Primitive A ngio sperms. 165 
(Fig. 9). The laterals are in contact with the double trace of the midrib, 
though they have not fused with it (Fig. 13, p. 170). The trace which enters 
the stele of the hypocotyl consists of a median double bundle with external 
protoxylem. A lateral bundle is in close contact with it on either side (cf. 
Gdrard, 27 , Fig. 24). The transition to a tetrarch root-stele is in essentials 
the same as that of the previous type (Figs. 14, 15 ; and Gerard, 27 , Figs. 
23-21). This type differs from the extreme forms previously described 
only in the closer approximation of the bundles in the petiole of each 
cotyledon. 
In the upper part of its vascular skeleton, the seedling of Althea does 
not differ greatly from that of Nigella. Each of the twin compound traces 
entering the hypocotyl from the cotyledons consists in both genera of 
a median double trace with a lateral trace approaching it on either side. 
In Nigella fusion of each pair of laterals with its median trace is complete 
though tardy. The root-stele is diarch. In Althea the laterals preserve 
their independence throughout the transition, and furnish an additional pair 
of protoxylem rays to the root, which is consequently tetrarch. Althea 
and the seedlings which resemble it clearly belong to the tetrarch type, but 
they show some approach to the diarch symmetry. Liriodendron , as 
figured by Miss Thomas (88, p. 81), though diarch, makes an even more 
decided approach to tetrarch symmetry. 
Here then is a series of forms linking the extreme tetrarch type 
(Figs. 10-12) to the extreme diarch (Figs. 4-6). Which of the two is the 
more primitive ? Is the tetrarch type derived from the diarch, or the diarch 
from the tetrarch ? 
The authors to whom I have been referring answer this question in 
their joint note ( 87 ). Miss Thomas gives their reasons more fully (88). 
They do not consider that comparison of different Dicotyledons with each 
other gives any decided clue. Both types are found in a wide range of 
families ; some in either list are probably ancient. But comparison with 
the seedlings of Gymnosperms is more instructive. Again both types are 
found, but here the tetrarch symmetry is characteristic of those Gymnosperms 
which botanists agree to consider primitive on other grounds. 
The seedling of the primitive Gymnosperm, then, probably possessed 
tetrarch vascular symmetry. But it must not be forgotten that a gap of 
enormous width separates the primitive Gymnosperm from modern Angio- 
sperms. Reference to the skeleton pedigree reproduced on page 137 shows 
the latest common ancestor of both groups to be some unknown Pterido- 
sperm. Assuming that Cycads and other primitive Gymnosperms derived 
the tetrarch symmetry of their seedlings from this Pteridospermic ancestor, 
we make a second and a greater assumption in supposing that the tetrarch 
symmetry of modern Dicotyledons has come down unchanged through 
countless generations of Hemiangiosperms to the Primitive Angiosperms, 
