556 TRANSACTIONS OF SECTION K. 



does not seem to me quite justified. Amphisporangiate forms may have been 

 preserved, or may have arisen anew in various groups of Pteridosperms or in 

 their descendants. Heterospory, we know, originated independently in at 

 least three of the great phyla of vascular Cryptogams, and originally, no doubt, 

 the same strobilus contained both macro- and micro-sporangia, as was the case in 

 Calamostachys Casheana, in the strobili of most Lepidodendraceae, and as is 

 still the case in the strobili of Selaginella and in Isoetes. Even in the existing 

 heterosporous Filicineae, micro- and macro-spores are found on the same leaf, 

 and on the same sorus; and though in the higher Cryptogamia and the lower 

 Phanerogamia there may have been a tendency to an iso-sporangiate condition, 

 yet, as the two kinds of spores are obviously homologous in origin, nothing is 

 more natural than an occasional reversion to a heterosporangiate fructification. 

 Thus in the group of Gymnosperms we have many instances of the occurrence 

 of so-called androgynous cones. In 1891, at the meeting of the British Asso- 

 ciation in Leeds, I described such amphisporangiate cones which occurred regu- 

 larly on a Pinus Thunbergii in the Royal Gardens of Kew, and only this spring 

 I was able to gather several hermaphrodite cones of Larix Europea. They 

 have, of course, been observed and described by many authors for a variety of 

 Gymnosperms. What more likely than that many extinct Gymnosperms may 

 have developed heterosporangiate fructifications? It is not necessary, there- 

 fore, to fix on one group of ancestors for the origin of all existing Angio- 

 sperms. Indeed, the great variety of forms, both of vegetative and reproductive 

 organs, which we meet with in the Angiosperms, not only to-day but even in 

 the Cretaceous period, in which they first made their appearance, warrants, 

 I think, the belief in a polyphyletic origin of this highest order of plants. 

 It is no doubt true, as Wieland points out, ' that the plexus to which Cycadoidea 

 belonged, as is the case in every highly organised plant type, presented mem- 

 bers of infinite variety,' and, indeed, so far as the vegetative organisation goes, 

 we know already, through the labours of Nathorst, of such a remarkable form 

 as Wielandiella angustifolia, while Wieland has shown us a further type in his 

 Mexican Williamsonia. Nevertheless, these diverse forms all agree in the struc- 

 ture of their gynaecium, the particular organ which is not so easy to Bring into 

 line with that of the Angiosperms. 



I am quite alive to, though somewhat sceptical of, the possibility of a direct 

 descent of the Ranales from the Cycadoideae, but my hesitation in accepting 

 Arber and Parkin's view of the ancestry of the Angiosperms is enhanced by 

 the consideration that it seems almost more difficult to derive some of the 

 apparently primitive Angiosperms from the Ranales, than the latter from Cyca- 

 doidea. Indeed, this common origin of Angiosperms from the Ranalian 

 plexus will, I feel sure, prove the stumbling-block to any general acceptance of 

 the Arber-Parkin theory. It is easy enough to assume that all Angiosperms 

 with the unisexual flowers have been derived by degeneration or specialisation, 

 from forms with hermaphrodite flowers of the primitive Ranalian type, but un- 

 fortunately some of these degenerate forms possess certain characters which 

 appear to me to be undoubtedly primitive. 



It is difficult for those who accept Bower's view of the gradual sterilisation 

 of sporogenous tissue not to regard the many-celled archesporium in the ovules 

 of Casuarina and of the Amentales as a primitive character, and though, as 

 Coulter and Chamberlain point out, this feature is manifested by several mem- 

 bers of the Ranunculaceas and Rosaceae, as well as by a few isolated Gamopetalae, 

 its very widespread occurrence in the Amentales seems to indicate its more 

 general retention in this group of plants, and does not agree readily with 

 the theory that these unisexual Orders are highly specialised plants, with much 

 reduced flowers. The possession of a multicellular archesporium is, however, not 

 the only primitive character exhibited by some of the unisexual orders of the 

 Archichlamydeae. Miss Kershaw 4 has shown, in her investigation of the structure 

 and development of the ovule of Myrica, that in this genus, which possesses a 

 single erect ovule, the integument is entirely free from the nucellus, and is 

 provided with well-developed vascular bundles, in both of which features it 

 resembles very closely the palaeozoic seed Trigonocarpus. The same features 

 were shown, moreover, by Dr. Benson 5 and Miss Welsf ord to occur in the 



* Annals of Botany, vol. xxiii., 1909. 5 Ibid. 



