126 
EDMUND W. SINNOTT 
the anatomical features of the seedhng, and that it may be used to 
distinguish large groups of plants. Of course the number of species 
studied is far too small to give an accurate idea of what the conditions 
are throughout the various families, but in those where the largest 
number of species has been recorded, nodal uniformity is very evident. 
In several cases both main types are found in the same family. The 
Ranunculaceae are prevailingly unilacunar, but certain species of 
Clematis seem from their description by Miss Thomas (5, p. 706) to 
have several gaps. We have recorded a similar situation in Berberis, 
the Magnoliaceae and the Euphorbiaceae. It seems very likely that 
in many of these less specialized families the seedling node may not 
be completely uniform. As to its general conservatism throughout 
the dicotyledons, however, there can be little doubt. 
A comparative study of the seedling node will evidently give us 
much valuable information as to relationships. The multilacunar 
condition seems to be quite absent in many of the great orders, even 
in those where the node of the mature stem has several gaps. In 
others it is invariably present in all species so far examined. In still 
others, notably the large and heterogeneous orders Geraniales and 
Sapindales, it characterizes certain families or groups of families but 
is.absent from the rest, thus providing a clue as to relationships within 
the order. In the Sapindales, for example, the Aceraceae, Hippo- 
castanaceae and Sapindaceae, grouped together by Engler as the sub- 
order Sapindineae; and the Melianthaceae, comprising the sub-order 
Melianthineae, are multilacunar. All the rest of the families ex- 
amined in this order, however, are unilacunar. 
As to what has been the evolutionary history of the seedling node 
we cannot be sure. Evidence has elsewhere been presented (2) that 
for the mature stem the trilacunar node was the primitive one among 
Angiosperms. The fact that this is present in the seedling node of 
only a few families, however, and that these are for the most part 
by no means primitive in their other characters, suggests that the 
unilacunar condition of the seedling is a persistence of an ancient 
gymnospermous condition and that the multilacunar type made its 
appearance in the node of the foliage leaf and has worked down from 
thence into the seedling. On the contrary, it may be argued from 
those cases where the lateral traces in the multilacunar condition are 
very small, that they are here dying out; and that the unilacunar 
type has arisen by the complete loss of lateral traces which originally 
were always present. 
