92 TERNSTKCEMIACE.E. 



long, filaments connate at the base. Ovary minutely rounded, 1 mm. long. 

 Styles 3, entirely connate, 3 mm. long. 



HAB. Wantang. 



Near T. Sasanqua, but differs from it in having rounded petals. The 

 petals of the present plant are usually oblong and rounded at the apex, 

 while there of T. Sasanqua are generally, if not always, notched at the apex. 



Thea biflora HAYATA Materials for a Flora of Formosa p. 44. Branches 

 fusco-ciuerasceut, or rubescent towards the apex. Leaves alternate remotely 

 arranged, oblong, 5 cm. long, 2 J- cm. broad, acute at the apex, retuse at the 

 extremity, obtusely rounded at the base, margin serrulate from middle to 

 apex, nearly entire downwards, pallido-flavescnt on both surfaces when dried, 

 costa, veins and veinlets elevated above, veins and veiulets inconspicuous be, 

 low. Costas and petioles shortly hirsute above, glabrous beneath, petioles 

 4 mm. long. Flowers in pairs at the apex of the branches, sessile. 

 Flower-buds oblong, sericeo-tomentose. Sepals 5-8, very iuequal, very much 

 imbricate, deciduous, 4-seriate, (flower-axis 5 mm. long,) broadly orbicular, 

 densely long sericeo-tomentose on the back at the centre and apex, coriaceous 

 at the centre, thinner on the margin, glabrous inside, the outermost the smal- 

 lest, the inner the larger, 15 mm. long. 2 mm. broad. Petals 5, inequal- 

 interior ones larger, obovately rounded, truucately rounded at the apex, 2 cm. 

 long, 1^ cm. broad, hirsute outside at the base, glabrous otherwise, inner ones 

 narrower. Stamens numerous half as long as petals, filaments connate to the 

 middway, anthers ovate, nearly 2 mm. long, 1^ mm. broad, emarginate on botli 

 ends, connectives more or less dilated. Ovary ovoid densely long toinentose, 

 hairs 2 mm. long, erect, patent, styles 4, connate to the middway or entirely 

 connate, erect, stigma on the outside at the apex of the style, obliquely 

 truncate. 



HAB. Kagi, Kodensho, by. T. KAWAKAMI and U. MORI. Oct. 1906, (No. 

 1758). 



There is nothing like this at Kew. I think it may be a species not yet 

 described. 



Thea shinkcensis HAYATA Materials for a Flora of Formosa p. 45. 



