Further Notes on Seed Structures. 



147 



extinct. And no reason is now seen to recede from this view. 

 For in reality Lagenostoma offers in the retention of reduced 

 inner flesh bundles, or an endovascular condition, an inter- 

 mediate stage between the large thick-walled amphivascular 

 seeds of existing cycads and the reduced or even truly mono- 

 vascular Cycadeoidea seeds ; while the archaic features in the 

 apical region of the latter must be of far more significance than 

 would be the retention of bundles, which were without a 

 shadow of doubt anciently present. And the closer we study 

 Cycadeoidea the clearer does it become that despite an enclosed 

 position and despite reduction the seed apex is after all of the 

 ancient multilobate type. Further, we think it can be accepted 

 that the stony palisaded layer of Lagenostoma loinaxi, though 

 following ripening and shedding of the seed the external one in 

 ali the lateral region of the seed, is none other than a middle 

 stone. The small tubercles of this layer (cf. fig. 4C) were 

 thought by Oliver and Scott (8) to have supported an outer 

 layer of thin-walled cells, while in the Cycadeoidea seeds one 

 may plainly see that the thin outer wall of the outer cells of 

 the stone laver has to do with its function as a bearing wall for 

 the "blow off" layer. In comparisons with Lagenostoma it 

 may be remembered, too, that various species are known, and 

 that in L. ovoides for instance, where both cell form and pre- 

 servation suggest decided resemblance to Cycadeoidea, the 

 indurated layer is several cells thick, so that this layer can vary 

 much in thickness and development in both these genera as 

 already known. To some the Lagenostoma cupule may seem 

 to prevent comparison, but it should be recalled that this 

 structure must have been under way of reduction, whereas the 

 tubular layer which occupies the same region in Cycadeoidea 

 and serves the same essential function as a basal cup or husk, 

 is evidently a last remnant of some ancestral leafy structure. 



So far as now recalled Polypterospermum of the French 

 Paleozoic is the only other ancient seed which has ever been 

 suggested as offering a distinct likeness to the Cycadeoidean 

 seed structure. Professor Lignier (3) in studying the four 

 (occasionally five) ribbed Bennettites Morierei, after pointing 

 out marked testal resemblance to Gnetopsis, attached too much 

 significance to apical appendages, reaching the conclusion that 

 the latter seed presented no true analogy. But he went on to 

 guardedly name an apparently related form Polypterospermum 

 as probably much nearer to Cycadeoidea than other ancient 

 types, a suggestion of relationship which seems to have been 

 little heeded. It is likely, however, - that restudy of the apical 

 region of the seeds of Bennettites Morierei in conjunction 

 with Polypterospermum will be well worth while, since this 

 region is not sufficiently known even in Bennettites Gib- 



