576 
MR, F. 0. BOWER ON THE COMPARATIVE MORPHOLOGY OF THE 
so close as was at first assumed by Sadebeck.* In these cases the two rows of pinnae 
correspond to the two rows of segments of the two-sided apical cell, though the 
individual pinnae have been shown by Kny not to correspond to the individual 
segments. It was thus suggested that a similar relation might be found in Osmunda. 
I have, however, been unable to discover any regular relation between the segments 
cut off from the apical cell and the individual pinnae : the position of the latter is 
such that they must arise partly from the tissues derived from the ventral series of 
segments, partly from those derived from the lateral segments (Plate 37, fig. 5). This 
conclusion may be put in relation with Kny’s observations on Ceratopteris, from 
which he finds that the pinnae do not necessarily coincide with the segments of the 
apical cell: in Osmunda, however, the absence of such coincidence applies not only to the 
cells of one marginal series of segments, but also to the series of segments themselves. 
The lower pinnae appear in their early stages of development as rounded masses 
of tissue, in which no marginal series of cells can be recognised even when the 
pinnae are viewed as in Plate 37, fig. 5. There is, it is true, a prevalence of cell- 
divisions in a direction perpendicular to the plane of the leaf over those parallel to 
that plane, but the distinction is not so complete as to mark off any marginal series. 
In the apical portions of pinnae in later stages of development, as well as in the 
apical portion of the phyllopodium when further advanced, and in the later developed 
pinnae, a marginal series may be recognised; in no case, however, is the marginal series 
so regular or so clearly marked as, for instance, in Aspidium Filix-Mas or Ceratopteris 
thalictroides. Thus, in Osmunda regalis the same mode of development by a 
marginal series of cells referable to a single plane is to be found as is prevalent 
throughout the leaf of the simpler Ferns. But in Osmunda it is relegated to the later 
stages only, viz. : at the apex of the phyllopodium and of the pinnae when they are 
far advanced, and also in the later-formed pinnae and the pinnules. 
It has often been described how the basal part of the leaf of Osmunda regalis is 
widened laterally so as to form a sheath, while in those leaves in which the upper 
portion is aborted and dried up, it is this wider basal portion which forms the scale- 
leaf first noted by Prantl. This widening is due to a transverse growth, chiefly in 
the peripheral tissues, including the superficial cells, the latter dividing repeatedly 
by periclinal walls. This growth in a transverse direction is most active at the lateral 
margins, that is at the points where a slight similar growth is found in other Ferns, 
e.g., in Onoclea germanica and Aspidium Filix-Mas : or where in other Ferns there are 
well developed lateral wings, similar to those in the upper parts of the leaf. There can 
thus, after comparing Osmunda with other Ferns, be no doubt that the massive lateral 
wings at the base of the leaf are homologous with the similar, but less massively 
developed wings at corresponding points in other Ferns. Moreover the massive wings 
at the base of the leaf in Osmunda may be traced as continuous in the young leaf from 
* Sadebeck, Verb. d. bot. Vereines. Prov. Brandenburg. Bd. 15, p. 129. Knt, Die Entw. d. Parkeriaceen, 
p. 40. 
