158 



Harper: Explorations in Georgia 1903 



noted, and Bartram {Travels, page 424) mentions a luxuriant 

 growth of C. indica along the Amite River in Louisiana in the 

 1 8th century, "*" at which time one would hardly suppose it could 

 have been an escape. Pursh also mentions C. angustifolia L. as 

 having been collected in Georgia by Enslen. 



Canna flaccida (or horticultural varieties of it) is often seen in 

 cultivation, but it is not certain how and when it first reached the 

 European gardens from which it has since been distributed. Its 

 occurrence in China early in the 19th century or before is a mys- 

 tery, unless perhaps it is indigenous there after all, like some other 

 plants of the Eastern United States. Another puzzling thing is 

 that most European authors have described the leaves of this 

 species as glaucous, but this does not seem to be true in the wild 

 state. 



- + 



In my 1903 collection Canna flaccida is represented by 710, 

 18 ig, collected on June 13 in Chatham County about nine miles 



west of Savannah. At this 



point it grows abundantly (and 

 often six feet tall) in low woods 

 and even spreads over an ad- 

 jacent railroad embankment. 

 (There can be no question 

 that it is indigenous there, 

 though.) I saw it from a train 

 at the same place two years 

 before, and in August, 1902, 

 I noted it in Charlton and 

 Camden counties, as well as in 

 the adjoining Nassau County, 

 Florida, where it grows in 

 swamps from which the La- 

 fayette formation seems to be 

 absent, as shown by the oc- 

 currence of Taxodiiim dis- 

 Figures Nearer view of a single sped. ^^v///.;;. and certain Other sig- 



men oi Canna Jiaccida [no. /8/g), 



nificant plants. The photo- 



graphs reproduced here are perhaps the first ever taken of this 

 species in a wild state. 



* Canna e!ega?is Raf. (Fl. Lud. 143. 1817) is based on Bartram's description 



