SEEDLING STRUCTURE IN THE LEGUMINOS&.EK. 85 
just below the level of the division of the third xylem. In Hippocrepis 
unisiliquosus and Coronilla montana the fourth xylem never appears as such, 
but is represented by its two halves, which arise above the level of forking of 
the third xylem. Finally, in //edysarum coronarium the whole of the 
hypocotyl is triarch, but the fourth xylem arises de novo in the cotyledonary 
tube. Thus we have a complete series of stages connecting triarchy and 
tetrarchy ; in which direction the series is to be read will be discussed 
later (p. 1055. 
In no case is there a reduction in the number of xylem poles as the 
hypocotyl is ascended. (Contrast certain Podalyriee, Galegez, and 
Trifolieze. ) 
The cotyledon trace usually consists of a median polar protoxylem and а 
pair of collateral bundles, the latter having one or two lateral half-protoxylems 
attached ; it usually gives off branches immediately on entering the 
cotyledon, even where this is very narrow. In Arachis hypogwa also the 
cotyledon trace ix essentially of the same type, despite the unusual form of 
the seedling. 
The level of the transition varies widely, being low in Arachis hypogea, 
the largest seedling studied ; intermediate in seedlings with diameters over 
1:0 mm. (except Scorpiurus sulcata) and also Desmodium canadense, and 
high in the slenderest seedlings. А certain parallelism can thus be 
recognised between the diameter and the level of transition. The exceptional 
‚сазе of the two Scorpiurus spp., 5. sulcata with a high, and S. vermiculatus 
with an intermediate, transition, both having the same diameter, may 
probably be explained by the much greater length of the hypocotyl in the 
latter species (cf. Genisteæ, p. 80). 
DALBERGIEJE. 
The tribe eontains exclusively arboreal species, some being trees, others 
lianes, often showiug anomalous secondary thickening in the stem. 
The germination is imperfectly known. De Candolle (1825) observed a 
single species, Lubbock (1892) three, and two others are recorded here. In 
each case the seedling is large and the cotyledons are epigeal; they are 
sometimes foliaceous; sometimes fleshy, not green and not expanding 
(ef. Phaseolew). The first two leaves are opposite, this being also comparable 
with the Phaseoleze. 
The root is tetrarch in the two species examined, all the four protoxylems 
being cotyledonary ; the transition in both is low. In details of structure, 
however, the two species show certain marked differences. 
VICES. 
The great majority of species of this tribe are herbaceous climbers, annua s 
or perennials : most climb by means of tendrils ; Lathyrus Nissolia supports 
