118 BULLETIN OF THE 
segmental duct, subsequent to its early development, occupies a position 
along the dorsal surface of the somatopleure (so’plu., Figs. 35 and 43), 
with which it lies in close juxtaposition. 
It should be noted in this connection that in the posterior trunk 
region the segmental duct is developed somewhat later than in the more 
anterior region, and its development is also slightly modified, owing 
to the great volume of the axial portion of the intermediate cell-mass 
(Fig. 37) and the retarded progress in the development of the lateral 
layers and intestine; but its source is exactly the same as in other 
regions, viz. from the lateral portion of the intermediate cell-mass. 
The fate of the axial portion of the intermediate cell-mass is not 
so easily traced as that of the lateral portion. I have not been able 
to demonstrate any differentiation of it prior to the beginning of the 
third day, and therefore only after the formation of the nephrostome 
and segmental duct in the anterior trunk region. That it is the 
source of the blood corpuscles seems to me almost beyond doubt, and 
that it also contributes the elements which form the walls of the sorta, 
at least in the middle and posterior trunk regions, seems to me 
equally certain; but what rôle it takes in the anterior trunk region, 
where it consists of relatively few cells which are brought in juxtapo- 
sition with elements from other sources, is not easily determined ; it is, 
however, highly probable that these elements have the same fate as 
those originating from the same structure in a more posterior region 
of the embryo, — that it gives rise to the blood corpuscles. 
It is worthy of note, that before the cells coming from the inter- 
mediate cell-mass have reached their position under the chorda, a few 
nuclei, whose origin I have referred to the ventral margins of the proto- 
vertebre, are seen in this position. These are the nuclei (nl.*, Figs. 
24-277) of cells which lie directly below the chorda and form a slender 
structure resembling the subnotochordal rod in those Chordata which 
are lower than the Teleosts; but the origin of the subnotochordal rod 
is in all cases, so far as I know, referred to the endoderm. Hence, if 
my supposition concerning the origin of these cells is correct, the two 
structures cannot be homologous, one being mesodermal and the other 
endodermal in origin. There is, however, this similarity, that this 
structure in Fundulus, as in case of the subnotochordal rod, does not 
exist until some time after the formation of the chorda, and that it en- 
tirely disappears at a later stage. In an embryo of three days (Plate II. 
Figs. 11-14) no trace of these nuclei can be seen between the chorda and 
the layer of entoderm below it ; about two days later, when the intestine 
