ANATOMY OF THE ROOTS OF PALMS. 453 
for the constitution of a “stele,” then if the cylinders of the base of the root, which, it 
must be remembered, extend in all cases into the apices of the young roots and often into 
the apices of older roots also as separate procambial strands, be not * steles," may they be 
termed “ bundles” ? : 
De Bary regarded the central cylinder of the normal root as a single bundle, and stated 
further: “In the thick roots of Pandanacee and the genus of palms, Jriartia, there are 
found a number of parallel bundles which converge at the growing point." His further 
remark, “it may be doubted whether these should be called parts of one very large divided 
oundle or so many individual bundles," does not concern us here. 
If, then, these strands at the base of the root are bundles, the root is * polydesmie " in 
tbis region, an interpretation leading to the result that the normal cylinder produced 
distally is “ gamodesmic," unless here again an endodermis surrounding each bundle be 
considered necessary for true “ polydesmic " nature. 
If such an investment be considered necessary, then the only conclusion to be drawn 
would appear to be that the root in its basal portion is purely * astelie," a “ meristelic " 
interpretation being out of court, since at no previous stage has any stelar strueture been 
present, ** meristeles " being the result of the splitting of a * stele.” 
Fusion of the separate strands distally, giving rise to series of ares of fibro-vascular 
tissue, provided on their peripheral surfaces with an endodermal layer, leaves the root 
still in an astelic condition, often retained even in very long roots of certain palms, but 
usually giving place distally, by further fusion of the arcs, to a typical root-cylinder 
surrounded by a complete endodermal investment. In this state the root might be 
described as locally monostelic for purely descriptive purposes, without any reference to 
the real mode of origin and relationship of its constituent parts. It is at least doubtful, 
however, whether Van Tieghem's conception of * monostely" is here attained, the 
* monostele " in the present case being the result of a gradual change from an * astelic " 
condition. 
'The stelar theory thus failing to explain satisfactorily the conditions met with in palm- 
roots, it becomes of interest to see how far Hanstein's idea of the root as composed of 
several independent zones arising from as many individual initial layers at theapex will 
here hold good. 
It has been already mentioned that in the apices examined (namely, those of Kentia sp. 
and Areca sp.) stratification is entirely absent, the apex in each case possessing a common 
group of initial cells, by division of which the ground-parenchyma of the root is formed. 
A short distance behind the apex the procambial tissue, from which the fibro-vascular 
elements are developed, appears in the ground-parenchyma, and any distinction of 
Hanstein’s initial layer is quite wanting. 
It is possible to extend this conception to include the structure in many ott ; 
among which may be mentioned the non-stratified root-apices of the Fapilionacere, 
described by Janezewski; stem-apices of the P teridophy ta, occupied by a single initial 
= cell or group of cells, which by division gives rise to a eee ground-parenchy mp, 
wherein arises the procambial tissue, in the so-called “ polystelic " forms, each procambial 
strand being the precursor of a “ stele.” 
her cases, 
9T2 
