M. M. METCALF ON THE EYES AND SUBNEURAL GLAND OF SALPA. 351 



flattened dorso-ventrally, forming a narrow slit separating the ventral 

 third of the brain from the dorsal two-thirds. As the cells of the dorsal 

 wall of the neural canal push up to form the dorsal part of the ganglion 

 they do not at once make a solid mass, but they leave between them- 

 selves fine, anastomosing, lucanar canals (Plate XLVII, Figs. 1 to 3, Z), 

 some of which open inward to the neural canal /, while others open 

 outward to the cavity around the brain, I'. These lacunae are not mere 

 chance spaces, but have a quite definite appearance and persist for some 

 time. In cross -section (Figs. 1 to 3, Plate XLVII) the lacunae I are seen 

 to be separated for the most part by single rows of cells resembling the 

 liver cells between the bile ducts of higher vertebrates. Soon the brain 

 becomes solid, the lacunae and also the central canal disappearing. 

 When this change is nearly completed (Fig. 5, Plate XLVII), we see a 

 condition almost exactly resembling that found in a nearly mature 

 doliolum. There is a slight antero-ventral process (shown in Fig. 5, 

 Plate XLVII, in longitudinal section, and in Fig. 6, v, in cross-section) 

 containing a remnant of the disappearing brain cavity, /. This is still 

 connected by a very fine canal with the wide lumen of the funnel. The 

 lumen of the canal does not show in the section drawn. In the later 

 development the brain cavity and the duct wholly disappears, and the 

 lumen of the funnel is the only remnant of the neural canal. The duct 

 disintegrates as if pulled apart by the elongation of that region of the 

 body, reminding one of the elongation of the duct in doliolum, by 

 which the cells are pulled out into an interrupted pavement epithe- 

 lium. Neither in the half-developed chain individual nor in the adult 

 can any trace of a duct opening into the funnel be found. 



Certain other structures develop later that seem to bear the closest 

 relation to the lateral communications between the subneural gland and 

 the peribranchial chamber, described for phallusia mammillata. At a 

 time when the embryonic eye-disk has taken its perpendicular position 

 (see the section on the development of the eyes of Cyclosalpa pinnata, 

 chain form) the wall of the peribranchial chamber, 1 which up to this time 

 has been close to the ventral surface of the brain (Figs. 1 to 3, Plate 

 XLVII), begins to separate from the brain. Two small areas remain 



1 Professor Brooks tells me that his study of the budding of salpa shows beyond 

 question that the dorsal part of the pharynx is at this time still quite wide, so that the 

 whole region beneath the ganglion and for a considerable distance laterally belongs 

 to the branchial chamber. The peribranchial chamber occupies anteriorly a very 

 narrow region on each side of the body. 



