730 MR W. T. GORDON ON 



chymatous pith with a system of internal tracheides. The other known zygopterid 

 steins are still more complicated and distinct. Thus, among the known Zygopteridese 

 the stem of D. romeri is the most primitive in its anatomy, and it is also among the 

 earliest known zygopterids in time. 



The position of this genus in the zygopterid alliance is exactly paralleled by that of 

 TJtamnopteris schlechtendalii (Eichwald) among the Osmundacea?. In this latter species 

 a solid protostelic stem-xylem is also present, but again it is not the most primitive 

 stele conceivable on theoretical grounds. The internal short tracheides are reticulately 

 thickened, but the outer long elements have multiseriate pits on their walls. Probably 

 more primitive examples of the Osmundacese possessed porose or reticulate tracheides. 



We have seen that the shape of the trace at different levels in the petiole is very 

 distinct. If sections cut near the base of the petiole were compared with sections cut 

 higher up, they would be classed in different genera according to current nomenclature. 

 Of course, it is just at the base of the petiole that we would expect ancestral characters 

 to be preserved, and this, I think, is what actually occurs. The trace at first is elliptical, 

 with a sunk protoxylem group at each end, but without any islands of parenchyma. 

 In one specimen examined, this elliptical shape is retained even after the islands of 

 parenchyma do appear, and then the trace bears a strong resemblance to that of Dineuron. 

 In all the other examples, however, the trace has become dumb-bell shaped before the 

 islands are present. At this stage (PL II. fig. 22) the trace resembles to a great extent 

 that of Clepsydropsis antiqua. The single elliptical specimen, on the other hand, is 

 distinctly similar in form to Dineuron. 



Returning to the normal Clepsydropsis stage we find that, at the first departure 

 from the petiole-trace, viz. the trace of the root-like body (or, it may be, very reduced 

 pinna-trace-bar), a constriction forms across the island of parenchyma. On this trace 

 departing from the petiole-trace the end remains open, whereas in Clepsydropsis the 

 constriction becomes complete and the departing trace is a closed ring, while the end of 

 the trace is also closed. Further along the petiole, in D. romeri. the short arms (PI. II. 

 fig. 24, and PI. III. fig. 33), which are formed by the first departing trace leaving the 

 end of the petiole-trace open, become larger and longer until a stage closely resembling 

 that of the normal petiole-trace in Zygopteris primaria results. Still passing upwards 

 in the petiole, the arms of the trace become more and more prominent, and the 

 horizontal bar less so, until the normal Diplolabis type is attained. In this one petiole, 

 therefore, the trace at different levels is exactly similar to the characteristic petiole- 

 trace in three of the more primitive members of the Zygopterideae (if we regard simplicity 

 of the xylem of the petiole as synonymous with primitiveness). There are thus the 

 ( Clepsydropsis, the Dineuron, and the Zygopteris stages. When the stems of any of 

 these three genera are discovered, we shall expect to find that their stem is at least as 

 simple in structure as that of D. romeri, and probably more so. 



In brief, the changes in the petiole-trace of Diplolabis, as we ascend in the petiole, 

 may be stated as follows: — The arms of the H or X become more and more important 



