610 FILICALES 
vascular type to that seen in the more complex Cyatheae, the probable 
progress has been as suggested in the ontogeny of 4. exce/sa, while in 
the sorus the basal indusium appears as a new structure, and the individual 
sporangia are liable to diminution in size and spore-output, as is exemplified 
in the extreme form in Cyathea dealbata. Thus there seems good reason 
to see in the Cyatheae a series having probable genetic relations with the 
Gleicheniaceae, but advanced on the one hand to the basipetal succession 
of the sori, and on the other to a high complexity of the vascular system. 
This conclusion is in agreement with the palaeontological facts, for 
representatives of the Cyatheae have been recognised as present from 
Jurassic times onwards. It is, moreover, specially interesting to note that 
the genus A/sophila is among the earliest of the fossils referred with certainty 
to this family, as exemplified by A. polonica, described by Raciborski 
from the fire-clay of Krakau.1 
Of Ferns in which evidence of a basipetal sequence of the sporangia 
in the sorus has been observed there remain Oxoclea, Sphaeropteris, and 
Diacalfe, all genera in which the position of the sorus is superficial and 
the indusium basal. The natural place for these genera appears accordingly 
to be in relation to the Cyatheae.2 The annulus in these Ferns is almost 
vertical: in Sphaeropieris it is slightly oblique, and may be traced as 
continuous past the insertion of the stalk of the sporangium, as is 
characteristic of the Gradatae; but in Déacalpe and in Oxoclea the annulus 
is interrupted at the insertion of the stalk. These genera appear to 
illustrate how, when the basipetal succession is not long continued, and 
the orientation of the sporangia not strictly maintained, the annulus is no 
longer markedly oblique, but passes over into the vertical position, and may 
even be discontinuous at the base. This will be noted also in the 
Dicksonia-Davatlia series. 
* 
SALVINIACEAE. 
It is impossible to leave the Gradatae without mention of the peculiar 
little group of heterosporous water Ferns, of the genera Sa/vinia and 
Azolla. They have been so exhaustively described elsewhere that it will 
be unnecessary to give any detailed account of them here, especially as 
they are in all probability a side branch from the main series. Examination 
of their sori shows resemblances to the sorus of the Gradatae: it is, however, 
difficult to attach them on these, or on more general grounds to any actual 
genus of living ferns.* It would seem probable that the type from which 
they sprang was homosporous, having an elongated receptacle upon which 
arose a basipetal succession of sporangia, with short thick stalks, and each 
containing 16 spore-mother-cells. That with the differentiation of the spores 
followed certain other modifications, such as a decrease in number of the female 
sporangia, and perhaps an increase of the male sporangia: the former is 
1 Abhandl. Akad. Wiss. Krakau, xviii., 1894. 2See Studies, iv., p. 55-58. 
