446 CONTRIBUTIONS FROM THE NATIONAL HERBARIUM, 
In September, 1916, the writer collected abundant material of the 
species of Berberis occuring at the Cascades or Great Rapids of the 
Columbia River. The original spot, where Lewis almost certainly 
collected his specimens, is now occupied by the Cascade Locks, but 
on the bench above, lying about 30 meters higher and one-fourth mile 
distant, Berberis is abundant. In this place two distinct species of 
Berberis are found: B. nervosa, abundant in underbrush; B. agui- 
foltum, with shiny leaves, abundant in shady places, and to a less 
extent in the open; and in addition in very open places and much 
scarcer occurs a somewhat dull or glaucous-leafed plant, which 
closely simulates B, repens Lindl. However, there are very numerous 
intermediates between the shiny-leafed species and the glaucous- 
leafed plant. The latter was at first taken to be B. repens Lindl., 
but microscopic characters hereafter discussed, as well as the form 
of the leaflets, point to its being a dull-leafed form of B. aquifolium 
Pursh. Specimens of 2. aquifolium and B. repens are usually at 
once distinguishable by the tall habit of the former and the low 
habit of the latter. No great stress can be put upon the character 
of creeping branches, which Lewis had mentioned in his notes on 
B. nervosa and B. aquifolium, as they occur at the mouth of the 
Columbia. This character was mentioned by Pursh in his descrip- 
tion of B. aquifolium, in which species it often occurs, as was espe- 
clally observed at the type locality; but creeping branches are far 
more developed in B. repens. Very careful comparisons of the de- 
tails of the flowers and seeds failed to disclose any characters that 
could be regarded as crucial. The leaflet characters of form and 
dentition suffice to separate the two species in the great majority of 
cases, but occasional specimens occur in which these characters are 
insufficient. The best distinguishing characters are those of the leaf 
surface. The leaflets of B. aquifolium are nearly always shiny above, 
but occasionally dull, and beneath pale green, but never glaucous ; 
while those of B. repens are nearly always dull on the upper surface 
and glaucous beneath. In ambiguous specimens the under surface 
of the leaflets when examined under a binocular supplies a critical 
difference to separate the two species, as first clearly indicated by 
Dr. Otto Stapf. 
In reference to the characters exhibited by the under surface of 
the leaves, Dr. Stapf writes as follows, October 30, 1919: 
“I have compared the anatomical structure of the lower epidermis of the 
leaves of B. Aquifolium and B. repens, and the other North American Ma- 
honias immediately allied to them, und huve come to the conclusion that the 
characters of the presence or absence of papillae is indeed a very great help 
in discriminating otherwise doubtful specimens. If they hybridize we might 
of course expect intermediate forms, but among the Kew material I have 
found none.” 
