350 



MEMOIRS OF THE NEW YORK BOTANICAL GARDEN 



[Vol. 8, No. 4 



surface proper; the pubescence is parted along the midrib. The blade edges are 

 always appressed-ci-liate, usually with the same type of trichome as the upper 

 leaf surface; aside from the indentations for these marginal trichomes, the mar- 

 gins are entire. Unless otherwise stated in the specific description, the density 

 and quality of the pubescence have been described from the surface hairs and not 

 from those on the veins. In the description of leaves and other organs, the term 

 "acute" has been restricted to angles of less than 90° , "obtuse" to angles of 

 90° -180° , and "truncate" to 180° . 



Inflorescence. The flowers are solitary or in 2-3-flowered dichasia, the di- 

 chasia sometimes being aggregated into panicles or corymbs. The solitary flow- 

 ers arise from pedicels or short branchlets in the upper leaf axils or on short 

 branches. Often these solitary flowers are in opposite upper leaf axils, and thus 

 "paired," but terminal vegetative growth usually leaves no doubt of their place- 

 ment in the solitary-flowered category in the species key. However, the basic and 

 most common inflorescence unit of the genus is the 3*flowered dichasium sub- 

 tended by leaves or variously reduced bracts or bracteoles; the next-lower node 

 below that of the dichasium often has two solitary flowers (occasionally also even 

 the node below this one). The peduncle, as herein used, bears a group of several 

 flowers, the pedicel a single, either solitary or dichasial flower. The peduncle is 

 sometimes differentiated from, and much more slender than, the supporting branch- 

 let, and then is pendent. In those species with a terminal dichasium and a pair of 

 solitary flowers at the node below, the peduncle has been measured from this node 

 to the dichasial node. Usually two opposite bracteoles are inserted on the flower 

 pedicel, rarely, in a dichasium, lacking. Whether these bracteoles are to be inter- 

 preted as of a different order from bracts or leaves is a debatable point; all grada- 

 tions can be found from obvious and persistent leaves, as in some solitary-flowered 

 species, to minute linear caducous bracteoles. Whether the inner pair of large 

 bracts closely investing the hypanthium of some species are modified pedicellar 

 bracteoles or directly modified leaves, with the true pedicellar bracteoles elimi- 

 nated by compression of the flower branchlet apex, is another unsettled question. 

 In those species with well-developed peduncles, the center flower of the perfect 

 (3-flowered) dichasium sometimes aborts; such abortions are obvious in Figure 23» 



Mery. The flowers are 4- or 5-merous, with very rarely a teratologic 6-merous 

 flower on an otherwise normal specimen. The stamens are 2n; the sepals, petals, 

 and carpels, n. Because of the limited number of flowers available for dissection, 

 the predominant mery in each species was determined by sepal number; fusion of 

 two calyx lobes has been observed very rarely, and then such fusions were obvi- 

 ous from the size of the lobe. Generally the change from 5-mery to 4-mery is 

 abrupt^ with the hypanthial ribbing agreeing with the sepal-mery; no obvious inter- 

 mediates were seen in extensive collections under one field number, such as the 

 large series available from several W. H. Camp collections. An analysis of the re- 

 lation of mery (sepal number), order of flowering, and position of the flower in the 

 inflorescence is presented in Figure 23 for two of these large collections, Camp 

 E-4871 (B. fraternum) and Camp E-5161 (B. seorsum). From the data presented for 

 B. seorsum, it can be seen that there is no absolute positional correlation with 

 either mery or order of flowering in this species; for the constantly 4-merous spe- 

 cies, B. fraternum, the center flower of the dichasium was the first-opening in all 

 observable inflorescences. « 



In the delimitation of species, mery has been used to separate closely related 

 populations in several instances. It is possible or even probable that such pairs 

 as B. alpinum and B. lindenii, B. rostratum and B. lutescens , or the 4- and 5" 

 merous elements of B. naudinii will eventually be treated as subspecies. 



