42 Description of a 4-mm. Human Embryo 
(Figs. 11 and 15). From the plexus on each side -branches extend ven- 
trally, those on the right ending blindly, but one on the left side joining 
by a narrow channel (Figs. 12 and 13, x) with another venous plexus 
(V. pl.) extending across the median line, ventral to the intestine, and 
posterior or caudal to the sinus venosus, from which it is entirely free. 
This ventral plexus is in intimate relation with the liver, and will be 
spoken of more fully in the next section. 
Posterior to the outgrowth of these plexuses the vitelline veins divide 
into many branches which spread out over the yolk sac, into which the 
intestine soon opens. Cut ends of some of these branches may be seen 
in Fig. 13. 
MODEL OF THE LIVER. 
This model is seen from the caudal end, and a little from the right 
side in Fig. 14. The intestine is a large tube much compressed laterally, 
and the vertical right side is shown. From the ventral border of this 
tube hangs a large outgrowth, extending first ventrally, then laterally, 
then dorsally, composed of an irregular mass of entodermal cells, which 
tends to break up into cords, often anastomosing with each other. Only 
the right side of this mass has been modelled. The cavity of the intes- 
tine passes for a short distance into this mass, but without subdivision, 
so that all the cords and irregularities are solid. In the model a part of 
the left wall of the intestine has been cut away to allow a view of this 
cavity. 
It has been stated by other writers that the liver arises by cords of cells - 
which grow into the territory occupied by a large vein, pushing before 
them, and thus becoming invested by the endothelium, and forcing the 
blood to run in small, anastomosing channels, the sinusoids. This state- 
ment would not be true of this embryo, for the liver cords are found 
growing into mesenchyma, at a level ventral to the vitelline veins; in this 
same mesenchyma, however, we find the branches of the vitelline veins 
ramifying, and forming plexuses, and in certain places these plexuses 
come into intimate relation with the liver cords. These points of con- 
tact, where veins and liver actually touch, are marked in the model of 
the liver (Fig. 14, V.), and are seen to correspond with the anterior 
surfaces of the portion of the plexuses marked in the model of the veins 
(Fig. 13, li.). These two models might be fitted together, in which case 
we should find the intestine lying between the vertical, mesial walls of 
the two vitelline veins, the liver spreading in the same mesenchyma that 
contains the lateral and ventral venous plexuses, between the sinus 
