292 DR. A. B. AVALKOM ON XC'MMCLOSPEIiMrM [vol. Ixxvii, 



There are indications, in some of the examples, of the nature 

 of the vascular system. In one or two specimens there is a distinct 

 ridge running from the base of the seed to the base of the nucellus 

 (PL XXI. fig. 2). This ridge gives of£ a pair of ridges, one on 

 each side, into the sarcotesta, and these two ridges continue 

 for some distance parallel to the junction between nucellus and 

 integument. This seems to be a plain indication of a single 

 vascular bundle entering the base of the seed, giving off a pair of 

 bundles running through the sarcotesta in the princijml plane of 

 the seed. It cannot at present be determined how far this pair 

 of bundles extends towards the apex of the seed ; but it is not 

 improbable that they will be found to extend almost, if not quite, 

 to the micropyle. Further, in some specimens, the body of the 

 nucellus shows a number of distinct vertical ridges ; it is not 

 unlikel}' that these represent a series of distinct strands arising 

 from an expansion of the main bundle where it reaches the base 

 of the nucellus, as is the case in Cardiocaiyus (see Oliver, Biblio- 

 graphy, No. 5, pi. xxiv, fig. 1), and in some Radiosperms. These 

 ridges extend almost to the upper margin of the circular portion 

 of the nucellus (that is, excluding the nucellar beak). 



The specimens exhibit very little variation in size, and it may 

 reasonably be concluded that they represent the dimensions of the 

 mature seed ; their diameter is about 10 mm. in the principal plane 

 of the seed. None of the specimens gives an idea of the thickness 

 perpendicular to this plane. 



Species associated with Nummulospermum bowenense. 



A number of specimens have been collected from Three-Mile 

 Creek, Bowen, and, in addition to the seeds just described, the 

 only species observed from that locality are GJossoiJieris indica, 

 G. ampla, and G. Dunstani, sp. nov. On the actual slab in which 

 the seeds are preserved there are only examples of G. ampla. 

 Apart from actual connexion of the seeds w^ith Glossopteris, it 

 Avould hardly be possible to obtain stronger evidence of close rela- 

 tion between them than this. But this is not quite all : the 

 evidence is somewhat strengthened by the following occurrences 

 among the Queensland collections : 



(a) Near Pentland, a single seed {Samaropsis Etheridgei, sp. nov.) occurs 



on the same piece of rock as fronds of Glossopteris indica ; the only- 

 other species noted from the same locality is G. ampla. 



(b) From Minnie Creek, there is also a single seed, probably an example 



of Nummidospenmim hoicenense. It occiu's in association with 

 Glossopteris Brotvuiana, Sphenopteris lohifolia, and scale-fronds of 

 G. Broivniana. 



(c) From Barwon Park, there are three flattened seeds, also probably 



N. hoivenense, occurring on the same specimen as G. Broivniana and 

 a Sphenopteris. The only other species from the same locality are 

 G. indica and ? G. tortnosa. 



Add to these David White's description (Bibliography, No. 8, 



