38 JOHN SZENTAGOTHAI 



crista having direct connexions only with two muscles; one of each eye 

 (Szentagothai, 1943, 1950, 1952). These connexions can be brought easily 

 into action by artificial endolymph currents evoked in any one of the semi- 

 circular ducts. Reciprocal inhibition in tliis reflex has a longer latency than 

 excitatory responses, and has apparently quite diff'erent pathways. If the 

 medial longitudinal fasciculus is transected between the vestibular and the 

 oculomotor nuclei bisynaptic excitatory responses are abolished, but reciprocal 

 inhibitions are preserved. If on the other hand the whole pons with the 

 exception of the medial longitudinal fascicle is transected the reverse happens, 

 i.e. all bisynaptic excitatory responses are preserved and reciprocal inhibitory 

 responses are abolished (Szentagothai, 1950, 1952). 



Tracing these pathways with degeneration methods shows that while 

 secondary vestibular neurons ascending in the medial longitudinal fascicle 

 reach the oculomotor neurons directly and establish synaptic contacts with 

 unusually large terminal knobs, secondary vestibular fibres ascending outside 

 the fasciculus in the reticular formation have no direct access to oculomotor 

 neurons. They terminate, however, with abundant synapses in the region of 

 the Darkschewitsch nucleus and generally in the central grey matter of the 

 midbrain just dorsally and cranially from the oculomotor nucleus. We have 

 been able (Szentagothai and Schab, 1956) to show that electric stimulation of 

 this region immediately abolishes the eff'ects of even supramaximal labyrinth 

 stimulation by artificial endolymph currents on oculomotor neurons (Fig. 9). 

 However crude this experimental approach, the results suggest as the most 

 simple explanation, that the simplest pathways in the vestibub-oculomotor 

 reflex are separated anatomically and the inhibitory pathway has an additional 

 short specific inhibitory neuron localized in the anterior midbrain central 

 grey matter. This situation would be analogous to the pathway of reciprocal 

 inhibition suggested by Eccles et al. (1954b). The situation is still more favour- 

 able, since electrolytic lesions can be placed into the Darkschewitsch nucleus 

 and the anterior central grey matter of the mid-brain without danger to 

 destroy any other pathway or fibre system leading to/through the oculomotor 

 nuclei, an experiment which cannot be performed of course in the spinal cord. 



Now again the degeneration findings are rather meagre. Akhough placing 

 lesions not farther from the border of the oculomotor nucleus than 200 n not 

 a single degenerated terminal knob could be found and only extremely fine 

 fibres around oculomotor neurons showed signs of degeneration (Fig. 5). 

 They were almost on the border of visibility with the best immersion systems, 

 and their branches can be assumed to be beyond their resolving power. The 

 resemblance between these fine degenerating fibres originating from a region 

 where specific inhibitory neurons must be assumed and the meshwork of the 

 fibres remaining intact in our isolated ventral horn preparations is striking. 

 Both observations suggest that inhibition might be exerted on motoneurons 

 by means of a submicroscopic type of synapse. 



