October 13, 1911] 



SCIENCE 



473 



to indicate its more general retention in 

 this group of plants, and does not agree 

 readily with the theory that these uni- 

 sexual orders are highly specialized plants, 

 with much-reduced flowers. The posses- 

 sion of a multicellular archesporium is, 

 however, not the only primitive character 

 exhibited by some of the unisexual orders 

 of the ArchichlamydeEB. Miss Kershaw' 

 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 en- 

 tirely free from the nucellus, and is pro- 

 vided with well-developed vascular bun- 

 dles, in both of which features it resembles 

 very closely the paleozoic seed Trigono- 

 carpus. The same features were shown, 

 moreover, by Dr. Benson'* and Miss Wels- 

 ford to occur in the ovules of Juglans 

 regia and in a few allied genera, such as 

 Morus and Vrtica. Also in a large num- 

 ber of Amentales with anatropous ovules 

 {Quercus, Corylus, Castanea, etc.). Miss 

 Kershaw has demonstrated the occurrence 

 of a well-developed integumentary vas- 

 cular supply. No doubt a further search 

 may reveal the occurrence of this feature 

 in some other dicotyledonous ovules, but in 

 the meantime it seems difficult to believe 

 that such a primitive vascular system, 

 which the Amentales share with the older 

 Gymnosperms, would have been retained 

 in the catkin-bearing group, if it had un- 

 dergone far-reaching floral differentiation, 

 while it had disappeared from the plants 

 which in other respects remained primi- 

 tive. It would be still more difficult to 

 imagine that it had arisen in the Amentales 

 subsequently to their specialization. 



There are other structural characters 

 and general morphological considerations, 

 which I have not time to deal with, which 



= Annals of Botany, Vol. XXIII., 1909. 



'Hid. 



underlie the belief in the primitiveness of 

 the Amentales and some allied cohorts, and 

 I trust they will be set forth in detail by a 

 better systematist than I can claim to be. 

 My object in bringing the matter forward 

 at all is to point out some of the difficulties 

 which prevent me from accepting a mono- 

 phyletic origin of the Dicotyledons through 

 the Ranalian plexus. 



One of these difficulties lies in the rela- 

 tionship of the Gnetales to the Dicotyle- 

 dons. Arber and Parkin have recently 

 made the attempt to gain a clearer insight 

 into the affinities of this somewhat puzzling 

 group by applying to it the "strobilus 

 theory' of Angiospermous descent." The 

 peculiar structure of the flowers of Wel- 

 ivitschia lends itself particularly well to a 

 comparison with those of Cycadoidea, and 

 a good case can no doubt be made out for a 

 hemiangiospermous ancestry of this mem- 

 ber of the Gnetacete, and by reduction the 

 other members, in many respects simpler, 

 might be derived from a similar ancestor, 

 though probably, as far as Ephedra and 

 Gnetum are concerned, an equally good, if 

 not better, comparison might be made with 

 Cordaites. But even supposing we admit 

 the possibility of a derivation of the Gne- 

 tales from an amphisporangiate Pterido- 

 sperm, I think the Amentales merit quite 

 as much as the Gnetales to be considered as 

 having taken their origin separately from 

 the Hemiangiospermse, and not from the 

 Eanalian plexus. I find this view has been 

 put forward also by Lignier* in his attempt 

 to reconstruct the phylogenetic history of 

 the Angiosperms, and I feel strongly that 



'Arber, E. A. N., and Parkin, J., "Studies on 

 the Evohition of Angiosperms, " " The Eelation- 

 shlp of the Angiosperms to the Gnetales," Annals 

 of Botany, Vol. XXII., 1908. 



' Lignier, 0., ' ' Essai sur 1 'Evolution morpholo- 

 gique du Regne vegetal," Bull, de la Soc. LinnS- 

 enne de Normandie, 6 s6r., 3 vol., 1909, reimprimfi 

 fevrier, 1911. 



