[Vol. » 

 238 ANNALS OF THE MISSOURI BOTANICAL GARDEN 



There are some reasons for suspecting that the original color 

 of the flowers of this genus were purple and that this color gave 

 rise to white. The best evidence to be obtained on this ques- 

 tion, however, comes from an examination of related genera and 

 will be considered in another paragraph. This much may be 

 said from a study of Thely podium alone: there is no reason to 

 doubt such an hypothesis even though there is little to confirm it. 



Inflorescence. — The characters of the inflorescence have been 

 utilized to considerable extent in specific delimitation. As in 

 most Cruciferae the flowers in this genus are borne in a raoeme 

 and are not subtended by bracts or leaves. Two different types 

 are recognized in Thelypodium. The first is characterized in the 

 key and descriptions as "racemose." By this is understood that 

 type exhibited by T. brachycarpum or T. Howellii for example. 

 In these species and some others the flowers when they open are 

 some distance below the apex of the cluster. In T. brachycarpum 

 the raceme is dense and in T. Howellii lax. This distinction ob- 

 tains between other species. The second condition is described 

 as "corymbose." It is illustrated by T. sagittatum and T. flexuo- 

 sum. In those species the flowers when they open are near the 

 apex of the inflorescence and form a somewhat flat-topped cluster. 

 In both the racemose and corymbose types the inflorescence when 

 mature is elongated and truly racemose. In the key and descrip- 

 tions "inflorescence" means the flower cluster— by "mature in- 

 florescence" is meant the arrangement of the pods upon the axis. 



There are also several good taxonomic characters to be obtained 

 from a study of the pedicels. The position is either horizontal or 

 ascending, and there is little variation between individuals of a 

 given category in this respect. Some species are well character- 

 ized by very short pedicels, others have slender ones which may 

 or may not be slightly flattened at the base. 



For several reasons it has been thought that the primitive in- 

 florescence was elongated even before an thesis, that it was prob- 

 ably rather dense, and that the pedicels were short and stout, 

 perhaps angled and probably horizontal. Such a condition may 

 be observed in T. eucosum, T. laciniatum and its variety strep- 

 tanthoidcs, T. brachycarpum, and others. This type seems to 

 predominate in those species that may be considered primitive 

 in other respects. It is also significant that in Stanleya, a closely 

 related genus and one that is more primitive than Thelypodium 

 in many respects, the inflorescence is universally dense and the 

 pedicels are horizontal. 



