Anatomy of Ter ato logical Seedlings. /. 523 
lag behind the members of the cotyledonary whorl, a feature which was 
evident to some extent in the tricotylous and hemitricotylous series, and 
which has already been noted by Compton ( 4 ) in Cannabis saliva. 
Discussion. 
Although the references to polycotylous seedlings are relatively numer¬ 
ous in botanical literature, they are as a rule merely incidental, and in the 
majority of cases the anatomy of such seedlings does not seem to have been 
investigated. This is especially the case with the earlier papers, which are 
little more than records with brief anatomical notes in one or two instances. 
Leger ( 14 ) has dealt somewhat fully, however, with the abnormal seed¬ 
lings which are frequently found in Acer platanoides, and Dangeard (6) in 
his investigations on the anatomy of coniferous seedlings has also attempted 
a classification of the seed leaves met with in the polycotylous genera of 
that group. 
The series of papers by Hill and de Fraine on the anatomy of gymno- 
spermous seedlings ( 9 ) is, however, the first in which a systematic attempt 
is made to show how the polycotylous condition may have arisen ; and 
although this occupies a subordinate position in the general scheme of their 
investigation, they elaborate a theory of whole cotyledons, half cotyledons, 
and subsidiary cotyledons, the basis of which is the part taken by the 
cotyledonary strand in root formation. Thus whole cotyledons are distin¬ 
guished by the fact that the vascular bundle of the cotyledon forms a root- 
pole quite independently; half cotyledons, on the contrary, being character¬ 
ized by their bundles uniting in pairs, each pair forming a root-pole; whilst 
the vascular bundles of the subsidiary cotyledons merely fuse on to the 
other bundles without taking any direct part in root formation. This theory 
they support by a considerable body of evidence, and they interpret certain 
cases as illustrating the ‘ promotion ’ of subsidiary and half cotyledons to 
higher rank as whole cotyledons. 
Compton ( 4 ), in that section of his paper dealing with schizocotyly, 
criticizes certain details of this scheme, taking exception to the class of 
subsidiary cotyledons, and to the idea of ‘ promotion and to these criticisms 
Hill and de Fraine ( 13 ) have replied in a further paper. 
In addition to the papers of Hill and de Fraine, and Compton, 
a number of isolated references to the anatomy of polycotylous seedlings are 
to be found in the majority of the more recent contributions to seedling 
anatomy, and these will be referred to when necessary. 
With regard to the seedling anatomy of the wallflower itself, as men¬ 
tioned above, that of the normal plant has been described by Thomas (20) 
and also by Scott ( 17 ). In addition Thomas ( 20 ) mentions tricotylous 
specimens of Cheira 7 ithus Cheiri and of Sisymbrium carpaticum y both 
showing persistent triarchy, and one of Matthiola tricuspidata in which a 
M m 
