50 

 no distinct walls, but are merely excavations in the funda- 



FiG. 2. 



Partly injected portion of the cord. 



(a.) Network of connective tissue cells, plainly 

 continuous with (b) darkly shaded portions, 

 representing channels filled with the in- 

 jection. 



Magnified 270 diameters. 



mental substance, seems especially inapplicable in the case 

 of a tissue where this substance is semifluid. Considering 

 the arrangement of the nuclei (mentioned above), which are 

 evidently imbedded in the Avails, the author thinks no other 

 hypothesis tenable than that the connective tissue cells them- 

 selves constitute the walls of the hollow spaces, which usually 

 go by that name ; and since there is no distinction between 

 the walls of these spaces and those of the channels which 

 connect them, the whole system must be enclosed in the same 

 way. The connective tissue cells would then be membranous 

 or epithelial nucleated structures which roll themselves up 

 to form a tube, just as some observers (Aeby and Eberth) 

 have supposed that the blood-capillaries are formed by ad- 

 herent e]3ithelial cells, only that for the short channels of the 

 connective tissue network a single cell suffices. The whole 

 system is thus, strictly speaking, intercellular, not intracel- 

 lular, as supposed by Virchow. 



Dr. Roster's memoir concludes with observations on the 

 epithelium of the umbilical cord. These coincide generally 

 with those of other observers, but he has observed some 

 peculiar structures among the cells of the lowest epithelial 

 layer. These are round or oval, sharply defined openings 

 {stomata) which are unequally distributed. They generally 

 contain a cell with granular protoplasma resembling the 

 contractile (migratory) cells contained in the plasmatic 

 channels of the cord. Small drops of mucus are also met 

 with in the stomata, and after injection they are often choked 

 with plugs of the same material. A still more complete in- 



