42 JOHN SZENTAGOTHAI 



problem. Sensory ganglia and/or spinal cord or medulla oblongata fragments 

 of the larval newt are deplanted into the dorsal fin of host larvae of the same 

 species {Triturus cristatus, and Pleurodeles waltii). An additional limb trans- 

 planted nearby serves as indicator of the activity of the isolated model centre. 

 When the preparation begins to function, the host's own spinal cord (i.e. the 

 part underlying the implants) is destroyed in order to exclude interference 

 from the host's own nervous system. In a number of such preparations func- 

 tion and histological structure have thoroughly been compared (Szekely and 

 Szentagothai, 1961), and the results can be briefly put as follows. 



The muscles of an implanted limb can be innervated successfully only by 

 prospective motoneurons (Weiss, 1950a). If the preparation is built up of 

 medullary neurons only, exclusively irregular spontaneous activity is ex- 

 perienced, which cannot be influenced by physiological stimuli. Reflex activity 

 with responsiveness to natural stimuli (touch) can be achieved if primary 

 sensory neurons (spinal or vagus) are incorporated into ihe preparation. Both 

 types of activity can be seen in preparations built up of from ten to twenty 

 neurons connected in completely haphazard manner, and without the slightest 

 indication of any organoid character. Terminal knobs are often seen in 

 contact with neurons of medullary origin if primary sensory neurons are 

 present (Figs. 13 and 14). Their preterminal fibres can sometimes directly 

 be traced back to their origin from sensory neurons. Spontaneous activity 

 occurs in preparations lacking terminal knob synapses, and they seem not to 

 be necessary even for true reflexes. 



Placing smafl pieces of filter paper soaked with strychnine on the skin of the 

 dorsal fin covering the deplanted centre greatly enhances spontaneous as well 

 as reflex activity in most of the preparations, in some of them no effect, 

 however, is noticed. No strychnine effect is generally seen during the first few 

 days of developing function, and it disappears again at the onset of meta- 

 morphosis, when due to breakdown of circulation in the fin the preparation 

 begins to degenerate. Spontaneous or reflex activity lasts generally for a week 

 or two longer. The strychnine effect is completely independent from the pre- 

 sence or lack of sensory neurons. 



Advantage has been taken here of the fact that neurons have early deter- 

 mined specific characters, which enable them to establish effective contacts 

 with certain other types of neurons and perform certain types of functions. 

 These specific properties are retained in spite of considerable disarrangement 

 of nervous structures and complete changes of environmental conditions. 

 They are generally thought to be rooted in the biochemical constitution of the 

 different neuron types (Weiss, 1950b; Sperry, 1955). Since there is reason 

 enough to consider strychnine as specifically blocking the transmission from 

 inhibitory neurons to motoneurons (Coombs et ah, 1955; Eccles, 1957), 

 some interesting conclusions can be drawn from observation of such simplified 

 "model nervous systems": (i) arrangement of neurons, cell processes, and 



