892 CONTRIBUTIONS FROM THE NATIONAL HERBARIUM. 
to recognize in it two distinct species, and partly from the fact 
that the fragmentary specimens sent out by Swartz happened to 
include fronds which were not really of the species represented by the 
bulk of his original material. To a certain extent, then, Swartz’s 
original specimens are a mixture, and the case on this account is 
somewhat unusual. But if it shows that, in rare instances, supposed 
portions of ‘‘original” specimens of species so small as this are not 
necessarily authentic, it points out at the same time the greater 
necessity that often exists for studying at first hand the actual 
type of a species. That blind adherence should be given to charac- 
ters offered by a figure or by a fragment of a purported ‘‘type,” if 
it be opposed to a satisfactory original diagnosis bringing out a 
very different set of characters, is scarcely to be thought of; nor on 
the other hand a similar adherence to a diagnosis that is obviously 
faulty, if there be available an illustration that is dependable. As a 
matter of fact it happens not infrequently that an examination of 
the actual type will harmonize errors of both the artist and the 
describer of a species. 
On every account, therefore, it-is of the highest possible necessity 
that actual type specimens shall not only be carefully preserved 
but very plainly indicated as such. Considering the far-reaching 
importance of the subject it is, furthermore, rather astonishing that 
in certain quarters the ‘‘type idea” should be so utterly ignored or 
even deprecated and in others so little understood as to be of no 
especial value to either an author or his readers. 
THE AMERICAN SPECIES OF OLEANDRA. 
In Christensen’s Index Filicum only two species of Oleandra are 
recognized from North and South America. One of these, Oleandra 
nodosa, is a common species which, as explained hereafter, must 
bear the older name 0. articulata. The other species, O. neritformis, 
was described originally from the Philippines; and the American 
plants associated hitherto under this name are not only distinct 
from the Philippine, as might be expected, but represent not less 
than eight readily recognizable species. Of these, two, O. pilosa 
Hook. and 0. trujillensis Karst., from the Guianas and Venezuela, 
were published long ago, and six, from Central America, Panama, 
and Colombia, must now be described as new. There are in addition 
two valid species allied to O. articulata: O. hirta Brack., from Brazil 
and the recently described O. bradei, from Costa Rica, making eleven 
in all from North and South America. 
These fall readily into two groups, as indicated in the key: The 
first, typified by O. articulata (O. nodosa), having the rhizomes 
‘ An additional species, 0. micans Kunze, from Peru, can not be determined from 
the short and wholly inadequate description. It is referred by Christensen to 
0, nodosa—that is, to O. articulata of this paper; but it is probably of the other group. 
