OF ARTS AND SCIENCES. 213 
KY. 
CONTRIBUTIONS TO AMERICAN BOTANY. 
By SeERENO WATSON. 
Presented May 14, 1879. 
‘ I. Revision of the North American Liliacee. 
Tue order Liliacee, as outlined by Dr. Gray in the last edition of his 
Manual (1867), and as now generally understood by botanists, pre- 
sents such a diversity in its characters and their combinations that it is 
by no means easy to satisfactorily group the genera according to their 
affinities, or to arrange them in any seemingly natural sequence or 
sequences. The difficulty is not much diminished, but rather increased, 
when the question is confined to the genera of a limited geographical 
area; hence, in the following attempt at a classification of the fifty 
genera that are found in North America, their relations to the rest of 
the order have been in some measure taken into account. 
If the character of a baccate as distinguished from a capsular fruit 
be considered a subordinate one (as seems to be necessary), a division 
of the genera may be made into three series or suborders, which, 
. notwithstanding exceptions, are on the whole pretty clearly defined. 
The first and largest of these is prominently distinguished by its 
scarious floral bracts, persistent nerved perianth, perigynous stamens 
with introrse anthers, an undivided persistent style, and a loculicidal 
fruit (if capsular). This includes much the larger portion of the 
genera which have usually been considered as belonging to the capsu- 
lar Liliacee, as well as most of the Asparaginee. Both of the other 
divisions have the stamens hypogynous or nearly so, with more or 
less extrorse anthers, and the floral bracts are more or less foliaceous 
or are wanting. Both also always have distinct perianth-segments 
and unjointed pedicels. But one has a nerveless deciduous perianth, 
the styles (when present) more or less united, and the fruit a loculi- 
cidal capsule or a berry; to this belong the Liliee or Tulipee proper, 
the Uvulariee and the Trillieg, as here defined, and a few other 
