4 







The pedunculate species of Trillium 



Henry Allan Gleason 



About forty specific names and a correspondingly large num- 

 ber of varietal ones have been proposed for the pedunculate species 

 of Trillium during the century and a half that has elapsed since 

 the publication of the Species Plantarum. Linnaeus himself left 

 the genus in some confusion by including two species under the 

 name T. cernuum. Others added to it by imperfect descriptions 

 and incorrect citations, and Rafinesque complicated matters still 

 further by the publication of a large number of species the identity 

 of which will probably never be satisfactorily established. Gray, 

 Watson, and Harbison have all contributed in bringing order into 

 the genus, and more recently Rendle has published such careful 

 descriptions of the types of the Linnaean species that they are no 

 longer in doubt. 



The species fall into three very natural groups separated by the 

 character of the stigmas and ovary. In the first of these, which 

 may be designated the T. erectum type, the ovary is deeply and 

 sharply six-lobed, and the members of each pair of lobes, corre- 

 sponding with one of the three carpels, unite above into a short 

 thick sessile recurved stigma,which tapers gradually to the tip. 

 In the second, the T. grandiftorum type, the ovary is less deeply 

 angled or even merely three-lobed ; it is truncate, rounded, or 

 somewhat narrowed at the end ; and the stigmas are slender and 

 of virtually uniform diameter throughout. T. rivale S. Watson 

 is included in this group because of Watson's statement that it is 

 related to T % nivale. The material available for examination was 

 too fragmentary to determine accurately the character of the stigma. 

 In the last, or T. Catesbaei type, the stigmas are likewise slender 

 and of uniform diameter, but they are united below into a definite 

 style. These distinctions are not only constant, but they are very 

 striking, and characters of no greater importance have been used 

 in other families for the segregation of genera. 



Within the three groups thus separated by the structure of the 



387 



