THE STRUCTURE AND AFFINITIES OF DIPLOLABIS ROMERI (SOLMS). 729 



distinguished from D. romeri, or at least, although the original specimens appeared 

 different, there are no characters which might not easily occur among different 

 individuals of the same species. The only difference between D. forensis and D. 

 romeri is apparently the greater length of the horizontal bar of the trace as compared 

 with the length of the arms, and its consequent relatively thinner appearance. This is 

 exactly the sequence shown in Diplolabis romeri in the same petiole, as we trace it 

 upwards. Indeed, PI. II. fig. 28, would be called D. forensis, and fig. 29 D. romeri. I 

 therefore propose to include D. forensis under the name D. romeri. 



I have already mentioned that D. romeri had a creeping rhizome, and the evidence 

 on which the assertion is made may be briefly stated here. The grounds for this are 

 (1) the internodes are very long; (2) the stem frequently branches; (3) the xylem of 

 the stem is about 4 mm. diameter while that of the petiole is 6 mm., i.e. the petioles 

 are very heavy in proportion to the stem ; (4) the pinnae are arranged in four ortho- 

 stichies, and the petiole must therefore have been held erect ; (5) the roots are given off 

 irregularly and in all directions ; (6) there is no evidence of any sclerenchyma in the 

 stem. Taken together, these afford sufficient evidence in support of the contention that 

 the stem was a creeping rhizome, the petioles being held erect as in Pteris aquilina. 



The elements of the stem-xylem are reticulate tracheides and there is no conjunctive 

 parenchyma. The xylem, however, consists of two zones, the inner of short elements 

 and the outer of long elements. The stem is therefore a good example of a protostele. 

 The elements are all reticulate, and therefore belong to the most primitive type known ; 

 but, since there are two distinct types of tracheide in the stem, the latter cannot be 

 considered the most primitive possible protostele. In more primitive stems we would 

 expect to find only one type of reticulate tracheide, namely, the long type, and, in that 

 the tracheides of the inner xylem of D. romeri are usually arranged in vertical rows, 

 this more primitive stem is distinctly indicated. It appears as if the short tracheides 

 were derived from long ones, by the latter becoming segmented. The stem-xylem of 

 D. romeri, therefore, probably represents a stage in the development, first, of a stele 

 with a mixed pith, and finally of a medullated monostele. 



The discovery of a solid zygopterid stem similar to that of the osmundaceous plant 

 Thamnopteris — i.e. with two types of wood and no xylem-parenchyma — has been foretold 

 by Kidston and Gwynne-Vaughan. The stem of D. romeri exactly suits this prophecy, 

 and its discovery is a striking testimony to the high probability of the truth of a theory 

 which they have recently elaborated. We would naturally expect to find the most 

 primitive types among the oldest rocks, and the stem of D. romeri also bears out this 

 expectation. The plant has been recorded in Devonian and Calciferous Sandstone 

 (Culm) rocks, all of which contain a more ancient type of flora than that of the Permo- 

 Carboniferous series. All the other zygopterid stems belong to higher horizons and 

 are less primitive. Among the Zygopteridese, the stems of Ankyroptems corrugata 

 (Williamson) and A. Brongniarti (Renault) show the strongest superficial resemblance 

 to D. romeri, but their tracheides are scalariformly thickened, and both possess a paren- 



