STRUCTURE OF RADIAL BUNDLES. 36 1 



the outside, while further inside a second occurs, which is elliptical in cross- 

 section, and is separated from the outer group by interstitial sclerenchyma, in which 

 vessels frequently lie. Thus both an inner and an outer row of phloem-groups are 

 here present in the otherwise typical bundle. 



It has been mentioned above that the xylem-plates not uncommonly converge 

 towards one or two axial vessels, though it may be without coming into immediate 

 contact with them. Such axial vessels frequently occur isolated, in the middle of a 

 thick parenchymatous or fibrous cylinder, and are separated from the inner edges of 

 the radial plates by many layers of cells. This kind of structure is found here and 

 there as an individual peculiarity of many roots, as for example in the Sarsaparilla 

 of Veracruz^; in Carex folliculata I found, on the same stock, roots of the structure 

 usual in Carex, with a thick dense sclerotic axial-cylinder, and others in which the 

 middle of the latter is traversed by about 5 moderately large, prismatic, pitted vessels, 

 which are in contact with one another. ^^^ 



These trifling forms of deviation constitute the transition to those more con- 

 spicuous cases, in which numerous vessels, as well as groups of sieve-tubes, occur 

 scattered in the whole of the cylinder inside the radial ring, a phenomenon which is 

 characteristic of the thick adventitious roots of many epiphytic Aroideae, of those 

 Musacea^ which have been investigated, of the Dracaeneae, Pandanese, (Pandanus, 

 Freycinetia, Cyclanthus), and of the Palms Iriartea exorrhiza and I. praemorsa. 



All the investigated roots of terrestrial Aroideae, as vvell as those of many epiphytic 

 species, present the usual typical structure; but in the thick aerial roots of other 

 forms, scattered wide vessels, and very large sieve-tubes, isolated or occurring in pairs, 

 and accompanied by cambiform tissue, are distributed throughout the wide and constantly 

 sclerenchymatous cyHnder inside the radial ring. The two kinds of elements do not 

 lie in the radial rows; Tornelia fragrans, Heteropsis ovata, Monstera surinamensis, 

 Adansonii, Raphidophora angustifolia, Scindapsus pictus, Philodendron micans, and 

 Anthurium digitatum^, are examples of this structure. 



The same occurs in species of Strelitzia, and no doubt in other JNIusaceae \ 

 Essentially the same arrangement is present in roots of Dracaenae and Pandaneae, with 

 the sole difference that the axial tissue in which vessels and sieve-tubes are distributed 

 is not homogeneous, but around the vessels and small phloem-groups consists of scleren- 

 chymatous fibres, while between them it is formed of parenchyma, in which, in the case 

 of Pandanus, lie wide intercellular passages containing air and scattered bundles of fibres. 

 The ring has likev\nse sclerotic interstitial tissue between the radial xylem- and phloem- 

 groups, the number of which even in moderately thick (1.5 cm.) roots of Pandanus 

 amounts to nearly 200 of each. The cross- section of such roots therefore presents first 

 the typical, relatively narrow ring, surrounded by a many-layered pericambium and an 

 endodermis, and then, inside this, a wide space filled by parenchyma, in which numerous 

 thick strands run longitudinally. Each of these strands consists of a many-layered mass of 

 sclerenchymatous fibres, in which are enclosed one or more isolated wide vessels or groups 

 of them, and one or more small phloem-groups separated from the vessels, while more 

 rarely only one, or neither of the two forms of tubes are present. The position of the 

 two in the stiand varies irregulai-ly. The distribution of the strands in the parenchyma 

 appears in equivalent roots to be somewhat different according to the species. Among 



* See Berg, Atlas d. pharmac. Waarenkunde, Taf. III. g. 



-' Van Tieghem, I.e. p. 149. 



' Compare Wittmack, Musa Ensete,' Halle (Linnxa), 1867, p. 62. 



