THE FOSSIL OSMUNDACE^. 771 



departure of each leaf-trace this internal phloem becomes continuous with the external 

 phloem along both sides of the leaf-gap (PL V., figs. 24, 25, and 26, l.g. ph.). The 

 external phloem presents the same peculiarities as that in the modern Osmundacese. 

 The sieve- tubes of the metaphloem are plentiful and large, and they are separated from 

 the tracheides of the xylem by a broad stratum of parenchyma four to five cells thick 

 (PL V., figs. 25 and 26, par.). There is no true protophloem, but the metaphloem is 

 immediately followed towards without by eight to ten layers of cells strongly elongated 

 in a tangential direction (PL V., figs. 25 and 26, p.l.). Their walls sometimes show 

 clear indications of a pitted or sieve-plate marking, and beyond doubt they correspond 

 to the porose layers first mentioned by Zenetti (7) in the living Osmundacese. The 

 more peripheral elements of this porose zone are narrower and less tangentially elon- 

 gated than the rest, and they may perhaps be counted as a normal pericycle. No layer 

 resembling an endodermis can be distinguished in the fossil, and, as stated by Penhallow, 

 it is practically impossible to set a definite limit to the stele. 



The internal phloem consists of metaphloem alone (PL V., fig. 27, int. ph.), and as 

 before the sieve -tubes are separated from the inner margin of the xylem by about five 

 layers of parenchyma (PL V., fig. 27, par.). The porose cells are entirely absent on the 

 inside of the xylem ring, nor are there any tissues present that can be identified with a 

 pericycle or an endodermis. 



The internal and external metaphloems are connected above the point of departure of 

 each leaf-trace by two bands of sieve-tubes extending along the two sides of the leaf-gap 

 (PL V., figs. 24, 25, and 26, l.g. ph.). These sieve-tubes are, as elsewhere, separated from 

 the xylem by about five layers of parenchyma (PL V., figs. 25 and 26, par.). The 

 median region of the leaf-gap is occupied by a large mass of starch-bearing sclerenchyma, 

 which extends inwards directly into the pith and outwards into the cortex in the axil 

 of the leaf-trace (PL V., figs. 24-26, l.g. sc). Above the leaf- departure the porose 

 layers very soon extend again across the leaf-gap so as to close it up on the outside 

 (PL V., fig. 26, p.l.). The sclerenchyma in the middle of the leaf-gap is separated from 

 the two bands of sieve-tubes that line the sides by about four layers of thin-wal]ed 

 parenchyma (PL V., fig. 25, par. 2 ). As the leaf-gap becomes narrower towards above 

 (cf. the diagram of Osmunda regalis, Plate VI., fig. 1), the sclerenchyma gradually 

 diminishes in quantity until at length the two lateral bands of phloem come into contact 

 in the median region of the gap (PL V., fig. 24, Ig. 1 and lg. 2 ). Still higher up, the two 

 xylem strands themselves fuse together to form a single one. The fusion is at first only 

 partial, beginning in their more central regions, so that the single strand thus formed has 

 a deep groove on its external surface that is still filled by the remains of the median 

 connecting band of phloem (PL IV., fig. 23, xy. 1 and xy. 2 ). 



In order to complete the description of the stele, it is necessary, first of all, to give 

 an account of the structure of the leaf-trace, and then to deal with the disposition of 

 the tissues in the regions lying below the leaf-gap after their entry with the leaf-trace. 

 The leaf-trace itself is very large, and it is already strongly curved, even while still in 



