HITCHCOCK AND CHASE NORTH AMERICAN PANICUM. 17 



GROUPING OF THE SPECIES. 



The genus Panicum, as here limited, contains two groups sufficiently 

 well marked to warrant their segregation as subgenera, namely, 

 Paurochaetium and Dichanthelium. Nearly all the remaining species 

 group themselves around the central idea of the genus as typified by 

 P. miliaceum. In order to avoid assigning to the main portion of the 

 genus a name of formal nomenclatorial significance, such as Eupan- 

 icum, this group is called true Panicum. There are fiye outlying 

 species which show no close relationships with the others, but are not 

 sufficiently distinct to be assigned generic rank. These are placed in 

 a final group under "Miscellaneous Species." It is probable that a 

 further study of the species in other parts of the world will indicate 

 that at least some of these species may, together with extra North 

 American allies, be placed in definite subgenera. 



The group true Panicum and the subgenus Dichanthelium are 

 susceptible of further division into minor groups, the names of which 

 are the plurals of the characteristic species of each group. These 

 names are not intended to be formal and should have no nomencla- 

 torial standing. The term Angustifolia is used as if we were to say, 

 P. angustifolium and its allies. A few tropical species of true Pani- 

 cum remain ungrouped, as they have no near allies in North America, 

 and there is no advantage in making a group for each species. 



The groups of these two main divisions are arranged to represent 

 our judgment as to their relationship, so far as this can be done in a 

 lineal sequence. In true Panicum the first group, the Geminata, is 

 furthest removed from the typical species, the inflorescence resem- 

 bling that of Paspalum. The Capillaria, Diffusa, and Virgata, typical 

 groups, are near the center of the series. In the same way, the 

 Depauperata are an outlying group of Dichanthelium, the typical 

 groups being the Dichotoma and Lanuginosa. 



The species of each group are also arranged to represent their 

 affinities, but it is impossible to indicate the difference in the degree 

 of relationship. Some of the groups are manifestly more homoge- 

 neous than others. The group Geminata, for example, includes two 

 closely allied species, P. geminatum and P. paludivagum, and one, 

 P. harbinode, in which the affinity is less evident. The latter species 

 is placed in the same group as the other two partly as a matter of 

 convenience. If there were several other species more closely allied 

 to this than to P. geminatum it would have been more convenient to 

 segregate these as a separate group. One more example will suffice 

 to illustrate the unequal value of the groups. Panicum reptans is 

 placed in the group Fasciculata because it is more closely allied to the 

 other species of that group than to those of any other group, yet the 

 41616°— vol 15—10—2 



