4 Contributions from the Gray Herbarium 
small yellowish or greenish flowers. The sepals and only slightly 
larger wings are subherbaceous and persistent (not deciduous as 
described by Chodat), the keel beakless and crestless, and the cap- 
sule (sometimes one-celled by abortion) is wingless with thick sub- 
coriaceous walls and only tardily dehiscent. The typical species, 
P. Penaea L., was with a few other species made into a genus by 
De Candolle, whish is retained by Bentham & Hooker, Urban, and 
by Britton, who has twice revised the group. In the lack of any 
really distinctive technical characters, however, it seems much 
better retained in Polygala. By Chodat it was sunk to subsectional 
rank under his section Hebecarpa, but it certainly merits equal 
recognition with his other primary divisions of Polygala. 
3. Hebecarpa. Of this very difficult group, which also occurs in 
the southwestern United States and South America, some sixty 
speces are here recognized. They are always perennial herbs or 
undershrubs, with herbaceous stems and usually fruticulose base, 
the flowers usually purple and medium-sized for the genus. The 
wings, except in one species and to a less degree in one or two others, 
are deciduous like the (very rarely persistent) sepals; the keel is 
neither beaked nor crested; and the fruit is usually narrowly mar- 
gined. The variations in itbiGng are here extreme and much use 
has been made of them in specific discrimination. Seeds of all spe- 
cies accessible to me at the present time have been figured, and it 
is hoped that at a later date the series can be made complete, since 
a correct idea of the seed-appendages can be given in no other way. 
Although in a number of cases I have described the flowers as ap- 
parently ochroleucous in dried material, it is not improbable that 
they are always of some shade of purple. 
The three sections into which the subgenus is here divided are 
based chiefly on the persistence or deciduousness of the sepals and 
wings, as indicated in the key on p. 18. The two larger subsections 
I have adopted under the main section (Euhebecarpa) are admit- 
tedly unsatisfactory as at present constituted, being founded 
chiefly on leaf-size. Had mature seeds of every species been avail- 
able, a more natural classification could undoubtedly have been 
made. In the typical species of subsect. Hebantha the corneous 
umbo of the aril is minute or almost wanting and the scarious 
border broad and variously disposed; in subsect. Microthrix the 
umbo is large and the scarious margin usually small or medium. 
