142 ASPLENIOID FERNS [ch. 



both the offices of nutrition and of propagation, the sori will tend to follow 

 the widening surface of the blade. Either they will be increased in number 

 by some form of duplication, or they will be themselves extended. Examples 

 of this appear in very striking form in the Marattiaceae, and they have been 

 traced in relation to the widening leaf-area in Vol. II, Chapter XX. They 

 illustrate the results of progressive extension of the sorus in the so-called 

 Archangiopteris, Protomarattia, and in Danaea: but subdivision probably 

 accounts for the spread of the numerous circular sori on the broadened leaf 

 of Christensenia. The same alternative was presented to the Ferns now 

 under discussion. The theory to be elaborated here is that while Dryoptei'is 

 in its broad-leaved forms repeated or reduplicated its individual sori, in 

 Asplenijim the individual sorus was subject to extension along the veins of 

 the widening area of the blade, appearing as the linear sori shown in 

 Fig. 669. The question will be what relation these linear sori of the 

 Asplenioid Ferns may bear to other well-known types of sorus elsewhere. 

 The enquiry will naturally be directed towards the Dryopteroid type, partly 

 by the similarity of vegetative structure in the two families, partly by those 

 hooked ends and other unusual soral forms already mentioned as occurring 

 in Aspleniiiiii. 



If we examine a large run of herbarium specimens of the genus, certain 

 species are found to present exceptional forms of sori in considerable numbers. 

 A specially favourable species is found in Diplazhnn lanceuiu Thunb., which 

 was figured by Hooker and Greville as A. siibsiimatwn {Ic. Fil. Vol. I, PI. 2y). 

 A single leaf of this species, now in the British Museum, collected by Wallich 

 at Napalia in 1829, yielded sori which appear to link the type o{ Asplenium 

 unmistakably with that of Dryopteris. Already in the Dryopteroid genera 

 Didyvioddaetia and McsocJdaena, and to some extent also in Luersseuia, an 

 elongation of the sori appears, following the outwards-running veins to the 

 margin of the widened blade. But here the receptacle remains fertile round 

 the distal curve, and the extension appears equal on both sides. If we imagine 

 it interrupted at the curve, the typeof sorus would approach thaXoi Diplasmm ; 

 if only one side were developed, the sorus would rank as Aspleniimi. All of 

 these conditions were found on a single leaf of Diplazium lanceiim (Fig. 672, 

 1-6). Occasionally the general form of the sorus did not appear to differ 

 materially from that of one of these Dryopteroids d): but others may be 

 unequally developed on the two sides (2, 3, 4). In others again the fertility 

 of the distal end of the receptacle may be lost, and the sorus takes the 

 character of Diplazium (5): or one side only (and here the anadromic side) 

 of the sorus is developed, it then has the character of Eii-Asplenium (6). 

 D. lanceuiu has an open venation, and the position of the sori relatively to 

 them is the same as that in Diyopieris. 



In all this we may see a natural development of the lop-sided or z}'go- 



