2 BULLETIN 772, U. S. DEPARTMENT OF AGRICULTURE. 
genera should enable the user to identify the genera of all our grasses, 
and the text under each genus should enable him to determine the 
species of our economic grasses. Under each genus are given the 
type? and the synonyms based upon American species, or the names 
that have been used in an American publication. 
The following generic names, of which the types have not been 
found, are based on American material but are unidentifiable from 
the descriptions: 
Anthipsimus Raf., Journ. de Phys. 89: 105, 1819. Based on A. gonopodus 
Raf., “ Dry hills of the Ohio.” 
Dactylogramma Link, Hort. Berol. 2: 248. 1833. Based on D. cinnoides Link, 
described from garden specimens grown from seed from western North 
America. 
Flexularia Raf., Journ. de Phys. 89: 105. 1819. Based on F'. compressa Rat., 
“Kentucky and Ohio.” 
The tribes have been arranged in a new sequence based on the com- 
plexity of the flower structure, the most primitive being placed first 
and the most highly developed last. It is impossible to arrange them 
in a lineal sequence and at the same time represent their relationships, 
as the phylogenetic development has not been along a single line. 
The present arrangement is the closest approximation to natural re- 
lationships that can be shown in sequence. The highest development 
in any given tribe may be far more complex than the most primitive 
example of the tribe placed above it, but the relative development of 
each tribe is believed to be fairly represented by its position in the 
sequence. The bamboos are placed lowest, as certain genera, such as 
Arundinaria, show the least differentiation in the floral structure. 
The Andropogoneae and Tripsaceae are highly specialized, as is 
shown by the great diversity and complexity of the floral structures. 
The tribe Oryzeae of most authors includes two groups of diverse 
genera, each worthy of tribal rank. Theallies of rice (Oryza) are here 
1The type species of a genus is the species or one of the species the author had chiefly 
in mind when describing the genus. Most authors of to-day designate the type species, 
but earlier authors usually did not. To make the application of generic names more cer- 
tain, old genera are now put on a type basis; that is, one of the original species is chosen 
as the type. If, then, a genus as originally established included species belonging in two 
or more genera the name of the genus goes with the type species and the species related 
to it. It will be seen that the type species of a genus as here given often bears a differ- 
ent generic name. (See Miegia, based on Arundinaria macrosperma, p. 22, and Panicu- 
laria, based on Poa aquatica, p. 34.) This means that the genus was based on a species 
previously described under a different genus. In this work the type species is given for all 
generic names, whether valid or synonyms. ‘The reasons for selecting a certain species as 
the type are given in each case. Among several species, the one most familiar to the au- 
thor of the genus may be selected as the type. Species described by Linnzus in his flora 
of Sweden (Flora Suecica) and in his flora of Lapland (Flora Lapponica) were familiar 
to him and are often the types of his genera. (For a further discussion of types, see 
Hitchcock, Generic Types with Special Reference to the Grasses of the United States, Amer. 
Journ. Bot. 5: 248-253, May, 1918; and Report of the Committee on Generic Types of the 
Botanical Society of America, Science, n. ser. 49: 333-336, Apr. 4, 1919.) 
2The form for literature citations throughout this bulletin is in the main that used 
in publications of the United States National Herbarium. 
