148 SEVENTEENTH REPORT. 



structure (PI. XIV, figs. 15-16). This group is, in its turn, un- 

 doubtedly descended from some of the Marattia-like Eusporangiate 

 ferns. We must assume that in the ancestral Pteridosj)ermeae some 

 such a tendency to strobilus formation arose as is shown among 

 modern ferns such as the Ostrich fern {Matteiiccia struthiopteris) 

 wliere the fertile leaves (sporophvlls) arise in a cluster together,' or 

 as in Osmnnda oimuimomea. Indeed, these first steps towards 

 strobily are found in many, if not most, ferns, for from a bud arise 

 usually sterile leaves first and subsequently fertile ones. In the 

 C.vcads we see in the genus Cycas the last steps in this same 

 tendency. Here the microsporophylls are already in a definite 

 strobilus but this is still rather lax as to the megasporophylls. 



"With the adoption of heterospory among the ancestral Euspor- 

 angiatae and subsequent acquirement of the seed habit both micro- 

 sporophylls and megasporophylls began to be modified. In the line 

 of descent that led to the Cycadineae the modification was about 

 equal in each and by some chance the strobili that developed were 

 unisexual. In another slightly different direction of strobilar de- 

 velopment a bisexual strobilus arose, but in this group the mega- 

 sporophylls were far more reduced than the microsporo])hylls and 

 we have arising the Bennettitineae. Probably from some point early 

 in this development line that ultimately produced the Bennetti- 

 tineae there started a third line of modification, like the latter in 

 producing a bisexual strobilus and in possessing a "perianth" but 

 differing in that the megasporophylls were not so completely re- 

 duced. These latter were probably more or less pinnate, much like 

 those of Cycas. and not reduced to a stalk and terminal ovule as 

 in the Bennettitineae (PI. XIY, fig. 17). On the other hand it may 

 well be that the microsporophylls had nearly lost their pinnate 

 structure or that the ])innae instead of being long with two rows of 

 sporangia were reduced to very short pinnae Avith but two sporangia 

 each. If the pinnae now became crowded by the shortening of the 

 microsporophylls and the microsporangia fused with one another 

 longitudinally, there would be produced a microsporophyll with 

 four longitudinal rows of fused sporanji^ia or four pollen-sacs such 

 as the typical stamen ])Ossesses. The slightly ])innate megasporo- 

 ])hylls by infolding would produce the closed pistils (PI. XIV, fig. 

 18). At first these were probably open at the tip to permit acce.is 

 of the microspores to the megasporangia (ovules) but as the micro- 

 spores increased their ability to produce pollen tubes the closure 

 gi-adually became complete. The Avriter's-" investigations upon the 

 development of the pistils in the Ranunculaceae, Alismataceae and 



20. Bessey, Ernst A. The Comparative Morphology of the Pistils, etc. 1898. 



