THE ANATOMY OF TODEA. 25] 
large cells not unlike the brown and thicker-walled cells of the subapieal * annulus " of 
an Osmundaceous sporangium. Some of the sporangia described by Williamson and 
others from the calcareous nodules of English coal-seams have been shown by Bower * 
to bear a close resemblance to those of Zodea; and Zeiller + has figured examples of 
sporangia found by Renault in quartz pebbles from Autun which recall those of 
recent Osmundacez.  Zeiller has also given an account of another type of sporangium 
borne on Sphenopteris-fronds from the Paleozoic plant-beds of Zongouldak in Asia 
Minor, which he refers to Stur’s genus Discopteris. One of these sporangia (D. Rallii) is 
shown in Pl. 27. fig. 12; the narrow cells running along the upper face mark the line 
along which dehiscence takes place. The sporangia are without an annulus, but large 
polygonal cells with slightly thickened walls occur on the dorsal face, and similar cells 
are shown at the distal end of the figured example. This occurrence of rather thick- 
walled cells led Stur to describe sporangia of the Discopteris type 1, which he placed in 
the Marattiaceee, as having a rudimentary apical annulus. Zeiller § interprets the 
structure of the wall as being intermediate in character between that of a sporangium 
in which there is no differentiation, and one which is characterized by a well-defined 
group of thick-walled cells such as occurs in the Osmundacee. 
Among the Cycadofilices, interesting examples are recorded of anatomical features 
recalling those of the recent Osmundaceæ. “The nearest analogy to Lyginodendron 
among recent Ferns is probably to be found in Osmunda.... The leaf-trace bundles, how- 
ever, unlike those of Lyginodendron, become concentric immediately on leaving the stele.” 
Another difference from the fossil genus is that in Osmunda the phloem forms a continuous 
ring, which does not seem to have been the ease with the primary phloem of Lygino- 
dendron. The position of the protoxylem is also different. In Lyginodendron, as we 
have seen, al! bundles are mesarch, with the protoxylem strands embedded in the primary 
wood nearer its outer than its inner surface. In Osmunda the bundles only have 
protoxylem at all in the upper part of their course. In the bundles, where they first 
enter the stele, the protoxylem lies on the inner concave side of the xylem horseshoe. 
Lower down, where the horseshoe becomes a closed curve, the protoxylem comes to be 
surrounded by wood, so that here we have a certain resemblance to the Lyginodendron- 
bundle, though perhaps only an accidental one|. The more pronounced mesarch 
character of the xylem in Todea superba and T. hymenophylloides offers a still closer 
agreement with the primary xylem of Lyginodendron than is afforded by Osmunda. 
A comparison as regards the disappearance of the protoxylem strands in the lower 
part of their course has been made by Scott {| between Osmunda and Poroxylon, another 
member of the Palzozoie Cycadofilices. 
The unfortunate scarcity of petrified plants of Mesozoic age almost restricts us to 
evidence afforded by carbonized fronds, but the preservation is frequently good enough to 
* Bower (1891). + Zeiller (1890), fig. 11, p. 16. t Stur (1885), p. 150. 
$ Zeiller (1899), p. 18. i Williamson E Scott (1895), p. 667. $ Scott (1900), p. 403. 
