58 TRANSACTIONS OF THE [JaN. 9 



are the oldest available for tlie various forms, and Lave suc- 

 ceeded in seeing most of the type specimens which would in- 

 fluence the result. I have adopted the earliest available name 

 in all cases, discarding such as have originally been applied to 

 different species than those with which they have been asso- 

 ciated by some authors, on the jDrinciple that a name once pub- 

 lished for an organism belongs to it and to no other. 



Michaux, who founded the genus * recognized four species, 

 all North American, viz : 1, L. sessilifiora, based on 31edicago 

 Virginica, L.; 2, L. procumhem^ ; 3, L. cajntata, and 4, L. poly- 

 Htachya, based on Hedymram hirtinn, L. 



Persoon, four years later, f accepted all of these excejot L. 

 sesdliflora, which he redescribed as L. reliculata, basing it on 

 Hedymrum reticidalum^'V^iWd.., admitted Hedymrinn violaceum, Ij. 

 into the genus as L. violacea, and added live Asiatic species, 

 together with one of whose halaitat he was uninformed Pursh, 

 in 1814, maintained eight sjDCcies, all of which are accepted 

 in the following pages, although mainly under older names. 

 Nuttall, in 1818, admitted eight, suppressing one of Pursh's 

 and adding L. Stuvei. Torrey and Gray in, 1840, reduced the 

 number to six, recognizing, however, a large number of varie- 

 ties. The genus was monographed by Maximowicz | in 1873, 

 who described 33 species, six of them North American, follow- 

 ing very closely the treatment of Torrey and Gray. Dr. AVat- 

 son's Bibliographical Index of 1878 admits nine species, includ- 

 ing the introduced L. striata, and in the sixth edition of Gray's 

 Manual he recognizes the same number. I am confident that 

 the difficulties found in naming these plants from descrijitions 

 are on account of too few species being admitted. It seems to 

 me that there are twelve distinct species in eastern North 

 America with a possibility of one or two more claiming recogni- 

 tion when more specimens of them are obtained. 



As to the characters which I have mainly relied upon to de- 

 termine species, I have not been able to detect a better 

 wherewith to effect a primary division of our native ones than 

 the short calyx lobes — shorter than the pod — taken with the 

 presence of cleistogamous flowers and nearly always purple or 

 pink corollas for one group, and the loiig calyx-lobes taken with 

 white or ochroleucous corollas (sometime tinged with ]iurple) 

 and absence of cleistogamous flowers for the other. The pe- 

 duncled or sessile clusters of flowers, shape of the leaves, erect 



* Flor. Bor. Am, il. 70. 



t Syn. ii. 318. 



+ Act. Hort. retrop. ii :!27-388. 



