EOOTS OF CERTAIN PALMS. 279 



sented in fig. 3, cut 77 millim. from the apex, sliovvs about one Iiundred c^roups of 

 protoxylem ; while a section 150 millim. from the apex (fig. 5) shows more tlian one 

 hundred and twenty distinct groups ; and cases more extreme could be cited. It is 

 scarcely conceivable that secondary changes could accomplish this increase of a tissue 

 like protoxylem-groups. 



Thus the first explanation seems to be inconsistent with the evidence from anatomy 

 and to have no parallel in any shoot. The second view, on the contrary, is consistent with 

 both lines of evidence. Developmental studies have made familiar the idea of the apical 

 meristem of a shoot altering its mode of differentiation while forming successively 

 younger parts. Por example, as regards vascular bundles with their protoxylem-groups, 

 there is a numerical increase, as the plumule of a monocotyledon undergoes develop- 

 ment ; and there is numerical decrease in later-formed portions of an axis of Equisetum. 

 Purther, as regards steles, Leclerc du Sablon has shown that in many ferns, sucli as 

 JPteris, a series of transverse sections of the same stem shows a change in number. It 

 is true that in this case the number of steles increases in successively younger portions ; 

 but there is ultimately reduction to a single stele in Nephrolepis^ and also in Gunnera 

 and in Primula Auricula, as recorded by Van Tieghem and Douliot. 



The following account of the changes in the mode of development of successively 

 younger portions of complex Areca-roots is true also of roots of certain Palms to be 

 mentioned afterwards. The changes are in part illustrated in the figures just referred to. 



As regards the structure of the apex of these roots seen in longitudinal section, thin 

 normal roots conform to the triacrorhize type usual in Monocotyledons ; thicker 

 roots show a structure which would doubtless be described by Van Tieghem as an 

 " enchevetrement de trois sortes d 'initiales," a condition which, from Sachs's standpoint 

 with regard to the disposition of walls in apical meristems, might well be expected. 



In the transverse section of the older part of a root there may be several perfect steles, 

 each surrounded by a complete endodermis, showing as circles ; and also imperfect steles 

 showing as arcs. These perfect steles are continued into younger parts in the form of 

 imperfect steles with incomplete endodermis. Further, the imperfect steles are con- 

 tinuous with the single central stele of still younger portions, which includes for some 

 distance isolated vascular strands in its pith. 



In the course of the gradual transition to simplicity of structure, Y- or V-shaped groups 

 of xvlem become l-shaped, as a limb may cease to be developed in the younger portions, 

 and correlatively the two phloem-groups that flanked the suppressed limb are, in the 

 succeeding portions, represented by a single strand. 



Further, two l-shaped groups of xylem with their surrounding sclerenchyma may be 

 convergent in the younger portions, thus forming Y- or V-shaped groups with one 

 surrounding mass of sclerenchyma. 



Vascular groups forming the edges of imperfect steles are continued towards the apex 

 as isolated strands traversing the pith ; these strands consist of xylem and sclerenchyma, 

 accompanied sometimes by phloem ; they gradually cease to be developed, phloem earlier, 

 xylem later ; see isolated strand, i.str. in tigs. 9, 8, 7. 



The endodermis accompanying each imperfect stele is incomplete as such, merging 



