374 MR E. C. DAVIE ON 



portion of this adaxial strand ; the rest simply separate away again and resume their 

 position as the median strand (text-fig. 14, g and h). Thus both adaxial strands are 

 reinforced by tracheides from the median strand during the period of pinna-trace 

 departure. This process may well be compared with that which we have already 

 considered in the supply of the pinnules by the pinna-traces of Davallia pallida and 

 Asplenium lucidwn, where the pinna-trace was reinforced by a set of tracheides 

 appearing on the aporachial side before the pinnules were supplied. It is noticeable 

 that, while the leaf-traces which supply their pinnae on the marginal system are 

 simpler in outline than those supplying their pinnae extramarginally, there is a 

 development usually of hooks on the abaxial ends of the leaf-trace strands {e.g. 

 Davallia, PI. XXXIII. fig. 2 ; Gymnogramma Pearcei, var. robusta, PI. XXXIV. 

 fig. 12). These probably serve the same purpose as the reinforcing strands of the 

 leaf-trace of Davallia solida, or of the pinna-traces of Davallia pallida and Asplenium 

 lucidum — they carry on the water-supply past the pinnae, and prevent a too great 

 reduction of the water-supply as we approach towards the tip of the leaf. 



These explanations of the outlines of various leaf-traces are by no means provable 

 in every leaf-trace which may be considered. But when they are taken in conjunction 

 with one another and in relation to certain distinctive types of Fern-leaf they afford 

 us the opportunity for a broad generalisation which finds confirmation in the arrange- 

 ment of the species of Ferns in the table on p. 354. 



But we have still to consider the possible advantage to be derived from the 

 change from the extramarginal to the marginal type of pinna-supply. It seems to 

 have occurred in a passage from a larger-leaved to a smaller-leaved form of the Fern 

 (cf. above, p. 373). On that basis its occurrence is fairly easily to be explained. For 

 in the small-leaved form the adaxial portion of the leaf-trace would be small and 

 thin. To supply pinnae from the back of a thin " hook would be to give them a 

 very exiguous pinna-trace. A long strip of tracheides can be more easily given off 

 from an unincurved leaf-trace than from an incurved one. The width of the back 

 of the hook is the dictator of the width of the pinna-trace in the extramarginal 

 supply ; the width of the marginal strip is the width of the whole margin of the 

 unincurved leaf-trace. The back of a hooked leaf-trace is always "thin" as regards 

 tracheides, and the pinna-traces which come off from it can give only a narrow strip 

 of tracheides. Once the difficulty of carrying forward the water from pinna-departure 

 to pinna-departure along the leaf was solved (by the development of hooks on the 

 abaxial ends of the adaxial strands) the marginal type of supply must have been 

 found an advantage rather than a drawback, because of its possibilities in supplying 

 a long strip of tracheides to a departing pinna. This may explain its retention in 

 the genus Polypodium, where the abaxial complications of the leaf-trace are often 

 very elaborate and involved, and in some species of Pteris (P. cretica, P. umbrosa, 

 etc.) where the abaxial arch and the lateral swellings of the leaf-trace are fairly 

 prominent. The margins of the leaf-traces could be extended almost indefinitely to 



