Ranales , Rhoeadales , and Rosales . 
699 
protoxylem group, which gives the most complete appearance of indepen- 
dence at the base. Thus, chance sections of the cotyledonary petiole of 
Ricinus communis certainly do not suggest any connexion between its four 
widely separated strands and a ‘double bundle’; nevertheless, the two 
central bundles unite higher up, and careful examination at a suitable age 
reveals the presence of the isolated protoxylem elements in the petiolar base. 
The anatomical arrangements found in hypocotyl, primary leaves 
(cotyledonary and plumular), and primary root, together with their inter- 
relationships, have proved very difficult of description, and there is not 
a little danger of being led astray by our own conventional even though 
quite necessary terminology. 
Perhaps the greater number of investigators, beginning with Van 
Tieghem, and including Sargant, Wright, Tansley and Thomas, Hill and 
de Fraine, and Lee, have described the transition features from above 
downwards, but not a few, including Smith, Chauveaud, and Compton, have 
described them in the reverse direction. Although I have been in the 
habit of describing the course of the vascular strands proceeding from leaves 
to root, in view of the many pitfalls involved, the reverse procedure is 
perhaps less open to objection. In both cases the hypocotyl is treated 
solely as a ‘ transitional ’ 1 region, and the inevitable impression of motion 
or change of position is exaggerated. From many points of view I am 
inclined to think that a truer concept is obtained by considering the hypo- 
cotyl as an axis with an anatomical structure of its own, having relation to 
that of leaves on the one hand and primary root on the other. 
The hypocotyl may be taken to be the region between the collet and 
the cotyledonary node. This region may be short or long, and at either 
end it will be influenced by its relationships with the other members. The 
most independent and uninfluenced part will be roughly half-way between 
these points. The cotyledonary node is always well defined morpho- 
logically, the collet far less so. The internal changes do not necessarily 
correspond in level with the external changes. This is true in general of 
the insertion of leaf-traces for fossil (e. g. Lyginodendron and Palaeostachya), 
and recent forms, as witnessed by de Bary’s twenty-three types of leaf-trace 
insertion, varying in distance from point of entry (i. e. node) and mode of 
union. 
In 1904 Mr. Tansley 2 and I drew attention to the phenomenon of root 
structure in the central cylinder of the hypocotyl of a number of forms. 
I think it is safe to say that a more or less root-like anatomy is character- 
istic of the hypocotyl of by far the great majority of the forms now known 
to us, that in short it is the rule and not the exception. Nevertheless, the 
1 This finds its most extreme expression in the work of Sterckx. 
Tansley, A. G., and Thomas, E. N. : Root Structure in the Central Cylinder of the Hypo- 
cotyl. New Phytol., vol. iii, 1904. 
