86 MR. R. H. COMPTON : AN INVESTIGATION OF THE 
itself by the tips of its phyllodes ; Abrus precatorius is a twining plant ; 
Vicia Faba does not climb. Abrus and Cicer tend to be woody. 
The great majority of the seedlings of this tribe show an extraordinarily 
uniform type of structure. Germination is hypogeal ; the cotyledons being 
large and fleshy, and but rarely becoming green even when exposed to light. 
The hypocotyl is very short, and in many cases the root-like surface extends 
up the axis to the very base of the cotyledons. Below the cotyledons the 
axis tapers continuously to a long primary root. The cotyledons are inserted 
on the axis at an angle of about 120? from one another, the re-entrant angle 
being bisected by the plane in which lies the first epicotyledonary bract. 
The anomalous genus Abrus shows a germination more nearly resembling 
that of certain Phaseolee ; it is epigeal, the cotyledons are opposite, and the 
first pair of leaves are opposite also. In most cases the root contains three 
similar primary xylem groups, of which one supplies each cotyledon and the- 
third passes above the cotyledonary node and enters the first. bract-like leat. 
In Vicia Faba, the largest seedling of this tribe studied, the structure is more 
complex: two of the root bundles enter the epicotyl, and connect with the 
first two braets respectively : the cotyledonary bundles may be two, in which 
case they behave exactly as do the corresponding bundles in triarch species ; 
or they may be more than two, in which case anastomoses occur below 
the node. In other large species, however, such as Pisum sativum and 
P. Jomardi, the root is triarch, and the structure of the hypocotyl is quite 
similar to that in such a slender form as Lathyrus Nissolia. 
The transition phenomena begin a few millimetres below the cotyledons, 
irrespective of the change of epidermis, and proceed rapidly ; the root 
protoxylem occupies, at the node, a somewhat median position with respect to 
the metaxylem, so that the transition is incomplete below the cotyledons. 
The leaf trace behaves a little differently from the cotyledon bundles ; its 
assumption of the centrifugal direction of differentiation takes place rather 
rapidly at the cotyledonary node, so that in the epicotyl it shows no trace of 
exarchy ; it is curious to notice, however, that it shows signs of doubleness 
which betray its root origin (cf. Cresalpinia sepiaria, p. 21). 
In those cases in which only the first bract-like leaf contributes protoxylem 
to the root—i. e. in ihe majority of species—it is noteworthy that the 
xylem trace of the second bract is represented below the cotyledons by 
metaaylem of the central triarch xylem ; a study of early stages of develop- 
ment shows that this metaxylem of the second-leaf trace is differentiated 
next after the three protoxylems in the hypocotyledonary axis, and before the 
tangential metaxylems whose fate is to form the faisceaux réparateurs. From 
this condition to that found in tetrarch Vicia Faba there is but a small step. 
The triarchy found in this tribe must be carefully distinguished from that 
seen in so many epigeal Trifoliew, Lotes, Galegeæ, and Hedysareæ, in which 
all the root protoxylems are cotyledonary. It is not justifiable to seriate the: 
