16 Lady Isabel Browne. 
typically dictyostelic, external and internal phloem being continuous 
through the gaps between the meristeles (28). Since the annular 
stele of the Osmundaceae appears to have been at first unbroken by 
leaf-gaps they suggest that the internal phloem probably arose at 
the branching of the stem, as in Osmunda cinnamomea, and was 
continuous with the external phloem through a ramular gap, and 
that the connection of internal and external phloem through the 
leaf-gaps, as seen in Osmundites skidegatensis , is a later modification 
connected with the development of foliar gaps (29). The view that 
the internal phloem is a recent acquisition is also supported by the 
absence of an internal endodermis in young plants of 0 . cinnamomea, 
and by the absence of any correlative reduction of the leaf-trace in 
recent Osmundaceae (43). 
The fossil Osmundaceae seem to be closely related to the 
Botryopterideae. The protostele of Zalesskya recalls that of the 
simpler Botrypterideae. This oldest Osmundaceous genus also 
approaches the simple Tubicaulis and Grammatopteris in that nearly 
the whole of the wood is centripetal, as it still is in Todea hymeno- 
phylloides. Most recent Osmundaceae are endarch and the whole of 
the centripetal wood has been replaced by pith. This gradual 
development of the central wood as non-conducting tissue and the 
correlated acquisition of centrifugal wood had already been initiated 
in Zalesskya, where the central part of the centripetal xylem was 
no longer water-conducting and the protoxylem no longer absolutely 
peripheral. The multiseriate pitted tracheae of the Botryopterideae 
are found in the Osmundaceae, and are confined among recent ferns 
to them and the Ophioglossaceae (22). The structure of the 
Osmundaceous sporangium in which the annulus is a group of cells, 
also favours the hypothesis that the Osmundaceae may have been 
derived from the simpler Botryopterideae or from their near allies. 
The Hymenophyllaceae differ strikingly from the Osmundaceae 
in the anatomy of their stems and fronds and in their sporangial 
characters, so that the assumption of a filmy habit by some 
Osmundaceae and the occasional production in this order of sub- 
filamentous prothalli are clearly due to development parallel to the 
Hymenophyllaceae. 
The discovery of Zalesskya has brought the Osmundaceae 
nearer to Lygodium, the most primitive of the Schizaeaceae, though 
the exarch homogeneous stele of the latter genus appears to be 
more primitive than the stele of Zalesskya. The writer cannot 
agree with Dr. Campbell, who regards the Schizaeaceae as the 
nearest relatives of the Osmundaceae (13). The structure of their 
