244 HEREDITY AND EVOLUTION IN PLANTS 



dicotyledons, as being derived from the latter by a process 

 of simplification (C/. p. 223). The structural and ana- 

 tomical evidence that eusporangiate ferns are more 

 ancient than leptosporangiate ferns is rendered more cer- 

 tain by the fact that the earliest fern fossils (in Paleozoic 

 rocks) are eusporangiate; the leptosporangiate forms do 

 not appear until later, and the fossils belong to families 

 closely related to the more ancient eusporangiate group, 

 while the fossils of more recent rocks show closer affinities 

 with the modern living forms. (See, however, p. 30.) 



In any tabular arrangement including all the great 

 groups or phyla every group must, of course, come defi- 

 nitely after some one group and precede another. Thus, 

 mosses logically fall between the Thallophytes and the 

 fern allies ; but there is scarcely any evidence that they are 

 phylogenetically related to the groups that follow them 

 in the table. Strangely enough, there are few well- 

 authenticated fossil remains of mosses (and those not 

 below the Mesozoic), and it has even been seriously 

 suggested that they may have developed from more 

 complex groups by processes of reduction and simplifica- 

 tions; but there is little, if any, evidence to indicate from 

 what higher group they might have been thus derived, 

 and the positive, though meager, fossil evidence is suffi- 

 cient to render highly improbable, if not to nullify, the 

 suggestion of derivation by reduction. 



The old group "Pteridophytes," of the manuals, 

 including the true ferns and their "allies" (horsetails, 

 lycopods, and little club-mosses), served a useful purpose 

 before the recent researches in fossil botany; but the 

 results of those studies have made it impossible consist- 

 ently to maintain the group longer in its former con- 



