6o4 MOSSES AND FERNS chap. 



extraordinarily like a developing ovule, and the closely invest- 

 ing indusium has all the appearance of an ovular integument. 

 The velum of Isoetes is possibly of the same nature. 



The development of heterospory in several unrelated groups 

 of Pteridophytes at once suggests the possibility of a multiple 

 origin for the Spermatophytes. The radical differences be- 

 tween Gymnosperms and Angiosperms, and the absence of any 

 truly intermediate forms, make it extremely probable that these 

 two great divisions have originated independently of one 

 another, probably from different stocks, and it is by no means 

 unlikely that the same may be said of the Cycads, Conifers, and 

 Gnetacese. 



The discovery of motile spermatozoids in Cycads and 

 Ginkgo (Ikeno (i, 2) ; Hirase (i) ; Webber (i)), and the re- 

 cent studies upon Palaeozoic seed-bearing plants all make it cer- 

 tain that the seed-habit has developed quite independently in 

 several widely separated groups. 



Except for their siphonogamic fertilisation, the Gymno- 

 sperms really are much nearer the Pteridophytes than they are 

 to the Angiosperms. As both the pollen tube and the seed- 

 formation are but further developments of heterospory, it is 

 quite conceivable that these might have arisen independently 

 more than once. The close resemblance between the Conifers 

 and the Lycopods, especially Selaginella, probably points to a 

 real relationship. The strobiloid arrangement of the sporo- 

 phylls, as well as the development of the prothallium and 

 embryo, are extraordinarily similar, and it is not unreasonable 

 to suppose that this is something more than accidental. The 

 strong resemblance between the method of the secondary thick- 

 ening of the stem in the arborescent fossil Lycopodinese, and 

 that of the Conifers, as well as the anatomy of the leaves sug- 

 gests a real affinity. It is known that some of these bore seeds, 

 which in structure and position may very well be compared to 

 those of typical Conifers. The prevailingly dichotomous 

 branching of Lcpidodcndron, however, is very different from 

 the type of branching in the typical Conifers. 



Recent studies on the Cycadofilices, and the discovery of 

 spermatozoids in the living Cycads, proves beyond a doubt the 

 origin of the latter from Fern-like ancestors. 



The most recent evidence seems to support the old view that 

 Isoctcs belongs in the series of the Lycopodinere ; nevertheless 



