NORTON: VARIATION IN TITHYMALOPSIS 457 



D. A plant of T. Ipecacuanhae flowering at the same time as 

 the earhest blooming T. corollata, found at the type station for 

 T. arundelana. 



E. Type of T. marylandica. 



F. Plants from two colonies showing relation to T. zinniiflora 



Small. 



G. Plants from the central U. S. prairie region. 

 H. Plants from dry, open woods. 



J. Plants from northern U. S. 



K. Plants from eastern U. S. not otherwise designated, nearly 

 all from Maryland. 



Seventy-two possible combinations of the variations in presence 

 and absence of appendages, length of stem, pubescence, and leaf 

 form are given in the tabulation. Thirty- three of these were found 

 in the field. But it will be seen that over one half the plants fall 

 into 3 or 4 of the ultimate divisions of the scheme. 



Though there are 6 groups in the T. Ipecacuanhae series, dis- 

 tinguished by the practical absence of stem, appendages or hairs, 

 most of the plants fall into the glabrous form with leaves broadest 

 above the middle. This might then be taken as the dominant or 

 typical form of this species in the region under discussion. 



No plants have yet been found in the next two series, those 

 without appendages and with long or short stems. 



The r. arundelana series, characterized by acaulescent plants 

 with white appendages on the glands, has 8 ultimate groups. 

 There are no very hairy ones and only 2 perfectly smooth plants, 

 though practically all fall into the essentially glabrous group. 

 One of the hairless individuals Is the single specimen of T. mary- 

 landica which I have seen, the plant from Dr. Greene In the New 

 York Botanical Garden herbarium. I do not know of any one 

 else other than Dr. Greene who has ever seen this species growing. 

 It differs further from any of its relatives In Maryland in having 

 unusually narrow linear leaves. A few plants otherwise similar 

 to this group and from the type locality of T. arundelana have 

 short stems and fall Into the next series. There is a possibility 

 that T. arundelana originated as a hybrid between T. Ipecacuanhae 

 and T. corollata, as all three occur together at the type station in 

 some abundance and variety and rarely all blooming at the same 

 time and are also found at every other place where I have dis- 



