i68 
Sargant . — The Reconstruction of a Race of 
type characteristic of the Tulipeae, can be derived from the seedling 
skeleton of Albuca or Galtonia , which must therefore be primitive in these 
tribes too. 
Thus the various types of seedling symmetry which have been found 
within the four tribes, Asphodeleae, Alliceae, Scilleae, and Tulipeae, appear 
to be founded on the Anemarrhena skeleton. But these four tribes occupy 
a central position within the family. Together with the Hemerocalleae, the 
Aloineae, the Yuccoideae, and the little-known Johnsonieae (or Aphyll- 
antheae), they constitute the Liliaceae in the narrowest sense. The eight 
tribes just mentioned are united by Mr. J. G. Baker in the sub-order 
Liliaceae verae (6, pp. 354-6, and 7, pp. 209, 210). The seedlings of the 
Aloineae and Yuccoideae which I have examined have been so far modified 
by the conditions under which they germinate that most primitive features 
are lost (72, p. 34). Hemerocallis fulva shows a general resemblance to the 
stouter Scilleae, such as Ettcomis . Nothing is known of seedling anatomy 
in the Johnsonieae. So far, then, it would seem that the Anemarrhena 
seedling presents the most ancient type of vascular symmetry found among 
the Liliaceae proper. 
The other tribes included among the Liliaceae by Baillon (5) or 
Bentham and Hooker have all been treated at some time or other as 
belonging to distinct families. Thus in the English translation of Le 
Maout and Decaisne’s textbook (52, p. 846) the family Liliaceae corre- 
sponds pretty closely to Mr. Baker’s sub-order Liliaceae verae, and the 
Junceae, Aspidistreae, Melanthaceae, Smilaceae and Asparageae are treated 
as distinct families. A note is added to the Liliaceae: ‘We have indi- 
cated the extremely close affinities between Liliaceae, Asparageae, and 
Smilaceae ; families which together form a group to which most other 
Monocotyledonous families may be linked, directly or by intermediates. 
Thus Junceae, which are near certain Melanthaceae and Liliaceae, connect 
these with other families with a free ovary ; and, on the other hand, those 
Amaryllideae and Dioscoreae which belong, the one to the Liliaceae, the 
other to the Smilaceae with an inferior ovary, connect them with the 
epigynous Monocotyledons.’ 
The seedling structure of Monocotyledons, so far as it has been 
examined, is not inconsistent with the view expressed in this quotation. 
The outlying tribes of Liliaceae in the larger sense possess seedlings whose 
vascular symmetry can be referred to one type or another belonging to one 
of the four central tribes. The Amaryllidaceae, Iridaceae, and Aroideae, 
seem similarly attached by seedling characters to the Liliaceae proper. 
The seedlings of Palmae, Scitamineae, and Gramineae, indeed, are very 
distinct, but they are all much modified forms, and there is nothing in their 
seedling structure to suggest a primitive symmetry other than that of 
Anemarrhena. Indeed, many points in the vascular skeleton of Elettaria 
