i 9 o8] THIESSEN—DIOON EDULE 371 



a larger development of metaxylem; also secondary wood may have 

 developed in the lower extremities, so that the transition from endarch 

 to exarch has become more prominent and can be made out with 

 greater clearness. The secondary wood accompanies the bundle as 

 yet only for a short distance, and ceases long before the transition 

 from endarch to exarch is complete. As the secondary xylem and 

 centrifugal metaxylem diminish, the centripetal xylem (which of 

 course is all primary) increases in bulk (fig. 9, a). In a plant three 

 years old no other secondary wood was present (figs. 22, 2j). Even 

 in the oldest leaves of quite old plants the secondary wood, which at 

 the origin of the strand is quite massive, decreases very rapidly, and 

 in the petiole just above the leaf base has thinned out to a few elements 

 (fig. 24), remaining quite uniform to the rachis, where it disappears 

 still more; while in the pinna no secondary wood whatever is present, 

 all the xylem being primary and centripetal. Although in the transi- 

 tion region the secondary wood diminishes in the same ratio in which 

 the primary wood increases, it must be noticed that the centrifugal 

 wood is not restricted to the secondary wood alone, as was shown in 

 the younger bundles of the embryo, where the transition is clearly 

 carried out in the protoxylem and metaxylem alone. Thus in the 

 seedlings the transition from centrifugal to centripetal wood is carried 

 on after the appearance of secondary wood, and is completed in the 

 primary wood. 



In the older strands where secondary wood has been developed, a 

 considerable amount of the centrifugal wood therefore is metaxylem. 

 This is shown by the amount of procambium that has been developed 

 into centrifugal xylem; as may be seen by comparing the younger 

 strands in figs. 21, 4, a, for example, where there is a certain amount 

 of centrifugal procambium, the amount depending upon the distance 

 from the point of egress from the central cylinder, with the older 

 strands in fig. 24. Sometimes all of the centrifugal procambium 

 has become xylem; more often, however, patches of procambium 

 or isolated cells of it are never lignified and retain their 

 nuclei, and are then referred to as the thin-walled cells. These 

 thin-walled cells do not necessarily lie against the secondary 

 wood, though they most often do, and become most evident 

 in the upper extremities of the transition. A series of cross- 



