FOSSIL GYMNOSPERMS 



141 



are very suggestive of Taxus. The terminal ovule upon a short 

 bractlet-bearing axis seems to be common to both. The simi- 

 larity is increased by the fact that in most of the Cordaitales 

 the testa develops into a fleshy outer layer and a stony inner 

 one, which suggests the fleshy aril and stony seed coat of Taxus. 

 Abundant remains of seeds are found, which apparently belong 

 to the Cordaitales. In most cases the testa develops as indi- 

 cated above, and closely resembles the seeds of Cycads and 

 Ginkgo, as well as certain Conifers. In other cases, however, 

 the seeds ripen dry and bear wings, which are exceedingly vari- 

 able in form and position. Curiously enough, while seed sec- 

 tions have brought to light most of the structures referred to 

 above, no embryos have been found. Probably their for- 

 mation was much delayed, as in Ginkgo, Gnetum, and some 

 Cycads. 



It seems evident that we are dealing here with a group of 

 Oymnosperms which deserves to be set apart as of equal rank 



TTlt 



aT» 





P 



Fig. 9] .—Cycadinocarpus A'ug'ustodunensis, showing upper part of ovule in longitudinal 

 section ; int, integument ; mi, micropyle ; pc, pollen chamber containing pollen 

 grains ; nu, remains of nuoellus ; ar, archegonia ; pr, endosperm.— After Eeitatilt. 



with Cycadales, Ginkgoales, and Coniferales. It certainly 

 presents important characters which serve to define it as clearly 

 as any of the living groups, and its ancient character suggests 

 that it must have been very intimately associated with the begin- 

 nings of the Gymnosperm phylum. 



