among Onagraceous plants to the indefinite stamens of 

 Myrtacese. The wliole genus, indeed, shews this to a cer- 

 tain extent, but in no case so distinctly as in that before 

 us. In the genus Eucharidium there are but four stamens ; 

 in Godetia there are eight ; in Clarkia pulchella there are 

 12, of which the first whorl is perfect, the second rudimen- _ 

 tary, the third in the form of narrow glandular scales at the 

 base of the perfect stamens ; in Clarkia elegans there are 16, 

 the first whorl of which is perfect, the second striate but not 

 rudimentary, the third in the form of small round hairy 

 scales at the base of the imperfect stamens, and the fourth in 

 the state of smaller scales at the base of the perfect stamens; 

 and in Clarkia rhoniboidea the two whorls of scales are 

 equally perfect, while the stamens are in the same state as in 

 C. elegans. 



This is indicated in the plate, where fig. I. is a section 

 of the tube of the calyx of C. elegans, fig. 2. of C rhom- 

 boidea, fig. 3. of C. pulchella. 



Nearly allied to C. rhomboidea is an unpublished species 

 in Douglas's Californian herbarium, the character and name 

 of which may be briefly stated thus — 



C. unguiculata ; foliis oblongis sessilibus dentatis, ovariis calycibusque villosis, 

 petalis unguiculatis limbo subsagittato rotundato ungue dup!6 breviore. 



