9 



A. montanay Wils., but the flowers are rather smaller and the 

 leaves on the flowering branches are often three-lobed. The 

 fruit is somewhat turbinate and trigonous^ about 2-5 cm. long, 

 wider than long, flattened and. of ten depressed at the summit, 

 slightly tapering to the pedicel with three slight longitudinal 

 and several irregularly transverse ridges. The fleshy part of 

 the fruit is thin, soft, and fibrous, and encloses 3-5 seeds, which 

 are smooth, subglobose, and 1-5 cm. in diameter. The fruit is, 

 therefore^ much smaller and more fragile than those of the two 

 Chinese species. The foregoing fairly rej)roduces Wilson's 

 •enumeration and description of his differential characters. 

 Although not accepting Wilson's description of the fruit of the 

 Japanese tree, I agree that the available evidence favours his 

 view that the Japanese A. cordata is specifically distinct from his 

 new Chinese species A. inontana. In this connection I have 

 re-examined all the old material at Kew, and have also examined 

 the additions since 1906, with the result that I can discover no 

 essential correlative differences in the leaves, inflorescences or 

 flowers. The leaves of the flowering branches in both Japanese 

 and Chinese specimens are mostly entire, but sometimes three- 

 lobed or, more rarely, five-lobed. The form and degree of 

 development of the petiolar and laminal glands of the leaves seem 

 to be equally variable in both, and the glands may be either 

 prominent or almost obsolete. Of the inflorescence and flowers 

 (as to size) the Kew material is insufficient for satisfactory com- 

 parison; but the Chinese specimens include such as bear rela- 

 tively large strictly terminal inflorescences, and such as bear 

 a number of smaller inflorescences proceeding from the axils of 

 the upper leaves. 



With regard to the size of inflorescences and individual flowers, 

 so much depends on the age of the tree and on local conditions 

 that little importance can be attached to size apart from other 

 characters. 



r 



The flowers from different sources present no differences beyond 

 degree in size, pubescence, etc., certainly nothing obvious in 

 shape or structure. I ^have examined flowers, partly male, 

 partly female, of Oldham 504, Formosa; Morse 542, Lungchow; 

 Oldham 632, Nagasaki; and Balansa 3288, Tonquin. All, 

 except Oldham 632, are referable to A, montana , yi i\s . , though 

 I have something more to explain concerning Balansa' s 



specimen. 



The characteristics of the flowers of all the specimens are : 

 Calyx spathaceous, entire or more or less deeply two- or three- 

 lobed. Petals usually 5, but varying occasionally up to 8, 

 spathulate, longitudinally striate, more or less hairy inside in the 

 middle portion, longer than the stamene and styles. Stamens 10, 

 in 2 series, the inner longer, filaments more or less hairy in the 

 middle portion. Staminodia or lobes of the disk 5; linear in the 

 male flowers and alternating with the outer stamens; in the 

 female flowers linear, with a broad base. Ovary more or less 

 hairy, three-celled; styles deeply bifid. 



Coming to the fruit, so far as it is known, the three proposed 

 species are easily di^]tinguished. A. Fordii has a spheroid fruit 



