THE PINNA-TRACE IN THE FEENS 363 



shows that at first the marginal type prevails. And an examination of the ultimate 

 branches of the Fern leaf reveals the marginal type in the great majority of species. 



We have then to explain how this variation of type has come to be, and whether 

 it can be related to any prominent features in the development of the vascular system 

 in the Ferns. To what is the variation due ? Can we trace it to the differences in 

 the construction of the stele ? Is it dependent on the length of the leaf, on the 

 number of pinnae, or on the state of division and subdivision of the pinnae themselves ? 

 Or is the outline of the leaf-trace responsible for the differing points of departure of 

 vascular supply for the appendages of the leaf ? 



Sinnott has based his classification of leaf-traces upon the number of protoxylem 

 groups present in them. On this basis he has separated Osmundaceae from Marat- 

 tiaceae, and these from all remaining forms. The Osmundaceae are protagonists of 

 the extramarginal type of pinna-supply ; the Marattiaceae are divergent but consistent 

 adherents of the same type ; the real differences occur in " all remaining forms." 

 The explanation of pinna-trace type cannot be in the number of protoxylem groups 

 in the leaf-trace. 



We are brought no nearer to an explanation by the work of Parmentier ('99), # 

 whose classification of the Ferns of France on the structure particularly of their 

 petioles is not too reliable. For his type-sections of these petioles have apparently 

 been made at varying levels, and his arrangement depends much upon the comparison 

 of sections taken at different distances from the base of the petiole. No attempt is 

 made to compare one leaf-trace with another as they run from end to end of the leaf. 



Two notable contributions to the knowledge of the evolution of the Filicinean leaf- 

 trace are those of Thom,e ('86) and Bertrand and Cornaille ('02). In both of these 

 papers a comprehensive survey of Fern-petioles is made and distinctive types are 

 contrasted with one another. 



From the leaf-traces of Phegopteris, Onoclea, Aspidium, Thom^e makes an 

 " A spidium- type," which he joins with the " Asj^lenium-type " and the intermediate 

 " Struthiopteris germanica-ty-pe " as a " Polypodiaceous type." 



Against this he sets the " Cyatheaceous type." The transition from the one to 

 the other is through Hypolepis tenuifolia,^ Pteris tremula, and Hypolepis repens. 



He makes the Marattiaceae and Cibotium exceptional modifications of these types. 

 He concludes that the different families of Ferns can be distinguished by the charac- 

 ters of the leaf-trace, but that a systematic arrangement of Ferns on anatomical grounds 

 is quite outside of any serious consideration, and that only broad distinctions can be 

 confirmed by the anatomy of the petiole. 



The most interesting feature of his classes is the selection of the bridge from the 

 Cyatheaceous type to the Polypodiaceous type. For we have seen that the pinna- 



* Cf. PELOURDE ('06). 



t In quoting the work of Thom^e and Bertrand and Cornaille I retain their nomenclature of the Ferns they 

 describe. 



