528 Brehner . — On the Anatomy of 
extends right across the stele and the cotyledon- trace is cut 
off. The phloem reappears on the side of the haplostele 
facing the leaf-trace, and the endodermis is differentiated in 
the ground-tissue between the two 1 . This process is repeated 
a variable number of times, the haplostele showing a fluc- 
tuating tendency to become crescentic, especially about the 
place of junction of a root. When the stele does become 
definitely crescentic, it is not due to the departure of a 
‘ sector ’ from a medullated haplostele, but because it is laid 
down in that shape by the differentiating meristem, hence 
there is no leaf-gap, in the ordinary sense, at first. After 
this crescentic stage has been reached, the next leaf-trace 
may go off from the convex side of the crescent (Fig. 28, 
diagram I), leaving two meristeles behind with a portion of 
extrastelar ground- tissue between them. The space 2 facing 
the previous leaf- trace is closed (diagram II) at the time 
a root comes in, and another leaf-trace is soon given off 
(diagram III). Thus the beginnings of the dictyostele are 
laid down, and it practically amounts to the branching of the 
haplostele through the formation of leaf-gaps. Up to this 
stage the leaf-trace is monomeristelic, only becoming dimeri- 
stelic through late dichotomy, but succeeding traces soon 
become dimeristelic from the first. Diagrams V, VI, VII, show 
how this takes place. After the crescent has been restored by 
the closing of the leaf-gap, the root R, coming in as usual, 
a portion is again separated off (m', diagram V), and rejoins 
the opposite horn of the horseshoe at a higher level (m\ diagram 
VIII). From each of the tips of the crescent or horseshoe 
a leaf-trace meristele is now separated off, and the leaf-trace is 
dimeristelic from the beginning (diagram VIII, Lt). Once this 
stage has been reached, or earlier in some cases, the central 
1 In the Figs, i, 2, 3, PI. XXII, the endodermis is not represented round the 
leaf-trace meristele, because the preparation did not show it, either through 
defective staining or because the seedling was too young. It was quite clearly seen 
in other cases, however, e. g. Fig. 7. 
2 This space, i. e. on the right of the diagram, is not a leaf-gap, but due to the 
change from a haplostele to a crescentic one, subsequently to the departure of 
the leaf-trace. 
