MORPHOLOGY OF THE VEKXEBRATE OLFACTORV ORGAN. 335 



words, whether the visceral clefts are formed between succes- 

 sive segmeniSj or through the middle of the segments- 



Though the former of these views is usually assumed to be 

 the true one, yet there are, I think, considerations of some 

 weight in favour of the opposite view. In the first place, we 

 must bear in mind that the original proto -vertebral segmen- 

 tation does not extend to the head, and that the secondary, 

 visceral cleft segmentation appears late, and differs totally 

 from the segmentation of the body, inasmuch as, instead of 

 starting on the dorsal side in the mesoblastic tissue on either 

 side of the neural axis, it arises in, and is limited to, the 

 lateral and ventral walls of the alimentary canal. Instead of 

 arising primarily in the mesoblast, it is a segmentation in 

 which the mesoblast takes no share whatever, except a 

 purely passive one; it is a segmentation produced by ilie 

 growth of diverticula from the alimentary canal, which come 

 in contax't with, and fuse with the external epiblast, and, 

 finally, by perforation of this latter open on to the exterior. 

 Now, It is more in accordance with what we know of the 

 occurrence of lateral diverticula of the alimentary canal iu 

 Inveitebrata, that these should be segmental rather than 

 intersegmental. 



Again, while each cranial segmental nerve supplies two 

 visceral arches, it only supplies one cleft; and, from the 

 analogy of Invertebrata, we should expect that the distribu- 

 tion of each nerve would be to its own sejjment ; while it 

 certainly would be a \ery remarkable fact that each seg- 

 mental nerve should supply adjacent halves of two seg- 

 ments. 



The distribution of the branchial arteries may also be 

 cited as additional evidence in the same direction ; the 

 corresponding vessels in Invertebrata do not occupy the middle 

 of eacli segment, but follow the intersegmental septa, which 

 septa, according to the view here advocated, would occupy 

 the centres of the vertebrate visceral arches, and so corre- 

 spond in position with the branchial arteries. On the same 

 view the skeletal elements of the visceral arches would also 

 correspond in position to the intersegmental septa, from which, 

 indeed, they may conceivably have been derived. 



This view acquires some additional interest in connection 

 with Ur. Dohrn^s suggestion that the visceral clefts are 

 homologues of segmental organs.^ 



The theory above propounded as to the morphology of the 

 vertebrate head, will, I venture to think, throw some liglit 

 on the nature of the skeletal elements of the head. I will 

 1 Op. cit., pp. 10, 11. 



