470 Bower —Studies in the Phytogeny of the Fili cates. 
than any significant fact for morphological comparison in Ferns at large. 
Such cases as these are few and isolated, so that they do not appear to 
indicate any stable modification of the usual marginal arrangement. 
But quite a different view must be taken of that gradual and very 
general modification of position of the sorus which is seen in the derivatives 
of the Dicksonioid series. Here, with the sori typically marginal in origin, 
as they are now demonstrated to be in Thyrsopteris , Cibotium , Saccoloma , 
Odontoloma , and Davallia , there is seen to be a phyletic drift towards 
a position of the sorus upon the lower surface. It is probably a biological 
adaptation by which better protection is secured. But even in cases where 
the indusial lips have become very diverse in size and structure, the ‘ upper ’ 
having grown to all appearance like a continuation of the foliar expanse, 
still the receptacle is as truly marginal in origin as in the types where both 
lips are practically equal. It is thus clear that in. these facts there is 
nothing to detract from the importance of the marginal position in its 
phyletic aspect. All that we see is a biological adaptation which tends to 
mask the original position of the sorus on the margin of the leaf. 
The converse, viz. the approach of the sori to the marginal position 
in certain Ferns where they are typically superficial, is seen in the Pterideae. 
In simple cases, such as Plagiogyria and Adiantum , it is quite obvious that 
these are merely special cases of the location of superficial sori near to the 
margin. The consequence is, however, that a curious convergent phylesis 
has occurred ; so that the appearance of some truly marginal types, such 
as Lindsaya , is very closely similar to that of some superficial types, such as 
Pteris. It may, indeed, be the case that sometimes the line of demarcation 
has not been rightly drawn. Developmental evidence from the earliest 
stages will have to decide in questionable cases. 
An almost equivalent convergence is seen between the Aspidieae, which 
arose from a superficial origin, and such genera as Nephrolepis and Oleandra , 
the origin of which appears to have been from marginal types. Here, 
again, it is upon a basis of observation of the earliest stages of development 
of the sori, combined with wide comparison by other criteria, that it may 
be possible to allocate such converging types to their respective phyletic 
sources. 
But the Davallieae, the Pterideae, and the Aspidieae are all relatively 
advanced types. None of the Ferns quoted as showing this convergence 
are really primitive. It follows that, notwithstanding such instances of con- 
vergence, and notwithstanding the modifications from type, sudden or 
gradual, such as have been above mentioned, a great body of fact establishes 
this conclusion : That from a very early period the Leptosporangiate Ferns 
have progressed along two parallel lines , the one characterized by a marginal , 
the other by a superficial position of the spore -bearing organs ; and that these 
lines have remained phyletically distinct throughout . 
