STUDIES ON THE TURBELLARIA. A451 
plasm is vertically fibrillated. Hach cell is perforated by the 
ducts of a number of rhabdite-forming glands. When these 
are empty they are clearly to be recognised in stained 
sections as perforations passing through the cell; when 
secretion has been passing through them at the time of 
fixation it becomes strongly stained, and the effect in hori- 
zontal sections is the appearance of a number of dark spots 
in the body of the cell, and in vertical sections of a number 
of narrow, vertical, deeply-stained strands passing right 
through the cell, and sometimes projecting slightly on the 
surface. A thin layer on the surface of each cell marked 
by innumerable fine vertical lines, corresponding apparently 
to the bases of the cilia, is probably of the nature of a cuticle, 
but it is not very sharply differentiated, and there are no 
indications that it is capable of being detached from the 
underlying layer. The epidermis is supported on a distinct, 
but thin, basement-membrane. 
The system of channels above alluded to is not equally 
distinct in all specimens, owing to the channels being 
sometimes more, sometimes less, dilated. In a favourable 
specimen the entire dorsal surface, and, in some cases, the 
ventral also, of the living worm is covered with a network of 
clear bands of varying width. When the surface is examined 
under a moderately high power of the microscope, it is found 
that these bands lie in the intervals between the cells of the 
epidermis, and represent branching vessels of varying calibre. 
In one or two of the specimens which I fixed and mounted 
whole these vessels are almost as distinct as they were in the 
living condition. In fig. 29 I have represented a portion of 
this network as seen in a surface view of one of these entire 
preserved specimens. Hxamination of the living specimen 
showed clearly that the appearance in question is not due to 
the presence of open clefts between the epidermal cells. In 
an uninjured animal no such clefts exist, the ciliated surface 
being smoothly continued without interruption from one cell 
to another across the clear interval. But in the majority of 
my series of sections, in cases in which the system is well 
