819 
Syncotyly and Schizocotyly. 
characterized by the bundle of each forming one pole of the root ; (b) half- 
cotyledons, which are recognized by the bundles of two of them being 
required to form one pole of the root structure ; and (c) subsidiary cotyle- 
dons, the strands of which have no influence on the number of bundles 
in the root structure ’} 
The fact appears to be that Hill and de Fraine are not justified 
in arguing from anatomy to morphology in this matter. Their categories 
apply only to the behaviour of vascular strands (and even here not by any 
means strictly 2 ) ; they cannot be transferred to the cotyledons . 3 
This may be illustrated by reference to the cases of polycotyly here 
described. As stated above, everything points to schizocotyly being the 
explanation of teratological polycotyly ; and yet in a great many cases the 
two vascular strands of the split cotyledon (or of two equal half-cotyledons) 
do not behave alike, but one of them has the characteristic upon which Hill 
and de Fraine define a subsidiary cotyledon. (See, for example, Phacelia 
tanacetifolia, ) In other cases (e. g. Cannabis sativa ), the vascular strands 
of the half-cotyledons are strictly equivalent, and unite to form jointly 
a single root-pole. It seems quite clear that the distinction between 
subsidiary cotyledons on the one hand, and half- and whole cotyledons on 
the other, cannot be maintained. The term ‘ subsidiary cotyledon ’ should 
be restricted to the other use which Hill and de Fraine make of it, viz. to 
denote a displaced plumular leaf. 
With the general conclusions of Hill and de Frame’s work I am 
in agreement, considering that the evidence for the derivation of polycoty- 
lous types from dicotylous ancestors is supported by highly cogent evidence. 
In addition to the arguments which have previously been brought forward 
in support of this view we have now the evidence that an increase in 
number of cotyledons by fission does take place as a more or less heritable 
anomaly in a great number of Dicotyledons, and that the anatomy of such 
schizocotyls shows features of close similarity with those described in 
normally polycotylous Gymnosperms. 
Moreover, in the relatively advanced Dicotylous families Proteaceae 
and Loranthaceae we find a few species whose seedlings are polycotylous 
and show remarkable resemblances to Coniferous polycotyls ; and here, as 
Fletcher 4 and Hill and de Fraine 5 have recently maintained, there can be 
practically no doubt that the ancestry was dicotylous. 
1 Hill and de Fraine (’09), p. 221. 
2 This becomes clear from a comparison of the different ways in which they explain the structure 
of, for instance, Cnpressus tonilosa, series C, Abies sibirica, series B, and Abies amabilis ((’08) p. 699, 
(’09) pp. 191, 195. 
8 It is clear that the term 1 half-cotyledon ’ implies that the structures in question have arisen by 
the fission of ‘ whole-cotyledons’. This, in itself, makes the idea that they can arise by ‘promotion’ 
of ‘ subsidiary cotyledons ’ unthinkable. 
4 Fletcher (’ 09 ), p. 877. 
5 Hill and de Fraine (T2). 
