part 4] BOWENENSE, GE]S^. ET SP. NOT. 293 



p. 559) of the association of seeds of somewhat similar type with 

 Gangamopteris fronds in Brazil, and Prof. Seward's record of the 

 association of Samaroj^sis Milleri with scale-leaves of the Glosso- 

 pteris type; wherefore the evidence of association seems sufficiently 

 strong to justify the statement that the seeds under discussion are 

 those of a species of Glossopteris. 



They would represent the megasporangia, the microsporangia 

 having heen discovered and described by Arber some fifteen years 

 ago (Bibliograph}^ No. 1, p. 324; No. 2, p. 41) on specimens 

 from Port Stephens, New South Wales. 



Of the Queensland localities mentioned above, Pentland is 148 

 miles west of Townsville on the railway-line ; Three-Mile Creek is 

 approximately 150 miles east of Pentland, and its position is 

 roughly 147°*^ 48' long. E., 20° 35' lat. S. ; Barwon Park is on the 

 Mackenzie Piver, and is about 180 miles south 25° east of Three- 

 Mile Creek; Minnie Creek is about 30 miles south of Barw^on 

 Park, and 10 or 12 miles from Blackwater Railway-station, which 

 is 125 miles from Pockhampton on the Central Railway. 



Prom this it is seen that the localities at which these seeds are 

 associated with Glossopteris are widely scattered, a fact tending 

 rather to strengthen the evidence. 



Scale-Fronds of Glossopteris. 



It may be noted here that the scale-fronds occui- in association 

 with Glossop)teris in Queensland. They cannot be distinguished 

 from specimens figured by Zeiller from India (Bibliography, No. 9, 

 pi. iii), and for comparison two of them are figured (PL XXI, 

 figs. 5 & 6). There is no doubt that those resembling PL XXI, 

 fig. 6, are in Queensland associated with Glossopteris Browniana ; 

 Aviiereas in India both the types here figured are referred by 

 Zeiller to G. inclica. This goes to indicate that probably these 

 two so-called ' species ' really represent portions of one very variable 

 species, Brongniart's original G. Broicniana. Further details of 

 these scale-fronds will be published with the full account of the 

 Queensland Permo-Carboniferous flora mentioned above (p. 289), 



Relationships of Glossopteris. 



Although we have advanced a stage in our knowledge of Glosso- 

 pteris, there is still much to be learned. The megasporangia 

 show that the genns belongs either to the Cycadofilicales or to the 

 Corclaitales, which two groups cannot be separated on the evidence 

 of their seeds alone. If we take into account other features in 

 connexion with Glossopteris, the balance of evidence appears to be 

 in favour of its reference to the Cj^cadofilicales. The association 

 of scale-fronds and perhaps the venation of the leaves support 

 this ; further, if we accept Vertehraria as being the underground 

 rhizome of Glossopteris, the indication is that it belongs to the 

 Cycadofilicales rather than to the Cordaitales. Parenthetically, it 

 may be remarked that our knowledge of the fossil Vertehraria is 



