io Chrysler .— The Nattere of the Fertile Spike in 
another strand ; the last two move towards each other, fuse, and supply 
the fertile spike. The plan shown in O. vulgatmn , as described by Holle 
( 14 ) and verified in my own specimens, is closely related to that in O . lasi- 
tanicum , differing chiefly in that the petiolar strands are more numerous, 
that is, the three strands break up into about ten (there may be as many 
as fifteen) and are disposed in an ellipse, or rather a semicircle, for the 
strands on the adaxial side generally form nearly a straight line. The 
latter strands represent the right- and left-hand edges of the original 
petiolar strand, and they higher up anastomose to form three to five 
strands which pass up into the fertile spike, while the remaining strands 
branch extensively as they pass up into the sterile segment. The plan 
here exhibited differs from that seen in B. ternatum in that the vascular 
supply of the petiole early breaks up into a number of strands instead of 
remaining a single U- or C-shaped strand. In both genera the vascular 
supply of the fertile spike is derived from the two free edges of the main 
petiolar system. This is well illustrated by the small species O. Bergianum , 
of which Bower (6, p. 463) says ‘ the single leaf-trace strand may long 
remain undivided, giving off two lateral strands which fuse on the adaxial 
side to form the supply for the spike : further up the strands of both 
sterile lamina and of fertile spike may branch again’. Campbell ( 9 ) has 
figured the disposition of the vascular bundles in the petioles of certain 
other species belonging to the section Enophioglossum. Assuming that 
the origin of the curved series of bundles in these petioles is the same as 
in those in which the origin has been traced, namely, from repeated 
division or branching of a curved leaf-trace, Campbell’s figures show that 
several other species, including O. moluccanum and O. californicum , agree 
with O. lusitanicnm in having the vascular supply of the fertile spike 
derived from the edges of the curved leaf-trace. 
In the section Ophiodernia the leaf-trace at its origin consists of from 
four to five strands arranged in a U, as is clearly shown by Bower’s figures 
( 5 ) of O. pendulum and O. simplex. In leaves which bear a fertile spike 
the edges of the U soon close in to form a circle, and several strands which 
represent the two free edges of the trace constitute the supply of the fertile 
spike, exactly as in Euopldoglossum. This may be seen from Campbell’s 
figures of O. intermedium and O. pendulum , though it is only fair to say 
that this writer draws a different inference from his studies of these species. 
The section Cheiroglossa , represented only by O. palmatum , is the 
most specialized in the genus. As in Ophiodernia , the vascular supply 
of the leaf arises as several strands ( 5 ) which form a leaf-trace, becoming 
more and more curved until a circle of over a dozen strands is formed. 
From the adaxial side of this circle break off about four strands to supply 
the fertile spike. So far the behaviour is exactly as in Ophioderma , but, 
as Bower’s figures show, from what now constitute the free edges of the 
