Further Notes on Seed Structures. 139 



vation will fully appear, if these two explanatory paragraphs 

 be borne in miud : 



Firstly .—The comparative study of all the known Cycad- 

 eoidea seeds and a fairly extensive list of existing gymnosperm 

 seeds, shows that the middle layer of the seed wall is essentially 

 a stony, rather than a merely indurated or suberous layer, and 

 much thicker in Bennettiles Morierei than in any American 

 species. Professor Lignier's description of this layer as fleshy 

 was given with some reservation, and it is merely necessary to 

 note that the stony layer shows varying degrees of induration 

 in allied species, throughout the gymnospermous series. 



Secondly. — In referring to the writer's own descriptions of 

 testal structure, as given in his American Fossil Cycads with 

 too few figures, all stress must be laid upon the fact that the 

 seed is there spoken of as it probably appeared on splitting 

 loose — as a shed seed — and that the form primarily dealt with 

 is Cycadeoidea Wielandi, which has only traces of ribbing and 

 now proves to have the most reduced testal structure known, 

 being almost, or wholly, without an outer fleshy, or " blow off 

 layer." Better sections from cones of other species show well- 

 marked ribbing and a heavy " blow off" to be commonly pres- 

 ent in the Black Hills Cj'cads. But so far there is no instance 

 in which this structure shows the fine detail in any silificied 

 specimens that it always exhibits in the ferrized B. Morierei. 

 In fact, the middle or indurated zone appears to have been far 

 more susceptible to preservation by silicification, whereas the 

 outer or " blow off" Jay er of the seed is infinitely better stained 

 and conserved by iron carbonate. "We think no one who sees 

 the original preparations can fail to be struck by this comple- 

 mentary reaction of the outer flesh and middle stone to these 

 two types of mineralization, as seen in the Calvados strobilus 

 and the Black Hills fruits. 



This explanation, then, entirely clears the way to a complete 

 understanding of the similarity of testal development in the 

 American and European fossil cycads. The seeds of the 

 Cycadeoidese containing embryos vary from one centimeter in 

 length in Amphibennettites Renaulti of Fliche (4), to three or 

 four millimeters in length in the smallest American forms, and 

 there is a remarkable homogeneity in the testal features of the 

 entire group. In fact, short of close study, the only differences 

 discernible are in the ribbing and size of the seeds with some 

 variation in the thickness of the tissue zones of the seed wall, 

 which is normally three layered with a distinct middle stone. 

 The outer flesh alone may, however, exhibit, aside from pecul- 

 iarities due to the unusual manner in which the seed is com- 

 pletely encased in the solid husk of interseminal scales, marked 

 reduction or even elimination. 



