﻿Yol. 64.] 



FOSSIL PLANTS FEOM SOUTH AFRICA. 



93 



consider the available data sufficient to enable us to draw a dis- 

 tinction between variations within one species and differences 

 worthy of specific rank. 



The piece of frond represented in PI. IV, fig. 1 (from the Molteno 

 Beds of Konings Kroon) has leaflets identical with those shown in 

 fig. 8, pi. vii, of my former paper ' ; but it exhibits an additional 



Fig. 4. — Thinnfeldia odontopteroides 

 {four-fifths of the natural size). 



feature in the dicho- 

 tomy of the frond- 

 axis, a character which 

 Potonie and other 

 authors are no doubt 

 correct in considering 

 as primitive. PI. Y, 

 fig. 1 illustrates an 

 example of a leaf from 

 Yaalbank, of which 

 the rachis bears pinnae 

 with pinnules of the 

 larger type charac- 

 terized by numerous 

 forked veins. Pin- 

 nules are borne also 

 on the rachis. In 

 another specimen from 

 the same locality the 

 segments are still 

 larger, one of them 

 (fig. 3 A, p. 91) having 

 a length of 1-8 centi- 

 metres and a breadth 

 of 1-5 cm.; this and 

 other fragments bear 

 a striking resemblance 

 to Feistmantel's larger 

 fronds from New 

 South Wales. Several 

 examples, such as that 

 shown in text-figs. 3 B 

 & 4, demonstrate the 

 passage, from linear 

 segments with a mid- 

 rib giving off" clusters 

 of forked veins and with a lamina showing difi'erent degrees of 

 lobing, to pinnae with short and broad ultimate segments agreeing 

 in all respects with those represented in PI. Y, fig. 1. The pinnule 

 shown in fig. 3 C (p. 91) occurs on an axis which bears also segments 

 identical with those represented in fig. 3B, and thus forms a 

 connecting-link between the types of frond illustrated in PI. lY, 



^ Seward (03) pi 55. 



