THE PINNA-TRACE IN THE FERNS. 365 



divergent pinna-traces of these species. These Ferns may all rightly be of the 

 Onoclean type, but that avails nought to explain why they supply their pinnae in 

 such very different ways. 



But the leaf-trace headlines Osmundean, Cyathean, Onoclean, and Marattian stand 

 clear and are suggestive. From the Osmundean and Onoclean groups there may be 

 drawn types of leaf-trace which agree in supplying their branches on a definite plan 

 which is seen in the earliest known Ferns, and which is found in those Fern-leaves 

 which occur first in the development and in those parts of Fern-leaves which are 

 regarded as primitive. And from the Onoclean group, too, can be taken other types 

 which are now known to be high in the scale and which supply their pinnae in a 

 manner, so to speak, improved upon that dominant in the Osmundean and Onoclean 

 groups. The marginal type of pinna-supply cuts up the Osmundean and Onoclean 

 groups when it is applied as a differential factor. And the marginal type is at once 

 the most primitive and the most advanced. 



This presence of the marginal type at the bottom and at the top of the evolutionary 

 ladder brings out prominently the need for discovering the factors that produced the 

 intermediate extrarnarginal type. This is altogether dependent on the presence of 

 the incurved hooks at the ends of the leaf-trace. It only appears where they are 

 present, though sometimes hooks are present in leaf-traces that supply their pinnae 

 marginally. The evolution of the hooked leaf-trace must be the first factor in this 

 development of the extrarnarginal type of pinna-trace. 



This evolution is suggested by Kidston and Gwynne-Vaughan ('08) in their 

 description of the leaf-trace of Thamnopteris Schlechtendalii, Eichw. (text-fig. 7). 

 There the leaf-trace arises as a swelling of the xylem of the stem. It passes off 

 without leaving any gap or depression in the stem-xylem, then protoxylem appears 

 almost centrally in it, an island of parenchyma occurs on the adaxial side of the 

 protoxylem, this island increases in size, the protoxylem groups widen along the 

 bay and the curved trace is produced, followed by a curving of the petiolar outline. 

 These changes in the individual leaf-trace are held to indicate the changes under- 

 gone in the ontogeny and phylogeny of the adaxially curved leaf-trace so repre- 

 sentative of the Filicales. The adaxial hooks thus remain as part of the adaxial 

 side of an originally solid leaf-trace. 



Such a solid leaf-trace does occur in several of the Zygopterideae. 



In Dineuron ellipticum, Kidston, and D. pteroides, Renault, the leaf-trace has 

 an elliptical strand of xylem grooved at the ends. The pinna-traces depart from 

 these ends as curved bars leaving no gaps in the tissues of the leaf-trace (text-fig. 8, a). 

 In Metaclep>sydropsis duplex, Williamson, the solid mass of tracheides in the leaf- 

 trace has the outline of an hour-glass. The pinna-traces go off from the dilated ends 

 of this leaf-trace (text-fig. 8, c). Diplolabis Romeri (Solms) supplies its pinnae from 

 the same relative position of the leaf-trace as do the two preceding species, but the 

 arms of the leaf-trace are highly developed here in the normal petiole, though at its 

 TRANS. ROY. SOC. ED1N., VOL. L. PART II. (NO. 11). 50 



