Sept. i9io.] FUJIL—SOME REMARKS ON THE ('RET. FLORA ETC. 207 



bus and on discussing its affinities (Stores and Fujir, 1910 p. 

 39.) a great stress has been laid on the structure of the seed 

 which seemed to resemble in some points the paleozoic seeds of 

 Trlgonocarpus-type ; and the free nucellus and the integument 

 which consisted of the fleshy outer and inner, and the stony 

 middle layers, were considered as the most important features. 

 Unfortunately the classical works of L. C. Richard (Commen- 

 tatio botanica de Conifereis et Cycadeis 1826) and Eichler 

 (Flora Brasiliensis, Conif.) have not been referred in that 

 paper. Recently, a number of seeds or ovules with the free 

 nucellus become known. Robertson (1906) announced such 

 structure of seeds in Phyllocladus alplnus, Kershaw (1909) 

 discovered such seeds in Myrica, Juglans, and a number of 

 other Monochlamydeous plants, and certainly many more ex- 

 amples are to be added to the list. Richard's above mention- 

 ed work contains a large number of figures of ovules or seeds 

 in longitudinal sections from various groups of Gymnosperms. 

 They represent many gradations in the freedom of nucellus with 

 regard to the integument. Even among one and the same genus 

 they are much variable. In Ephedra americana the nucellus is 

 free from the integument only as far as one fifth of its length, 

 in B. vulgaris it is free till about midway towards the base, 

 while in B. altissima it is free as much as nearly % of its 

 whole length. The papers of Land (1907) and of Berridge 

 and Sanday (1907) give further gradations to this series. 



Although I can not enter into details in the present 

 paper, there is among Conlferae themselves a complete series 

 of gradations of the freedom of nucellus. Taxodium dlsticum 

 is at one extremity of the series, where very little of the 

 nucellar apex is free ; this is to be confirmed from Coker's 

 figure (Coker, 1903) ; species of Plnus come next to it, then 

 several plants of various families and genera among Conifers are 

 to follow ; towards another extremity are to be ranked Tor- 

 reya nuclfera [?], Phyllocladus rhomboldalls (where the nucellus 

 is nearly free, but not right down to the base as in Ph. alplnus 

 studied by Robertson), some species of Thuya, Callitrls quadrl- 

 valvls, Araucarla Dombeyi, etc., where the nucellus is free to a 



