8 4 
Ethel N. Thomas. 
extended bundle, which in Torreya , recalls that of Ginkgo at the 
base. 
The state of affairs in Ginkgo biloba, and the fact that the 
Araucarias may also show hexarch and tetrarch stages before 
assuming the final diarch form, taken in conjunction with the 
marked tendency to reduction seen in the Cycads, are very strongly 
in favour of the primitiveness of the tetrarch type, and the deriva¬ 
tion from it of the diarch type by reduction, at any rate in the 
Gymnosperms. 
That the types are not directly correlated with habk is very 
evident from Araucaria, Ginkgo and Torreya , the seedlings of which 
have the habit of a Cycad, with more or less completely Coniferous 
anatomical features. 
IV.— The Case for Reduction in Dicotyledons. 
Any attempt to assign taxonomic grade to a given character, 
is, at the present time, peculiarly difficult in the Angiosperms, for 
there is so little unanimity as to the relative primitiveness of the 
various alliances, that the evidence drawn from the antiquity of any 
one group, which was justly laid stress upon in connection with 
the Gymnosperms, can only be used with the greatest reserve 
among Angiosperms. Under such circumstances, there is con¬ 
siderable danger of arguing in a circle, and of assigning primitive¬ 
ness to a character, because it is found in plants otherwise 
suspected of being primitive, and then using this character as fresh 
evidence of the phylogenetic position of the group. At the same 
time, if this danger be only properly appreciated, it should not 
prevent us from giving due weight to the coincidence of a sup¬ 
posed primitive character in a supposed primitive plant, if the 
primitiveness of both he also independently supported. 
Such difficulties of necessity arise in the early stages of the 
minute and comparative study of groups, during which not only the 
characters likely to be of service in unravelling phylogeny are being 
determined, but at at one and the same time, the criteria by which 
they are judged are deduced. 
One may regard the study of the Angiosperms as being largely 
in this stage, for, considered in relation to their great number, and 
the complexity of their structure, comparatively little exact work 
has yet been done on them. As far as seedling structure goes, it is 
felt that the number examined up to the present is very inadequate. 
With this preface one may conveniently consider first, the 
