Bower . — Studies in the Phytogeny of the Filicales. 453 
ture ; and this goes along with other primitive characters named. We may 
now therefore conclude more confidently than before that in the Cyatheae 
the dendroid type is secondary and derivative. 
DlCKSONIEAE. 
But there is the other family of dendroid Ferns to be considered, 
viz. the Dicksonieae. They are clearly marked off from the Cyatheae 
by the position of their sori. While these in the Cyatheae are constantly 
superficial in origin, as in the Gleicheniaceae, the Dicksonieae have their 
sori as constantly marginal in origin, corresponding in this feature to the 
Schizaeaceae. Prantl has demonstrated the marginal origin of the sporangia 
in the various genera of the Schizaeaceae in his great monograph on 
the family (‘ Schizaeaceen,’ pp. 39-45). But the detailed evidence has never 
been fully given for the Dicksonieae, so as to relate the origin of the recep- 
tacle of the sorus, or the sporangia themselves, distinctly to the marginal 
series of cells of the developing leaf. It has been stated by Burck (‘ Indusium 
der Varens,’ p. 43) and by Gluck (‘ Die Sporophyll-Metamorphose,’ Flora, 
1 895, p. 19) that it is so ; but the mere statement that it is so, without the 
recognition of the segmentation which leads to it, amounts to nothing more 
than a bare assertion. On the other hand, the present Fig. 11 of Dicksonia 
Sckeidei demonstrates the segmentation of the leaf-margin, and how while 
the lower and upper indusium flaps arise as upgrowths from the surfaces of 
the leaf, the receptacle originates from the marginal cell itself of the section. 
It will be shown below that the same holds also for various members of the 
Dicksonieae and other related Ferns. We may provisionally accept the 
essential distinctness of the Cyatheae from the Dicksonieae, as based upon 
this constant difference of origin of their sori. The question will now 
be taken up of the habit and structure of the Dicksonieae for comparison 
with what has been seen in the Cyatheae. 
Gwynne-Vaughan, in his papers on Solenostelic Ferns (‘Ann. of Bot./ 
vol. xvii, p. 689, &c.), has dealt with various types which belong to the 
Dicksonieae, and in fact his best examples have come from that family, or 
its immediate derivatives. He has described for Cibotium (. Dicksonia ) 
Barometz how the stelar condition ‘ must be regarded as dictyostelic, 
although it is very near solenostely \ The overlapping of the leaf-gaps, 
which he noted, depends on the degree of elongation of the creeping axis ; 
where the internodes are relatively long the solenostelic state, without any 
medullary strands, is typically seen in the bulky axis. As he points out 
(1. c. p. 709), lateral shoots are formed ‘ at the back of the leaf-trace * ; in fact, 
they correspond in position to those of Lophosoria , or Metaxya. Thus 
C. Barometz may be held to have, as regards habit and vascular structure, 
a similar relation to the dendroid Dicksonieae to that which Metaxya 
bears to the dendroid Cyatheae. The examination of its runner, or lateral 
