BRIEFER ARTICLES 



NOTES ON CLEISTOGAMY OF GRASSES 



(with five figures) 



The genus Triplasis Beauv., so far as we can find, has not been recog- 

 nized as one of the many genera of grasses producing cleistogamous spike- 

 lets. Its nearest relative recognized as having this habit is SieglingM 

 decumhens (L.) Ktze. This and other cleistogamous grasses are dis- 

 cussed by Professor Eduard Hackel (Oest. Bot. Zeits. 56:81-88, 143-154. 

 180-186. 1906). Specimens of Tripla'sis in the National Herbarium col- 

 lected in autumn show reduced panicles whoUy or partly included in the 

 sheaths, and bearing few to several cleistogamous spikelets with glumes 

 and awns much reduced, but othenvise like those of the terminal panicle. 



Specimens of T. purpurea (Walt.) Chapm. coUected on the low dunes 

 of the Isle of Palms, S. C, October 18, 1907 {CMse 4524), show an addi- 

 tional form of cleistogamous spikelet. This is larger, solitary, and sessJe 

 in the base of an indurated prophyllum, in the wings of which it is clasped 

 together with the first intemode of the branch, lying between the branch 

 and the back of the prophyilum. The 



sections faQing to show even traces of them. The spikelets are one-flowered, 



.wn 



^^ are 



h reduced. (The 

 from the terminal 



ngures snow tiic u-iiicicuvt u»-i.**»_wi^ v^^w w^ ^ 



panicle, also the comparative size of the two developed grams.) Ihis 

 second form of cleistogene was discovered in the field, and plenty of matenal 

 for dissection was coUected. Most of the culms produced two to eight or 

 nine of these spikelets, beginning usuaUy at the second node. The spite- 

 lets in the upper sheaths have smaller florets and rddimentarj' glumes 

 and -r 'e into the ordinary form of cleistogene. Since herbarium matenai 

 cannot be ruthlessly disjointed, only a few sheaths of autumnal specunens, 

 which by their swoHen appearance gave promise of contammg these cleis- 

 togenes, were examined. Of these, one specimen of T._ amenmmBe^nr. 

 from Florida {Comhs 871) showed this form, the glumes wanting, the am-n, 

 which in this species is 5"^ or more long, nearly obsolete. A dupkcate 



[^taffllcal Gazette, toI. 45 



