114 CONTRIBUTIONS FROM THE NATIONAL HERBARIUM. 
there may be several specimens from which the type must be selected 
by comparison with the description (Panicum barbulatum Michx., 
page 148). These and other difficulties complicate the study and 
make it necessary to examine carefully all the evidence. This evi- 
dence not infrequently shows that a species has been misunderstood. 
The original description may be insuflicient to identify the species, 
but the identity can be established by the type specimen (Panicum 
nitidum Lam., page 148). Tradition may have attached a name to 
one species, while the description and the type specimen show that 
the name belongs to another species (Cenchrus tribuloides L., page 127; 
Agrostis aspera Michx., page 150). 
In the following account I have considered each case upon its 
merits and have presented the evidence upon which I have based 
my decision. It will be seen that usually the apparent difficulties 
disappear and we are able to determine the specimen the author 
had chiefly in mind when he wrote the description. The earlier 
authors, especially Linneus, frequently cited descriptions or plates 
which they considered as referring to the same plant they were describ- 
ing. Linneus even based his binomial upon the description or plate 
of another author. If an author quotes the diagnosis of a species 
described by another author and gives a name to this, but has no 
description of his own, the type of the older author becomes the 
type of the later (Panicum capillare, L., page 118). Linneus often 
gave binomial names to species described by others. But if Linnzus 
wrote a description and there has been preserved a specimen which 
the evidence shows must have been seen by him when he drew up 
the description, this specimen is the type, and not the specimen 
which is the basis of the synonym (Panicum latifolium L., page 118; 
Paspalum paniculatum L., page 116). The danger of placing too much 
weight upon cited synonyms as evidence is shown by the fact that 
Linneus sometimes cited a given Sloane plate under different species 
in different works or even in the same work (Panicum sanguinale 
L., page 117); or the synonyms may be quite different from the species 
under which they are cited (Andropogon nutans L., page 125). 
Fortunately the grasses left us by the older authors, though often 
fragmentary, are in a satisfactory state of preservation, and it is 
usually possible to determine their identity with certainty. 
THE AMERICAN GRASSES DESCRIBED BY LINNZAUS. 
The herbarium of Linneus, preserved at the rooms of the Linnean 
Society of London, Burlington House, Piccadilly, contains most of 
his types. In the following article I have considered only those 
species based wholly or in part upon American material, nearly all 
of which was furnished by Kalm, Gronovius, Sloane, or Browne. In 
the case of Old World species the specimens preserved by Linneus 
