65 



species of Machaeranthera. They were Aster Patter sonii Gray 

 2Lnd Aster coloradensis GvdLy. A. coloradensis seems to be related 

 to A. Pattersonii and if that is placed in Machaeranthera I 

 think A. coloradensis should also be placed there. 



Of the species with shorter green tips of the bracts, M. 

 spectabilis Greene and M. Selbyi are very similar; the characters 

 which distinguish them are slight. M. viscosula Rydb. is 

 characterized by very narrow leaves. 



II 



The species under my second division are equally difficult 

 of discrimination. M. canescens and M. superba seem distinct 

 enough; but M. pulveriilenta, M. divaricata, M. subalpina, M. 

 ramosa, and ilf. g/a6g//a are more difficult of separation. That 

 they are related to M. canescens seems evident, and that they 

 are quite closely related to each other also seems evident. 

 Whether there are characters by which they may be separated 

 into distinct species is rather doubtful. In the Fl. of N. Am. by 

 Torrey and Gray the genus name for these plants is Dieteria, 

 and under Dieteria divaricata it is observed that "these species 

 are so nearly related that they may hereafter be found to pass 

 into each other." 



Windsor, Col. 



THE FRANKENIACEAE AS A LINK IN THE 

 CLASSIFICATION OF DICOTYLEDONS 



Alfred Gundersen 



An interesting difference between the system of Bentham and 

 Hooker and that of Engler appears in the position of the Pink 

 Family, the Caryophyllaceae. In the former system this family 

 is classified with Frankeniaceae and Tamaricaceae, in the latter 

 under Centrospermae with Chenopodiaceae, Aizoaceae and 

 others. The many detailed characters in common between the 

 families Caryophyllaceae and Frankeniaceae on the one hand, 

 and between the Caryophyllaceae and Chenopodiaceae, etc. on 

 the other, suggest that both the above interpretations may be 

 correct. But such a double connection would require a re- 



