34G JOHN WARREN 



seen only at early stages. He does not believe in the ones de- 

 scribed by the above writers. Neal agrees with him and states 

 that the so-called fore brain and mid brain neuromeres are not 

 true neuromeres owing to their shape, structure and late time of 

 appearance. He prefers to homologize the fore and mid brain 

 vesicles with hind brain neuromeres and to regard these segments 

 as secondary subdivisions. Hill found that the five neuromeres 

 of the mid and fore brains disappeared early before the pri- 

 mary fore and mid brain vesicles were developed. The future 

 subdivisions of these he does not consider true neuromeres on 

 account of their time of appearance and dorsal position. 



As regards the structure of the segments shown in the models 

 and figures I agree with Orr that they do not conform to regular 

 hind brain neuromeres, though there is some resemblance. The 

 diencephalic segments are certainly much better marked in the 

 dorsal than in the ventral zone, especially the synencephalic seg- 

 ment and they appear much later than the neuromeres of the hind 

 brain. From the evidence it seems clear that there is a primary 

 set of neuromeres of which the majority of writers assign three to 

 the fore brain and two to the mid brain. These apparently dis- 

 appear when the mid and fore brain vesicles develop. The fore 

 brain vesicle then divides later into the telencephalic segment and 

 two diencephalic segments. This is clearly shown by the models. 

 It is conceivable that, if the fore brain was primarily subdivided 

 in three parts and that the lines of demarcation were later obliter- 

 ated by modifications in the shape of the neural tube, that a sec- 

 ondary subdivision occurring here would tend to follow the lines 

 of the primary subdivision, as the limits between neuromeres 

 might be considered in a way as lines of least resistance. The 

 relative shape and size of the secondary subdivisions or seg- 

 ments would naturally be modified and distorted by the unequal 

 growth of the parts. This of course is merely a supposition and 

 I do not think that a direct connection between the primary and 

 secondary subdivisions has j^et been clearly proved. I am inclined 

 to agree with von Kupffer's conclusions (57), p. 166,when he says 

 that there are six neuromeres behind and five neuromeres in front 

 of the fissura rhombo-mesencephalica in the region of fore and 



