the Ophioglossaceae . 9 
the U-shaped petiolar trace, and hence must be considered a single 
pinna. 1 
Cases where one or more pinnae of the sterile segment bear sporangia 
have been earlier figured for several species, especially B. L unaria (see 
Bower, 6, p. 443) ; Fig. 25 represents a case of this kind occurring in 
B. obliquum. The branch of the sterile segment which appears to be 
median in this figure is really a lateral branch which has been turned 
upward in preparing the herbarium specimen. An examination of the 
specimens of Botrychium in the Gray Herbarium shows that (1) almost 
all species occasionally show the feature illustrated in Fig. 25, and (2) in 
the smaller-leaved species the sterile leaflets frequently bear a few spo¬ 
rangia along the margin. The latter case resembles the half-sterile pinnae 
which may be seen on almost any plant of Osmunda regalis. These 
cases indicate that the whole leaf was at one time fertile, and that in the 
course of evolution a division of labour has occurred (cf. Osmunda Claytoniana ), 
resulting in the sterilization of all of the pinnae except the basal pair, which 
have become specialized by turning upright (as is the case in Aneimia) and 
finally fusing. But in certain cases reversions occur, giving a clue to the 
probable evolutionary history of the genus. 
Bower rejects the evidence afforded by these ‘ sports ’, considering the 
cases of double spikes merely instances of splitting or chorisis. I do not 
think that the cases here adduced can be disposed of in this simple way. 
If three or four spikes arose from the same point, the objection might be 
more significant, but in the plants of B. obliquum with three spikes two 
are paired and the third larger one arises from a lower point on the petiole. 
It seems clear that in this case there are two pairs of fertile pinnae, the 
basal pair of which are fused. Even in case two spikes are present they 
may arise from different points and may have a single or double vascular 
supply, as has been shown. The structural features are completely and 
easily explained on the view that a spike represents one or more pinnae 
of a leaf. Bower admits the cogency of the argument derived from the 
occasional fertility of pinnae which are usually sterile, but refers to the 
view of Goebel that sporangia are parts sui generis, and hence may occur 
on members of different rank. Apparently Bower does not endorse this 
view, and surely such an argument should be reserved as a last resort, 
for to admit it would rob comparative morphology of much of its meaning. 
Among the species of Ophioglossum which have had their petiolar 
structure examined, O. lusitaniemn has a comparatively simple structure. 
Prantl’s account ( 18 ) shows that the leaf-trace early divides into three 
strands, one central and two lateral; these three arrange themselves as 
the angles of a triangle, and from each lateral strand there breaks off 
1 See Note at end of paper. 
