THE DEVELOPMENT OF THE EARTH-WORM. 243 



migrate from the ventral surface. Then a fissure appears 

 between the cells of the median septum ; at first it is con- 

 fined to each segment, but later extends the whole length 

 of the nervous chain ; this is the ventral fissure of the 

 subintestinal nerve cord of the adult. 



My researches have fully confirmed Kowalewsky's im- 

 portant discovery that the subintestinal nervous chain of 

 the annelids arises solely from the ectoderm. Semper^s state- 

 ment that it is made up of an unpaired median thickening 

 of the ectoderm^ comparable to the medullary groove of ver- 

 tebrates, and of two cords of mesoderm, corresponding to 

 the spinal ganglia, is definitely contradicted by the develop- 

 ment of that apparatus in the Ltimhricini. And further, 

 this dogma, which had for its object the reconciliation of the 

 differences in the structure and development of the nervous 

 system, observed in annelids on the one hand, and verte- 

 brates on the other, has missed the mark, since we have 

 learned from the excellent researches of Balfour that the 

 spinal ganglia of vertebrates are not derived from the 

 mesoderm. 



Semper has already found an opponent in Hatschek, Avho 

 upholds for Lumbricus the origin of the entire ganglionic 

 chain from ectoderm. But, beyond this, his own not very 

 clear descriptions appear to me to be erroneous. We have 

 already noticed that, according to this author, the ganglionic 

 chain is formed from two prolongations of the cephalic 

 medullary plate; that is to say, only the lateral parts of it, 

 parts, which he calls lateral cords, since a medullary groove, 

 very similar to that of the vertebrates, is then formed between 

 them. It is true that in the development of the medullary 

 plate a fissure, sometimes a very deep one, is seen (or), rather 

 there are two, diff'ering in time of appearance and in mode of 

 formation. To my view Hatschek's figs. 2, 3 and 4 ^ would 

 represent the first. But here certainly is not a case of 

 invagination ; the groove besides is only the space which 

 from the beginning separated the primitive cords, deepened 

 in consequence of the great thickening of their neighbour- 

 ing sides. With the development of the medullary plate 

 this groove disappears. On the other hand, the fissure in 

 fig. 6 cannot but be the second, whose formation we have 

 described above, and Avhich accordingly has nothing to do with 

 the first. But what I do not know how to explain is the 

 fact that Hatschek represents the walls of the fissure as 

 being very obviously separated from the lateral cells of the 

 plate ; in Lumlricus trapezoides there is not the least trace 



^ Ijoc. cit. 



