a large axon which crosses the brain and descends the opposite side of the spinal cord, co-ordinating somatic motor activity. Papers discussing the presence of these cells in the lampreys are reviewed: two different pairs in the lamprey of ‘Miiller’s cells’ (giant co-ordinating cells with homolateral axons) have been considered homologous with Mauthner cells, but recent textbooks accept neither homology.
A pair of very large neurones, having the Mauthnerian characteristics, is described in the embryonic and early larval stages of Lampetra planeri and L. fluviatilis. The neurones are illustrated by figures in different planes. They are not one of the pairs of Millier neurones described by previous workers.
From this and similar evidence it is suggested that in all non-amniote vertebrates the earliest swimming movements under the control of the brain are effected through variations from one homologous co-ordinating system, of which the Mauthner cells form part.
Introduction
An earlier generation of zoologists sometimes grouped together the more 2A. primitive, aquatic vertebrates as the ‘ichthyopsids’. The term covers lampreys, cartilaginous and bony fishes, the tadpole stage of amphibians, and adults of aquatic tailed amphibians. These animals have certain significant characters in common ; in particular, they possess a lateral-line sensory system and their locomotion is achieved typically by serial contractions of trunk and tail myotomes.
The organization of neurones and of somatic muscle which brings about the swimming movements of ichthyopsid animals may therefore be homologous in them all. If this is so, a close comparison between different members of the group is justifiable, and should give valuable information on several aspects of the working of the nervous system.
On the other hand, it may be that the developmental sequence by which the swimming mechanism is achieved has followed different paths. Even within an homologous morphological system of myotomes, peripheral nerves, and central tracts, the neurones forming the units of this system may have combined into functional patterns which, although they all provide for swimming movements as part of the behaviour pattern, have evolved independently in different groups, e.g. in the different vertebrate classes.
It is therefore necessary to see whether there exists a sound basis for a detailed comparison of neurones and sensorimotor arcs among the ichthyopsids.
The simplest living example of the ichthyopsids is the newly hatched or ‘pro-ammocoete’ larva of the lamprey. The lampreys, although of cephalaspid rather than pteraspid affinities, have probably diverged little from the structure of the agnathan ancestor of the jawed vertebrates. The organization of the cranial and spinal nerves and of the gross morphology and tract-systems of the lamprey brain appears to be primitive among vertebrates, with few specializations: this has been shown especially by the work of Johnston (1902, 1905, 1912). The pro-ammocoete is without some specialized features of the ammocoete and adult stages: its brain is not yet compressed from in front by the backward growth of the naso-hypophysial tube, the eyes and eye-muscles are not degenerate, and the spinal cord is cylindrical, and not ribbon-shaped as in the ammocoete or adult. .
The neuromuscular organization of the trunk of the pro-ammocoete is comparable with that of higher ichthyopsids such as the larvae of urodele amphibians on which Coghill made his now classical researches on the relation of b ehaviour with neuromuscular anatomy. The neurones which could effect a contralateral response develop very early, as they do in urodeles. These neurones are : (a) Rohon-Beard somatic-sensory cells (these resemble dorsal-ganglion cells but the cell-body lies within the cord); (Z>) ‘primary’ somatic-motor neurones; (c) large internuncial neurones providing a contralateral linking of the sensorimotor pattern. Types (a) and (6) are strikingly similar to the corresponding types described by Coghill. I have described this organization in the pro-ammocoete previously (Whiting, 1948). The physiology of the neuromuscular system also seems to be similar to that of higher ichthyopsids (compare Harris, 1955).
It seems probable that a contralateral sensori-motor arc, developing first at a post-otic level and then rapidly at successive cephalo-caudal levels, is the first functional system of neurones in all ichthyopsids. The facts, on which this view is based, are given elsewhere (Whiting, 1955). Now, in the lamprey embryo, as cells of the types (a), (6), and (c) develop at more caudal levels of the cord, so there descend to these levels the growth-cones of axons from the first coordinating cranial neurones (Harris and Whiting, unpublished). These neurones, when functional, will bring the movements of the somatic muscle under the control of the brain ; they are known as the giant Müller neurones : the cell-bodies, of which there are several pairs, lie in the mid- and hind-brain floor and none of the axons cross the mid-line as they descend the spinal cord. In higher ichthyopsids, on the other hand, growth down the cord of axons of coordinating neurones occurs relatively later, and one pair of axons is particularly prominent. These arise from a pair of giant Mauthner neurones.
After comparison of the earliest sensori-motor arcs ((a), (b), (c)) of the trunk of the lamprey with those of higher vertebrates, it is important to establish whether the brain of the lamprey then co-ordinates spinal cord activity in the same way as in higher animals. The Mauthner neurones provide an important element in this second step.
The Mauthner pair has many distinguishing features: the cell-body lies at the level of the otic-capsule, certain constant relations of a few prominent dendrites include a direct synaptic connexion with incoming fibres of the VHIth nerve, the axon, of outstanding diameter, crosses the floor of the hindbrain to descend the length of the cord in synaptic relation with the contralateral primary somatic-motor neurones. Intensive study, by Beccari, Bartel-mez and Hoerr, Detwiler, Stefanelli, Bodian, and Leghissa among many others, has revealed important structural and developmental characters of the Mauthner neurones. This information has helped our understanding of nervous function, and has shown, I think, that the Mauthner apparatus is a part of a specific co- ordinating mechanism, much as a well-defined pattern of key indicates the existence of a specific lock. This specificity is indicated also by the existence of many other pairs of co- ordinating neurones which, while less striking, are equally constant in position (compare Bartelmez, 1915, on Amiurus, and Stefanelli and Camposano, 1946, on Anguilla).
Functionally, these cells have been shown to play a part in the development of effective swimming movements in the larval stages (Detwiler, 1933). The Mauthner cell is important in the movements of the tail by which the animal maintains equilibrium: the connexions of one of its dendrites with the vestibular division of the VHIth nerve suggests this, as does the absence of the cell in those fish in which the tail is lost (Mola) or does not assist in equilibrium (some bottom-living fish). Control of the tail cannot be its only function, since it is in synaptic relation with motor neurones at all levels of the cord. More detailed information on the functions and connexions of the Mauthner cell is given by Kappers and others (1936), Beccari (1943), Piatt (1948), Leghissa (1946, 1947), Stefanelli (1951), Bodian (1952), and Cordier (1954). Cordier’s figures (pp. 250 and 327) give the general relations of the cell to the rest of the brain in an adult fish or amphibian.
Mauthner cells have been described in the holocephalan Chimaera, in the lungfish Neoceratodus, in the sturgeons Acipenser and Polyodon (Johnston, 1901, Hoogenboom, 1929), in the great majority ofthe teleost fishes which have been examined, in the larval stages of urodele, anuran, and apodan Amphibia, and in adult aquatic urodeles. They are found in a reduced form in some adult frogs. They have not been found in elasmobranchs, whether sharks or rays. The most complete survey of their distribution is given by Beccari (1943), who, however, omits the sturgeons from his list.
It is therefore important to know whether Mauthner neurones, in addition to Müller cells, exist in lampreys. For the presence here of Mauthner cells would then suggest that the whole mechanism by which the brain co- ordinates the activity of the spinal cord is homologous in all ichthyopsids : their absence from elasmobranchs would be secondary. It would become justifiable to attempt a detailed comparison not only of the behaviour but of the sensorimotor arcs and the cell-types found in the development of all ichthyopsids up to the relatively late stage at which swimming locomotion under the control of the brain is achieved.
The studies of neurones of the central nervous system of the lamprey, which have been made over the past 80 years, might be expected to have settled this question a long time ago, but in fact the results have been very conflicting. The present account provides evidence that the pro-ammocoete stage of the lamprey does possess true Mauthner neurones.
Previous Work
Ahlborn (1883), surprisingly detailed work for the period, found several pairs of Müller neurones and one pair of Mauthner neurones in young ammo-coetes of 15-20 mm length. His plan-view diagram shows the Mauthner cell-bodies at the level of the VHIth nerve and the Mauthner axons forming a chiasma and then descending the spinal cord on the side opposite to the cell of origin.
Since then there have been many accounts of the co- ordinating apparatus in the brain of the lampreys ; fortunately, descriptions of the Müller cell part of it agree upon the facts. In older ammocoetes and in adults, the organization of the Muller cells has been found to have a well-defined pattern of the form shown in plan view in fig. 1, which follows Stefanelli (1934). The pairs of Müller cells 9 and 10 in this figure are the important ones for the present problem. The neurone K is not part of the known co- ordinating system but is a hypothetical cell which will be discussed later. The Müller cells are shown in lateral view in the book by Kappers and his colleagues (1936, fig. 316), where cells 9 and 10 are the upper two at position ‘B’.
The shape and position of each Müller cell-body is constant within a species, and readily comparable between species; the position is either along the viscero-motor column, or, more medially, on the somatic-motor column. These results have been established by the work of Johnston (1902), Tretjakoff (1909), Saito (1928), Stefanelli (1934), Pearson (1936), Woodburne (1936), and Barnard (1936). It is agreed that each Müller axon runs along the spinal cord on the same side as the cell-body from which it comes. Tretjakoff has mentioned that some neurones of the medial group have processes, apparently axon-collaterals, which cross the hind-brain floor to form a commissure, the chiasma fibrarum Milllerianunr, but Johnston (1910) reaffirmed that no Müller axons decussate in the hind-brain.
A typical Mauthner cell with a contralateral axon, as described by Ahlborn, was not found again until Whiting (1955) briefly stated that a homologue of the Mauthner neurone exists in the lamprey embryo and showed, in a diagram, the relative positions in the spinal cord of the Mauthner and Müller fibres.
Other recent work on the nervous system of the head of the lamprey (Leghissa, 1942; Larsell, 1947; Heier, 1948; Lindstrom, 1949) does not give any new information about the Mauthner or Müller cells.
Material And Methods
Early motile stages of Lampetra planeri and L. fluviatilis were used. No differences were found between the two species. Specimens were taken from eggs which had been fertilized on the nest or in the laboratory. Serial sections of many embryos and larvae were prepared ; over 20 specimens were carefully studied in working out the connexions described below.
The stages of development were determined by the criteria given by Damas (1944). Observations were made on animals of 3-5 to 8-0 mm in length, i.e. motile stages of the embryo, stages 7 to 11 of Damas ; newly hatched larvae, stages 12 to 14; and an immediately following stage, larvae of 6-5 to 8-o mm in length, which may be termed stage 14a since it is much younger than Damas’s stage 15, The 8-mm larvae still have well-developed eyes, a cylindrical spinal cord, and a naso-hypophysis opening at the front of the head (not dorsally) ; in the free state they would not yet have burrowed into the bed of the stream : they have not changed from ‘pro-ammocoetes’ into prides (‘ammo-coete’ larvae).
Impregnation of nervous structure was effected by Bodian’s protargol or Holmes’s silver nitrate techniques used after Nonidez’s fixative and Lang’s method for dehydration (Whiting, 1948; Harris and Whiting, 1954). It was difficult to get good impregnation of thin sections but these were easier to study; the most useful were 8 μ, in thickness. Non-nervous detail is shown well in embryos fixed in Susa, dehydrated by Lang’s method and stained with Mallory’s phosphotungstic acid haematoxylin, but the histology of muscle- and nerve-cells is shown in good protargol preparations.
Preparations of the brain of adult Lampetra were compared with the account by Johnston (1902). His 19 figures of transverse sections of Golgi preparations were photographed and set up as a three-dimensional model, with the functional components traced out in threads of different colour.
Final stages of the work were done with camera lucida drawings made with a Zeiss-Winkel drawing apparatus, which is accurate over its whole field : in this apparatus the relative brightness of the drawing surface is varied by means of two rotatable polarization filters. Nerve-processes can be traced from the lowest optical plane of one section to the highest plane of the succeeding section by this means, if the points of reference for the figure are well chosen. This procedure can be repeated through many sections. Drawings of important series of sections were made on sheets of Kodatrace, and fibres were followed on these.
Personal Observations
The organization is best seen at about stage 14. Where convenient, earlier stages are referred to, as well as the time at which cells, once identified, can be first found.
In transverse sections of stages 14 and 14a, a pair of very large fibres was observed to cross the floor of the brain at a level slightly posterior to the otic capsules, in a marked and symmetrical chiasma. The fibres were difficult to follow forward, in sections cut in the usual planes, because they turned obliquely upwards and outwards. A series of sections cut parallel to the long axis of the animal, but at about 350 to its vertical axis, proved to be in the plane of orientation of the cell-body and of the anterior part of the axon, and is illustrated in figs. 2 and 3.
Fig. 2 shows the organization of the head at this time ; the dimensions have been confirmed by comparison with larvae of the same age, cut sagittally. The arrangement of nerves and visceral arches illustrates a primitive vertebrate condition in very beautiful fashion. The profundus ganglion and nerve run anteriorly forward, dorsal to the optic nerve and eye, while the maxillomandibular nerve runs ventral to those structures. The facial nerve runs downward along the hyoid arch, entirely behind the first gill-pouch, which corresponds to the spiracular pouch of fishes. The first pair of gill-pouches contains the pair of muscular velar folds which can be clearly seen in rhythmical pumping action in the living pro-ammocoete ; these gill-pouches do not form definitive gill-openings, as do the succeeding seven pairs. Further details of the general morphology will be found in the figure of Damas (1944, plate III, 14).
The main result may be seen from fig. 2. The cells of origin of the two fibres lie above the posterior third of the otic capsule. The axons run caudally and medially, crossing the midline at the level of the glossopharyngeal ganglia. After the chiasma so formed, each of the two, now contralateral, fibres runs among and parallel with the ipsilateral Müller fibres (not shown) in the medial longitudinal bundle of that side, as far as the level of the fourth gill-pouch. At this point two fibres, apparently the same pair of contralateral ones, swing out of the closely packed bundle into a more dorsal and lateral position in the spinal cord and then continue caudally again, parallel with, but now distinct from, the Müller fibres. Other series of sections confirm that it is the contralateral pair of fibres which swing away from the Müller fibre bundle in this fashion. The pair of cells which have this relationship are considered to be the homologues of Mauthner cells, and will be given that name for convenience in further reference.
Further detail of the Mauthner cell is shown in fig. 3. Two large cells with characteristic dendrites lie farther forward in the brain (cell A in the midbrain and cell B at the front of the hind-brain) : these cells and the Mauthner cell are seen in both figures and show the relation between them. The outlines of the brain and notochord, and the background shown within them, were drawn from a single optical level of one section of the series used in fig. 2 : the neurones shown in detail, including the left A, B, and Mauthner cells, and the right Mauthner axon, were drawn from the whole thickness of the section. Further detail of the most ventral part of the brain was added from the adj a-cent section on the medial side. The outlines of brain and notochord were very similar in the two sections, except at the tip of the notochord, so that alignment was not difficult. Cell-walls of part of the notochord were drawn from the second section in stipple: the stippled lines coincide, as they should if the sections are correctly aligned, with continuous lines drawn from the same cellwalls seen in the first, more lateral section.
The left Mauthner axon runs caudally and ventro-medially from its cellbody, in the first section; in the next medial one, it crosses the right Mauthner fibre in the prominent chiasma seen in the figure; continuing caudally and to the right of the mid-line, it enters a’third section (not drawn). From the chiasma in the second section, the right Mauthner fibre enters the first section and joins the bundle of left-side Müller fibres, as shown.
The Mauthner cell-body lies dorsal to the nearer Müller cells, one of which it straddles. From the cell-body, two axon-collaterals run ventro-medially, becoming very slender as they approach the mid-line ; they pass very close to the outer side of this Müller cell. There are few dendrites on the Mauthner neurone, compared to the prominent array running into the white matter from a Müller neurone: one main dendrite runs dorsally, as shown, to the point of entry of the VUIth nerve.
The figure also shows that the Mauthner neurone is, at this period, of about the same size as the Müller neurones, but the oblique angle of its cell-body makes it inconspicuous in sections cut in the usual planes. This can also be seen from fig. 6, A, which is of a different larva, and at a higher magnification.
Fig. 4 shows the Mauthner cells of a stage 11 embryo projected, from five horizontal sections of 16 /x, on to the lowest: this passes through the middle of the otic capsules, the most anterior trunk-myotomes, and all but the tip of the notochord. The facial ganglion is still small and is ventral to the level of the section. The fourth (first post-otic) myotome is divided into a part medial and a part lateral to the otic capsule : the lateral part is divided into an upper (shown) and a lower half: divisions no doubt due to the relative expansion of the otic capsule. The contracting units are horizontal muscle-plates containing myofibrils ; separate muscle-fibres cannot be seen (compare Brachet, *935)-
The figure shows both Mauthner cells lying obliquely, lateral to the nearest Müller cells; the Mauthner axons crossing the mid-line at the level of the glossopharyngeal ganglion, and then running caudally among ipsilateral Müller fibres; and the position of the larger Müller cells in this stage of embryo (projected from all five sections). The Mauthner axons could not be traced farther caudally, because they are very close to the Müller fibres, and are at this time of the same calibre as these fibres. Most of the Müller axons are omitted, for clarity.
Only the anterior axon-collateral is shown in the figure. On one side, this process appears to be traceable into the opposite ventral motor column at about the level of entry of the maxillo-mandibular nerve ; however, this collateral is slender and is close to an arcuate fibre crossing the mid-line. (Higher magnifications could not be used because their field did not include suitable reference-positions.)
The Mauthner dendrites are above the level of the otic capsules, but the capsule ends dorsally at this stage in a vertical endolymphatic duct; this enabled the two upper sections to be aligned with those below. The main dendrite continues the line of orientation of the cell-body, running dorsally, laterally, and anteriorly towards the point of entry of the VIIIth nerve. Horizontal sections of older (stage 14) material show that this thick dendrite sends out short branches from its lateral aspect: most of these are directed laterally, but a few point longitudinally; these are probably in contact with a column of Vth nerve fibres, as explained below.
This figure also shows that the Millier cells are taking up the positions and the symmetry described by Stefanelli and others (compare fig. 1, p. 166).
In fig. 5, the Mauthner neurones have been drawn from transverse sections, cut at 8 p, of a stage 11 embryo. The section on to which the neurones are projected passes through the outer parts of the fourth myotome on both sides, through the front end of the inner part of this myotome on the left, and just anterior to the inner part on the right ; this is the level of the upper part of the Mauthner cell-body and the base of the main dendrite.
Mauthner dendrites extend, anterior to the section drawn, through one section. The chiasma of the Mauthner axons lies in the same section as the glossopharyngeal ganglion, eight sections posteriorly to that drawn. Below the Mauthner cell-bodies, the position of the Müller cells nearest to them is shown.
This figure confirms relations shown in figs. 2-4. The obliquely sited cellbody is dorsal to Müller cells lying along the somatic-motor and visceromotor columns. The large main dendrite ascends dorsally to the position of entry of the VUIth-nerve fibres, where its thickness makes it prominent in transverse sections: here three or four smaller dendritic branches extend laterally. These branches do not pass beyond the contours of the brain; no synapses are yet observable on them, but they are of course in close contiguity with incoming fibres running parallel with them. The animal is too young for all the separate, very complex, divisions of the lamprey’s acoustico-lateral system to be safely distinguished (compare Kappers and others, 1936, p. 43943), but these dendrites are probably contiguous with the vestibular division of the VUIth nerve.
The axon and its branches have the course described for the preceding series. The lower branch of the axon-collateral ends rather abruptly, as shown: the upper branch continues as a slender process, towards the mid-line ; its full extent could not be shown in the figure. The axon continues caudally beyond the chiasma, as a fibre of constant and prominent thickness.
Examination of pro-ammocoetes of stage 14, cut in orthodox planes, has given further information. In a transverse series, the Mauthner axons have been followed from the cell-body across the chiasma into the midlateral position in the cord, as far as the level of the anus, where this series stops. The graduation in calibre between the different longitudinal fibres is greater by this stage : tracing of individual fibres for long distances becomes practicable. When the Mauthner fibres have reached the mid-lateral position in the cord, in the innermost part of the white matter, they are easily distinguished from all the smaller longitudinal fibres near them (fig. 6, E). I have already illustrated them, and described their course in the spinal cord, without being aware of their identity (Whiting, 1948, figs. 10 and 11 and p. 374).
Parasagittal sections of a stage 14 animal (fig. 6, B) have also been drawn by camera lucida. This series confirms that the Mauthner fibres in the floor of the hind-brain are continuous caudally with the mid-lateral fibres in the spinal cord, across the gap shown in fig. 2.
The course of a Mauthner fibre along the spinal cord will bring it into contact with the dendrites of successive somatic-motor neurones down the trunk (fig. 6, c, D). There appear to be two types of somatic-motor neurones, primary and secondary : the primary motor neurones only reach the position of the Mauthner fibres through the dorsal end of the main dendrite, while the dendrites of the secondary system are much closer to them (Whiting 1948, *955)-
The Müller fibres, lying ventro-medially in the cord, will effect their coordinating function differently, either through funicular cells in the ventral part of the cord or directly upon the primary somatic motor neurones by way of the ventro-medial process of these neurones. This process is probably an axon-collateral rather than a dendrite; also, it often extends into çlose connexion with the contralateral motor column,. It is, in any case, separate from the main dendrite at this time and apparently remains so (compare Kappers and others, 1936, p. 153 and fig. 67, B).
The earliest identification of Mauthner neurones is in a stage 7 embryo, where the Mauthner neurones and one pair of Müller neurones can be seen. The cell-body is high on the side of the brain, the chiasma is visible, and the axon is beginning its descent of the spinal cord in a mid-lateral position. However, the Müller axons are also in a relatively lateral position at this time.
The Mauthner neurones are extremely constant in their relation to the rest of the nervous system. This can be clearly seen in fig. 7, A-c. Fig. 7, D, E amplifies some features of figs. 3-5 : particularly the relatively large calibre of the Mauthner fibres at the chiasma, compared to other crossing fibres, and the isolated position of the Mauthner cell-bodies, dorsal to those of the Müller neurones.
CONCLUSIONS
These results show that the brain of the embryonic and early larval lamprey contains a pair of co- ordinating neurones having the main features of the Mauthner neurones found in fishes and amphibia : a very large cell-body lying in the hind-brain at the level of the otic capsule ; a dendritic system in direct connexion with the VHIth nerve; a very large axon crossing the floor of the hind-brain and forming a symmetrical chiasma with its fellow; this axon descends the contralateral side of the spinal cord.
No other pair is set apart in this way. No Müller neurone has a well-marked direct connexion with the VHIth nerve yet, although we know from the work of Johnston, Pearson, and Stefanelli that cells 9 and 10 will develop this connexion later. No Müller neurones have axons descending the contralateral side of the cord: from stage 7 onwards, the axons descend the ipsilateral motor tracts, as so often described in older lampreys.
Axons, other than those of the Mauthner pair, which do cross the floor of the hind-brain, have not yet been traced to their full extent. All at present appear to be typical arcuate fibres, derived from cells dorsal to the Müller cells. The larger arcuate fibres are derived from levels of the brain too far above or below the otic capsule for the cell of origin to be in direct communication with the VHIth nerve.
The interpretations made by previous workers will now be briefly stated, and assessed in the light of the results given above.
Johnston (1902) considered that a pair of Müller neurones lying ‘in the lateral motor column at the level of VII’ is ‘possibly directly homologous with Mauthner’s cells in Acipenser’.
Stefanelli considers that the Müller pair 9 ‘is homologous, from its locations and connexions, with the M(authner) cells, but its morphological characteristics are similar to those of other giant pairs’ (Stefanelli, 1951). Stefanelli and Camposano (1946) agree that the axon of cell 9 follows an ipsilateral course—’decorra lungo il midollo omolateralmente’—but compare these Müller cells with the Mauthner cells of Anguilla and consider that the small but crossing axons of Anguilla bridge the gap between normal Mauthner cells and this Müller pair. Pair 9 is also illustrated by Addens (1933, fig. 59), Barnard (1936, fig. 2), and Woodburne (1936, fig. 3A): its identity is not in doubt since it is described lying between the motor centres of the trigeminal and facial cranial nerves.
Pearson (1936), working partly on the same material as Barnard and Woodburne, but studying different connexions, describes a Müller neurone distinct in position and connexions from that of Barnard and Woodburne. This cell lies at the same transverse level as the ventral nucleus (B, fig. 1), the chief vestibular centre in the lamprey’s acoustico-lateral area ; it is also at the same level as the incoming VHIth nerve fibres, with which it is in direct synaptic relation. This is not cell 9, but is almost certainly cell 10. Comparison of Pearson’s figure with that of Johnston (1902, fig. na) shows that this is the cell which Johnston considered as the Mauthner homologue, and Pearson also concluded that the relations of this cell ‘strongly suggest it is the forerunner of Mauthner’s cell’. Stefanelli (1937) considered that Pearson’s theory concurred with his own. I think Stefanelli was mistaken: two distinct pairs of Müller cells have been proposed as homologues of Mauthner cells.
Kappers and his colleagues (1936, p. 442) have put forward a third possibility. The forerunner of the Mauthner neurone may be one of the larger arcuate cells in the ventral nucleus in the acoustico-lateral area (this nucleus probably corresponds with Deiters’s nucleus of higher vertebrates; compare Beccari, 1943, and Cordier, 1954). These cells are concerned with vestibular reflexes and some of the axons cross the hind-brain to descend the contralateral side of the spinal cord in the medial longitudinal bundle : a hypothetical neurone K, having this relationship, is shown in fig. 1.
This possibility is strengthened by conditions in Myxine (Jansen, 1930), where some of the Müller-type neurones have a contralaterally running axon : these cells are only distinguishable from arcuate cells by their size. But the acoustico-lateral system is greatly reduced in hagfishes, so that a detailed comparison of cells in this system with similar cells in other animals would be inappropriate.
Recent accounts of the vertebrate nervous system do not accept a homology between the Mauthner cell and any particular neurone of cyclostomes or selachians. Beccari (1943, p. 187) contrasts the Müller cell with the Mauthner cell, which ‘differisce essenzialmente per il comportamento del neurite’ and emphasizes the special relation of the Mauthner neurone to the nucleus of Deiters. The same two points appear from Cordier’s diagram (1954, fig. 133) which includes both Müller and Mauthner neurones.
The validity of these theories may now be assessed. Although the fate of the Mauthner neurone during ammocoete and adult life is still unknown, it now seems improbable that this cell can become transformed into either of the Müller cells with which Stefanelli, Johnston, and Pearson wished to homo-logize it. However, statements about this particular pair of neurones form only a minute part of the array of facts presented in these previous contributions to our knowledge of the nervous system of cyclostomes.
On the other hand, the identification by Ahlborn of Mauthner fibres in his material must now be considered as probably correct. It is significant that he was working on younger material than that studied by the later workers. It is therefore possible that the Mauthner neurones have degenerated or become much less prominent, in the stages studied by Tretjakoff, Johnston, Pearson, and Stefanelli.
It is therefore reasonable to conclude that the pair of neurones which has been termed the Mauthner pair in this account is homologous with the Mauthner neurones of fishes and Amphibia and should correctly be given this name.
This in turn suggests that the whole mechanism by which the brain coordinates the activity of the spinal cord, and so of the trunk-myotomes, must be homologous in the ichthyopsid vertebrates, up to at least the stage of development when they can first swim, for the reasons given in the introduction.
On the basis that this homology is accepted, a more detailed comparison may be made. During further development from the 4-mm embryo shown in fig. 5, the neural canal will become broader and shallower and the grey matter will be rotated outwards and downwards about a centre at the median raphe. An intermediate stage in this process is shown in the 8-mm larva in fig. 6, F, and a final stage in a section of the adult brain at this level, e.g. figure 225 of Kappers and others (1936). Consequently the main dorsal dendrite of fig. 5 will be rotated so that it points laterally, while the axon will come to point more medially and less ventrally. These are the relations shown in most figures of gnathostome Mauthner cells, where older stages are usually being described, e.g. the Mauthner neurone of Salmo alevin larvae of Beccari (1943, fig. 145) or Kappers and others (1936, fig. 194). The main or dorsal dendrite of the pro-ammocoete corresponds then with the main or lateral dendrite of typical Mauthner cells.
In fishes and Amphibia, the lateral dendrite is in synaptic relation with vestibular fibres of the Vlllth nerve, and passes close above the ‘descending column’ (‘spinal V’) of the Vth nerve, which runs longitudinally just below the entry of the VUIth nerve, carrying the somatic sensory component of the snout to vagal and spinal levels. These relations of the Vth and VUIth nerves are also found in lampreys (fig. 6, G and Woodburne’s figures). The main dendrite of the pro-ammocoete Mauthner neurone sweeps over the ‘descending V’ column of fibres on its way to the VIHth-nerve vestibular fibres, in just the same way.
In fishes and Amphibia, there is also a large ventral dendrite which runs ventrally among the many longitudinal columns of fibres, especially trigeminal and tecto-bulbar groups. This dendrite leaves the cell-body close to the axonhillock. In younger stages, when the axon was directed more ventrally, for the reason given above, the axon and ventral dendrite would arise from a common stem, just as the axon and axon-collaterals do in the Mauthner cell of the pro-ammocoete. These axon-collaterals correspond in origin, course, and apparently in connexions, with the ventral dendrite of other ichthyopsids : the forwardly directed process from the axon-collateral, shown in fig. 4, is probably coming into contact with the tecto-bulbar system.
There is one striking difference between the Mauthner cells of the pro-ammocoete and those of jawed vertebrates : the position of the Mauthner fibres in the spinal cord. After the pro-ammocoete period, the spinal cord becomes flat and ribbon-like : in some species, two or three pairs of giant fibres then run in a more lateral position (compare Kükenthal, 1929, fig. 218). The lateral position of the Mauthner fibre in the pro-ammocoete may be linked with the approaching flattening of the cord—the growth-factors showing their effect upon this fibre first.
A similar change of position has been described in the hagfish Myxine by Jansen (1930). Some large co- ordinatory fibres, descending the hind-brain floor in the median ventral column, turn outwards and dorsally as they enter the cord, where they run with a motor column which is sufficiently lateral and dorsal in position to lie dorsal to the motor roots. This appears to be a closely parallel example.
Finally, the facts given here have supported the theory of Kappers and his colleagues in that the pro-ammocoete has a Mauthner neurone with the characteristics of their theoretical cell, termed here cell K. But this theory postulated that a unique and giant pair of cells has differentiated from among a number of smaller, similar cells, presumably at the phylogenetic juncture between Ag-natha and Gnathostomata. The Mauthner cell was in fact developed at a much earlier phylogenetic stage.
It seems probable that a course of evolution may have occurred that is opposite to that proposed in their theory. The Mauthner neurone may be derived from a cell with an arcuate fibre, which developed its special Mauth-nerian characteristics before there were any other neurones in the centre— ‘Deiters’s nucleus’—in which it now lies: the Mauthner neurone would be a precursor or pioneer-neurone for the other cells in the same centre, not a derivative of them. Large and distinctive neurones are found more often among primitive chordates or younger stages than’ among advanced or adult ones. The ontogenetic and phylogenetic history of Rohon-Beard neurones, from the lamprey to man, is an example of evolution in this direction. The Rohde cells of Amphioxus, cells A and B mentioned in this paper, and the Müllerian cells of cyclostomes, Amiurus, Anguilla, Gymnarchus, Xenopus, and Tropidonotus provide similar evidence. The ancestor of the vertebrate would possess numerous large and distinctive neurones, paired or unpaired, carrying out between them sensory, motor, correlating, and co- ordinating functions: such a team, measured in hundreds and not millions, might at first be the predominant part of the central nervous system, as it is still in the motile stages of lower vertebrate embryos.
ACKNOWLEDGEMENTS
I have great pleasure in thanking Professor J. E. Harris, F.R.S., for much help and advice. I am also indebted to Mr. G. L. E. Wing and Mr. J. K. Wood for technical assistance.