29 



que, section of this is produced, and at the same time the wall of the 

 stomach posterior to the entrance of the oesophagus is cut as the figure 

 represents. The obliquity of the section causes the oesophageal band 

 to appear somewhat thicker than it is in reality, though not greatly so. 



I think it highly probable that the oesophageal band is, func- 

 tionally at least , an endostyle , meaning by this that it performs the 

 same office as the similarly situated and similarly named organ in 

 Tunicates. That it is a ridge rather than a groove as the endostyle in 

 both Tunicates and Amphioxus is, does not, I think, present any se- 

 rious objection against supposing, it to have such a function. If such 

 be its nature, it is certainly in a very primitive stage, and embryology 

 makes us familiar with many instances where a local thickening of an 

 epithelium is an initial step to the formation of a pit or furrow. It will 

 be noted that I have said that functionally this ridge may be an 

 endostyle. Whether it be homologous with this organ in the Chor- 

 data, that is quite another matter. In view of the several remarkable 

 chordate characters possessed by Balanoglossus, the possibility of such 

 an homology inevitably suggests itself. A decided opinion on the 

 subject would certainly be premature till the fate of the ridge after 

 metamorphosis, is more fully studied. There certainly does exist an 

 area of modified epithelial cells on the floor of the branchial section 

 of the digestive tube of the young Balanoglossus — this attracted my 

 attention before 1 had made any sections of the Tornar ia. However, I 

 have not stages enough to enable me to obtain definite knowledge as 

 to the relation of these to the ridge in the larval oesophagus. 



This suggestion of the presence of an endostyle in Balanoglossus 

 is not a new one. Bateson 1885 has described and figured a groove 

 in the position where the organ would be looked for in the young B. 

 Kowalevskii, though he seems not to consider it as having such a sig- 

 nificance. More recently, however, Schimkewitsch 1888 has ob- 

 served a similar groove in B. Mereschkovskii ^ which he regards as 

 homologous with the endostyle of vertebrates. I reproduce his words : 

 »Der Kiementheil zerfällt in zwei Theile , und zwar einen oberen mit 

 der Epibranchialleiste und einen unteren, der die Form einer kleinen 

 Rinne hat, deren Boden mit Papillen besetzt ist. Diese Rinne ist dem 

 Diverticulum des Kragentheiles ähnlich und beide können als Homo- 

 logon des Endostyles, der Hypobranchialrinne und der Schilddrüse 

 der Cyclostomen angesehen werden.« 



The only other point that I touch upon in this communication is 

 the development of the dorsal nerve cord. This is not accomplished 

 by a delamination of an inner portion of the ectoderm along the mid- 

 dorsal line, as described by Bateson foi B. Kowalevskii , but by a 



