EVOLUTION OF SEED-PLANTS 113 



bination of this character with indefinite growth 

 in the thickness of each cylinder is peculiar to 

 the fossil family and seems to us a cumbrous 

 arrangement. The great leaf-stalks, with their 

 numerous vascular strands and strong mechan- 

 ical construction, resembled those of a recent 

 Cycad. The plants were thus of great complexity 

 in then- vegetative structure. 



As regards the reproductive arrangements, a 

 number of seeds, which there is good reason for 

 referring to this family, are well known. Gener- 

 ally speaking, we may say that the seeds of Neu- 

 ropterideae were very much like those of modern 

 Cycads in dimensions and structure. They had 

 a thick, fleshy outer coat, and a "stone" inside 

 it, within which was the central body or nucellus, 

 containing the embryo-sac. The nucellus had a 

 well-marked pollen-chamber at the apex, much 

 like that of a Cycad. The seed was well supplied 

 with vascular strands, which formed a double 

 system, one set running through the outer flesh, 

 while another set traversed the nucellus. There 

 are some differences of detail, but on the whole 

 these seeds are the nearest among fossils to those 

 of the recent Cycadacese. 



A very interesting point in connection with 

 seeds of this group is the wonderfully perfect 

 preservation of the pollen-grains found in the 

 pollen-chamber. It can be seen quite clearly that 



