596 
Davey. — Seedling Anatomy of certain Amentiferae. 
groups and from which differentiation proceeds in a more or less tangential 
direction on either side’. This statement may be applied equally to 
the Amentiferae at present under consideration, the structure of which, 
therefore, lends support to the view that the unit of vascular structure in 
the hypocotyl is the triad unit as above defined. The triads are continued 
upwards into the cotyledonary strands (and sometimes into plumular leaf- 
traces), where they constitute the familiar double bundles of so many 
authors. 
In the Amentiferae, examples of the double bundle or triad struc- 
ture in the traces of earlier plumular leaves are frequent ( Juglans , 
Corylus, &c.). At their base in the region of the cotyledonary node these 
traces usually die out or become merged with cotyledonary strands in the 
hypocotyl. In certain cases, notably Juglans nigra and Cary a olivaeformis , 
the triads of the first and second plumular leaves extend downwards 
into the hypocotyl as independent strands related to root poles in exactly 
the same way as the cotyledon traces. 
In Castanea sativa there are interpolated between the cotyledonary 
traces a number of additional strands (usually seven or eight), which 
are well differentiated as triad units in hypocotyl and root, and corre- 
spond in position to the plumular leaf-traces or desmogen strands found 
at the cotyledonary node. It seems that there is continuity of unit strands 
between hypocotyl and plumular leaf. This would indicate that a large 
number of units in root and hypocotyl may be differentiated in connexion 
with rapidly developing plumular leaves. The number of leaves is not so 
great as might appear, since the three strands which enter the node from 
any one leaf may remain as independent centres. 
In the genus Juglans there are shown varying degrees of relationship 
of two hypocotyl units with the strands of the first and second plumular 
leaves. In Juglans nigra the two intercotyledonary poles are in continuity 
with the triad units of plumular leaves. In J. Sieboldiana the central part 
of the strand is plumular, while the lateral portions diverge into the cotyle- 
dons ; in J. cinerea the poles are entirely cotyledonary. 
It is possible that a similar series might be established in the genus 
Cary a, in which, so far, only two species have been examined : C. olivaeformis, 
which is octarch with its intercotyledonary poles entirely plumular, and 
C. amara , which is hexarch owing to the absence of poles in this position. 
Similar phenomena 1 have been described by Compton 2 as occurring 
in members of the Leguminosae. Independent plumular intercotyledonary 
poles like those of the above-mentioned species of Juglans and Carya have 
been found by him in Pithecolobium Unguis-cati and in Caesalpinia 
sepiaria . 
1 It is possible that the 1 accessory ’ root bundles described by Miss W. Smith in Bumeliatenax 
(loc. cit, p. 192) may bear the same interpretation. 2 loc. cit., pp. 9 and 21. 
