82 
F. 0. BOWER. 
transverse sections of the roots, in the former case of Equi- 
setum hiemale, in the latter of Pteris hastata, at points 
immediately below the apical cell. The correspondence of 
arrangement of the cells in these sections with that of the apical 
meristem of Osmunda, as seen in figs. 6 — 8, is undeniable, 
though certain irregularities, which are less common in Equi- 
setum and Pteris, are frequent in Osmunda. Thus it will 
be observed that the sextant walls in the former usually curve 
all in the same direction, and are homodr omous, according 
to the terminology of Naegeli and Leitgeb (1. c., p. 105) ; they 
accordingly insert themselves successively on the three prin- 
cipal walls. This is not the rule in Osmunda (figs. 6 and 7), 
but a similar irregularity is to be observed in the fig. 9, of 
Equisetum, in which two of the sextant walls are inserted on 
opposite sides of one of the principal walls. This hetero- 
dromous arrangement (Naegeli and Leitgeb, 1. c., p. 105) of 
the sextant walls, which appears to be the exception in Equi- 
setum, is not unfrequent in Osmunda. The further con- 
sideration of this interesting point must be deferred for the 
present. 
Before leaving the study of transverse sections it may be 
observed that they show that the divisions of the segments 
which go to form the root-cap appear to be so arranged that, 
whatever the form or number of the initial cells, each successive 
addition to the root-cap consists of a group of four cells. This 
is illustrated in figs. 3 and 7, in which the group of four cells 
may be recognised, though derived from initial cells of very 
different types. A similar result was obtained by Naegeli and 
Leitgeb, in their observations on Equisetum and the 
Polypodiacese. 
Lastly, I may state that no clear case of four initial cells, 
such as have been described by Schwendener for the Marat- 
tiacese (1. c.), has been observed by me in any of my prepara- 
tions from the root of Osmunda regalis. 
