162 
Sargant . — 77 /<? Reconstruction of a Race of 
( 87 , 88). The generosity of the authors has allowed me to verify details 
by examination of their original preparations. Their researches deal, as 
already stated, with Gymnosperms as well as Dicotyledons. 
Next in importance comes recent work inspired by the same phylo- 
genetic aim, but dealing with comparatively small groups (T. G. Hill, 40 , 
41 ; A. W. Hill, 39 ; Matte, 60 ; Hill and de Fraine, 42 ). Finally we have 
the anatomical literature just quoted as corroborative evidence. 
After my own experience of Monocotyledons with their endless variety 
in detail, the uniformity of Dicotyledons is very striking. The hypocotyl 
and primary root form part of the main axis of the plant. In annuals, 
biennials, some perennials, and in most trees, they last its life. They 
commonly survive for several seasons even when not permanent. Accord- 
ingly these organs become adapted to their functions in a way rarely found 
among Monocotyledons. The survival of variations in their structure de- 
pends on conditions affecting the plant during its whole life, not on those 
alone which immediately succeed its germination. We have in consequence 
a tendency to uniformity in the primary structure of the hypocotyl and 
primary root, because it reflects the permanent conditions of life, and these 
are uniform, too, in the long run. Moreover, in Dicotyledons the primary 
structure of root and stem alike is subordinate to the secondary structure, 
and this is uniform throughout the class. 
Accordingly the Dicotyledons as yet examined display but two distinct 
types of seedling structure. Both are widely distributed within the group. 
The extreme forms of each type are connected by intermediate links. The 
types maybe distinguished as tetrarch and diarch (Thomas, 88, pp. 79-81 ; 
Dangeard, 20). 
In both types the blade of each cotyledon possesses a midrib, and, as 
a rule, two main lateral bundles. All three bundles enter the petiole and 
commonly fuse there, but occasionally they are found as distinct traces within 
the hypocotyl, and do not unite even during their insertion on the stele. 
In the diarch type the lateral traces always unite with the midrib before 
its insertion on the stele, though complete fusion is sometimes postponed 
very late. Mechanical reasons may sometimes determine the level of the 
junction. Thus in Nigella daniascena the petioles of the cotyledons are 
quite 10 mm. long (Fig. 7). The midrib runs down the middle of each 
petiole, and a lateral bundle stiffens either edge (Fig. 8). In the much 
shorter petioles of Delphinium requienii (Fig. 1) the laterals are inserted on 
the midrib at the base of the blade (Fig. 3). The fusion bundle can be 
followed down the petiole, and in transverse section it shows the curious 
double structure (Fig. 2). The phloem groups are distinct : the metaxylem 
groups are partially divided from each other by a single group of 
protoxylem, common to both divisions of the bundle. In Nigella the 
midrib trace shows a similar double structure at the base of the petiole 
