GYMNOSPERM^. 



117 



pollen-saos were, according to M. Renault, disposed radially in double ranks upon, 

 and partly embedded in, the surface of a not greatly modified peltate leaf. The 

 pollen grains were very large, formed of many nearly equal cells, and doubly dehiscent 

 longitudinally. The female organ, discovered by M. Grand'Eury, is an orbicular, 

 scarcely metamorphosed leaf, hollowed in the centre to receive a single oval seed. 

 In this primitive type the pollen-sacs are evidently honiologues of the microsporangia of 

 vascular Cryptogams ; the pollen grains are the microspores, while the seed corresponds 

 to a macrospore, germinating in position, and supported on a carpellary leaf. 



Next to Bolerophyllum must be ranged the still more imperfectly known Canno- 

 phr/llites of Brongniart, a plant with large leaves cut up into numerous segments, with 

 median veins and numerous oblique veinlets, recalling leaves of Scitaminea. In C. 

 Virleti, Brongt., the leaves must have measured several feet in length and the stems have 

 been also of great strength. Their nearest affinities are believed to be with Fsygmo- 

 pjiyllum, a prototype of Ginkgo. 



Still more developed, and far more perfectly studied, are the Cordaites. This impor- 

 tant group first appears in the Carboniferous and did not outlast the Permian, its extinction 

 seeming at once to make way for a great development of the true Gymnosperms. They 

 were large trees, varied, distributed into many genera, and possessing characters common 

 to the Cycadacese, Taxeas, and Gnetaceaj, yet being as a whole inferior to true Gymno- 

 sperms ; a group developed rapidly, and which has left no direct descendants. The stem 

 was large and repeatedly branched, and in some cases bore leaves several feet in length. 

 The leaves were coriaceous and possessed a multitude of equal subparallel veins. Three 

 different types are known and are placed in separate genera, the commonest form being 

 relatively broad and blunt at the end. The leaf-scars are transverse and discoidal, or 

 bent and elliptic, and disposed in quincuncial order on the stem. The simple, or rarely 

 composite fruit-spikes, furnished with spikelets in two rows, occur among the leaves 

 without definite order. The stem was composed of a central pith ; a woody region 

 possibly identical in some cases with plants described under the name Badoxijlon ; 

 and a thick cortical region, divided into an inner parenchymatose, and an outer 

 denser zone, traversed by fibrous cells and resinous canals. The reproductive zone was 

 at the inner periphery of the cortical region or bark, and corresponded with Cycadean 

 and certain monocotyledonous stems, rather than to those of acicular-leaved Coniferge. 

 The leaves had the progymnospermous structure which is still retained in the Cyca- 

 dacese. The male and female organs were inserted in the axils of bracts on a floral 

 spike. In the male the axis is short and furnished with bracteoles, and carries a 

 number of staminal leaves either in a terminal cluster, or else spirally disposed. The 

 staminal leaves bear three or four erect and terminal pollen-sacs, corresponding morpho- 

 logically to microsporangia. The pollen grains contained in them are large and elliptic, 

 composed of a finely reticulated integument that splits longitudinally, and an endospore 

 composed of as many as ten cells. In the female, carpellary leaves corresponding to 



