63. RUBIACE.E. 85 



10. Ophiorrhi&a LINN. 



Sir J. D. HOOKER remarks, in his " Fl. Brit. Incl. III. p. 77," that the 

 characters taken from the position of the stamens on the corolla-tube, the 

 hairiness within of the latter, the length and smoothness or hairiness of the 

 style, and the form of the stigma are probably sexual to a very great degree. 

 In the course of my study in the fields, I have found that HOOKER'S opinion 

 is quite correct. When I was collecting in the Arisan ranges, last spring, I 

 found two different forms of flowers on two very similar plants. This diver- 

 gence of flowers may be attributed partly, in some cases, to the sexual, 

 and partly, in the other, to the specifical difference. 



Dichotomous Key to the Species. 



( 1 ) Bracteoles 0, flowers very much smaller, less than 5 mm. in length. (2) 

 Bracteoles conspicuous or small, flowers larger, longer than l- cm. (3) 

 ( 2 ) Plants longer, nearly 50 cm. long, leaves larger 12 cm. long. 



0. parviflora and liukiuensis. 



Plants smaller at most 20 cm. in height, leaves smaller 5 cm. long. 



. . . 0. injlata and pumila. 



( 3 ) Stigma-lobes linear. (4). 



Stigma-lobes rounded. (6). 

 ( 4 ) Leaves oblong, lanceolate, usually abruptly acuminate towards the apex, 



attenuate at the base. (5). 

 Leaves ovate, obtuse or acute at the apex, rounded or abruptly acute 



at the base 0. japonica form, longistigma. 



( 5 ) Leaves pale glaucous beneath in dried specimens. . . 0. dimorphantha 



form, longistigma. 



Leaves dark greenish above, reddish purple beneath, in dried specimens. 



.... 0. monticola, form, longistigma. 

 ( 6 ) Leaves oblong or lanceolate. (7). 



Leaves linear-lanceolate 0. stenopJitjlln. 



( 7 ) Leaves quite glabrous or very thinly pubescent. (8). 



