An anatomical and electrophysiological study of the centrifugal visual system in the lamprey (Lampetra fluviatilis)

Brain Research, 292 (1984) 41-56 Elsevier

41

An Anatomical and Electrophysiological Study of the Centrifugal Visual System in the Lamprey ( Lampetra fluviatilis) N. P. VESSELKIN l, J. REPERANT2.4,*, N. B. KENIGFEST1, D. MICELI 3, T. V. ERMAKOVA 1and J. P. RIO 2

ILaboratory of Comparative Physiology of the Central Nervous System, Sechenov Institute of Evolutionary Physiology and Biochemistry, Academy of Sciences, 44 Thorez A venue, Leningrad 194223 (U. S. S.R.), 2Laboratoire de Neuromorphologie, INSERM U-106, H6pital Foch, Suresnes 92150 (France), 3Groupe de Recherche en Neuropsychologie Expdrimentale, Universitd de Quebec, C.P. 500, Trois Rividres, Que. G9A 5H7 (Canada) and 4Laboratoire de Psychophysiologie Sensorielle, Universit( Paris V1, 9 Quai St Bernard, Paris 75005 (France) (Accepted May 31st, 1983)

Key words: centrifugal visual system - - lamprey visual system

The centrifugal visual system of the lamprey Lampetra fluviatilis was investigated using various neurohistophysiological methods: intraocular injections of [3H]adenosine, fluorescent tracer (Evans Blue) and the iontophoretic deposit of HRP on the optic nerve. Retrogradely labeled neurons were identified bilaterally within the nucleus M5 of Schober and contralaterally in the reticular mesencephalic area (RMA). Comparison of the various orthograde and retrograde labeling results indicated that the neurons of M5 and RMA were labeled via retrograde axonal transport of the different tracers in the retinopetal system and not by orthograde transneuronal processes or from extraretinal pathways. Part of the anatomical data regarding RMA as a site of origin of the centrifugal visual system was confirmed using electrophysiological techniques involving evoked potential and unit cell recordings in RMA following electrical stimulation of the optic nerve. The experiments were performed in the curarized animal under conditions of either normal blood circulation, perfusions of adapted physiological saline, or with a solution known to block chemical synaptic transmission. Various electrophysiologicalcriteria, including the results obtained during the conditions of reversible chemical synaptic blockade, indicated that the responses in RMA reflect an antidromic process. The anatomical organization of the centrifugal visual system in the lamprey is compared to that found in different gnathostome vertebrate species. Several hypotheses concerning the marked interspecies differences related either to the number and the topographical location of the centrifugal neurons as well as the evolution of this system are advanced. INTRODUCTION Over n u m e r o u s decades the question of the existence of centrifugal visual pathways in non-avian vertebrates has been the subject of strong controversy (for reviews see refs. 53, 60, 61 and 74). Only in recent years, and particularly with the development of new neurohistophysiological techniques based upon the retrograde axonal transport of various tracers, has this problem been reconsidered with a greater degree of success 3.9.11,13,15,17.20.23,29,30,35.39-42.54-60~ 65.67,72,77,78. Some recent experiments using the H R P method 29,30.59.60.78 have provided evidence for the possible existence of a centrifugal visual pathway in Larnpetra * To whom all correspondence should be addressed in Suresnes. 0006-8993/84/$03.00(~) 1984 Elsevier Science Publishers B.V.

fluviatilis. Following the intraocular injection of the enzyme, a bilateral though p r e d o m i n a n t l y contralateral labeling of cell bodies was observed in the nucleus M5 of Schober66 and within a neighboring region, the reticular mesencephalic area ( R M A ) . However, based on various methodological considerations, the possibility that the observed n e u r o n a l labeling was the result of either some recently described transneuronal processes 10.14.16.19,24.25,43,44,76,83 or uptake of the tracer material from extra-retinal axon terminals could not be ruled out completely. Therefore, in the present study, other experimental approaches, both anatomical and electrophysiological, have been undertaken in order to confirm the existence of a centrifugal visual pathway in the lamprey.

iii!iii

~ii!i

!~i~ili~iiiiij!iii!!ii!iiiii!i ¸

~i iiii iiiiiii!i~i¸ i i i ~ ! ~ i i i ~

J

i!i!i'~ " ~

c~

~

~ ~i. . . . . . . . . .

~ ~ ~ ~~

~ ~i~i~i~ii~ ~i~

~, ~ !~i ~'i'i~!ii~'~iii~iiii!i!!!ii~ii' ¸¸~i,~i~!iiii~i~ i~,!i~i,

43

Plate I. A-E: photomicrographs of transverse sections showing HRP-labeled retinofugal (A-E) and retinopetal (C, D) systems, 4 days following iontophoretic deposit of the enzyme on the right optic nerve (Mesulam method). A: diencephalon level (x 200). B: pretectal level (× 200). C: tecto-tegmento-mesencephalic level (x 200). D: higher magnification (x 580) of the preceding section showing in the TO the labeling of the retinofugal projections (SFCE) and of the retinopetal axons (arrow). E: high power micrograph (× 580) taken at the thalamic level showing the labeling of the retinofugal system as well as that of some NGL neurons. F: photomicrograph of a transverse section taken from a monitor screen using a videocamera showing Evans Blue-labeled retinopetal neurons (M5 and RMA) 5 days following injection of the fluorescent dye into the right eye (× 310). G and H: photomicrographs of transverse sections showing the retinopetal neurons (M5 and MRA), 5 days after the injection of [3H]adenosine into the right eye (G, × 190; H, × 330). Abbreviations: AOA, accessory optic area; AOT, accessory optic tract; AXOT, axial optic tract; E, ependymal layer; M5, nucleus M5 of Schober; NGL, nucleus geniculatus lateralis; RMA, reticular mesencephalic area; SCFP, stratum cellulare et fibrosum periventriculare; SFCC, stratum fibrosum et cellulare centrale; SFCE, stratum fibrosum et cellulare externum; SM, stratum margihale; TO, tectum opticum; TOM, tractus opticus marginalis; V, ventricle.

44 ANATOMICAL STUDIES Different tracers (HRP, [3H]adenosine, Evans Blue) known to be transported by retrograde axonal flow 1.22,32.33 were used at the level of the peripheral optic system in order to identify the cells of origin of the centrifugal visual system.

Material and Methods (1) HRP experiments HRP (Sigma type VI) was iontophoretically deposited on the optic nerve. The animals (13 specimens) were anesthetized by gill flooding with a 0.3% urethane solution. After enucleation the optic nerve was gently exposed, then isolated from the oculomotor nerves which were covered with paraffin oil. A glass micropipette (20-50 p m in diameter) filled with 4% HRP in 0.1 M phosphate buffer, pH 8.0, was connected to a current generator and placed in contact with the sectioned optic nerve. An electrode positive 10/~A current was delivered during 20 min. Later, the animals were reanimated in oxygen-saturated water. They were allowed to survive for 4-14 days in water kept at 5-7 °C and were sacrificed by a transcardial perfusion of 0.7% saline followed by a 25% diluted Karnovsky solution 26. Forty/~m thin brain sections were obtained on a freezing microtome and processed for the HRP reaction according to the B D H C Mesulam technique 38. The sections were examined both under light- and dark-field conditions.

(2) [3H]Adenosine experiments Ten lampreys received a unilateral eye injection of [3H]adenosine (10-20 pCi, specific activity 30-40 Ci/mmol). The animals were allowed to survive in fresh-water for either 1 day (n = 2), 2 (n = 1), 4 (n = 2), 6 (n = 2), 8 (n = 2) or 10 days (n = 1). They were then perfused transcardially with 0.7% saline followed by a 25% Karnovsky solution26. Their brains were embedded in paraffin and sectioned serially (12 /~m) in the transverse plane. The sections were coated with Ilford K5 emulsion, dried and stored at 4 °C in light-tight boxes for a 4-5 week exposure. Next, they were developed in D19 fixed in sodium hyposulfite and counterstained with cresyl violet. The sections were examined under light- and dark-field illuminations.

(3)Fluorescent tracer experiments Twelve lampreys received a unilateral ocular injection (10pl) of Evans Blue, a fluorescent dye which had been dissolved in distilled water (10% w/v solution containing 1% poly-L-ornithine). After 3-6 days survival at 5-7 °C, the animals were anesthetized and perfused transcardially with 0.7% saline followed by a 10% formol solution, then 30% sucrose solution. The brains were removed and sectioned at 40/~m in the frontal plane on a freezing microtome. The sections were immediately mounted onto slides, airdried and coverslipped using a low fluorescent mounting medium (Eukitt). They were studied using a Leitz Ploemopack fluorescence microscope equipped with a filter-mirror system providing an excitation wavelength of + 550 nm (system N2). The structures containing the fluorescent label were mapped on a monitor screen using a videocamera.

Results (1) HRP experiments Labeling of the primary visual system (Plate L A-E). After intraocular injection of the enzyme, orthograde H R P labeling was observed within the primary optic system (POS) in 60% of the experimental cases. However, the retinal projections were never revealed entirely29. 30.78. In contrast, following the iontophoretic deposit of H R P on the sectioned optic nerve, the whole extent of the POS was labeled in all cases. The optic fibers displayed a homogeneous dark bluish-green coloration throughout the full length of the optic tract. Similarly, the optic centers upon which the optic fibers projected were intensely stained in toto and optic terminals were represented by an extremely dense accumulation of fine peroxidase granules. Optimal results were obtained 4-6 days following the iontophoretic H R P deposit. The present data concerning the bilateral projections to the postoptic area, the nucleus geniculatus lateralis, the area pretectalis and the superficial layers of the optic tectum, are similar to those which have been reported in previous studies on this same material using either the anterograde degeneration method28,29,45 or the autoradiographic technique 56,57,78. However, the present material also provided some complementary information regarding the trajectory of several optic tracts, and particularly of the accessory optic tract

45 (AOT) which, to date, has not been accurately described. The A O T emerges from the ventral border of the marginal optic tract (TOM) at the level of the posterior thalamus. This tract is made up of large diameter optic fibers and runs anteriorly along the latero-ventral border of the mesencephalon. It then turns ventro-medially to terminate diffusely within the accessory optic area, a large zone of the ventral mesencephalic tegmentum (Plate I, A - C ) .

Somatic labeling of M5 and RMA neurons (Plate L C, D). The iontophoretic deposit of H R P on the sectioned optic nerve produced, 4-6 days after and in all cases examined, a bilateral though predominantly contralaterai labeling of neurons within M5 and also of contralateral R M A neurons. These results are comparable to those previously obtained following the intraocular injection of H R P and using the DAB method ~9,3o,6°,7s. However, the labeling observed here was far more intense and the cellular morphology more apparent since the peroxidase material invaded the rather long dendritic extensions and sometimes a segment of the axon proximal part. The cell bodies of M5 were pyramidal in shape and larger than those observed in R M A which generally appeared fusiform. The dendrites of M5 and RMA neurons were mainly oriented towards the optic tectum. Some of them penetrated the deep and intermediate layers of the optic tectum and occasionally extended as far as the superficial layers where retinal fibers terminated. The labeled axons of M5 and R M A neurons followed a latero-medial course reaching the deepest layers of the optic tectum through its ventrolaterai flank. They then crossed the intermediate layers extending to the zone of terminal optic labeling situated in the superficial tectal layers. It was not possible to follow their trajectory beyond this region. Somatic labeling of neurons in the NG L contralateral to the optic nerve injection (Plate L E). In 40% of all cases examined, and particularly after long survival periods, some small neurons containing H R P label were found in the contralateral N G L generally situated in immediate proximity to the TOM. No other neuronal labeling was observed in the brain. (2) [3H]A denosine experiments One to two days after the intraocular injection of

[3H]adenosine, the primary visual system was lightly labeled. Elsewhere, a weak somatic labeling was observed in the neuronal bodies of M5 and RMA. Four days following the [3H]adenosine injection, the labeling appeared heavier in M5 and R M A (Plate I, G, H) as well as in the primary visual centers. In the latter regions, the label was situated within the neuropii but a small accumulation of silver grains was also found overlying the different cell bodies. With survival periods of 6--10 days, the silver grain concentration increased in all of the above-mentioned structures, however a significant amount of label was also detected in the regions adjacent to those containing optic boutons. Here, the label was localized in the neuropil and within the different neuronal and glial cell bodies. In all cases, such an aspecific somatic labeling appeared much less pronounced than that observed in the neuronal bodies of M5 and RMA.

(3) Fluorescent tracer experiments Four to 6 days following the intraocular injection of Evans Blue (EB), a somatic labeling was identified essentially within M5 and R M A (Plate I, F). Moreover, no label was observed within the primary optic system. When viewed with the N2 filter-mirror system, EB cells displayed a bright homogeneous red fluorescence. The total number of neurons labeled within M5 and MRA was approximately 750 and was constant regardless of the tracer used. Discussion The morphological data obtained previously using intraocular injections of H R P 29,30.6°,78 showed labeling of neurons within M5 and RMA. Furthermore, various experimental controls demonstrated that this cellular labeling could be attributed neither to the presence of endogenous proteins apt to give an HRPpositive reaction nor to the retrograde transport of the enzyme from those regions lacking a blood-brain barrier TM. It was thus concluded that the observed labeling was probably a direct result of H R P transport from the centrifugal fiber endings located in the retina. However, the possibility that these neurons could have been labeled via the retrograde transport of the enzyme from extra-retinal axon terminals could not be discounted entirely. The hypothesis that

46 the labeling of M5 and R M A cells was due to transneuronal transport of the tracer from orthogradely labeled optic terminals could also not be ruled out. On the basis of the different morphological data accumulated in the course of this study, an attempt will be made in the present discussion to provide some arguments against both of these hypotheses and sufficient proof that the centrifugal visual system of the lamprey does in fact originate in M5 and RMA. (1) Do M5 and R M A neurons project to extra-retinal structures such as ciliary ganglia, and the oculomotor or corneal muscles? Various observations do not support the existence of such connections: (a) agnathans are the only vertebrates which lack ciliary ganglia63; (b) oculomotor neurons have been identified using the H R P technique 12 and are located in regions which are rather distant from M5 and RMA; (c) the corneal muscle which is characteristic of lampreys 63 is innervated by a branch of the Xth nerve 34 whose motoneurons are situated in the brainstem; (d) lastly, in the present experiment using the iontophoretic deposit procedure, the labeling of M5 and RMA cell bodies could only result from the uptake of HRP by either retinopetal or retinofugal axons contained within the optic nerve.

(2) Are M5 and R M A neurons non-retinopetal? Different investigators using H R P 10,16,19.25,43,44,76 have suggested the possibility of somatic labeling of neurons by means of transneuronal processes. Several studies have also indicated that the somatic neuronal labeling with [3H]adenosine can occur via retrograde labeling 22,54,55,58,69.8°.81, but also through the transneuronal transport of the nucleoside 2z.54-5~. 6s,so,sl. Evans Blue has been shown under certain experimental conditions to be capable of labeling glial cells transneuronally 3e. Elsewhere, HRP 5z and adenosine 22,54,58,68.~1 have been reported to be transported via the orthograde axonal flow. It should be noted that in the present HRP and [3H]adenosine experiments, the primary optic tract and the neurons of M5 and RMA were found labeled in the same preparations. Following the iontophoretic deposit of H R P on the optic nerve, the dendrites of these neurons were also seen to reach some regions of the superficial layers of the optic tectum where optic fibers terminate.

Thus, it is conceivable that the somatic labeling of M5 and RMA neurons stemmed from the transneuronal transport of tracer material subsequent to its release at the optic terminals, then taken up by dendrites and transported to the somas of M5 and RMA. Various data derived from different experiments do not support this hypothesis: (a) the labeling appears extremely constant and reproducible with all of the different tracers used and occurs at very short survival times (case of [3H]adenosine); (b) in those cases where the somatic labeling does result indirectly from transcellular transport (case of [3H]adenosine), the labeling is weak and aspecific, affecting for example virtually all of the cell bodies in the optic tectum, and this is observed after long survival periods; (c) the concomitant labeling of optic fibers and somas of M5 and RMA renders the process of transneuronal transport of the tracers improbable as this process requires a certain length of time; (d) in some preparations following the intraocular injection of HRpT8, or when Evans Blue was used, no labeling of the optic tract was noted (particularly of the retinotectal pathway), however labeled somas were observed in M5 and RMA. (3) Significance of the neuronal labeling within the N G L contralateral to the optic nerve injection of H R P The results derived from different experiments suggest that such a labeling occurs through the transsynaptic transport of the enzyme from the optic terminals: (a) in those cases where a labeling was noted (40% of the cases examined), the number of labeled neurons varied from one animal to another (6-20 cells) and their topographical location along the anterior-posterior axis also differed markedly; (b) such neurons were always positioned adjacent to and against the tractus opticus marginalis (TOM), which showed a very strong peroxidase reaction; (c) labeling of the NGL was never observed in the absence of orthograde axonal transport of HRP within the TOMT8; (d) correspondingly, no labeling of NGL neurons was observed following the intraocular injection of Evans Blue which appears not to be transported orthogradelyX.6,79; (e) finally, with long survival periods, the intraocular injection of tritiated adenosine produced a light labeling which appeared transsynaptic in nature 5~ and was observed over all somas of the NGL.

47

(4) Conclusion The somatic labeling of M5 and R M A neurons obtained using the different techniques can only involve the retrograde axonal transport of these tracer materials from axon terminals situated in the retina., Correspondingly, the centrifugal visual system in the lamprey originates in the mesencephalic tegmentum at the level of M5 and RMA. ELECTROPHYSIOLOGICAL STUDIES This series of experiments was undertaken with the aim of verifying the preceding conclusions by attempting to demonstrate antidromic responses in one of the involved centrifugal structures (RMA) after electrical stimulation of the optic nerve.

Material and Methods Twenty-one lampreys were used during the course of this study. The animals were immobilized by intramuscular injections of a curare solution (0.2 mg/I). The respiration was maintained by gill flooding with superoxygenated water. The ocular globe was retracted and the ocular muscles sectioned. A stainless steel hook was introduced into the ocular globe allowing the eye to be distended slightly in order to adequately expose the optic nerve. Using this preparation, the optic nerve constituted the sole remaining link of the ocular globe. After sectioning the skin, the cephalic muscles and the fibrous roof which covers the brain, both the dorsal and lateral aspects of the diencephalon and mesencephalon were exposed. An active stainless steel stimulating electrode of 0.2 mm diameter was positioned directly upon the optic nerve at a point proximal to the ocular globe. The stainless steel hook which was implanted into the ocular globe served as the reference electrode. The electrical stimulation consisted of rectangularwave impulses of 16--20 mV amplitude and 0.2 ms duration. The stimuli were presented through a radiofrequency block. Evoked potentials or single unit responses were recorded at the level of the mesencephalon using glass microelectrodes filled with a solution of Procion-brilliant red diluted to 4%. The electrode resistance varied between 3 and 5 MQ for recording focal evoked responses and between 10 and 15 Mr2 for single unit activity. The reference elec-

trode consisted of a steel needle placed in the animal's buccal, cavity, The recorded signals were led into a high impedance input amplifier and photographed on an oscilloscope screen. A bandpass filter of 0.1-500 Hz was used for the evoked potential recordings and 0.1-5000 Hz for unit responses. The recordings were performed on either (a) normal animals, (b) animals perfused with physiological saline suited for cyclostomes (Solution A 71) or (c) animals perfused with a solution known to block chemical synaptic transmission (Solution B75). In order to perfuse the animals by these various means, a cannula was inserted into the aorta and the auricle incised to permit elimination of the perfusants. Solution A contained, in ~mol: NaCI, 86.3; KC1, 2.5; NaHzPO4, 8.0; CaCI 2, 4.7; MgC12, 0.9; N a H C O 3, 0.8 and C6H120 6, 5.5. The solution was cooled to 4-6 °C and saturated with 98% 02 and 2% CO z to achieve a pH of 7.4-7.6. In order to obtain a long-lasting immobilization of the animals for the duration of the experiment, 0.15 mg/! of curare was added to solution A. The perfusion was performed at a rate of 40-50 drops/min. The blockade of chemical synaptic transmission was obtained by replacing Solution A by Solution B, in which Ca 2+ was replaced by Mn 2÷ ions. This solution was cooled to 4-6 °C and contained, in ~mol: NaCI, 119; KCI, 2.5; MgCI 2, 0.9; NaHCO 3, 4.0; C6H1206,5.5 and MnCI z, 2.0. For the histological localization of the recording site, Procion-brilliant red stain was released iontophoretically at the electrode tip with an electrode negative current of 3-6/~A lasting 10 s. The animals were then perfused with 10% formol, and the brains removed and postfixed in a formoi solution containing 30% sucrose. They were cut serially on a freezing microtome at 50-60 ~m thickness in the transverse plane (see Fig. 8).

Results (1) Recordings in the normal curarized animal Electric shock stimulation of the optic nerve produced a characteristic response within the contralateral optic tectum (OT) and mesencephalic tegmentum (MT). Surface potentials on the optic tectum consisted of either negative or negative-positive waves with latencies of 10-12 ms. These waveforms were often preceded by a short-lasting presynaptic

48

C

Figs. 1-3. Responses recorded in the mesencephalon of the lamprey to electrical stimulation applied to the contralateral optic nerve. Fig. l'. Schematic representation of a transverse section ~through the lamprey mesencephalon showing the recording sites within the tectum opticum (TO, striped zone: a and b) and reticular mesencephalic area (RMA: c). Fig. 1. Recordings in different parts of the mesencephalon of the curarized lamprey under conditions of normal blood circulation. The potentials shown were obtained from the tectal surface (a), the inner strata of the tectum opticum (b) and the RMA of the mesencephalic tegmentum (c). Fig. 2. Evoked potentials (a) and single unit (a') responses recorded in the tectum opticum during conditions of perfusion with Solution A. Fig. 3. Changes in the tectal response under different perfusion conditions: a, Solution A; a', Solution B; and a", 5 min after replacing Solution B by A. Calibrations = 100 mV and 10 ms.

49 component (Fig. la). As the electrode penetrated into OT, an inversion of the response was observed throughout the full thickness of the structure (Fig. lb). Upon reaching the MT, a persistent and marked increase in the response amplitude (positive wave) was noted, the maximal response (Fig. lc) was obtained at a precise zone of the dorsomedial MT (150/~m in height and 200/~m in length). The histological localization of this zone showed it to correspond essentially to RMA (Figs. 1' and 8). The latter zone was the same as that found to contain neurons which were labeled following the intraocular injection of different tracers, but also constituted a site of passage for the dendrites of cells located in M5 which were also labeled by the tracers (Fig. lc). The strong response of complex waveform persisted in MT following the ablation of the contralateral optic tectum. A retinal projection to the MT has been described previously (see p. 54). Consequently, the evoked potentials recorded in this region might represent an orthodromic response generated by elements which were activated by retinal afferents. However, it is conceivable that this orthodromic response may mask a response which is antidromic.

the MT but never in RMA. Moreover, when animals were placed under conditions blocking chemical synaptic transmission (perfused with Solution B), an evoked response was recorded in RMA following the stimulation of the optic nerve. The response consisted of a negative wave whose latency was approximately 10 ms, with an amplitude of 100-120/tV and lasting 5-8 ms (Fig. 5c). With optic nerve stimulation, unit spike activity could also be recorded in the RMA and coincided with the appearance of the focal or slow-wave response (Figs. 5c', c"). The latter response recorded in RMA did not disappear with stimulation frequencies of 80-100 Hz (Fig. 7). Following stimulation of the contralateral optic nerve an evoked response was recorded in the dorsolateral MT which corresponded to an area of passage of optic fibers (TOM: d, Fig. 4). Its amplitude was relatively higher than that of the preceding responses and comprised a short triphasic wave of 4--6 ms latency (Fig. 6). Simultaneously recording potentials from RMA and TOM at the same anteriority showed the latency of the response within the former structure to be 5-6 ms longer than that in the latter. Discussion

(2) Recordings in animals perfused with either Solution A or B Under conditions where animals were given a trans-aortal perfusion with the physiological Solution A, the brain conserved normal electrical activity during several hours. The properties of the focal responses in the OT and MT following stimulation of the contralateral optic nerve did not appear to differ from those recorded under conditions of normal blood circulation in the curarized animal (Figs. 2a, 3a and 4a). Furthermore, in the animals perfused with Solution A, very distinct cell responses were produced by optic nerve stimulation in the various structures involved (Fig. 2b). Replacing Solution A by Solution B, which is likely to block chemical synaptic transmission21, 27, completely abolished, after 5 min, the evoked responses of OT generated by the stimulation of the contralateral optic nerve (Fig. 3a'). Correspondingly, when Solution B was replaced by Solution A, a reappearance of the evoked tectal activity was noted after 5 min (Fig. 3a"). The abolition of the evoked responses could also be obtained in several regions of

In order to provide supplementary evidence for the existence of a centrifugal visual system in the lamprey, it is of primary importance to show that the evoked potentials recorded in RMA under conditions of chemical synaptic blockade represent antidromic responses which are generated by optic nerve stimulation. Consequently, the following two hypotheses must by discounted: (A) the response recorded in RMA constitutes an action potential of retinofugal fibers, and (B) that the response is generated by elements within RMA which are excited transsynaptically. (1) Hypothesis A Stimulating the optic nerve produces a large synchronous discharge of the retinofugal fibers in the TOM which is situated very laterally in relation to RMA. The TOM response appears earlier than that of RMA and displays a markedly different waveform. In the TOM, a characteristic triphasic wave is recorded when the propagation wave of excitation passes through the recording site. In contrast, in

Cd

-

4’

51 sponses r e c o r d e d in R M A do not constitute action potentials of retinofugal fibers.

(2) Hypothesis B The signals r e c o r d e d in R M A are g e n e r a t e d by elements which are intrinsic and excited transsynaptically by way of the retinofugal system. This would p r e s u p p o s e that Solution B does not induce a total blockade of chemical synaptic transmission. The latter seems i m p r o b a b l e as it is generally believed that replacing Ca ions by Mn or various other bivalent ions leads to a virtual complete b l o c k a d e of chemical synaptic transmission 2t,27. F u r t h e r m o r e , the fact that Fig. 8. Photomicrographs of a transverse section showing the histological localization of the recording site in RMA. Procionbrilliant red stain (arrow) was released iontophoretically at the electrode tip. TO, tectum opticum; RMA, reticular mesencephalic area (x 185).

R M A the potentials a p p e a r m o n o p h a s i c which implies that the zone of electrogenesis is not very distant from the recording site. These various observations are evidence in favor of the notion that the responses of R M A cannot be attributed to the activity of T O M optic fibers. It is conceivable that the signals r e c o r d e d in R M A reflect the responses of retinofugal fibers exhibiting slower conduction velocities c o m p a r e d to those of T O M and are g r o u p e d into a compact tract. However, this is unlikely on the basis of various neuroanatomical data regarding the organization of the visual system in the l a m p r e y (refs. 28, 29 and 79 and present results) which have never described the existence of an optic fasciculus projecting to R M A . T a k e n together, these various r e m a r k s indicate that the re-

the potentials in R M A persist during optic nerve stimulation with frequencies of 80-100 Hz constitutes additional evidence that the R M A excitation is not p r o d u c e d transsynaptically. That a given response is capable of following stimulus frequencies of 100 Hz is usually considered an a d e q u a t e criterion indicating an antidromic process7E Finally, it is also possible that the responses in R M A are g e n e r a t e d by postsynaptic elements which are activated via electrical synapses. Such synapses a p p e a r to be quite numerous in the CNS of the l a m p r e y 70. H o w e v e r , this hypothesis can also be discounted on the basis of what was said earlier concerning the absence of a retinal projection to R M A .

(3) Conclusion The present results indicate that the responses evoked in R M A following contralateral optic nerve stimulation and after the b l o c k a d e of chemical synaptic terminals are antidromic in nature and imply that the underlying process involves the activation of the centrifugal visual system.

Figs. 4--7. Responses recorded in the mesencephalon during the trans-aortal perfusion of either Solution A or B. Fig. 4'. Schematic representation of the lamprey mesencephalon showing the location of the recording sites at the tectal surface (a) in the tractus opticus marginalis: TOM (d), and in the reticular mesencephalic area: RMA (c). Fig. 4. Tectal responses recorded during the perfusion with Solution A, 5 sweeps superimposed. Fig. 5. The responses recorded in RMA during the perfusion with Solution B. c, evoked potentials, 5 sweeps superimposed; c' and c", unit responses. Fig. 6. The responses (d) recorded in TOM during the perfusion of Solution B. Fig. 7. Recordings obtained under conditions of perfusion with Solution B; responses in RMA. c, to a single shock stimulis; c', to stimuli applied at a rate of 80/s. The first potential in the sweep was recorded after 5 s of stimulation. The dots indicate the stimulus and the arrows the peaks of the waves. Calibrations = I50/~V and 10 ms.

52 GENERAL DISCUSSION The various histophysiological and electrophysiological data derived from the present study offer substantial evidence in favor of the existence of a centrifugal visual system in the lamprey. In total, 35,000 fibers have been counted in the optic nerve of Lampetra fluviatilis46. In the present study, and using either of the different histophysiologicai methods, the estimated number of labeled neurons within M5 and RMA was around 750 which would correspond to about the same number of centrifugal optic fibers. Thus, the centrifugal visual component would appear to represent approximately 2% of the total number of fibers contained in the optic nerve. Furthermore, the present morphological observations showed that the centrifugal visual axons leave the optic tectum at the level of the superficial layers and attain the retina through the marginal optic tract. Moreover, the electrophysiological data indicate that these axons have a smaller diameter than those of the retinofugal fibers. The histological data clearly show that numerous dendrites of centrifugal visual neurons, and particularly those of RMA, extend as far as the superficial layers of the optic tectum, a region which also receives retinofugal endings. It is therefore conceivable that these neurons are activated more or less directly by the axons of retinal ganglion cells. Such an organization suggests a closed-loop feedback system comparable to the retino-tecto-isthmo-retinal system of birdsS,36,37. The existence of a centrifugal visual system in numerous vertebrate groups has been actively investigated in recent years. In the following discussion, an attempt will be made to relate the results of the present study with those obtained in other vertebrate species. It was towards the end of the last century that the existence of a centrifugal visual system was demonstrated for the first time in birds using the retrograde or orthograde degeneration techniques 47.82. As in the lamprey, the structure of origin of the centrifugal visual pathway, the nucleus isthmo-opticus (NIO), is located within the mesencephalic tegmentum. Also, there is some evidence that the NIO projects mainly upon the contralateral retina and, more precisely, upon the internal plexiform layer at the amacrine cell levels 5~7,5o,51. The use of the H R P method at the ret-

inal level has made it possible to confirm that the NIO is the main source of the centrifugal visual pathway 33. Moreover, this same anatomical technique has demonstrated the presence of other centrifugal visual neurons situated in the CG (substantia grisea centralis), close to the NIO which display a bilateral, though predominantly contralateral ocular projection18. In different species belonging to most of the major taxonomic groups (teleostean, selacian, amphibian, reptilian, mammalian), it has be.en possible to identify a number of centrally labeled structures after the intraocular injection or iontophoretic deposit of various tracer substances (HRP, [3H]adenosine, fluorochromes, cobaltous-lysine). The latter labelings have generally been interpreted as representing the sites of origin of centrifugal visual pathways. However, it is important to note that the topographical localization of these different structures varies greatly from one species to another and also within the same taxonomic group. In teleosteans, 21 species belonging to 11 different families have been investigated using H R P or cobaltous-lysine methods. The results obtained were extremely varied. In the Ameiridae and Noptopteridae, no retrogradely labeled structures have been observed 42. In the Cyprinidae the labeling of tectal 65 (TO) or olfactory bulb (OB) neurons 72 has not been confirmed (refs. 42, 48, 73 and 77 for TO; ref. 42 for OB). In the other families investigated, the number of contralaterally labeled structures has differed between one (Aluridae 77, Pantodontidae15), two (Anabandidae 42, Centrarchidae42), three (Mockolidaeg. 39, Poecilidae42), four (Tetraodontidae 9,4°, most Ciclidae 42) and five (in other Ciclidaeg). The latter are situated either within the tectum opticum (stratum griseum et fibrosum superficiale and stratum griseum centrale: Tetraodontidae9, 40, Mockolidae9, 39, Ciclidae9), in the pretectum (subnucleus of the pretectal c o m p l e x 9,39,4° o r pretectal nucleus42: Tetraodontidae 9~4°, Mockolidae 9,39, Ciclidae 9,41,42, Anabandidae 42, Centrarchidae 42, Poecilidae41,42), in thalamus (dorsomedial optic nucleus: Mockolidae 9,39, Tetraodontidae9, 40, Ciclidae9); corpus geniculatus lateralis ipsum: Tetraodontidae9, 40, Ciclidae 9, in the ventral hypothalamus (Poecilidae 41,42) and the preoptic region (preoptic area: Aluridae77; nucleus intrachiasmaticus: PantodontidaelS). Other similarly labeled

53 structures have been found in the ventromedial anterior telencephalon (nucleus olfactoretinalis42 or the telencephalic optic nucleusg: Ciclidae9,41,42, Poecilidae41,42, Centrarchidae 42, Anabandidae42; nucleus of the medial olfactory tract and the rostral olfactory bulb: Ciclidae42). A weak ipsilateral labeling has been reported rarely (only in Centrarchidae 42) and situated mainly within the nucleus olfactoretinalis. The presence of labeled neurons within the optic tectum (zona externum of the stratum cellulare externum) following the contralateral intraocular injection of HRP has been described in the selacian

Ginglysmostoma cirratum 35. In amphibians negative results have been obtained in anurans 1°,13,64,83whereas in urodeles and following the intraocular injection of HRP a bilateral but predominantly contralateral label has been observed within the pretectal area lying near the commissura posterior 13. In reptiles, the location of labeled neurons after the intraocular injection of either HRP, [3H]adenosine or various fluorescent tracers also exhibits a marked degree of variation. For example, such labeled neurons may be situated either in the mesencephalic tegmentum (in the crocodilian Caiman crocodilus 11, in the turtle Pseudemys scripta 3, in the lizards Gerrhonotus coeruleus 17 and Varanus exanthematicus20), in the ventral thalamus (nucleus of the ventral commissure TM or centrifugal optic thalamic nucleus55: in the snakes, Crotalus viridis67, Vipera asp/~ 54,55,60, Thamnophis sirtalis17, Tharnnophis radix 17 and in the lizard Cordylus cordylus 17) or in the telencephalon (basal areas: in the snake Python reticulatus20). In most species the labeling is bilateral but exhibits a contralateral bias. Investigations in mammals regarding the origin of the centrifugal visual system and employing the intraocular injection of a variety of tracers (HRP, [3H]adenosine, fluorescent substances) have often been unsuccessful 4,6,31.49,52,53,59,60,79,81. Only recently, using the HRP technique in the rat, has a bilateral, though predominantly contralateral projection, been reported originating from the medial pretectal area 23. The topographical localization of structures supposedly at the origin of centrifugal visual pathways in the different species examined therefore appears extremely varied. Consequently, it is very difficult to

establish adequate homologies between these structures and, more specifically, in relation to the RMA and M5 of the lamprey. At least three hypotheses which attempt to explain either the high degree of variability or the a priori contradictory results can be advanced. (1) The centrifugal visual system of vertebrates is polyphyletic in origin, having appeared on several occasions and independently in different vertebrate lineages and arising from rather diverse structures. As such, the wide range of centrifugal visual structures identified may be functionally different and cannot be considered as being truly homologous. Instead, they may be referred to as being homoplasic 2. It should be noted, however, that a centrifugal visual system is present in one of the most primitive living vertebrates, the lamprey, and consequently represents a very primitive anatomical arrangement which probably arose at the agnathan stage. Elsewhere, if one assumes that gnathostome vertebrates stemmed from the main lineage of jawless vertebrates62, the phylogenetic hypothesis implies that this specific anatomical arrangement was lost at the gnathostome stage, but was then re-established during the course of evolution in the different lineages of gnathostome vertebrates. (2) According to Ebbesson 8, neuronal systems evolve by a process involving the loss of connections instead of the creation of new connections with hitherto unrelated targets (parcellation theory). For example, the extensive retinopetal system seen in teleosts may reflect a primitive organizational arrangement and the reduction of such connections which have been described in more advanced vertebrates (reptiles, birds, mammals) may be the result of the neuronal system evolution by a loss of pathways 9. This evolutionary hypothesis can be questioned based upon the fact that in lampreys, which constitute a more primitive vertebrate form than the teleost fish, the number of structures of origin of centrifugal visual pathways appear more reduced (2) than those found in these fish (as many as 5), which represent the most evolved forms of the actinopterygian lineage 62. (3) The labeling of neuronal cell bodies observed within various structures following the intraocular injection of different tracers does not result from their retrograde axonal transport after being taken up by

54 the intra-retinal terminals of the centrifugal system.

gian lineage and the dipnoans of the sarcopterygian

Therefore, such structures cannot be considered to

lineage. This would provide additional information

truly represent the sites of origin of centrifugal visual pathways. Because of insufficient controls in several

for testing either of the above m e n t i o n e d hypotheses. Furthermore, a comparative analysis of these struc-

experiments previously reported, the neuronal label-

tures at both the hodoiogical (intra-cerebral connec-

ing which has been observed may in fact be the result

tions, precise localization of their retinal projection

of either the transsynaptic transport of tracer sub-

zones) and functional levels should offer a better un-

stances from the optic terminals, or the retrograde transport of tracers from extra-retinal axon terminals which are intra- or extra-ocular (see p. 49).

derstanding of the basis for such marked inter-species variation and its phylogenetic significance.

Consequently, it would be difficult given our pres-

ACKNOWLEDGEMENTS

ent state of knowledge to clearly explain the existing interspecies differences in the topographical location

The authors thank F. Roger, G. Sanchez and L.

of the structures believed to be retinopetal. The questions must be reconsidered using new experimental approaches. For example, neuroanatomical and electrophysiological investigations with techniques similar to those employed in the present study in the lamprey should be extended to identify such

Tremblay for their excellent support, D. Le Cren for his skillful photographic assistance, Dr. P. Paulhus, Director of the A q u a r i u m de Qu6bec and Prof. A.

centrifugal structures in other primitive forms such as the chondrosteans and holosteans of the actinoptery-

Barets of the Universit6 de Bordeaux, for providing some of the specimens. This work was supported by the U.S.S.R. Academy of Sciences, the D G R S T , (R.G. 82E 683) the 1NSERM, the CNRS (France) and the C R S N G (Canada).

REFERENCES 1 Aschoff, A. and Holl~inder, H., Fluorescent compounds as retrograde tracers compared with horseradish peroxidase (HRP). I. A parametric study in the visual system of the albino rat, J. Neurosci. Meth., 6 (1982) 179--197. 2 Campbell, C. B. G. and Hodos, W., The concept of homology and the evolution of the nervous system, Brain Behav. Evol., 3 (1970) 353-367. 3 Cervetto, L., Cited in Marchiafava, P. L., The centrifugal actions on amacrine and ganglion cells in the retina of the turtle, J. Physiol. (Lond.), 255 (1976) 137-155. 4 Colman, D. R., Scalia, F. and Cabrales, E., Light and electron microscopic observations on the anterograde transport of horseradish peroxidase in the optic pathways in the mouse and rat, Brain Research, 102 (1976) 156-163. 5 Cowan, W. M., Centrifugal fibres to the avian retina, Brit. med. Bull., 26 (1970) 112-118. 6 Davis, J. N. and McKinnon, P. N., Anterograde and transcellular transport of a fluorescent dye, bisbenzimide, in the rat visual system, Neurosci. Lett., 29 (1982) 207-212. 7 Dowling, J. E. and Cowan, W. M., An electron microscope study of normal and degenerating centrifugal fibre terminals in the pigeon retina, Z. Zellforsch., 71 (1966) 14-28. 8 Ebbesson, S. O. E., The parcellation theory and its relation to interspecific variability in brain organization, evolutionary and ontogenetic development and neuronal plasticity, Cell Tiss. Res., 213 (1980) 179-212. 9 Ebbesson, S. O. E. and Meyer, D. L., Efferents to the retina have multiple sources in teleost fish, Science, 214 (1981) 924-926. 10 Ermakova, T. V., Kenigfest, N. B. and Vesselkin, N. P., Transneuronal distribution of peroxidase in the brain of the

11 12 13 14

15

16

17

18 19

frog Rana temporaria, Z. Evol. Biokhim. Fisiol., 17 (1981) 95-96. Ferguson, J. L., Mulvanny, P. J. and Brauth, S. E., Distribution of neurons projecting to the retina of Caiman crocodilus, Brain Behav. Evol., 15 (1978) 294-306. Finger, T. E. and Rovainen, C. M., Retrograde HRP labeling of the oculomotoneurons in adult lampreys, Brain Research, 154 (1978) 123-127. Fritzsch, B. and Himstedt, W., Pretectal neurons project to the salamander retina, Neurosci. Lett., 24 (1981) 13-17. Gerfen, C. R., O'Leary, D. D. M. and Cowan, W. M., A note on the transneuronal transport of wheat geriu agglutinin-conjugated horseradish peroxidase in the avian and rodent visual systems, Exp. Brain Res., 48 (1982) 443-448. Gerwerzhagen, K., Rickmann, M. J., Meyer, D. L. and Ebbesson, S. O. E., Optic tract cells projecting to the retina in the teleost, Pantodon buchholzi, Cell. Tiss. Res., 225 (1982) 23-28. Gomez-Bamos, P. and Rodriguez-Echandia, E. L., Retrograde axonal transport and transneuronal transference of horseradish peroxidase in the rat ciliary ganglion, Experientia, 37 (1981) 1337-1338. Halpern, M., Wang, R. T. and Colman, D. R., Centrifugal fibers to the eye in a nonavian vertebrate: source revealed by horseradish peroxidase studies, Science, 194 (1976) 1185-1188. Hayes, B. P. and Webster, K. E., Neurones situated outside the isthmo-optic nucleus and projecting to the eye in adult birds, Neurosci. Lett., 26 (1981) 107-112. Hongo, T., Kudo, N., Yamashita, M., Ishizuka, N. and Mannen, H., Transneuronal passage of intraaxonally injected horseradish peroxidase (HRP) from group lb and II fibers into the secondary neurons in the dorsal horn of the

55 cat spinal cord, Biomed. Res., 2 (1981) 722-727. 20 Hoogland, P. V. and Welker, E., Telencephalic projections to the eye in Python reticularis, Brain Research, 213 (1981) 173-176. 21 Huckett, J. T., Selective antagonism of frog cerebellar synaptic transmission by manganese and cobalt ions, Brain Research, 114 (1976) 47-52. 22 Hunt, S. P. and Kiinzle, H., Bidirectional movement of label and transneuronal transport phenomena after injection of [3H]adenosine into the central nervous system, Brain Research, 112 (1976) 127-132. 23 Itaya, S. K., Retinal efferents from the pretectal area in the rat, Brain Research, 201 (1980) 436-441. 24 Itaya, S. K. and Van Hoesen, G. W., W G A - H R P as a transneuronal marker in the visual pathways of monkey and rat, Brain Research, 236 (1982) 199-204. 25 Itaya, S. K., Williams, T. H. and Engel, E. L., Anterograde transport of horseradish peroxidase enhanced by poly-L-ornithine, Brain Research, 150 (1978) 170-176. 26 Karnovsky, M. J., A formaldehyde-glutaraldehyde fixative of high osmolarity for use in electron microscopy, J. Cell Biol., 27 (1965) 137A-138A. 27 Katz, B. and Miledi, R., The effect of divalent cations on transmission in the squid giant synapse, Publ. Staz. Zool. Napoli, 37 (1969) 303-310. 28 Kennedy, M. C., Rubinson, K. and Stone, D. J., Retinal projections in larval transforming and adult lamprey, Petromyzon marinus, J. comp. Neurol., 171 (1977) 465--480. 29 Kosareva, A. A., Retinal projections in lamprey (Lampetra fluviatilis), J. Hirnforsch., 21 (1980) 243-256. 30 Kosareva, A. A., Vesselkin, N. P. and Ermakova, T. V., Retinal projections in the lamprey Lampetra fluviatilis, as revealed by horseradish peroxidase transport along the optic nerve, Z. Evol. Biokhim. Fisiol., 13 (1977) 405-407. 31 Kruger, L. and Saporta, S., Axonal transport of [3H]adenosine in visual and somatosensory pathways, Brain Research, 122 (1977) 132-136. 32 Kuypers, H. G. J. M., Catsman-Berrevoets and Padt, R. E., Retrograde axonal transport of fluorescent substauces in the rat's forebrain, Neurosci. Lett., 6 (1977) 127-135. 33 LaVail, J. H. and LaVail, M. M., Retrograde axonal transport in the central nervous system, Science, 176 (1972) 1416-1417. 34 Lindstr6m, T., On the cranial nerves of the cyclostomes with special reference to n. trigeminius, Acta zool., 30 (1949) 315-458. 35 Luiten, P. G. M., Two visual pathways to the~telencephalon in the nurse shark (Ginglymostoma cirratum). I. Retinal projections, J. comp. Neurol., 196 (1981) 531-538. 36 McGill, J. L., Powell, T. P. S. and Cowan, W. M., The retinal representation upon the optic tectum and isthmo-optic nucleus in the pigeon, J. Anat., 100 (1966) 5-33. 37 McGill, J. L., Powell, T. P. S. and Cowan, W. M., The organization of the projection of the centrifugal fibres to the retina in the pigeon, J. Anat., 100 (1966) 34--49. 38 Mesulam, M.-M., The blue reaction product in horseradish peroxidase neurohistochemistry: incubation parameters and visibility, J. Histochem. Cytochem., 24 (1976) 1273-1280. 39 Meyer, D. L. and Ebbesson, S. O. E., Retinofugal and retinopetal connections in the upside-down catfish (Synodontis nigriventris), Cell Tiss. Res., 218 (1981) 389--401. 40 Meyer, D. L., Fiebig, E. and Ebbesson, S. O. E., A note on

41

42

43

44

45

46 47 48

49

50 51 52

53

54

55

56

57

the reciprocal connections between retina and the brain in the puffer fish Tetraodon fluviatilis, Neurosci. Lett., 23 (1981) 111-115. Mfinz, H. and Claas, B., Centrifugal innervation of the retina in cichlid and poecilid fishes. A horseradish peroxidase study, Neurosci. Lett., 22 (1981) 223-226. Mfinz, H., Claas, B., Stumpf, W. E. and Jennes, L., Centrifugal innervation of the retina by luteinizing hormone releasing hormone (LHRN)-immunoreactive telencephalic neurons in Celeostean fishes, Cell. Tiss. Res., 222 (1982) 313-323. N/issel, D. R., Transneuronal labeling with horseradish peroxidase in the visual system of the house fly, Brain Research, 206 (1981) 431-438. N/issel, D. R., Transneuronal uptake of horseradish peroxidase in the central nervous system of dipterous insects, Cell Tiss. Res., 225 (1982) 639-662. Northcutt, R. G. and Przybylski, R. J., Retinal projections in the lamprey Petromyzon marinus L., Anat. Rec., 150 (1973) 400. Ohman, P., Fine structure of the optic nerve of Lampetra fluviatilis (Cyclostomi), Vision Res., 17 (1977) 719-722. Perlia, R., Ueber ein neues opticus Centrum beim HiJhne, Arch. Ophthal. (Von Graefe's), 35 (1889) 20-32. Peyrichoux, J., Weidner, C., Rep6rant, J. and Miceli, D., An experimental study of the visual system of Cyprinid fish using the HRP method, Brain Research, 130 (1977) 531-537. Pickard, G. E. and Silverman, A. J., Direct retinal projections to the hypothalamus piriform cortex and accessory optic nuclei in the golden hamster as demonstrated by a sensitive anterograde horseradish peroxidase technique, J. comp. Neurol., 196 (1981) 155-172. Ram6n y Cajal, S., La r6tine des Vert6br6s, La Cellule, 9 (1893) 119-225. Ram6n y Cajal, S., Histologie du Systdme nerveux de l'Homme et des Vertebras, Maloine, Paris, 19ll. Rep6rant, J., The orthograde transport of horseradish peroxidase in the visual system, Brain Research, 85 (1975) 307-312. Rep6rant, J. and Gallego, A., Fibres centrifuges dans la r6tine humaine, Arch. Anat. micr. Morph. exp., 65 (1976) 103-120. Rep6rant, J., Rio, J. P., Peyrichoux, J. and Weidner, C., Orthograde and retrograde axonal movement of label and transneuronal transport phenomena after intraocular injection of [3H]adenosine in a poikilotherm vertebrate (Vipera aspis), Neurosci. Lett., 16 (1980) 251-255. Rep6rant, J., Peyrichoux, J., Weidner, C., Miceli, D. and Rio, J. P., The centrifugal visual system in Vipera aspis. An experimental study using retrograde axonal transport of HRP and [3H]adenosine, Brain Research, 183 (1980) 435--441. Rep6rant, J., Vesselkin, N. P., Ermakova, T. V., Kenigfest, N. B. and Kosareva, A. A., On anterograde and retrograde transport of [3H]proline in the visual system of the lamprey Lampetra fluviatilis, Z. Evol. Biokhim. Fisiol., 16 (1980) 417-418. Repdrant, J., Vesselkin, N. P., Ermakova, T. V., Keniglest, N. B. and Kosareva, A. A., Radioautographic evidence for both orthograde and retrograde axonal transport of label compounds after intraocular injection of [3H]proline in the lamprey (Larnpetra fluviatilis), Brain Research,

56 200 (1980) 179-183. 58 Rep6rant, J., Vesselkin, N. P., Kenigfest, N. B., Miceli, D. and Rio, J. P., Bidirectional axonal and transcellular transport of [3H]adenosine from the lamprey retina, Brain Research, 246 (1982) 305-310. 59 Rep6rant, J., Vesselkin, N. P., Rio, J. P., Ermakova, T. V., Kenigfest, N. B. and Peyrichoux, J., Evidence for visual centrifugal system in different groups of nonavian vertebrates, Neurosci. Lett., Suppl. 4 (1980) $287. 60 Rep~,rant, J., Vesselkin, N. P., Rio, J. P., Ermakova, T. V., Miceli, D., Peyrichoux, J. and Weidner, C., La voie visuelle centrifuge n'existe-t-elle que chez les Oiseaux? Rev. canad. Biol., 40 (1981) 29-46. 61 Rodieck, R. W., The Vertebrate Retina, Freeman and Co., San Francisco, 1973. 62 Romer, A. S., L'origine des classes de Vert6br6s, La Recherche, 33 (1973) 347-361. 63 Rovainen, C., Neurobiology of lampreys, Physiol. Rev.. 59 (1979) 1007-1077. 64 Scalia, F. and Teitelbaum, I., Absence of efferents to the retina in the frog and toad, Brain Research, 153 (1978) 340-344. 65 Schmidt, J. T., The laminar organization of optic nerve fibres in the tectum of goldfish, Proc. roy Soc., 205 (1979) 287-306. 66 Schober, W., Vergleichend-anatomische Untersuchungen am Gehirn der Larve und adulten Tiere von Lampetra fluviatilis (Linn6 1758) und Larnpetra planeri (Bloch 1784), J. Hirnforsch., 7 (1964) 107-209. 67 Schroeder, D. M., Retinal afferents and efferents of an infrared sensitive snake, Crotalus viridis, J. Morph., 170 (1981) 29-42. 68 Schubert, P. and Kreutzberg, G. W., Axonal transport of adenosine and uridine and transfer to postsynaptic neurons, Brain Research, 76 (1974) 526-530. 69 Schwarz, D. W. F., Schwarz, I. E. and Tomlinson, R. D., Avian efferent vestibular neurons identified by axonal transport of [3H]adenosine and horseradish peroxidase, Brain Research, 155 (1978) 103-107. 70 Shapovalov, A. I., Interneuronal synapses with electrical and chemical mode of transmission and the evolution of the central nervous system, Z. Evol. Biokhim. Fisiol., 13 (1977) 621-633.

71 Shapovalov, A. I. and Shiraev, B. I., Dual mode of junctional transmission at synapses between single primary afferent fibres and motoneurones in the amphibian, J. Physiol. (Lond.), 306 (1980) 1-15. 72 Springer, A. D., Retinopetal cells in the goldfish olfactory bulb, Invest. Ophthal. Vis. Sci., 22, Suppl. (1982)23. 73 Springer, A. D. and Gaffney, J. S., Retinal projections in the goldfish: a study using cobaltous-lysine, J. comp. Neurol., 203 (1981) 401-424. 74 Stell, W. K., The morphological organization of the vertebrate retina. In M.G.F. Fuortes (Ed.), Handbook of Sensory Physiology, VII~2, Springer-Verlag, Berlin, 1972, pp. 111-213. 75 Towe, A. L., Patton, H. D. and Kennedy, T. T., Properties of the pyramidal system in the cat, Exp. Neurol., 8 (1963) 220-238. 76 Triller, A. and Korn, H., Interneuronal transfer of horseradish peroxidase associated with exo/endocytotic activity on adjacent membranes, Exp. Brain Res., 43 (1981) 233-236. 77 Uchiyama, H., Sakamoto, N. and Ito. H., A retinopetal nucleus in the preoptica area in a teleost, Navodon modestus, Brain Research, 222 (1981) 119-124. 78 Vesselkin, N. P., Ermakova, T. V., Rep6rant, J., Kosareva, A. A. and Kenigfest, N. B.. The retinofugal and retinopetal systems in Lampetra fluviatilis. An experimental study using radioautographic and HRP methods, Brain Research, 195 (1980) 453-460. 79 Weidner, C., Miceli, D. and Rep6rant, J., Orthograde axonal and transcellular transport of different fluorescent tracers in the primary visual system of the rat, in preparation. 80 Wise, S. P. and Jones, E. G., Transneuronal or retrograde transport of [3H]adenosine in the rat somatic sensory system, Brain Research, 107 (1976) 127-131. 81 Wise, S. P.. Jones, E. G. and Berman, N., Direction and specificity of axonal and transcellular transport of nucleosides, Brain Research, 139 (1978) 197-217. 82 Wallenberg, A., Das mediale opticus Bfindel der Taube, Neurol. Zbl., 17 (1898) 532-537. 83 Wilczynski, W. and Zakon, H., Transcellular transfer of HRP in the amphibian visual system, Brain Research, 239 (1982) 29-40.