110 



CONTRIBUTIONS FROM THE NATIONAL HERBARIUM. 



say, very indistinctly connected at the base — and, if we adopt Nut- 

 tail's definition, "each terminated by two central filaments"; 

 lastly, the stigma is obtusely conical and more or less distinctly 

 bilobate. In Roulinia, the corolla is rotate, with reflexed lobes, and 

 the scales of the crown are distinctly connate at the base, ending with 

 a long, single appendage;'' the stigma is depressed, with a scarcely 

 conspicuous apex. 



Now, the features that constitute the special characters of Enslenia 

 are met with so far only in Enslema albula Nutt. and E. volubilis 

 Karst., some details of the floral structure of which are reproduced 

 here after Karsten (fig. 17). As to E. ligulata Benth., the description 

 agrees decidedly better with that of a Roulinia. It is true that the 

 corolla is said to be campanulate, but this term can easily be applied 

 to a rotate corolla in its first opening stage, just as the lobes of a 

 campanulate corolla may be reflexed so as to give it a rotate appear- 

 ance in a late stage of development. On the other hand, the scales 



W 



Fig. 17.— Klower parts of Enslenia albida and /•:. volubilis. < )f former, n, segment of corona; 6, translatorlum 

 Of latter, c, segment of corona; d, translulorium. a, c, Scale. IS; ft, d, scale (i4. 



of the corona are described as "briefly connate at the base," and 

 divided into 3 lobules, the lateral ones very short and obtuse, the 

 middle one (single) protracted in a long, lanceolate tip; the stigma is 

 pulvinate (but not conical nor bifid), umbilicate, and obscurely pen- 

 tagonal. 



Bentham's description applies quite well to specimen no. 11021 of 

 Pringle, collected at Yautepec in the State of Morelos, Central 

 Mexico, the only noticeable difference being that the elongate rachis 

 of the raceme often bears more than 3 flowers (3 to (>) at a time, and 

 that the dimensions of the several parts of these are much less. But 

 as to these dimensions there is evidently a mistake, for a corolla 7 



"The Genera reads (loc. cit.): "Corona squamae 5, basi tubi staminci affixae, 

 ereetae, latae, membranaceae, apice subtruncatae el medio in ligulam integrant 

 2-fidam v. duplicem productae." The interpolation of the word "integram" is 

 evidently intended to insure the place of Enslenia? ligulata Hentham, but is also an 

 unjustified swerving from one of Nuttall's best characteristics of his genus. 



f>The original description of Decaisne {loc. cit.): "Foliolis ovato-rotundis crassi- 

 usculis in acumen erectum arcuatum v. inflexum v. bifidum desinentibus" is not 

 supported, as to the italicized word, by his further diagnoses of 11 species. 



