218 BULLETIN OF THE 



On the contrary, the lumen is already potentially present, as shown by 

 the arrangement of the nuclei before any actual separation of the walls 

 occurs. 1 am of opinion that, in the cases referred to, the separation is 

 largely artificial, and that the ruptures take place most frequently at the 

 nephrostomes for the reason that the walls, which elsewhere form a closed 

 ring, here have in cross section the form of a sharp re-entrant angle bor- 

 dering on a large open space. It is evident that in the former region the 

 walls would be less liable to be torn apart in the preparation of the ma- 

 terial than in the latter. In general, however, it must be admitted that 

 the development of the lumen, like that of the system as a whole, 

 actually advances from anterior to posterior regions. 



The fundaments of the three pronephric tubules shown in Figm-e 11 

 are not to be regarded as outgrowths from the somites. They are, it is 

 true, very closely related to the segments in their arrangement, but, as 

 transverse sections prove (Plate I. Fig. 6, and Plate II. Fig. 15), they 

 lie wholly ventral to the lower boundaries of the protovertcbrse. The 

 frontal section figured (Fig. 12) was chosen for the reason that it was 

 the one which indicated most precisely the course of the fundaments of 

 the three tubules. The plane of the section is parallel to a tangent 

 to the dorsal margin of the structure, and passes only a little below that 

 margin, not through the nephrostomes. These begin in a more ventral 

 unsegmented region. 



In the oldest embryos of this stage, the fundament of the duct has 

 developed very rapidly. Anteriorly, it has in cross section a distinctly 

 elliptical outline, and its cells have, with reference to the major axis of 

 the ellipse, the same arrangement that I have described for the inter- 

 segmental regions of the pronephric pouch. On following the structure 

 backwards, this distribution becomes less and less obvious, until the 

 cells seem to have no definite arrangement. In this region the funda- 

 ment of the duct is in far more intimate union with the somatopleure 

 than was the case in anterior somites. In the region of somite IX. 

 the last trace of the structure is to be seen as a simple thickening of 

 the somatopleure, similar in form to that described and figured in the 

 youngest embryos of this stage (Fig. 17), for a region just back of 

 somite VI. The region in which the duct is formed is throughout im- 

 mediately ventral to the constriction separating the protovertebrse from 

 the lateral plates.-* 



1 In sections from tbe posterior end of tlie embryo, it is necessary to guard 

 against the false appearances which arise from the obhquity of the plane of the 



