Johnston, Morphology of the Head. 189 



extends only to the dorsal mesoderm, the segmentation of the 

 lateral mesoderm is determined or dominated by the arrange- 

 ment of the gills. Aside from the fact that this would scarcely 

 be expected when we consider the character of invertebrate 

 segmentation, the implications are inconsistent with the facts 

 of segmentation in Amphioxus. Here the whole mesoderm is 

 at first segmented and the lateral segments secondarily fuse 

 into continuous lateral plates. A reminiscence of this is pre- 

 served, as we have seen above, in the head mesoderm of cyclo- 

 stomes, selachians, and amphibia. If this be taken to indicate 

 the primitive condition in vertebrates, then the idea of an inde- 

 pendent branchial segmentation is hard to accept. It seems 

 much more probable that the position of the entodermal gill 

 pouches and of the ectodermal furrows was at first determined 

 by the law of least resistance, the clefts between mesodermic 

 segments offering favorable places for the breaking through of 

 gill slits. The fact should not be lost sight of that in larval 

 Amphioxus the branchial segmentation agrees with the meso- 

 dermic. 



Fortunately we have in the work of Koltzoff on Petro- 

 myzon direct evidence bearing on this question. The primitive 

 character of Petromyzon gives greater weight to this evidence. 

 The mandibular arch is connected with somite 2 and the hyoid 

 arch corresponds exactly to somite 3. The third, fourth and fifth 

 visceral sacs on account of the thickness of the gut wall, are longer 

 than the corresponding somites so that when five visceral sacs 

 are formed the fifth lies beneath somite 9. While further vis- 

 ceral sacs are forming the gut wall is thinning out and the vis- 

 ceral sacs occupy less space. No visceral sacs are formed be- 

 hind somite 9, but when the whole number, eight, are formed 

 they occupy the space of eight somites. Immediately behind 

 them somite (myotome) 10 sends down a ventral bud to enter 

 the subbranchial musculature. Then at a later time the devel- 

 opment and expansion of the gills carry them back until the 

 last sac lies beneath somites 15-16, and by this movement the 

 subbranchial muscle buds are cut apart from the myotomes 10, 

 II, etc. A complete correspondence between the branchial 



