1898] THE GROUPS OF ANGIOSPERMS 341 
] believe it a misconception to regard the Scitaminee as 
direct descendants of the Liliacee or of the: Amaryllidacee. 
- Though the Scitaminee do resemble Dracena and: Cordyline, 
stronger:characters argue against deriving them from the Lilii- 
fore, viz, the. formation of endosperm and perisperm, the 
smooth exine and thick intine of the pollen grains, the presence 
of simple:and coraplex starch grains in the albumen.. 
The Microspermz comprise the Burmanniacee and Orchida- 
tet, whose relationship to each other is based solely upon the 
tumerous small ovules and the parietal placentation. The Bur- 
manniacee are like the Orchidacez in that they attain a degree 
of zygomorphism which is found in few monocotyledons, and a 
careful review of characters shows no possible connecting link 
| between Orchidacee and the Liliiflore. The Microsperme may 
fal, therefore, into the. two subseries, Burmanniinee and 
ynandre, 
DICOTYLEDONS. 
The dicotyledons as here presented show many departures 
Eichler’s system. One of the most essential changes is 
ie breaking up of the series of Amentacee and. the interpola- 
: ~ ~ various places of the miscellaneous families designated 
| Eichler as Hysterophyta. That the Amentacee of Eichler 
: hot represent one circle of relationship has become apparent 
: ea investigation of the ovules and the processes of fer- 
' a tion. That the establishment of the Hysterophyta was 
~ Provisional is evident from their long-known variability as 
€s; the Passiflorine are brought under the 
tire} the series. Terebinthine and Tricocce have been 
‘he Ger broken up and their separate families referred partly to 
aes, partly to the Sapindales or to the Esculine, the 
