i6o C;VMNOSPERMS 



vegetative leaves is a crown of much reduced leaves, the male sporo- 

 phylls, bearing sporangia, much as they are borne in a fern. Inside 

 the crown of male sporophylls is a crown of still farther reduced 

 leaves, the female sporophylls, bearing the ovules. The change from 

 the bisporangiate condition to the dioecious is one which can be seen 

 in all stages of transition in living angiosperms; and the evolution of 

 the compact cone from a loose crown of sporophylls is seen in the 

 living cycads. 



From the condition shown in fig. 173, the Bennettitales may have 

 developed as in fig. 174; and the Cycadales as in fig. 175. 



From the Cycadofilicales to such a form as Cycadospadix hcnno- 

 quci, the transition is not hard to imagine; and Cycadospadix might 

 well have been included in the living genus, Cycas, for its megasporo- 

 phyll differs less from that of Cycas revoluta than the sporophyll of 

 C. revoluta differs from those of some of the other species of the genus 

 (fig. 176). Cycadospadix milleriana,'^^^ with sporophylls loosely ar- 

 ranged, but stopping the growth of the axis, shows a condition in- 

 termediate between the loose crown of sporophylls of Cycas revoluta, 

 and the compact cone (fig. 177). 



The theory that the Cycadales may have come from the Bennet- 

 titales has already been referred to. The trunk and leaves are similar 

 in the two lines and some of the Cycadales, like the Bennettitales, 

 have axillary strobiU. But the ovules, even in the earUest known 

 Bennettitales, have followed a very different Hne of evolution. They 

 have, in all cases, retained the terminal ovule and lost the lateral 

 ones; while all of the Cycadales, even those known only as fossils, 

 have lost the terminal ovule and retained the lateral. The Bennet- 

 titales could not have transmitted what the line had already lost. 

 The multilocular microsporangium of the Bennettitales is so dif- 

 ferent from the unilocular sporangium of the Cycadales that they 

 could hardly be related. Lines characterized by unilocular and mul- 

 tilocular sporangia may have been as distinct in the Paleozoic as 

 they are in Angioptcris and Marattia today. Both types, in the Pa- 

 leozoic, may have developed heterospory ; and the hetcrospory of the 

 fern may have progressed into the seed habit of the Cycadofilicales, 

 so that one section of the Cycadofilicales may have given rise to the 

 Bennettitales, and the other, to the Cycadales. 



