300 BOTANICAL GAZETTE [NOVEMBER 
separated Hippurus from the Haloragidaceae on the basis of its 
morphological characters, and made a new family entitled Hippuri- 
daceae. Further rearrangement no doubt will follow later, when 
more morphological work has been done on the group. 
Viewing the Umbelliferae from a morphological standpoint, 
it is clear that separate petals are the only character which they 
have in common with the Archichlamydeae, and they even lack 
sympetaly, if the Umbellales are taken as a whole on the basis of 
VAN TIEGHEM’S work on the Pittosporaceae; for, according to 
BESSEY (5), ‘one organ.may be advancing while another is retro- 
grading.” 
“The complete cyclic arrangement of floral members associated 
with definite numbers” (12), the single integument (fig. 17), the 
anatropous ovule (fig. 14), the absence of parietal tissue of the 
megasporangium (fig. 18), the small nucellus (fig. 18), and the 
complete tetrad of the megaspores (fig. 19) of the Umbelliferae, 
all of which are general characters of Sympetalae in contrast with 
those of the Archichlamydeae, from which the Umbelliferae stand 
so stiffly apart, prove that the Umbellales in reality belong among 
the Sympetalae. 
The question now arises, if they are to be placed among the 
Sympetalae, what is their relative position? This is not difficult to 
answer, for the epigynous nature of their flower places them surely 
above the Tubiflorales, and their other floral characters put them 
below the Campanulales. Hence, a position in the neighborhood 
of the Rubiales is without question. C. E. Brssry (5), E. A. 
BEssEY (6), and WETTSTEIN (84) consider the Umbellales as giving 
rise to the Rubiales. WerrnuaAm (82) notes that “within both 
cohorts,”’ as he calls them, “the progress from polycarpellary to 4 
bicarpellary gynoecium is observable; in both the ovary is only 
very rarely unilocular; and in both the androecium is primitively 
isomerous with the corolla, and the latter primitively regular.” 
Accordingly, he recognizes the ‘‘ Umbelliflorae as the representatives 
of a side branch from the calycifloral (rosalium) plexus, and the — 
Rubiales as another such side branch of this stock.” All this is 
to fit in with his scheme of a polyphyletic origin of the Sympetalae. 
