Development o/'Pyrosoma. 243 



I was not able to discover any connexion between the axial 

 canal of the mesoderm and the enteric cavity, nevertheless 

 the position of this canal renders it extremely probable that 

 we have in it the nearest representative of the chorda dorsalis 

 of the Ascidian embryos. The beautiful investigations of 

 E. van Beneden and Julin have taught us that the notochord 

 of the embryos of GlaveUina is, in the earlier stages of deve- 

 lopment, represented by a tube which is situated between the 

 two mesodermic diverticula, and therefore exhibits the same 

 relations as we find in the axial tube of the embryos of 

 Ptjrosoma. In the case of Pyrosoma the tube in question is 

 a transitory structure and lasts but a very short time. 



In spite of the similarity between the coelomic sacs of 

 Pyrosovia and those of the Ascidian embryos, to which I have 

 just alluded, the two structures nevertheless exhibit an 

 important difference in their histology. This consists in the 

 fact that, while the coelomic sacs of the Pj/rc>so?>ea-embryos 

 are bounded by a multilaminar tissue, those of the Ascidian 

 embryos have unilaminar epithelium-like walls. In the later 

 stages of the Pyrosoma-embryos, however, this difference is 

 removed ; for, in the course of growth, the coelomic sacs 

 likewise become bounded by a single layer of cells. The two 

 sacs, right and left, are at first precisely similar, and are 

 symmetrically placed with regard to the longitudinal axis of 

 the germinal disk. This condition, however, is of but short 

 duration. As early as the stage of the first appearance of the 

 peribranchial tubes, the two sacs exhibit important differences 

 from one another ; with this a second period in the develop- 

 ment of the coelomic sacs is inaugurated, which may be termed 

 the metamorphosis of the coelomic sacs. While the left coelomic 

 sac has greatly increased in size and forms a spacious cavity, 

 the cavity of the right sac has almost entirely disappeared 

 and appears as a tiny lacuna which adjoins tiie septum 

 between the two primitive coelomic spaces. The entire distal 

 portion of the right coelomic sac is now represented by a 

 solid mass of cells, from tlie periphery of which some cells 

 are in the act of being liberated. In the stage at which the 

 peribranchial canals deepen and form little blind tubes, the 

 whole of the right coelomic sac has completely disappeared 

 and has broken up into little cells. The development of the 

 left coelomic sac, on the contrary, rapidly proceeds. The 

 circumference of the sac increases, and it subsequently divides 

 into two portions : its proximal portion becomes incrassated, 

 and forms a swelling situated beneath the endostyle which is 

 in process of formation ; ia its further development it plays 

 a very important part in the formation of various mesodermal 



