604 HAROLD HEATH, 



determined. At a later period the anterior arms of the cross divide 

 but how their cleavages compare with the posterior I am unable to say. 

 Whether this rosette series forms the cerebral ganglia cannot be 

 decided definitely but there are two or three facts which lead one to 

 believe such may be the case. In the first place the slight in- 

 dentations over the place where the cerebral ganglia are forming, 

 which stain intensely black with osmic acid, appear where the rosette 

 series disappeared, and their outer extremities are some distance 

 removed from the velum , as was the case with the tip of the arms 

 when the rosette series was last seen. Certainly it can scarcely be 

 questioned that the inner ends of the ganglia in contact with the 

 apical sense organ arise from this group of cells which occupies the 

 exact centre of the velar field. Also the rosette series shows throughout 

 its development a remarkable independence ; its cell cleavages in no 

 way appear closely correlated with those of the arms of the Molluscan 

 cross, and as far as it can be traced it exhibits a quadriradial sym- 

 metry in its outline. And it is certainly a significant fact that ap- 

 pearing exactly as do the cerebral ganglia are two smaller areas 

 posterior to the sense organ and connected with it. I have examined 

 hundreds of embryos at this stage and this character is almost always 

 present. I feel positive that small masses of cells proliferate in these 

 areas which uniting with the inner ends of the cerebral ganglia lie 

 beneath and in contact with the apical sense organ. In rare cases sur- 

 face views show the two posterior tracts extending to a greater distance 

 and in three embryos I have found them to extend to as great a 

 distance as the cerebral ganglia. They were slightly depressed and 

 lacked the perfect sharpness of outline characteristic of the cerebral 

 ganglia but otherwise exhibited much the same appearance as these 

 latter structures. 



The thought suggests itself that the rosette series with its 

 independence of development and radial form is to be 

 considered as the fundament of a quadriradial nervous 

 system which in the primitive radial ancestor develop- 

 ed in all quadrants but since has partially degene- 

 rated owing to the change from a radial to a bilateral 

 form. Nervous systems similar to this are found in the Polygordius 

 trochophore and the Ctenophores but the manner in which these 

 develop is unknown, yet from the behavior of the rosette series and 

 the development of the nervous system I am led very strongly to the 



