THE PAST HISTORY OF MONOCOTYLEDONS 23 1 



is thought to have been homologous with that of Welwitschia. 

 Gnetuin and Welwitschia are derived from the Angiospermic stock, 

 but before the carpel became the [)ollen-receiver. 



In the Penaeaceai, a group of dicotyledons, Miss E. L. Stevens ^ 

 finds that there is no polarity in the early stages of development of 

 the endosperm, as in most Angiosperms. The embryo-sac contains 

 not 8 but 16 nuclei which form four egg-apparatus. By the 

 fusion of 4 nuclei a definitive nucleus results. By this reductional 

 process the same features met with in Welwitschia are recalled. A 

 very similar process is recognised in Gnetum by Lotsy in the germina- 

 tion of the megaspore or embryo-sac, and the development of the 

 endosperm. 



5. Conclusions. 



Summing up the evidence afforded by the different views held 

 by the writers cited, and by personal observation and experience, 

 it seems to us that systematic researches have failed to appreciate 

 the results of palaeontology and of morphology, and in considering 

 Panda/ms, etc., as the oldest group sufficient attention has not been 

 given to the comparative results of floral structure or of morphology. 



As to the theories based upon either embryology or morphology. 

 Dr. Worsdell's view seems to us to be based upon the most stable 

 evidence. 



The view that there is only one cotyledon in both groups, indeed, 

 is perfectly reasonable ; and if the cotyledon and the seta of Bryo- 

 phytes are conceded to be homologous, then their respective order 

 in time is also practically settled, for the single cell in the latter 

 gives rise to two, and so on until, just as in Capsella, PL II. fig. 3 

 (part i.), p. 180, we arrive, in due course, at the bifid cotyledon. 

 The fact of the subdivision from an original single cell presupposes 

 the primitive character of the latter, and the derived nature of the 

 double row of cells. 



As to the position of the cotyledon, if the above view is accepted 

 the originally terminal character of the cotyledon also follows. The 

 lateral position subsequently, if we adopt the sympodial arrangement, 

 does not affect the question. The lateral appendage view does more 

 harm, it seems to us, in obscuring the sequence of the ontogeny 

 (and so the phylogeny), and the morphological significance of organs 

 than any other theory promulgated. 



The reversionary nature of the monocotyledonous type in Ranales 

 is in itself more or less favourable to the earlier age of monocoty- 

 ledons and the derivation of dicotyledons from them. 



Miss Sargant's views are in direct contrast to those of Dr. Worsdell's. 

 This is readily understood when it is seen that she has based her 

 evidence mainly upon the earlier embryological features of the 

 species studied, too early in the ontogeny for a proper recapitulation 

 to be observed. Dr. Worsdell, however, takes the most mature 



^ " Proc. Roy. Soc. S. Africa," 1908. 



