650 GEOLOGY 



are its connections with the true ferns. Initial forms have been 

 identified from the Devonian and Lower Carboniferous, and a 

 few are found in the Permian; but the group is essentially 

 Pennsylvanian. 



The Cordaitales. This was a remarkable, family (now extinct) 

 of gymnosperms (p. 944), having alliances with the " seed- ferns,'' 

 conifers, and ginkgos, and yet many distinctive features of its own. 

 The Cordaites were tall, rather slender trees, reaching two feet or 

 more in diameter, and 90 feet or more in height. The wood was 

 of the coniferous type, coyered, as in so many other plants of the 

 period, by a thick bark. The trunks had a large pith. The leaves 

 were parallel veined, suggestive of monocotyls of the yucca type, 

 and sometimes attained a length of six feet and a width of six 

 inches. They were preserved in great abundance, and make up a 

 large part of some beds of coal. The leaf-structure combined char- 

 acters now possessed by certain conifers, with others possessed by 

 certain cycads. In one form, the leaf had a distinctly fleshy char- 

 acter, as if adapted to xerophytic life. 1 The floral organs were 

 peculiar to the family, and have been worked out with marvelous 

 success, even the structure of the pollen having been determined. 

 The inflorescence took the form of separate male and female catkins, 

 arranged on slender stalks attached to the stem between the leaves. 

 The seeds (Cardiocarpus) were of the cycadian rather than of the 

 coniferous type, and were very abundant and sometimes winged, 

 as if for wind transportation. 



It is doubtful whether conifers existed in the Pennsylvania^ 

 period, though they were probably represented in the Permian. 

 The upland vegetation is not known, and it is not impossible that 

 conifers, a type especially suited to an upland habitat, prevailed 

 there. 



Cycadales have been commonly reported from the Carboniferous, 

 but the evidence remains inconclusive, and the fossils concerned 

 are probably Bennettitales, rather than true cycads. 



The Coal flora of North America and that of Kurope were strik- 

 ingly similar, implying close geographic relations and like condi- 



1 Scott, op. cit., p. 425. 



