Primitive Angiosperms. 173 
joined by traces from the rudimentary plumule. The protoxylem of each 
cotyledonary trace branches in three directions : the four phloem groups 
remain distinct for some time. The resemblance of the stele at this level in 
transverse section (Fig. 20) to that of Anemarrheua (Fig. 17), and Althea 
(Fig. 14), is very clear. But in Podophyllum as in Erantliis the four 
phloem groups do not remain distinct. They unite in pairs lower down, 
opposite the intercotyledonary poles of protoxylem (Fig. 21), and the latter 
finally disappear, leaving a completely diarch root-stele. 
In Podophyllum and Eranthis the two cotyledons are united for at 
least four-fifths of their whole length, and in both the vascular skeleton 
shows signs of reduction from the tetrarch type. This reduction seems to 
have begun as in Althea , but to have proceeded further. The fusion of the 
lateral traces with the midrib is more complete in the petioles. The for- 
mation of a tetrarch root-stele is merely indicated : the final form is diarch. 
Now the Anemarrheua symmetry can also be considered as a reduction 
from the Althea type. The reduction in the root-stele is less complete 
than that of Podophyllum . On the other hand, the bundles of the petiole — 
so clearly double in Podophyllum (Fig. 19) — are single in Anemarrheua 
(Fig. 16) : their double structure is indicated only just before the transition 
begins. In spite of these differences, the general resemblance between the 
vascular skeleton of the Anemarrheua seedling and that of Podophyllum or 
Eranthis cannot be denied. It appears, then, that the partial union of 
cotyledons in a tetrarch seedling has actually operated to reduce the 
vascular skeleton to a form very near the Anemarrheua type. Whether 
the facts imply any real relationship between the Ranales and the Liliaceae 
or not, they do at least show that union of the cotyledons in a tetrarch 
dicotylar seedling might bring about the formation of a monocotylar type 
with a reduced vascular skeleton resembling that of Anemarrheua. 
Phylogenetic Schemes. 
The two characters which separate Monocotyledons from Dicoty- 
ledons most completely have now been discussed. Primitive Angiosperms 
appear to have resembled their Dicotyledonous descendants in both. 
There can be little doubt that they possessed a cambium. The evidence is 
perhaps less conclusive with regard to the number of cotyledons, but still 
there seems good reason to think that Primitive Angiosperms had two. 
The embryological evidence on which this conclusion is chiefly based 
requires also that the single cotyledon of Monocotyledons should be 
regarded as a fusion of the ancestral pair. 
Monocotyledons, however, differ from Dicotyledons in a number of 
other characters. From the systematic point of view they are of minor 
importance, because they are far less constant than those already discussed. 
Leaves with parallel venation are, for example, the rule among Mono- 
