328 JOHNS HOPKINS UNIVERSITY MORPHOLOGICAL MONOGRAPHS. 



running obliquely downward and backward toward the dorsal wall of 

 the cloaca (Plate LVII, Figs. 1 and 2). The walls of the invagination are 

 pressed together so that the whole structure appears to be merely a 

 double fold of ectoderm with no lumen between. Fig. 4, a diagrammatic 

 view from the dorsal surface, shows the general outlines. The double 

 fold of ectoderm is crescent-shaped, the horns of the crescent projecting 

 forward on each side of the ganglion a little beyond its middle point. 

 It is wider than the ganglion and a little longer antero-posteriorly. 

 Figs. 1 and 2, longitudinal and cross-sections respectively, show that the 

 ectodermal fold runs down toward the wall of the cloaca but does not 

 quite reach it. It approaches the cloaca and not the pharynx, since its 

 posterior end is behind the junction of the gill and the dorsal wall. 

 In histological character the cells of the invagination resemble the 

 ectodermal epithelium. This structure arises as a mere pit in the ecto- 

 derm, following the lateral and posterior boundaries of the ganglion 

 (Plate LVII, Fig. 3). In the earliest stages, when the ganglion is just 

 commencing to enlarge, the dorsal ectoderm is an even surface. As the 

 ganglion grows it pushes the ectoderm upward, forming a rounded 

 hillock. In later stages of development the ectodermal epithelium still 

 clings to the whole dorsal surface of the ganglion. That portion of the 

 ectoderm behind the brain, however, grows rapidly and pushes up over 

 the brain, overlapping from behind forward the epithelial covering of 

 the brain. In this way there is formed the double fold of ectoderm 

 described. It arises, then, not as a direct invagination, but by an over- 

 growth. I do not know its fate, since I have had no opportunity to study 

 the later stages of development or the adult of the solitary form. I am 

 unable to see that this structure has any morphological or physiological 

 significance. It does not seem to be worthy of the name organ, though 

 possibly, if present in the adult, it may serve as a slight protection for the 

 ganglion. 1 



SALPA COSTATA-TILLENII. 



In the chain form of this species (Plate LVI, Fig. 2), which shows a 

 still less definite structure of the rod cells in the larger eye, there are 



' Oka (13) describes peculiar ectodermal proliferations on each side of the hypo- 

 physical tube in the buds of Botryllus, which have about the same position and appear 

 at about the same stage of development as does the ectodermal invagination described 

 in the embryo of Salpa hexagona. Neither structure is sufficiently well understood 

 to justify a comparison. 



