462 Holden and Daniels . — Observations on the 
previous investigation has shown, solely in the epicotyledonary modifications, 
which are strikingly similar. It will be obvious that the mode of origin of 
seedlings of the second type raises a point of some theoretical interest. 
If they are produced by cotyledonary suppression, then syncotyly and 
heterocotyly coexist in the same species : if they represent the ultimate 
phase in a syncotylous series, then the two cotyledons exhibit so perfect 
a dovetailing of parts that all traces of double origin are lost. 
It was realized that suitable material for the determination of this point 
would be scarce, and in all some 118 seedlings have been microtomed with- 
out yielding absolutely conclusive evidence. As was inevitable, however, 
collateral problems of interest have arisen, and these have also been followed 
up and the results included in the present paper. 
The direct evidence which bears upon the origin of the seedlings of the 
second type falls under two heads, namely : 
(a) The behaviour of the ‘double’ bundles in certain undoubted syn- 
cotyls. 
(b) The variations in the cotyledonary axillary buds. 
Two clear cases of complete midrib fusion were obtained (Seedlings A 
and B), and these will be described in detail. In ‘ Seedling A’ (Fig. 1), in 
which a closely syncotylous condition obtains, the respective median strands 
travel down the compound lamina in close proximity and, whilst still 
separate, develop ‘double’ bundle structure (Fig. 16). They are inclined 
slightly to each other, and for some distance subsequent to their union two 
mesarch protoxylem groups are evident in the common metaxylem, with 
a phloem group on either flank and one in the plane of junction (Fig. 17). 
The fate of the median phloem is somewhat curious, the major part passing 
over to unite with one of the lateral groups, while the remnant dies out. 
The compound bundle loses all traces of its double origin, and forms a single 
pole both in the hypocotyl and root. The marginal strand and the lateral 
of one side fuse and, near the apex of the hypocotyl, unite with the 
marginal strand only from the opposite side, the composite bundle becoming 
exarch and constituting a second pole (Fig. 18). The protoxylem of the 
lateral which remains independent persists in an endarch position for a short 
distance and then disappears, being followed at a lower level by the 
metaxylem, so that below this point the hypocotyl is diarch for a period 
(Fig. 19). The two xylem poles associated with the cotyledonary strands 
are augmented by two which rise de novo in the plane at right angles to 
them, and a tetrarch condition is thus produced, giving rise to four lateral 
rootlets in the root-whorl, and below these to a tetrarch tap-root (Fig. 20). 
In ‘ Seedling B ’ the march of events is somewhat different, but is also of 
interest. The seedling (Figs. 2 a, 2 b) is an obvious syncotyl, but in the 
bottom third of the compound lamina the midribs, which are represented by 
collateral bundles at this stage, approach each other at an angle of almost 
