670 CONCLUSION 
Ophioglossum vulgatum (Figs. 260 and 260 d7s, p. 466) and in Botrychium 
(Figs. 261, 262, p. 468) the apex originates from approximately the centre 
of the epibasal hemisphere, and notwithstanding that there is an early 
displacement owing to the precccious development of the first root. ‘The 
facts thus suffice to support the general statement, that whatever the other. 
fluctuations of form'of the Pteridophyte embryo may be, all the exactly 
investigated types show the apex of the axis to originate in close relation 
to the intersection of the epibasal octant-walls. 
It is accordingly recognised that the very first step that can be 
observed in the embryogeny involves the definition of its polarity, and 
that the apex of the shoot bears a constant and close relation to the 
centre of the epibasal hemisphere. The base of the primitive shoot thus 
defined is the suspensor where that part occurs; where it is absent the 
centre of the hypobasal hemisphere may be held to mark the base of the 
primitive axis.1 The whole embryo thus appears from the first as a 
radially constructed spindle upon which appendages may be borne: these 
are of the nature of leaves and of accessory roots, and they may vary in 
number, and in position and time of origin, causing thereby marked 
variations in the early structure, which are for the most part open to 
biological explanation. The embryo is, however, subject also to early 
distortion in various ways, through the formation of swellings of the nature 
of haustoria, or of storage tubers: or it may be that modifications of 
form arise in relation to the precocious development of some one 
appendage and the correlative delay or diminution of another, or even 
of the axis itself. Though such modifications are probably secondary, 
yet they have produced such peculiarities of form and aspect in the 
embryos in which they appear that the originally radial form of the shoot 
is disguised, and its morphology has consequently been misunderstood. 
Examples will now be qpoted illustrating these various points. 
The origin of the cotyledon in Ferns is constant in time and place: 
this is probably related to the prone position of the embryo, and to its 
importance for early nutrition in replacing the supply derived from the 
small and evanescent prothallus. But in other cases there is less constancy : 
in ELguisetum there may be sometimes two, though usually three cotyledonary 
leaves in the first whorl. In Z. Se/ago and Phlegmaria one cotyledon takes 
precedence, soon followed by a second leaf (Figs. 183, 184, 185), but 
in LZ. clavatum two equal cotyledons are formed (Fig. 186). Again, in 
Selaginella Martensit two equal cotyledons appear very early; in S. 
spinulosa Bruchmann specially notes that though the two cotyledons may 
1Some previous writers have held the primary axis to run from the stem apex 
obliquely to the apex of the first root. ‘Reasons will be shown below for regarding the 
root as an accessory part, commonly latéral, and not determinate in position. Its 
growth may in certain cases approximate to the original axis of polarity of the shoot, 
as it does in the embryos of Ferns; but this is held to be an occasional and accidental 
rather than an inherent character, as is indicated by a comparison with the more bulky 
embryos of Agzzsetum and the Ophioglossaceae. 
