28 WILLIAM F. ALLEN 



recorded for tlie pig embryos. In view of the facts, this expla- 

 nation of the formation of the roof expansion seems more ten- 

 able to the writer than to attribute all, as His has done, to the 

 action of a pontine flexure. For beyond question, a considerable 

 expansion of the roof plate in the human medulla takes place 

 before the pontine flexure appears. 



2. Description of three roof plate expansions of the spinal cord in 



the 20 cm. Polistotrema series 



An interesting variation (abnormality) was found in the spinal 

 cord of a single specimen of Polistotrema (Bdellostoma), other- 

 wise unusual. Certain structures appeared in the roof of the 

 cord that were very similar to the tela chorioidea of the fourth 

 ventricle, and will be described because of the light that they 

 may throw on the origin of the tela chorioidea of the fourth 

 ventricle. 



What has been designated as the first roof plate expansion of 

 the 20 cm. Polistotrema series is shown in the photographs of 

 models 1 and 2 (figs. 1 and 2, 4 and 5, R.Ex.) to be an immense 

 outcropping of the roof plate ependjana. Most unfortunately 

 the anterior portion of this specimen, from which the series 

 through the tail was taken, has been lost, so that it is impossible 

 to state how much further cephalad this expansion of the roof 

 plate extended, or whether there were other outcroppings of 

 the roof plate in front of it as there are behind. It appears in 

 the first model (fig. 2, R.Ex.) and in the first transverse section 

 (fig. 10, R.P.Ex.) as a median mass, covering about one-half of 

 the dorsal surface of the spinal cord; it then shifts gradually 

 over to the right side (figs. 2 and 11); then gradually attains a 

 median position. In this position it continues as far as the 

 middle of model 2 (fig. 5), covering a large part of the dorsal 

 surface of the spinal cord. 



For the most part the roof plate expansion contains a cavity 

 of considerable size, which is shown anteriorly in the cast (fig. 



3, C.C.Ex.) and in transverse section (figs. 10 to 12) to be in 

 direct communication below with the central canal. Posteriorly 



