2 6 COMPARATIVE PHYSIOLOGY OF THE BRAIN 



strips of a mammalian heart to beat. He also draws 

 from his observations the conclusion that there 

 is no reason for assuming the existence of a centre of 

 a coordination (6). In Medusae, also, a synchronical 

 contraction of all the parts takes place if the stimulus 

 from the portion first active can travel rapidly enough 

 to the rest of the margin. This is only possible when 

 the margin is uninjured. It is evident, however, that 

 the neighbouring tissue as well as the nerve-ring is in- 

 volved, because the radial incision must reach well 

 toward the centre of the bell if we wish to stop the 

 coordination. In this case the wave of stimulation 

 must pass around the incisions, a process which in- 

 volves so much time that the separate parts are able 

 to contract independently, and the synchronism is 

 lost. In injured or dying Medusae, where the contact 

 of the cells is less close, uncoordinated, rhythmical 

 activity occurs. 



In order to test this idea further, I proposed to 

 Dr. Hargitt, who was working in my laboratory, that 

 he attempt to graft two Hydromedusae, and observe 

 whether they continue to contract synchronically or 

 independently after healing. For this purpose it was 

 necessary to remove the margin of the Medusae. 

 Two of them were then placed with their wounded 

 surfaces in contact, and kept in this position. Figure 

 5 shows two Gonionemi grafted in this way. They 

 grew together along the entire line of contact with the 

 exception of a small part at O. New tentacles would 

 probably have developed there in time had we not 



