70 



SAXIFRAGES. 



The leaves of our plant are 

 much smaller than those of 

 the Himalayan. Since the 

 Chinense plant is proved to 

 be different from the Him- 

 alayan, it is very likely 

 that our plant should also 

 be different from the latter, 



as is conjectured from its 



Fig. 12 A, Pnnsepia scandens HAYATA ; a, a flower ; b, 



different habit of the tree the same, seen from back ; c, a petal ; d, vertical section of a 

 , ., T^ . f ,-, flower; e, another vertical section of the same, ovary and disc 



and the different size ol the ' ' . J 



shaded; i, g, h, pistil, seen from dinerent sides; i, collateral 

 leaves. ovules ; j, one of the same ovules, much more magnified. 



Saxifrageae. 

 Hydrangea LINN. 



Hydrangea Integra HAYATA Fl. Mont. Formos. p. 90, t. 7 ; Ic. PI. 

 Formos. III. p. 107, fig. 17. 



HAB. Mt. Arisan inter Nimandaira et Heishana, leg. B. HAYATA et TAKEO 

 ITO, Mart. 1914. 



The specimens which we collected in the mountains are wanting marginal 

 flowers. Yet in every 

 respect they are iden- 

 tical with H. Integra. 

 It may be ascribable 

 to the season of flow- 

 ering, whether the 

 plant has marginal 

 flowers or not. Calyx 

 tubes are fresh-green, 

 the lobes greenish 



Fig. 13, Loropetalum chinense DC.; a, a leaf; b, a flower; c, 

 wnite, Wnile petals t ne same , seen from side ; d, the same, seen from front ; e, a petal ; 



and stamens are f ' 8 ' h ' stamens seen from different sides ; i, Ovar 7 ; j k > a stellate 



hair, seen from different sides ; a, natural size, others variously 



purely white. magnified. 



