ORIGIN OF MBTAMERIC SEGMENTATION. 79 



theory of the growth of the Vertebrate embryo is untenable. 

 It seems to me that the advocates of that theory have mixed 

 up three distinct embryonic structures, the mesoblastic bands, 

 the primitive streak, and the ridges of the medullary groove. 



The primitive streak is in most forms at first a median 

 structure. I agree vrith the current view as to its nature as a 

 rudiment of the blastopore, and I suggest a reason for its 

 persistence. 



I ought particularly to mention that I regard the Annelid- 

 origin of the Vertebrata and Arthropoda as untenable. This 

 will be obvious to anyone reading the following pages. 



I oflFer no suggestion as to thephylogeny ofMesoblast. I 

 agree entirely with the current view that it has ai'isen from 

 both of the primary layers. 



Mesenchyme is obviously merely precociously developing 

 mesoderm, and is particularly developed in free larvae. 



Finally, I may add that I do not put forward these hypo- 

 theses in a dogmatic spirit, and that I fully recognise that 

 theories dealing with the complicated facts of morphology can 

 only have in most cases a very temporary value. The main 

 idea of the comparisons discussed below first occurred to me 

 some years ago, when investigating the development of the 

 Vertebrate excretory organs; but they have received such 

 striking confirmation from Hatschek's work on Amphioxus, and 

 more recently from a study of the embryo ofPeripatuscapen- 

 sis, that I have at length decided to publish them, hoping that 

 they may at least excite criticism and so lead to the increase 

 of our knowledge, and to the greater definition of our ideas. 



PART I. 



In the discussion which followed the communication of the 

 late Professor Balfour's notes and drawings of the early 

 embryos of Peripatus Capensis, to the Eoyal Society 

 (December, 1883), I drew attention to the great resemblance 

 between the embryo of P. Capensis with its elongated blasto- 



