i)IDYNAMIA ANGIOSPERMIA. Ill 



Root perennialj creeping. Stem about a foot high^ obtus'ely 4-angled, 

 occasionally branched- Leaves all obtuse, with a margin slightly undu- 

 late, the lower ones nearly round. Cali/x with a short tube, the segments 

 subulate, almost setaceous, hispid; in the former species the calyx is gene- 

 rally divided to the base. Border of the Corolla equally 5-cIeft, slightly 

 emarginate, pale blue or purple, spotted with a dusky yellow. Stamens 

 shorter than the corolla. Anthers sagittate- Germ surrounded at basd 

 with an orange coloured glandular ring. Style a litde longer than the sta- 

 mens. Stigmas simple- Seeds few in each cell of the capsule. 



■ 



4 



The R. Biflora of Linnaeus probably belongs to this species, I have 

 omitted the name as evidently incorrect; the habit oi the plant is to pro- 

 duce in the first instance one flower in each axil, if it grows luxuriantly 

 two lateral opposite flowers are next produced, so that the axils are 1 or S 

 flowered and may increase afterwards regularly by pairs- It may occa- 

 sionally happen that one of the lateral buds will prove abortive,* or one 

 may shoot up and expand before the other, in either of these cases a biflo* 

 reus specimen may be collected, but this is accidental and liot the habit of 

 the genus. 



Grows in sandy pine barrens. 



Flowers from May to the close of the summei?. 



5* HuMISTRATA. MicH* 



F 



R. glabriiisciila, dif- 



fusa, radicans ; foliis 

 m petiolum longiiis 

 cule ani^ustatis, ovali 



Glabrous, dilTuse, 

 radicant ; leaves at- 

 tenuated at base into 

 a long petiole,, oval. 



bus, obtusis ; floribus I obtuse ; flowers near- 

 subsessilibus ; capsu- j ly sessile j capsule liu 

 lis linearibus. | ear* 



Mich. 2. p. 23. Pursh 2. p. 421. 



Found by Michaux in the Southern parts of Georgia. I have founil no 

 spwies exactly agreeing with the description. 

 Flowers probably through tlie whole sunimer. 



The plants described under this head will undoubtedly belong to Ruel- 

 "a, however the genus may be limited. In fact they agree so much 

 among tlieraselves, that it is difficult to find specific distinctions. But be- 

 tween the almost campanulate flower of the Ruellia and the bilabiate some- 

 wnat nngent, corolla of thci Justicia, at least as the species are presentcil 

 !?. "^ '" ^'^'^ country, the diflerence is so great that nothing bat the capsii^; 



See Smith's observation's on RL'MI.- 



appears to connect the two genera. 



^■Vjn Rees' CyclopaOia. 



