1886. | BOPANICAL GAZETTE. 275, 
tions. These cells vary in shape, some being almost wedge-shaped. 2. The 
next region (II) contains four or five layers of thin-walled cells, which are 
somewhat irregular, sometimes five or six-sided in section, separated by a few 
intercellular spaces, and containing a small amount of coloring substance. 3. 
The third region (III) is decidedly a pigment layer. The cells aresmall, regular 
and thick-walled. 4, The innermost region (IV) contains two layers of nearly 
empty cells. They are thin-walled, with the exception of the wall next the nu- 
cellus (V), and somewhat rectangular. It will be noticed that the “crystal 
layer” found in many of the harder seeds is entirely wanting. —Cuas. U, Srock- 
BARGER, Wabash College, Indiana. 
e notes on Hypericum.—Since the publication of my revision of 
North American Hypericacee in the BorantcaL Gazette for April and May, 
1886, I have received some very icing material from Dr. A. Gattinger, of 
Nashville, Testy ie has for many years been making a careful study of the 
state flora, ms to be a peti for this group, where northern and 
southern forms mingle. Eighteen species of Hypericum are found within its 
borders, and it is not wonderful that in some of its almost inaccessible regions 
@ new species has been discovered. 
ery interesting discovery is that of H. Kalmianum L. in the oakbarrens 
of Tullahoma, Middle Tennessee, July 10, 1882. Heretofore thought to be re- 
stricted to the region of the great lakes, its occurrence in this widely separated 
locality is very unexpected. Asa rule the specimens seem more robust than 
their northern representatives, but not more so than some specimens I collected 
last August at Point Abino, near Buffalo. It would be interesting to learn 
more of the surroundings, but “oakbarrens” give us probably the same condi- 
tions of soil as are found to favor the northern forms. 
In the revision referred to a separation is made between H. Kalmianum 
and the group containing H. prolificum and H. densifloram upon the basis of 
five- and three-celled capsule. Undoubtedly this distinction occasionally breaks 
down, as H. Kalmianum is found with capsules four- to six-celled, and the cap- 
sule of H. densiflorum is often four-celled, while the new species described be- 
low combines characters of both groups, and forms a complete transition from 
H. Kalmianum to the species that follow. While these exceptions show that 
the division is not an absolute one, it still is the rule, and furnishes as good a 
distinguishing character as can be e xpected in species so closely allied. For 
the present, then, the new species, while it is undoubtedly most closely related 
to H. densiflorum and H. prolificum, will be grouped most conveniently with ~ 
H. Kalmianum, on the basis of a five — capsule, as follows: 
of Ek um Gattinger, n. sp.1_ Shrub, five to seven feet high, with 
prolificum : fovert as ae doen: page 
ule tw hree li ong. 
upright branches : leaves oe in tf, 
not foliaceous, | 
tapering to the long strong ee completely fivecelled ond deeply five-lobed, 
in most cases the fi falling away from 
a baer op axis. eae sway lands, in the Orange *sand formation, near Hol-_ 
sti net, 
1Announced to the Botanical Club of the A. A. A.S., Buffalo meeting, 1586. 
