116 SPOLIA ZEYLANICA. 
secondary marginal glands, as they may be termed, appear to have 
several characters which are intermediate between those of the 
erythrophile glands and the true marginal glands; for instance, 
they are larger than the first, but much smaller than the second ; 
they are stained less intensely with eosin than the first, and are also 
stained very lightly with hematoxylin. The erythrophile granules 
are more minute than those of the true erythrophile gland. As fig. 8 
represents, these secondary marginal glands form a thin layer (two 
or three cells thick in section) immediately below the glandular 
ridge. The two zones in which both sorts of marginal glands open 
to the exterior are separated from each other by a narrow area of 
the epidermis, which is characterized by containing a few closely- 
set rhabdites (see Plate IV., fig. 8, rd). There is a similar distinct 
layer of rhabdites more ventral to that just mentioned (see Plate IV., 
fig. 8,7d). Between the two rhabdite layers is interposed a narrow 
clear space of the epidermis (see Plate IV., fig. 6 and 8, sm), probably 
corresponding to the so-called sensory margin known in many land 
planarians. 
The structure, which may be compared with the ‘‘ Sinneskante,”’ 
is seen, in transverse sections through the glandular ridges, to be a 
narrow clear space of the epidermis (about 0°02 mm. wide) lying 
between the two rhabdite-layers. Its most characteristic features 
are the total absence of nuclei in the epidermis of this region, and 
also the absence (probably apparent) of the basement membrane 
(see Plate IV., fig. 8, sm). The epidermis is here represented by a 
group of faintly stained fine threads which frequently anastmoose. 
These threads are directly continuous with a small, clear, and com- 
pact mass of elongated cells, which are certainly nucleated. The cells 
run parallel to each other and obliquely to the epidermal surface. I 
have not been able to detect any distinct connection of the above 
cells to the nerve-fibres, which abound in the neighbouring parenchy- 
matous tissue. Nor have I found any ciliary appendages either on the 
sensory margin or on the whole epidermis, excepting that of the sole. 
Although there are still some points to be elucidated in the 
histology of the sensory tissue described above, I cannot entertain 
any doubt as to the homology of this structure with the “ Sinnes- 
kante”” observed by von Graff in many land planarians. Three 
chief points of homology may be noticed : (1) the absence of nuclei 
in the external plasmic layer; (2) the direct connection of the 
latter to the underlying nucleated spindle-shaped cells ; and (3) the 
absence or imperfect formation of the basement membrane. If 
the above homology be accepted, the species under examination 
differs in a striking manner from other land planarians with regard 
to the relative positions of the sensory and glandular margins. In 
all previously known cases the former invariably lies to the latter, 
while in the present case the relative positions are reversed. Lastly, 
it must be added that that horseshoe-shaped depression which is 
