34 MR. R. H. COMPTON: AN INVESTIGATION OF THE 
tangentially and then divides into two. While this is happening, a fourth 
xylem plate appears, first as a projecting group of metaxylem vessels which 
then acquire a protoxylem (fig. 81). The base of the hypocotyl is tetrarch, 
but the polar xylems are better developed than the lateral. 
The primary xylem becomes mixed with a little parenchyma and so broken 
up into detached pieces. No true pith appears, however, until we have 
ascended about seven-eighths of the hypocotyl, the centre of the stele being 
occupied by one or two large metaxylem vessels. As we pass up the hypo- 
cotyl all the primary xylem vessels become spirally thickened. About 2 mm. 
below the cotyledonary node a pith appears, and the xylem separates into 
two V-shaped masses. These gradually contract into pairs of loose xylem 
groups, joined externally by the polar protoxylem and having the halves of 
the lateral protoxylem attached internally. In this condition the cotyledon- 
tube is entered. Here each cotyledon trace consists of a pair of wedge-shaped 
xylems, with phloems obliquely dorsal to them and with the crushed proto- 
xylem in between. 
As the cotyledons become detached the bundles become more collateral and 
endarch, and fuse together in the median plane, shortly afterwards beginning 
to branch in a pinnate manner. 
MEDICAGO TURBINATA, Willd. 
Diffuse annual herb. Cotyledons ovate. 
The material was unsatisfactory. Triarchy was found in the root, changing 
to tetrarchy higher up. In the upper part of the hypocotyl the inter- 
cotyledonary root-poles were absent, the transition being completed on the 
diareh plan and being high. Thus the comparison appears to be with 
Ononis rotundifolia, at least in some respects. 
MEDICAGO FALCATA, Linn. 
Hypocotyl tapering slowly to the root. Cotyledons ovate, not jointed to 
the petiole. 
The;structure of the seedling was investigated by Gérard (1881, p. 349). 
From his account of the mode of transition the present differs in certain 
respects. Gérard describes a triarch root of which two xylems are 
cotyledonary, while the third changes the direction of differentiation from 
centripetal to centrifugal, acquires phloem, and passes asa collateral stem-like 
bundle into the first leaf. 
In the material described, however, two chief differences were observed 
from Gérard’s account. The root was found to be triarch, but in ascending 
the hypocotyl there took place changes like those described in Medicago 
lupulina, &c., resulting in the production of a tetrarch symmetry. Further, 
it was found that the lateral xylems become divided into halves and that 
these pass out into the cotyledons with the polar xylems, the whole forming 
