290 DR. A. 15. WALKOM OX XrMMrLOSPERMUM [vol. Ixxvii, 



These are, so far as I know, the most important contributions 

 hitherto made towards the elucidation of the real nature of 

 Glossojyferis and Gangamopteris. 



A detailed description is appended of the Queensland seeds, for 

 Avliich the new genus JsummnAos'j^ermv.m is proposed, and 

 • this is followed hv a discussion of the o^rounds on which their 

 connexion with GJossopteris is based, 



XuMMULOSPEE-MUM, gen. nov. 



This name is proposed for a number of seeds occurring in asso- 

 ciation with Glossopferis at Three-Mile Creek, Bo wen. Three- 

 Mile Creek is a small tributary of Pelican Creek, which is a branch 

 of the Bowen River. The question of their occurrence and asso- 

 ciation is discussed below. 



Seed platyspermic (?) : circular or oval, with broadly acute 

 apex. Integument consisting of wide outer sarcotesta and (?) nar- 

 row inner sclerotesta. Xucellus circular, with prominent beak 

 projecting into the narrow micro2:)Yle. Vascular system consist- 

 ing of a single strand entering the base, giving of£ a pair of 

 strands in the sarcotesta, then passing through the sclerotesta and 

 (?) dividing into a number of strands enveloping the megaspore, 

 and extending nearly to its apex. 



Genotype : Jsnmmvlospermv.m howenense, sp. nov. 



The foregoing description summarizes what appear to be the 

 essential characters of these seeds, which are discussed more fnlly 

 below under JS'. hoicenense. 



The characters of the seeds seem to be sufficiently well indi- 

 cated to warrant some more distinctive name than Samaropsis, 

 and at the same time the structure does not conform to that of 

 any of the described fossil genera known to m.e. 



NrMMULOSPEEMrM EowEyEXSE, sp. nov. (PI. XXI, tigs. 1-4. 

 & text-fig., p. 291.) 



Characters as described for the genus. 



Type : P 1900 in the collection of the Queensland Geological 

 Survey, from Three-Mile Creek, a tributary of Pelican Creek. 



On a large specimen (F 1900) fi'om Three-Mile Creek, Bowen 

 (PL XXI, fig. 1), there are remains of at least twenty-eight of 

 these seeds, and all of them present the same view of the seed. 

 Prom their general appearance, and the fact that all the seeds 

 known are preserved showing the same aspect, one might reason- 

 ably conclude that the seeds were platyspermic, and had split 

 along a princijDal plane. We must remember, however, that in 

 several of its characters the genus JS^ii.mnntlospermv.m shows close 

 affinities with members of the Trigonocarpales, in which the seeds 

 are radially s^'mmetrical (Badiosperms). 



The impi-essions as preserved appear to represent the features of 

 the seeds as seen on a principal plane, along which they apparently 



