198 



D. H. Scott. 



same nature as the cupules of Lyginodendron oldhammm, but in Stur's 

 specimen the seeds had been shed — perhaps prematurelj'. 



Mr. Arber has recently described a seed, Logcnostoma Sinclairi, 

 Kids ton. evidently allied to the seed of Lyginodendron, and like 



Fig. 29. Diagrammatic transverse sections of the seed of Lyginodendron; their 

 position in indicated in Fig. 28. A (micropyle) shows pollen- chamber, canopy, and 

 free lobes of cupnle. B (body of seed) shows integument with bundles, and cupule 

 breaking up into lobes. C (chalaza) shows central supply-bundle, chalazal tissue, 

 base of integument and continuous, furrowed cupule. D, section of pedicel, uesembling 



the rachis. After Oliver. 



the latter, invested in a cupnle (see Figs. 31 & 32). Here again, 



the cupulate seeds are borne on a branched rachis (Arber, 1905). 



Lastly, M. Grand' Eury has observed six-lobed cupules, in some 



cases still contaiuing the seeds, situated at the extremity of long, 



