344 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1918. 



the interseminal scales, between which the micropylar tubes of the 

 seeds protrude. The microsporophylls are only preserved in im- 

 mature strobili. They have disappeared at a later period, but their 

 former presence can usually be inferred from the basal shoulder 

 that marks their insertion on the receptacle. The probabilities are 

 that all the cycadeoids were bisporangiate, although this is disputed 

 in the case of some species. The seeds are very small as compared 

 with those of living cycads and have a two or three layered testa, 

 prolonged upward into a micropylar tube, the outer palisade layer 

 of which at the shoulder where it joins the seed being expanded by 

 radial extensions, forming five or six radial wings. The cavity of 

 the seed is completely filled by a large dicotyledonous embryo which 

 in its development, unlike any living gymnosperms, destroys all of 

 the endosperm. 



A variety of cycacl fronds are found in the deposits which have 

 yielded the petrified trunks, but our information regarding their 

 actual foliage is furnished by small unexpanded petrified material 

 which shows pinnate fronds -of the Zamia type, circinate in verna- 

 tion, with erect pinnules. The number was small in Cycadella. In 

 Cycadeoidea ingens there were 60 to 100 pairs of linear or slightly 

 spatulate pinnules, and the truncate pinnate fronds are estimated as 

 having been several feet in length at maturity. It seems probable 

 from the large number of cones in the same stage of development in 

 some trunks and their complete absence in other large trunks that 

 seed formation was the culminating event of many seasons of vege- 

 tational activity, possibly occurring but once at the maturity of the 

 plant, but more probably repeated at increasing intervals through- 

 out life. In the historic Johns Hopkins Cycad shown in the accom- 

 panying figure 17 there are 58 strobili on the face figured. 



Fronds indistinguishable from those of cycadeoids are cosmopoli- 

 tan from the Triassic to the Upper Cretaceous, but there is no means of 

 determining to what an extent they belong to the cycadeoids, and the 

 probabilities are strongly in favor of considering most of them as the 

 fronds of the order that I have termed the Williamsoniales. 



The Williamsoniales comprise a much more protean and long-lived 

 group of forms, unfortunately known almost entirely from impres- 

 sions. They appear in the later Paleozoic in the genera Plagiozamites, 

 Pterophyllum, and Sphenozamites ; are very prominent in the Tri- 

 assic, Jurassic, and Lower Cretaceous; and are still present but in 

 reduced numbers in the Upper Cretaceous after the cycadeoids had 

 become extinct. The nature of their remains precludes a concise sum- 

 mary of their features. Both the very abundant foliar remains as 

 well as such fructifications as have been preserved indicate a much 

 wider variability than among the other two orders of cycadophytes. 

 The sole feature that at present can be considered as probably char- 



