Franklin P. Mall 289 
units together with a terminal portal twig. Now these arterial capillaries 
communicate with the capillaries of the lobule, as do all of those that 
encircle the bile ducts. At no point does any of the injected mass enter 
the portal branches within the liver unless it is as a backward injection, 
in a direction the blood does not circulate normally. To test this 
question thoroughly the whole portal tree was first plugged with a thick 
granular injection mass made by mixing lamp-black or baryta with gela- 
tin. The venous tree thus plugged did not cut off the capillaries, but pre- 
vented a second injection into the artery from spreading in the vein. In 
Fic. 58. Picture to show the termination of the hepatic artery. X 40. 
P, portal vein; h, hepatic artery. The portal vein was first plugged with a 
granular mass and then an aqueous solution of Prussian blue was injected 
into the artery. The extent of the capillary injection with the blue is shown. 
general it was found that baryta gelatin for the portal vein and aqueous 
Prussian blue for the artery gave the most satisfactory result. All the 
vessels and capillaries injected with either fluid are always sharply de- 
fined. In such an experiment it is possible to force the blue fiuid through 
all of the capillaries arising from the artery and in case they collect within 
the capsule of Glisson which empty into the portal vein, their beginnings 
should be marked. But at no place were such veins found; the blue had 
only percolated through the white mass in the interlobular veins. The 
capillary plexus around the bile ducts communicated directly with the 
