igi2] JONES— DIANTH ERA 7 



at right angles to the main bundle, and being inserted on the sides 

 of the leaf traces of the first pair of leaves. There is thus formed 

 a complete ring of vascular tissue in the first epicotylar node. 

 Rarely the forks at the base of the bundles are of very unequal 

 size, or one of the bundles may even fail to divide, in this case 

 merely bending aside, and becoming inserted on one side or the 

 other of one of the cotyledonary traces. The direction of differ- 

 entiation of these side bundles seems to be acropetal, but this 

 could not be made out definitely. The cross-arms connecting the 

 bundles with the traces of the first pair of leaves, however, are 

 undoubtedly younger than the main part of the bundles. Fig. 20 

 shows the course of the bundles at this stage. 



stage in 



With the development of the second pair of opposite leaves at 

 right angles to the first pair, and directly oyer the cotyledons, we 

 have the third stage (fig. 22). At the insertion of each of these 

 leaves, there starts to differentiate a single bundle, passing outward 

 into the leaf, and downward through the stem, becoming inserted 

 in the crotch of the fork of the bundle just developed at the close 

 of the preceding stage (fig. 21). 



If we now examine this latter bundle (fig. 23), we find that it 

 possesses three protoxylem elements, one being derived from the 

 newly developed trace of one of the second pair of leaves, the other 

 two connecting, one through each of the horizontal connecting 

 branches, with the outgoing leaf traces of the first pair of leaves. 

 At the lower end of the bundles (figs. 24 and 25), one of the pro- 

 toxylems passes out along one of the forks; the other two, which 

 have been closely applied to each other throughout the whole 

 length of the bundle, pass out the other. These bundles at first 

 have the appearance of being double, the two parts being separated 

 by a narrow parenchymatous (medullary ?) ray, but being, later at 

 least, entirely surrounded by a complete endodermis. 



A cross-section (fig. 28) through the middle of the second 

 epicotylar internode shows an epidermis, or rather protoderm, of 

 thin-walled cells, sharply marked off from the internal cells. 

 Within this protoderm is a meristematic tissue of cells practically 

 all alike, showing no differentiation into cortex and central cylinder, 



