790 
the lobules where the inter-lobular connective tissue penetrates into 
the margin of the lobules, and where the cells of Kuprrer are said 
not to occur. 
The smaller forms nearly always lie single, or at the most in 
in groups of twos, while the larger ones, even those with a distinct 
cell outline, may often be in threes, or even in groups of fours, taking 
up a considerable space between the hepatic cells, or in the connective 
tissue. 
With rare exceptions we have been unable to obtain any staining 
of the connective tissue cells in the inter-lobular spaces by this method, 
the few that were sufficiently distinctly brought out by the stain, ap- 
pearing to be about the same size as the medium granular cells, but 
at the same time more regular in form and less coarsely granular. 
The extensions from all the varieties of the granular bodies are 
ordinarily quite short, and rapidly attenuate along the side of the 
vessel, or between the hepatic cells, but more rarely long processes, 
equally granular with the body, may be found, which seem as if they 
were conjoined with the processes of other peri-vascular cells, forming 
for some distance a continuous line of granular particles along the 
vessel’s wall. 
With the intra-acinose nerves, the perivascular cells have no ap- 
parent relation. At long intervals the finer nerve fibres cross them, 
or run by them; but there are no terminal endings adjusted to the 
bodies, nor are they surrounded by a terminal network, and in their 
vicinity, as well as in the longitudinal direction of the portal capil- 
laries, nerves are very sparingly seen. 
KUPFFER sought to place his star-cells among the number of the 
nerve cell structures, and later his pupil RoTHE, though discovering no 
relation between the Sternzellen and nerve fibres, likens them to the 
smaller ganglia of MEISSNErR’s plexus in the stratum submucosum of 
the intestine, but so far as the silver method permits us to see, they 
have no distinct relation to any of the nerve elements, and belong 
only to the list of the perivascular cells. 
The gall capillaries have absolutely no relation to our granular 
cells beyond occasionally crossing them as they pass from one colum- 
nar row of liver cells to another. 
While it is positive that the livers used in our investigation, 
beyond occasionally containing a few scattered psorosperms (whose 
locality was carefully avoided), were normal, and belonged to young 
though full grown animals, and though the liver structures showed an 
absolutely normal aspect so far as the hepatic cells, vessels, and con- 
