114 THE EVOLUTION OF PLANTS 



each pollen-grain contained a number of cells, 

 with delicate walls separating them, which broke 

 down to let out the contents. There can scarcely 

 be a doubt that the bodies emitted from these 

 cells were active spermatozoids; in fact, from the 

 structure of some of these fossil pollen-grains, 

 Renault deduced the probable presence of sper- 

 matozoids long before they were discovered in 

 living Cycads or in Ginkgo. The many-celled 

 structure of the pollen-grain extends to various 

 other Palaeozoic families. There is no proof, in 

 any case, that a pollen-tube was formed, and it 

 is quite likely that this structure had not yet 

 been evolved, the spermatozoids relying more on 

 their own powers of movement in the Palaeozoic 

 plants than is the case in the later Cycadophyta. 



Besides the Lyginodendreae and the Neurop- 

 terideae there were a great number of other 

 Fern-like plants of Palaeozoic age which appear 

 to have been reproduced by seeds. In two genera 

 (Pecopteris and Aneimites) the seeds have been 

 found in connection with the fronds, and are re- 

 markable for having a flattened, instead of a 

 rounded form; the importance of this point is 

 that it shows a very near approach to the seeds 

 of the Cordaiteae, a Palaeozoic family which had 

 much in common with the Coniferse. 



Our knowledge of the Pteridosperms is still 

 very imperfect, but nevertheless the discovery of 



