Bower.—Studies in the Phytogeny of the Filicales. VII. 43 
the whole width of the receptacle, as seen in vertical section, may be 
occupied by them (e). Where a sporangium is included in the section the 
hairs may be on either side of it ( d ). Now these facts do not fit with 
Mettenius’s theory of the origin of the inner indusium by a fusion of hairs— 
a theory which is quite out of harmony with the comparisons which are 
being developed in this memoir. Using the method of Mettenius it would 
be equally possible to suggest that the hairs seen in Fig. 30, c, were in 
course of formation of a transverse indusium ! But what is fatal is that 
there is no constancy of the hairs in the position where the inner indusium 
should be : nor has any webbing been observed in them. It is not intended 
Fig. 30. a-c. Sections through the young sorus of Anopteris hexagona (L.), C. Chr. ( = P ter is 
heterophylla , L.), showing the origin and relation of the soral hairs. ( a , b , x 125 ; c, x 85.) 
to deny that any of the hairs can be held as representing an inner indusium. 
But if any of the hairs are of that nature phyletically, they represent not an 
inner indusium in the making, as Mettenius suggested, but the degenerate 
relics of an inner indusium which, being of no further use, is in course of 
dissolution. The view which I take, however, is .that the hairs seen in 
the sorus of Anopteris hexagona are simply paraphyses scattered over the 
receptacle, as they are in other species of Pteris ; though here they are more 
closely arranged than usual. 
In Pteris serrulata , L. fil. (= P. multifida , Poir.), the sporangia are 
again of intramarginal origin, and are produced almost simultaneously 
in considerable numbers (Fig. 31, a) ; but the mixed character of the 
sorus becomes almost at once apparent ( b ). The segmentation clearly 
shows that it is the actual margin which grows directly into the protecting 
