464 STUDIES IN FOSSIL BOTANY 



Sutcliffia, however, the anatomy of the stem, peculiar as 

 it is, can only be compared with that of a protostelic 

 Fern, while the concentric foliar bundles accentuate the 

 relationship. There is nothing as yet to indicate from 

 what group of primitive Ferns the Medulloseae may 

 have sprung, and indeed the great complexity of the 

 seed shows clearly enough how remote the Fern- 

 connection must have been. On anatomical grounds it 

 seems not unlikely that this family may have had a 

 common origin with the Lyginodendreae from some 

 unknown protostelic type. 



On the whole, the Medulloseae or Neuropterideae 

 strike one as a more advanced group than the Lygino- 

 dendreae. Their seeds, if we may take Trigonocarpus 

 as a type, perhaps approach nearer than, any other 

 Palaeozoic seeds to those of recent Cycadaceae, as shown 

 by the differentiation of the integument into a sarcotesta 

 and a sclerotesta, the double vascular system and the 

 form of the pollen-chamber ; the chief difference lies in 

 1 the apparently free nucellus of the fossil seeds, a feature 

 with which . other distinctions may be correlated. As 

 regards anatomical characters, the petiole and rachis of 

 Medullosa very closely resemble those of a Cycad in 

 structure ; the same may be said of the root. The stem- 

 structure, however, throughout the Medulloseae is 

 essentially different, for neither in the almost protostelic 

 Sutcliffia, nor in the more complex polystelic stems 

 of Medullosa, do we find, as it seems to me, any 

 fundamental agreement with the stem either of recent 

 Cycadaceae or of Mesozoic Bennettiteae, though certain 

 analogies can be traced. On the whole of the evidence 

 an affinity between the Medulloseae and the Cycado- 



